Class I1l M ilk Pricingt An Evaluation of Assumptions and CalculationsDavid M . Barbano, Ph.D.Professor of Food Sdtnte

Corndl Universits Ithacay NY 14853dmb37@cornell.edu

Version patt: April 24, 2000The lnform ation contained in this docum ent will be prestnttd at the USDA Prlce Hearing

that starl on M ay 8, 2000.

M y areas of expertise are in cheese and whey processing teclmology, milk componentanalysis, cheese characteristics, milk composition and quality, cheesa yield formulas, factorsiM uencing cheese yield, and cheese manufacturing costs. I teach a course on the chemistry of dairyproducts and carry out research on these topics as part of my responsibilities as a faculty member atCornell University. Ireceived my Ph.D. in Food Science from Comelluniversity in 1978 and I havebeen on the faculty at Comelluniversity since 1980. 1a111 not representing any company orproducergroup at this hearing. I do not own or operate a farm, cheese company, or any other dairyyroductmanufacturing business. Mypurpose inpresentjng this information is 1) toprovlde the dairy mdustryand USDA with a critical review of the current system of Class 11l price calculation and assumptionsused in the calculation and 2) to offer a different approach and some new ideas for calculating theClass l1I price. The approach that l will present is derived h'om the Vanslyke Cheddar cheese yield

fonnula. The objective of this approach is to provide betler economic signals between processorsand milk producers. Hopefally, a more fair and equitable 'reflection of changes in milk values for both

producers and processors can be aohieved.

lntroduction:Historically, the basis for a national Chss f1l milk price wms the M irmesota-W isconsin Price

Series for manufacttuing grade milk. W hen there was a large volume of çitmregulated'' millt formatmfacttlring beingsold for cheese manufacttuingsthis price reflected the unregtzlated ïfee marketvalue'' of milk for cheese maldng. M ilk used for Class 11 or Class 1 pmducts would have a highervalue. Over the years the quantity of millc represented by the M irmesota-W isconsin Plice Seriesdecreased. In the 1990's the validity of using the M innesota-W isconsin Price Series as the basis forsetting the uniform Class III milk price throughout the USDA Federal Milk M arketing Orders was

questioned.Because of changes in industry stmcture within the US, tlze US Congress mandated that the

USDA Fedex'al M ilk Marketing Orders reorganize to better reflect the current milk marketing areaswithin the US in the 1996 Farm Bill. At the same time, the Congress provided that USDA may makerevisions to the milk pricing system to ensure that fair and equitable prices are paid to milk producersin 411 regions of the country and to harmonize the provisions of the system of milk pricing in differentregions of the coutztry. ln doing this, a fundamental change was made to the method for establishingthe Class ll1 plice for milk within the Federal Orders. M ilk in Class l1l would be priced based oncomponent values. The Varlslyke cheese yield fom mla was used to calculate the butterfat andprotein factors (i.e., 1.582 and 1.405) used to anive at a protein value in the Class l11 pricecalculation (1).

1

EXHIBIT

599z6 f?-/x-p7

Starting January 1, 2t*0, the monthly Class HI prke has been calculated as follows:

A. Trtze Protein PHce ($/1b) is determined in tsvo steps)

1. Calculation of the Value of the Protein in Cheese:

(NASS Monthly Cheddar Price) minus (Cheddar Make Allowance) x (1.405)

The 1.405 factor is derived 9om the Vanslyke Cheddar cheese yield formula and isdesignedto reflect the expected increase itA Cheddar cheese yield that would occurfora tmit increase in true protein contcnt of milk. To calculate this factor the followingparameters used in the Vanslyke formula calculation are needed: fat and trtze proteincontent of the milk, the percentaje fat recovery in the cheese, the proportion of tnzeprotein that is caseino and the mmsttlre content of the cheese. Selection of a differentset of asslmptions for these parameters will ptoduce a different factor than 1-405-

2. Calculation of the Extra Value of Protein Due to Fat:

(INASS Monthly Cheddar Price) minus (Cheddar Make Allowance) x (1 .582)) minus(the butterfat price ) x (1.28).

lt is my understanding that the primary reason tlzis calculation is done js to reflect theadded Nalue of milk <at in obevxse in tbe absenve of a diszTeet pùc,e fo: milk f'at usedin Class 111.

The 1.582 factor is also derived hom the Vanslyke Cheddar cheese yield formla andis designed to reflect the increase in cheese yield from a tmit increase in milkfat.Again, to calculate this factor the following parameters used in the Vanslyke fommlacalculation are needed: fat and trtze protein content of the milkp the percentage of fatrecovered in the cheese, the proportion of tt'ue protein that is casein, and the moisturecontent of the cheese. Selection of a different set of assumptions for these parameterswill produce a different factor than 1.582.

The 1.28 is not derived directly from the Van Slyke cheese yield formula. lt is mytmderstanding that this factor is supposed to reflect the amount of milk fat that onepound of true protein in milk catl hold in Cheddar cheese. For the gttrpose ofcalculations in Federal Order Reform and calculation of the Class I11 pnce, a milkcontaining3.s% fat, 2.9915%trueprotein, and 5.6935% othersolids (i.e., 3.10% trueprotein and 5.90% other solids in the sltim portion) has been used fol' calculations ofthe Class Illprice. However, relative to the 1.28 assumption in the calculation of theextra value of protein due to fat, the average ratio of fat to tnze protein that exists inthe milk supply will probably be lower than this value all year. ln a national IIIICcomposition stcdy of commingled milks in cheese factories in the US in 1984, it was

2

found that the ratio of fat to trtze protein varied throughout the year with valuesranging from 1.145 to 1 .18. Generally, the fat to casein ratio is lowest in June

, July,and August.

These two values (i.e., Value of Protein in Cheese and the Extra Value of Protein due to Fat)are added together to anive at the tl'ue protein price. The 1.405 and the 1.582 factors weredtrived 9om the Van Slyke Cheese Yield Fom mla.

Sample Calculation: (March 1999 Prices)NASS Cheese Price $1.3064/1bCheddar Make Allowance $0.1702/1bNASS W hey Powder Price $0.1917 /1bW hey Powder Malte Allowance $0.137/1b

Calculate Tnze Protein Price1. ($1.3064/1b) - ($0.1702) x (1.405) = $1.5964/1b2. (($1.3064/lb) - ($0.1702) x 1.582) - ($1.4487) x (1.28) = $0.4464/1b3. $1.5964 + $0.4464 = $2.0428/1b of true protein

B. Other So- li-ds Price Calculation1. (NASS Dry Whey, $/1b) mhms (whey make a11owance)/(0.968)

($0.1917) - ($0.137)/(0.968) = $0.0565/1bThe 0.965 factor is used to veflect tlàat on average th'y whey is 3.2% by weigbt moisttve.

C. Class I1I Skim Price is Calculated tat 3.1% true protein and 5.9% other solids) As Follows:(True Protein Price) x (3.1) = $2.0428 x 3.1 = $6.3330(Other Solids Price) x (5.9) = $0.0565 x 5.9 = $0.3334($6.3327 + $0.3334) = $6.6664 per hundred weight

D. Class H1 Price at 3.5% fat is calculated As Follows:Cliss Ill Sldm Price $6.6664 x 0.965 $ 6.4331 (Skim portion)Butted-at Price $1.4487 x 3.5 + $ 5.0705 (Fat portion)Class 111 M ilk Price at 3.5% fat $11.5036

Behavior of Class IlI W ltole M ilk and Sldm Prices W hen Fat Value Changes.

ln my opinion, when the Class 1l1 milk price calculation (described above) is used to calculate thewhole and skim milk values in Class III with changing butterfat prices and millc compositions, thechanges in tht milk prices inrelation to changes inmilk fatprice do not give sensible economic sir alsto milk producel's. The fundamental problem (in the carrent Class l11 price calculation) is that whenthe value of milk fat goes ug (driven by an increasing butter price), the calculated trlze protein valueit4 dollars perpound of proteln goes down, and it goes decreases at a fastermte than the value of milkfat increases. Thus, when the price of butter increases, the Class II1 milk price (i.e., milk price paid

3

by cheese makers) for a milk tllat has a fat to protein ratio of less than 1.28 will go down. 1 willuse several examples to illustrate this point.

ln Figure 1, the butter plice is increased from $1.00 per pound to $1.90 per pound. For a producerwith a milk that contains 3.83% fat, 2.99 tnle protein, and 5.68% other solids, the price paid for milkby the cheese maker willremain constant as the butter price increases from $1.00 to $1.90 perpound.This means that the price for the skim portion paid to this producer is going down at the same ratethe fat value in that milk is increasing. Thus, despite the fact that butter is short and the price is high,the price (at constant milk composition) that a fanner (with a ratio of milk fat to true protein of lessthan 1.28) receives for milk decreases with increasing price of butter fat. As seen from Figure 1, aproducer with a 1.36 ratio of fat to true protein, the milk price goes up by about $0.30 kcwt when thebutter price increases âom $ 1.00 to $1.90 per potmd. However, for the producer with a 1.00 ratioof fat to protein and a 2.99% fat test, tbe price of milk goes down by about $1.00 /cwt as the butterprice increases from $ 1.00 to $1.90 per potmd. This is not the correct economic signal to send todairy farmers in this simation. lf a plot of skim milk value instead of whole milk value is made, thedecrease in skim milkvalue as butterprice increases (and cheese price remains constant) is even moredramatic than that shown for whole milk in Figure 1 (shown below).

Producer Milk Valueslcwt) at Various Butter PricesAssum ing $1.303b Cheese Price and 19$3b Dry W hey

and Components for Producers ofVaring B.F, 2-99 Protein and 5-68 O.S-

4 2 50

52 00

': ,1 sot!l ,1.eo

10.60

10 00

1.01 5.*10 1.20 1 30 ,1 .40 1 S0 1 60 1.70 1.80 4.95

Buster Prpce

4

How does this impact producers in a market? Figure 2 is a frequency distribution of fat to trueprotein ratio for producer milk from the Southwest Federal Milk Market Order (provided by themarket administrator). The distribution repxesents 16,230 observations in one federal order in 1999for the average fat to tl'ue protein ratio. rl'he distribution of fat to casein ratios is relatively normalin shape, btlt the median ratio of milk fat to true protein is 1.17. Only about 5% of the producer milksamples had a fat to tnze protein ratio that was higher than 1.28. Thus, when butter price increases,the average chss 11I price for the group of producers with a fat to true protein ratio less than 1.28willdecrease at constant cheese and whey powder price. At first glance, one might say moving the1.28 factor to 1.17 will tix the problem for tbis population of producers. Howevex, what this will dois make the price paid to half the producers go up and half the producers go down when butter priceincreases. When milk fat value (NASS AA butter price) incxeasesp the price paid to every producerfor milk should go up to reflect the increased value of the fat portion of the milk.Figure 2

Distribution of Butterfat to True Protein Ratip in Ptoducer Milk In the Soutbwest - 1999

1200

' k

1000

800 .

