6810 1 2 3 4 5 6 7 8 9 NATIONAL FEDERAL MILK MARKETING ORDER 10 PRICING FORMULA HEARING 11 12 DOCKET NO.: 23-J-0067; AMS-DA-23-0031 13 14 Before the Honorable Jill Clifton, Judge 15 16 ---o0o--- 17 18 Carmel, Indiana 19 October 4, 2023 20 21 ---o0o--- 22 23 24 25 26 Reported by: 27 MYRA A. PISH, RPR, C.S.R. Certificate No. 11613 28 6811 1 A P P E A R A N C E S: 2 FOR THE USDA ORDER FORMULATION AND ENFORCEMENT DIVISION, USDA-AMS DAIRY PROGRAM: 3 Erin Taylor 4 Todd Wilson Brian Hill 5 Michelle McMurtray 6 FOR THE AMERICAN FARM BUREAU FEDERATION: 7 Roger Cryan 8 FOR THE MILK INNOVATION GROUP: 9 Ashley Vulin (Remotely) Charles "Chip" English 10 Grace Bulger 11 FOR THE NATIONAL MILK PRODUCERS FEDERATION: 12 Nicole Hancock Brad Prowant 13 FOR SELECT MILK PRODUCERS, INC.: 14 Ryan Miltner 15 FOR EDGE DAIRY COOPERATIVES: 16 Dr. Marin Bozic 17 FOR INTERNATIONAL DAIRY FOODS ASSOCIATION: 18 Steve Rosenbaum 19 20 ---o0o--- 21 22 23 (Please note: Appearances for all parties are subject to 24 change daily, and may not be reported or listed on 25 subsequent days' transcripts.) 26 27 ---o0o--- 28 6812 1 M A S T E R I N D E X 2 SESSIONS 3 WEDNESDAY, OCTOBER 4, 2023 PAGE 4 MORNING SESSION 6815 AFTERNOON SESSION 6931 5 6 7 ---o0o--- 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 6813 1 M A S T E R I N D E X 2 WITNESSES IN CHRONOLOGICAL ORDER 3 WITNESSES: PAGE 4 Dr. Marin Bozic: 5 Clarification Testimony 6815 6 Dr. Peter Vitaliano: 7 Direct Examination by Ms. Hancock 6820 Cross-Examination by Mr. English 6842 8 Dr. Charles Nicholson: 9 Direct Examination by Ms. Hancock 6916 10 Cross-Examination by Mr. English 6961 Cross-Examination by Mr. Miltner 7000 11 Cross-Examination by Mr. English 7009 Cross-Examination by Mr. Miltner 7011 12 Cross-Examination by Mr. Rosenbaum 7012 Cross-Examination by Ms. Taylor 7021 13 Cross-Examination by Mr. Wilson 7048 Cross-Examination by Mr. English 7052 14 15 ---o0o--- 16 17 18 19 20 21 22 23 24 25 26 27 28 6814 1 M A S T E R I N D E X 2 INDEX OF EXHIBITS 3 IN CHRONOLOGICAL ORDER: 4 NO. DESCRIPTION I.D. EVD. 5 297 Edge-15 Corrected 6815 6818 6 298 Edge-15B Corrected 6815 6818 7 299 Testimony of 6820 Dr. Peter Vitaliano 8 300 MIG-28 6883 9 301 MIG-29 6884 10 302 Testimony of 6917 7053 11 Dr. Charles Nicholson 12 303 Summary of Testimony 6917 7053 13 14 ---o0o--- 15 16 17 18 19 20 21 22 23 24 25 26 27 28 6815 1 WEDNESDAY, OCTOBER 4, 2023 - - MORNING SESSION 2 THE COURT: Let's go back on record. 3 We're back on record. It's 8:03 in the morning on 4 October 4, 2023. It's a Wednesday. 5 Dr. Bozic, would you identify yourself, please? 6 DR. BOZIC: Dr. Marin Bozic for Edge Dairy Farmer 7 Cooperative. 8 THE COURT: And you remain sworn. 9 MARIN BOZIC, 10 Having been previously sworn, was examined 11 and testified as follows: 12 THE COURT: And we have been anticipating that you 13 would clear up some questions that we had left over. How 14 would you like to proceed? I have got two new exhibit 15 numbers to give you. I have got, for example, 297, which 16 could be for your Edge-15 corrected, and 298 for your 17 Edge-15B corrected, if that's what you want. 18 THE WITNESS: Yes. Thank you, Your Honor. 19 THE COURT: Very good. 20 You may proceed. 21 (Exhibit Numbers 297 and 298 were marked for 22 identification.) 23 THE WITNESS: This should take only four minutes. 24 If we can have the slides on the screen. 25 Mr. Wilson asked me yesterday morning how come 26 that my baseline PPD was much higher than was published in 27 the Journal of Dairy Science. Upon the urging of Your 28 Honor, I did some forensics, and turns out that because I 6816 1 wanted to use the version of the file that hasn't been 2 modified or further automated since the Journal of Dairy 3 Science article was published, I had on this slide a mix 4 of two different orders. 5 So the first four columns, the baseline PPD, 6 trends, III/IV spreads and advanced price -- the first 7 five columns -- advanced prices and Class I reform, all of 8 that was for Southwest Order, and then the depooling and 9 actual PPD were for the Mideast order. 10 THE COURT: No wonder you couldn't figure it out. 11 THE WITNESS: So to err on the side of 12 transparency, I've now included -- I have obviously 13 modified all of these slides to properly be titled 14 Southwest, and the last two columns corrected. But I also 15 included additional slides for Mideast, as while I was 16 testifying on Monday I read off some pooling numbers that 17 were for Mideast. So just to, you know, err on the side 18 of transparency, I included everything. 19 No conclusions are changed, no numbers presented, 20 relative size of these numbers or the absolute values of 21 those numbers, none of that changes. That's all from the 22 file that has been previously part of a package that was 23 reviewed for the Journal of Dairy Science. It was just 24 wrong. The order of the -- the title of the order was 25 wrong, so I now have both Southwest and Mideast. 26 In the Exhibit 15 I only included Southwest. And 27 in the Exhibit 29- -- 28 THE COURT: 298? 6817 1 THE WITNESS: This is 298? 2 THE COURT: Oh, okay. So, I'm sorry. In 297, we 3 have Exhibit 15 corrected -- 4 THE WITNESS: Yes. 5 THE COURT: -- and in 298, 15B corrected. 6 THE WITNESS: Thank you, Your Honor. 7 In the Exhibit 297 I have modified only slides -- 8 only the titles and appropriate sentences to -- to clarify 9 that it's the Southwest, not Mideast. 10 In the Exhibit 298, I have both Southwest and 11 Mideast. 12 Those are the full extent of my corrections. 13 THE COURT: Now, for our viewing audience, who 14 haven't seen anything up on the screen, that will come. 15 THE WITNESS: Yes, these exhibits will be posted 16 on the Federal Order website sometime today, I assume. 17 THE COURT: Very good. 18 And in the meantime, I still would like to 19 entertain moving them into evidence, even though, we 20 know -- we know what to expect. We just haven't seen all 21 of it yet. 22 What did you do in an attempt to alert the people 23 who are in the room? 24 THE WITNESS: This morning, between 6:30 and 25 7 o'clock, I sent these files to the -- to the AMS team as 26 well as the counsels for all other parties. 27 THE COURT: What questions would anybody like to 28 ask Dr. Bozic about this topic? 6818 1 There are none. 2 Is there any objection to the admission into 3 evidence of Exhibit 297, which is the corrected Edge-15, 4 and includes the corrections to the -- did you say titles? 5 THE WITNESS: To the titles and some text in 6 the -- on the -- related to the waterfall charts, related 7 to the depooling analysis. 8 THE COURT: Is there any objection to that 9 document being admitted? That's Exhibit 297. 10 There is none. Exhibit 297 is admitted into 11 evidence. 12 (Exhibit Number 297 was received into 13 evidence.) 14 THE COURT: With regard to Exhibit 298, which is 15 the Edge-15B corrected, including both Southwest and 16 Mideast orders, is there any objection? 17 There is none. Exhibit 298 is admitted into 18 evidence. 19 (Exhibit Number 298 was received into 20 evidence.) 21 THE WITNESS: Your Honor, this is -- I anticipate 22 this is my last time on the stand, so I just want to 23 express my gratitude for the hard work of your colleague 24 that was here the first few weeks, yourself, AMS team, as 25 well as all of the parties, and all I can say is we should 26 do this more often. Thank you. 27 THE COURT: That's great. 28 You know, every participant in this hearing is so 6819 1 valued. The Secretary cannot possibly address these 2 issues without hearing from different parts of the 3 country, different aspects of the business, and everyone 4 is valued. And I appreciate your collegiality, and I have 5 enjoyed, Dr. Boze (sic), welcoming all different people's 6 explanations, and trying to puzzle through it, and trying 7 to help the Secretary find something that would work. And 8 I appreciate very much his good humor. 9 THE WITNESS: Thank you very much. 10 THE COURT: You're welcome. 11 Now, Dr. Vitaliano. 12 And I'm going to take just a minute to talk about 13 the FEMA Emergency System trial that will happen at 14 2:20 Eastern today throughout the entire country, an 15 alert, to see if it works as an emergency alert. So help 16 me remember that we should all go off record about 2:15, 17 so wherever our devices are showing the alert, I think it 18 will be on television as well as devices. So we'll see 19 how that works. But help me be off record by 2:15, if you 20 will. 21 Would you state and spell your name, please? 22 THE WITNESS: Peter Vitaliano, P-E-T-E-R, V as in 23 Victor, I-T-A-L-I-A-N-O. It's the word "Italian" with a V 24 on the front and an O on the back. 25 THE COURT: Have you previously testified in this 26 proceeding? 27 THE WITNESS: I have, Your Honor. 28 THE COURT: You remain sworn. 6820 1 THE WITNESS: Thank you. 2 PETER VITALIANO, 3 Having been previously sworn, was examined 4 and testified as follows: 5 MS. HANCOCK: Good morning, Dr. Vitaliano. 6 Just for the record, Your Honor, not only has he 7 been previously sworn in and testified, but he's also been 8 previously designated as an expert in this matter. 9 THE COURT: Excellent. 10 DIRECT EXAMINATION 11 BY MS. HANCOCK: 12 Q. Dr. Vitaliano, did you prepare Exhibit NMPF-35 in 13 support of National Milk's proposals related to Class I 14 differentials? 15 A. I have. 16 MS. HANCOCK: And, Your Honor, I believe we're at 17 Exhibit 299? 18 THE COURT: Yes. 19 MS. HANCOCK: If we could mark that exhibit. I 20 missed the 300 by one exhibit. 21 (Exhibit Number 299 was marked for 22 identification.) 23 THE COURT: Well, you want to step down for a 24 minute? 25 THE WITNESS: In the interest of moving the 26 hearing along, I will forego that privilege. Thank you. 27 BY MS. HANCOCK: 28 Q. Okay. Dr. Vitaliano, would you proceed with your 6821 1 written testimony, please? 2 A. Yes. 3 This is a -- I have testified on all five of 4 National Milk's proposals. My written testimony has 5 followed the same form in all five. A -- start -- begins 6 with an introductory section of a few pages describing our 7 process of arriving at our package of proposals, and has a 8 section later on on economic impact. 9 My original testimony on Proposal 1 on the very 10 first day of this hearing -- seems like ages ago -- I read 11 the full testimony into the record. Those repetitive 12 parts I have not read subsequently, and I will follow 13 basically that same procedure. 14 But since this testimony on Proposal 19 bookends 15 that original one and the whole series, I will re-read a 16 few selected paragraphs from those common sections to kind 17 of refresh your memory and for the benefit of Your Honor. 18 I will note that the version -- the short version 19 of my written testimony that we're handing out is just the 20 textual part. The full version on the website is about 21 80-some pages and contains the full list of the 3100-some 22 counties, city, and parish differentials that we are 23 proposing. 24 Q. And, Dr. Vitaliano, I forgot to mention. We 25 originally submitted this in September at some point, I 26 can't remember what the deadline was, and you have since 27 amended just the counties, which is that last part of your 28 testimony that begins on page 12; is that correct? 6822 1 A. That's correct. 2 Q. And what did you change in the counties that was 3 resubmitted? Do you recall? 4 A. From our technical group that put these together, 5 I received only two, believe it or not, two corrections, 6 to two counties in Texas. And the version that's posted 7 on the website as Exhibit NMPF-35 has the corrected 8 versions. I'm not sure whether the Appendix A version has 9 them yet. 10 The two corrections are, for those of you who have 11 those, Comanche County in Texas should be $3.85. 12 THE COURT: Do you know what page? 13 THE WITNESS: I don't have that. If you give me 14 the page number for the -- I don't have the full version 15 in front of me. 16 MS. HANCOCK: Your Honor, your version should be 17 corrected. I think he's just noting the difference that 18 happened. But it should be on page 69. 19 THE COURT: Very good. Thank you. 20 And say it again, please, Dr. Vitaliano? 21 THE WITNESS: Comanche County, Texas should be 22 $3.85. 23 And then a few pages later, Travis County, 24 Texas -- there are a lot of counties in Texas, it takes up 25 several pages -- it should be $4.35 -- 26 MS. HANCOCK: And that's on page 73. 27 THE WITNESS: -- instead of 4.70 -- $4.70. 28 With that, let me begin my statement. 6823 1 I'm Peter Vitaliano, Vice President of Economic 2 Policy and Market Research for the National Milk Producers 3 Federation. 4 Skipping to the last paragraph on page 2, those of 5 you following. 6 NMPF has engaged in an almost two-year 7 comprehensive study of needed updates to the Federal Order 8 pricing formula provisions. NMPF has undertaken this 9 important activity with the essential and dedicated 10 assistance of dozen of marketing experts from the staffs 11 of its member cooperative marketing associations. 12 In a series of well over 200-monthly virtual 13 meetings by this mostly virtual meetings, this team 14 examined every detail of the current federal pricing 15 formulas of the Federal Order uniform pricing regulations 16 in 7 CFR, paragraph 1000.50 through 52. 17 The goal was developed -- to develop a 18 comprehensive, integrated, and balanced program of updates 19 to these formulas, to realign them more fully with the 20 structural realities of the current dairy industry, and to 21 address the disorderly marketing conditions which the 22 growing misalignment has allowed to develop. This effort 23 included considerations of mechanisms for making further 24 updates in the future as the industry continues to evolve. 25 The comprehensive package which resulted includes 26 seeking additional legislative authority for USDA to 27 conduct mandatory studies of manufacturing costs and 28 product yield factors, seeking a change via ordinary rule 6824 1 making for the regulation implementing the dairy product 2 mandatory reporting program, and five recommendations for 3 amendments to the uniform pricing provisions for all 4 Federal Orders. 5 The NMPF Board of Directors unanimously approved 6 this package of recommendations, including the five 7 recommendations for proposed amendments to all Federal 8 Orders which NMPF has submitted as Proposals 1, 3, 7, 13, 9 and 19. 10 This testimony today is in support of Proposal 19 11 concerning the Class I and Class II differentials. NMPF 12 requests that the Secretary amend 7 CFR 1000.50(b) and (c) 13 and 1052 applicable -- 14 THE COURT: Sorry, that's .52. 15 THE WITNESS: .52. 16 THE COURT: And please go very slowly through 17 this. This is very hard to capture just by hearing it. 18 THE WITNESS: Okay. 19 -- applicable to all Federal Milk Marketing 20 Orders, as well as 7 CFR, paragraph 1005.51(b), paragraph 21 1006.51(b), and paragraph 1007.51(b), as specified at the 22 conclusion of this testimony, which would increase the 23 Class I differentials for all counties, parishes, and 24 cities of the 48 contiguous United States to reflect the 25 current cost of providing adequate supplies of fresh milk 26 to fluid processing plants. 27 The majority of Federal Order Class I 28 differentials have remained unchanged since Federal Order 6825 1 Reform, as reviewed and revised by Congress. The 2 differentials in the Appalachian, Florida, and Southeast 3 orders were modestly updated in 2008. 4 Just as the Make Allowances embedded in the milk 5 component pricing formulas are out of date, so, too, are 6 the underlying cost assumptions embedded in the Class I 7 differentials. Since the current Federal Order Class I 8 differentials were established, one of their key 9 determinants, fuel costs and the basic per mile cost of 10 hauling milk, have increased significantly. Truck driver 11 per-day hours have been reduced, which has required more 12 truck drivers and investment in more rolling stock. 13 Additionally, federal requirements for in-truck electronic 14 driver and truck logs were implemented during this period. 15 Higher capital investments have also driven up overall 16 milk hauling costs. 17 Other structural changes have increased both the 18 costs and general availability of milk hauling, including 19 increased road tolls, restrictive and variable road weight 20 limits, labor shortages, and truck, trailer, tire, and 21 replacement parts and shortages, as well as significant 22 diesel fuel cost increases. 23 THE COURT: I'm sorry, go back again to the "and 24 replacement parts," and finish from there, please. 25 THE WITNESS: And replacement parts and 26 shortages -- 27 THE COURT: So -- 28 THE WITNESS: -- of replacement parts. 6826 1 THE COURT: So replacement parts costs? 2 THE WITNESS: Costs and shortages thereof. 3 THE COURT: Thank you. You may continue. 4 THE WITNESS: Thank you. 5 Driven by the increase -- 6 THE COURT: I'm sorry, I didn't mean for you to 7 abandon the rest of your sentence. 8 THE WITNESS: Oh, okay. 9 -- as well as significant diesel fuel cost 10 increases. 11 Driven by the increased cost of hauling milk per 12 loaded mile, the cost per hundredweight for 100 miles has 13 almost tripled since the current Class I differentials 14 were established. Compounding this greater expense, 15 opportunities for reducing costs through backhauls have 16 become more limited. 17 For example, in the Florida Order, the marketing 18 area most distant from a reserve milk supply, backhauls of 19 orange juice and orange juice concentrate used to be 20 common. However, today, reduction in the Florida citrus 21 industry and the availability of juice concentrate from 22 other countries have nearly eliminated juice backhauls out 23 of Florida. Where backhauls may still be an option, 24 processors often forbid the possibility by requiring 25 tanker trailers to remain dedicated to delivering milk and 26 dairy products only. 27 Changes in the relative locations of farms and 28 fluid milk processing plants have also increased the cost 6827 1 of delivering Class I milk to markets. Development in 2 exurban fringes has displaced dairy farms. The location 3 of milk production is increasingly distant from human 4 population centers, while Class I processing plants remain 5 in cities due to the higher per unit cost of transporting 6 packaged fluid milk relative to bulk unprocessed milk. 7 The miles that bulk raw milk must travel to get from dairy 8 farms to processing plants have increased. 9 The combination of increased miles milk must move 10 to serve Class I markets and the significant increases in 11 the per mile cost of moving milk is threatening the 12 reliability of milk supplies for Class I use in many 13 Federal Orders. The Class I differentials which continue 14 to be the fundamental regulatory mechanism of the Federal 15 Order program for attracting an adequate supply of farm 16 milk for fluid milk processing remain largely unchanged 17 since Federal Order reform 23 years ago. 18 In addition to increases in milk hauling costs 19 since 2000, all contributors to the costs of producing 20 Grade A milk at the farm have also increased. Class I 21 prices are the only Federal Order prices for which the 22 cost to producers is taken into account, albeit in an 23 indirect fashion. 24 The Federal Order base Class I differential has 25 historically recognized that there has been a difference 26 in the cost of producing milk solely for manufacturing use 27 and the cost of producing for daily delivery to the 28 Class I market. Over time, and with the Federal Order 6828 1 reform changes in manufacturing class use prices 2 eliminating any competitive milk procurement factor in a 3 base milk price, the Class I differential base price now 4 represents a modest nod to production costs at the 5 producer level. 6 Since 2000, those costs have risen far more than 7 the limited increase in the base Class I differential from 8 $1.60 per hundredweight to $2.20 per hundredweight as 9 embedded in the NMPF proposal, Proposal 19 that is. The 10 base Class I differential also plays an important role in 11 reducing instances of class price inversions, the 12 importance of which the Department stressed in Federal 13 Order reform, as previously reviewed in my testimony on 14 Proposal 13 earlier in this hearing. 15 NMPF recognizes and supports USDA's longstanding 16 policy of maintaining federally-regulated prices as 17 minimum prices and allowing market forces to fine-tune 18 market prices. However, structural changes in the 19 industry are limiting the reach and effectiveness of 20 over-order pricing for milk used in fluid milk products. 21 Larger fluid milk plants, higher costs of hauling, 22 increased distances raw unprocessed milk must travel to 23 supply Class I processing needs, and growing resistance by 24 handlers to accept over-order prices are leaving many 25 costs of serving Class I processors increasingly 26 uncovered. The result is disorderly marketing conditions. 27 As costs increase and the capacity for over-order prices 28 to keep up with these costs wane, pricing equity between 6829 1 competing processing plants is threatened. Worse, dairy 2 farmers are subsidizing shortfalls of Class I prices to 3 cover the full cost of supplying Class I milk to 4 processors. 5 Taken together, milk transportation costs, 6 producer production costs, and other factors have created 7 a market environment in which the Federal Orders operate 8 which is antithetical to the goals of the Federal Order 9 system. That is, ensuring adequate supplies of milk for 10 fluid processing, equitable treatment of producers and 11 processors, and providing for the orderly marketing of 12 milk. It is important for USDA to ameliorate this, as 13 well as other changes that are eroding the effectiveness 14 of the Federal Order system. 15 Our proposed solution to update the current 16 Class I differentials for all counties, parishes, and 17 cities in the contiguous United States. 18 NMPF's proposal to address these multiple 19 challenges and to help alleviate the economic stresses on 20 milk marketers who have accepted the responsibility of 21 supplying the marketplace with milk for Class I use is to 22 update the adjusted Class I differentials for every U.S. 23 county, parish, and city currently listed 7 CFR, paragraph 24 1000.52. 25 The method NMPF has followed to develop its 26 proposed update to the Class I differentials follows the 27 general process previously used during Federal Order 28 reform. This method also follows certain precepts of 6830 1 price alignment accepted by the Secretary in the 2 Southeastern Order pricing hearing held in 2007. 3 In brief, NMPF commissioned an update to the 4 University of Wisconsin, previously Cornell University, 5 national price surface model using 2021 model input data, 6 including milk supplies, dairy product demand, cost of 7 processing milk, and the cost of transporting milk and 8 dairy products. The model has been greatly expanded to 9 include many more supply and demand points, as well as 10 considerably more point-to-point road mileages. 11 NMPF used the model outputs from the University of 12 Wisconsin model as a starting point. NMPF then applied 13 local knowledge of milk movement, plant locations, and 14 historic price relationships to refine the model results 15 and prepare a rational regulated Class I value surface, 16 using time-honored Class I price alignment techniques and 17 processes. NMPF's final Class I recommendations deviated 18 somewhat from the model results due to a variety of 19 real-world milk movement considerations, as will be 20 addressed in further hearing testimony. 21 In all locations, as would be expected given the 22 substantial increases in the cost of milk hauling, the 23 recommended regulated Class I differential surface 24 increased versus the current regulated Class I 25 differentials. The tilt, or slope, of the price surface 26 from reserve supply points to Class I demand points has 27 become steeper, and the geographic locations representing 28 the reserve supply of milk have generally shifted toward 6831 1 western states. Similar to the general nature of the 2 existing Class I differential price surface, the updated 3 price surface slopes from lower values in the Northwest 4 and West areas of reserve supply, with increasing values 5 when moving toward the milk deficit areas of the 6 Southeast. 7 The updated Class I differentials, as proposed, 8 which resulted from this NMPF analysis, reflect less than 9 the full cost of moving milk, and thereby maintain the 10 Department's longstanding principle of minimum prices. In 11 developing this proposal, NMPF used the expertise of 12 numerous individuals responsible for marketing milk in 13 NMPF member cooperatives, as well as others that have 14 longstanding expertise in the national Class I price 15 surface. Their expertise was used to further refine the 16 model results to develop the proposed pricing surface that 17 best fits the reality of today's marketplace. As such, 18 the proposal does not follow the model's results in every 19 instance, as there are both positive and negative 20 deviations from the model results to better support a more 21 orderly marketing system. 22 The results of the NMPF study, analysis, and price 23 alignment processes are included in Figure 1 below. It is 24 a color-coded representation map, as shown, that visibly 25 presents the 3,108 counties, parishes, and independent 26 cities and each civil district's Class I differentials. 27 Exhibit USDA-46, which is Hearing Exhibit 28 Number 46, provides a summary of the Proposal 19 national 6832 1 average Class I differentials by Federal Order. 2 This testimony provides an overview of NMPF's 3 justification for adoption of Proposal 19. More detailed 4 testimony will follow that supports all or key portions of 5 Proposal 19, including testimony provided by Jeff Sims, 6 representing NMPF member cooperative Lone Star Milk 7 Producers, as well as an expert witness from the 8 University of Wisconsin who will testify about the 9 national price surface model used to develop Proposal 19, 10 also, other members of the NMPF task force that developed 11 NMPF's Federal Order modernization proposals, and 12 producers who are members of NMPF member dairy 13 cooperatives. 14 I will read a few more paragraphs from the 15 following section on the economic market impacts of NMPF's 16 proposed changes, starting in the top of page 8: 17 Figure 2 provides a perspective on the key issue 18 of the impact of NMPF's proposals on consumers of the 19 Federal Order program and potential changes to the 20 regulatory provisions of that program. This figure charts 21 the monthly consumer price indices (CPIs) reported by the 22 U.S. Bureau of Labor Statistics (BLS) over the past decade 23 and a half for all items -- which is the line in red -- 24 which is the general measure of overall consumer price 25 inflation, also referred to as the overall cost of living, 26 together with the aggregate CPIs for all food and 27 beverages shown in green, for all dairy products shown in 28 the bright blue, and for all fluid milk products shown in 6833 1 a sort of darker shade of blue, the principal -- which is 2 the principal regulatory focus of the Federal Order 3 program, that is, fluid milk. 4 These CPIs reflect actual retail prices paid in 5 all U.S. cities, but they are expressed in the form of 6 indices with their respective U.S. average retail prices 7 during the 36-month period of 1982 through '84, each set 8 to the value zero to facilitate comparisons. 9 THE COURT: Now, mine doesn't say "zero," so you 10 will need to explain to me. 11 THE WITNESS: Oh, set to the value 100, to 12 facilitate -- 13 THE COURT: So, it's not me. 14 THE WITNESS: Excuse me? 15 THE COURT: I'm just making a joke. 16 THE WITNESS: Oh. 17 THE COURT: So I don't quite understand that, but 18 I'm sure I will before we finish. 19 THE WITNESS: Yeah. The use of indexing is a 20 standard method when you want to compare something like, 21 you know, the cost -- the cost of a product A might have 22 been $2 each in a base period, and the cost of another 23 product B that you want to make a comparison to may have 24 been $3, and so if -- if both of them have increased at, 25 let's say the same rate, in ten years from that base 26 period, they will still be different. 27 If you want to show how the price of both products 28 changed relative to each other, you would take that base 6834 1 period, you would divide the $2 product price by $2 and 2 get it down to say, you know, basically 100. If you 3 divide $2 by $2, you don't get a dollar figure, you just 4 get an index number. That would be 100, or basically, you 5 know, 100% starting out. You would divide the $3 product 6 price, product B, by $3, and put its higher price down to 7 an index value of 100. So the two of them would start out 8 at the same relative price. 9 So over time, let's say ten years out, if 10 product A price stayed fixed, and product B's went up by 11 10%, at the end of that ten years, the index of product A 12 would still be 100, the index of product B would be 110, 13 so you could see instantly that product B inflated more 14 than product A. Whereas, if you looked at a chart of $2 15 and $3 starting out, and $2 and, you know, 3 -- you know, 16 3.30, in ten years, it wouldn't be quite as obvious how 17 they changed relative to each other. So it's just a way 18 of putting them on a common denominator so you can 19 compare. 20 And so index -- or Figure 2 then shows the 21 relative rate in which general inflation in red, all food 22 and beverage inflation, which is an aggregate number, 23 everything put together, that's the way CPIs are done, 24 where some of them are very broad and some of them get 25 more and more specific. You can get down to -- you know, 26 the price of butter has its own CPI. And then you can 27 compare how a somewhat more disaggregated category, like 28 dairy, which is embedded in that all food and beverages, 6835 1 you can see how dairy in the brighter blue line has 2 inflated relative to all food and beverages, and you can 3 see how dairy, over time, has gotten less and less 4 expensive relative to all food and beverages. 5 They started out the same at 100 during that base 6 period in the early 1980s, but as I'll show in my 7 continued text here, going back to 2008, although that's, 8 what, 25 years after the base period, they still -- all 9 four of these indices were pretty close to each other. 10 That means in 2008, the overall cost of living was -- had 11 gone up about the same as all food and beverages, which 12 was about the same as all dairy, which was about the same 13 as fluid. They were all about the same. You could 14 have -- you could have updated them to an index of 2008, 15 and this chart would look very similar. That's just 16 indexing chart mathematics -- arithmetic is what it is 17 really. 18 THE COURT: Whoa, thank you. I would never have 19 been able to figure it out without your explanation, and I 20 appreciate that. 21 THE WITNESS: I'm happy to give it an explanation, 22 because it's a pretty simple concept once you look at it. 23 So what this chart shows us is over these 24 15 years, how much have consumers needed to pay -- how 25 much more have consumers needed to pay for the overall 26 cost of living, everything they spend money on? How much 27 more they have had to pay for food and beverages, how much 28 they have had to pay for all dairy as an aggregate 6836 1 category, how much more they have had to pay for fluid 2 milk. 3 So continuing to -- Figure 2 shows that the retail 4 prices represented by all four of the measures pictured 5 had increased as of 2008 by about the same amount, 6 slightly more than doubling during the quarter century 7 since the index base period. 8 And that is because the index period they were all 9 at 100 in the early '80s, in 2008 they were about 210, 10 meaning they had, both slightly -- all of them had 11 slightly more than doubled. That's what going from 100 12 index value to 210 means. They had all kind of gone up 13 the same. 14 From 2008, the overall cost of living and the cost 15 of all food and beverages have both continued to increase 16 at a relatively steady pace, which accelerated during the 17 recent bout of inflation, mostly last year and the year 18 before, 2021/2022, at a relative -- with food and beverage 19 prices slightly outpacing the overall inflation rate 20 particularly in recent months. 21 And that's where those -- toward the right-hand 22 side where everything started going up faster, that was 23 the inflationary period that we have all read about in the 24 last couple of years. 25 The less aggregated dairy and fluid milk CPIs have 26 shown a greater sensitivity to the price of producer milk, 27 including the 2009 price plunge, the price spikes of 2014 28 and 2022, and the stagnation of prices between these two 6837 1 peaks. This closer connection between farm and retail 2 prices for dairy stems from the fact that the cost of raw 3 milk has averaged about 31% of the retail value of dairy 4 products since 2002, while the farm value of most food and 5 beverage products represents a much smaller share of the 6 total retail value of the finished products, which 7 accordingly, reflect more closely the main drivers of 8 overall retail price inflation, including such factors as 9 energy, labor, and transportation. 10 That means when you have such broad categories, 11 all food and beverages, it's a very specific part of the 12 economy, but it's so broad that the rate of inflation 13 that's affected food is not all that different, slightly 14 faster, than affecting everything in the economy, which is 15 a much broader measure. Because it's -- food and beverage 16 is such a big category by itself, whereas dairy is more 17 specific and a little different because the value of the 18 raw product, milk itself, raw milk, is a much bigger 19 portion of the retail price than, say, how much a box of 20 corn flakes -- how much the price of raw corn affects a 21 price of a box of corn flakes, which is much smaller. 22 However, those factors have also caused retail 23 prices, price inflation for dairy products, to outpace 24 general and food/beverage price inflation during the 25 recent bout of general price inflation. That's 2022 26 particularly. But also, it's caused dairy prices to 27 recover more quickly from that bout of inflation with 28 dairy product retail prices actually dropping this year, 6838 1 while the two more general CPIs, overall inflation and 2 food and beverages, continue to increase. 3 And you can see from this figure that food and 4 beverage inflation has -- has actually recently outpaced 5 overall inflation slightly, that's the green line, 6 diverging above -- increasingly above the red line, 7 overall inflation. 8 But you will note that the dairy line has 9 actually -- it went up faster during the -- than the 10 another two broader categories, during this recent 11 inflationary period, but it's now dropping. And fluid 12 milk has stayed generally below the overall dairy rate of 13 inflation during most of this period. It experienced a 14 bout of increased inflation along with all these other 15 categories, but is now dropping down again below the 16 overall dairy line. 17 Of particular significance to the -- for the 18 current purpose, the overall cost -- and this -- general 19 purpose, we have had a lot of discussion about the impact 20 of prices to consumers and its effect on fluid milk 21 consumption. The overall cost to consumers of dairy 22 products, and fluid milk products in particular, has 23 declined during the illustrated period relative to both 24 overall inflation, as well as general food and beverage 25 price inflation. 26 One noteworthy datum is that the simple difference 27 by which the monthly CPI for all fluid milk has fallen 28 below the monthly CPI for all food and beverages reached 6839 1 its highest level ever in July 2023. That's the 2 difference over on the far right-hand side between the 3 green line where it -- where it ends against that right 4 margin, and the duller blue line representing fluid, which 5 is the lowest of these. You can see how that fluid milk 6 has diverged more and more below the overall cost of food 7 and beverages. 8 Agricultural production enjoys built-in 9 productivity advantages due to its biological basis, which 10 can generate increases in production per animal, or 11 increases in production per planted unit as a result of 12 genetic improvements and other productivity, which are 13 enhancements unique to biological production processes. 14 These advances generate unit cost reductions which the 15 competitive nature of farming passes on up the various 16 agricultural and food marketing channels, eventually to 17 consumers. This consumer cost reduction aspect of 18 agriculture varies in direct relation to the proportion 19 which the basic agricultural commodity represents of the 20 total retail value of the resulting food products, which, 21 as mentioned, is relatively high for dairy products. 22 This aspect of agricultural production, coupled 23 with the great productivity of U.S. agriculture, has 24 resulted in the general cost of food representing one of 25 the smallest proportions of total consumer income in the 26 United States compared to that in all other countries. 27 It is, therefore, very difficult to consider the 28 facts presented in Figure 2 which reflect the relative 6840 1 influence of all economic factors at play in producing 2 general, food and beverage, overall dairy product, and 3 fluid milk product price inflation over the past decade 4 and a half, which is a period that includes the continuous 5 operation of the Federal Order program -- to go back and 6 repeat the beginning -- it is, therefore, difficult to 7 consider these facts and conclude that Federal Orders have 8 had a deleterious effect on consumer welfare via the 9 retail price of dairy products and retail prices of fluid 10 milk and retail prices of dairy products in general. 11 Skipping then to last section on the bottom of 12 page 10. 13 NMPF sincerely wishes to thank Secretary Vilsack 14 and the Department for holding this important hearing, and 15 for thoughtfully considering adoption of its proposed 16 amendments to the Federal Milk Marketing Order 17 regulations. NMPF has devoted considerable time and 18 resources to thoughtfully considering and recommending the 19 important changes it considers necessary to correct the 20 growing misalignment between the dynamic changes in the 21 U.S. dairy industry since Federal Order reform and the 22 largely unchanged factors in the critical Federal Order 23 component and Class IV class price formulas originally 24 adopted at that time, the time of Federal Order reform. 25 Together, NMPF is requesting the Secretary to 26 amend certain provisions of 7 CFR 1000.50.52 -- excuse 27 me -- dash, 52, those three sections, 1000.50, 1000.51, 28 and 1000.52, which are applicable to all Federal Milk 6841 1 Marketing Orders, and 7 CFR 1005.51(b), paragraph 2 1006.51(b), and paragraph 1007.51(b). The changes in 3 these regulations that Proposal 19 would entail are as 4 follows, which includes, as we have always portrayed 5 them -- the proposed regulatory changes we are portraying 6 in our testimony at this hearing, include all of the five 7 proposals' language. We have not singled out a single one 8 of them. 9 And the changes that Proposal 19 would bring are 10 relatively simple on this -- this page. 11 Section (b) of 1000.50, class prices, component 12 prices and advanced pricing factors, (b), Class I skim 13 milk price: The Class I skim milk price per hundredweight 14 shall be the adjusted Class I differential specified in 15 paragraph 1000.52 -- strike "plus the adjustment Class I 16 prices" in those three sections indicated, which our 17 Proposal 19 would propose that those separate -- separate 18 amended -- or increased Class I differentials in the three 19 Southeastern Orders be reincorporated back into 1000.52. 20 Since our proposal redoes the entire differential surface, 21 there's no need to keep those separated. And to simplify, 22 those would be struck, plus "the simple" -- the higher of 23 the advanced pricing factors, etcetera. That's our 24 language -- proposed language for Proposal 13. 25 (c), the Class I butterfat price: Similarly, the 26 Class I butterfat price per pound shall be the Class -- 27 adjusted Class I differential specified in 1000.52 divided 28 by 100, strike the language that adds those three 6842 1 Southeastern Order butterfat differential sections. 2 They'd be reincorporated into section -- 3 paragraph 1000.52, along with the skim milk price factors. 4 And then we would delete -- propose to delete in 5 their entirety, Sections paragraph 1005.51(b), 6 paragraph 1006.51(b), and paragraph 1007.51(b). Those are 7 the sections in the three Southeastern orders that specify 8 those -- those adjustments to the base Class I 9 differentials that would now -- we would roll into 10 1000.52. 