2t.: GGIjR

l

: .200 . . ,

r

Aq V,N T> *,% k7 k,q qN .% 0* AA (4% os o% & & og %* '& s% '.A 'xe 4% 45 qfa 4 %m s* o'h ob 07 nq h'y w* éa ft *9O O B I B B O I L U Q N N N N N N N N N N N N N N N N N N N N N N N N N

Butterfat tl 'frue Protfm *.!1.10

5

W hile the current system for Class 1Il plice calculation rem esents a trtmendous amount of thinkingand development by the industry arld USDA staft in my opinion, the current system for the Class I1Iprice calculation is not providing the correct signals from processor to producer wben the marketprices of various products change, particularly milk fat. There are additional and more subtle issuesin the current Class l11 price calculation that trouble me, but in my view the one illaskated in Figures1 and 2 is the major one and it needs to be corrected. Thus, I have come to this hearing to presentsome ideas on how to eliminate some of the short comings of the current metllod of calculation ofthe Class I11 price for milk within the Federal M ilk M arket Orders.

How Should The Dairy Industry M odify the Class II1 Price Cakulation to Eliminate theseShort Comings?

The approach that l propose is also based on the Vanslyke Cheddar cheese yield formula. TheVanslyke formula works well for ftzll fat Cheddar cheese made from milk that is not fortified withnonfat milk solids. For other cheeses and cheeses made using fortified milkn, other yield formulaswould be more appropriate (2) for prediction of cheese yield. J.n my opinion, the selection ofCheddar cheese made without nonfat solids fortitication of milk for cheese making is the right choiceas the basis for the Class I1I minimum tmiform milk price oalculation. Below l will provide the fulldetail for the basis of a different method of Class 1II uniform minimum price calculation.

First, 1 would like to explain the Vanslyke Cheddar Cheese Yield Formula.

((% fat recovery x % fat) + (0.78 x cnzde protein - 0. 1)q 1.09

l - (cheese moisture/loo)

The valut selected for % fat recovery (in the cheese) for the calculation can be debated,however a 93% fat recovery in the cheese is achievable with modern cheese maldngequipment and was achievable in the mid 1890's when Vanslyke developed his cheese yieldformulabased on observations of Cheddar cheese making practice inmany factories in centralNew York over a two year period.

The 0.78 times cnzde protein is a substittzte for a measurement of the casein content of themilk. The origînal Vanslyke formula uses percent milk casdn. The industry has used anassumption that 78.00% of the crude protein content of millc is casein. For a 3.67% fat milkwhh a 3.1762 crude protein (3.1% tlue protein in the skim portion) and 2.9862 trae protein,the multiplier (mathematically equivalent to 0.78) for estimation of casein from tnle proteinis 0.8295.

The minus 0.1 used in the equation reflects an expected fixed loss of casein into whey that willocc'ur dming cheese making, regardless of starting milk composition.

6

The 1.09 factor accotmts for the nonfat, noncasein milk solids expected to be retained in themoisture phase of the cheese and the added salt in the cheese. The constant 1.09 value% sumes that the tinal cheese conuins about 1.7% salt. Thus, the numerator in the Vanslykeequation calculates the weight of milk solids plus added salt that is expected to be collectedas cheese given the milk composition values used in the calculation.

The denominator of the Vanslyke equation simply adjusts the calculated total yield ofChtddar cheese to the target moisture petcentage used in the formula. Thus, the formula canpredict expected Cheddar cheese yields from milks of different fat and protein contents atselected m oisture contents.

The following are a11 the parameters where assumptions and values are needed in the calculation ofClass 1Il price that I propose. These values and this format are part of a spreadsheet that I have usedto summarize a11 the values used in the calculation. The values in blue are values that can be varied(for sensitivity analysis) and the numbers in black are tixed values that are calculated as htennediatesin my calculatîons.

l have been told that different values than the ones I have mentioned above were used in theVanslyke Cheese Yield Formula when the protein and fat factors (i.e., 1.405 and 1.582) weredelived to use as the basis for calculation of protein value in the pricing system initiated on Januarylp 2000. It is my understanding that for the current pricing system, the value used for fat recoveryin the cheese is 90%, the value for casein as a percentage of crtlde protein is 0.75 and the value formoisture content of Cheddar cheese is 38% .

7

Table 1. CompositionAssumptions and Values Used in the Current Class III Price Calculation(M arch 1999 data used for this txample).

Mllk Fat Content, % 3.5000Crude Proteln, OA 3.1815Casein as percentage of crude protein,o/à 75.0000True Protein, % 2.9915Caseln as percentage of true protein,% 79.7635Mllk Caseln, % 2.3861Mllk Serum Protelns % 0.6054Mllk Other Sollds Content, % 5.6935Mllk Total Solids Content, % 12.1850Fat Recovery in Cheese 0.9000NonFat,Noncasein Solids Factor for Vanslyke Yield 1.0900

Vanslyke Cheddar Cheese target % moisttlre, lb/cwt 9.5571Yield at

Fat in the Cheddar cheese, lbs 3.1500True Protein in the cheese, Ibs 2.2861True Protein not in the cheese, Ibs 0.7054

NASS Cheddar Prlce, $/Ib 1 .3064Cheddar Cheese Make Allowance, $/Ib of cheese 0.1702Cheddar Cheese Composition

Fat, % 32.9599Protam, % 23.9208Skim Portlon, 9/o 67.0401Moistur ,e % 38.0000Fat on Dry Basls, % 53.1612Other Sollds, % 0.3268

NASS Whey Powder Prlce, $/1b 0.1917Moislure Test of W hey Powdery% 3.2000Whey Powder Make Allowance, $/Ib of whey powder 0.1370Yield of W hey Powder at 3.2% molsture 6.2728

pounds of true proteln in whey powder 0.7054pounds of other solids ln whey powder 5.3667

EXPLANATION OF VALUES IN TM LE 1.

M ilk Com position:

The basis for the milk composition values shown in Table 1 are as follows: The trtze proteinand other solids values are values from the Federal Orders that are thought to represent the armualaverage skim milk composition in the United States. The tnze protein (2.9915%) and other solids(5.6935%) values for the 3.50% fat milk corresptmd to a 3.1% true protein and a 5.9% other solidscontent in the skim portion. The crude protein value is calculated assuming that there is 0.19%protein equivalent as nonprotein nitogen in the average milk. A value of 75% of the crtzde proteinwas used by AM S for calculation of the casein cfmtent of milk in tlle ctuwnt pricing system

, so l

8

have used this as a default value for my first calculations. This value was used to calculate theequivalent value of casein as a percentage of true protein. For this example, the casein as apercentage of tnze protein is 79.7635% . I show both values because the industry has only recentlystarted working with trtle protein as the basis for payment and there is a need to show how a valueequivalent to the 75% casein as a percentage of crude protein was derived.

In Table 1, the milk serum protein percentage is simply the tnze protein mhms the casein percentage.The milk total solids content is calculated from the sum of fat, true protein, and other solids.

Cheese Yield Formula;

The Vanslyke formula is as described above. The value for fat recovery in the checse usedin this example is 90%. The nonfat, noncasein millc solids plus salt retention factor in the cheese is1.09. The cheese yield value given is the value calculated at this milk composition for a cheese with38% moisture. The pounds of fat in the cheese andpotmds of true proteinin the cheese come directlyfm m the numerator of the cheese yield equation. The pounds of tnze protein not in the cheese iscalculated as the differtnce between the pounds of true protein in the milk minus the potmds of trueprotein retained in the cheese. This will be the pounds of tnle protein that goes into whey powder.

NASS Prices:

The NASS Cheddar cheese price ls a value calculated by the USDA Dairy Programs basedon the wetkly survey of cheese prices. The price survey data has the following characteristics.

Block Cheddar: The moisture content of block cheese reported in the sun'ey is notreported to NASS. One can assume that it is less than the legal maximum moisture forCheddar of 39$$. NASS specifies that the moisture content of the blocks shall not be lessthan 36.51$. lt is assumed that the cheese meets the minium requirement for fullfat Cheddarof 50% fat on dry basis. The price reported by NASS blocks includes the cost of packagingthe 49 lb blocks as described in the instructions and the cheese is colored to between 6 and8 on the National Cheese lnstitute Color chart. The price should re:ect cheese wrapped insealed, airtight package in conugated or solid fiberboard containers with reinforcing innerliner or sleeve. A11 other packaging costs are excluded hom the reported prices. The sale iswhen a transaction is complete (i.e., cheese is shipped out and title transfer oceurs). Inka-company sales, resale of cheese, lansportation, and clenring charges art not included in theprice. Price is L o.b. the processing plant or storage center. Blocks must meet W isconsinState Brand, UDSA Grade A or better. Blocks of cheese made for aging are not included inthe survey.

Barrel Cheddar: Cheese reported as barrel cheese cannot exceed 37.7% moisture content.This is based on a Chicago M ercantile Exchange rtzles which state that chtese exceeding thismoisture content cannot be invoiced on a moisture basis. The moisture content of barrel

9

cheese is lazown and reported in the NASS sunrey results. The fat on a dry basis for theban'el cheese is not known, but it must exceed 50% to comply with standard of identity forCheddar eheese. The reported cheese prioe by the manufacml'er foç ban'd cheese is at theacttzal moismre test of the cheese reported and this price includes no packagisg costs. NASScalculates a moisture adjustment to bring allprices to a 39% moisture basis fbrbarrel cheese.The cheese is white and must meet W isconsin State Brand, USDA Extra Grade, or better.The sale is when a transaction is complete (i.e., cheese is shipped out and title transferoccms). lntra-company sales, resale of cheese, transportation, and clearing chrges are notincladed in the plice. Price is feo.b. the processing plant or storage center.

M onthly NASS Price Used In the Class IlI M ilk Price Cakulation: The wcighted averagemonthly Cheddar cheese price used fn the Class Illprice fonnula is computed by the USDAAgriculttlral M arketing Service per the provisions of the order. A weighted average iscomputed for blocks and barfels each, using the applicable weeldy pcices and weights. Theprices aro oomputed to 4 decimal points. No adjustments are made to the published NASSprices. Three cents are added to the banel average and then the bloc,k and barrel averages areweighted using the monthly weights. This price is rounded to four decimal places and is usedin the Class 11I price calculation, Tbe average moisture test of the cheese tbat correspondsto the combined block plus ban'el Cheddar cheese price is not ltnown, but given theinstructions in the sunrey it must be be- een 36.5 and 39?4. lf the amount of barrel cheesei.ls the sun'ey for a month is abeut 62% of tlze ttltal weight oî cbeese in the sunrey and weassume all the block cheese is at the minimum moisture, then the moismre content of thecheese represented by this price would be about 38.05% .

In my opinion it would bc of beneft to the dairy jndustry if moisture data werecollected for the block Cheddar represented in the NASS survey. This would allow thecheese price produced by the NASS survey to be associated wlth a specific m oistureconttnt that wold be known. W îth this information, the moisture content inthe cheese yieldfonnula used to calculate the Class 11I pricz would m oduce prives fol' fat atld protein iqCheddar cheese tbat are in harmony with the moisture basis for the NASS chetse price.

NASS W hey P/wder Price Used ln the Clasy III M ilk Price Calculatitm:

The product is USDA Extra Grade edible nonhygroscopic dty whey. The price is f osb. theprocessing plant/storage center. Prices are reperted for a1l 25 kilovam, 50 lb bag, tote, andtanker sales, The following al'e excluded; lansportation charges, sales of Cyrade A dry whey,sales of dry whey more than 180 days old, intra-company sales, resales of purchased drywbey. The current Class III price calculation fol' other solids value assumes that whey

powder contains 3.2% moismre.

Cheddar Chetse Composition:

A value for Cheddar cheese moisture content must be selected for use in the cheese yield

10

calculation. In the default values used in Table 1, the value is set at 38% moisture. This value wasused by USDA to calculate the grotein and fat factors in the current pricing system. However, asalready mentioned in the discusslon of NASS cheese prices, the moisture content selected for use inthe yield calculation should be consistent with the moisture content of the cheese included in theNASS survey. Once a target moisture value is established, then the Cheddar cheese compositioncan be calculated 9om the millt composition values and the cheese yield formula. The fat and tnleprotein content of the cheese (Table 1) is the pounds of fat and true protein retained in the cheesedivided by the cheese yield and multiplied by 100. The salt content assumed (as part of the 1.09value) in the Vanslyke Cheddar cheese yield fonnula is 1.7%. The sldm portion of the Cheddarcheese (Table 1) is 100% minus the pelvent fat plus percent salt in the cheese.

W hey powder yield:

The weight of tnze protein in the whey powder (Table 1) is the weight of tnze proteincontained in the milk minus the weight of tt'ue protein contained in the cheese. The weight of othersolids in the whey powder (Table 1) is the weight of other solids in the milk minus the weight of othersolids retained in the cheese. The weight of other solids in the cheese (Table 1) is calculated bytaldng weight of solids in the cheese minus the weight of fat plus trtze protein plus salt in the cheese.The calculation assumes that the cheese contains 1.7% by weight of salt. This number is the amountof other solids retained as dissolved solids in the water portion of the cheese. The sum of other solidsplus true protein in the whey powder divided by one minus the peveent moisture in the whey powder(e.g., (1 - (3.2/100)j provides an estimate of whey powder yield at 3.2% moisture.

Cheddar Cheese and Dry W hey M ake Allowances:

These values (Table 1) are defmed as fixed values that are used in the calculation of the Classl11 price by USDA. They are based on input 9om industry da> for Cheddar cheese mmmfacttuingcosts. Dry whey manufacturing costs are based on a smdy conducted at Com ell University. Itwould be useful to have a clear and complete description of what is included and what is not includedin the cheese and whey make allowances.

The make allowances are expressed as $/1b of cheese. However, a high percentage of themake costs are fixed and relate better to hundred weight of liquid in the vat not directly to a poundof chetse. Thus, when milk composition varies within normal ranges and products calculatedchanges in cheese yield the true make costs for cheese do not increase or decrease as much withchange as cheese yield as one would calculate. Thus, caution must be used when calculating rettzrnsto a cheese maker when milk composition tand fherefore theoretical cheese yield) varies withchanging milk composition.

1 1

M ETH O D PROPO SED BY D AVID BARBANOFO R CALCULATION OF THE CLASS llI M ILK PRICE

The input data shown in Table 1 are used as current default values for the pmpose of comparison ofthe Class H1 price at 3.5% fat by the calculation I proposed versus the Class I1I price calculatedtmder the current millc pricing system at the same milk composition. To the best of knowledge, thedefault values shown in Table 1 represznt the values currently used by USDA and the prices are fromM arch 1999. This does not mean that l agree with the current default values being used by USDA.That issue will be addressed later in my discussion.

The new method of calculation that I propose has three steps. These steps and a sample calculationare shown on the spread sheet that is provided with this description.

Step 1. Class III Fat Value Equals NASS Cheddar Cheese Prke (for this example thevalue would be $1.3064). The current Class I1I milk plicing system initiated as a result of FederalOrder Reform stnzggles with this issue. The current system does not establish a separate Class IlIprice for butterfat, but instead adds a fat value to the protein price. This is the fundamental cause ofthe problem with thz current pricing system that was demonstrated in Figures 1 and 2. Therefore,in Step 1 of the proposed calculation, the milk fat used in Class 1ll is priced at the same value in $/lbas the NASS price for cheese. Thls /,& the key new step used in my approach to calculate a ClassHlprlce based on t/ze#rfcc ofcheese. Once a price per pound of cheese is established al1 parts ofthat cheese have that value in the market place when it is sold. Therefore, l assign the cheese priceper potmd to the fat and calculate the portion of the total value of a potmd of cheese that is fat. Theresidual weight of the nonfat portion of the chees, takes on the remainder of the value per potmd ofcheese and a11 of this value (minus a make allowance) is allocated to the protein retained in thecheese.

Cheddar cheese has a defned minium fat content of 50% on dry basis. Jn reality Cheddarcheese of acceptable quality for processing can be made in the range gom 50 to about 55% fat on drybasis. Thus, the selling price of the oheese is the price that the cheese maker receives for the fat soldin the cheese. If milk fat has a higher value in other utilization classes, then the cheese maker willhave a sigtml to remove fat 9om milk, as cream, itl excess of that needed to achieve 50% fat on dxybasis. lf milk fat has a lower value in other utilization classes than Class 111, then the cheese makerwill have a sipaal to keep more fat in the cheese, up to the limit that acceptable cheese quality willallow. This should contribute to the development of reduced volatility of fat prices in the long rtm.W ith respect to the use of whey oream in the manufacture of Cheddar cheese for processing, whenthe price of milk fat in other classes fs low, there will be an incentive for the cheese maker to try torecover fat from whey cream and incoporate it in the cheese. lf the value of fat in other classes ishigher and if the value of whey cream that could be sold outside the plant exceeds its use value ascheesea then whey cream will move into the market to provide an increased supply of fat forutilization in products (e.g., ice cream, cream cheese, etc.) in other classes when cream is tight.

12

Step 2. The Value of the True Protein in the M ilk Equals the Value of the True Proteinin the Cheese Plus the Value of True Protein in Dry W hey. First, the value of the skim portionof tbe cheese is calculated. The skim portion in a pound of cheese is the fat and salt portion in apotmd of cheese subtracted from 1. ln the example (Table 1) the skim portion of the cheese is67.0401% of the cheese. This value is divided by 100 and multiplied by the NASS cheese price perpound (i.e.r (67.0401/100) x $1.3064), or $0.8758 ptr potmd of protein. The full Cheddar cheesemake allowance ($/1b) is subtracted from this value ($0.8758 - $0.1702) to give a value of the tnteprotein in one pound of cheese as $0.7056. The value of protein in a pound of cheese divided by thepounds of true protein in a pound of cheese ($0.7056/.239208 1b) equals the true protein value($2.9498) per pound.

Like the value offat in cheese, the value oftrueprotelnperpound ln the wheypowder lsasshned the same value as the NASS whey powder price (i.e., $0.191 7Z/â ln thls examplehAgam, this is an important assumption, that relates the value of true protein in dry whey directly to

the changes in value of whey powder in the market place. The remaining value of the whey powderis assigned to the other solids âaction of milk.

The value of tnze protein in the milk is calculated as the sum of the value of tnze protein inthe cheese plus the tnze protein in the whey. The weight of tnle protein in the cheese (2.2861 1b)divided by the weight of true protein in the milk (2.9915 1b) multiplied by the tnze protein value inthe cheese ($2.9498 per pound) plus the weight of true protein in tlle whey (i.e., true protein not inthe cheese = 0.7054 1b) divided by the weight of h'ue protein in t'he milk (2.9915 1b) multiplied by thevalue of true protein per potmd in the dry whey ($0.1917 per pound) equals the value per pound oftrue protein in the milk ($2.2994).

Step 3. Calculation of the Other Solids Value. The method of calculation of the othersolids value is also different than in the current system. First, the yield (6.2728 1b) of whey powder(the calculations as described earlier in the description of values in Table 1) is multiplied by the priceper potmd ($0.1917/lb) of whey powder. This provides the total dollar value ($1.2025) of the wheypowderproduced per htmdredweight of milk. Second, the manufacteng costperpotmd ($0.l37/1b)of whey powder multiplied by the yield (6.2728 1b) of whey powder is equal to $0.8594 and this issubtracted from the total value ($1.2025) of the whey powder. This provides a net value ($0.3431)of the whey powder after removal of manufacturing cost. The value of protein in the whey powderwas previously assigned (in Step 2, above) the value per pound of the whey powder ($0.1917). Thisis multiplied by the weight (0.7054 1b) of trtze protein in the whey powder to give a total value($0.1352) of the true protein in the whey powder, which is sublacted from the net value of the wheypowder ($0.3431) after removal of matmfacturing cost. This provides the residual value in the wheypowdtr for other solids ($0.2079). The residual value of the other solids is divided by the originalpozmds of other solids in the milk (5.6935 lbs) to give the value per pound of other solids($0.0365/lb).