11 And then the adjusted Class I differentials, 12 adjusted for location to be used in 1000.50(b) and (c) 13 shall be as follows: We would delete everything that 14 follows in that -- in section -- in the language of 15 paragraph 1000.52 and substitute the list that is on 16 page -- starting page 12 through page 82 of Exhibit 299, 17 which includes the recommended price surface -- Class I 18 differential price surface that National Milk is proposing 19 in Proposal 19. 20 So that concludes my read/spoken testimony. 21 MS. HANCOCK: Your Honor, at this time we would 22 make Dr. Vitaliano available for cross-examination. 23 THE COURT: Thank you. 24 MR. ENGLISH: Good morning, Your Honor. 25 CROSS-EXAMINATION 26 BY MR. ENGLISH: 27 Q. My name is Chip English for the Milk Innovation 28 Group. 6843 1 Good morning, Dr. Vitaliano. 2 A. Good morning, Mr. English. 3 Q. So let me start off, and I might end here as well, 4 is this the last time you will be presenting for National 5 Milk at this hearing as far as you know? 6 A. As far as I know, this is the last time I will be 7 presenting testimony. With this hearing, I have stopped 8 making predictions. 9 Q. And so I thank you for that. 10 So let me begin at the bottom of page 2 of your 11 statement, and that is the discussion about a two-year 12 long comprehensive study. 13 When precisely did National Milk Producers 14 Federation begin the comprehensive study? 15 A. I began probably in the summer of 2021, or two 16 years ago, by looking at all of the current Federal Order 17 product price formulas as shown in the USDA AMS fact 18 sheets that were handed out here, the Class I, Class II, 19 Class III, Class IV. Looked at each of those pieces in 20 every -- every part of those proposals, and looked at, you 21 know, what -- what might need to be updated. 22 And then there was discussion with it. It 23 probably started rolling into a higher gear in late 2021, 24 when we formally put together a task force of our member 25 specialists. We hired a consultant, Mr. Jim Sleper, to 26 manage that process. 27 So there was not a kickoff date where we said, 28 we're now in the process. But by the end of 2021, we were 6844 1 fully engaged in this process. 2 Q. Would it be fair to say that that was also in line 3 with when the International Dairy Foods Association 4 engaged about Make Allowances? 5 A. I don't know your exact timeline, but I know that 6 was a -- that was an effort that the International Dairy 7 Foods Association undertook. 8 Q. So when did Class I come into the equation, at the 9 same time or after the Make Allowances conversation? 10 A. The Class I surface discussion was the last piece 11 because it was going to be the lengthiest, and we needed 12 the University of Wisconsin study results to begin that. 13 And so that -- that part could not proceed in earnest 14 until we had -- you know, turned out to be the third 15 iteration of the model results to -- to work with. 16 Q. So -- so actually, there's a lot to unpack there, 17 so I appreciate it very much, because you anticipated 18 about the next eight or nine questions. 19 So when did National Milk Producers Federation 20 retain the University of Wisconsin to perform the model 21 study -- the first model study? 22 A. I can't give an exact time, but I would say it was 23 basically in -- sometime in the springtime of 2022. 24 Q. Which would be consistent with the fact that the 25 runs that Dr. Nicholson provided were for May of 2021 and 26 October 2021, correct? 27 A. Yes. We wanted to include to -- to include a -- 28 you know, a recent period, but we wanted to avoid using 6845 1 2022 numbers that were probably subject to this recent 2 bout of inflation that I illustrated in Figure 2, on the 3 assumption that that might be a little non-representative. 4 We may end up being wrong there. But we intentionally did 5 not take the most recent highest cost in that current bout 6 of inflation. We intentionally limited it to 2021. 7 Q. So when did National Milk Producers Federation 8 receive the first iteration of the model results? 9 A. Probably would have been sometime in the spring of 10 2021, 2022. 11 Q. And what did the University of Wisconsin do for 12 that study, that model run, if you know? 13 A. That first model run, the University of Wisconsin 14 crew -- there's a long pedigree to that model. They have 15 updated it and run it for various purposes several times, 16 I think including several times since 2021. The model has 17 grown in size and complexity as the computing power in a 18 laptop has grown. 19 And so I think during -- during recent years, the 20 keepers of that model have updated a lot of the components 21 of it, even prior to us engaging them. But the one 22 thing -- among others, the one thing that they really 23 looked to our help for was to update the plant list, 24 because our task force had a lot of knowledge of current 25 plants, plants that were going to be closed, plants that 26 were soon coming online. 27 And so the first model run, the University of 28 Wisconsin folks, you know, running the model, had -- had 6846 1 updated a lot of the parts of it. Then, you know -- well, 2 go ahead. So that answers that question. 3 Q. Okay. So -- so you received the results of the 4 first iteration. 5 And what, if anything, did you ask with respect 6 to, say, maybe this plant information with respect to the 7 second run -- now, I am focusing on the second iteration 8 right now rather than the third. What, if anything, did 9 you ask with respect to the second iteration? 10 A. We took a look at it, at the results, and even 11 after the first run, we concluded that the model results 12 even of that first run were a relatively good 13 representation of what our specialists, with all of their 14 local knowledge, understood might be a -- you know, a 15 reasonable current Class I differential surface. 16 We -- we didn't -- other than providing some 17 updated plant information, I don't recall we made any 18 major changes. We will -- our -- our next witness is one 19 of the -- is the current keeper of that model, and so he 20 probably would be better -- you know, better informed in 21 terms of what we fed back to them at that time. And so I 22 would recommend you keep that question, make sure you ask 23 him that question as well. 24 Q. I appreciate your attempt to deflect to him, but 25 for the moment, if I may, I at least want to explore, 26 since you are the witness for National Milk Producers 27 Federation, and the other witnesses, 20 or so, are either 28 Dr. Nicholson or individual NMPF members. I'm trying to 6847 1 just focus and understand from everybody. So I get it if 2 you don't have the precision, that's fine, I just want to 3 understand what you recollect. 4 So when did you receive the results of the second 5 iteration? 6 A. It would have been several months after that, 7 after we received the first one. It took some time 8 between -- between the iterations of the model. 9 Q. So early summer, mid-summer 2022? 10 A. Probably around that time, yes. 11 Q. And then after you received the second iteration, 12 what did you ask of University of Wisconsin before it ran 13 the third iteration? 14 A. We might have added a few more -- given a few more 15 updated plant information, but I don't recall there was 16 anything of great significance that we -- you know, we fed 17 back. There was some -- you know, all of our individual 18 task force members looked at the numbers in their 19 particular regions because they were gearing up for the 20 task of taking the final run and working through the 21 process of up- -- of adding their institutional knowledge 22 of -- of, you know, the realities of the industry in their 23 regions to those results. 24 So they wanted to make sure that -- that the model 25 results were -- were reasonably correct, because we did 26 not -- we specifically did not want to end up making major 27 changes to what the model showed. So we wanted to just 28 make sure that the results in all of those areas looked 6848 1 reasonable enough so that when we applied the -- the art 2 part of the process, that we would stick as close as 3 possible to what the model results showed. 4 Q. So a few minutes ago you mentioned, whether it was 5 the second iteration or the third iteration -- and I'm 6 going to take it apart -- at first you provided University 7 of Wisconsin information with respect to closed 8 facilities, correct? 9 A. Yes. 10 Q. So those were plants that were closed, not 11 closing, correct? 12 A. There may have been a few that were -- we knew 13 were going to be closing, and so we didn't, you know -- to 14 the extent that we were sure that they were going to 15 close, we -- we felt it was not going to be useful to have 16 that in the model results. We wanted to have it as 17 current as possible. 18 Q. And what about the plants that you understood to 19 be being built, what -- what categories of plants would 20 that include? 21 A. I remember there was, I think, a butter powder 22 plant. But again, I would refer you to our task force 23 members' testimony because they would know much more 24 specifically what they fed into that process in their own 25 regions. 26 Q. You mentioned the art. Other than the plants that 27 were closed or closing, or the plants that were planned or 28 you thought would open, and recognizing I should ask 6849 1 others the details, did you, for the third iteration, 2 provide the University of Wisconsin with any information 3 about the art? 4 A. No, because we understood particularly by the 5 third run, what the model could do, which was amazing, all 6 the detail that it could do. But everybody who was 7 involved in that art part had done this sort of thing 8 before, and they knew the kinds of things that was just 9 not likely to be incorporated in the model. Because we 10 had a very good idea of what the model could and what the 11 model couldn't do, and we were planning to, and preparing 12 for, and did, apply that institutional knowledge that the 13 model was not able to take into account. And there will 14 be plenty of testimony about what those things are 15 originally. 16 Q. I'm very well aware. 17 A. Yes. 18 Q. But -- but what kind of experience -- so let's 19 backtrack for a minute. 20 The last time the Class I differentials were 21 updated nationwide was during Federal Order reform, 22 correct? 23 A. That's correct. 24 Q. And then the only other changes were from the 25 Southeast hearing, which was decided at the end of 26 February 2008, correct? 27 A. That's correct. 28 Q. So did these -- when you say that these people had 6850 1 been involved, are you saying that outside the Southeast, 2 these were people who had been involved in Federal Order 3 reform in this art? 4 A. They were people who were aware of that process of 5 the model results, and -- and what would -- what was 6 generally needed to -- to work with the model results 7 supplying that institutional knowledge of their local 8 areas. 9 Q. And is it National Milk Producers Federation's 10 view that the use of the art made modest changes to the 11 model? 12 A. Yes, we think it has. 13 Q. So to the extent that we have seen on the USDA 14 website the results of the model, you would agree with me 15 that even after the third iteration, the model used the 16 current base Class I differential of $1.60, correct? 17 A. That's correct. 18 Q. When did the concept of increasing the base price 19 from $1.60 to $2.20 arise? 20 A. Throughout the current, the entire process of 21 working with the various runs of the University of 22 Wisconsin model, the University of Wisconsin personnel, 23 Dr. Nicholson, Dr. Mark Stephenson emphasized to us what 24 we already kind of knew, that the model did not solve for 25 the base -- the lowest differential, that it only solved 26 for relative differences. That the model basically came 27 out with -- and, again, the differences between the 28 various locations. And that they continuously asked us 6851 1 what should we set the base differential, the $1.60, even 2 though that did not affect -- they did not need that 3 information for their actual analysis of the spatial 4 differences. 5 Our group was preparing to do the hard work of 6 looking at taking the geographic spatial relationships 7 and -- and modifying them for things that the model could 8 not do. We specifically put off the discussion of what 9 the minimum differential should be until we completed that 10 other process, and then turned our attention to what it 11 should be. The $1.60 was maintained through the model 12 runs because it was what was in the current Federal Order 13 provisions. 14 Subsequent discussions with working with the model 15 results led us to conclude that since the $1.60 was based 16 on several cost factors, and those cost factors had gone 17 up, just like the cost factors affecting the spatial 18 differences had gone up, that we needed to look at 19 modifying that $1.60, and we concluded that that should 20 now be raised. The lowest Class I differential should be 21 raised to $2.20. 22 Q. So isn't it true that one of the considerations 23 for National Milk was that when you saw the third 24 iteration, or maybe even the first and second at $1.60, 25 there were locations, especially in the West, Southwest, 26 where the differential went down from the current 27 location? 28 A. I believe there were some of those, yes. 6852 1 Q. So how specifically did National Milk Producers 2 Federation develop the base $2.20 used in the model to add 3 the $0.60 to the $1.60? 4 A. There will be, again, extensive testimony on that. 5 But we used the basic framework that USDA used, that USDA 6 and Federal Order Reform identified three components of 7 that $1.60, and we basically updated those three 8 components. 9 Q. So just to be clear, in the $1.60 today, 10 transportation costs are not part of the $1.60, are they? 11 A. No. They are part of the spatial differences. 12 Q. Did National Milk, in considering the development 13 of the $2.20 instead of $1.60, include transportation 14 costs in any way in that $2.20? 15 A. I'm not sure that it officially incorporated 16 transportation costs. There is a component of the cost of 17 assuring a supply of Class I milk in one of the three 18 factors. But our concern -- our -- our feeling was 19 transportation costs were properly covered in the spatial 20 differences that were solved for the University of -- in 21 the University of Wisconsin model, as modified in some 22 cases by the further work of our task force members who 23 had knowledge of local market conditions that would not be 24 reflected fully in the model. 25 Q. And I appreciate that. And, yeah, again, I'm 26 going to have that opportunity to examine the other 27 witnesses for National Milk, the members of National Milk. 28 And again, I'm just trying to understand from you what 6853 1 your understanding was. 2 So -- so you mentioned that there's the three 3 elements of the $1.60. 4 As I read the testimony, there's extensive 5 discussion of the issue of Grade A in the testimony that's 6 going to follow you, correct? 7 A. That is correct. 8 Q. And there's testimony about the inversion issue, 9 correct? 10 A. That's correct. 11 Q. But I did not see any discussion about the other 12 two factors in the base; is that correct? 13 A. What are other the two factors? 14 Q. I think one is viewed as balancing, and one is 15 viewed as the cost of the incentive to get milk away from 16 manufacturing facilities. 17 A. Okay. 18 Q. I did not see that in the discussion. 19 A. I thought those were two of the three. 20 Q. Right. Those are two of the three, and the 21 Grade A is the third, correct? 22 A. Correct. 23 Q. I just wanted to be clear that as I read the 24 testimony -- and there's a lot of testimony here and I 25 could miss something -- as I read it, the discussion is 26 focused on the Grade A and then separately this issue of 27 inversion. 28 Am I correct in the universe there as I understand 6854 1 it? 2 A. In terms of our testimony, I can't answer that. I 3 think you will have to wait for the testimony to follow to 4 speak for itself. 5 Q. Okay. 6 THE COURT: Mr. English, can you remember where 7 you are and let us take a ten-minute break? 8 MR. ENGLISH: Absolutely, Your Honor. 9 THE COURT: Excellent. Let's go off record -- 10 well, first of all, when you come back. Come back at 11 9:30. Let's go off record at 9:17. 12 (Whereupon, a break was taken.) 13 THE COURT: Let's go back on record. 14 We're back on record at 9:35 a.m. 15 Mr. English, you may proceed. 16 MR. ENGLISH: So I want to start where I left off, 17 and then go backwards just for a couple seconds. 18 BY MR. ENGLISH: 19 Q. I understand and appreciate your comment that you 20 believe there are other people who know more or the 21 details of the 2.20, correct? 22 A. Yes. 23 Q. There's a lot of testimony. 24 So could you help me, which of the witnesses who 25 are going to come after you are the best ones, in your 26 view, to talk about the 2.20? 27 A. The first NMPF witness, Mr. Jeffrey Sims, will 28 spend time in his testimony on the 2.20. 6855 1 Q. Okay. 2 A. And Mr. Eric Erba is going to spend some time on 3 that as well. 4 Q. So you think -- I -- I'm not saying others can't 5 or won't, but those two -- 6 A. Right. Those two are going to hit, you know, the 7 main substance of that. 8 Q. I forgot to ask earlier because -- well, whatever 9 reason. 10 When did National Milk receive the third iteration 11 of the model? 12 A. That was, I think, as late as early October of 13 2022. 14 Q. And in reference to a question I asked, you said 15 that there were some modest number of changes to the 16 model. 17 Is there anywhere we can find a summary of the 18 changes? 19 A. Probably the best way to get that information 20 is -- would be to -- to ask Dr. Nicholson from the 21 University of Wisconsin because he received those changes 22 and was responsible for making them, so he has that 23 knowledge of the actual work of doing that. 24 But mostly it was basically plant lists, updating 25 plant lists, and specifically the recommendation on the 26 fuel costs to use. Which, as I had mentioned earlier, we 27 intentionally did not want to have -- we wanted to have 28 2021 fuel costs, which was, you know, prior to a major 6856 1 escalation of those costs that occurred in 2022. We did 2 not want to use the higher numbers of 2022. 3 Q. Do you recall whether he declined to accept any of 4 your suggestions? 5 A. I don't recall that he declined to accept any of 6 them. He was very interested in our knowledge for 7 updating the model, and he very specifically did not 8 indicate in any way that he thought any of the things that 9 we provided him in the way of updated data were 10 inappropriate for -- you know, for the purposes of his 11 analysis. 12 Q. And I apologize because we actually just 13 digressed, which was my fault. 14 What I was referring to was, when we were talking 15 very briefly and just initially about the art that was 16 applied after the third iteration, I thought we talked 17 about the fact that there was sort of a modest number of 18 modest changes; is that correct? By -- from the art? 19 A. How would you describe -- define changes? 20 Q. Okay. So you received the third iteration -- 21 A. Yes. 22 Q. -- and after you received the third iteration, 23 that's when your experts got together and consulted and 24 applied, I thought you used the word art, correct? 25 A. Yes. 26 Q. And I thought, and I might -- we might have 27 misunderstood or miscommunicated -- that you said that 28 there were -- you know, those -- that art resulted in 6857 1 modest changes to the results of the model. 2 Did I have that right? 3 A. Yes. But there were modest -- modest changes to 4 the numbers that the model came out. We did not ask that 5 any of those changes were incorporated back into the 6 model. 7 Q. Okay. 8 A. The model was an objective thing. We were -- we 9 were doing that art part, you know, before -- you know, 10 during -- starting with -- with some of the earlier 11 iterations, because we were under a timeline to present 12 our final recommended price surface to our decision-making 13 bodies. 14 Q. So -- thank you. 15 And I think you answered a question that you -- 16 you did not go back to the University of Wisconsin, 17 when -- again, when you made the art changes, correct? 18 You did not? 19 A. No, we did not. 20 Q. And so when I asked my question imprecisely and 21 asked about the National Milk modifications to the numbers 22 from the model, I was asking is there somewhere where 23 there's a one-page or two-page or whatever summary of what 24 those modifications were, or are? 25 A. I don't know of one that is publicly available. 26 Q. So essentially, one needs to read the 20, plus or 27 minus, testimonies that are about to follow in order to 28 get all of that? 6858 1 A. The testimonies that follow will describe in 2 considerable detail where those -- those changes to the 3 model results were made, which is very, very regional, 4 very local, and therefore, the most pertinent. 5 Q. So a moment ago you said that the process to 6 provide the National Milk modifications started at some 7 point prior, I think to the third iteration. 8 Do you know when they started? 9 A. Probably would have started sometime during -- 10 during the summer of 2022. I did not keep a log of all of 11 these changes. 12 Q. Were there central principles involved for the 13 changes? 14 A. The central principles were basically understood, 15 you know, by the folks -- the task force members that were 16 specifically going to work on that in their regions. And 17 they were made based -- you know, by people who had done 18 this sort of thing before. I can't tell you exactly 19 which -- which process and procedures they were used for, 20 but the people involved had experience with this, and so 21 they kind of knew what was involved. 22 You take, in this case, you know, the results of a 23 computer model that does a wonderful job of getting you, 24 pick a number, 90% of the way, but there inevitably -- 25 when you are doing something as important as setting -- of 26 recommending what the Class I differential should be, you 27 cannot take the results of a model, no matter how 28 wonderful it is, without adding some particular -- 6859 1 particular things to it that based upon the institutional 2 knowledge of experts who know about moving milk in their 3 particular areas. 4 Q. Was somebody overall in -- for want of a better 5 phrase, in charge of these committees? 6 A. Mr. Jeff Sims was formally the chair of the 7 Class I surface working group, but there was no, you know, 8 master plan. There was basically -- it was primarily, you 9 know, getting -- getting the folks to get the work done 10 and putting their individual expertise in. And 11 particularly in the areas where there was no one person 12 who had the detailed institutional knowledge of what the 13 changes to the model for the Class -- you know, for the 14 differentials in a particular region. Nobody in the group 15 had that knowledge of every one of the 3100-some county, 16 city, and parish differentials. It was a rather 17 decentralized process. 18 Q. But were there sort of common precepts? 19 A. Yes. 20 Q. What were the common precepts? 21 A. The common precepts were things such as -- let me 22 give you an example of the ones, because I was not 23 detailed involved in a lot of this. 24 As I recall, the model showed that there should be 25 a different differential for the cities in Texas of Dallas 26 from Fort Worth, because those cities are some distance 27 apart with respect to the major milk supply serving those 28 cities in West Texas to the Texas Panhandle. For 6860 1 institutional purposes, it was decided to -- that the 2 differential, despite what the model said, should be the 3 same for those two cities because of historic price 4 alignment. 5 There was a considerable -- without being enslaved 6 to the past, there was considerable effort and care taken 7 to make sure that the updating did not do -- you know, I 8 might say, you know, disruptive -- make -- make disruptive 9 changes to existing price relationships, particularly, you 10 know, amongst plants that are located relatively close to 11 each other. We tried to respect the fact that the 12 existing differential surface, even though it was 13 outdated, imposed certain competitive relationships that 14 we did not want to be disruptive of, to the extent 15 possible. 16 Q. Okay. Anything else? 17 A. Oh, things like there was a feeling that the 18 differences between cities or plants where there was a 19 mountain range in between, where -- where travel times 20 would be, you know, more difficult than -- than would have 21 appeared based on the model results, some of those things 22 needed to be modified. 23 And, again, you will -- you will receive 24 voluminous testimony from those who have the expert 25 knowledge in their areas of those -- exactly those kind of 26 things. 27 Q. Believe me, I'm aware there's voluminous 28 testimony. 6861 1 Is it your understanding that the model does not 2 take into consideration issues like mountain ranges? 3 A. It must -- it -- it takes it into account -- the 4 model basically uses standard road mileages between point 5 to point, and it has several millions of those 6 point-to-point arcs. It does not necessarily reflect 7 differences in travel time for terrain. There are a lot 8 of areas where there's a lot of congestion on roads where 9 the travel -- travel distance would be a lot slower, and 10 therefore more costly in terms of driver time than the 11 model was able to take into account. 12 Again, the model does an incredible job of 13 incorporating an awful lot of complexity, but there is 14 another level of complexity that really needs to be -- to 15 be taken into account to accommodate some changes from the 16 results of even an almost perfect model. 17 Q. Anything else? 18 A. Those are things that I would mention at this 19 point. And, again, you will hear many of them in the 20 subsequent testimony. 21 Q. So on page 6 of Exhibit 299, in the middle of the 22 page, which is the fourth paragraph, and you refer to 23 National Milk used the expertise of numerous individuals 24 responsible for marketing milk in National Milk Producers' 25 member cooperatives, as well as others that have 26 longstanding expertise in the national Class I price 27 surface. 28 Who were those others that have longstanding 6862 1 expertise in the national Class I price surface who are 2 not members of National Milk? 3 A. I did not imply that those were members outside of 4 National Milk. 5 Q. Okay. 6 A. Those were particularly people who were -- who had 7 knowledge of things like transportation, trucking costs, 8 as opposed to people who necessarily were involved in the 9 daily movement of milk. The people who were responsible 10 for moving milk on a daily basis are the people we relied 11 on to really have that -- that -- that boots-on-the-ground 12 type knowledge of these sorts of things. 13 Q. Now, were trucking costs considered in the model? 14 A. The costs of transportation were included in the 15 model, and I assume that that included however you define 16 trucking costs. 17 Q. So I think I'm trying to ask -- and, again, maybe 18 imprecisely -- with respect to the modifications, the art 19 that National Milk employed after the model numbers came 20 out, were there persons outside of National Milk who 21 assisted you in providing analysis for those 22 modifications? 23 A. I guess, how would you define outside of National 24 Milk? I think the vast majority of those -- those changes 25 were made by people who worked for National Milk 26 cooperatives and were direct -- as directly involved in 27 moving milk as -- as we thought was necessary for that 28 purpose. 6863 1 Q. So for instance, was Select Milk Producers 2 consulted? 3 A. No. 4 Q. Was Edge Cooperative consulted? 5 A. No. 6 Q. What about the fluid milk proprietary customers, 7 those who are members of IDFA, those who are members of 8 the Milk Innovation Group, who are -- 9 A. The only members of IDFA that I know we -- were 10 involved were those who were also members of National Milk 11 Producers Federation at the time. 12 Q. Was Organic Valley consulted? 13 A. Not to my knowledge. 14 Q. So do you know precisely who it was from National 15 Milk who was in each back room where it happened? 16 A. I'm not aware there were any back rooms. 17 Q. Okay. These were closed-door meetings of National 18 Milk members, correct? 19 A. Can you define closed-door? 20 Q. Well, we have just said that Select wasn't 21 invited, correct? 22 A. Select was not a member of National Milk during 23 most of that time. 24 Q. So what I'm getting at is, you know, you talked 25 about employing the expertise of industry, while I think 26 what you are telling me is the expertise in the industry 27 was limited to National Milk members. 28 A. The expertise was based upon the task force that 6864 1 we put together for National Milk, because we figured we 2 had all the information, the objective information that we 3 needed for this process. 4 Q. But, in fact, you know, you excluded two 5 cooperatives, Organic Valley and Select, correct? Or 6 three, actually, Edge. Edge, Organic Valley, and Select, 7 correct? 8 A. We -- we invited anybody who wished to participate 9 in the process, in the task force process who was a member 10 of National Milk and was willing to supply the 11 expertise -- time and expertise of their members -- of 12 their staff that had the knowledge we needed. 13 Q. Was an invitation issued to anybody who was not a 14 member of National Milk? 15 A. Not that I'm aware of. 16 Q. Wouldn't it be fair to say that entities like 17 Select Milk Producers, Edge, Organic Valley, and members 18 of IDFA who are proprietary operations would also have 19 local knowledge of the markets? 20 A. Well, let's say if we wanted to have an open 21 seminar or workshop and invited everybody in the country 22 that might have been able to contribute, we would have had 23 a much bigger process. 24 We felt that we had all the expertise we needed. 25 We were not trying to exclude anybody. We were trying to 26 get a job done, and we felt that we had the resources to 27 do that. 28 Q. Wouldn't you agree that there's at least an 6865 1 appearance of unfairness when some members of the industry 2 get to give input to change the model results and others 3 don't? 4 (Court Reporter clarification.) 5 BY MR. ENGLISH: 6 Q. Wouldn't you agree that there's at least an 7 appearance of unfairness when some members of the industry 8 get to give input to change the model results and others 9 don't? 10 A. I don't think there's any -- any reason why 11 that -- why that sense would be -- would be significant, 12 you know. If you are telling me that you have that sense, 13 that's your privilege. 14 Q. Would National Milk Producers Federation accept a 15 model that has been modified by Select to specifically 16 reflect markets where it has plants and understands the 17 conditions in the market? 18 A. If Select chose to forward a model of that sort, 19 we would take a look at it and see if -- and take a 20 position on it. But that's -- there's -- that is not a 21 proposal at this hearing. 22 Q. We have to -- we appear to have -- and I said 20 23 earlier, and maybe that's because I was counting some 24 other witnesses -- 17 National Milk witnesses discussing 25 different regions on the departures from the model. 26 Are there others involved in the National Milk 27 Producers Federation meetings only who made red-pencil 28 adjustments who are not testifying? 6866 1 A. There may be some. I don't have a full list of 2 those folks. But we feel that there's a very generous 3 number of our task force members who are involved in the 4 process who are going to provide extensive testimony on 5 what they did in their area, and they will all be 6 available to be cross-examined. 7 Q. So what kind of horse trading went on in the back 8 rooms given that some members operated Class I plants and 9 others don't? 10 A. I'm not aware of any horse trading. There were -- 11 no horses were involved, just colored pencils, electronic 12 versions. 13 Q. So when you look at the model results, the model 14 provided by University of Wisconsin, gave you a May 15 number, which is spring, and an October number, which is 16 fall, correct? 17 A. Correct. 18 Q. And then National Milk calculated an average, 19 correct? 20 A. That's correct. 21 Q. When National Milk made its modifications, did it 22 consistently use one, that is to say, all spring, all 23 average, or all fall? 24 A. Could you define what you meant by all spring, all 25 average, all fall? 26 Q. All right. So you will have an opportunity in a 27 moment to look at the spreadsheets. 28 The University of Wisconsin provided you a column 6867 1 of spring numbers, correct? 2 A. Correct. 3 Q. Which are generally, not exclusively, but 4 generally lower than the fall, correct? 5 A. Correct. 6 Q. And then it provided a column of fall numbers, 7 correct? 8 A. Correct. That's -- that's the way the model 9 usually -- 10 Q. Yes. 11 A. -- is run. A spring flush month and a fall -- 12 fall -- 13 Q. Whatever. 14 A. -- tighter supply period month. 15 Q. As I asked, and you agreed, that National Milk 16 added a column that was average, correct? 17 A. Right. 18 Q. Okay. 19 A. Because we knew that the Class I differentials in 20 paragraph 1000.52 were a single number. They are not -- 21 they are not seasonably variable. So we knew that we had 22 to work with a single number that combined the two, and 23 the easiest way to do it was to take a simple average. 24 And all of the art part of the process was based upon 25 that -- the average numbers. 26 Q. Are you sure? 27 A. As far as I know. But you can, again, ask the -- 28 ask the individual groups that -- that made the 6868 1 modifications to the model results. But to the best of my 2 knowledge, we always worked with the average because we 3 knew we had to come up with a single number. 4 Q. So you are not aware whether, for some locations, 5 spring value was selected? 6 A. Not to my knowledge. 7 Q. And similarly, you are not aware whether in some 8 instances the fall number was selected? 9 A. No. I don't recall where the difference between 10 the spring and fall numbers was considered of great 11 significance or taken as a major factor that was used in 12 adjusting the numbers. 13 Q. What ultimately is the purpose of the model if it 14 is so significantly altered? 15 A. Define significantly. 16 Q. I'll move on. 17 So on the bottom of page 2 -- the good news is I 18 have moved on to part 2. This is page 2 of Exhibit 299, 19 at least I thought it was. 20 You reference in your testimony that some precepts 21 were followed from the Southeast hearing in 2007, the 22 decision in 2008, correct? 23 A. That experience was -- was available to members of 24 the task force, yes. 25 Q. Are you aware that in that case -- I think there 26 were three people in this room who were at that 27 proceeding -- in that case, SMA, followed by USDA, applied 28 an 80% of hauling cost concept? 6869 1 A. I'm not aware of the details of the considerations 2 in making that 2008 Southeast region differentials. 3 Q. To your knowledge, is there any 80% of hauling 4 cost concept applied in the National Milk Producers 5 Federation modifications? 6 A. I don't recall a fixed number because the 7 transportation costs -- the primary impact of 8 transportation costs in the National Milk recommendation 9 in Proposal 19 came from the model, which is based upon 10 the road network, the fuel costs, labor costs, and the 11 like. It was basically from public sources. We did not 12 dictate a particular transportation cost number to the 13 University of Wisconsin personnel. 14 They -- we wanted their objective model results. 15 The one thing we did ask was that they use fuel costs 16 pertinent to the 2021 months and not the higher costs of 17 the 2022 months that were available at that time. 18 Q. And candidly, that would make sense, because if 19 you are using May and October 2021 data, you would want to 20 have the data match up, correct? 21 A. Yes. We wanted it to be consistent. We did not 22 dictate anything of what we wanted the model to show. We 23 simply provide updated plant information, made the 24 recommendation on using the cost from 2021, and most of 25 the other data was already in the model. 26 Q. So on the bottom of page 4 of Exhibit 299, the 27 last paragraph, you state, "The combination of increased 28 miles milk must move to serve Class I markets and the 6870 1 significant increases in the per milk cost of moving milk 2 is threatening the reliability for milk suppliers for 3 Class I use in many Federal Orders." 4 So first I note you say "many Federal Orders," 5 which is not the same thing as all. 6 So in what Federal Orders is the increased cost of 7 moving milk threatening the reliability of milk supplies 8 in Class I? 9 A. You know, can you repeat the question again? 10 Q. Given the fact that you say "many" rather than 11 "all" in this paragraph, which Federal Orders -- in which 12 Federal Orders is the increased cost of moving milk 13 threatening the reliability of milk supplies in Class I? 14 A. So in which orders it is threatening -- 15 Q. Yes. 16 A. -- as opposed to -- 17 Q. Yes. 18 A. Well, the Texas order is one that we have -- you 19 know, that came to mind particularly, and there will be 20 testimony on that. 21 The Texas market, population is growing. The main 22 urban centers are in east and south Texas, and the milk 23 supply in Texas, the local milk supply is moving from 24 areas closer to those population centers, is moving out 25 pretty -- pretty specifically to the Panhandle area in 26 West Texas. 27 In areas closer to Dallas/Forth Worth, Houston, 28 San Antonio, those more local milk supplies are declining. 6871 1 And that's -- that was a kind of a -- a -- you know, a 2 major example of areas where the milk supplies were moving 3 to areas more distant from the consuming centers where the 4 fluid milk plants were. Those hauling distances are 5 increasing. 6 And there happens -- because the West Tex- -- the 7 Texas Panhandle is an area of production growth, it is a 8 fact of the current dairy industry that new plants are 9 being built in areas where the milk supply is growing. 10 Particularly, as the general patterns of consumption of 11 dairy products are shifting from fluid to manufactured 12 products such as cheese, butter, and ingredients for the 13 growing export market and growing food manufacturing uses 14 domestically. 15 So we have a situation where the milk supply is -- 16 the availability of manufacturing plants near the areas of 17 milk supply is growing, and the availability of milk 18 supplies closer to the fluid milk consuming areas is 19 declining. And, therefore, hauling distances from where 20 the milk is produced to where it's needed for Class I use 21 are increasing. And you will see that in many of the 22 testimonies to follow. 23 Q. I'm going to come back there, but -- okay. 24 Many Federal Orders, so after Texas, which is the 25 Southwest order, what other orders are the increased costs 26 of moving milk threatening the reliability of milk 27 supplies in Class I? 28 A. I would leave that to the individual testimony. 6872 1 You will have all the information you need on that. 2 Q. Okay. So let me come back to Texas. And I'm 3 going to try to avoid pulling Exhibit 39 again, but you 4 have been here for much of the hearing when we have talked 5 about performance standards, correct? 6 A. Yes. 7 Q. And I need -- even just yesterday, I pulled 8 Exhibit 39, which is the changes in performance standards, 9 correct? 10 A. Can you define the performance standards, then? 11 Q. This is the order provisions with respect to what 12 percentage of the milk needs to be, you know, shipped 13 or -- to Class I plants, diversion limits -- 14 A. Yes. 15 Q. Okay. That's what I mean. 16 A. Yes. I don't have particular expertise in 17 applying those because we don't -- I don't -- we don't 18 move milk in National Milk, but I'm aware of those -- 19 those provisions. 20 Q. But you know, one, that there's been no call for a 21 hearing in Order 126, which is the Southwest Order, since 22 some time in the mid-2000s to change those performance 23 standards, correct? 24 A. Not that I'm aware of. 25 Q. And there's been no increase in the performance 26 standards by the Market Administrator, correct? 27 A. Generally, the -- I'm not aware of increases in 28 performance standards, but I -- I would not swear to that 6873 1 under oath. 2 Q. Okay. So isn't it true that to the extent Class I 3 handlers, who do not have an opportunity to depool, to the 4 extent there is any quid pro quo for paying a Class I 5 differential, that the point of that is to get milk to 6 their plants, correct? 