The valuts per pound of each componentlfatstrue protein,and other solids) calculatedin steps 1, 2, and 3 provide the values used to calculate the Class III price for m ilk of any

13

compositlon in that month. A calculation of net return to the cheese maker f@r milk with3.5% fat, 2.9915% protein, and 5.6935% other solids is also shown in my example. Thepurposv of sltowilg tlle calculation net returns to a cheese maker is to ensure that the newsystem is working correctly. W hen the calculation of the fat, protdn, and other solids pricesis working correctly, it produces a net revenue of zero when the Class IH price is calculated.The net revenues on milks-of ather clm positions other than th-e milk com positiomused in th-ecalculatlon of tlle fat. true proteim other solids prlee: will nlt-be zerm This will l)e explainedlater..

This calculation (rounded to two decimalplaces) anives at a Class III Price at 3.5% fat, 2.9915% tlueprotein, and 5.6935% other solids of $11.66, while the current system arrives at a uniform milk priceof $11.51 using M arch 1999 data and the same default assumptions. The difference between thecun'ent system Class IIl prices and the system that 1 have proposed in this presentation will vary frommonth to month (when using al1 the snme default values), but for the most part on average they willtrack about the same. Thus, the two calculations produce a similar Class llI milk prict whenthe same assumptitms are used in b0th thepropostd and turrent mdhods of caltulation. I willleave it to others to calculate the comparison of the Class Illprices under thv current systtmand my proposed calculation across a period of time using different monthly prices.

At a fat test of 3.67% (with 3.1% true protein and 5.9% other solids in the skim portion) the currentsystem produces a Class I1l price of $11.74/cwt (when using the March 1999 data), while thepropesed new caleulation produces a yrice of $1 1.88/c% (a smaller difference). This is caused bythe fact that tlle plice per potmd of fat m the current system is ($ 1.4487/1b) higher than the price perpound of fat ($1.3064) in the calculation that I have proposed (when using the March 1999 data).This relationship will vary from month to month. W hen data from other months are used for thecalculation, this relationship between the two methods of calculations will change because in thecttrrent system of calculation the variation in the butter fat price used in the Class 11l calculation isnot determined by, and does not vary in direct proportion to, variation in the cheese price.

An important point is that the system that I have proposed 5vil1 reduce volatillty in the proteinand fat prices compared to tht currtnt system. The new system solves the problem describedearlier in the current system as it will not produce a reducfion in protein prke ptr pound andskim value, wlten fat value increases. The fat and protein prices for Class III will move togetherwith the cheese price. The sensitivity analysis presented in the next section willprovide an evaluationthe default values that have been assumed in botb methods of calculation.

A copy of tlle spreadsheet used for these calculations is provided.

14

?oa Pu $ s? m 4o ol o p- to to >. Ja o = to = ?w o t/ tv = m o o e7 >. w- o cz to p s. to ol = v' % = x oP m j 0* w'?- *s. OJ î*y j Y.v *j j j ''g.- j 1 Q'a 'g,o j j gv CX

. j Bvq j j j Oxg j j Ug j v-eaea Yj **' 2. j tw %s. j Wm ogom =' w m cq e!r j! t) * 4p (n o. txf e4 N o to tu o m o cn e) tx o v- o w x. to o go o o co o o o to .e. tu o 44 % v. o oo xh zy B c- xt .- co tu to e) to w. w

= > s .jtn <

EX eoArmmooooNem o m * % m o % o e ls N o c m o o m m r o G > O e

V>* o m o x m n o m o N o m m N Q D * G >o#m >8>*m* k*o o>N DtL d. g , G o m o o a oy * g g s.o xpw xo y g y pg g om g g ay g p x g) *x. g g j m S S œu g xy y* 4J m t% EX W V O VD r%P > n C) m - P1 G1 lV o * (N D - o (D O P1 D - D D * = QD P1 D o tg C> ID D to 7 e: o ** * G ts o- $ 45 t- * m 17 CN < GJ A < v

*r m :7W U Y *o

P (kb. 3$9 >X toJ> v s4 F2= m <m s' œ

to o up o to a m o cp Ap >. cz g e> *4 %> N to zvp rw o g g >. g o xt oz el tp : t/ xr w' o j: o**VO Mwomoo mooom x mw == uexo g >-m-* om g- o

Xm/ 8 oo oozmoomooswm %N xmoog aaoomxsnw og o

mxD nc m .=zvaVa. o=== >a o o o . o m e o e = . . 0 * o o o o ov œu i m m m - w - a o o - o - o m o w o . 0 o o& o N= om o o o m G w o ' % *. d 'o

C = a - - o = owooo V .o c. y p :2 n vn o mse. u R > <

X V*=.*&SX m'o mœ*s ooooeoxam RoA oxev a A R N r o o m m a R o o xœ o o r o

% O N of3 (5 : k cycy g 8 & 40 'd 'Mug j N j co Yo

*,koa M. u'u'w

'-ox j -<. j j j ! vj 6vo j wotx P-r bicq Noo. jj j QJ. $ 4 Om IC vs Oou a .y to m = o p. m tv wJ v' to r/ e* 9i cp e) u: o, o, e: = to ru <> v cp zn ro nt <:> v. cp v. <' to jj <p o o co o rp = g; v ru o r xr .. = o% t R * a s. >- = ro tx to m

m-= = 'u o 'voqx:*Q.ps&j !# r

c= a re fc .t5m a ,- we =mN o e o * & o o e

vovz *0 >O=oN* Noo**N%+* *% oR cl o a .a S m. e v. vo Y to c: 2 c, o v. >- = to ua o v, q> c? o ,z w v o >.- . (u -œ$ $g -= e,- (p -& G = c, oQ' u =<c : #$ e v.* <2 * s.s. to* .% G'G' *v. G7 œo 6/6/ 'D'D 'r/m tutu Gs. % b'm oyic? 'J Mg &= =x- Wco GJ :4 t'J ''-m u'a epep =q 04 'oo ''>m 3 M *q Oo'D* 'j a. k5 t.p kn to el o, ol o zo o4 o .- o cn co cq o v- o ot e, w. <' s m =, ez o ko o m v. kq qx r w .- u om ê u s. s. .- rp w rv ov

.-tz > -,7 m /u fa tu oro to .w o * em r r .e.w w .. Jes e

ym o. = .$to mo

G < -c >) t!

sa. p v.

a a.- œ o 2 m

% u e

o. *E >

xa pq m c> f.l

y .-. ..c jE 1 a 0. z g sa > .& m o'-x m @ $ u w = * : %o o Y g = : : a a txl t; >

.4m : r e z' m

p x> = y..$ >.

s m $ tl p : vc vq u, -Q >* e e

m 4 m * & <: 2. = -J E< (s q$ .y. w 0. g ./0 >. t:u o cy x p ry x5 m z: %- â

* >' >' o = w. .p -<

o aa i5 - & w, p v o c' o s. . xo :> ,s . o ',5

p .c .: vo to s =: r r s b'Q <: q' q? > n G 2t a: v e, ez > : .'.e. z qh s s > e, . .= .

s s a s s m .0 sk pt aw x. 1: p A œ m o

vy.: v :> = : l 15 a m . m - = n !: s w u =: o r :

2 a t) u =. '6 'c f =w s , : v.g $.y= o a x v> .j -J .v% ue Y o t J? I : o s tz %% iu .n - a E a o e z' '-a @

-5 o = = r: o e w s o r s7 s a $ o zq a. w l :p : t: o.a o * e g av s tutv >. = g j o %e . p: y r txt o k .o s. rao m qa oo s xx4 m o o.L= = w . w

>o Y o o % %

œu m yr es mn-œ; d 0* *o$X#Gw<R Xw o =* *w *1**0 IXX œO :a * a gYTbX QWVOY % Gb W y==9 Qro o * 'a o = o - .# s ogo o K N< SY Y=xRGGA - œ%*r=m = y gccœo os22 amw xy o vgzuzmmwoooo y-=rooœ 9 a y .* o . x x o s / = w = vp=4asxoo% a2a

<x2 = = e. docoysoo wn auyo. yp wo uo a o m m wz xxa ..o jAIZN m y awo mmwyn; owox qa sy. osxo

oo c u c s o u o wï-nm eMz ay M

coxw ywgmxwsxo a-yxwopmmsso jwopooeacw bg2zX o* oœ-- oow m ooo o g ) gaaœec œwob x= rz xgw w

o o o d E o o mgco ;oo *ay$ acaw: ama EoYm2g, eo>aso m.. = o gmwaa i

oo a

*. œ=s:g*j:m= = u Ra. s tvoo vsxaq ov) r % o a. . J jo o- SQ m N = .1 (p '$ ': w a xs Y

a. . .. o ej xv.