7 A. Can you repeat that question, please? 8 Q. You agree -- I'll break it up. You agree that 9 Class I plants are the ones who are captive to the system 10 and must always be in the pool, correct? 11 A. Pool distributing plants must pool their milk, 12 yes. 13 Q. And whether explicit or implicit, the quid pro quo 14 for that payment of a Class I price that is higher or at 15 least generally higher than the other class is that they 16 will have priority to get milk to the fluid plants, 17 correct? 18 A. Could you define the parties to the quid pro quo 19 that you are referring to? 20 Q. The order expressly provides, one, that Class I 21 handlers will pay a Class I differential, the very thing 22 that's at issue in Issue 5, correct? 23 A. Correct. 24 Q. Okay. The order also provides performance 25 standards, that is to say if you want to be in the pool, 26 for those people who don't have to pool, you have got to 27 do certain things, correct? 28 A. That's correct. 6874 1 Q. The point of those performance standards is to 2 move milk to Class I plants, correct? 3 A. That's correct. 4 Q. Okay. I think we have heard a fair bit of 5 testimony this hearing, maybe not by you, that a purpose 6 of higher Class I plants is to cause or otherwise -- I 7 think one person used the phrase "force" -- other classes 8 of milk to pool. 9 Isn't it the case that since that pooling is 10 voluntary, when you say that the increased cost of moving 11 milk is threatening the reliability of milk supplies of 12 Class I in Texas, what you really mean is that the Class I 13 differential that is already being charged is so diluted 14 that the people actually incurring the cost of delivery 15 don't have an incentive to do so? 16 A. I don't know all the mathematics of that, but -- 17 but the -- our members are telling us that the return they 18 are getting from supplying Class I milk, which is 19 expensive, is not returning enough revenue given all of 20 the costs that they are incurring to do it. 21 There's a parallel that I have pointed out, and 22 you will hear it in other testimonies, that just as IDFA 23 has provided testimony that the cost of manufacturing 24 dairy products has increased and is not being covered by 25 the current Make Allowances, a point in which our members 26 generally agree, similarly, the cost of supplying Class I 27 milk to fluid plants has increased, and that -- and the 28 fundamental mechanism for ensuring that fluid plants get 6875 1 adequate supplies of Class I milk are the Class I 2 differentials. That's the basic foundation of the Federal 3 Order program. Those current differentials are no longer 4 adequate to the task, and we're proposing that they be 5 adjusted for -- to conform with current realities. 6 Q. What I'm sort of specifically getting at here, is 7 my understanding was when you say, on the bottom of 8 page 4, that the combination of increased miles milk must 9 move to serve Class I markets, and the significant 10 increases in the per milk cost of moving milk is 11 threatening the reliability for milk suppliers for Class I 12 use, that you are making that statement as a justification 13 for modifications of the University of Wisconsin model 14 results. 15 Am I correct? 16 A. We believe the University of Wisconsin model 17 results reflect that the reality of supplying milk, the 18 cost of supplying milk to -- for Class I plants throughout 19 the United States, and we use that as a basis to come up 20 with our recommendations in Proposal 19. 21 Q. But as you have stated -- and I'll have the 22 pleasure or opportunity, and so will Mr. Sims, to discuss 23 at some length Texas -- you have gone through -- you, 24 National Milk members, have gone through some significant 25 effort to justify modifications to Texas from the 26 University of Wisconsin model results, correct? 27 A. Correct. In general, the modifications that 28 National Milk made based on the institutional knowledge of 6876 1 their -- you know, our members' staffs that have expertise 2 in their local markets were relatively modest compared to 3 the -- to the results of the University of Wisconsin 4 model, which is a greatly expanded and improved version of 5 the model that was used by USDA to establish the current 6 Class I differentials. 7 We didn't come with -- we didn't invent this 8 process of using the University of Wisconsin -- previously 9 Cornell University -- models as the basis and then making 10 some fine-tuning adjustments from that. That was -- that 11 was the procedure that the Department initiated in Federal 12 Order reform to come up with the current differential 13 structure which was considerably different than the 14 previous one, which kind of zoned everything out of Eau 15 Claire, Wisconsin. 16 Q. Well, let's not talk about Eau Claire. 17 So I think I'll probably move on, but I confess, 18 I'm very confused about what's going on in Texas. And for 19 those who know me, when I entered this wonderful business 20 in 1985, it's because of Texas. 21 So I -- I -- what I'm trying to get at is if, as 22 you say, there's all this new cheese production coming on 23 in the Southwest because of the value of milk used in 24 cheese versus fluid milk, why when we have declining milk 25 supplies -- I'm sorry -- declining fluid milk consumption, 26 if that is the case, why are we further increasing Class I 27 prices? 28 A. We are proposing an increase in Class I prices to 6877 1 account for the increased costs of supplying milk to 2 Class I fluid plants for all the reasons of I have 3 outlined, and you will hear in great detail by further 4 witnesses. 5 The fundamental purposes of the Class I 6 differentials is to provide -- facilitate the provision of 7 an adequate supply of fluid milk for -- for Class I 8 manufacturing. And, therefore, we are basically just 9 updating the standard procedures for evaluating and, you 10 know, the proper level of the Class I differentials, which 11 have not changed, mostly, in almost a quarter of a 12 century, while the costs that -- that -- and the 13 structural changes in the dairy industry that are 14 pertinent, directly pertinent to the proper level of the 15 Class I differentials, have not changed. 16 We're simply proposing an update to the -- to the 17 current Class I differential structure based upon the 18 provisions of the -- of the Federal Order and its 19 principles. 20 We are -- we are not aware that the 1937 Act 21 indicates that the Federal Order program is responsible 22 for making changes in the consumption of Americans -- of 23 the American population of fluid milk. 24 Q. But shouldn't it be relevant -- you said in the 25 quarter century since they've been modified, those costs 26 have gone up. The same time in that quarter century, 27 Class I utilization in Federal Orders, which now includes 28 California, is down to 28%, and if you exclude Federal 6878 1 Orders, it's 18%, correct? 2 A. Yes. 3 Q. So leaving aside all the testimony we have had 4 about the Southeast, don't we really have a plentiful 5 supply of milk, it's just that the incentives we have 6 aren't getting to where it's needed? 7 A. Define plentiful supply of milk. 8 Q. I'd say 82% of milk being used in other than 9 Class I is plentiful. 10 You don't agree? 11 A. No, I don't agree with that. Because 12 manufacturing those -- transforming milk into those other 13 dairy products in fluid is just as important to the -- you 14 know, to the dairy industry as trans- -- as transforming 15 that milk into fluid products. 16 Q. So both in your testimony on page 6, second 17 paragraph, and in response to some of my questions, you 18 have referred to alignment as one of the criteria for the 19 National Milk modifications, correct? 20 A. Say that word again? 21 Q. Alignment. 22 A. Alignment? 23 Q. Correct. 24 A. Yes. 25 Q. Are you quite certain the National Milk has 26 honored alignment in its private meetings? 27 A. Yes, to my knowledge. Our members who are 28 actually responsible for supplying milk for Class I use 6879 1 are acutely aware of the disruptions that can be caused by 2 Class I differentials in, you know, nearby counties being 3 out of alignment -- 4 Q. Are you -- 5 A. -- and they have sought to correct some of those. 6 Q. Are you aware of examples where National Milk 7 Producers Federation's intent to ensure historical price 8 alignment were made even if the model concluded that 9 the -- that the values were significantly different? 10 A. I'm not aware that we made major changes in the 11 alignment from the model results to the final Proposal 19 12 results. I mentioned an example in Dallas/Fort Worth 13 where the model showed, as you would expect, you know, 14 where there's, what, a 30-some-mile difference between 15 those, that there would be, you know, a small difference 16 in the model results. For -- for -- you know, for other 17 reasons we decided to -- to make them the same. We did 18 not consider that to be a major deviation. 19 I'm not aware of anything where we -- where the 20 model said the two nearby areas should be, you know, the 21 same or, you know, roughly similar, and we ended up making 22 them vastly different. We respected the general alignment 23 scenario that the model gave us in almost all cases. 24 Q. For the red -- I think you used the word 25 electronic pens, or computers, alterations, was there a 26 limit on the modification size? That is to say, could it 27 be more than $0.10? 28 A. We looked at trying to keep the modifications from 6880 1 the model as minimal as possible, but we did not, to my 2 knowledge, say this is the maximum. We had a general 3 sense of that and -- and again, the results show that 4 those changes from the model results were relatively 5 modest, particularly as a percentage of the Class I price, 6 but I'm not aware that there was a binding limit, you 7 know, you cannot -- you cannot come up with a change that 8 was more than X dollars per hundredweight, or cents per 9 hundredweight. 10 Q. So in answer to questions from your counsel, you 11 indicated that the pages 12 through 82 of Exhibit 299, 12 marked originally as Exhibit National Milk Producers 13 Federation 35, contains the proposed county-by-county 14 Class I differentials with the two corrections you made 15 today, correct? 16 A. Correct. 17 Q. Where in National Milk Producers Federation's 18 pre-submitted testimony can I find the county-by-county 19 Class I differentials that resulted from any -- one or 20 more -- of the University of Wisconsin model runs? 21 A. We submitted in our -- in everything we submitted 22 to USDA in our petition, in our testimony, we basically 23 used the structure of the Federal Order regulations in 24 paragraph 1000.52 as a model, which -- which did not -- 25 which basically stated these are our recommendations for 26 the differential. It was not a didactic exercise that we 27 supplied that information where we wanted to show 28 everything we did. We're not trying to hide anything, but 6881 1 we did not feel in our formal request to the Department 2 and our testimony that it was necessary to provide all 3 that information. 4 Q. So let me just be clear. The one set of numbers 5 that I believe -- and I could be wrong -- that are 6 pre-submitted or at this point you have submitted as 7 Exhibit 299 for the proposed class and differentials and 8 any justification for them in terms of the -- as opposed 9 to what testimony I'm going to get -- is found on pages 12 10 to 82 as corrected of this exhibit, correct? 11 A. That is our Proposal 19. You will hear plenty of 12 testimony from -- from task force members in their own 13 areas of the specific changes they made to the model 14 results and how they modified them based upon their 15 additional information. 16 Q. Now, to be clear, as you referenced a minute ago, 17 the petition that you made to USDA and information 18 supplied to USDA, absent somebody putting that into this 19 record, whatever you filed with your actual petition and 20 the backup materials that might have been submitted, are 21 not part of the record unless somebody makes them part of 22 the record, correct? 23 A. The only thing I'm aware of as part of the record 24 is the differentials that we proposed in Proposal 19. 25 Q. Which are found in Exhibit 299, correct? 26 A. Correct. 27 Q. But, in fact, National Milk Producers Federation 28 submitted to USDA significant spreadsheets with respect to 6882 1 the model runs and National Milk's -- back in May and 2 June, correct? 3 A. Yes. I believe that we provided our -- you know, 4 the model information to the Department as pertinent 5 information to support our proposal. 6 Q. Okay. But at least as of this moment, they are 7 not in the record, correct? 8 A. I have not seen them in the record. 9 Q. Is National Milk planning to put them in the 10 record? 11 A. I can't answer that because I don't know the 12 answer to that question. 13 Q. All right. 14 MR. ENGLISH: Your Honor, it may make sense to go 15 off the record as I pass out -- we pre-filed MIG-29 and 16 MIG-30, I have lost track, Monday night, so it's been 17 available since, you know, USDA posted it at least Tuesday 18 morning. And, of course, what I will pass out has been on 19 the USDA hearing website, if not an exhibit, since this 20 summer, so it's not a surprise to anybody I believe. But 21 if I may hand them out. 22 So pursuant to the rules, my understanding is that 23 we must provide four printed copies as a courtesy. 24 Notwithstanding the expense, we have 25 with us. We can't 25 share, you know, one for every single person, but we 26 wanted to make available obviously to Your Honor, the 27 witness, myself, and a few others. But we have the four, 28 I believe in color, and to save a few pennies, the others 6883 1 are black and white. So if we can go off the record to 2 distribute these. 3 THE COURT: All right. Let's do. I need to 4 stretch some, too, so let's take ten minutes. So please 5 be back and ready to go at 10:40. 6 We go off record at 10:27. 7 (Whereupon, a break was taken.) 8 THE COURT: Let's go back on record. 9 We're back on record at 10:41 a.m. 10 MR. ENGLISH: So, Your Honor, we have passed 11 things out. I think USDA is still maybe marking. I don't 12 believe you have a copy at the moment. I don't believe 13 the witness has a copy. 14 THE COURT: So mine comes from Emily. And are 15 they already marked? Okay. 16 MR. ENGLISH: Okay. So I earlier said MIG-29 and 17 30. I should have said MIG-28 and 29. 18 So I would ask that MIG-28 be marked as 300. I 19 believe National Milk is disappointed but -- you know, 20 they are not 300 on this but -- and that MIG-29 be 301. 21 Is USDA going to supply a copy to the witness or 22 do we need to provide that? 23 MS. TAYLOR: We can give a copy. 24 THE COURT: Okay. We remain on record. I just 25 want to state how we mark these. It is, as Mr. English 26 requested, Exhibit MIG-28 is Exhibit 300. 27 (Exhibit Number 300 was marked for 28 identification.) 6884 1 THE COURT: Exhibit MIG-29 is 301. 2 (Exhibit Number 301 was marked for 3 identification.) 4 THE COURT: I am one of the blessed people who has 5 colored copies, but there's not much colored, actually, is 6 there, Mr. English? The person that has black and white 7 is not disadvantaged. 8 MR. ENGLISH: Well, since I don't have one in 9 black and white, I can't say. Even with my eyesight, I 10 believe they are not disadvantaged. 11 And by the way, I want to note that as we passed 12 them out, we passed them out in one binder clip so that -- 13 that those people in the audience should note that MIG-28 14 is the first 54 pages of what was passed out with one 15 binder clip, and then MIG-29, which is now Exhibit 301, is 16 the 54 pages that follow. So if you are confused because 17 you have only one big document, it's because we -- in 18 order to produce them and pass them out, we did it that 19 way, but they are two separate documents. 20 THE COURT: And just for the record, Mr. English, 21 what size paper is this that they are printed on? 22 MR. ENGLISH: I believe it is 11x17. 23 THE COURT: Thank you. 24 MR. ENGLISH: And the people who actually know say 25 I'm right. 26 THE COURT: Very good. 27 MR. ENGLISH: So as I said before we went off the 28 record, Your Honor, electronic versions were submitted to 6885 1 USDA Monday night, but also I would note that these were 2 submitted by National Milk to USDA. 3 It's my understanding that Exhibit 300 was 4 submitted in May of this year, and that Exhibit 301 was 5 submitted in June of this year. But I can -- I can ask 6 the witness some questions. 7 Further, Your Honor, I represent that these 8 documents that were submitted, were downloaded from the 9 USDA website, and the only change is the header and footer 10 where we added MIG Exhibit Number 28 or 29, and pages 1 of 11 54 as requested by USDA for submissions. 12 I also note that each document at the bottom has 13 the URL where they can be found -- very small print, but 14 it's there -- and I'm not going to attempt to read that, 15 as they are on both the paper copies and the electronic 16 version. 17 THE COURT: Thank you, Mr. English. And you may 18 continue to question. 19 BY MR. ENGLISH: 20 Q. So, Doctor, do you recognize Exhibit 300? 21 A. Yes. 22 Q. And this was submitted to USDA by National Milk in 23 May of 2023? 24 A. Yes, I believe so. 25 Q. And similarly, do you recognize Exhibit 301? 26 A. Yes. 27 Q. And was that submitted to USDA in June? 28 A. I will take your word for it, those dates of 6886 1 submission. We did -- we did supply this information to 2 USDA. And to my understanding, that the -- Dr. Nicholson 3 is intending to also enter these similar information into 4 the record. 5 Q. Well, actually, that anticipates my next question, 6 because it is -- would you agree with me that -- so let me 7 say for the record that there are column letters A through 8 S on Exhibit 300, and column letters -- well, it goes 9 through S, but there's no numbers past O, so A through O 10 on 301. 11 And so when you say that Dr. Nicholson will supply 12 something, in fact, he can supply only a part of this, 13 correct? Because -- 14 A. I'm not sure what he's planning to supply, but in 15 terms of the basic information, we -- we have not intended 16 to keep this private. This is -- we have made this 17 information available. 18 Q. Sir, I did not mean in any way, shape, or form to 19 imply that's what it was. I, frankly, was concerned -- 20 lest somebody think it was part of the record, I have had 21 an off-the-record conversation with one of our colleagues 22 here who was, like, oh, I didn't realize this wasn't in 23 the record. So it certainly is not implied. Obviously we 24 have had access to it, so I don't disagree that it's been 25 public. 26 A. Yep. 27 Q. Absolutely. But -- but let me -- let me see if I 28 can be clear. And so let me run across the columns with 6887 1 what has been called for me, my magic decoder pen. 2 So I want to start and discuss Columns A through 3 E, and then Columns F and G. 4 Column A is simply a model county identification 5 number, correct? 6 A. Yes. Sequential numbers 1 through presumably 7 3100-something. 8 Q. And Column B is the county -- county name, 9 correct? 10 A. County, city, or parish. 11 Q. County, city, or parish, thank you for the 12 clarification. 13 And Column C is the state name, correct? 14 A. Correct. I'm working from the first page. 15 Q. The state abbreviation, correct? 16 A. Correct. 17 Q. And then the Column D is actually the full state 18 name, correct? 19 A. Correct. 20 Q. Then we have column E which is called the FIPS 21 code. 22 Do you -- we may have to ask Dr. Stephenson, but 23 do you know what the FIPS code is? 24 A. It seems to be a code that identifies individual 25 counties. 26 THE COURT: And just for the record, would you say 27 the letters that comprise "FIPS"? 28 MR. ENGLISH: F-I-P-S. 6888 1 BY MR. ENGLISH: 2 Q. Okay. And those all came from the University of 3 Wisconsin model, correct? 4 A. Correct. 5 Q. They were delivered to National Milk as a -- 6 A. They correspond to what's currently listed in 7 terms of identifying county, cities, and parishes in 8 paragraph 1000.52. 9 Q. And then Column F is the model result for the 10 spring, or May of 2021, correct? 11 A. Yes. 12 Q. And this is the result of the third iteration, 13 correct? 14 A. I believe so. 15 Q. And so Column F came from the University of 16 Wisconsin, correct? 17 A. Correct. But you will need to direct that 18 question also to Dr. Nicholson to confirm. 19 Q. And I -- I have a cross-examination for him. So, 20 yes. Thank you, though. 21 But -- but you -- your understanding is that 22 Column F came from the University of Wisconsin model? 23 A. This is the way we received the model results. I 24 cannot confirm every single number in there. But I -- I 25 assume that this is -- if it came from the website, I 26 assume this is the correct final model results. 27 Q. Okay. And then we have Column G, which is the 28 equivalent of Column F, but this time, however, it's the 6889 1 fall or October 2021 University of Wisconsin model result, 2 correct? 3 A. Correct. Correct. 4 Q. Okay. Am I correct that once we get past 5 Column G, everything else on columns -- Exhibits 300 and 6 301 were derived not from the University of Wisconsin 7 directly, but from National Milk? 8 A. It appears to be so. You are taking differences 9 between the May and October results, you are taking 10 differences between the May results and current, and same 11 thing with October. 12 Q. So to be clear, if we just put, you know, 13 something over the document, you know, everything left of 14 the line between G and H came from the University of 15 Wisconsin, correct? 16 A. Specifically E -- excuse me -- F, G really were 17 the main things that came from the model. 18 Q. Okay. 19 A. Everything else, you know, the model results 20 included a lot of these calculations, but the guts of what 21 came from the model are Columns F and G. 22 Q. Okay. And I guess what you are saying is 23 Columns A through E are basically effectively lining up 24 with the Federal Order language? 25 A. Labels. 26 Q. Labels, okay. 27 A. Yes. 28 Q. And then -- but everything to the right, so to 6890 1 speak, so Columns -- on 300, Columns H, I, J, K, L, M, O, 2 P, Q, R, S, were added by National Milk, correct? 3 A. I assume so, because they are -- just knowing how 4 spreadsheets work, these look like they are fairly simple 5 calculations from Columns E, F, and the current 6 differentials, and basically what's currently in 7 paragraph 1000.52. 8 Q. And so maybe this would be the better way to ask 9 the question. 10 Dr. Nicholson did not provide the information in 11 those calculations done in Columns H through S, correct? 12 A. I don't recall exactly what -- what -- what was -- 13 there were some calculations that the -- Dr. Nicholson's 14 provided, just as output. But these were -- in all cases, 15 those are simple comparisons, very simple calculations. 16 And anybody -- whoever made them, they were pretty 17 straightforward calculations. 18 Q. So while you and I may believe they are 19 straightforward calculations, for purposes of the record, 20 let's see if we can quickly go through. 21 So Column H is labeled October to May differences. 22 So what is that, exactly? 23 A. That is a difference between the numbers on each 24 line and from -- between Columns G and Column F. 25 Q. And Column I labeled current differential at -- is 26 basically if you go to part 1000.50 adjusted for the 27 Southeast in 51, that's the current differential, correct? 28 A. Yes. Adjusted for the Southeast, yes. 6891 1 Q. And then Column J says May-current. 2 What is May-current in Column J? 3 A. That's the difference between the number in 4 Column A on each line and the number in Column I. 5 Q. I'm sorry, did I hear you say A or did you mean -- 6 A. Excuse me, F. Column F and Column I. 7 Q. And then -- so K would be the difference between 8 Column G and Column I? 9 A. That's correct. 10 Q. Okay. And then Column L, what is Column L? 11 A. Column L should be the average of the numbers in 12 Column F and Column G. 13 Q. And then so Column M is the difference between 14 Column L and Column J? 15 A. Column -- Column I. 16 Q. I. Thank you. 17 THE COURT: State again in one sentence what it's 18 the difference of? 19 MR. ENGLISH: Thank you. 20 BY MR. ENGLISH: 21 Q. Column M is the difference between Column L and 22 Column I, correct? 23 A. That's correct. 24 Q. And Column N is certainly the Federal Order number 25 where the county is located, correct? 26 A. Yes. If the county is located in the marketing 27 area of the Federal Order, that Federal Order number is 28 given in Column N. 6892 1 Q. Okay. And then Column O is proposed Class I, 2 correct? 3 A. Yes. That is the -- that is the proposed number 4 that was in Proposal Number 19. 5 Q. So I'm a little confused. Column O is labeled 6 Proposed Class I, and Column S is New Proposal. 7 How are Column O and Column S different, if you 8 know? 9 A. Can you repeat that? 10 Q. So I'm looking at Column O, which is labeled 11 Proposed Class I, and then I look over at Column S, where 12 the label is New Proposal. And I don't know if they are 13 duplicative or not. 14 Can you explain why there are two columns and 15 whether or not they are the same or different, if you 16 know? 17 A. No, I don't know. They appear to be the same. 18 Q. Then Column P is proposed versus current, which 19 would be, I believe, Column O minus Column I, correct? 20 A. That's correct. 21 Q. And then Column Q is proposed versus -- so it says 22 proposed versus model average, which I take it would be Q 23 minus L; is that correct? 24 A. Yes. 25 Q. And then there's a column labeled R, average 26 monthly pounds, 2022, in millions. 27 Can you please explain that? 28 A. I would assume that that is the average monthly 6893 1 pounds that the -- that the model had assigned to each of 2 those individual counties, cities, and parish. But that's 3 a question, again, for Dr. Nicholson. 4 Q. Well, are you sure it's for Dr. Nicholson? 5 Because I don't know if he provided that data or you did. 6 A. I don't recall that we went through and -- and 7 interpolated the more aggregated numbers that were 8 available for the pounds of milk. I assume that refers to 9 pounds of milk. 10 Q. Okay. 11 A. I'm not aware that National Milk did a 12 disaggregation to the county, city, parish level for all 13 3100-plus counties. 14 Q. Okay. 15 A. My sense is this calculation was made by somebody 16 else, but that can be clarified -- that can be clarified 17 if you ask it to enough of our witnesses. 18 Q. Thank you. 19 THE COURT: Mr. English, I want you to go back to 20 Column Q and again ask the witness how that is calculated. 21 MR. ENGLISH: I believe, but the witness can 22 correct me, that Column Q is Column P minus Column L. 23 THE WITNESS: Minus Column? 24 BY MR. ENGLISH: 25 Q. L. 26 A. Yes. 27 Q. Am I right? 28 A. That's correct. 6894 1 Q. Okay. I'm sorry, it's Column O. I apologize, 2 it's Column O -- 3 A. Column O minus Column L. 4 MR. ENGLISH: I'm not sure how many times I'm 5 going to get that wrong, Your Honor, so let me try it 6 again. And I thank my extremely helpful colleague. 7 BY MR. ENGLISH: 8 Q. So Column Q is Column O, labeled Proposed Class I, 9 minus Column L, which is labeled UofW v3 -- for I think 10 iteration 3 -- average. 11 Would that be correct? 12 A. That's correct. So, for example, that very first 13 line of Autauga County, Alabama, that indicates that the 14 changes made to the final model results resulted in a 15 lowering of the differential in Autauga County, Alabama by 16 $0.20. 17 MR. ENGLISH: May I consult with my colleague for 18 one moment, Your Honor? 19 THE COURT: Certainly, yes. 20 Let's go off record. It's 11:00 a.m. 21 (An off-the-record discussion took place.) 22 THE COURT: And let's go back on the record. It's 23 still 11:00 a.m. 24 THE WITNESS: Time is standing still. 25 BY MR. ENGLISH: 26 Q. So as it happens, I needed a tiny bit of help from 27 my consultant, and I probably should have known myself, 28 given the fact that I am from the Commonwealth of 6895 1 Virginia. 2 (Court Reporter clarification.) 3 BY MR. ENGLISH: 4 Q. So I think, again, for the benefit of the record, 5 when we turn -- because Your Honor noted color, but I 6 think there's some modifications. 7 Pages 49, 50, and 51, do have some additional 8 color, not blue, but yellow or orange. And I believe 9 you're closely enough connected to the Washington, D.C., 10 Metropolitan Area that you can probably understand where 11 I'm going with this. 12 A. I don't have a color copy. 13 THE COURT: The witness should have a color copy. 14 MR. ENGLISH: Can I hand it to him for a moment, 15 Your Honor? 16 THE COURT: No, I'm going to exchange. I'm going 17 to take what he's got. 18 THE WITNESS: If you are going to ask me a 19 question about colors, I need to see what the colors are. 20 MR. ENGLISH: I apologize. 21 THE COURT: And I don't -- I can follow along 22 without them, and I have not marked them in any way. 23 Thank you so much. 24 THE WITNESS: Thank you. 25 BY MR. ENGLISH: 26 Q. So if we look -- let's just start with 49, and 27 I'll try to keep this really short. 28 THE COURT: But don't go fast. 6896 1 MR. ENGLISH: Point taken, Your Honor. 2 BY MR. ENGLISH: 3 Q. So Virginia is a jurisdiction where there are 4 cities that -- and counties, and sometimes cities are 5 inside counties, correct, Doctor? 6 A. Virginia has cities and counties. I'm not sure 7 that the cities are incorporated in the counties or 8 whether they are separate. 9 Q. Okay. 10 A. Like, I live next door to City Falls Church and 11 Fairfax County, but I do believe that those are separate. 12 Q. They can be separate. But you can live in Falls 13 Church for Postal Service purposes, and yet be in Fairfax 14 County, correct? 15 A. I'm not aware of exactly what's the territory of 16 Fairfax County and whether it incorporates it. 17 MR. ENGLISH: Your Honor, maybe I'll just shorten 18 it. I don't think there's any controversy here. 19 THE WITNESS: If you know the facts, I will accept 20 your word. 21 BY MR. ENGLISH: 22 Q. I want to make a representation for the record. 23 Having, you know, gone to the University of 24 Virginia, which is in Charlottesville, Virginia, and 25 Charlottesville, Virginia is inside Albemarle County, and, 26 in fact, the Albemarle County courthouse is across the 27 street from the Charlottesville courthouse. 28 A. I will take your word for it to speed things 6897 1 along. 2 Q. And similarly, as I look through what are marked 3 as yellow and orange, every single one of these is an 4 instance where there's a -- there's a city that is located 5 in or connected to a county. And I also grew up in Falls 6 Church, but I grew up in the part of Falls Church that is 7 part of Fairfax County. 8 And so I think to simplify the conversation, it 9 would just say that when we see these yellow and oranges, 10 they are not significant in any material way because they 11 just describe a peculiarity of the Commonwealth of 12 Virginia. 13 MR. HILL: Mr. English? For those following 14 electronically, could you tell us what line you are on 15 rather than the page number -- 16 MR. ENGLISH: Okay. 17 MR. HILL: -- to the left. 18 MR. ENGLISH: Okay. So I'll go through the line 19 numbers: 2790 is Alexandria City; 2801 is Bristol City; 20 2805 is Buena Vista City; 2811 is Charlottesville; 2812 is 21 Chesapeake City; 2815 is Colonial Heights; 2816 is 22 Covington; 2820 is Danville City; 2823 is Emporia. 23 THE COURT: Now, let me stop you. Even with my 24 black and white copy, I can see the highlighting of 25 everything that you are reading, so I don't think you need 26 to read them all, but -- 27 MR. ENGLISH: I'm fine stopping, if that's okay. 28 MR. HILL: It's just that online it doesn't have 6898 1 the page numbers. I just needed to know where you were. 2 MR. ENGLISH: Okay. So it's Virginia, starting at 3 2789, which is Albemarle, and it runs through the end of 4 Virginia, which is line number 2920, York. 5 MR. HILL: Thank you very much. 6 MR. ENGLISH: Okay. Thank you, sir. I'm happy 7 not to read them all in. 8 THE COURT: And is the Commonwealth of Virginia 9 the only batch of lines that has this polarity? 10 MR. ENGLISH: Looking through it very quickly, 11 Your Honor, yes. 12 THE COURT: Makes you proud, doesn't it? 13 MR. ENGLISH: Wahoo-wa. 14 BY MR. ENGLISH: 15 Q. All right. Okay. So I am really not going to 16 spend a lot more time on all this, but I am going to try 17 to clarify. 18 So let's start with -- let's turn to Exhibit 301. 19 A. Do you want me to keep the pages that you just 20 referred to on number 300 or are we done with those? 21 Q. We can give those back to the judge and switch if 22 you want. 23 THE COURT: No, no, no, I want the witness to keep 24 that. Thank you. 25 MR. ENGLISH: I just remembered the answer. All 26 right. 27 BY MR. ENGLISH: 28 Q. So we turn to 301, first it's labeled at the top, 6899 1 June 2023, at the very top in the header. 2 Are you in 301? 3 A. Yes, I am. 4 Q. Okay. And so would this refresh your recollection 5 that if it's labeled June 2023, it was probably submitted 6 in June 2023? 7 A. I see it listed as such, yes. 8 Q. I should have started there. So this -- this is 9 different in one, at least to me, obvious respect, which 10 is that through Columns A through Column N appears to be 11 identical to Columns A through Columns N on Exhibit 300. 12 Do you agree? 13 A. It would appear such. 14 Q. Okay. And then it appears that this omits 15 Exhibit 300, Columns O, P, Q, R, and substitutes with a 16 caveat Column S for Column O. And I'll come back to the 17 caveat in a second. 18 Is that -- is that correct that -- that maybe 19 there's some differences between S and O in between 300 20 and 301, but in essence, you have fewer columns and you 21 have omitted O, P, Q, R from Exhibit 300? 22 A. Columns O, P, Q, and R are basically -- Column O 23 is significant because that's the final results on 24 Exhibit 30, the rest are just calculations. But I cannot 25 testify exactly what -- in Exhibit 300 is -- Column S is 26 labeled as New Proposal and Column O is listed as 27 Proposal. 28 Q. Okay. And do you know -- if you don't, that's 6900 1 fine -- whether there's any differences between the 2 numbers that appear in Column S and the numbers that 3 appear in -- Column S, Exhibit 300, and Exhibit 301, 4 Column O? 5 A. I do not know the answer to that question, because 6 I have not had time to go through and compare them line by 7 line. 8 THE COURT: And, Dr. Vitaliano, you mentioned 9 Exhibit 30, and you were looking at 300 at the time. 10 THE WITNESS: 300 and 301. 11 THE COURT: Thank you. 12 MR. ENGLISH: Your Honor, I only have a few more 13 questions, but for housekeeping, I move the admission of 14 Exhibits 300 and 301 having laid, I think, a sufficient 15 foundation. 16 THE COURT: Is there any objection to the 17 admission into evidence of Exhibit 300? 18 There is none. Exhibit 300 is admitted into 19 evidence. 20 (Exhibit Number 300 was received into 21 evidence.) 22 THE COURT: Is there any objection to the 23 admission into evidence of Exhibit 301? 301? 24 There is none. Exhibit 301 is admitted into 25 evidence. 26 (Exhibit Number 301 was received into 27 evidence.) 28 /// 6901 1 BY MR. ENGLISH: 2 Q. So keeping the two exhibits handy -- and I want to 3 go to -- 4 THE COURT: Be closer to the mic. 5 MR. ENGLISH: Thank you. It's a little hard, the 6 size of the documents, but thank you. 7 I want to go to FIPS, F-I-P-S, Column E, code 8 27053, which is on page 23 for those in the hearing room. 9 THE COURT: The line number? 10 MR. ENGLISH: I'm getting there. 11 THE COURT: Oh, okay. 12 MR. ENGLISH: I'm sorry, I had it and then I lost 13 it. 14 27053 is -- so it's line number 1307 ironically 15 under ID, because there's one number off, it's 1307, which 16 is Hennepin County, Minnesota, otherwise known as 17 Minneapolis. 18 THE WITNESS: I see that as 1308. 19 BY MR. ENGLISH: 20 Q. I understand it's 1308 on the line number. So 21 let's just omit Column A for this, because the way the 22 line numbers work -- 23 A. Okay. And they are sequential. 24 Q. They are off by precisely one between the ID and 25 the line number. 26 So let's use line number 1308. And the Column E 27 is FIPS Code 27053, and it's Hennepin County, Minnesota. 28 And I note, you would agree, that the model spring 6902 1 is 2.60 in Column F, correct? 2 A. Correct. 3 Q. And for Row 1308, Column G, the October number in 4 Column G is 2.70, correct? 5 A. Correct. 6 Q. And -- 7 THE COURT: Now, do you want the transcript to 8 show $2.60? 9 MR. ENGLISH: Yes, $2.60 and $2.70. All of these 10 numbers are in dollars. I will try to remember to say 11 that. 12 THE COURT: Thank you. 13 BY MR. ENGLISH: 14 Q. And if we go over to Column L, which is labeled 15 University of Wisconsin version 3, average, the average is 16 2.65, correct? 17 A. Correct. 18 Q. And if we go all the way over to the right in 19 Column S, the proposal, under 2 -- new proposal, it's 20 2.80, correct? 21 A. Correct. 22 Q. Now let's go to Exhibit 3. So let's remember 23 that's 2.80 from Exhibit 300. 24 So let's please go to Exhibit 301, the June 25 submission, same FIPS code. So I will repeat, 27053, 26 page 23 for those who have a copy here, line 1308, 27 Hennepin County. And I would ask you to go all the way 28 over to the right in Column O, and you see $3, correct? 6903 1 A. That's correct. 2 Q. So what changed, in National Milk's view, from the 3 May submission to the June petition where Proposal 19 went 4 to $3, which is $0.20 higher than was submitted in May? 5 A. Well, to get the definitive answers, again, you 6 need to direct that question to the witness who will 7 testify in that area. But it was an iterative process. 8 Anybody who's done this kind of analysis knows that you 9 don't often get the perfect number the first time. You 10 need to double-check, you need to look at a number of 11 things, and, you know, those things will change. At some 12 point you have to say this is final and -- and submit, in 13 this case, your final numbers in the form of a proposal. 14 And so that number changed by $0.20, given these 15 documents, and you'd have to ask the person who was more 16 directly involved what caused that change. 17 But as an analyst, there's nothing very surprising 18 about this process to me. 19 Q. Now, if we stay on Exhibit 301 in the same FIPS 20 code, for Hennepin County, you have a proposal for $3, and 21 a Column I, current differential, and $1.70. 22 So you would agree that you are proposing, in 23 Proposal 19, to increase the Class I differential in 24 Minneapolis by $1.30, correct? 25 A. Okay. Which exhibit are you in? 26 Q. I'm still in Exhibit 301. 27 A. What's the MIG number? 28 Q. 29. 6904 1 A. Okay. Are you still on Hennepin County? 2 Q. I'm still on Hennepin County. In Column I is the 3 current differential is $1.30; in Column O is $3. 4 And you would agree with me that that difference 5 is $1.30 higher, correct? 6 A. I see Hennepin County, Column I is $1.70. 7 Q. Right. And Column O is $3, correct? 8 A. Correct. 9 Q. And maybe the easiest thing we're going to do 10 today is you subtract $1.70 from $3 and you get to $1.30, 11 correct? 12 A. That's correct. 13 Q. And that's an increase, correct? 14 A. That's an increase in the number in the column 15 labeled Proposed from the current differential. 16 Q. Okay. Going back to our conversation before 17 either break, given the Class I utilization in the Upper 18 Midwest, which I think is around 5 to 8%, what is the 19 justification for increasing the Class I differential in 20 Minneapolis by $1.30? 21 A. The justification is basically the purpose of 22 price alignment. We had to look -- each county, 23 particularly counties with a -- with a city, or you know, 24 milk plants in them, had to be aligned with those from 25 other areas, and that was one of the overriding 26 considerations in coming up with our proposed 27 differentials. 28 Again, in terms of the specifics, you need to 6905 1 direct that question to the person who will be testifying 2 specifically in that region that includes Hennepin County, 3 Minnesota. 4 Q. And I am reminded that maybe I have been imprecise 5 in my questions, so let me backtrack for one moment. 6 Is it your understanding, if we look at MIG-29, 7 which is what you have, which has been marked as 8 Exhibit 301, that except for the two changes you told us 9 about earlier today, what is found in Column O is -- in 10 your understanding, is what is NMPF-19? 11 A. NMPF Proposal 19? 