.g tk =* w m (> tx tu s m to .0 .y w e q, x , o x) -a

o. a .z .w= = 1J z'4

- a oO a oN z'S- .'' A- d 71 co *. r:0 < .@ .m é xo p= sf 3u= o! uq j.o yl j m b.o s, x. x:* i! > > > > tu = fa > z (A oW . j Q fa >vR! *

w X u9 er B .ç -te <>

2 p I w.

e -9 # .5* u = O

> œ * >

V>&A o w a = o- w o < v m = o o w w - - w > % = oS Q) : f: g to m tox g ooo g pg V p

ao .07 y y tg g ojo y w xrv ovn j go * uz (o >. v- y1oz p? 00 o #> tp v' N q +: e: v' to v Gl tj p) o os tn <' v kr >J tp e? oP NM - <> o :9 c> rv o <> o t> *y h- o (v ev tp m + o o

œ? zv :5 J R ov v. o uy 7, uy vy * <. v, .

u t.p u J; u / vy= & -o < > 4

oY W m tp tD nl * 'e b' @ï u: tV m m = v * o v' m to to m fk t% 4/ 4% o

LQ yv v. * Cq o tH <> vm h. lp * v' >. to = xe N o v' e: ua w o :% >. otx < rR N 6V o * (N * >: &* tN v. tp ju @ gk 4.* 4 >s. x.o xc $9$9 * *0 C>

Q * o o s. w v; a w. 4p v- 4v o

tq qD) > - tRQ Q y

v o o o - = (N *- <> o o o o + to v' p. N v. c >* *- = IpJ ï o 4* * * * * e *A to fa * tp + >j. <A y m vy- * *O Qt *

tkzq *ya Om œ tnm œ o

.*iv m a u a o m cq o >. >. 'e p- Ae tx to - * e) * o o o o k o

L) = - = - e/ o o o t. m Ko U N q: x.. e. * <p jj p w to u

q 12 tx ID s. oo (J e (k Q! 3 may 7r Rp @ mx- ;; / m'e Om g wm f*q c s tu too ow sc @ k e)- no % so* V 4 V '$ J o o a N o r% = c q> o o o > = c)

r >. $'4 m o # v' o o

=* a 2. = 4+ * * *

w m to m w? fa tv * t,v tez v, Ap *= v D = =

Q e(J < > q *

#z n'q &-p*Y2a *o mmos : o Nm * ON ON mo * **o # g ooo+yaoo mo m ov R -*4mN > œ

eN r smow g0= ; < *- < >* G o m w o o o œ o w

o * NU e n * < - œ o a N * o m q X-m h x x- wv zoo

Ylle > o o yo - y=o o o ymo

o N muo '%G oCopï * * * w *

* * * e * p w u u- *-=-= oe1<X<

!; = o= > a.a Y . w * tp tty <* ew * kD * Cq G m % to cp G h. jg A m çc * =to to o to o g rn ol oï 1V rw vo txl n 4o ol s. T o ej o* r G J; o >. n % oh o

to % (c o o >. s. o ro <' to o 4 m to oc) T wo o 40 =) = pw m v; (u eq to tp m PJ o o * o o v m e: Tp @) o1i D tp cz o el o tw g o o o a a % m o ;: sv e: <' * g; tp oY - w- D fo Rz eo * * tzz

te iey vy u to s t.* ; :

ot= 'R % f;t) IJ e- O

Y' + + Y **

ex=

>,S S .q: E .7

)$ 4:'g. >' = ct E2 * *

, * î o a$:) y: tly 1$, -*-. ,;It= tp a --. .* .Q .1 (1a. m p. a g * gm o e; - Ye % Y .-% c, -B 1y 22?-. q' ./1 rv = :Q 4 2 x

.% -; -*= = = w s. cv = w tx

o ?w u tx s. pw yy s%- o

.1 to et m o * q,: (j o =. -. to g a s .em ao .s g zj vj .lj= g >< oï; B é tT tz >G - fv 23 = @ > = u= O .,w G .ç w $ Ww o g = ae Tnm g = ='o. (h 4o g. o $ * oR x u, .- m 'u = oGrx

l'No s . f .:a à = = p; tg .43 ug .== .% t: .* .% tn = qxs co o, m *x p o : > o

. n o. u = ot o u to =

..c w* o uv- i a

. E E >' * t> .)$ to : m $* =Kj a <, .G ql % O G Q' 2 $72 j .

:2 G t' *ï Y .4 ge $t 'j s * @ v, s lr

oS z 4) .,> w zx a .S, .5. w a. b-s tq .. v j =17 s 6 *. r = (u V = xy &* Q = m t4 to .2 ptG

w c m s g : yo y .s y s p. cx a'2 * 2 % * > * . - o >; .. jg pw % -q

s (p 2 ro o * *. > 3: s .P a .@ o .: .q .2 w rx e) o a

p. '!J = .e c w'- : m 8 >, x u g E * 8 5 =e g, o to .v

m J a .y p. g; .0 n . o. a . o v x xo

o = > * s .,

ï s $r g. * e o .e e = x q. p e .j a : oag w g a.. s çav % z j s v. a qg a atv wz a, a s

:5 r;

=o > G *to = o œ w 4 * o u. -Fu m > * g

o s m L t> tK a o .s ! Y N G a a2 4 os 6 2

= = up c x .. o

t4 0. x o y mpw p = a g, kx s g .> x, : ::r a x: s g çs o tv =

w : s ck - rs m .e a

N : v o = m .y .$: > <1 u x cp .

.?o .= o u (M l a. s s u y ar: tx çx <p o n

. .ç .2 = a! y b :E ck J! 47 * r .-. a rx -= = = a +r yv u w o / c. c u u a < = p œ % u ck 4. ..m X Q' .: = * w.

o o = >( o u .- o o :Q rç m o v, -Q

Y X .% 2 G E o, : >. o o qa :4 ax a g. * R s. o a. .c, a tz o oo o o c: s A . qI 'j o. a! o. e .j s* : c s y /; c =s xe s. xs = oz tz .,z

@. = a p * w o

o u mo = o , u o. a ï o*4 : sjsorcs yvssm u oyy: # o

s2 = t e o s - a e .= :e * > ao m-- ou co vo ea K.

o o 5 = :

> > w = x wgzawo oxvzs g zozyvowxooa :

> R o w y zm xz e :

<.. a oou : Wo = o m

m : o oz n- o-o = z =< . m m m & e-* v x > /1* = x= =amm- g œo woaam n a œ

=x . m::K*or j ss .o * o % w

a y - gyrro xloop = o ocD:

œ 2 o 4. Y = oAs m. roc o n z>

o* m a

o o aX Gw oo:io 2***2 % w o ox a

*v u> *< œ G a s .: $ fm=pO4*. :**.*0*-. n wo

-œ0*w,%o > = > m

x s =>l a a s oo

o o a > a 25 / * m =. a** % x

w oxm vœ V A l1

: p, sPa G rr /la - â> > e=k=. o 5 o j o. :y axa a jia- >*: =e ae a Y w :. wx , oet: x cR o o 0 o o: 1 2:o a =m. a%> =o . >. . G ia a œ - y : qauw =

z2. c > c z z< o : *:%! ; Rovjloym v.

xœ o o o w

m g a = w wx . ,=-w= a m xR ? 2 *> == =r% o aj sjj / mvmo=couBtdno. l x

o ! j= s - : * px e am # 0. oo . o r o o

s m > œ o n w o x w2 n % R 2 G*w Y *.> = ,

**24:2* A m * *. o w o . o

p m . .o *SJ o œ co d*o s x E m gu y > gœ = m gsm z z-nœœ l w m> l g y .H j s ov o oR -aa = so8xœx-c% * l n * * = Em c w a : o > -x mo w a xx w s 4 1o œ o x r . = s: a: o y op o gs o a - o:* 1 a e e m=* m * * Vx. % ==!Rjo2 . wsw o gyœo2 o w = == s 22 *o* - FP G o

x xo * m p o oo 2 B o>W 9* G >œ :

% - o ooï wo z e ï= n >> >*'D p œv Lmsm* $ x zm % vz..

' o - o - so * te o >*Z*2o- œ œo - =kûs . ? 11 u *3 ow z c

c o aa s = <e >

a x# t ** K

v mr œlr > 22 P v1%a % -

2 . a ' , m œ x o Y w p s 3r fo o 4, m o x j..4p = $x . = x = * q, qj o. n = = = as o'** 3 d =Y 'em 3 4 .,5 ..5 B 3 7) S -0 -tv 15 % %t: e) (r1 b- > m > J-. +. !- v> %. 1 > > a: o

o

SENSITIV ITY ANALY SIS OF FA CTO RS IN CLUDED IN TH E CLA SS lIlPRICE C ALC ULA TIO N TH AT I H A VE PRESEN TED .

W hen the uniform price is calculated for Class II1 milk at 3.5% milk fat, 2.9915% tnze protein, and5.6935% solids not fat, the uniform 3.50% fat milk price is established in dollars per hundred weightand a value of a pound of fat, a pound of tnle protein, and a potmd of other solids are establishedfor that time period based on the NASS Cheddar cheese price and the NASS dry whey price. TheVanslyke theoretical cheese yield equation is used in these calculations. The Vanslyke formula wasdesigned for full fat Cheddar cheese with a moisture of about 36 to 37%. Other cheese yieldequations are available that have been optimized to work with other cheese vmieties and underconditions of milk fortification (2). The factors that influence the calculated Class l1I price and thevalues of fat, tnle protein, and other solids can be separated into three different categories: 1)technical factors in the Vanslyke cheese yield equation that influence the calculation of protein valuein the cheese, 2) make allowances, and 3) NASS cheese and whey prices. Once the Class 1l1 valuefor a pound of each of the components is determined, then the Class lll price for any milk can becalculated.

In this sensitivity analysis, 1 look at sensitivity of the milk price to changes in various factors andprices. A comparison of the sensitivity of the Class Illprice to vriation for differentparameters mayhelp direct the attention of the industry to those that are the most important and avoid too much timebeing spent on factors that have little impact.

Technical Factors in the Cheese Yield Equation and Calculation of the Protein Price.

The Vanslyke Cheddar cheese yield formula is used for calculations in the current Class IIImilk pricing system and I have used the same formula in the system described in this presentation.A review of cheese yield formulas has been presented elsewhere (reference).