12 Q. Yes. 13 A. I'd have to double check that to give you an 14 affirmative answer, but I have no reason to question why 15 that is not the case. I don't have -- I don't have the 16 numbers in Proposal 19 as I submitted them in front of me. 17 Q. Okay. So I want to turn next, I want you to 18 remember what we did in Minneapolis with the $1.30 19 increase, which, by the way -- let me go back. Let me 20 strike that. 21 So I want to go to FIPS code 12086, which is 22 Miami-Dade, Florida. 23 THE COURT: Which is what? 24 MR. ENGLISH: Miami-Dade, Florida. 25 MR. HILL: Line number? 26 MR. ENGLISH: I'm getting there. 12086 for those 27 in the room. It's on page 6, and 12086 is line 28 number 335. 6906 1 MR. HILL: Thank you. 2 BY MR. ENGLISH: 3 Q. I'm looking on Column F, FIPS code. And I'm only 4 looking at Exhibit 301, which is Exhibit MIG-29. 5 And I want to walk you through -- so under 6 Column F, from the University of Wisconsin, for May we 7 have $7.40? 8 A. Yes. 9 Q. And for Column G, we have $8.40? 10 A. Correct. 11 Q. And then the average under Column L, University of 12 Wisconsin average, is $7.90, correct? 13 A. Correct. 14 Q. And Column O is also $7.90, correct? 15 A. Correct. 16 Q. And so for all of this conversation in this 17 hearing about the need for more milk in the Southeast, if 18 you look at Column M, you are increasing the Class I 19 differential in Miami by $1.90, correct? 20 A. That's correct. 21 Q. As compared to raising Minneapolis by $1.30, 22 correct? 23 A. Correct. 24 Q. Why is it that for the greatest, as you said in 25 your testimony, the milk that needs the milk to -- the 26 county that needs the milk to move the farthest from the 27 farthest reserve supply, you used the average in Column O, 28 but for Minneapolis you use a number higher than either 6907 1 Column F or Column G or the average for Minneapolis? 2 A. We felt that the model results for Miami-Dade were 3 adequate for the purpose of price alignment, all of the 4 purposes we looked at for which we commissioned the model 5 and made adjustments to it. We chose not to make 6 adjustments to the model results for Miami-Dade. We chose 7 to do those for Hennepin County, Minnesota. 8 Q. Wouldn't it, if we need to move milk to Florida, 9 make more sense to increase that spread as opposed to 10 decrease that spread? 11 A. That question would spring from a much, much 12 simpler understanding of the whole process. And, again, I 13 then -- then we used -- and I would recommend you direct 14 that question to Dr. Nicholson first and -- to speak for 15 the model, and to the person who is -- to the people who 16 are going to testify on those two different regions for 17 the modifications that National Milk made to the model 18 results. They will give you much better answers to those 19 questions. 20 Q. But we already said that Dr. Nicholson didn't 21 calculate the average or any of these columns included in 22 the proposal, correct? 23 A. What average are you referring to? 24 Q. Column L, University of Wisconsin didn't provide 25 you Column L, did it? 26 A. Probably not. But they provided the two numbers 27 that we decided to use as the starting point in the 28 process of making further adjustments to the model 6908 1 results. We did not necessarily use the average in all 2 cases, but we calculated the average as a starting point. 3 For the primary purpose that our eventual proposal had to 4 have one number, the Class I differentials listed in the 5 Federal Order regulations have a single number, they do 6 not permit us to use seasonal numbers. And we saw no 7 reason to recommend further disaggregating 8 paragraph 1000.52 into regional -- or excuse me -- 9 seasonal -- seasonal parts. 10 Q. I want you to turn to FIPS, F-I-P-S, 80 -- I'm 11 sorry, whoa, 08031, which is Denver, Colorado, and I will 12 give page and line as soon as I have it. 13 The very bottom on page 4, line number 233. 14 A. Are we still just on Exhibit 301? 15 Q. Yeah, we'll stick -- unless I say otherwise -- and 16 thank you very much. Yes. I -- let's -- I -- I think 17 that -- I only have a very few more of these, but I think 18 from now on we're not going to look at the fact that there 19 was a change, we're just going to look at 301. 20 So if we look at 301, and line 233, Denver, you 21 would agree that Columns F and G are identical at $2.50, 22 correct? 23 A. Correct. 24 Q. And if you look at Column I, the current 25 differential is $2.55. And so if you look at Column J -- 26 I'm sorry -- well, yeah, Column J or Column K, the -- it 27 actually is down $0.05, correct? 28 A. Correct. 6909 1 Q. And the average in Column M is down $0.05 on the 2 average, correct? 3 A. That's correct. 4 Q. So how did we get from a model number that went 5 down to 2.50 to an $0.80 increase to $3.30 in Column L? 6 A. As I explained, the model results were very 7 accurate in many cases, very, very, very close. If you go 8 through and look at those differences, you will see that 9 they are generally pretty modest, but -- but in some 10 cases, based upon the institutional knowledge of our -- 11 the members of our task force that were looking with 12 expertise in those regions, we chose to make a change. 13 And, in general, those changes were -- were relatively 14 modest, but were not in all cases modest, and you will 15 have to direct that question to the witness that speaks to 16 the changes made to the Colorado numbers. 17 Q. So this allows me, I think, to ask and hopefully 18 get an answer to a question that had puzzled me for a 19 while until Ms. Keefe helped me understand it. 20 To the extent that a base price increase occurred 21 from $1.60 to 2.20, I don't see that directly reflected in 22 MIG-29, which is 301. I believe it appears for the first 23 time in what I think is a hard code in the Excel 24 spreadsheet in Column O. 25 Would I be correct? 26 A. Could you repeat that question? In other words, 27 you are asking about the $1.60 base differential? 28 Q. No, the 2.20. I'm asking about a change -- so you 6910 1 agree with me, you said earlier, that the model results 2 from the University of Wisconsin, so Columns F and G, were 3 run in each of the three iterations at $1.60, correct? 4 A. That's my understanding, yes. 5 Q. Okay. 6 A. But you will have to confirm that with 7 Dr. Nicholson, because we did not -- I don't recall that 8 we -- that we decided on the 2.20 until after the model 9 runs were made. 10 Q. And my point is, am I correct that there's no 11 column that delineates a change in a base price from $1.60 12 to 2.20? There's just no column that says, here it is. 13 A. I don't see one in these documents. 14 Q. And so would I be right that the place you need to 15 look at in order to find and maybe then backwards derive 16 what the base price increase is would be Column O? 17 A. Column O was the final number. And I'd have to 18 confirm that with looking at my list of these counties, 19 cities, and parish numbers. But my understanding is that 20 our final numbers would include the 2.20. But, yes, that 21 would have been because it was June 2023. 22 Q. So it does, doesn't it? It must include -- 23 A. Sure. 24 Q. Yes, correct? 25 A. It must include the 2.20, yes. Again, I'm seeing 26 all these numbers for the first time in this particular 27 spreadsheet format. 28 Q. But you, National Milk submitted it, correct? 6911 1 A. Yes, it's labeled National Milk. 2 Q. Okay. 3 A. I have no reason to doubt that that's those 4 numbers. 5 Q. All right. I only have two more of these for you. 6 Let's go to FIPS code 48453, which is Travis 7 County, Texas, also known as city of Austin. And as you 8 say, there's a lot of pages. So I think we're on page 47 9 for those in the room, and we're looking at line 2717, 10 which is Travis, Texas, and I represent to you that it is 11 Austin. 12 And if we do what we have done before exhibit -- 13 and that's MIG-29, 301 -- if we look at Column F for 14 Travis, it was a $4 from the run for May, 4.20 for the run 15 for October, and back under Column L it's 4.10 for an 16 average, correct? 17 A. That's correct. 18 Q. And before your counsel made corrections, or you 19 made corrections with your counsel, the number under 20 Column O for Travis County is $4.70, correct? 21 A. It is listed as -- as such in Exhibit 301. 22 Q. And you have corrected it to 4.35, correct? 23 A. That's correct. 24 Q. Can you explain what happened between June and 25 now, if you know? 26 A. No. I assume that -- no, I don't know. I do not 27 know the reason for that correction. There were only two 28 such corrections out of all of these numbers, so it's not 6912 1 surprising that some further examination determined that 2 further adjustments those two corrections needed to be 3 made. 4 Q. And finally, let us turn to FIPS code 06065, 5 Riverside, California, which is another state with a lot 6 of counties but not as many as Texas, I think. 7 So page 4, line 191, FIPS code 6065, and we see 8 for Column F, $2.30, for Column G, $2.50, correct? 9 A. Correct. 10 Q. The model average under Column L is $2.40, and the 11 proposal under Column O is $3, correct? 12 A. Correct. 13 Q. Which is $0.60 higher than the average, correct? 14 A. Correct. 15 Q. So California has a lot of milk, doesn't it? 16 A. In parts of the state, yes. 17 Q. So why is Riverside, California $0.60 higher than 18 the model average; Denver $0.80 higher than the model 19 average; and Miami, Florida, which we have heard a lot 20 about for being the biggest deficit, the model average? 21 A. Well, your question derives from probably a 22 somewhat too simplistic understanding of what the whole 23 process was. But I would again direct you to direct those 24 questions to the witnesses that are going to be testifying 25 specifically to the changes in those regions. 26 Q. Is it too simplistic to think that Miami should 27 have gone up the most? 28 A. Not necessarily. 6913 1 MR. ENGLISH: Your Honor, I am going to go to a 2 different section. I am mindful, I think, of certain 3 travel plans. I don't know whether we need to have a 4 conversation, and so maybe we need to confab, but my 5 understanding is that Dr. Nicholson needs to be done -- 6 done today, and if it's -- 7 THE WITNESS: Same here. 8 MR. ENGLISH: Well, okay. I will let National 9 Milk decide. I mean, if -- I have more to go. I am more 10 than halfway done, but not two-thirds. On the other hand, 11 this was probably the longest section. So I just want to 12 be courteous and try to give National Milk an opportunity 13 to figure out what they want to do. 14 THE COURT: Ms. Hancock, are you able to talk to 15 us now about a proposal as to how we proceed with those 16 two witnesses? 17 MS. HANCOCK: Yes, Your Honor. We have -- 18 Dr. Vitaliano will be back next week. He does have to 19 leave today for another commitment that he has, but 20 Dr. Nicholson also needs to be done today, if possible, 21 and he won't be back. 22 So I think if everybody is okay, we could put 23 Dr. Nicholson's primary testimony on before lunch, and 24 then after we return from lunch, have his 25 cross-examination conducted, and then Dr. Vitaliano could 26 pick back up next week when he returns. 27 THE COURT: Let me first ask, Dr. Vitaliano, how 28 does that sound to you? 6914 1 THE WITNESS: That sounds fine. I can be here 2 until -- I have a flight at 6:15 from this airport. 3 THE COURT: So you possibly could be recalled 4 today, but it's kind of unlikely. 5 THE WITNESS: Depend -- given the length of the 6 cross-examination, particularly of these key witnesses, I 7 would guess that if Dr. Nicholson goes on -- which I'm 8 happy to yield my time to him -- he will be kept occupied 9 until he has to leave for his flight. 10 THE COURT: You yield back to the gentleman from 11 where? 12 THE WITNESS: Wisconsin, sorry. 13 THE COURT: That sounds the smartest. Do you 14 agree, Ms. Hancock, just to have Dr. Vitaliano be 15 interrupted now to be resumed next week? 16 MS. HANCOCK: I think that's fine. And further 17 optimism, maybe that will help truncate some of his 18 examination. 19 MR. ENGLISH: Your Honor, first of all, I'm happy 20 to yield. I think what happens when you are the first 21 witness, you don't know who else is going to say things. 22 And I actually agree. I think it may very well be the 23 case that if -- assuming Mr. Sims gets on and off, and 24 Mr. Erba gets on and off, you know, I may have fewer 25 questions. So I do think it would make sense. 26 Besides which, I think we have routinely in this 27 proceeding recognized that the -- you know, the non-member 28 witnesses, like Dr. Nicholson and others, should have some 6915 1 priority. Not quite the same priority as dairy farmers, 2 but I think next up. 3 So I am prepared to mark where I am. And I do 4 promise, I really do, that if I get the answers to the 5 questions before he gets back on, I will subtract them. 6 All right? 7 THE COURT: If you get the answers to the 8 questions what? 9 MR. ENGLISH: The questions that appear on 10 pages 11 through 17 of my cross-examination, having 11 finished 10, if I get those answers prior to his coming 12 back on -- the reason I'm asking him is he's the National 13 Milk witness, he's the first witness. I don't know for a 14 fact what other people are going to say. If I get answers 15 to questions that are otherwise posed for him, I will not 16 duplicate him. I make that assurance for everybody. 17 THE COURT: Very good. Well, let's take a 18 ten-minute break while everyone repositions. Is that a 19 good idea? No? Well, yeah, that's a good idea. 20 We can at least -- if we take, what, you want a 21 five-minute break? Can you be ready? 22 Okay. We'll take a five-minute break now. Please 23 be back and ready to go at 11:46. 24 We go off record at 11:41. 25 (Whereupon, a break was taken.) 26 THE COURT: Let's go back on record. 27 We're back on record at 11:46 a.m. 28 Would you state and spell your name for us, 6916 1 please? 2 THE WITNESS: My name is Charles Nicholson, 3 C-H-A-R-L-E-S; Nicholson, N-I-C-H-O-L-S-O-N. 4 THE COURT: Thank you. I'd like to swear you in. 5 Would you raise your right hand, please. 6 CHARLES NICHOLSON, 7 Being first duly sworn, was examined and 8 testified as follows: 9 THE COURT: Thank you. 10 DIRECT EXAMINATION 11 BY MS. HANCOCK: 12 Q. Good morning, Dr. Nicholson. Thank you for being 13 here. Did you just provide your address? 14 THE COURT: I did not ask. 15 BY MS. HANCOCK: 16 Q. Sorry, for some reason I couldn't remember if you 17 just did. Could you provide your business address, 18 please? 19 A. My business address is 1675 Observatory Drive, 20 Madison, Wisconsin, 53706. 21 Q. Thank you. 22 And did you prepare Exhibits 36 and 36 -- well, 23 did you prepare Exhibits NMPF-36 and 36A in support of 24 your testimony today? 25 A. Yes, I did. 26 Q. Okay. And is Exhibit NMPF-36, is that the full 27 and complete written testimony that you have provided? 28 A. Yes, it is. 6917 1 Q. And is 36A a summary that you are intending to put 2 into the record today in support of the full testimony? 3 A. Yes, it is. 4 MS. HANCOCK: Your Honor, if we could give those 5 Exhibit Numbers 302 for NMPF-36 and 303 for NMPF-36A? 6 THE COURT: We shall. Thank you. 7 (Exhibit Numbers 302 and 303 were marked for 8 identification.) 9 BY MS. HANCOCK: 10 Q. Before we turn to your statements, I'm wondering 11 if you can provide an overview of your educational 12 background. 13 A. Okay. I have a bachelor's degree in economics and 14 statistics from the University of California at Davis. I 15 have a master of science degree in agricultural economics 16 from Cornell University. And I have a Ph.D. in 17 agricultural resource and managerial economics, also from 18 Cornell University. 19 Q. And can you give us an overview of your 20 professional career? 21 A. So I have, post-Ph.D., now experience going on 22 close to 30 years. Much of it has been devoted to 23 economic analysis of dairy industry issues, both in the 24 United States and globally. 25 Q. And we heard yesterday about a group or a kind of 26 a brain trust of agricultural economists. 27 Do you belong to that group as well? 28 A. Yes. So that is what is now known as the Program 6918 1 on Dairy Markets and Policy, primarily led out of the 2 University of Wisconsin. Prior to that, when it was based 3 at Cornell University, it was known as the Cornell Program 4 on Dairy Markets and Policy, and it was a group of 5 academics who met to discuss dairy industry issues and 6 offered an annual workshop for dairy economists and policy 7 an analysts. 8 MS. HANCOCK: And, Your Honor, at this time we 9 would offer Dr. Nicholson as an expert in -- as a dairy 10 economist. I should expand that for all the other areas 11 he's testified to as well but primarily for our 12 purposes -- 13 THE COURT: I will write them all down, so go 14 ahead and say what else. Dairy economist? 15 MS. HANCOCK: And applied economics. 16 THE COURT: Applied economics? 17 MS. HANCOCK: Or agricultural and applied 18 economics. 19 THE COURT: And any others? 20 MS. HANCOCK: Any others you would like to be 21 characterized as an expert for? 22 THE WITNESS: I would actually say that supply 23 chain management would be an area of expertise. 24 THE COURT: Good. 25 Does anyone like to -- would anyone like to voir 26 dire Dr. Nicholson about his qualifications to be accepted 27 as an expert witness in the areas of dairy economist, 28 applied -- no, agriculture and applied economics, and 6919 1 supply chain management? Is there any objection to my 2 accepting him as an expert in those three fields? 3 There is none. I accept Dr. Nicholson as an 4 expert witness in those three areas, as a dairy economist, 5 as an agricultural and applied economist, and as a supply 6 chain management expert. 7 MS. HANCOCK: Thank you, Your Honor. 8 BY MS. HANCOCK: 9 Q. Dr. Nicholson, would you proceed in providing us 10 your testimony? 11 A. Yes. Thank you very much. 12 So, Judge Clifton and personnel of AMS Dairy 13 Programs, I am appearing before you to offer a summary -- 14 Q. And I'm just going to interrupt you really quick. 15 We have a court reporter who is taking down everything 16 that we say, and so if you could read at a much more 17 moderated pace, that will help ensure that she captures 18 everything. 19 A. Thank you for that. 20 I am appearing before you to offer a summary of my 21 written prepared statement describing in more detail the 22 results of a recent research project that analyzed 23 differences in the spatial values of milk in the 24 contiguous United States, in particular the spatial 25 differences in values at fluid milk processing plants. 26 I'm an agricultural economist with more than 30 years of 27 experience in the analysis of dairy markets, including the 28 spatial evaluation of milk values. 6920 1 Importantly, I am not here to advocate for any 2 specific policy action, but rather to offer my insights 3 into the spatial differences in the economic values of 4 milk. This is a summary of research performed in 5 collaboration with Dr. Mark Stephenson, who recently 6 retired as the director of Dairy Policy Analysis at the 7 University of Wisconsin, Madison, but also does not 8 represent an official statement of the University of 9 Wisconsin, Madison. 10 The analyses that I will report are based on 11 spatial economic models that have a long history of 12 development, beginning in the 1980s at Cornell University. 13 Earlier versions of these models have provided evidence 14 about spatial milk values for previous Federal Milk 15 Marketing Order hearings, notably in 1998. 16 For the past 20 years, I have been the lead 17 researcher responsible for the further development and 18 updating of data for these detailed spatial economic 19 models, again, in collaboration with my former Cornell and 20 UW colleague, Dr. Mark Stephenson. Analyses based on 21 these models have appeared in refereed academic journal 22 articles -- a number are cited in footnotes -- and book 23 chapters -- again, cited in footnotes -- and have been 24 used by state government and industry groups to support 25 investment decisions. 26 A summary of the key findings of this research is 27 as follows: 28 1. Analysis with a detailed spatial economic 6921 1 dairy supply chain model that accounts for all sources and 2 uses of milk and dairy components, provides 3 location-specific milk values consistent with the lowest 4 possible systemwide costs, providing a competitive 5 benchmark for those values; 6 2. The analyses suggests that there are 7 considerable differences between the values of milk at 8 fluid plants derived from the spatial economic modeling 9 and the current values of Class I differentials, 10 differences as large as $3 per hundredweight; 11 3. These differences between current spatial 12 economic values at fluid milk plants and current Class I 13 differentials arise due to substantive changes over time 14 in the locations of milk production, the composition of 15 dairy product demand, changes in the location of demand 16 for dairy products given regional population shifts, and 17 the costs of transporting farm milk to plants, 18 transporting dairy products between plant locations, and 19 distributing products to final demand locations; 20 4. Review and adjustment of spatial values 21 generated by the model for the purposes of revising 22 Class I differentials are appropriate to account for local 23 circumstances and institutional factors not included in 24 the model analysis. Any quantitative model is, by 25 definition, a simplification of reality, and the USDSS 26 (U.S. Dairy Sector Simulator) does not directly represent 27 existing commercial relationships that can be important 28 determinations of the locations and volumes processed in 6922 1 existing operations. 2 I would now like to move to a description of the 3 U.S. Dairy Sector Simulator. 4 Spatial milk values are calculated using the U.S. 5 Dairy Sector Simulator. The USDSS is a highly-detailed 6 mathematical spatial optimization model, but at its core 7 it solves a practical problem, how to get milk from dairy 8 farms to plants to be processed into various dairy 9 products, and distribute those products to consumers with 10 the lowest cost possible. The model takes the total milk 11 supply, plant locations, and product mix, and consumer 12 demand as it existed for an individual month. It 13 indicates how to move that farm milk to plants via the 14 existing road network and distributes the finished 15 products to consumers, also according to the road network. 16 For the U.S. dairy industry as a whole, the USDSS 17 minimizes the systemwide cost of assembling milk at 18 plants, making final and intermediate dairy products, and 19 transporting them to other plants and locations of final 20 demand. The model includes the principal cost between the 21 farm gate and the retail locations for the consumer. The 22 model minimizes this total cost subject to the physical 23 constraints, such as mass balance and required product 24 composition that we have imposed upon the system. 25 The most recent spatial milk values derive from 26 two versions of the USDSS model: A large version with 27 data disaggregated at the county level, 3,108 counties, 28 and a smaller version with a few hundred multicounty 6923 1 regions. Both the large and small models yield similar 2 quantitative values and patterns of spatial milk prices. 3 Three, I'd like to talk about the USDSS model 4 outputs. 5 There are two types of results that are provided 6 by the USDSS. One is a primal solution, and the other is 7 a dual solution. The primal solution describes the 8 physical flows of product through the dairy supply chain 9 network. The dual solution represents the relative 10 monetary values of milk and dairy products at each model 11 location. 12 An example of the primal output from the smaller 13 USS -- USDSS model -- Figure 5 in the full written 14 testimony -- and now if we can go to the slide that should 15 be there -- is shown here. This shows milk assembly 16 flows, processing locations, and distribution flows to 17 final demand locations. The green lines represent milk 18 assembly flows from farms to plants, whereas the orange 19 lines represent the distribution of finished products from 20 plants to demand locations. The plants are shown as black 21 triangles. The size of the assembly and distribution 22 flows are represented by the relative thickness of the 23 lines, the green and orange lines. And the size of the 24 plant location triangles indicates the relative volume of 25 product processed at each plant. 26 And you will see that this figure is actually 27 showing the milk assembly at fluid plants and packaged 28 milk flows for May of 2021. 6924 1 The dual solution shows the spatial value of milk 2 or, more specifically, the marginal value of milk at a 3 processing location for a supply location for raw milk. 4 Thus, the dual values provide estimates of the spatial 5 value of milk and are the key results reported for the 6 purposes of this component of the hearing. 7 Dual values are calculated by the USDSS at all 8 milk plant locations across the country, although our 9 focus here is on the values for fluid milk processing 10 plants. This price surface indicates estimated spatial 11 values of milk for each county location in the contiguous 12 United States, consistent with the spatial aggregation 13 used for Class I differentials. However, the indicated 14 spatial milk values should not be interpreted directly as 15 Class I differentials. The values should be thought of as 16 price relatives, that is, the difference in values across 17 locations. 18 The Agricultural Marketing Service of USDA used 19 results from a previous version of the USDSS model results 20 as input into the 1998 Federal Order hearings. 21 Differences between the model-generated relative spatial 22 values of milk compared to those of the current Class I 23 differentials suggest a potential need to modify Class I 24 differentials. 25 Four, factors affecting the price relatives in the 26 USDSS model. 27 The USDSS shows the spatial milk values at a given 28 point in time, but it is also relevant to consider the 6925 1 drivers of changes in these values. Three factors 2 constitute the important causes of change in the spatial 3 milk values, the price relatives. These factors are 4 changes in the milk supply, demand for dairy products, and 5 transportation costs. 6 The detailed written statement describes the 7 substantive changes in the location of U.S. milk 8 production during the past decade. It also documents 9 changes in the product mix for U.S. industry and in the 10 locations of the population. Transportation costs have 11 changed over time due to the cost of purchase or lease of 12 the vehicle, driver wages and benefits, and fuel costs. 13 I'd now like to discuss specific results for the 14 spatial milk values at fluid milk plants. 15 The USDSS was simulated using both the smaller 16 multi-county and large county-level versions with 2021 17 data with similar quantitative results and patterns. The 18 models are run for the months of May 2021 and October '21, 19 to represent both the flush and the short months of the 20 year. 21 The general pattern is lower values in the north 22 and western regions, and rising into the south and eastern 23 areas of the U.S. The pattern of these values mirrors the 24 current Class I differential structure and reflects the 25 relative surplus and deficit regions of milk. However, 26 the current differentials range from $1.60 to a high of 27 $6, while the model suggests that the price surface is 28 steeper moving towards the Southeast, high values more 6926 1 than $7, reflecting both changing regional production and 2 demand and higher transportation costs. 3 Spatial milk values for October '21 have a pattern 4 similar to that in May 2021, but with the spatial values 5 in the Southeast indicating an even steeper price surface 6 and reaching a maximum value of more than $8. 7 The seasonal differences in value, which are 8 Figure 17 in the original full written testimony, indicate 9 a fairly steep rise in values from St. Louis through 10 Atlanta, and down to Miami, along the I-75 corridor. The 11 western portions of the U.S. show very few seasonal 12 differences in the calculated spatial values of milk. 13 The differences between the May 2021 spatial 14 values and the current Class I differentials are 15 considerable. 16 Let me refer, then, to the second of these 17 figures. In particular, there's a band from about 18 Norfolk, Virginia, through Montgomery, Alabama, where the 19 current Class I differentials appear to be well below the 20 model calculated spatial value of milk at the assumed 21 $1.60 per hundredweight minimum differential. There are 22 also a few cities, such as Charleston, West Virginia, 23 Cleveland, Ohio, and Chicago, where Class I differentials 24 are considerably below USDSS model estimated spatial 25 values. 26 The U.S. is roughly divided between east and west 27 approximately along the Mississippi River, which separates 28 regions where differentials are modestly low, west up to 6927 1 about $0.80, to areas where the difference may cause 2 difficulties encouraging milk to move where it is needed. 3 Probably the reason that there is a ridge where there is a 4 northern edge in the Southeast where current differentials 5 are significantly below calculated values is because of 6 the changes made in 2008 to the previous 2000 7 differentials. 8 At that time, the biggest changes, up to $1.80 per 9 hundredweight, were made to Florida values. More modest 10 increases were made to Georgia and Alabama, and even less 11 to states further north. So a similar pattern of 12 differences exists between USDSS-calculated differentials 13 for October '21 -- 2021 -- show that here in this 14 figure -- and the current Class I differentials, but with 15 somewhat smaller differences in Florida, Georgia, 16 Tennessee, and Kentucky. 17 Okay. So my concluding comments. There have been 18 formal studies of the spatial value of U.S. milk for about 19 a century. However, it has been approaching three decades 20 since nationwide spatial values of milk have been 21 systematically evaluated using the U.S. Dairy Sector 22 Simulator (USDSS) model. Over this time, there have been 23 considerable changes to where milk is produced and where 24 population growth has taken place. There have also been 25 substantive changes to transportation costs. Milk supply, 26 demand, and transportation costs all have an impact on the 27 spatial value of milk. 28 The USDSS captures many aspects of these 6928 1 fundamental determinants of values in U.S. dairy supply 2 chains to estimate spatial milk values that can inform the 3 setting of Class I differentials. The USDSS provides a 4 competitive benchmark for the differences in spatial milk 5 values, and analysis for two months in 2021 indicates 6 value at fluid milk plants considerably different from the 7 current Class I differentials. 8 As noted, the differentials arise from the 9 combined effects of changes in the locations and amounts 10 of milk supply, changes in the nature and location of 11 dairy product demand, changes in the locations and 12 capacities of dairy processing facilities, and changes in 13 transportation costs. 14 The USDSS provides evidence of the need for a 15 change in Class I differentials because it represents an 16 economic -- a spatial economic benchmark, but other 17 factors such as existing commercial relationships can be 18 important determinants of spatial organization. The model 19 results provide relevant input for differences in county 20 values, but may need to be adjusted based on additional 21 information about the characteristic of the particular 22 locations. 23 Any quantitative model is, by definition, a 24 simplification of reality, and the USDSS does not directly 25 represent existing commercial relationships that can be 26 important determinants of the locations and volumes 27 processed in existing operations. In fact, a review of 28 results from a previous version of the USDSS model was 6929 1 used as input into an adjustment process employed by AMS 2 to specify differentials in 1998. 3 And because I'm a weather nerd in addition to 4 being a modeling nerd, I would like to use an analogy 5 here. So there's an analogy here to the use of models 6 that generate the weather forecasts familiar to all of us. 7 The outputs of large-scale weather models are used as key 8 inputs, but forecasters often adjust this so-called model 9 guidance with professional judgment to arrive at a more 10 accurate forecast for a particular locality. 11 That concludes my statement. Thank you very much. 12 Q. Thank you so much, Dr. Nicholson. 13 MS. HANCOCK: Your Honor, I just have a few direct 14 examination questions before we turn him over for 15 cross-examination, but because we're after noon, and I 16 fear that Dr. Nicholson probably read that as fast as he 17 could without getting reprimanded by us, it might be a 18 good time for lunch, and then come back. 19 THE COURT: That sounds good. Now, I spent all my 20 time looking at 302. And I, most of the time, found out 21 where you were. 22 Did you also cover 303 while I was in 302? 23 MS. HANCOCK: He read 303. 24 THE WITNESS: I read 303, and 302 is the full 25 written testimony. 26 THE COURT: Well, it was a lot more fun to be in 27 302. 28 THE WITNESS: That's what people say about 6930 1 economic analysis all the time. 2 MS. HANCOCK: Nope. 3 THE COURT: Very good. Let's see. 4 Agricultural Marketing Service, that sounded like 5 a good plan, yes? 6 MR. HILL: That's fine. 7 THE COURT: All right. Great. So we'll just 8 leave everything where it sits, we'll have lunch, and we 9 normally take an hour. Is that still good? 10 MR. HILL: Yes. 11 THE COURT: Good. Please be back and ready to go 12 at 1:15. We go off record at 12:11. 13 (Whereupon, a luncheon break was taken.) 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 6931 1 WEDNESDAY, OCTOBER 4, 2023 - - AFTERNOON SESSION 2 THE COURT: Let's go back on record. 3 We're back on record at 1:16 p.m. 4 Ms. Hancock. 5 MS. HANCOCK: Thank you. 6 BY MS. HANCOCK: 7 Q. Dr. Nicholson, thanks for being back here with us 8 and providing your testimony. Just a few questions to 9 help clarify some of the things that you have in your 10 statement and the work that you did. 11 The USDSS model, I'm wondering if you can talk 12 about the dual values that are -- that's utilized in 13 that -- in that model. Maybe we can start there. 14 A. Okay. So as I indicated in a previous statement, 15 the dual values are the things that are providing us with 16 the spatial milk values, and in particular, that lead to 17 the mapping of the pattern of spatial milk values across 18 the United States. 19 And without trying to get too much into the 20 complications, essentially what those dual values 21 represent are, in the math of the model, we have a large 22 number of constraints. And the constraint would be 23 something like, if you are going to make a dairy product, 24 you have to have a sufficient amount of milk to be able to 25 make that dairy product based on the physical yield 26 relationships. Okay? 27 So the dual value is essentially saying, how much 28 would a dairy plant be willing to pay to have an 6932 1 additional hundred pounds of milk at that plant, based on 2 the mathematics of that particular constraint? And so it 3 is kind of a mathematical result. But it indicates the 4 marginal value of milk, how much more would a plant be 5 willing to pay for milk at that location. 6 Q. And so, for example, if you had two plants that 7 were across the street from one another, so same location 8 essentially, and you had a cheese plant on one side and a 9 butter nonfat dry plant on the other, how would the model 10 take those into account to quantify that? 11 A. Okay. So in the example of a cheese plant and a 12 fluid milk plant right across the road, the model has a 13 fairly myopic view of what the value difference would be 14 between those two plants. We know that the component 15 utilization in the cheese plant, like, in terms of the 16 butter and the protein and the other nonfat solids, would 17 be a little bit different at the cheese plant from the 18 fluid milk plant, but we also know that the model is 19 really only taking into account the transportation costs 20 between those two plants based on difference. 21 And if we, dare I say it, imagine a thought 22 experiment as a logical outgrowth of previous 23 conversations, we would think about the fact that if those 24 plants are really right across the road from one another, 25 the transportation cost difference that would be captured 26 by our estimate of transportation costs would be really 27 pretty small. So we would have every expectation that 28 just looking at those values from the model, they ought to 6933 1 be fairly similar. And that is, in fact, what we see. 2 Q. And if -- if -- if instead of being a cheese plant 3 and a butter nonfat dry milk plant, what if it was a 4 Class I plant and a Class III plant, would that output -- 5 would that outcome change? 6 A. No. Essentially if you have two plants 7 essentially of any type across the road from one another, 8 again, the model is only taking into account to estimate 9 those marginal or dual values. Any difference in 10 component utilization and the transportation cost 11 difference, that will be very minimal between those two 12 plants. 13 So, no, we would, again, not expect, regardless of 14 the two plants that are being compared, that we would see 15 a significant difference between those two plants in the 16 marginal value of milk. Thank you. 17 Q. Okay. I want to talk for just a second about this 18 model, this -- you understand that previously Class I 19 differentials were set based on 1998 modeling that was 20 done previously; is that right? 21 A. Modeling done in 1998 or prior to 1998, I 22 understand was one of the inputs into the Class I 23 differential surface that was promulgated in 2000. 24 Q. And what is your understanding about how the 25 current model that you have deployed compares to the model 26 that was used in 1998? 27 A. Well, there are some basic similarities in the 28 sense that we're using the same mathematical approach to 6934 1 try and minimize the cost systemwide of moving milk from 2 farms to plants and to final products. But there are a 3 number of significant refinements that have taken place 4 since 2000 that make this model much more appropriate for 5 today's dairy industry. 6 So some of those refinements are increasing the 7 number of spatial locations where farms, plants, and 8 consumers can be located. This most recent version of the 9 model went from a number in the hundreds of multi-county 10 aggregations for milk supply, for example, to a milk 11 supply in every county, 3,108, a significant scaling up of 12 the analysis. 13 Also, and perhaps more importantly, what we're 14 looking at is a very different set of product categories 15 that are now in the model that the 1998 model version, if 16 I'm recalling correctly, really only had four products to 17 represent four product classes. We now have in the tens 18 of different product categories, and we account for a 19 larger number of what we call intermediate products, 20 products that flow from one dairy plant to another, such 21 as the use of nonfat dry milk and cheese making. 22 So other refinements on this really relate to the 23 data which we have updated. We have a very different set 24 of population distributions, we have a very different set 25 of dairy product demands, and we have a very different set 26 of both farm production locations and components in farm 27 milk than we had back in 1998. 28 So all of those things represent a major overhaul 6935 1 to the model structure since the time it was originally 2 used. 3 Q. And National Milk approached you, was it in 2021, 4 to help them with some modeling work? 5 A. So my recollection is that National Milk 6 approached Dr. Mark Stephenson in March of 2022 to ask 7 about the possibility of us updating the USDSS model from 8 its 2016 database to 2021. 9 Q. And how is that you got involved to take over that 10 modeling work? 11 A. So I originally became involved in taking over the 12 responsibility for the USDSS when I went to Cornell 13 University in 2000. I had previously done graduate work, 14 as I noted in the earlier session this morning. I came 15 back to be a senior researcher at Cornell and was tasked 16 with the job of updating the particular model that we're 17 talking about, the USDSS. 18 So since 2000 I have been the primary programmer 19 of the model, and the primary person who has collaborated 20 with others to put together the datasets that we need to 21 run the model. 22 Q. And so at some point -- so Dr. Stephenson asked if 23 you would do the modeling work that National Milk was 24 asking to be done because you were the one that was in 25 charge of this database and the information? 