The Vanslyke Theoretical Cheddar cheese yield fonnula is as follows:

(40.93 x % fat in milk) + (% casein in milk -0.1)) 1.09

1 - (target cheese moisttzre/loo)

The casein content of milk is not as easily measttred as the fat content of milk. Howevzr, inrecent time both the crude protein arid more recently the tl'ue protein content of milk have beenroutinely measlzred with both chemical reference methods and electronic milk testing equipment. Ithas been common industry practice to use a factor multiplied by the crude protein content of milk toestimate the casein content of milk. The most conmmnly used factor seems to be 0.78 times crtzdeprotein. However, the average value for the US milk supply is probably between 0.77 and 0.78. Ina National Milk Composition Study that I conducted in 1984 for the US milk supply (3), the averagecasein as a percentage of cnlde protein was 77.93% and the average casein as a percentage of tnze

17

protein was 81.95%. At the time of the 1984 study, the current ofticial AOAC methods for caseinand nonprotein nitrogen were not in place and the methodology was a little different than that usedin a more recent stttdy. Since 1992, my laboratory has monitored the casein as percentage of crtldeprotein and tnze protein for milk from several factories that participated in the 1984 sttzdy. I haveseen no trend for a decrease in the casein as a percentage of true protein in these milk supplies. Ifanything there has been a slight tendency for the casein as a percentage of protein to increase. Thisincrease has probably been due to tlze atlentitm that has been focused on improving mflk qudlity (e.g.,reducing psychrokophic bacteria cotmts and somatic cell count) for cheese maldng. Improvementin these quality parameters for a milk supply would tend to increase the casein as a percentage ofprotein, because of reduced enzymatic damage to casein. M ore recently, my laboratory hasmonitolvd the casein as a percentage of true protein in bulk milk supplies in NYS at three largecheese factories. These data were reported in October 1999 at the Cornell University AnimalNutrition Conference (4). Test values reported for the 1992 to 1998 period below were determhedusing the oficial AOAC Kjeldahl methods that are in place today (5,6,7). Composite monthly rawsilo milk samples were tested monthly for cmde protein, true protein, nonprotein nitrogen, and caseinfor each factory from 1992 through 1998. Over that seven year period, the average nonproteinnitrogen content of the milk was 0.192%. The average annual casein as a percentage of true proteinfor the milk supplies in the three factories was 82.17, 82.17, 82.42, 82.15

, 82.12, 82.31, and 82.19fo< a seven year average of 82.22% casein as a percentage of tl'ue protein.

The influence of the selection of constants for use irl the Vanslyke cheese yield equation (forfatrecovery in the cheese, the nonfatpnoncasein solids retention factorinthe cheesezmoisture contentof the cheese, and casein as percentage of true protein in the milk) on the calculated Class lI1 tmiformprice atld the net retunzs to a Cheddar cheese maker are shown in Table 2.

Fat Recovery in the Cheese. An expected fat recovery in Cheddar cheese is used as an inputvalue in the Vanslyke cheese yield formula. The current pricing system uses a value of 90% fatrecovery in the cheese for calculation of the base prict. As is shown in Table 2, an increase in fatrecovery value assumption of 1% causes an increase in the Clmss I1I milk price of $0.024. Fatrecovery in the cheese is a parameter in cheese making that the cheese industry monitors closely. lnmany factories, the fat content of the whey as it is being removed from the cheese vat is determinedas an index of fat loss. A value of 93% fat recovery in the cheese is achievable at a commercial level,however not a1l factories achieve this. Recent advances in design of large scale, enclosed cheese vatshave been able to achieve fat recoveries in the cheese that approach 93% . The value of 90% fatrecovery in the cheese is probably low for large scale modern cheese factories. ln my opinion,the most appropriate value to use as a default value currently is between 90% and 93% . Astechnology of cheese making continues to advance, these values may change and they may need tobe re-evaluated periodically.

Nonfat, Noncasein. Solids Recovery in the Cheese. The l .09 factor in the Vatlslyke equationassumes that there will be 1.7% salt in the cheese and that some nonfat, noncasein, milk solids (i.e.,other milk solids) will be retained in the cheese. The cmrent pricing system uses a 1.09 factor andthat value has been used traditionally for Cheddar cheese that contains about 36 to 37% moisture.This value is used in the current Class lll price calculation. As can be seen from Table 2, the

1 8

calculated Class 1lI milk price is sensitive to this coeffcient in the equation. A change of 0.01 in thiscoefficient causes the milk price to change by $0.0966. ln my opinion the value of 1.09 is a goodvalue for a Cheddar cheese that contains about 36 t037% moisture and 1.7% salt. Given a constantsalt content of 1.7%, the true value of the 1.09 factor will increase with increasing moisture contentof the cheese. This happens because there are other milk solids dissolved in tlae free moisture portionof the cheese and as moisture content of the cheese increases so does the nonfat, noncasein milksolids content of the cheese. The actual moisture content of the barrel cheese reported in the surveyis usually between 35 and 36%. The moisture content of the block cheese reported in the NASSsurvey must be greater than 36.5% moisttu'e. Thus, in my opinion the 1.09 factor is probablyclose enough, given the importance of some other factors that will be distussed.

M oisttzre Content of the Cheese. A value for the target moismre content of the cheese isused in the cheese yield calculatitm. Cheese yield is very sensitive to moisture content, with cheeseyield increasing with increasing moisttzre. Therefore, one would expect a change in the assumptionfor cheese moisture content in the Class 111 price calculation to have a large influence on the milkprice. As seen in Table 2, an increase in moisture content of 1% causes a $0.1608 increase in the m ilkprice. The Cheddar cheese moisttu'e assumption in the cunvnt Class lllpricing system is 3S% and1 have used that value as at'f assumed value in my proposed price calculation. However, the mostimportant point is that the value assumed in this calculation and the moisture value for thecheese and price for the cheese induded in the NASS survey must match. Unfolttmately, onlythe moisture content of the barrel cheese included in the NASS survey is known currently. I thinkthe dairy industry would be betler served if the moisture content of a11 cheese in the survey wasreptvted and a cheese price calculated at moismre content that is the sam e for il0th NASS moistw eadjustment and the Class Il1 yield formula calculation. The sensitivity analysis in Table 2 uses aconstarlt cheese price for a11 moisture contents and therefore shows a significant variation in milkprice. The magnitude of the milk price changes sbown in Table 2 acmally demonstrate whathappensto milk price when the moisture content of the cheese included in the NASS survey does not matchtlze assumed value used in the cheese yield formula. However, as 1 explained earlier in this reportthe tnae average of tbe 39% moismre adjusted barrel cheese and the block cheese of unknownmoisture content is probably near 38% and, thereforerunder the currentprice calculationthe moistare

adjusted cheese price, the mean moisture adjusted basis for the cheese in the NASS survey, and thecheese moisture assumption in the current Class 111 price calculation seem to be comparable at about38%.

Casein as a Percentage of Txue Protein. The current Class lIl micing system used 75% ofprotein as casein to arrive at the protein factor (equivalent to a factor of 79.76% of true protein).Second, this value (75% of cnlde prottin) is, in my opinion, too low. ln the past several yems, Ihave been approached by cheese makers that have been concerned that the casein as a percentage ofeither cmde or true protein is lower than normal. ln every case that I have been involved with, thelow values bave been traced to improper methodology for measuring casein or poorhandling of milksamples during collection and the time immedîately prior to analysis. A paper on the proper handlingof milk samples for casein analysis and the a description of the chemical methods for caseindetermination in raw milk has been published (reference).

19

Typically a value such as 0.78 times percent crude protein in the milk has been used in the cheeseyield eqtzation as a substimte for a casein percentage. In Table 2, 1 have shown values for 75 to 79%of crude protein acd the corresponding values for casein as a percentage of true protein. The valueof 0.78 on a crude protein basis is almost equivalent to 0.83 on a tnte protein basis (Table 2). Astht default value fo< casein as a percentage tnte protein is increased, the Class 11I milk priceincreases. The milk value increases by $0.0616 for every one percent increase in casein as apercentage of cnzde protein. The value would be slightly larger on a trtle pzotein basis. ln myopinion, a value of 82.2 to 82.4% for casein as a percentage of true protein is prohably acorrect value for this param eter. This is quite different than the asNumption in the currentprice calculation that wa: used to derive the protein fador.

Cheese and Dry W lxey M ake Allowances.

Cheese. 'I'he calculated Class 111 milk price in the current milk pricing system and the ClassHl price calculatitm proposed in this document are both sensitive to the make allowances selected asdefault values. The sensitivity of the Class I1I price in the system that I have described is shown inTable 3. As make allowance for cheese changes by $0.01, the milk price at 3.5% fat will change by$0.0956 per htmdrttl weight. W hile cheese manufactaring eost is a very important prameter, itchanges with changing economic conditions, scale of production, and advalzces in technology.Therefore, surveys and collection of actual data are probably the best approaches to keep thisassumed value cuaent and realistic with conditions in the industry.

W hev. The make allowance for dry wbey is also an important component of the Class 1lI milkprice in the current Class lI1 pricing system and the one that I have proposed. As the makeallowance for whey increases, the milk price paid to the farmer decreases. As make allowancechanges by $0.01, the Class IIl milkprice changes by $0.0627 in the Class lllprice calculation systemthat 1 have proposed when al1 other assumptions are the same as the current system.

NASS Cheese and W hey Prices.

Cheese Prices-. The Class I1I milk price is extremely sensitive to change in Cheddar cheesepdce, as it should be. As can be seen from Table 4, an increase in the cheese price of $0.10 perpotmd will increase the Class l1l milk price by $0.9907. Since the value for fat in Class 1Il isdetermined directly by the cheese price in the approach that 1 have presented, it eliminates thedecrease in Class II1 milk price to producers with a fat to protein ratio less than 1.28 when the fatvalue in Class 1l1 decreAses. Thus, changes in clxese price will cleax'ly dtive ehanges in Class III milkprice. The accuracy and represvntativentss of the NASS cheese prke is critkal. Alst), theharmonization of the clœese price and the moisture basis is extremely important.

The pay price to a farmer at constant milk composition will increase when cheese priceincreases and decrease when cheese price decreases. The calculatcd Class I1l milk price using thecalculations that a 1 propose and in the current calculation is the most sensitive to change in oheeseprices. Therefore, big changes in cheese prices in the market place will drive big changes ill the milk

20

price, as it has in the past. However, changes in fat value in other milk utilization classes will notcause the slcim value to change in the Class IlI price calculation that l have proposed

.

W hev-prices. W hilenot as important as cheese price, the whey price does influence milk pricein this system. ln the calculation that 1 proposed, the whey price directly influences the value of thetl'ue protein from the milk that goes into the whey. As can be seen in Table 4

, an increase of $0.01per pound in the whey price will increase milk price by $0.0627.

Calculatitm of M ilk Prices ill the Proposed System and the Current System -

Questioning Some of the Defaults.