26 A. Typically this has been a team effort between 27 Dr. Stephenson and myself. My role has been primarily to 28 make sure that the programming code and the model results 6936 1 are run. I'm the guy who flips the switch on the model on 2 a computer to make it actually generate the numbers, and 3 I'm the one who knows how to do that. 4 We have shared responsibilities for the collection 5 and updating of the data that the model needs to actually 6 do its magic in a particular month of a particular year. 7 So it's been a shared effort. 8 And so when Mark was approached as the better 9 known of the duo doing modeling by National Milk, he would 10 need to ask me if I'm interested in collaborating to make 11 that happen. 12 Q. Because you are the one that's in charge of the 13 model? 14 A. I'm the one who is in charge of the model. I'm 15 the one who has the model on my laptop, and I'm the one 16 that needs to flip the switch. 17 Q. And did National Milk give you any kind of 18 directives or guidance or any kind of outcomes that they 19 were hoping to achieve when they asked you to perform this 20 modeling work for them? 21 A. What I recollect is, whenever somebody has asked 22 us to do an update to the model, and this has happened on 23 a number of occasions, they always want to know what's the 24 latest information that you can use so that it's the most 25 recent. 26 So we had to have a little bit of a conversation 27 about what year. And that's why, even though we were 28 midway into 2022 at that point, the data availability was 6937 1 such that we could only do 2021. 2 We also have typically, as I noted in the 3 statement this morning, used the approach of doing two 4 months within a given year to represent more of a flush or 5 surplus season for milk, and a fall season in which milk 6 is in shorter supply, so that we have the contrast between 7 those two months. 8 So we have, on occasion, used months other than 9 the May and October that you see reported, so we had a bit 10 of a conversation about, are May and October okay? And 11 that seemed to be okay. 12 But other than that, it was up to us to update the 13 data and provide some initial results to the National Milk 14 team for discussion. 15 Q. Okay. So did they tell you, we're hoping to at 16 least increase it by a certain amount, or here's some 17 information that might help influence where the numbers 18 are going to go? 19 A. No. The initial model analyses that we undertook 20 were completely independent of any direct input from 21 National Milk, other than the things that I have 22 mentioned. 23 Q. Okay. And so that first time that you ran it in 24 May of 2022, you were using 2021 data; is that right? 25 A. Yes. 26 Q. And what was the transportation -- what was the 27 transportation costs that you were using in the initial 28 run? 6938 1 A. So the transportation costs that we have in the 2 model, what we do is create a large matrix of costs that 3 link every origin point to every destination point in the 4 model. To do that we use a transportation cost function 5 that relates distance to the amount that it costs to move 6 either a hundredweight of farm milk, or an intermediate 7 plant product like cream, or a distribution route like 8 packaged milk. 9 The things that are easier for us to change to 10 make it more applicable to 2021 would be something like a 11 fuel diesel price and a wage rate. And so we did adjust 12 the diesel price to 2021, and we did adjust the wage rates 13 to 2021 based on Bureau of Labor Statistics data and data 14 from -- I think it's the U.S. Department of Energy on the 15 fuel costs. 16 Q. Okay. And -- and that was so that you were 17 matching the transportation costs with the year in which 18 you were evaluating the other data as well? 19 A. Yeah. To be consistent, I mean, one can imagine 20 running a scenario, we had originally did this, where we 21 used the much higher 2022 diesel fuel prices, which would 22 generally tend to raise the nature of that price relative 23 surface. In part based on our assessment and in part 24 based on conversations with the folks at National Milk, we 25 decided it was probably better to be consistent because we 26 had 2021 data for farm milk supplies, dairy product 27 demands, and processing plant locations, and used 2021 28 diesel fuel values, even though they were lower than the 6939 1 high values that we saw in 2022. 2 Q. And did you feel like that was National Milk 3 attempting to be fair about ensuring that the data that 4 they were using in the model was going to be more 5 accurately representative of the 2021 calendar year? 6 A. It struck me as being both fair and also more 7 consistent, given that the rest of the data in the model 8 were 2021. 9 Q. Okay. And certainly not National Milk trying to 10 puff up its numbers, right? 11 A. I guess I don't want to speak to what National 12 Milk's intentions were, but I can say I never had any 13 impression that -- other than providing us with relevant 14 input to help us do our job, that they were trying to 15 influence the result in a particular way. 16 Q. Okay. And after running those initial results, 17 National Milk asked you to run them again in June of that 18 year; is that right? 19 A. That's correct. 20 Q. What additional information did National Milk 21 provide you, or what additional guidance did National Milk 22 provide you in order to have you re-run the numbers? 23 A. Yeah, maybe it's helpful here for me to point out 24 that we have always tried to be folks that are more in the 25 role of an analyst and not of an advocate. So when 26 someone asks us to do modeling work such as this, we can 27 have legitimate discussions about what scenarios are 28 relevant, that is, what assumptions we will use when we're 6940 1 going to run a model like this. And an example would be, 2 should we use the 2021 May and October diesel price or 3 should we use the 2022 diesel price as a way of evaluating 4 what that price surface would be? 5 So we had conversations with them about what 6 scenarios would be most appropriate, and one of the 7 conversations related to the diesel price. 8 We had another conversation that was related to 9 what we call the plant lists, which is the processing 10 plant locations and processing capacity values that we 11 have for plants in the model. And at that point, the 12 National Milk team had reviewed our plant list and 13 suggested that maybe we should not use the plants that had 14 already closed. And so we made some minor modifications 15 to a couple of specific plants based on the area of 16 specific knowledge of the National Milk team in order to 17 run an additional set of scenarios. 18 Q. And that original plant list that you had for the 19 initial run, did -- was that something that you already 20 had prior to National Milk? 21 A. Yes. So that plant list has been something that 22 has been developed over a long number of years and adapted 23 from information on various sources. And that can 24 actually include personal contacts with people in the 25 industry, where they will say there's going to be a plant 26 coming online, it's going to make these products. We also 27 have some states where there are lists of licensed plants. 28 Like, in Wisconsin, we have a licensed dairy plant list 6941 1 that we can look at. And we also have information that 2 comes from public press announcements about plant closures 3 or plant openings. 4 And so we had developed that list over a long 5 number of years and agreed to share that list with 6 National Milk to solicit their input on how we could make 7 the list more accurate. 8 Q. And so National Milk, in providing you with some 9 updated information about plants opening or closing, did 10 you take that into account when you re-ran the model? 11 A. So we kept the results from the first of our 12 simulations, and then through two iterations, we made 13 adjustments based on allowing plants that were scheduled 14 to come online to be included, so that meant we had to add 15 those into the entire structure of the model for the 16 analysis, and then also to sort of disallow processing of 17 facilities that had either already been closed or slated 18 to be closed, so we made those adjustments. 19 I don't remember the exact number, but I think 20 we're probably talking about a total of six to eight 21 plants switching from one category to another out of 22 several hundred, that are across the different product 23 categories. 24 Q. And how much did this affect the model results? 25 A. So not -- not very much. And actually, I remember 26 remarking to Mark that we were going to have to do a lot 27 of work to review the entire plant list when it wasn't 28 going to make a dang bit of difference. So it can make 6942 1 some difference if you have a very large plant in a 2 specific location, that can be impactful. 3 We did analysis when there was a Kraft Foods plant 4 in Canton, New York that was going to shut down and stop 5 making cheese, and we saw that right in that localized 6 area, yes, it actually had a fairly significant impact, 7 like $0.50 hundredweight, on the producer value of milk in 8 that area, but it did not affect the overall price 9 patterns for the United States. And the same is true for 10 the changes that we made to either close plants, not allow 11 them to be part of the model solution, or to allow plants 12 to enter. 13 Q. And notwithstanding the additional work that it 14 required from you, and even understanding that it might 15 have nominal or no effect on the results, you understood 16 that National Milk was just trying to get to an accurate 17 result? 18 A. Yeah. I think -- so one of the things that 19 happens when we do a model like this, is people often want 20 to know why is that number $3.50 at this particular 21 location? And that's a difficult answer to give for this 22 kind of modeling approach, because there are millions of 23 pieces of information that all come together and interact 24 to create that $3.50 number. 25 So when National Milk team reviewed the model 26 results, people are always trying to wonder, "In my 27 particular area, why did you get that $3.50?" And that 28 leads them to say, "Well, did you have the right plants in 6943 1 place for the analysis? And if you'd had that one open, 2 then you shouldn't have. And if you had that one closed, 3 then you probably should open it up." 4 And I think that was the motivation for making the 5 plant list more up to date, even though we recognize that 6 in the broad picture, it was not going to change the 7 nature of the results we were going to get. 8 I think that was what motivated in part, let's 9 make sure that you are not getting an answer for my 10 particular part of the world that I'm familiar with that 11 is different than what I think because you don't have the 12 right plants. 13 And so we appreciated the fact that we could 14 update and make more accurate the plant list, and also try 15 to say, we still think we're getting the right numbers for 16 the right reasons. 17 Q. And what are the major drivers of the model's dual 18 price results? 19 A. So the model, again, has both the dual results, 20 which have no values, and then also the primal results, 21 which are the physical flows through the supply chain. 22 And as I noted in the statement this morning, 23 there are some key things that are part of those millions 24 of pieces of information that drive that. So the key 25 things really are, where do we have milk, and what is its 26 composition spatially throughout the United States? Where 27 is the milk located? What's its composition? What's the 28 composition and location of dairy product demand? What is 6944 1 the location and processing capacities of different dairy 2 processing facilities? 3 And what are the transportation costs that link a 4 farm to a processing plant in terms of milk assembly, the 5 movement -- excuse me -- of intermediate products from one 6 dairy processing facility to another, and the 7 transportation costs associated with distribution. 8 So all of those are part of the core database that 9 make up the USDSS analysis, and all of those things are a 10 part of why we get the spatial price surface that we get. 11 Q. And I think in your testimony throughout, you 12 refer to the model results as a benchmark. 13 Why do you consider them to be a benchmark? 14 A. So the terminology that I have used is a 15 competitive benchmark. And in this case I'm kind of 16 drawing upon the economic idea of perfect competition 17 where we don't -- we say, everybody is sort of equal, they 18 are all small, they all take the same price or receive the 19 same price from people, and that means that we're not 20 really fully accounting for a number of institutional 21 factors that could be relevant to refining the model 22 results to come up with what might be a more appropriate 23 industrywide Class I price surface. 24 So what I'm saying competitive benchmark, what I 25 mean is, this is sort of like the lowest possible 26 systemwide cost that we can imagine in a perfect world. 27 Right? And so we recognize, though, that that perfect 28 world isn't the world in which the dairy industry lives. 6945 1 There are lots of other factors that might be important, 2 even if this provides a basic scaffolding for thinking 3 about what those price relatives should be. 4 Q. Okay. And we heard Dr. Vitaliano talk about 5 some -- some -- what I would -- that he called art, or 6 what is an overlay over the numbers that -- that come out 7 of the benchmark. 8 Do you recall him talking about that? 9 A. Yes, I do. 10 Q. And is that the additional information that you 11 believe would -- is -- is used or applied to the model 12 results that come out of this model? 13 A. I guess I -- I don't know exactly what information 14 was used in the process, not having been a part of any of 15 the discussions of what has been called the 16 colored-pencils sort of adjustments. All I can do is 17 comment on the things that I think the model does not 18 fully incorporate that might be relevant. 19 Q. Okay. And what would those be? 20 A. I think they come into maybe three categories -- 21 well, four. 22 So one is really we use average transportation 23 costs on the basis of difference in distance between a 24 start and an endpoint for moving milk. We do actually 25 adjust those for local conditions in the sense of having a 26 different fuel cost and a different wage cost. But what 27 we don't account for, for example, is like the density of 28 the milk supply in a particular county. 6946 1 So I used to work at Penn State University in the 2 top-ranked supply chain management department, and from 3 that I know that there are counties in Pennsylvania where 4 there are a lot of plain sect folks, Amish, who have small 5 farms. And our model would say, all that milk is at one 6 location in the county, and to move that county down to 7 the next county would all be the same costs. And the 8 reality is that if you're trying to serve that particular 9 set of farms, the cost would probably deviate a bit from 10 what the model would say would be the cost to move it from 11 one county to another. 12 Another example from when I worked in California 13 is I'm quite familiar with how traffic can be in the Los 14 Angeles area. So our model assumes all the costs are on 15 the basis of a distance movement, which would say there's 16 such and such a distance going from Bakersfield to Los 17 Angeles, and the cost would be this, but we don't account 18 for the fact that that time cost and the driver cost 19 associated with it could be much different. Right? 20 So those are transportation cost examples that are 21 probably more widely relevant for places that I haven't 22 lived and worked that the others from the National Milk 23 team may want to speak to. 24 Second thing is that the model has no compassion 25 about keeping plants open because there's always been a 26 plant there. In presentations that I have given about 27 this model previously, I like to use the example of a 28 model being a dairy dictator, like the Vladimir Putin of 6947 1 dairy supply chain allocations. And it would say, if you 2 have a plant that's not in a good location, the model is 3 not going to keep that plant operating. But for an 4 individual company, that would probably not be an easy 5 decision for them to make, especially in the short-term. 6 So the model doesn't account for that existing capacity 7 that an organization would want to keep using. 8 Another example is commercial relationships. 9 Again, we're hardhearted, we just want to get the milk and 10 the dairy products from the farm to the plant, and to the 11 consumer as a low cost as possible, with the analytical 12 approach we're using here. We don't know anything about 13 the commercial relationships that might link a particular 14 farm milk supply to a farm, to a plant that actually has a 15 contractual obligation on that milk. All right? So the 16 model is going to show more flexibility than the real 17 world in terms of not respecting that contractual 18 obligation. 19 And the one last thing that's kind of important 20 that often people have maybe been a little bit confused 21 about is we use the model to generate these price 22 relatives to provide a base of information for Class I 23 differentials, but the model itself is a competitive 24 benchmark from a supply chain perspective, it does not 25 know anything at all about Federal Orders. It does not 26 know anything about pooling provisions, it does not know 27 anything about current order boundaries. 28 And so one of the things that can arise -- and 6948 1 although I was not a part of the team at Cornell that did 2 the modeling work in 1998, my understanding was that when 3 the folks at Dairy Programs AMS were doing their version 4 of the adjustment process to the model results, one of the 5 things that they were interested in understanding and 6 making sure was okay was sort of price alignment at Order 7 boundaries. So we don't have any Order boundaries in the 8 model, and therefore, we could come up with price 9 relationships in nearby space that would be perfectly fine 10 from a model perspective, but may not be acceptable from 11 an Order boundary or price alignment perspective. 12 So we have sort of those four things that I think 13 are relevant for why adjustments might be necessary to the 14 raw results from the USDSS model that include some more 15 detailed knowledge of local transportation conditions, the 16 existing contractual arrangements, the existing capacity 17 in wanting to maintain open a plant that you have invested 18 in, and the issue of price alignment across orders in 19 particular. 20 Q. Okay. And these four areas, these are the areas 21 that you believe would be taken into account on top of the 22 model results which are the benchmark that you have 23 described? 24 A. Now, again, I can't say what was taken into 25 account in coming up with any differences between the 26 model results and the proposal that's being put forward by 27 National Milk. What I'm trying to do is point out that 28 there are factors that I would consider relevant factors 6949 1 that would mean adjustments to the model would be 2 appropriate. 3 Q. Okay. And one of -- one of the elements that we 4 have heard about, and maybe you will hear more about as 5 your cross-examination continues, is what the base was 6 that was included in your model. 7 Can you talk about that? 8 A. Yeah. Sure. So another thing that's important 9 for me to maybe clarify is when we run this model, we get 10 a series of price relatives, as I've said. And it's 11 basically about how steep is the price difference, so the 12 marginal value difference between two locations. 13 So typically what we need to do to actually 14 convert that to something that is equivalent to what we 15 might think of as the current Class I differential surface 16 is we need to establish $1.60 as the minimum. So it would 17 be fairly typical in a model simulation run to have one 18 location that says the marginal value of milk is zero. We 19 don't need any more milk here. There is no additional 20 value from having another hundredweight of milk at this 21 location. 22 Well, we don't fully believe that the value of 23 milk at any location is zero. And so what we do to come 24 up with the results that have been shared in the written 25 testimony, and parts of here in the oral testimony this 26 morning, is if we have a value of zero, we say, to align 27 that with the current Class I differential surface, we're 28 going to add a value of $1.60 per hundredweight to that 6950 1 and every other location. So it maintains the price 2 relatives the same, but it takes the level, the minimum 3 level, up to the current minimum level of Class I 4 differentials of $1.60. 5 And that's important, in part, to be able to 6 compare the apples to apples that you have. Our model 7 simulation results start with $1.60 per hundredweight, so 8 do the Class I differential current surface, and then it 9 makes it a lot more consistent to evaluate the differences 10 between the spatial values of milk in our model and the 11 current Class I differentials. 12 Q. Okay. And that's what you did here in this case? 13 A. Yes. That's what we did. 14 Q. And so for all of the different iterations that 15 you ran, did you always use that $1.60 base? 16 A. Yes. So we always made sure that the minimum 17 marginal value of milk was $1.60 throughout the entire 18 U.S. for fluid milk plants. 19 Q. And then at some point did National Milk come to 20 you and say, "We would like you to increase that to 21 $2.20"? 22 A. No. I have no idea where the $2.20 number came 23 from. 24 Q. Okay. So that's not something that National Milk 25 tried to direct you to do? 26 A. They did not direct, and it did not happen, I 27 guess. 28 Q. And we also -- and maybe along those same lines, 6951 1 was there anything that National Milk ever told you to try 2 and influence your results? 3 A. So I mentioned before that the only influence was 4 really on the design of the experiments that we were going 5 to do, these scenarios. And those were really limited to 6 what were the months and year we were going to look at, 7 what was the diesel price, and let's be sure that we have 8 the appropriate plant list that is consistent with updated 9 information. Other than that, the scenarios that we ran 10 were entirely based on our own data. 11 Q. Is there anything that you could have done 12 differently in any of the iterations that you ran that 13 would have made it more accurate or more reflective of the 14 market conditions? 15 A. I think we have about as accurate a representation 16 as we can with the available information, and it did 17 actually help, even though it did not change the model 18 solutions very much at all, and created a lot of 19 additional work on a weekend that I didn't want to do, to 20 have the additional information to update the plant list. 21 And so that was the source of information that we were 22 able to tap into the knowledge of the National Milk team 23 to be able to improve in that sense. 24 Q. And Mr. English, when we was conducting the 25 cross-examination of Dr. Vitaliano, he looked at a change 26 that your model had predicted, or that your model had -- 27 the model results for Miami and the increase that was -- 28 the increase that was proposed by National Milk based on 6952 1 that model, and then as compared to Minnesota. 2 Were you in the room when he was asking those 3 questions? 4 A. I had the pleasure of hearing that discussion. 5 Q. Is there any insight you can provide as to, if -- 6 if the Southeast is in such dire need of milk, why the 7 results didn't come up with something even more 8 significant? 9 A. I guess I'd make two points in that regard. One, 10 in the testimony that I gave this morning I noted that the 11 largest divergence between the spatial models predicted by 12 USDSS and the current Class I differentials are not in 13 Florida, they are north of Florida. And I also offered a 14 suggestion that one reason for that may have been that the 15 differentials in that part of the country were already 16 adjusted in 2008. 17 Q. Okay. 18 A. And one other point that I make on that is we 19 still actually do see the largest spatial values of milk 20 in that South Florida area, up to $8, and so there's a 21 considerable difference between what the model is 22 suggesting would be the spatial value of milk at that 23 location and the current Class I differential. 24 Q. Okay. 25 A. Just not the biggest divergence at this point. 26 Q. Meaning that area specifically had already had an 27 update since the 1998 model results. 28 A. We can't correctly analyze that with the model. 6953 1 But in comparing the current Class I differentials to the 2 model spatial values, we can begin to understand that that 3 is a possible explanation for why those -- the differences 4 are higher north of that area than they are in that area. 5 Q. And would you mind pulling up your Figure 3 from 6 your testimony in Exhibit 303? 7 A. I can ask the -- there you go. 8 Q. Can you talk about whether this helps illustrate 9 what you were just describing? 10 A. Yeah. This is the pattern that I talked about in 11 the summary this morning. The darker colors there, the 12 oranges and the reds, are the places where there is a 13 larger divergence between the current Class I 14 differentials and the model-generated values. 15 And so you can see in that area down in Florida, 16 that green area, it's a little bit hard to see the scaling 17 on this, but that kind of generally falls in the $1.50 to 18 $1, maybe $2 range. Whereas, north of that we actually 19 get up into things that look more -- well, definitely 20 above $2, maybe 2.50 to 2.75. And the brightest red spot 21 there, which I think is around Charleston, West Virginia, 22 is the thing that I cited as the largest of the difference 23 of $3. 24 THE COURT: What location is that red spot? 25 THE WITNESS: I don't know if I have my geography 26 right, but I'm thinking it's Charleston, West Virginia? 27 THE COURT: We're getting nods "yes." 28 THE WITNESS: Okay. Thank you for helping my 6954 1 geography-challenged brain. 2 BY MS. HANCOCK: 3 Q. And when you were talking about the transportation 4 costs that are built in to the current model, were those 5 same transportation costs built into the model back in 6 1998 when it was originally run? 7 A. So some of like I mentioned before, the basic 8 structure of the model had some similarities. But I was 9 not part of the modeling team in 1998. What I do 10 understand is the initial version of that model had a 11 straight line transportation function, where the cost of 12 transportation increased linearly mile by mile. 13 One of the things that we have, I think, learned 14 through the additional analysis of data on the 15 transportation costs is, at least up to a certain point 16 where you might run into an hours-of-service limitation, 17 the costs increase with distance, but they don't increase 18 linearly. They taper off. They increase a little bit 19 more slowly because you have covered some fixed costs 20 initially, right? If you do hit that hour-of-service 21 limitation, and you've got to go another day or have 22 another driver, then actually that could make that cost go 23 up again, but that's not captured directly in our 24 transportation cost analysis. 25 Q. Okay. 26 A. So there's a very big difference in terms of the 27 data that's been used, and also the form of the 28 relationship that determines the cost between two 6955 1 locations. 2 Q. Okay. And back to my original question, which 3 was, the modeling that was -- this model that was used in 4 1998 to set differentials, that, likewise, took into 5 account transportation costs, it's just that the 6 methodology of how it was taken into account has become 7 more precise with updates to the system? 8 A. Yes. Both models includes some representation of 9 the transportation costs for farm milk assembly, 10 interplant flows, and distribution routes. It's just that 11 the nature of the estimation and the updating is 12 different. 13 Q. Okay. And in this very tight-knit world in which 14 we live in the dairy industry, you're familiar with 15 Dr. Stephenson using the modeling in support of MIG's 16 proposals for their differentials; is that right? 17 A. I recently became aware of the fact that 18 Dr. Stephenson had used model results to provide input to 19 the MIG proposal. 20 Q. And I think you said earlier that you're kind of 21 the keeper of the model. 22 Did he have to come to you and ask you for some 23 information? 24 A. So we have shared a lot of the information, both 25 the inputs and the outputs, throughout the modeling 26 process that was undertaken for National Milk. 27 In regard to this particular question, I shared 28 information with Dr. Stephenson to allow him to confirm 6956 1 that he had the correct values of spatial milk values from 2 the model. I did not realize the purpose to which that 3 information would be put. 4 Q. Okay. And what do you understand is the 5 difference in the methodology that he's deploying as 6 compared to what you are doing? 7 A. Well, the same model is generating the 8 information. And what's happening, somewhat like I 9 described, we make a calculation that makes sure we have a 10 $1.60 minimum Class I differential. 11 Dr. Stephenson is taking the information from the 12 same model and using it to do some alternative 13 calculations and for a different purpose. 14 Q. Okay. And what do you understand is the 15 differences in how he's doing his calculation? 16 A. Okay. So what I understand is a core part of the 17 analysis that's been submitted is to consider the 18 model-generated differences in spatial milk values at 19 Class III and Class I plants, without incorporating the 20 $1.60 differential that is included in our analyses. 21 Q. Okay. And in your opinion, is it appropriate to 22 use a Class I and Class III comparison in order to 23 evaluate these numbers? 24 A. So it's a perfectly fine calculation to do to look 25 at the difference between a Class III price and a Class I 26 price, not including what the $1.60 differential would be. 27 Where I think I have a bit of a difference of 28 opinion is that we have never used this model to try and 6957 1 determine what that minimum Class I differential should 2 be. That is, we have never used this model to try and 3 determine whether $1.60 is an appropriate number. And 4 part of the reason that we have not done that is the model 5 does not really represent the factors that underlie the 6 justification for that $1.60 minimum Class I differential. 7 So my assessment is, given that the model was not 8 really designed to evaluate what the minimum differential 9 should be because it doesn't incorporate those factors, it 10 is probably not appropriate to use the difference between 11 a Class III model-generated value and a Class I 12 model-generated value to suggest what the minimum Class I 13 differential should be. 14 Q. Okay. And then I want to take us full circle, 15 which was all the way back to my very first question that 16 I asked when we started, which is, now we're back to we 17 have two plants across the street from one another. And I 18 posed you the question early on, if you have a Class I 19 plant and a Class III plant across the street from each 20 other, how the model impacts the decision to go one way or 21 the other. 22 Do you remember that? 23 A. Yes. 24 Q. And I asked you, well, if you just took those 25 plants and you replaced them with a cheese plant and 26 butter nonfat dry milk plant, would the results change? 27 And what was your answer? 28 A. My answer was that regardless of the plant types, 6958 1 because of the factors that are included in the model, we 2 would expect to see very similar differences regardless of 3 what class plant or what product plant type that would be. 4 So we would not expect to see large differences based on 5 the factors that are accounted for in the model for the 6 hypothetical situation where a cheese plant and a fluid 7 milk plant are right across the road from each other. 8 Q. So is the point there that this model doesn't tell 9 you one way or another which one is the bigger driver 10 between the -- between the classes of milk? 11 A. So I guess I would say that the model is not going 12 to accurately represent what a fluid milk plant should pay 13 to get milk into the plant relative to what a cheese plant 14 should pay. It's really good at describing how the 15 differences across space exist for different fluid milk 16 plants, but it's not designed to account for the fact -- 17 or the factors that affect what that minimum Class I 18 differential should be. 19 Q. And your role here today, Dr. Nicholson, are you 20 here as an advocate for National Milk's proposal or to 21 object or oppose any other proposal? 22 A. No. So I -- you had mentioned earlier that I was 23 a part of something called the Program on Dairy Markets 24 and Policy, and that was a group of academic economists 25 with an interest and focus on dairy. And one of the 26 things that was a requirement for membership in that group 27 was some kind of commitment to the fact that we want to 28 play an educational role and we want to play an analytical 6959 1 role by providing information that can help the industry 2 to make better decisions, and that we were not to be 3 advocates for any particular position, in part, because as 4 you mentioned, this is a small industry. We have worked 5 with people on all different sides of different questions. 6 We wanted to maintain the credibility that we were a 7 neutral, unbiased source of information. 8 So I'm here at the request of National Milk, but 9 I'm not actually here to say I think the National Milk 10 proposal is a wonderful idea or it's a bad idea, or MIG's 11 proposal is good or bad. I'm here to try and help provide 12 some insights about the spatial milk values and how they 13 changed over time. 14 Q. And we heard Dr. Bozic here yesterday, or a few 15 days ago, I can't remember when it was, and he said that 16 he's going to be stopping the work that he's doing within 17 that Program for Dairy Markets because he's leaving the 18 academic side; is that accurate, your understanding of 19 what's happening? 20 A. With all due respect to Marin, who I've known for 21 a long time, that was an appropriate decision. And it's 22 not as if we kind of are gatekeepers. We actually sort of 23 semi-lost a number of people through retirement that go 24 back many years in my time. Most recently we can still 25 count a little bit on folks like Mark Stephenson and Andy 26 Novakovic, but essentially it's down to Dr. Chris Wolf at 27 Cornell and myself that are the ones that are trying to 28 make that program happen. 6960 1 Q. And that's because Dr. Stephenson is retiring as 2 well; is that right? 3 A. Is retired. 4 Q. Okay. 5 MS. HANCOCK: I have no further questions at this 6 time. We would make him available for cross-examination, 7 Your Honor. 8 THE COURT: Dr. Nicholson, would you state again 9 and spell the names of your two colleagues who still work 10 with you in this program? 11 THE WITNESS: Okay. So there's one colleague who 12 is definitely not retired and still working, and his name 13 is Christopher, C-H-R-I-S-T-O-P-H-E-R, Wolf, W-O-L-F. 14 Members that are still available to us, although 15 they have retired, are Mark Stephenson, I can spell that 16 if you wish. 17 THE COURT: I don't need that one. 18 THE WITNESS: Okay. And Andrew Novakovic. I can 19 spell those if you would like as well. 20 THE COURT: Yes, please. 21 THE WITNESS: Okay. I hope Andy is not listening 22 because if I get his name wrong, it's going to be trouble 23 for me. So Andrew, A-N-D-R-E-W, and Novakovic is 24 N-O-V-A-K-O-V-I-C with a special Serbian accent over it. 25 THE COURT: So -- so I got N-O-V-A. 26 THE WITNESS: K-O-V-I-C, I believe. Help me out 27 in the audience if I'm not getting it right. 28 THE COURT: N-O-V-A-K. 6961 1 THE WITNESS: O. 2 THE COURT: V-I. 3 THE WITNESS: V-I-C. 4 THE COURT: Pronounced? 5 THE WITNESS: "Novakovich," proud Serbian 6 heritage. 7 THE COURT: Excellent. Now, I see we already have 8 someone else who wants to ask questions. 9 Would you identify yourself, please, sir. 10 MR. ENGLISH: Certainly, Your Honor. My name is 11 Chip English for the Milk Innovation Group. 12 And I had a little time to get up here because you 13 were going through some spelling of -- of names. 14 CROSS-EXAMINATION 15 BY MR. ENGLISH: 16 Q. Good afternoon, Dr. Nicholson. 17 A. Good afternoon, Chip. Mr. English, excuse me. 18 Q. So I think, although I may get corrected, I just 19 want to have, I think, one question based upon the last 20 line of questioning, and I'll let Dr. Stephenson speak for 21 himself. 22 But when you talked about the data -- or the 23 request for information, I want to be clear that my 24 understanding is that because of proprietary information, 25 that is to say the work you did for National Milk was 26 proprietary, that the data that Dr. Stephenson used was 27 not the 2021 data, but rather 2016 data for another 28 project; is that correct? 6962 1 A. Yes. Excuse me for that. Yes, that's correct. 2 Q. Okay. So I want to be very clear that whatever he 3 did was not use the data that was paid for by National 4 Milk, correct? 5 A. Correct. 6 Q. Okay. And that was important to him, correct? 7 A. Yes. 8 Q. So I am going to try very hard to shorten my 9 cross-examination because a number of questions that I had 10 were questions that National Milk Producers' counsel 11 asked. And forgive me if I do duplicate, but I'm going to 12 try hard not to. 13 A. Thank you. 14 Q. So you were not, as you said, part of the work for 15 Federal Order reform, correct? 16 A. The 1998 effort? No. 17 Q. Yep. And in fact, just to be clear, the 1998 18 effort was USDA's proposed rule, but it was actually based 19 upon a report dated July 1997, correct? Do you remember? 20 A. That's the best of my knowledge, yes. 21 Q. And are you familiar with that report? 22 A. I have reviewed that report, but it has been quite 23 some time. 24 Q. But you reviewed it, you certainly reviewed it in 25 light of the fact that you have been, over time, updating 26 the underlying data, correct? 27 A. Yes. 28 Q. And you couldn't update that without having looked 6963 1 at it some time in the past, correct? 2 A. At some time in the past, yes. 3 MR. ENGLISH: So, Your Honor, one of the things I 4 want to try to do -- and I believe I have agreement of all 5 counsel so I got to speed it up -- but that's -- is that 6 rather than asking him to go through in some level of 7 detail that report, I represent the following: 8 The 1997 report, July 1997, known as RB 9709 -- 9 and why don't I make this simpler by handing you a copy. 10 THE COURT: Thank you. So that's RB, as in boy. 11 MR. ENGLISH: Yes, 9709. Entitled -- 12 THE COURT: That's 97-09. 13 And then what were you going to say? 14 MR. ENGLISH: It's entitled "A Description of the 15 Methods and Data Employed in the U.S. Dairy Sector 16 Simulator, Version 97.3," authors: James Pratt, Phillip 17 Bishop, Eric Erba, E-R-B-A, Andy Novakovic, whose name you 18 just got spelled, and Mark Stephenson. 19 This study, Your Honor, is cited six times in the 20 Federal Order Reform proposed rule. But I also specify, 21 and let me read from preface, page ii: "Funding for this 22 project has been provided by the U.S. Department of 23 Agriculture through the National Institute for Livestock 24 and Dairy Policy and through USDA's Agricultural Marketing 25 Service Dairy Division and Federal Milk Market 26 Administrators." 27 As -- as such -- and, by the way, it is cited, 28 like I said, six times in the Federal Order reform, and it 6964 1 is specifically footnoted. And within Federal Order 2 reform it says, "We, USDA, had two partnerships, one with 3 Cornell University and one with Texas A&M to assist us 4 with Federal Order reform." 5 Rather than making -- 6 THE COURT: To assist us with what? 7 MR. ENGLISH: "To assist us with Federal Milk 8 Order reform." 9 I -- I believe that this is basically a public 10 document funded by the federal government, submitted to 11 the federal government, recognized by the federal 12 government. 13 Rather than making it an exhibit and, you know, 14 helping the paper companies sell more paper, I simply 15 propose -- with a citation I'll give in a moment -- to 16 take official notice of it. I believe I have agreement of 17 all the parties. Maybe it will shorten my 18 cross-examination by 30, 45 minutes. 19 THE COURT: Now, are you going to put this -- if I 20 take official notice of it, are you going to submit it so 21 that it's part of what USDA receives as a document, or are 22 you just going to leave the citation in the record be 23 adequate? 24 MR. ENGLISH: Yes. The second is the case, Your 25 Honor. 26 THE COURT: All right. 27 MR. ENGLISH: So the citation I have is -- 28 THE COURT: Just one second before you do that. 6965 1 I want to ask if there's any objection to our 2 proceeding in this manner? 