As mentioned earlier, the two methods of price calculation (current and the one l havepresented) return similar total Class lIl milk prices when they start with the same assumptions.However, the two systems anive at different fat and protein values. The system l have proposedtliminatts the decrease in milk protein and skim price whtn milk fat price goes up and visa versa

.

lri my opinion, some of the default assumptions need to be evaluated frem a technical basisfor their correctness. Changes in these default values will cause the same direction of Class I11 pricechange in both the current system of milk pricing and the system that 1 have proposed

. To illustmtethese changes in default values that I think need to be evaluated, 1 will present 5 colunms of da= inthe form of a spreadsheet (pages 15 and 16) and calculations on the spreadsheet that illustate theimpact of the default value selected for each paramder.

The first column of data reflects the current default values as used in the current Class 1l1milk jrice calculation. Some of the default values were used as the basis for the derivation of theprotem and fat factors in the current system so they are part of the assumptitms

, even though theymay not be visible in the routine calculation each month in the current pricing system . Thecalculation of mllk price using the M arch 1999 data using my price catculation produze a milk prioeof $1 1.66 for a milk of 3.5% fat. This is $0.15 higher tban tbe price calculated using the currentsystem for the March 1999 data. As mentioned eadier, this difference between the two caleulationswill vary from month to month and other people are calculating tbose relationships.

The second column of assumptions and data represents the outcome of a change in theassumption for cheese moisture and cheese price that corresponds to that moisture content

. In thesame fashion as NASS does when they calculate a moisture adjuse ent of bm-relcheese compositionand price &om the level of 34 to 35% (its actual moisttve at production) to the value of 39$$, 1 haveadjusted the yield and prke per pound of cheese back down to 36% moisture. In reality tbe cheesewas never made at 39% moisture and never had as higlz a cheese yield as indicated itl the first colunm .

By xaising the moismre content to 39% and lowering tht pdce per pound of cheese, the fixed cheesemake allowance ($0.1702/lb) is sublacted from a lower cheese price. In my opinion, this results intoo much make allowance being subtTacted off the cheese price. l have lowered the assumption forthe moisture content of the cheese from 38% to 36% moist'ure and adjusted the price per potmd ofcheeseupwm'd from $1.3064/114 to $1.3485/1b to reflectthe highervalueperpound of cheese at lower

2 l

moisture and then recalculated the Class IlI milk price. This price per potmd of cheese ($1.3485)is closer to theprice that was repoled in tlje NASS suwey before moisture adjtts%ent. To what maybe the surprise of some individuals, this change in assumption at the ppint of calculation of the perpotmd values of protein and other solids produces a higher Class I1I milkprice not a lowerprice. Thecalculated Class II1 pxice increases from $ l l .659 l to $ 1 l .7240 or about $0.0649/cwt. lf one goesback to the cuzrent pricing system and makes the same changes to moisture for use in the calculationof the fat and pretein factors and then makes t'he same moisture adjustment to the cheese pzice, theClass l11 milk price also incrtases. Since tht barrel cheese was never made at 39% moisture, 1 seeno basis for adjusting the moisture up to 39% and the price per potmd of cbeose down. This inflatesthe cheese yield to a value that never existed and then allows for a make allowance bmsed on a higheryield of cheese.

The third column demonstmtes the impact of changing the casein as a percentage of truepxotein to a value that is mol'e <ep<esentative of tlxe trtve value in the mitk supply. The oliginal valueof 75% of cnzde protein (i.e., 79.76% of true protein) is not consistent wjth thenormal values fozmdin the milk supply when fresh milk is analyzed by the official reference methods for tnze protein andcasein analysis (7). The data referenczd earlier in this presentation has demonstrated that a moreapm opriate assumption for this value is about 82.2% of the tme protein is casein. If this assumptionis used in the proposed new calculation system (third column) it produces a milk price of $11.8664when coupled witb tbe mevious change in moismre basis from 38 to 36% moisture. The priceincrease due to this change in assumption would be about $0. 1424/cwt. lf this same change inassumption for casein as a percent of tnze protein is used to recalculate tl)e protein and fat factors inthe current system, the milk llrice will also increase in the current milk pricing system.

The fpurth column demonstrates the impact of changing the assumption for fat recovery intlle cheese from 90% to 91.5?4. This change produces a higher Class l1l ealculated plige in lnoth thecurrent system of price calculation alzd will produce a price increase in the new system. The pricechange is about $0.0344/c% due to this change. A value of 91.5% fat recovery in the cheese maybe more representative of fat recovery performance in modem well managed cheese plants, Somefactories will perform better than this and some will perfol'm worse.

The fifth column deals with the issue of tbe selection of a milk composition at which tocalculate the per potmd values of fat, tlue protein, and other solids. ln my opinion, the milkcompositionused for this calculation should represent the avemge of the raw milk supply as it wouldbe received at cheese factories. An estimate of this average is 3.67% fat, 2.9862% tme protein and5.6835% other solids (protein and other solids are based on a 3.l % tnle protein and a 5.9% othersolids content in tht skim portion). This estimate is taken only for the purpose of example. Adetermination of the average milk composition should be used as the base. W hen the previouschanges in assumptions are used with this milk composition, the calculated Class 111 miA price is$12.2221/cwt versus $1 1.9008 at 3.5% fat. The key point is that the ealculated pxice per polmd offat and other solids are unchanged by tbis difference in the selection of ti)e default milk composition,however the price per poxmd of protein increases by $0.0356/1b of protein.

22

W àat is imlmrtant ahout the selection t,f a milk composition for calculation t)# thisprict? The milk eomposition selected becom es the Rpivot point'' for ntt revenues for thecheese maker. A milk with a composition lower than average will produce a negative net retlma forthe cheese maker (relative to the pivot point composition) and a milk with a composition higher thanaverage will produce a positive net rettlrn, if the processor's cheese making performance meets theassumptions. Placing the pivotpoint of net retum atthe average milk composition for the milk supplystill gives the cheese maker the incentive to buy higher solids milk to improve profitability, as is thecase in the cunvnt system . W ith respect to the mtio of fat to tnle protein, the cheese maker willhavea positive net revenue when the fat to ttqze protein ratio is higher than the average of the milk supply.lf tbe fat is too 1ow for the amount of protein in the milk, then the cheese maker will have theincentive to add cream to maintain the level of fat on a dry basis in the cheese that is as high as isrealistic with respect to quality of ftlll fat Cbeddar. This is not different than the signal in the current

system .

This demonskates changes in net revenue behaviors resulting 9om milk pricing that happenboth in the current Federal Order system and in the new system of calculation that 1 propose. If thecomposition for & producer's milk is higher th% those assumed for the milk in the Class II1 profeinvalue calculation, then the cheese maker willget a higher net return on that milk. On the other hand,if the milk from a producer has a milk compositifm lower than the assumptions in the Class lllcalculation, then this producer's millt will cause a lower net return for the cheese makcr thanpredicttd in the cakulation. Again, this is not different thm1 what is happening in the carrent pt-icingsystem. The slope of these relationships are fairly steep and the slope will be influenced by theabsolute level of the cheese price. Also, if a cheese factory happens to have a milk supply that islower in composition tharl their competitor, t'hen they have a built in disadvantage in net retum eventhough their milk price was lower. This would indicate that these plicing approaches over payproducers with milk composition below the Class lllmilk composition assumptions used to calculateprotein and other solids values and tmder pay producers that have milk composltions that are higherthan the Class lII milk composition assumptions. 80th the current Class 111 system and the new ClassI1l price calculation 1 have proposed (that calculate a fixed price for a potmd of protein) do notaddress this issue, Thus, endproductpricing would correct this problem and would deliverpaymentsto each producer that would be linked respond directly to the value of cheese and whey that couldbe produced from each moducer's milk.

M ilk Prke Caleulator:

l have included a milk price calculator in the spread sheet. It uses the fat, true protuin, andother solids plices per potmd that are determined in the calculation done in column five. I haveshown the calculated milk price for five different milk compositions. ln addition, I have shown thetotal returns from cheese plus dry whey plus whey cream. I have not deducted the make allowancefrom these totalreturns. The make allowances are used in the calculation of the millcprice and shouldrepresent the make costs and some retm.n to the cheese maker at the milk composition used tocalculate the values of a potmd of fat, true protein, and other solids. As discussed earlierp since mostof the cests in the make allowance are in reality fixed with respect to the volume of milk processed

23

and do not vary with yield of cheese and whey product, it is not meaningful to calculate a differentcheese and whey powder make cost for each different milk composition. The total revenues perhtmdred weight of milkprocessed forthe cheese maker increase or decrease, respectively, as the milltcomponent concentrations increase or decrease. To maximize the total retttrn on the milk used tomake cheese, the Cheddar cheese maker must control casein to fat ratio in the vat. This is nodifferent than tmder the current pricing system.

IDEAS FOR THE FUTURE AS THE DAIRY INDUSTRY CONTG UES TO ADOPT NEWTECHNOLOGIES.

It is possible mathematically to keep the net returns to a cheese maker constant across a1l milkcompositions, without producing the decreasing protein (skim) value when fat values increase as itoccurs now in the current milk pricing system. However, the approach that would be used toachieve this would calculate a protein 'Ivalue'' for 'çeach'' milk instead of calculating a fixed proteinprice per pound that is applied tmiformly to milk from all producers. The disadvantage of thisapproach is that it would be harder for producers to understand, unless the price calculation wasconverted to a cheese yield and whey powder yield basis to commtmicate to the farmer. If this wasdone, it would be very easy for a farmer and the cheese industry to tmderstand the milk price. Anapproaoh that would ktep net vetttms to the processor constant (given a constant make allowancein $/1b ofproduct for cheese and whey) on eachproducer's milkwould increase the difference in milkprice between milks tlAat have 1ow versus high protein and fat concentration. This approach wouldalso more correctly return to each farmer the true value of that milk in Class 111.

An approach to pricing that holds net revenues for the cheese maker constant on a11 milkcompositions would also put cheese companies that happen to have different average compositionitl their milk supplies on a more equal playing field. The cheese factory receiving a milk with higherconcentration of fat and protein would still have some competitive advantages with respect tomanufacturing efficiency, but the competitive advantages for that cheese manufacttlrer that arecreated by the pricing system would be eliminated. Cheese factories that have lower marmfachuingcosts per potmd of product would still have competitive advantages over those with highermanufacturing costs. The interface of this approach for Class 11I pricing with other classes could beproblematic unless the value of components is more completely reflected in other products. Adiscussion of tbis topic is outside the scope of this hearing. The dairyproductmanufachuing industryis not at this level yet, but may be someday. At that time, many of the limitations in the ourrentsystem of milk pricing calcula:ons will be more easily resolved.