3 There is none. 4 Thank you, Mr. English. Now you may tell me 5 everything that you want in the record about this citation 6 of this document of which I will take official notice. 7 MR. ENGLISH: All right. The citation is 8 dairymarkets.org/pubPod/pubs/RB9709.PDF. 9 THE COURT: Great. 10 THE WITNESS: So, Mr. English, if I may? So that 11 is an online reference through the DairyMarkets.org 12 website, which, since Mark has retired, has been less well 13 maintained. But that document, the RB stands for research 14 bulletin, and it is also available perhaps in a more 15 permanent and accessible form at the Charles H. Dyson 16 School of Applied Economics and Management web page under 17 research bulletins. So just in case there would be any 18 issues with the link that would make that available, there 19 is an alternative source. 20 MR. ENGLISH: I'm grateful. The link worked this 21 morning. 22 THE WITNESS: Excellent. 23 MR. ENGLISH: I don't guarantee it works this 24 afternoon. 25 THE WITNESS: I don't either. 26 THE COURT: I'd like you to spell the name that is 27 part of the identification of where a person would find 28 this report. 6966 1 THE WITNESS: So this is Cornell University, the 2 Dyson School, D-Y-S-O-N, School of Applied Economics and 3 Management. And I imagine that a Google search of 4 research bulletin, Cornell Dyson, would take you somewhere 5 close to accessing it through that set of links. 6 THE COURT: Excellent. Thank you both. 7 I do take official notice of this document, 8 R.B. 97-09, the name of the document -- well, first of 9 all, the date of the document is July 1997 (printed 10 December 1997). Name of the document is "A Description of 11 the Methods and Data Employed in the U.S. Dairy Sector 12 Simulator, Version 97.3." Down at the bottom it says "A 13 Publication of the Cornell Program on Dairy Markets and 14 Policy." 15 BY MR. ENGLISH: 16 Q. So now, Doctor, counsel for NMPF took you through 17 some of the updates since that time, correct? 18 A. Yes. 19 Q. So I just want to, as quickly as we can, 20 nonetheless, talk about the robustness subject of the 21 updates, nonetheless so this record has the robustness of 22 what was in there in 1997, to the best of your 23 recollection. 24 I'm going to give you some ideas, and you can tell 25 me whether I'm right or not. Does that work? 26 A. Yeah. Okay. 27 Q. And part of that is, and for your Honor's benefit, 28 since this was -- you know, we had informal rulemaking for 6967 1 Federal Order reform. That is to say Congress passed a 2 statute that said, you know, go do all this stuff, but 3 don't be here in this hearing room. And for whatever 4 reason, people decided they prefer this process. 5 And so -- but this is a different process. It's a 6 formal rulemaking, and so things that were just -- there's 7 no -- there's no hearing record, to my knowledge, from 8 that proceeding, but there's a hearing record here, so 9 I -- just bear with me, and again, I'm trying to move as 10 fast as I can. 11 THE COURT: I don't want you to go fast. I know 12 that Dana Coale wants you to go fast. 13 MR. ENGLISH: And I think Dr. Nicholson wants me 14 to go fast. 15 THE WITNESS: And my students, and the Department 16 of Ag in the State of Wisconsin, but other than that, 17 we're fine. 18 MR. HILL: Mr. English, this is Brian Hill. 19 Before you get started, I know Your Honor wanted to have a 20 break at some point because of the test. It's now 2:14, 21 and so I wanted to alert you to that before we started 22 moving. 23 THE COURT: Six minutes until we're to get FEMA's 24 emergency system test -- 25 MR. ENGLISH: Turn off all our phones. 26 THE COURT: -- this is just a test. We're 27 supposed to have, yeah -- I'll go off record, you can move 28 around, and see if you get it. It's supposed to go on to 6968 1 televisions, mobile devices, and the like. It's just a 2 test. 3 All right. Let us do go off record now at 2:14. 4 You are free to move around. Come back by 2:25. 5 (Whereupon, a break was taken.) 6 THE COURT: We're back on record at 2:25. 7 MR. ENGLISH: Thank you, Your Honor. 8 And thank you, Mr. Hill, for reminding us about 9 how our phones are going to blow up. 10 BY MR. ENGLISH: 11 Q. So before the break I was going to attempt to 12 summarize with the witness some of the robustness subject 13 to -- robustness of the 1997 materials which have been 14 subsequently updated, some of which the most recent 15 materials came from counsel for National Milk Producers 16 Federation. 17 So as I understand it, there were objective 18 functions, such as raw milk assembly costs, correct? 19 A. Are you looking at the Table of Contents? 20 Q. Yes. 21 A. Is it possible to have you guide me through that a 22 bit? 23 Q. Yes. So I'm thinking about page -- starting on 24 the Table of Contents, United States Dairy Sector 25 Simulator, Explanation of Objective Function and 26 Constraints, the fifth line down, there's a series of 27 functions listed, and one of them is, you know, Raw Milk 28 Assembly. 6969 1 A. So the -- I guess what I would like to perhaps 2 help distinguish is the objective function is the overall 3 set of costs for everything in the model that has a 4 particular equation. It is not similar to the remaining 5 ones which are constraints which must be satisfied. So -- 6 Q. I appreciate -- 7 A. -- they are all -- they are all equations, but 8 they are of a different type when you go from the first 9 line to the next line. 10 Q. All right. Then in that case, help me out 11 understanding what they are, because that's exactly what 12 I'm having a problem with. 13 A. Okay. 14 Q. Just identify the reason why you are testifying 15 and I'm not. 16 So let's start with that. So what is a function 17 versus a constraint? 18 A. So a function is any mathematical relationship, 19 and in an optimization model like the one that provides 20 the information from the United States Dairy Sector 21 Simulator, the objective function has got a combination of 22 the variables in the model. And typically, in this case, 23 it's also going to have the associated costs that go along 24 with the variable. 25 So, for example, a variable would be the movement 26 of milk from one county location to a plant in a different 27 county location. A cost associated with that would be the 28 cost to move the milk that distance if it's a farm milk 6970 1 assembly movement, right? So you would have a combination 2 of a cost per unit, times a volume of milk flowing from 3 one location to another. You multiply those two things 4 together, and you actually get a dollar value. And then 5 you do that about 6 million more times, and you've got the 6 objective function for the current version of the USDSS. 7 So it's adding up the total value in terms of the 8 costs for the month either of May or October 2021. 9 Q. And so when you said 6 million more times, that's 10 just for the raw milk piece? 11 A. Those are all the variables that are part of the 12 objective function, which would include the milk assembly 13 flows, the processing at different locations, the 14 distribution flows, the interplant flows, all those things 15 are variables that are included. 16 It's not -- that 6 million number is not just for 17 milk assembly. 18 Q. Okay. So that's -- the milk assembly, the 19 receiving of milk components at plants are included? 20 A. So -- so the objective function is all in terms of 21 values. The other functions tend to be in terms of 22 physical quantities, because they are putting constraints 23 on that for the most part represent a mass balance that 24 says, if you are going to have so much cheese come out of 25 a cheese plant, you have to have so much milk and other 26 ingredients come into that cheese plant to be able to 27 mathematically describe the relationship between the milk 28 inflow and other products I mentioned already, nonfat or 6971 1 cream, and the cheese product that comes out. 2 Most of the other of those equations that are 3 described as the constraints are in terms of physical 4 units. The main one that's in terms of dollars is the 5 objective function. 6 Q. And is the restriction on use of components from 7 intermediate products a constraint? 8 A. Yes. So that would be, for example, you can't 9 make cheese entirely from nonfat dry milk. 10 Q. And when you talk about -- so actually, I'm going 11 to turn to the data now. 12 So what is involved in these data and how many 13 data points are there, if you know? 14 A. I can talk about that. Do you want me to talk 15 about the 1997 version or talk about the -- 16 Q. I'm happy -- I'm happy -- I'm happy to have you 17 sort of -- 18 A. Okay. 19 Q. -- combine. The whole point was to give people a 20 chance to see how much was there, but I do think we want 21 to talk about present time rather than 1997. 22 THE COURT: You are just going too fast, 23 Mr. English. 24 MR. ENGLISH: That's because I'm not reading 25 anymore. I will slow down. 26 THE WITNESS: Okay. So I had previously mentioned 27 that the key data inputs in the current version of the 28 model are the amounts of milk and their composition at 6972 1 different supply locations, either the multi-county 2 regions or the 3,108 individual counties, many of which in 3 the United States will have zero milk, the locations of 4 plants of different processing types and their capacities 5 to the extent that we know them, and the dairy product 6 demand at different locations based on population and per 7 capita dairy product demand values that we have calculated 8 based on publicly available data. 9 The other part of the information that is included 10 in our version of the model is the transportation costs, 11 which I have described a little bit before as being based 12 on functions that differ for farm milk assembly and differ 13 for interplant flows from, say of cream, from a fluid 14 plant to an ice cream manufacturer, and for the 15 distribution routing of final products from a plant to a 16 customer location. 17 So if it's -- if it's helpful for me to go through 18 and say which of the things that are listed on data here 19 are things that are data that are included in the model, I 20 could do that. 21 BY MR. ENGLISH: 22 Q. That would help, yes. 23 A. Okay. So cities and distances, yes, we have a 24 network of cities, and we have the distances that connect 25 them. It's greatly upscaled in the county level version 26 of the current model. I have mentioned previously farm 27 milk supply, the areas of quantity, and composition. I 28 have mentioned previously processing locations, and 6973 1 actually maybe if it's helpful, I can refer to the page 2 numbers that are cited on this, if that helps those 3 following along. Okay. 4 All right. So cities, and if we're on page iii of 5 the Table of Contents, down at the bottom under Data, the 6 Cities and Distances are listed on page 26 of this 7 document. Okay. Are you with me? Okay. 8 Farm Milk Supply, Areas, Quantities, and 9 Composition, 30. 10 Processing Locations, I have mentioned, page 34 11 described here. 12 Intermediate Products, Description and 13 Composition, yes, although the number and form of 14 intermediate products was greatly expanded in the current 15 version of the model. 16 Consumption Areas, we have demand locations. 17 That's what I would call those consumption areas. 18 We have the Consumption of Final Dairy Products -- 19 Q. Wait a minute, that was page 41. 20 A. Page 41, excuse me. 21 Consumption of Final Dairy Products, that's 22 essentially the demand that needs to be met at different 23 geographic locations in the United States. Here you see 24 that's divided into two basic product categories, fluid 25 and manufactured dairy products. 26 In the testimony, the written testimony submitted, 27 there's a complete listing of both the final intermediate 28 and tradable products that are included in the current 6974 1 version of the model. 2 Dairy Product Composition, so Components in Fluid 3 Milk Products, line 56. Okay? Components in Manufactured 4 Products. 5 One difference that I'll note with the current 6 version of the model is, I believe I'm correct in 7 stating -- and I'd have to go to page 56 to be sure about 8 it -- that fixed values were used for the composition of 9 fluid and manufactured dairy products. 10 One of the modifications that we made -- and I 11 recall this in part because it took a great deal of 12 programming effort and time -- was to actually make the 13 product composition, say, of cheese be endogenous to the 14 ingredients that were used at the processing facility. 15 So there are a number of different ways in which 16 you can make cheese. Not that I'm expert in that 17 category, but I've studied enough the math of the cheese 18 yield process, and we decided that milk coming into a 19 plant would not be the same -- not sort of yield the same 20 product yield as milk from another farm that had a 21 different composition. So we expanded the role of 22 components to account for the fact that milk with a 23 different farm composition could result in a different 24 product yield or require a different sort of make formula 25 that is alternative ingredients to be used to account for 26 an appropriate composition of a product, particularly 27 cheese, that would be complicated mathematically in this 28 case. 6975 1 So I believe I'm correct in recalling that the 2 original formulation was for fixed components, and one of 3 the things that the updated version of the model does is 4 allow those components to be used more appropriately to 5 indicate the product composition when the product is 6 manufactured. 7 Q. If I could interrupt for one second. 8 A. Yes, please. 9 Q. Turn to cost data and transportation cost. And 10 before you go into what was, I think my understanding is 11 that there's been a significant change since 1997 that, at 12 some point, there was a separate transportation function 13 created, separate function model created, that I think 14 probably replaces all of this material about 15 transportation; is that correct? 16 A. So I would have to go remind myself of the 17 specific equations that are here. 18 Q. And I don't want to ask about the specific 19 equations, and I think they're proprietary, and we respect 20 that. 21 So I just -- it's my understanding that some time 22 after 2010, USDS created a separate transportation model 23 that is used to input into the USDS; is that correct? 24 A. Yes. It's different than what was used in this 25 version of the model in 1997. 26 Q. Okay. And so I would rather, in this case, not 27 talk about what was in 1997, because I think what's 28 relevant -- because I assume that that model was then used 6976 1 for purposes of what has been submitted for this hearing, 2 correct? 3 A. Yes. 4 Q. Okay. And I do not, first of all, think I 5 understand all the equations. Secondly, I don't think 6 that they are necessarily public, and so I don't want to 7 go into equations. 8 But if you could generally describe the broad 9 parameters of the equations and what they cover from the 10 separate transportation model. 11 A. Okay. So one of the things that's important that 12 I have somewhat noted before is that this was a joint 13 effort to develop the data sources that were used for any 14 of the updates that we have done to the USDSS since the 15 1997 model formulation. And that joint effort has 16 involved both Dr. Stephenson and myself, and it may be 17 helpful to clarify who did what. 18 So whenever we did this, we relied upon 19 Dr. Stephenson's expertise in looking at farm milk 20 production data and allocating that farm milk to places 21 where the model needed to have a value appropriately in 22 assessing the composition of the farm milk at those 23 different locations. And Dr. Stephenson also contributed 24 to the development of the transportation cost function, 25 and I'll come back to that in a moment if I may. And 26 also, Dr. Stephenson contributed the data on the 27 processing costs for the different products at the 28 different facilities. Okay? 6977 1 So to come back to the transportation costs. What 2 I understand Mark Stephenson to have done is there is a 3 program that is available that -- actually it was an 4 outgrowth originally of an extension related 5 transportation calculator for milk haulers that allows one 6 to estimate the costs of transportation, particularly for 7 milk assembly, but also for final product distribution and 8 for interplant flows. 9 And that cost of transportation function would 10 take into account the core costs for a trucking company 11 that would include overhead and maintenance and 12 replacement of equipment. It would include some notion of 13 the fuel costs, it would include some notion of the driver 14 time required. And Dr. Stephenson would, in a sense, 15 simulate the values of transportation that were required 16 for a particular set of routes, some number of them, and 17 then he would develop a statistical relationship that 18 would show the cost relationship to distance travelled by 19 that particular transportation movement. 20 And so over time, we have used that same basic 21 approach, but to update the transportation cost functions 22 we go back and look at new data for things like fuel costs 23 and wages and -- and overhead and tires and things like 24 that, that are associated with usual trucking costs. And 25 I have to say I work a little bit in transportation 26 logistics, but I don't know how to operate a trucking 27 company. So I'm not sure all the things that might be 28 included in that, but I trust Mark's judgment for that 6978 1 particular thing. 2 Q. Do you understand that tire costs are included? 3 A. Tire costs are included. Apparently that's some 4 kind of big deal. You blow out a lot of tires on a heavy 5 truck and trailer unit. I see them on the road. 6 Q. So Mark was very comprehensive, correct? 7 A. I believe these are reasonably comprehensive 8 estimates of what the transportation costs would be. 9 Q. And in fact, do you know whether for you or for 10 Dr. Stephenson, the model -- this particular -- the 11 transportation model, has been used to consult with 12 members of the dairy industry to help them understand 13 their hauling costs? 14 A. I don't know if anything like the current version 15 of this program has been used to consult. I know that 16 Dr. Stephenson has maintained contact with trucking 17 companies who haul milk in Wisconsin and Minnesota. I 18 don't know if they were providers of information or 19 whether they were receivers of information or both. 20 I do know that a much earlier version of this was 21 actually an extension tool that was available online to 22 help hauling companies understand what their costs might 23 be in part, so that they could set appropriate contractual 24 rates to avoid going out of business. I think that was 25 the original purpose for the tool. 26 Q. So I understand and appreciate National Milk 27 Producers' counsel's questions to you about using 2021 28 data. 6979 1 How much impact does hauling cost have on the 2 actual model results, say, the relationship between 3 locations? 4 A. So that's actually a little bit of a difficult 5 question to answer, because it's fairly common for people 6 to ask me, as I mentioned earlier, why did this change in 7 this way? And so I can say intuitively, if we have higher 8 transportation costs, that will tend to raise the 9 steepness of the pricing surface. 10 But to actually say something other than like, 11 what effect does the change in the diesel price have, and 12 to generalize that to what effect does transportation cost 13 have, to come up with the best answer I would actually 14 need to go away and do two model versions where I have one 15 set of transportation costs, so it includes all the stuff, 16 including the tires, not just the diesel, and then run 17 another of those, and then I can say, here's what 18 difference this makes. 19 If it's helpful, for the purpose of this question, 20 what we do find is that something like a diesel fuel price 21 can have a significant impact on the steepness of the 22 slope but does not tend to change the basic patterns that 23 exist in the spatial milk value surface. 24 Q. I think that was the question I was trying to ask. 25 A. Okay. 26 Q. It's fair to say that the current model has a lot 27 of constraints, correct? 28 A. Yes. I -- I can't exactly remember the number, 6980 1 but I believe it's something like 250,000, maybe more for 2 the bigger of the models. 3 Q. And it has a lot of variables, correct? 4 A. In the millions, yes. 5 Q. And it also has, I think, something called 6 activities. 7 A. Activities are the synonym in linear programming 8 world language for variables. 9 Q. Okay. 10 A. So they would include things like how much milk 11 moves, how much milk is processed at a plant, how much 12 product is produced at a plant, how much product is 13 consumed at a location. 14 Q. Does the model make any assumptions about whether 15 it is the producer or the processor who bears the hauling 16 transportation cost? 17 A. No, it does not. 18 Q. So in terms of, say, maybe one example of what has 19 changed over time, I think I understand from one 20 conversation I had with somebody that -- and I don't know 21 when it changed, but at one time, the model produced a 22 surprising result in terms of moving milk from the west 23 side of one of the Great Lakes to the east side. 24 Does this -- do you have a recollection about 25 this? 26 A. I can envision this being a kind of conversation. 27 I don't remember a specific instance that I can say, yeah, 28 it was this location to that location. 6981 1 Q. I think -- so, for instance, one of the -- the 2 kinds of things you do is when you learn about that, you 3 say -- you put a limitation in that says, you can't cross 4 a lake without a bridge, correct? 5 A. Well, let's -- maybe I can clarify. That the 6 current version of the model is based on the existing road 7 network. We don't assume that the truck is going to get 8 on a ferry and go across Lake Michigan, even though maybe 9 in principle that's possible for some trucks. So it does 10 rely on the existing road network. 11 And if I'm remembering correctly, there were times 12 in past years where there was milk that could not be 13 processed in Southern Michigan and actually ended up going 14 into Wisconsin. So it went around the lake and through 15 Chicago and up into Wisconsin. 16 But the -- all the movements have to be consistent 17 with the existing road network. And so there's no kind of 18 imaginary line that connects the city on one side of the 19 lake with another side of the lake. 20 Q. So recognizing -- so let me go back to federal 21 order reform. 22 Do you know that USDA itself, in Federal Order 23 reform, took the results of the model and made 24 adjustments? That is to say, USDA made the adjustments 25 based upon its knowledge. Do you know that? 26 A. So I was not a part of the process of doing the 27 modeling work that contributed to this document in 1997, 28 and to the discussions, and I certainly was not in the 6982 1 room, if there was a room when AMS was having 2 conversations about this. So the knowledge that I have is 3 secondhand knowledge that comes from the people who were 4 involved in that process. And that's the basis for my 5 statement that I believe there's a similarity between the 6 use of model results and making adjustments and -- in that 7 era and what is being done here. 8 Q. The difference being that one thing maybe -- if 9 USDA is doing it versus private industry. 10 A. I would say that's a difference. 11 Q. So if an adjustment is made, say, to, you know, 12 one county, and just by hypothetical, $0.50 to the value, 13 does that then alter the entire model because you are 14 talking about the relative value of milk, or does it 15 create some -- if you don't build it back in, some 16 inconsistencies because the model would have said it 17 should be X and now it's Y? 18 A. Again, maybe it's helpful to clarify that, again, 19 the model doesn't know anything about Federal Orders, it 20 only knows really about transportation costs. And also, 21 all of the adjustments that were made, again, as I 22 understand it in the previous 1997, 1998 process, and the 23 adjustments that were made here, were made after the fact 24 based on the existing model results. Okay? 25 So what we can't say with this particular model 26 and tool is if you added $0.50 here, is that going to mess 27 everything up in Federal Orders, because the tool is 28 simply not designed so address that question. 6983 1 Q. I understand. 2 But leaving Federal Orders aside for a moment, 3 does it impact the value in other counties if you have 4 altered one county but not others? In the model itself, 5 if you left out Federal Orders, would it alter the value 6 in other counties? 7 A. The model does not allow us to actually change a 8 value in one county arbitrarily. The model provides a 9 result of what the value in the county would be, and it is 10 what it is, again, based on the millions of interactions 11 that are part of the model structure. 12 So there's no way to go in -- you can make a 13 change to something like there's no plant there anymore, 14 or there is a plant there, or we lost all the farms in 15 that county, and you can evaluate what the impact of that 16 would be in a multi-county area. But we can't actually go 17 in and change one value and then see what happens with 18 everything else. It's just not the way the model is 19 designed. 20 Q. Okay. So I want to turn to your discussion with 21 the counsel for National Milk for Canton, New York. And 22 part of it is, I want to make sure I understand it, and 23 then I want to understand if there is a limitation. 24 It sounds like by making a change for an 25 individual plant, it wouldn't change the big picture. It 26 doesn't alter the big picture as you said, that's -- 27 that -- I think that's understood; is that right? 28 A. So if you were to open or close one plant of a 6984 1 modest or moderate size, it would not change the bigger 2 picture. 3 Q. But would it change the localized impact? Would 4 it have a localized impact? 5 A. So changing the availability of a plant, either 6 bringing one online or taking one offline, can have an 7 impact close by that plant, but it depends on a number of 8 factors. 9 One factor would be, are there any other plants 10 anywhere near that location? In the Canton case, that 11 really was the only plant within a number of counties, and 12 it was actually processing a fairly large volume of milk. 13 So if there are other plants that can take up that 14 milk, then the impact is likely to be fairly small. And 15 it also, as I mentioned, depends on how big a plant was 16 that. If it was a fairly large plant, the impact would be 17 likely to be larger; if it was a fairly small plant, then 18 even the localized impact -- there will be some. The 19 numbers will always be somewhat different. But in 20 general, it takes a fairly large plant in a location where 21 there aren't alternatives to have a significant localized 22 impact. 23 Q. And what happens if you add a demand center in a 24 county that doesn't otherwise already exist? 25 A. I guess we'd actually have to be creating people 26 somehow to make that happen. But essentially we have -- 27 we have full coverage of the population of the demand for 28 dairy products now. We could do something, and we have 6985 1 done things like this in the past, where we say, add an 2 export demand. 3 It's actually one of the things that's also 4 different from the current version of the model to the 5 previous version of the model. We actually account for 6 the specific product exports by port district in this 7 model, and we allow dairy product imports. The export 8 side, of course, is much more important now than it used 9 to be. 10 So if we had, say, a scenario where there was 11 increased export demand, we could actually evaluate sort 12 of the systemwide impacts of that, and we could actually 13 do it either in a particular port location like Los 14 Angeles, or we could do it spritzing a percentage increase 15 across all the different port districts. And we have done 16 some of those kinds of analysis as a part of work that we 17 did for the state of Pennsylvania in 2017. 18 Q. So I want to go back to your answers to questions 19 about the sort of the base. 20 A. Okay. 21 Q. How would the gradient change if the base was 2.20 22 versus $1.60? 23 A. So the price relatives are the same regardless of 24 you choose $1.60 base or 2.20 base. Because the model 25 starts from the assumption that we have a minimum value 26 that would be, in the case of what we have actually 27 presented, $1.60, and then we build the price relatives 28 off that. 6986 1 So on a percentage basis, maybe the slope changes, 2 but the slope would actually be the same regardless of 3 what the base would be. And that's kind of consistent 4 with the idea that this model was not designed to figure 5 out what the base should be. 6 Q. The gradient is -- a gradient is independent of 7 the base, correct? 8 A. Yes. 9 Q. Okay. So could you go back to showing Figure 3 10 that National Milk just counsel brought up? 11 So I think maybe there was -- perhaps, on 12 somebody's part, and it may be mine -- a misunderstanding 13 of what I thought I was driving at with respect to the 14 questions about Miami. 15 And I certainly understand the fact that I was 16 there in 2007, so I know about the Southeast decision. 17 And I know, I think, what that creates, a bit of a ridge 18 to the north of it. 19 But Miami is green -- I know I'm -- my eyesight's 20 not great, but Miami is in the green section, correct, of 21 the map? 22 A. Yes. 23 Q. And Minneapolis is somewhat in a blue section, 24 correct? 25 A. Yes. 26 Q. And blue is a smaller increase in this than green; 27 is that correct? 28 A. Yes. Confusingly, based on the color scheme in 6987 1 this diagram, it's not a defined gradient, yes. 2 Q. Okay. So I think the point of my questions was 3 what about -- well, let me start over. 4 The model suggests that the Miami price should go 5 up more than the Minnesota price, correct? 6 A. Well, the model suggests that the difference 7 between the model-generated spatial of milk price values 8 in the current Class I differentials is larger in Miami 9 than it is in Minneapolis, yes. 10 Q. And Denver is in a blue -- I think dark blue area, 11 correct? 12 A. I actually -- I can't exactly tell. I think it's 13 to the west of that dark blue -- 14 MR. ENGLISH: Blue or purple -- 15 (Excessive crosstalk.) 16 THE COURT: Which one of you would like to talk? 17 MR. ENGLISH: He would like to talk. 18 THE WITNESS: I believe that the Denver area is to 19 the west of the blue. 20 BY MR. ENGLISH: 21 Q. So in the purple? 22 A. It's purple or some shade of purple, yeah. 23 Q. And purple's even -- 24 A. Purple is a smaller value than the blue -- 25 Q. Okay. 26 A. -- in this scheme of colors. 27 Q. Okay. 28 A. Which, by the way, I must give full credit to Mark 6988 1 Stephenson for also doing the mapping. 2 Q. He'll appreciate that. 3 So that would suggest, as with Minneapolis, that 4 the relative value spread between Denver and Miami has 5 gotten larger, correct? 6 A. Yes. Because that is a smaller value difference 7 there than in Miami. 8 Q. And finally, for Riverside, which I believe is 9 also purple, it's a -- that would be the same as Denver, 10 correct? It should be a greater spread? 11 A. Yes. Roughly approximately equal to the Denver 12 value. 13 Q. So I would like now to have you turn to 14 Exhibits 300 and 301. 15 A. I was hoping you would ask. 16 THE COURT: So you have those? 17 MR. ENGLISH: He has his own set, but I guess Your 18 Honor may need a set again. 19 THE COURT: I have two sets. Does anybody need 20 one? I do. I have two. 21 MR. ENGLISH: My briefcase is full enough. All 22 right. 23 BY MR. ENGLISH: 24 Q. So first, as I talk about this, I have been told 25 to stop calling them line numbers and call them row 26 numbers. 27 So you were here, as you said with counsel for 28 National Milk, riveted in the conversation I had with 6989 1 Dr. Vitaliano, correct? 2 A. I was hoping you would do more line numbers, yes. 3 Q. Well, I'm actually hoping not to do any line 4 numbers. I think I mostly want to do column numbers, or 5 column letters. 6 A. Okay. 7 Q. Can you confirm with me -- and I'll try to shorten 8 it -- that Columns A, B, C, D, and E are literally there 9 because of that's how the Federal Order uses them? 10 A. Yes. Can I actually ask, though, are you 11 referring to 300 or 301? 12 Q. For now, 300. 13 A. 300. Okay. 14 Yes. Those are reporting columns. 15 Q. Okay. And then the results of your model were 16 Columns F and G, correct? 17 A. Yes. 18 Q. Have you ever seen, before today, columns after 19 Column G? 20 A. Yes. So in fact, in the results that were 21 provided to National Milk, we had Columns F, G, H, which 22 is the October to May differences; Column I, the current 23 differentials; Column J, May current; and Column K, 24 October to current difference. That was information -- 25 the core information that was derived from the model were 26 Columns F and G; the core information from the existing 27 Class I differentials was Column I; and the others were 28 basic calculations reporting differences. 6990 1 So Columns F through K were things that were 2 reported to National Milk from us. 3 Q. Okay. Thank you for that clarification. 4 But after that, did -- did you do Column L? 5 A. No. 6 Q. Okay. Or Column M? 7 A. No. 8 Q. Column N -- well, that's easy, that's just an 9 order number. 10 A. Yeah. 11 Q. Basically, to your knowledge, none of the other 12 columns after you get past Column K? 13 A. No. None of the other columns past Column K 14 were -- had anything to do with the information we 15 provided. 16 Q. Let me specifically ask about Column R. This is 17 the average monthly pounds you -- have you ever seen that 18 column before? 19 A. So we actually do have that information in the 20 model because we have to have -- actually, no. Sorry. 21 No. Column R is not something that we reported, because 22 we only have monthly information on pounds of milk for the 23 months of May and October 2021. So we -- and I also do 24 not believe that we provided that information on the milk 25 pounds for those two months to National Milk. So Column R 26 is something different. 27 Q. Okay. So do you have with you, or can you make 28 available if National Milk permits, what you gave National 6991 1 Milk so the record can have that? 2 A. So I actually was checking my computer to see what 3 I did and did not have. It turns out that the model can 4 run on a laptop, but actually runs on a University of 5 Wisconsin laptop that has more bandwidth, and most of the 6 files that are associated with that are on that laptop and 7 not on this one. 8 So the question of whether or not it can be made 9 available, I suppose, is up to discussion by the National 10 Milk team since they had sponsored this particular 11 research. 12 Q. I only asked -- just -- this is not for you, this 13 is for the record. In terms of the foundational piece, I 14 just -- it seems to me it would help for foundation, but I 15 leave that to National Milk. 16 A. May I also add, though, that I don't have any 17 reason to doubt that the information that's here is what 18 was provided. 19 Q. And that's fine. Thank you. I appreciate that. 20 And I think -- I think I got an answer from 21 National Milk, but -- and maybe there was also a question; 22 I may have missed it from counsel. I don't believe that 23 National Milk consulted with you about your columns in 24 terms of any modifications that they proposed, correct? 25 A. No. 26 Q. Okay. Are there any areas from your model results 27 that indicated -- apparently one of us had a double 28 negative or something. 6992 1 THE COURT: What -- it was unclear exactly what 2 you just established. 3 MR. ENGLISH: So that's two people who advised me 4 of that. Obviously it was. 5 BY MR. ENGLISH: 6 Q. National Milk Producers Federation did not consult 7 with you about any of their modifications, correct? 8 A. That is correct. They did not consult with either 9 me or Mark Stephenson about any of their modifications. 10 Q. Are there any areas, any counties for your model, 11 that indicated using the $1.60 base, the Class I values 12 would decrease rather than increase? 13 A. Yes. 14 Q. Hasn't production grown faster than milk 15 requirements for Class I in a number of areas? 16 A. Can I ask you to repeat that? Sorry. 17 Q. Hasn't milk production grown faster than the 18 requirements for Class I in a number of areas? 19 A. That may be true, but I guess I don't have data 20 before me to help establish that. 21 Q. Would -- would a value at a $1.60 that is lower 22 than the present suggest that at least as to that 23 location, production has grown faster than the 24 requirements for Class I? 25 A. So if we're looking at -- let me again, maybe make 26 a little bit more of a lengthy explanation of this. 27 So what determines the values at a particular 28 location, I have previously said is somewhat difficult to 6993 1 assess because it relies on not just the values at that 2 particular location, but also the system values, which, 3 again, involves these millions of points of data. 4 So the factors that influence whether that value 5 would go up or down or be higher or lower than the current 6 Class I differentials do depend on the local milk supply 7 and demand balance at that location, but they also do 8 depend on the interaction of all of the other connections 9 within the modeling structure. 10 So I guess I would say it's maybe a little bit 11 oversimplistic to say, well, if it went down by a nickel, 12 that must mean that there was more than enough milk 13 available at that location, because it depends on the 14 systemwide impacts to really be able to say that. And as 15 I indicated before, highlighting the change at any one 16 location is actually pretty difficult given that it's part 17 of this broader system set of outcomes. 18 Q. Thank you for that correction/clarification. 19 A. Trying not to be too much of a lecturing 20 professor. 21 Q. So could you put Figure 3 back up again? That was 22 fast. 23 So given what you see there in the purple sections 24 in the central part of the country, does that at least 25 suggest that, say from the Great Plains from the north to 26 south, at least until you get to Central Or East Texas, 27 that the value for milk in Class I relative to other 28 locations has gone down? 6994 1 A. I'm sorry, can you maybe parse that question out 2 into the different components for me again? 3 Q. So going back to Colorado for a moment. 4 A. Okay. 5 Q. Relative to places to the east where you see the 6 green, I think what we're talking about -- you know, for 7 Miami would apply also, say, comparing to, say, where we 8 are today in Indiana, that the value -- the relative value 9 of milk used in Class I has gone up more in Indiana than 10 in Colorado, correct? 11 A. Again, the difference in the spatial milk values 12 between the model and the Class I differentials, yes, is 13 larger in Indiana than in that part of the country. And 14 maybe it also helps, I like to think about this as having 15 the slope is now steeper, right? 16 Q. The slope is steeper -- 17 A. The slope is steeper in the model than it is in 18 the Class I price surface as we have it today. 19 Q. It is steeper moving west to east. 20 A. Yes. I guess either way, it depends whether you 21 are going up slope or down slope. But, yes. 22 Q. Steeper up? 23 A. Steeper up. 24 Q. You agree with me that prices in the -- that 25 values in the fall are generally going to be higher than 26 the spring, correct? 27 A. So we actually report a seasonal price difference, 28 I think it's in the column here on H on Exhibit 300, and 6995 1 we also have a mapping of that in the full report. It's 2 at testimony. And, yes, that generally is the pattern, 3 although there can be exceptions where there is no change. 4 And I believe there may be a couple of places where there 5 is a decrease in between October compared to May, but 6 generally the pattern is higher values in October. 7 Q. So recognizing the model doesn't know what Federal 8 Orders exist, nonetheless, this is likely to be used in 9 some way for Federal Orders, correct? 10 A. I guess I have no idea whether this will be used 11 or not. I guess I kind of like to think there's some 12 input that's valuable. 13 Q. If we set prices based upon the fall run when 14 values are generally higher than the spring, given the 15 fact there's a minimum pricing system, would that suggest 16 that we result in pricing being too high during the spring 17 flush? 18 A. I guess we can't really conclude that on the basis 19 of the model analysis, because the model does not know 20 about Federal Orders. 21 And I will remark here that about five years ago 22 we recognized that one of the limitations of a tool like 23 this is that it's primarily got a supply chain value 24 focus, and so it does not allow us to examine the 25 implications of making changes to the Class I differential 26 surface in terms of what that would mean in terms of blend 27 prices or supply responses or demand responses. 28 And so with some fairly minimal support from AMS, 6996 1 we actually began working on an alternative to that that 2 would allow for more of that responsiveness, in essence, 3 to develop a model that would allow me to give you a 4 better model-based answer to your question about, is this 5 kind of change too high? 6 But we actually can't do it on the basis of the 7 analysis that we have here, because it relies on 8 understanding the impacts of that change, and this model 9 is just reporting what is, not what would happen. 10 Q. Since this is a minimum price, since the Federal 11 Order system is a minimum pricing system, do you have a 12 view about using the minimum over the average or the high 13 value from your model? 14 A. I don't have a recommendation to make in that 15 regard, because I think I -- as I mentioned, this model 16 doesn't allow us to think through from a research 17 perspective what the implications of using one of those 18 values versus another would be, because it doesn't include 19 the regulation structure under orders. 20 Q. So how is organic milk accounted for in the model? 21 A. It's included in the total milk supply, and it's 22 not treated any differently than the conventional milk 23 supply. 24 Q. Would you expect -- I understand it's included 25 that way, but if it were separated out, would you expect a 26 different result for organic milk? 27 A. So what I know of the distribution network for 28 organic milk and its processing differences from 6997 1 conventional milk are that it tends to move longer 2 distances, because there are smaller processing volumes 3 and smaller shipments going into individual stores. That 4 would tend to suggest that there would be a higher supply 5 chain cost for that distribution network, but I haven't 6 done any work to analyze what that difference would 7 actually be. 8 Q. And following up on -- on some of the questions 9 from National Milk, would you agree that any model has 10 inherent limitations? 11 A. Yes. 12 Q. Can this model tend to reinforce current market 13 conditions? 14 A. So I'm not sure this model has a lot to say about 15 whether it reinforces current marketing conditions or not. 16 Because as -- again, as I mentioned, it's not actually 17 modeling a market, it's modeling the logistics of milk 18 assembly and distribution. So it's providing what I think 19 of, again, as a competitive benchmark for the costing 20 structures. 21 And it -- again, it -- it knows something about 22 what milk goes into what plant. The only place that it 23 really recognizes a class distinction is really through 24 the addition of the 1.60 minimum value that's applied to 25 Class I milk. Otherwise, it doesn't really know or 26 doesn't really care where the milk is being used or what 27 the current market structure is or whether some product 28 was at a Class IV plant or Class III plant. 6998 1 Q. And as you have stated to me and also to counsel 2 for National Milk, the model doesn't consider the 3 existence of Federal Milk Marketing Orders, correct? 4 A. That's correct. 5 Q. And so it doesn't consider depooling or 6 performance standards or PPDs or anything like that, 7 correct? 8 A. It does not. 9 Q. So would it be fair to say that the USDSS is more 10 like a traffic cop but not a driver of the cars? Sort of 11 reports on what's going on, but it doesn't actually drive 12 the cars? 13 A. So the USDSS takes existing market conditions and 14 provides us with an assessment of what the spatial milk 15 value surface would look like. And in economics, we can 16 talk about different kinds of models where a key 17 distinction is whether or not the model has a behavioral 18 component, right? 19 And by that I mean people will base their 20 decisions or the behaviors on the outputs of the model. 21 In this case, this is a pure, let's minimize costs 22 throughout the system. I described it earlier as being an 23 impassioned -- or a dispassionate dairy dictator that did 24 not care about where a plant was currently located, it 25 wants to get the lowest overall costs. 26 In that sense, this is not a behavioral model, 27 because the only behavior that's being represented here is 28 that ruthless cost reduction idea. So it doesn't include 6999 1 the responses of any potential dairy producer or dairy 2 processing facility or dairy consumer to the imposition of 3 these Class I differentials. But we actually don't have 4 any model that would allow us to analyze those questions 5 at this point, which is the reason that we began working 6 on an alternative several years ago. 7 Q. So I want to go back to your very last 8 statement -- your last sentence in your statement, which 9 sort of used an analogy to weather forecasting. 10 Do you remember that? 11 A. I absolutely do. 12 Q. And you reference sort of like the national 13 weather forecast, and then there's the local forecasters, 14 correct? 15 A. Yes. 16 Q. You are a fully neutral expert applying this 17 model, correct? 18 A. Yes. 19 Q. And in the case of local weathermen, they don't 20 have a financial stake in whether or not it rains or the 21 sun shines, do they? 22 A. I guess it depends on whether they have invested 23 in solar panels or something like that, but generally I 24 would say, no, they do not. 25 MR. ENGLISH: Thank you. That's all I have. 26 THE COURT: Well done, Mr. English. Thank you. 27 Do we need five minutes to stretch before the next 28 cross-examiner comes forward? 7000 1 MS. TAYLOR: Yes. 2 THE COURT: Yes. Okay. Please be back and ready 3 to go at 3:22. We go off record at 3:16. 4 (Whereupon, a break was taken.) 5 THE COURT: Let's go back on record. 6 We're back on record at 3:26. 7 Who next has questions for Dr. Nicholson? 8 Mr. Miltner. 9 CROSS-EXAMINATION 10 BY MR. MILTNER: 11 Q. Good afternoon, Dr. Nicholson. 12 A. Good afternoon. 13 Q. This is a change. My name is Ryan Miltner. I 14 represent Select Milk Producers. 15 So I want to start by perhaps exposing my 16 ignorance. You noted that the model does not, you said, 17 doesn't recognize Federal Orders or it doesn't take into 18 account Federal Orders; is that correct? 19 A. It does not take into account the full set of 20 regulations under Federal Milk Marketing Orders. It does 21 address different class plants as a part of calculating 22 what a spatial price surface would be, but it doesn't 23 represent the full range of incentives that would be 24 available to dairy farmers, dairy co-ops, dairy 25 processors, or consumers. 26 Q. That's helpful. 27 When the model attempts to create or determine the 28 spatial value of milk, is it looking to determine the 7001 1 value of that milk to a plant that is purchasing it or the 2 value received by the producer who might or might not be 3 pooled on a Federal Order? 4 A. I guess I'd say the model is not really 5 representing either of those things. What it's looking at 6 is what I call the price relatives: The margin, the 7 differences in marginal value, the value of an additional 8 hundred pounds of milk at a location. So it's not, for 9 example, representing the pay price that a producer would 10 get or the payment that a Class I plant would make. It's 11 only looking at based on essentially the supply chain and 12 logistics costs that are relevant there what the 13 differences are in spatial values of milk. 14 Q. So for that additional hundred pounds of milk, 15 when it's -- when it's determining that value -- and I 16 hope I'm -- we're not ships passing in the night or 17 whatever -- but is it -- the value that is determined, 18 whose -- whose value is determined? 19 A. Yes, I come back. It's no one's specific value 20 because it's a marginal value of having an additional 21 hundred pounds of milk at that plant at that time. 22 Q. Okay. I'd like to ask about the $1.60 base. 23 And do you still have a copy of Exhibit 300 with 24 you? 25 A. Yes, I do. 26 Q. Excellent. I brought my copy as well. 27 So this is Row 518, Ada County, Idaho. 28 Which one is Ada -- that's where you live -- 7002 1 that's where Ms. Hancock lives, so -- 2 A. Must be an excellent location. 3 Q. I'm sure it is. 4 When I look at the row for Ada County, Idaho, I 5 see as I look down Column F, that the May '21 model 6 returned $1.70 as its output, and then October '21 7 returned $1.60 as its output. 8 Aside from the fact that that -- that's a bit of 9 the inverse that the fall would usually be a higher value 10 than the spring, does that reflect that that value for Ada 11 County in October was no higher than the base value? 12 A. So -- yes. So the $1.60 implies that that is the 13 value with the addition of $1.60 to the original model 14 result. 15 Q. Okay. Granted there's over 3,000 rows to look at, 16 and I suppose I could have done some sort of sort function 17 too, but I did not see $1.60 show up anywhere else in 18 either of May '21 or October '21. 19 Do you happen to recall if there was another 20 instance where $1.60 appeared? 21 A. So I have to admit that I don't recall of the 22 3,108 different values that we're looking at in two 23 months, so over 6,000 values, but perhaps I can shed a 24 little bit of the intuition of what is happening there. 25 Q. Great. 26 A. So that October value for Ada is suggesting that 27 at that month, at that county location, the model was 28 saying there is no positive marginal value for milk in 7003 1 that month at that location. That is, nobody would want 2 to pay to have an additional hundred pounds of milk at 3 that location in that month, right? 4 So that is the raw output of the model that would 5 be used to adjust to the current Class I differential 6 surface by adding $1.60. So essentially the model says 7 that value for us at that month is zero. To make it 8 consistent with the minimum current pricing system so that 9 we can compare apples to apples throughout the spatial 10 analysis, that value is set to 1.60. 11 I don't know if there are other values for May 12 that are equivalent to 1.70, but what that's suggesting is 13 that the value of an additional hundredweight of milk at 14 that location in May was $0.10. Somebody would be willing 15 to pay only $0.10 to have an additional hundred pounds of 16 milk at that location at that time. And then we added the 17 1.60 to get to 1.70. 18 THE COURT: Mr. Miltner, may I ask probably an 19 ignorant question, but I -- I am just puzzled. When an 20 area is not regulated, where do you get your inputs for 21 the model? 22 THE WITNESS: So you are saying that to me? 23 THE COURT: Yes. 24 THE WITNESS: Okay. So, again, the model does not 25 rely on information about whether an area is regulated or 26 not regulated, it only relies on the core data of the milk 27 supplies and composition, the population and the demand 28 for dairy products, the plant locations and their 7004 1 capacity, and the transportation network that connects 2 them. 3 So the model itself is not trying to capture the 4 current full regulation, it's trying to represent a 5 competitive cost benchmark in the absence of regulation. 6 THE COURT: Is milk composition readily available 7 online for an unregulated area? 8 THE WITNESS: So much of the milk composition 9 information comes from state-level national statistics 10 offices. They -- for major dairy states, they tend to 11 report more information than for states that are not major 12 dairy states. For some states what we need to do is use a 13 statistical relationship that compares the butterfat value 14 to the other components in the milk because that 15 information is not reported. Information is also 16 sometimes reported through the Federal Milk Marketing 17 Order statistics for things like milk composition that we 18 can also use to develop the composition at different 19 locations, whether they are regulated or not. 20 THE COURT: Thank you very much. 21 THE WITNESS: Thank you. 22 MR. MILTNER: Changing gears a little bit. 23 BY MR. MILTNER: 24 Q. Mr. English asked you about the relative values of 25 milk in Colorado and here in Indiana. 26 Do you recall that questioning he asked of you? 27 A. Yes. 28 Q. Okay. I understand that the model when making -- 7005 1 if you are comparing those two locations, what the model 2 shows is that the change in the incremental value of milk 3 in Colorado is a lower magnitude than the change in the 4 incremental hundredweight in Indiana. 5 A. Again, I think I would be careful about using 6 change so much as saying there is a divergence between the 7 current Class I value at that location that is bigger in 8 Indiana than it is in Colorado. So -- because we're not 9 actually making a change in the values, we're just looking 10 at the fact that there is a difference between those 11 values and those locations and the size of the difference 12 is different. 13 Q. And so it's not that the -- it's not that the 14 value -- I'm not going to get my terms right. It's not 15 that the value of milk in Colorado has declined since 16 1998, it's that the rate or the magnitude of the 17 difference between the measurement in '98 and the 18 measurement in '21 is smaller in magnitude than the same 19 measurements for Indiana? 20 A. Again, so if we take the current Class I 21 differential surfaces representing 1998, but actually we 22 should also probably incorporate the changes that were 23 made in 2008. That's why I want to be a bit careful about 24 this. 25 What it's saying is that the divergence has become 26 bigger in a place like where we are here in Indiana than 27 it became in -- in that supply location in Colorado for 28 that plant location in Colorado. 7006 1 Q. Much more articulately stated than I tried to. 2 A. I'm trying. 3 Q. Okay. Do you have a copy of your testimony as you 4 presented it, NMPF-36A? 5 A. Yes, I do. 6 Q. Great. I'm looking at Figure 1 there. And so I'm 7 looking at the state of New Mexico, and you have got a 8 plant there in Albuquerque. And, again, the yellow lines 9 show the flows of the product from the plant to the 10 consumer, correct? 11 A. That's correct. 12 Q. Okay. Now, I note that for that plant in 13 Albuquerque, as least as it's presented here, there are no 14 green lines. 15 Can you explain how -- how milk inflows to plants 16 are represented on here? 17 A. Okay. So the absence of a green inflow line of 18 that Albuquerque location is indicating that the milk that 19 went into that plant originated at that same location. So 20 one of the things that's important to note about this 21 particular figure is that the mapping is based on the 22 smaller model configuration that did not have -- it had 23 multiple counties aggregated into one supply location, 24 like at Albuquerque, and therefore, it doesn't pick up the 25 full disaggregated information that would be all of the 26 counties which would actually definitely show a green line 27 if that was going to be a supply plant that was 28 distributing to those other locations. 7007 1 And the reason to use this small model version 2 rather than the big model version, is if you have 3,108 3 different lines going to every consumption point in the 4 United States, you basically can't discern any reasonable 5 patterns with that information. 6 So this is trying to represent the kind of 7 information that is provided by the model, but there's a 8 much more detailed distribution routing that would be 9 present in the full 3,108-county model that's not shown 10 here because it's just too messy. 11 Q. So, for example, in as -- if the milk supplies in 12 Albuquerque and Bernalillo County had -- had declined as 13 farms had gone out of business and milk was being pulled 14 from the Roswell area in Eastern New Mexico, those 15 counties could be aggregated under -- under the model as 16 it was used to create this map, and it wouldn't show that 17 milk movement across the state graphically? 18 A. Yes, that's correct. So it's not going to show 19 the movements of milk that would come from anything other 20 than the multi-county region that's represented here. 21 Q. And the same would be true for the plants in El 22 Paso there, correct? 23 A. Yeah. A similar situation applies there, though 24 I'm -- I think there may be a green line underneath that 25 that's may be coming from Las Cruces going down to El 26 Paso. 27 Q. There could be. 28 A. I can't tell. Yes. 7008 1 Q. Okay. Now, New Mexico is the ninth largest state 2 in dairy production, and they have lost production that's 3 down about 10% year over year in '23. 4 So that loss of production, since it occurred 5 since this model was run, wouldn't be incorporated into 6 the model, correct? 7 A. That's correct, because the model is based on 2021 8 data. 9 Q. Time still flies. I mean, from -- because it's a 10 spatial model, if -- if a significant milk shed had a 11 decline of 10% of its milk supply, would that -- do you 12 think that would have a meaningful effect on the output? 13 A. So it could have a meaningful effect on the 14 output. It does depend on the -- as you mentioned, the 15 percentage decrease, the magnitude of that, but also what 16 the base was that we started from. So I don't know what 17 the milk production is in Rhode Island, but if we lose 18 100% of it, it doesn't make any difference to anybody. 19 Q. Right. 20 A. So to give you a better answer to the question, we 21 would actually have to look at that magnitude of milk 22 production reduction, preferably in the full county model 23 to have the better spatial disaggregation, and we can 24 assess what the impact would be. In general, we would 25 expect that the marginal value of milk in that region 26 would be lower, sorry, might be higher with the lack of 27 milk that's available from that location. 28 Q. Mr. English did a fine job of eliminating lots of 7009 1 my questions. Actually, I think he's done a fine job of 2 getting rid of the rest of my questions, so I don't have 3 anything further. 4 MR. MILTNER: Thank you very much, Dr. Nicholson. 5 THE WITNESS: Thank you. 6 THE COURT: Mr. English? 7 CROSS-EXAMINATION 8 BY MR. ENGLISH: 9 Q. So a bit of a follow-up, and from what Mr. Miltner 10 just asked, I'm not sure if we need to bring the map up or 11 not. 12 How big are the multi-county regions? 13 A. They vary depending on the state location. I 14 actually wonder if there is -- yeah. So I do have in 15 97-09, R.B. 97-09, these have been modified somewhat for 16 the small version of the model, but you get the basic idea 17 that we have denser representation. 18 Q. Sorry, what page are you looking at? 19 A. Sorry, 32, and thank you for asking me about that. 20 MR. HILL: What exhibit? 21 MR. ENGLISH: This is not an exhibit. This is the 22 official notice document. 23 THE WITNESS: Yeah. This is not the exact replica 24 of the multi-county regions that are present in the small 25 version of the model, but it indicates that areas where we 26 have a greater proportion of the milk supply, at least at 27 that time, received smaller aggregations of counties into 28 the multiple-county region. 7010 1 BY MR. ENGLISH: 2 Q. So that's, of course, back in 1997. 3 What are the multi-county regions now? 4 A. Well, they are similar to this, but I can't tell 5 you just by looking at this what the differences are 6 between the current version of the multi-county, what we 7 call the small model, and the county-level analysis. 8 Q. But we know, of course, it has definitely shifted 9 west, correct? 10 A. Yes. So the dots that are represented on that 11 Figure 6 do indicate the locations and the relative 12 magnitudes of the milk supplies in those multiple-county 13 regions. 14 So for example, if you look at the Central Valley 15 of California, it's nearly every county has its own 16 representation. And each of those dots is large. There's 17 a somewhat similar story in Wisconsin, and there's a 18 similar story in New York. 19 So the principle that's being applied here is the 20 same. I can't tell you without actually doing further 21 checking what the differences are. 22 Q. So just to be clear, we don't have in the record 23 what the multiple-county supply areas are in 2021 for the 24 purpose of the model? 25 A. For the small version of the model, we do not. 26 Q. Okay. I'm sorry. I may have missed something. 27 Is there a large version of the model that we do 28 have it for? 7011 1 A. Yes. Actually in my -- well, we don't have a 2 map -- well, we do not have, I think in the record, a map 3 that shows the county-level milk supplies in a way similar 4 to that Figure 6 on page 32 of this document. We also 5 don't have the similar version for the multi-county 6 regions for the updated version of the model. 7 All I was trying to do by showing you that figure 8 is to illustrate the basic idea that back in 1997 we 9 couldn't solve a model that was bigger than this, and it 10 was a challenge at the time. It required actually like 11 mainframe computing. And so the multiple-county regions 12 were put together to allow the problem to be tractable, to 13 be solved, to provide information. 14 Q. So how many -- there were 240 multiple-county 15 supply areas back in 1997. 16 How many are there today, do you know? 17 A. I don't know off the top of my head. 18 MR. ENGLISH: Thank you, sir. 19 THE WITNESS: Thank you. 20 CROSS-EXAMINATION 21 BY MR. MILTNER: 22 Q. Just to quickly follow up on that. The 23 calculations that are really the output that National Milk 24 has utilized, that was based on an analysis of all 25 counties, right? The aggregation of counties was really 26 limited to providing that figure showing flows of milk; is 27 that correct? 28 A. That's correct. Although I would note that the 7012 1 differences between the analyses, we actually ran the 2 smaller model first because we had difficulties getting 3 the larger model to solve because of its extended size. 4 But we also recognize that it's a lot easier graphically 5 to show the flows from that model, even if it loses some 6 of the spatial detail. 7 I'd also -- I guess I would say that the flows are 8 there more to illustrate the kinds of outputs from the 9 primal side of the model than to be a focal point for what 10 the marginal values of milk would be. 11 MR. MILTNER: Thank you. 12 THE WITNESS: Thanks. 13 THE COURT: Mr. Rosenbaum. 14 CROSS-EXAMINATION 15 BY MR. ROSENBAUM: 16 Q. Steve Rosenbaum, International Dairy Foods 17 Association. 18 So you described the map as establishing, if you 19 will, the value of marginal milk at a particular location, 20 correct? 21 A. It's the marginal value of milk at a particular 22 location. 23 Q. But, obviously, for these purposes, we're 24 considering whether to what extent to use these for 25 purposes of setting Class I differentials, correct? 26 A. That's my understanding. 27 Q. So that at that point, they become an actual 28 payment obligation by processors, correct? 7013 1 A. Well, so -- 2 Q. Differentials -- 3 A. -- the -- let me be clear. The model is 4 generating spatial milk values at a differential surface. 5 Whether that is used to create a system that results in 6 processor obligations is not -- that's not part of what I 7 have analyzed. It's -- it's not part of what the model is 8 designed to do. 9 It's suggesting that there are differences in the 10 spatial value of milk that might be accounted for in 11 setting or making adjustments to a Class I price surface. 12 Q. And I mean, to the extent that they are relied 13 upon for that purpose, they will then help set the actual 14 minimum price that has to be paid, right? 15 A. That depends on the extent to which any of the 16 results are relied upon for that purpose, yeah. 17 Q. But you do understand that's the reason we're 18 looking at this information, right? 19 A. I do understand that there is interest in 20 evaluating whether our current Class I differential price 21 surface is appropriate in the world of now as exemplified 22 through model analysis from 2021. 23 Q. Okay. But you're here presumably because you 24 think this information has something to contribute to the 25 conclusion? 26 A. I'm here because I think this information has 27 something to contribute to the decision-making processes 28 related to whether or not those Class I differentials 7014 1 should be modified. 2 Q. Okay. And -- okay. And just so we're clear, to 3 the extent that they are modified based upon this 4 information, then that becomes a mandatory payment 5 obligation for Class I handlers, correct? 6 A. So I can certainly say that if the Class I 7 differentials are changed, it becomes a mandatory 8 obligation on handlers. 9 Q. And when you are assessing value, let's say at a 10 place like Miami, that value reflects the cost of getting 11 the milk to Miami, correct? 12 A. So as I have stated on a number of kind of 13 previous responses, yes, that's one of the factors that 14 affects what the value is in Miami. But it's not just 15 what happens in Miami. It's not just about the 16 transportation flow from any particular location to Miami. 17 It's about the broader systemwide interactions that create 18 that value at Miami. 19 Q. One of the factors affecting the value in Miami is 20 the cost of getting milk to Miami; is that fair? 21 A. Yes. I previously noted that both the farm milk 22 assembly, the interplant transportation costs, and the 23 distribution costs are part of the transportation costs 24 that are considered in the analysis. 25 Q. And -- and captured by your -- to the extent -- 26 hopefully captured by the numbers that your model 27 establishes as values, correct? 28 A. They are definitely a part of the computations 7015 1 that lead to that value being established. 2 Q. Okay. So -- and you have said more than once 3 and -- that the model is, if you will, ignorant as to the 4 terms of the Federal Order system, correct? 5 A. The model does not take into account the -- what I 6 would call the behavioral incentives that Federal Milk 7 Marketing Orders provide. It is simply a supply chain 8 model designed to evaluate milk values. 9 Q. Does it take into account the provisions of the 10 orders with respect to, for example, transportation 11 credits? 12 A. No, it does not. 13 Q. Okay. So that -- okay. So that -- that sort of 14 goes beyond behavioral aspects of the model to payment 15 obligations of the order -- let's start that question 16 again. 17 That goes -- that goes beyond the behavioral 18 aspects of the orders to the payment obligation to the 19 orders; is that right? 20 A. So this model does not determine the payment 21 obligations under the orders because it does not contain 22 most of the Federal Order structure. It's simply a supply 23 chain model. It's trying to come up with the spatial 24 value differences for milk used in different classes. 25 Q. To the extent that the model would be relied upon 26 to set Class I differentials, the people using the model 27 in that fashion should recognize that the model does not 28 reflect, for example, whether there are transportation 7016 1 credits provided for in the orders; is that fair? 2 A. Yes. As I have previously stated, this is a -- 3 what I think of as a competitive benchmark that is 4 ruthless about trying to have the system costs be the 5 lowest possible that they can be. And so it does not take 6 into account, as I mentioned before, anything related to 7 passion about keeping a particular plant open, it does not 8 think about the pooling provisions, or any of the pool 9 dollars that are generated, and how that might influence 10 behavior. 11 Q. And so to the extent that, for example, there is a 12 recommended decision, awaiting final decision to require 13 Class I handlers in the Southeast orders to pay for the 14 transportation of milk into those locations, wholly apart 15 from the Class I differential, that's -- that's something 16 that your system -- your model just doesn't account for at 17 all, correct? 18 A. So the model accounts for the transportation costs 19 that would be -- and throughout the dairy supply chain, 20 but it is not dealing with specific obligations on the 21 part of anyone to make that payment. 22 Q. It -- so it does not account for the fact that 23 Class I handlers may be required by law to pay for those 24 transportation costs under the recommended decision, 25 wholly apart from whatever their Class I differential 26 obligations are; is that right? 27 A. It does not include any of the payments that would 28 be required under Federal Milk Marketing Orders because it 7017 1 focuses only on logistics costs. 2 Q. And -- all right. Are you willing to go so far as 3 to say it would be wrong to use the values established by 4 your model in setting Class I differentials to the extent 5 that there's another provision in those three Southeast 6 orders which is on the cusp of being adopted that requires 7 Class I handlers separately to pay for the cost of getting 8 milk to those locations? 9 A. Actually, first, I have tried to state previously 10 that I'm here to provide information as an analyst. So 11 questions of right or wrong are not really, I think, 12 generally within my purview. 13 And second, if I were trying to understand the 14 implications of having a transportation payment 15 requirement in addition to the transportation costs that 16 are represented in this model, I would actually want to 17 have the ability to model that and analyze it to come up 18 with a better answer to say, here's what I think the 19 implications would be. And that's kind of different than 20 saying right or wrong. 21 Q. I take it you were not asked by National Milk to 22 take a look at the recommended decision regarding 23 transportation credits in the three Southeast orders and 24 determine to what extent that might affect how the work 25 you had done that we have been looking at today might need 26 to be adjusted? 27 A. So I was not asked to look at that decision. But 28 I also think that the nature of what is included in the 7018 1 model and what is not included in the model really implies 2 that it would not be possible to consider that even if I 3 had reviewed that decision. 4 It's actually not saying who pays what is 5 essentially what we're talking about here. So it's not 6 saying who pays those things. It's considered as a broad 7 systemwide cost. 8 Q. Right. But I mean, we both know this whole 9 exercise is being undertaken for the purpose of 10 determining what Class I differentials should be, correct? 11 That's why we're here, correct? 12 A. I would not disagree. 13 Q. Okay. Let me switch to a different topic, which 14 is the question of adjustments being made after your model 15 was completed, okay? Do you know what I'm talking about? 16 A. Uh-huh. 17 Q. You should say yes or no. 18 A. Yes. 19 Q. Okay. So your model attempts to determine value 20 based upon the current locations of milk supply. 21 Obviously, you do a lot more than that, but I 22 mean, that's part of what you are doing, correct? 23 A. Yes. That's part of what we're doing. 24 Q. Let's assume a situation which 25 years ago when 25 the model was last -- the precursor model was last used to 26 determine what the Class I differentials would be. And 27 let's assume at that time you had two large cities, 28 30 miles apart, and the milk was being supplied, at that 7019 1 time -- I'm going to make the example somewhat 2 oversimplified -- from the south of those two cities, 3 okay? And those cities lay east and west of each other, 4 okay? 5 So under that scenario, it would be -- it would 6 have been the case the model would have potentially showed 7 values the same in those two cities because the milk was 8 coming from south, and cities were equal distance from 9 those sources of milk; is that fair? That would be how 10 the model would work? 11 A. So I think you have captured some of the logic. 12 But I do have to remind us all that we can talk about one 13 milk supply and two cities, but the values and the surface 14 are determined by the interaction of all of the millions 15 of variables and possibilities. 16 So in general, we see a certain kind of gradient. 17 We might say if we had a different demand at a different 18 location, that could modify somewhat what that gradient 19 might be. But it's difficult to state if you have two 20 cities and one milk source, what the implication of that 21 would be in the basis of a modeling event, even like what 22 we did in the simpler version in 1997 -- well, they did in 23 1997. 24 Q. What's the -- there's a term -- I think I could 25 say it, but I'm going to ask you to say it instead -- in 26 Latin, where economists use it all the time, all other 27 things remaining the same? Is it ceteris paribus? 28 A. Ceteris paribus. 7020 1 Q. Okay. Ceteris paribus? 2 A. Uh-huh. 3 THE COURT: Do you know how to spell that in 4 Latin? 5 MR. ROSENBAUM: C-E-T-E-R-U-S (sic)? Ceteris 6 paribus? 7 THE WITNESS: C-E-T-E-R -- and I think it's "I," 8 but you're saying "U" -- I-S, paribus, P-A-R-I-B-U-S. And 9 that's some card I should carry in my wallet as an 10 economist that I always have it available to me to spell 11 that, but I don't. 12 THE COURT: And it means everything is always the 13 same? 14 MR. ROSENBAUM: Everything else remaining the 15 same. 16 THE WITNESS: Other things being equal. 17 THE COURT: Oh, of course. 18 MR. ROSENBAUM: Yes. 19 THE COURT: Thank you. 20 BY MR. ROSENBAUM: 21 Q. So if -- if now the milk is coming from the west 22 rather than from the south to those cities, then other 23 things remaining the same, your model today could well 24 produce a value in the -- in the western of the two cities 25 that would be lower than the value in the eastern because 26 now the milk has to go an extra X miles to get to the -- 27 to the eastern city. 28 Is that -- is that just a fair way to -- a 7021 1 simplified way to think about how the model would work? 2 A. It's a simplified enough example, but I get the 3 idea. So generally if you have to move milk a longer 4 distance along the same trajectory from the same source, 5 you would expect to see a higher marginal value of milk at 6 that city location than a city closer to that same supply 7 source, again, ceteris paribus. 8 MR. ROSENBAUM: That's all I have. Thank you. 9 THE WITNESS: Okay. Thank you. 10 THE COURT: Are there other questions before I 11 call on the Agricultural Marketing Service? 12 There are none. I invite the Agricultural 13 Marketing Service to question Dr. Nicholson. 14 CROSS-EXAMINATION 15 BY MS. TAYLOR: 16 Q. Good afternoon. 17 A. Good afternoon. 18 Q. I was thinking we might not be done by 5 o'clock 19 today, but you might luck out. 20 A. I'm counting on USDA to come through for me. 21 Q. Okay. Well, let's see how we can go through these 22 questions. 23 I'm going to try to stick to questions out of 24 Exhibit Number 302, which is your longer statement. 25 A. Is it okay if I pull it up on my computer? 26 Q. Yes. Because I'll probably refer to some page 27 numbers, etcetera. 28 A. I have that up now. 7022 1 Q. Okay. Some of the questions is just to help us 2 kind of synthesize what we have heard over the past few 3 hours just to make sure we're all clear, and there will be 4 some other kind of more technical questions. 5 So in the start, in your summary you talk about 6 how the model is -- produces these location-specific 7 values to be used as a competitive benchmark. And I think 8 what I heard you define that earlier was it's kind of like 9 in the perfect world that's the lowest cost solution? 10 Okay. 11 A. Yes. 12 Q. And your number 3 listed in your summary of key 13 results talked about some of the reasons that we will see 14 differences in these spatial values. You list a number of 15 things, supply lo- -- changes in supply, changes in demand 16 locations, etcetera. 17 And one of my questions was to ask you, is there a 18 way we could kind of delineate those factors. But I think 19 heard you say before that that's not really possible. If 20 we wanted to see how much is the change in supply or the 21 supply is now and figure out what impact did that have, we 22 are not able to do that with these results currently? 23 A. With the current results, that's correct. So in 24 principle, and we have done analyses along these lines, we 25 could say what if we have the equivalent 2016 supply, 26 which is the last time prior to this that we updated it, 27 and in 2021, and the same thing we would do in terms of 28 changing demand. 7023 1 I think the key is, it's possible to do that, but 2 it has to be sort of done one thing at a time to 3 understand the implications of, say, a supply change 4 versus a demand change versus a transportation cost change 5 versus a plant location type change. 6 So in principle, it's possible to use the tool to 7 do that. It's a bit challenging to think about whether we 8 should use, like, a change over six years from 2015 to 9 2021 to accomplish that. That would only tell you in 10 these six years, this is what happened. 11 Q. Okay. So on the next page, page 4, you are 12 talking about how the large model is 3108 counties. The 13 smaller model, I think, had about 100 or so different -- a 14 few hundred multi-county regions is how you describe it in 15 your paper. 16 A. Yeah. And actually if you have a copy of the 17 R.B. 97-09, again, that Figure 6 provides a rough idea of 18 the way in which the multi-county regions were aggregated. 19 Q. And so when you have a sentence that says, "the 20 smaller model also allows more direct comparison with 21 prior analyses," are you talking about the '98 run, or are 22 you talking about what you just mentioned, which was 2016? 23 What is the prior one you are talking about? 24 A. More like 2016. So when we began developing the 25 database for this, one of the things we do is run a series 26 of different model checks to make sure that we have at 27 least reasonably consistent data, and then one of the 28 things that we also typically do is go back and look at -- 7024 1 once we have generated initial price surface, we go back 2 and look at a previous year's price surface, in this case, 3 the last one prior to that was 2016, and say, is there 4 something crazy going on here that would suggest that 5 there is an issue with the model data. 6 Q. Okay. 7 A. So we typically do that kind of check as an 8 informal visual thing, just to highlight whether or not we 9 should go back and look at any particular aspects of the 10 data inputs for the model. 11 Q. And you did this in this case? 12 A. Yes. For both the spring and the fall months. 13 Q. Okay. And if I heard correctly from some previous 14 questions, one of the reasons you do both the large and 15 the small versions of the model is, one, it's -- as 16 according to your paper, it's quicker to do the smaller 17 version, but also, you can see your product flows on the 18 map because you have less lines, let's say. So visually 19 you get an idea of what's going on? 20 A. Yeah. So, for example, I didn't even ask -- 21 Dr. Stephenson is the one who does the mapping for these. 22 I didn't even ask him to do the map for the fluid milk 23 distribution routing because it would basically be this 24 massive, ugly, spiderweb of orange lines, and it wouldn't 25 help us to -- and the sizes of the flows are also 26 indicated by the thickness of the lines. And so you would 27 have a lot of lines that were a mess, and then they would 28 also be very similar in size because of the smaller 7025 1 amounts of milk going to individual counties. 2 Q. Okay. And so when you did the different model 3 types, is it just the large model run results that you 4 gave to National Milk? 5 A. So, yeah. So we did both of those, and the 6 assessment is that they generate very similar outcomes. 7 When we have the 3,108 model, we actually have more data 8 points in a sense to work with, and so that was -- that 9 was what we did and reported only the mapping for the 10 smaller versions of the model to give a -- and, again, the 11 primal part is really to focus on providing an example of 12 what the kind of information that the model provides, not 13 so much to be important for the setting of Class I 14 differentials. 15 Q. Okay. And so for the way you ran the model, I 16 think it's in here somewhere, but I want to make sure is 17 it clear, that are the results constrained by the capacity 18 of each processing location? I know you talked about how 19 you went through and you talked some in cross and you got 20 help from National Milk on plant locations, for example, 21 on which plants might be closing. 22 But how did you come up with processing volumes, 23 would be one question? And the second question is, does 24 the model -- is the model constrained by those volumes? 25 A. So, yes, let me start with the second of the 26 questions. 27 So the model is constrained in the sense that 28 plants can only process up to the allowable volume, and 7026 1 this is expressed in terms of a total milk volume, farm 2 milk volume that can go into the plant, not in terms of 3 the products that can come out of it. 4 So the data for that are, to be honest, 5 incomplete. So we have estimates that primarily were 6 developed by Dr. Stephenson over time of the capacities at 7 most of the larger plant locations. But particularly for 8 smaller plant locations, there's some ice cream maker in 9 Wisconsin, we don't really know what their capacity is. 10 So we allow the ones for which we have information to be 11 constrained, and that's, you know, the largest proportion 12 of the milk and products that would be produced are under 13 those constraint plant locations. Which actually then 14 kind of conditions what the rest of the model is going to 15 do, because that milk will probably go there first, and 16 then only be available to some of those smaller facilities 17 like -- it's not after in the sense of time, but would be 18 made available to those plants that are not constrained. 19 So since we have some plants that are constrained 20 when we have the data, and we have other plants that are 21 not constrained, but the not constrained plants tend to be 22 the smaller plants. 23 Q. And so they would kind of be filled second in a 24 way; is that what I'm hearing? 25 A. It's kind of along those lines. Basically if you 26 have a favored plant location, then you would want to use 27 all the available capacity at that plant location, and 28 then once you can't do that anymore, then you got to think 7027 1 about, where else do I go with that milk, if you will. 2 But it depends crucially on that constrained plant 3 location being a location that the model thinks is good in 4 this broader system to minimize costs. 5 Q. And so could you give an estimate of -- of the 6 plants that were accounted for in the data, how many of 7 them do you think had volumes so they were constrained, 8 and percentage-wise maybe how many were -- or maybe by 9 volume, total volume of milk? 10 A. Yeah. I'd actually have to try and go back and 11 look at that. But I think that we do have the majority of 12 milk going into the different plant types, particularly 13 fluid, cheese, butter, and powder plants were under 14 constraints. For an MPC product, or ice cream product, we 15 have less coverage for those, but they tend not to be the 16 big volume drivers. 17 Q. Okay. Your model took into account imports and 18 exports -- 19 A. Yes. 20 Q. -- locations. And exports have certainly changed 21 a lot since 2000. 22 So I'm curious if you can maybe opine a little bit 23 on how you think that might have influenced the results or 24 had an impact on the results, if at all? 25 A. Yeah. I guess this falls again in the general 26 category of it's a little bit hard to determine what 27 causes what without doing a little bit more parsing out of 28 the different influences. But I can say that one of the 7028 1 key differences between this model version and even some 2 of the previous ones in the 2010 era were that we used to 3 have -- because when exports and imports were less 4 important to the U.S., we used to have three import and 5 export nodes, like major ports. I think they were L.A., 6 Houston, and New York. And with the growing amount of 7 export volume going out through different ports, you think 8 about a port like Seattle taking a lot of powder products 9 to Asia, we decided that we would actually use Census 10 Bureau-generated port district data to assign the actual 11 export quantities for the different products to those 12 locations. So it's like an extra demand at that 13 particular demand node. Right? 14 So why it's difficult to answer your question 15 about what difference did exports make is that the total 16 volume of exports is up, but also so is the total volume 17 of milk supply. And so we can try and, like, take out the 18 milk equivalent of those exports and see what difference 19 it makes, but without doing something like that, it's kind 20 of difficult to parse out exactly what that difference 21 made. But we did want to account for that to recognize 22 that that is a part of this broader system that determines 23 those price relatives. 24 Q. Okay. Which would probably make sense if we're 25 exporting close to 20% of our milk to account for that 26 demand, right? 27 A. Yes, that's why we did it. 28 Q. Okay. I wanted to turn to page 8. This is your 7029 1 table of product categories. And I just want to make sure 2 we're clear on the record what two column headings mean. 3 A. Okay. 4 Q. So the first column is the product, and then the 5 next two columns indicate whether it's a final product or 6 it's an intermediate product. And so I just want you 7 to -- so the fourth and fifth column, if you could just 8 make clear for the record what those two columns mean. 9 A. Okay. So if it's okay, I'll say the final product 10 in this model is things that went to final demand. We can 11 think about that being consumers, but we can also think 12 about it being foodservice or institutional buyers. 13 An intermediate product in the terminology that 14 we're using here is a dairy product that is used by 15 another dairy manufacturing process but came from a dairy 16 plant, and I have got examples there of the different 17 products. 18 So the IP allowed to make this product is 19 basically saying, for fluid, that that can be a 20 combination of cream and skim based on the idea that when 21 most milk hits a processing plant, it's often separated 22 into those two different entities and then brought back 23 together in the correct proportions to make the different 24 fat milk that's sold at retail. Right? 25 So something like ice cream, we actually have a 26 separate product category for ice cream mix that can be 27 produced at particular plants. That is the input into ice 28 cream manufacturer. 7030 1 Nonfat dry milk, the IP allowed to make this 2 product, in our terminology, skim milk and cream are 3 considered these intermediate products. So to make 4 nonfat, we would dehydrate the skim milk, and the -- I 5 guess that's the fifth of the columns, says product 6 allowed as IP in, basically we're saying that you can use 7 nonfat dry milk in the manufacture of fluid milk. And 8 actually I should note that that's really only in 9 California to meet their higher composition standards. 10 But it can be used to make yogurt, it can be used to make 11 American cheese, other cheese, casein, and ice cream mix. 12 So does that give you enough examples of what the 13 definitions are? 14 Q. It does. Thank you. 15 A. Okay. 16 Q. I want to turn to page 11. 2.8, Processing and 17 Transportation Costs. You mentioned how Dr. Stephenson 18 was -- you know, you all kind of separated assignments 19 when it comes to data, and I think you said that 20 processing costs came from Dr. Stephenson. 21 And you mention -- 22 A. Yes. 23 Q. -- that these costs are based on previous cost of 24 processing studies updated to reflect 2021 cost 25 structures. 26 There's been a few studies of his on costs put 27 into this hearing record, so I'm trying to figure out 28 which one you are talking about. 7031 1 A. So the input for processing costs in this model 2 derived from sort of a long series of our looking at the 3 evolution of processing costs over time. And essentially, 4 what we did was look at the 2016 processing costs that we 5 had, and Mark provided an adjustment factor that he 6 thought was relevant to bring that to 2021. 7 And I imagine that he used information available 8 from at least the first of the costs of processing studies 9 that he had done, but I don't know specifically how he 10 arrived at what that adjustment factor should be. 11 Q. Okay. Thanks. 12 And then down when you are talking about 13 transportation costs, and you used a standalone 14 transportation cost simulation program. I was wondering 15 if you could expand on that. 16 And in my mind, what I think I hear you saying, 17 which is something that Dairy Programs can do itself 18 sometimes when we do modeling, we figure some other things 19 out externally, and then input those results into the 20 model, so I -- that's what I'm interpreting that as what 21 you did, but I would like you to kind of expand on how 22 that was done. 23 A. Yeah, you have the basic idea. We used a separate 24 model to provide the transportation cost inputs for this 25 model. 26 And I think I mentioned earlier that there's been 27 an evolution of a model. It started out as an 28 extension-based tool to help haulers understand the full 7032 1 cost of moving milk from a farm to a processing plant, and 2 that has been refined to be a little bit broader structure 3 to allow the assessment of the different costs that I 4 mentioned. 5 And I think I mentioned previously that the way in 6 which this tool was used was -- again, this is something 7 that my good friend Dr. Stephenson is responsible for -- 8 was to run a large number of different possible routes 9 with that standalone transportation costing tool, and then 10 to understand what the statistical relationship was 11 between the distance of those routes -- and these are the 12 actual road mileage type routes, somewhat similar to how 13 we have done this in the model here -- and then to 14 establish a statistical relationship that is typically a 15 non-linear relationship that would look at all of those 16 different cost points relative to the distance and 17 establish what sort of a mean value would be at a distance 18 of X number of miles. 19 So does that provide enough information to help 20 you understand how we use this approach? 21 Q. It does. 22 And I had another question kind of later on. I 23 think you mentioned you used updated fuel cost data to 24 reflect 2021 diesel prices. Are those factors in this 25 model or that just gives you the relationship and then 26 your USDSS model puts in diesel prices separately, for 27 example? 28 A. Yeah. So let me try and clarify that, because I 7033 1 think it's an important point. 2 The initial transportation cost model is used to 3 generate what we call a cost matrix that has a cost to go 4 from any origin of the 3,108 to any destination of the 5 3,108, which relates to that function of distance. Okay? 6 We then -- and that actually includes the operating costs 7 that would include wages and tires and fuel in that 8 initial estimate. 9 But then in order to better reflect regional 10 differences in fuel costs and wages, we adjust that by a 11 factor that shows how the average -- or how a wage or a 12 fuel cost in a particular location is related to the 13 national average. 14 So if it's 95 -- if diesel is 95% of the national 15 average cost, that 95% is used to reduce the diesel cost 16 at the starting location where that pattern exists. And 17 it could be, you know, 5% more in which case you would 18 multiply it by 1.05. 19 So we start with a base of the transportation cost 20 from the model -- sorry, I'm using my professor hand -- 21 and -- 22 Q. I appreciate this lecture, so this is good. 23 A. Okay. And so we then adjust that in the -- in the 24 actual model simulations with the USDSS model to account 25 for the regional differences in wages and in fuel costs. 26 Q. Okay. And so I think I remember you stating you 27 used DOE data for fuel costs? 28 A. Right. And so that would actually, excuse me, 7034 1 come in where the average -- that would be the thing that 2 describes what the average national price would have been 3 in October for diesel, and then we have a regional 4 adjustment based on the DOE data. 5 Q. Right. Okay. 6 So the data you did -- you use for kind of like 7 your beginning index, I'll call it that, is for May of 8 2021? 9 A. May and October for preceding months, yeah. So 10 the base, if you will, that you are going to adjust up or 11 down is that May and October value. 12 Q. Okay. And then the wage data, what -- you used 13 wage data from BLS. I'm assuming that was specific to 14 trucking? 15 A. Yeah. I don't remember the specific BLS category, 16 but it was designed to be specific to trucking labor. 17 Q. It's a little bit different as Mr. English talked 18 about, and before -- when we did reform, it was informal, 19 and we kind of knew these things or could work with the 20 people. This is our only opportunity to ask these 21 questions. 22 A. I appreciate that. 23 Q. Okay. I want to turn to the Figure 5, which is on 24 page 13. 25 A. Yes. 26 Q. And I think in that figure what you said, because 27 this is using the smaller run, less locations, so kind of 28 plant locations might be grouped together in a larger 7035 1 triangle to represent a few plants that are there. 2 A. Absolutely. Yes. 3 Q. I did have a question. If we noticed -- well, let 4 me -- I just want to get your feel on how kind of accurate 5 you think those plant locations are, if, for example, 6 based on looking of these things -- you know, our own 7 knowledge, and not my knowledge, but the people sitting 8 behind me's knowledge, working out in our Federal Order 9 offices see a dot, let's say in the part of South Dakota 10 where they're not quite sure there's a plant there, but no 11 dot maybe in North Dakota where we think there is a plant 12 there, I mean, I think you talked earlier about the plant 13 list was probably the hardest part of this whole -- 14 putting together the plant list is probably the hardest 15 part to put together for the model. 16 A. It is. It's the place where the least amount of 17 publicly available information exists. I think is -- that 18 makes it the bigger challenge. 19 Q. Okay. So one question is, is kind of these 20 differences in where maybe a plant is or isn't, because 21 this is a smaller model, is that like a mapping problem 22 based on the fact that you just used the smaller runs to 23 do this, not necessarily all the plant locations that you 24 did in the larger run? 25 A. So this has been a question that we have been 26 getting -- well, at least I have been getting since the 27 time I started working on this model in 2000, and they 28 actually got it before in 1997. It has to do with exactly 7036 1 what you are talking about, the aggregation of plants that 2 are at other specific locations to the locations that are 3 available in the model. 4 And so I can't remember the specific instance, but 5 it was -- there was a conversation they were having, and 6 somebody said, "You don't have my plant on there. My 7 plant's over here, it's not over here." I said, "Yes, we 8 know, because that actual triangle location is accounting 9 for all the processing capacity, whether it's right there 10 or not. It's like, in the multi-county region, this is 11 the location that we chose to aggregate where that plant 12 location -- or plant capacity was available." 13 Q. And that multi-county region is kind of going back 14 to the study from 1998 and the circles, right? So that 15 region could be a large circle where that dot represents 16 or a small circle? 17 A. Yes. And as I noted, and in locations where there 18 was a good deal of milk supply in a dense area, like the 19 Central Valley of California, it's pretty much county by 20 county. In other regions where at least at that time 21 there was less milk supply, you had less dense coverage, 22 and you were aggregating more plant capacity at a given 23 location. So I can understand how the specific flows here 24 would make it look like it's not representing the plants 25 that you are aware of. 26 Q. So we had one question, if we can compare Figure 5 27 to Figure 4. So Figure 4 is on page 7. 28 A. Yep. Thank you. 7037 1 Q. And Figure 5 is on page 13. 2 And I want to point to the area in North Carolina 3 on the coast where there's on Figure 5, there's a triangle 4 there. So that would represent some fluid plant in that 5 area? 6 A. Let me look at the Figure 5. 7 Okay. I don't know the reason why there isn't a 8 blue dot where there is a dot in Figure 5. So, again, 9 these maps were, again, generated by Dr. Stephenson. 10 Q. Maybe I'll get the privilege to ask him that 11 question. 12 A. Perhaps you will. 13 Q. Okay. I want to make sure it's clear for the 14 record, because we just had the question come up. I think 15 what I heard from you in regards to -- earlier, I think 16 this was questioning for you from Mr. English -- or 17 Mr. Rosenbaum, while if there's a kind of a plant there in 18 your map that's not really there, or a plant -- a plant 19 missing that is there, in the big picture that doesn't 20 really change your results? 21 So, for example, maybe the missing plant in North 22 Carolina or something, that doesn't -- in the big 23 picture -- doesn't in the big picture kind of change the 24 results out of the model? 25 A. Actually, I think the Figure 5 is actually showing 26 that the supply plant is there. It's just omitted in the 27 general graphic that's talking about the dairy process -- 28 and actually, I think maybe one of the reasons that that 7038 1 may be omitted is we didn't have a capacity for that 2 plant, and therefore it didn't show up in the capacity 3 graphs that are shown in Figure 4. I don't know. I'd 4 have to actually better understand that. 5 But -- so the plant is definitely there for the 6 purposes of processing something in the analysis that we 7 did. What seems funny is that it doesn't show up as 8 listed in Figure 4. 9 Q. Okay. Okay. I'm on page 14, into 15. So here 10 you are talking about milk assembly -- well, that's the 11 figure, Figure 6. I don't want to talk about 6. 12 But I did want to move to the second -- page 15 in 13 that top paragraph. The second sentence from the bottom 14 of that paragraph says, "The model results are not 15 sensitive to changes of plus or minus 5%, and demand 16 values are estimated transportation costs." 17 So I just wanted to make sure it's clear, what you 18 are saying is if -- if there's only a small change, the 19 model's not going to pick that up or won't change its 20 results if there wasn't the change greater than 5% in one 21 of those variables? 22 A. Yeah, that's the correct interpretation. And 23 perhaps a little bit of background on that is helpful to 24 have. 25 I mentioned in the starting statement that I made 26 that this has actually been peer-reviewed research that's 27 been published. And one of the things that the reviewers 28 required us to do was that kind of analysis to assess the 7039 1 sensitivity of our results to those kinds of changes. 2 And so I mentioned earlier that we can tweak those 3 things to evaluate their impacts. And in general, what we 4 find is it takes a really large change, like the shifts in 5 population that we have seen and the locations of milk 6 supply and the transportation costs being markedly 7 different, for us to see a markedly different result from 8 the model. 9 Q. On page 17 you have Figure 8, milk production by 10 region. And you list the four regions, but I don't think 11 they are defined anywhere of what encompasses those 12 regions. 13 So could you elaborate on that a little bit? 14 A. Yeah. So, again, we talked about the division of 15 labor between Dr. Stephenson and myself, and I mentioned 16 that he was the one who was responsible for pulling 17 together the milk supply data. 18 I hope you will have the opportunity to rigorously 19 question him about Figure 8. 20 Q. Well, hopefully the person calling him on the 21 stand might know that we were going to ask him these 22 questions, so thank you. 23 This is -- I don't -- I'm trying not to go back 24 and forth. I want to go to page 21 where you show the pie 25 charts of the changes in the transportation costs as a 26 percentage of the total from 2011 to 2021. 27 So, for example, fuel costs represented 35% of 28 transportation costs in 2011, while they only represented 7040 1 25% in 2021. 2 We were wondering if you have kind of what are 3 the -- what are those values for those two years? 4 A. I agree that the breakout is perhaps relevant to 5 show the shifts among costs, but actually figuring out 6 what the total costs would be is also important. 7 And I will again ask you to refer to my esteemed 8 colleague, Dr. Mark Stephenson, to provide additional 9 details because he did that component of it. 10 Q. Okay. I will wait. And I wrote Mark on my 11 sticky, so we're going to come back to that. 12 And I assume then that he also did the Figure 13? 13 A. Yes, he did. 14 Q. This is why we're going to get done by 5:00. 15 A. Excellent. 16 Q. I did have a question. So your results that you 17 gave to National Milk, they are in Exhibit 300 that we 18 have discussed, include the base differential of a $1.60. 19 A. That's correct. 20 Q. The proposed results that are kind of -- from 21 National Milk's proposal results in Column S on 22 Exhibit 300 include a 2.20 base differential. 23 A. That's my understanding, but I have not had any 24 direct input at how that was calculated. 25 Q. But assuming that's their -- that's what's in 26 that -- assuming that the 2.20 is what they put in their 27 base differential, then is it right to say that $0.60 -- 28 part of the $0.60 difference between what the model came 7041 1 up with versus what they proposed could be attributed to 2 the different base differential that you -- that each 3 party assumed? 4 A. Again, I -- I don't know enough about how that 5 process was done to say you are just adding $0.60. I 6 actually don't think that was what was done. But I don't 7 know exactly how that set of values was assigned and the 8 role that the 2.20 played in that. 9 Q. I'll look forward to asking National Milk about 10 that. 11 A. I think you will have an opportunity. 12 Q. Okay. On the bottom of 23 you talk about Kriging, 13 K-R-I-G-I-N-G, is employed. And I think what I think you 14 are saying is, right, we get all these values at certain 15 locations, and you put it in this -- use this method to 16 kind of make a one continuous map. Even though there 17 might not be milk or a plant in a lot of locations, you 18 still assign a value in that? 19 A. Yeah, that's -- that's exactly the idea. It is an 20 algebraic algorithm that allows you to extrapolate values 21 based on -- if you think about -- say you have got a 22 series of marginal values at Class III plants, but you 23 don't have one in every county that you can use to say 24 that would be the value in that county. 25 This is an algorithm that allows you to take those 26 existing values, use different weighting schemes, and 27 assign the value to every county. And it's fairly 28 commonly used in this kind of spatial mapping type work. 7042 1 I guess one other thing I'll add is, that in 2 contrast to some of the previous work that we have done, 3 there are different ways to implement this algorithm. And 4 the way we initially were implementing it when we started 5 this process, we'd say if a -- if -- it's the linear 6 distance between two points, so if you are going to cross, 7 like, Lake Erie, you would say you can just take the value 8 and go right across there, and we realized that that was 9 not appropriate. 10 And so it's been modified to reflect the fact that 11 it actually has to respect the borders that are part of 12 that mapping system so that you are not coming up with 13 funny values for things that look to be closer than they 14 are, because the straight line distance is not the same as 15 what you have to have happen by moving the product. 16 Q. And when you say that's what you did initially, is 17 that what was done back in the '90s or what -- what that's 18 the initial period? 19 A. Well, initially in the '90s, but also in the early 20 iterations of this model in the 2000 era we were using 21 that without recognizing the importance of accounting for 22 that geography. 23 Q. Okay. We're looking at Figure 15 on page 24 with 24 all these pretty lines, pretty wavy lines. And this might 25 be a Mark thing -- 26 A. It is a Mark thing. 27 Q. Okay. 28 A. Can I say that to everything and then we'll be 7043 1 done really quickly or -- 2 MR. ENGLISH: No. 3 THE WITNESS: I guess I'm under oath. Sorry. 4 BY MS. TAYLOR: 5 Q. Well, I'll ask it anyways in case you know, but 6 I'll write it down here to ask him. 7 Is the difference between lines, like, a 10% 8 change in -- is that -- because this -- yeah, $0.10, 9 sorry, not 10% -- $0.10 change? Because some of these 10 numbers are kind of hard to read. 11 A. Yeah, this is actually the way that we used to map 12 it was only with the lines and then some colored 13 gradation. What I think is a little bit challenging about 14 this is you actually have the individual county levels 15 being mapped, and on top of it you have this line surface. 16 And technically, we wouldn't need the line surface 17 to help us interpret that so much except for the fact that 18 there's kind of this very broad range of values that goes 19 again from 1.60, you know, up into the 7s. And I think in 20 putting the lines in there, Mark was trying to help guide 21 the eyes to kind of where the break points were as opposed 22 to just using the colors. 23 Q. Okay. Okay. Another question for you, and you 24 talked a little bit about this with somebody, about kind 25 of the art that goes into kind of taking the model results 26 and then trying to bring them into the real world, not 27 just what the model spits out. 28 And you talked about what kind of things might 7044 1 people look at, factors go into different changes. And 2 one, you talked about competitive relationships that 3 currently exist, which the model does not account for. 4 You also talk about places where geography gets in the 5 way, so maybe that's mountain ranges or a lake. 6 A. So let me expand on that just a little bit, if I 7 may. 8 So the geography that we have here is the road 9 network. So if the road is going over a mountain and the 10 milk and products are moving over a mountain, they are 11 going over the mountain or they are not. It's not whether 12 or not there is a mountain range. We don't account for 13 any differential costs on a movement that would be going 14 over a mountain range versus traveling flat across the 15 plains once you got east of there. But we can have 16 geography if it's based on the existing road network. 17 So when I was thinking about things that were more 18 related to the transportation network, I was kind of 19 thinking about we don't account for traffic congestion in 20 metropolitan areas, for example. 21 Q. And maybe, back to my mountain example, maybe you 22 don't count for it might cost more to go over that 23 mountain, even though -- 24 A. That's correct. 25 Q. Uh-huh. And that's something that -- that's kind 26 of the art of people with knowledge of that marketing area 27 might be able to attest to? 28 A. Yeah. And I should also say there's -- and I grew 7045 1 up in San Diego, so I'm a little bit familiar with the 2 California geography, but we can also think about how many 3 arteries there really are to move product from a location 4 like the west, and there aren't that many. So that 5 actually might account for greater congestion on those 6 routes, and we didn't account for that in the system that 7 we use, which is sort of like this average costing of 8 routing plus an adjustment for fuel and wages. 9 Q. Okay. I think in your statement somewhere you 10 define disorderly marketing can result when differentials 11 are greater than transportation costs. 12 Does that ring a bell? 13 A. I'm not sure I used the word disorderly marketing. 14 Let me have a look and see. 15 Q. My notes say it's on page 23, so we'll all flip 16 there. 17 It's on the bottom of 22 into 23. So the sentence 18 says, "If these values" -- and you are talking about 19 differentials -- "were larger than the cost of 20 transportation, then 'disorderly' marketing conditions 21 could result with excess milk trying to find its way to 22 the higher valued plants." 23 A. I'm still looking for where you are. 24 Q. At the bottom of page 22. 25 A. Sorry, I was on 23. 26 Q. Yeah, sorry. And then the sentence starts there 27 and goes on to 23. 28 A. Thank you. 7046 1 Q. Yep. 2 A. Okay. So -- 3 Q. The question is, can you comment if you think you 4 see where -- if there are any instances of that in your 5 model results or you feel confident that your results are 6 not overstating transportation costs? 7 A. So I guess I haven't done a systematic analysis to 8 look at, like, whether a Class I differential for a 9 particular movement was a bigger value than the 10 transportation cost. I think generally what we see in the 11 model results is that we're not really actually talking 12 about the disorderly marketing per se, we're talking about 13 can we evaluate what those transportation costs would be 14 and how they affect what the marginal value of milk would 15 be, for example, at that Class I plant. 16 But we don't have any kind of more systematic 17 analysis of is the model saying something is too big 18 relative to what the current Class I differential is. 19 Q. Okay. And then for your -- particularly for the 20 fluid plants that you had in your data, does the model 21 differentiate between ESL plants and traditional HTST 22 plants? And do you -- if it does, do you think it 23 impacted the results in any way considering they have kind 24 of different distribution networks? 25 A. It does not currently make a distinction between 26 those two milk products. Again, I mentioned earlier, nor 27 does it deal with the differences in distribution for 28 organic milk. But we have actually started talking about 7047 1 that as more ESL milk has come online, do we need a 2 separate category for that, and the basic challenge is 3 being able to come up with the demand data to support 4 that. 5 Q. Okay. Not going to ask you questions I don't 6 think you're here to answer. Save that. Let's see. 7 So you talked a little bit about it, and I know we 8 looked at the results from 1998, but does your -- did your 9 model put out different surfaces for different product 10 categories, the four classes of products, for example? 11 A. Yes. 12 Q. And does each one of those kind of have a zero 13 point? I mean, they are all spatial values, so I assume 14 just like the Class I surface you produced, those also 15 have kind of like a zero starting point somewhere. 16 A. I guess I would have to go back and look at the 17 data. My recollection is we have fewer of those zero 18 points in part because of the way that manufacturing 19 capacity is spread around the country. 20 Q. And do you -- would you know maybe what the range 21 was in those surfaces? 22 A. So if we're thinking about whether or not there is 23 a -- dare I say it -- a Class III price surface, the 24 spatial analysis, which suggests that there is, my 25 recollection is that surface largely goes from west to 26 east -- and this is an approximate value, and I'm not 27 remembering it so much from this as from previous work -- 28 that might be like $0.50. 7048 1 Q. Okay. And does the model solve for all four of 2 those surfaces simultaneously? 3 A. It's generating a marginal milk value at all 4 locations where plants are processing. We then can 5 actually look at -- for the plants that would be doing 6 products that would be in any of the classes, we can 7 assign a value for that. 8 And so all of the information is being generated 9 at one time, just as I have talked about everything kind 10 of happens as a part of this broader system. And we don't 11 typically map any of the other surfaces other than 12 Class I. We have, for our own interest, occasionally done 13 that, and then most people do not want to talk to us about 14 a Class III price surface. 15 MS. TAYLOR: I think Mr. Wilson has a question for 16 you. 17 CROSS-EXAMINATION 18 BY MR. WILSON: 19 Q. First of all, I'm not an economist to any degree, 20 so there are terms that I don't know if you have used them 21 today, but other people have used them in the past, 22 related to shadow pricing. 23 A. Yes. 24 Q. That's equivalent to the -- to the marginal 25 spatial price cost that you are talking about, right? 26 A. So the base marginal values generated in the model 27 would be called by economists shadow prices or shadow 28 values, and the only place where we depart from that is 7049 1 when we add the $1.60 to get to something that's 2 comparable to the current Class I differential surface. 3 Q. Are there occasions where that shadow price is, on 4 Class I, a lot different than the other II, III, or IV 5 shadow prices? And if they are, what's your view on 6 what's causing that? 7 A. So I think you are -- maybe you are talking about 8 a broad set of differences between a Class III and a 9 Class I value? 10 Q. Correct. 11 A. Yeah. So I mentioned numerous times that the way 12 this model works is it's looking at all this information 13 simultaneously. So as an example, we might expect that a 14 Class III value would be different than a Class I value, 15 except for that 1.60, in a location where there's a lot of 16 cheese plants and a lot of milk going into cheese plants 17 that are satisfying a demand. And it'd be particularly 18 the case if there isn't a lot of demand for milk to go 19 into a fluid plant. And I'm kind of thinking about in my 20 mind an example of Idaho and Montana, right? 21 So there you could actually get differences 22 between the need for milk to make cheese, which is big and 23 strong, and saying, I need this right here based on the 24 capacity, and not so much demand for fluid milk, and also 25 a much sparser network of fluid plants in that region that 26 mean you actually have to move that farm milk a lot longer 27 distance to get it to a fluid plant. 28 Q. So if you had that county in Idaho that maybe had 7050 1 a fluid plant sitting next to a cheese plant, and one of 2 them was a lot more -- if Class -- if the cheese value 3 shadow price was significantly greater than the fluid 4 value, what's that telling you? What's that -- what's 5 that telling us? 6 A. Okay. So I can imagine a situation in the real 7 world where that might happen. In the model world, if you 8 have a cheese plant right next to a fluid milk plant, the 9 model only knows about the value differences due to what I 10 mentioned before, like, there's a little bit of different 11 component mix that's going into those different plants, 12 and it only knows about what would the transportation 13 costs be for me to go from one to the other. 14 So the model won't generate big differences in a 15 Class III value and a Class I value, again, ignoring the 16 1.60 part, if they are right next to each other, because 17 it's only accounting for those specific differences in the 18 component use and in the transportation that would take 19 you to go kind of this hypothetically across the road from 20 one plant to another. 21 So the model won't generate something that looks 22 like a big difference between those two values, other than 23 that 1.60 that would come into play. And so the model 24 can't really inform very much about what would happen if 25 we saw that. 26 Q. So in looking back at some of the maps that you 27 didn't create, we see these assembly points for the fluid. 28 Okay? And so I think of these shadow prices for the fluid 7051 1 side as milk is trying to get to the Southeast of the 2 United States, that shadow price gets larger and larger, 3 meaning that it's costing more money to get that milk to 4 processing. And so it's an assembly. It's -- it's -- 5 that shadow price is driven from the assembly side of farm 6 to processor -- transportation, not assembly, right? 7 A. Yeah. So it -- again, I apologize for kind of 8 keep saying this, but we have to think about this as a 9 broader system that includes all the different elements. 10 I agree with the basic idea that the model is 11 saying milk wants to move to the Southeast, it needs to 12 move longer distances. And that's actually both the 13 longer distances that we're looking at for a farm milk 14 assembly to a plant, as well as the distribution routings 15 that would take place, right? 16 So the basic idea I think is there. What exactly 17 causes that has a broader set of factors. But the map, 18 actually even for the small model, tends to suggest we 19 have a lot of milk that wants to move through kind of a 20 stair-step to the Southeast, and that's part of what's 21 generating those larger differences between the current 22 Class I differential and the model-generated marginal 23 values at that location. It's just more costly to get the 24 milk into those locations. 25 MR. WILSON: Thank you, Dr. Nicholson. 26 THE WITNESS: Thank you. 27 MS. TAYLOR: That's it from AMS. How about that? 28 THE COURT: Mr. English? 7052 1 CROSS-EXAMINATION 2 BY MR. ENGLISH: 3 Q. I have one follow-up question. 4 On page 14, Figure 6, a question that Ms. Taylor 5 asked a couple of different times about North Carolina. 6 Isn't it, sir, that that is likely the Port of Wilmington, 7 North Carolina, and is an export node? 8 A. Let me look. So I'm looking at Figure 5 -- 9 Q. Yes. 10 A. And I'm looking at the fact that that is 11 representing a fluid plant, which is not an exportable 12 product in the model. And it's also apparently, at least 13 if my eyes are not deceiving me, distributing milk. 14 Q. Okay. 15 A. So I think the omission is on the previous figure 16 where it wasn't represented with the blue dot, it's not 17 the omission for that, and it's not just an export 18 location. Although Wilmington is an export location in 19 the model. 20 MR. ENGLISH: All right. Thank you, sir. 21 THE WITNESS: Sure. 22 THE COURT: Ms. Hancock? 23 MS. HANCOCK: Thank you, Your Honor. We have no 24 further questions. Appreciate your time. 25 We would move for admission of Exhibits 302 and 26 303. 27 THE COURT: Is there any objection to the 28 admission into evidence of Exhibit 302? 7053 1 There is none. Exhibit 302 is admitted into 2 evidence. 3 (Exhibit Number 302 was received into 4 evidence.) 5 THE COURT: Is there any objection to the 6 admission into evidence of Exhibit 303? 7 There is none. Exhibit 303 is admitted into 8 evidence. 9 (Exhibit Number 303 was received into 10 evidence.) 11 MS. HANCOCK: And may the witness be excused, Your 12 Honor? 13 THE COURT: Is there anything you would like to 14 add? 15 THE WITNESS: No. I thank everyone for the 16 opportunity to make this presentation, and I thank you for 17 helping me meet my other obligations. 18 Thank you very much. 19 THE COURT: Wonderful. Thank you, Dr. Nicholson. 20 We really appreciate it. 21 And the witness may be excused. 22 THE WITNESS: Thank you. 23 MS. HANCOCK: I would just maybe tell you what I 24 understand or -- I don't know if, Erin, if you want to go? 25 MS. TAYLOR: Sure. So on my list of people that 26 need to go tomorrow, from National Milk, a Dr. Koontz, and 27 from MIG, Dean Sommer. 28 THE COURT: I'm sorry, say those again? 7054 1 MS. TAYLOR: Dr. Koontz, K-O-O-N-T-Z. And then 2 that would be a witness for National Milk. A witness for 3 the Milk Innovation Group would be Dean Sommer, I think -- 4 did I write that down correct? 5 MR. ENGLISH: You did, but it's really IDFA. 6 MS. TAYLOR: That would be an IDFA witness. I 7 think Mr. English said his name earlier. 8 THE COURT: Dean Sommer. 9 And then I had written down last night Jeffrey 10 Sims? 11 MS. HANCOCK: Yes. Well, for the -- he didn't 12 have the need to go and be off by tomorrow, but assuming 13 that we get Dr. Koontz on and off, and Mr. Sommer on and 14 off, we would be prepared to put on Jeff Sims, Dr. Eric 15 Erba. I think that Sally Keefe still at some point would 16 like to go on. 17 MS. TAYLOR: That will take us through tomorrow 18 and part of Friday. 19 And then Friday I have one in-person dairy farmer 20 that I know from National Milk that's coming, and we have 21 six farmers signed up to virtual testimony in the 22 afternoon. 23 THE COURT: You know, that is wonderful because 24 you just announced and you said, if you would like to 25 testify, let us know by noon or something, and you filled 26 it up. 27 MS. TAYLOR: People are listening. 28 THE COURT: I love it. 7055 1 All right. Is there anything else before we go 2 off record for the day? 3 There is none. See you tomorrow morning at 4 8:00 a.m. 5 We go off record at 5:02 p.m. 6 (Whereupon, the proceedings concluded.) 7 ---o0o--- 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 7056 1 STATE OF CALIFORNIA ) ) ss 2 COUNTY OF FRESNO ) 3 4 I, MYRA A. PISH, Certified Shorthand Reporter, do 5 hereby certify that the foregoing pages comprise a full, 6 true and correct transcript of my shorthand notes, and a 7 full, true and correct statement of the proceedings held 8 at the time and place heretofore stated. 9 10 DATED: November 29, 2023 11 FRESNO, CALIFORNIA 12 13 14 15 16 MYRA A. PISH, RPR CSR Certificate No. 11613 17 18 19 20 21 22 23 24 25 26 27 28