CONCLUSIONS:

1. The current milk pricing system produces a decrease in protein and slcim value as the milk fatvalue used in the current class I1I calculation increases. This results in abnormally highprotein prices (and skim milk price) to cheese makers when fat value is low and the reverse

24

wben milk fat prices are high. This produces decremsing milk price to producers with a fatto tl'ue protein ratio of less than 1.28 when fat value goes up. This causes higher volatilityin milk pxotein price than there should be and it sends a confttsitlg pdce sigrtat to mostproducers. In my opinson the autom atic decrcase in protein value with inereasing fatvalMe for a producer with a fat to protein ration of Iess than 1.28 that happens in thecurrent Class lll prke calculation needs to be eliminated and one way to do tllis is bychanging the method of Class HI price calculation.

2. The method that l have proposed to calculate Class Il1 rdies on the samt foundation of theVanslyke cheese yield formula as the current system and when using the same assumptionsas the current system, the proposed method retums amilkprice that is about $0.15/c% higherat 3.5% rat for the M arch 1999 data t14% the current system. This difference will vary frommonth to month. In my opinson, the system of Class III m ice calculation that l lmveproposedwould redacevelatilityof protein prites, itwould establish afatvaluein Class1I1 that is tied directly to the NASS cheese price, and it wfmld eliminate the Class IIImllkprlce behavlor of decreasingproteln values caused directly byincreased fat valuesthat sends a confusing price signal to producers in the current system .

3. The parameters used as default values for the NASS cheise price and moisture adjustmentshould be reevaluated. The values for NASS cheese price and moisture that are used in thecheese yield calculation to determine true proteinprice shouldreflectthe average compositionof the cheese as it is made, not a 39% moisture. This will allow a more correct makeallowance adjustment. ln my opinion, the cheese prke used in the Class lII prlce qhouldbe a price per pound of cheese at a mojsture test that more tlosely represents the actualmoistare at which the clleese wa: produced and that sam t moisture assnmption shouldbe used in the cheese yield formula for the price calculation.

4. The default assumption used in tlae current price formula for casein as a percentage of crtzdeprotein of 75% , which equates to a value of 79.76% on a tnze protein basis is too low. Inmy opinion, the btst analytical data at the present time would indicate that a morecorrect value for the assumption of casejn as a pereentage of true protein is

approximately 82.2% .

5. The default assumption for fat recovery in the cheese of 90% is 1ow in relation to averagecheese industry performance using average modern cheese making tecbnology. In myopinion,amorerepresentativeaveragevalueforlargemodern Cheddarcheesefactorieswould be 91.5% . Some factorie: have higher fat recovery in cheese than this, others

have Iower.

6. In my opinion, the default milk compodition (at which the protein and other solidsprlces per pound art cakulated in the Class lII price calculation method that I havepropese) should representa milkcomposition thatis the avtragt milkcompositionwithrespect te fat, trlze pretein and other solids eontent present in the milk supply used by

25

cheese m akers.

26

cO

9 m cp t'q-'d .N s. % w too tq o m (oD E <' * 'e to CD to txlO $ to TN try * * to 'DE * = w. xr ko 'qt m v' o tN= uo x. o x- o ruO G (wzx r) .- o- c u QE

o * m o.= .S m o74 B o * .tR 'l1 e E 'Q * to t.-w , rl to xr o:= tu s :Q e > M' >- p- sY * pv't$ >- * .-= G a tf e :> = >- 'B = o= t * = v w. a = z ,-. *-o m .i x zz fp . wB a * i- :5 2 =- Y > '-' 2 *o œ o tx ,-< .0 s o o txl: * s 'f w (o o txt (? v.1 e E = t.a tp .. ts aj m-

..z o a mY 11 G o 9 12 v. u. .* .* G IB ** :s v; o > % *> -e -m G œ O o .2) =o. 7#k m = * s jo ro a D = m : o o= o

.(: u) o = o o' o' oo q) * fz f2 o. v. a q) o o (z. to ruo u >. Y K m m o w .c .0 to (on Q >$ Y. tfA Qq:: t@ = m ko o: u = to co u s M' 'q- = ': k z; ay o .s v $n q o x- v.& kt tg o 's o . o gx3 aa != E < tD IJ o n- x-= 'i2 fo utz Y -' %- v- œ t! Ao = > sO 1.2' = )E rm (/ u. .- * (7: .0 œ o o o o E : o == = o E G 4xa 4/ o o % < tz C fa a tocg o. E = > 2 m o toY tl 1 - Y .$: o uo x 'eo o m = to w

.r * o 'd E 9 .câ a. % *4 *- m (@: = o m a o = (o + m û) œUo = a cx > 11 o o : :& = ë 'B = * *. 'S : : Y w = 5Qo tp f, m o o m .. mo ;z w = = = -Q tu E E o .= m r t.u tu q

e4 9 o .m F- ï : z m .,x $:4 < oj cu x- q o o o- 'y 'J c: c co o w >. p-TD o : * :, u 2 cy 0. ol txl 'r' 9 p- >. o v. m= 0) w. y o m - t n. o o .- x- o o o= ..z tp r.- s. UMœ ? S = t7 (: i! Y. o = = * o c7 t;> * ..' * ù *2 ..1 t U> Q o o k Ow .t$ p =z k5 - v) = o o tzl 'a o a tp E c: xr .., ot;l - 2 * g? * o tp - % uï V c *3 * * Y = 2 w œ a>> n o r e) = 2j I : o s oOo = o 1: C: ro .2? S j m .., u c 't; .r m (u u o tp œ:E O

.$t O O .ta 1y G 6 < V m *> m y c> o o t) Or qy kn .2 = = *E o (p e a o E b, o c o o u u. o w m e (s. o s g s s sm = e -* o cs x > 01, s % k u wn : *= *= * ql v- (A 43 s . s u , m r a tu a E J>:B m o s t4 (u (p > u o *t: s s œ 1- Eu t7 K o tv

.' E e e ..8 o -t: * o 'Uql o a, o s to E ' o =w .1.2 u. E E m m m v : ': xe <' >.*-0 >. = (0 x . o = a kw E E = oy * z x w. w s. e' qlql s (x. tu qo co *41 E *t r * ï e *d e .(g s m o to x. %. p. ol mm .q> o. u. to to $o S .c a .y .y om g 4tv o S '/ = tv 'e' o a tyj q) y Q *(J to qo =

.c: . o o 4 oo *(5 -'d a œ z .: * = qj -Q n- m to v- a ./ et co . .x o z z x- = =c tq v/ o :: q' E e * .= o o $ o = o v? .go = E o 41 > o *tz a = *- o / qd *d E* tn = c) 2t 2t x- v. = .. g (q $ $I .b 2! ç: G : a .= m m o .- <- v- ow *ez -< *'Q o s b o o o + .x: tz e fv (u (0 's b x z x (a (7

$5 o qj E '# = o. otp .s d ? w. * c: = a x x x o o w tn txrn c s = s : c: - a o <q g w= $7 z o (o a E m x X 1 f-l t7 w (x ua ol o o *'Q O = e o o : tn o o o w o s. xe to.> Ja *t1 o u ç; m n (! Q) w (x to txl . s. w to o ro m- .- m m tx mfe (o > n. o m. ': o v c: .u to o o * = =o to ':5 :1 k; o >- 'qr to g m v: = v' x- t'4 * 2 <' t' 0. D..e $ .f2 5 q: 'c3 .a 8 fl o to v; m x- v- c4 v) = o o. o. = o.u a a %>' 'v c cu vo = rs c o.= w- R o v. o n. o. = z: s s E g) co EJa > o o o - :z 2 rs ç: * .r s a. ra E s= o gz m y o o. p a o o. 2 .R r/l toqy ro : 6a ïp .'a tx m: ,s rp ço y E co a lo

q2 o. k :7 te D k- O o: = E s to o *> >' = wa *> * m G x ë ttà w. w. >.n - s o. o (x. o a o o o. qpS Y .2 E mo G z' li i! Y o * *' N'ov No, .$î .; tl = 4: -q w ps a. s, * m(@ r :f : J) G (v ta e n o o 1. p. w v =.(: * u 0. 2 trz n o. o Q .o o a s.: E o. e 8. u s = %- s.a r e z s cx w ar o .$2 G * P O .j = .- 2 s.T vs to .- .s o z ., Y o > kr;

% '- E b (o s e * ZH = = -: =f; m = O m Y R 2 O v.= s m .'z * 1J r = = s a.2 *4 X .R .s % * q: * çF 62 a S )>v (p z v * E2, *gt ft -Q * E: 82 S; -c5 0. * nr n (; r g. x o. e - s F-$- !- u î-

Table 2. Sensltlksty Analysls - Vanslyke Yleld Equabon ParametersUslng March 1999 Prlce Data and Current Defauit Assumptlons

% Fat Class I11Recovery Prlce ($/cwt) atin Cheese 3 5% fat

88.0 11 611089.0 1 1 .635090.0 1 1.659091 .0 11 683092.0 1 1 -707093.0 J 1.731094-0 1 1 7550

SohdsRecoveryFador

-! 07 1 1 .46551.08 1 1 -56251 09 11 6591t .10 J J .75571 .1 1 1 1 8523

MoistureContent. %

35 0 1 1 $ 57535,5 1 1 237936 0 1 1 31 8336.5 1 1 .390J37.0 1 1 479137.5 1 1 559538 () 1 1 639938.5 1 1 720339 () 1 1 8007

Casem as a Caseln as apercent of percent of

Crude Proteln True Proteln75.00 79.76 11 659175.59 80.3 1 1 .689976.00 80 83 1 1 .720776.59 81 36 1 1.751577.09 81.89 1 1.782377.59 82 42 1 1 813178.00 82 95 1 1-843978.50 8349 11.874779.09 84.02 1 1 .9055

Table 3 Senslbvlty Analysls - Make AllowanceUsmg March 1999 Pfice Data and Current Default Asstlmpbons

Class 1llMake Allowance Price ($/cwt) atCheese, &/lb 3.5% fat

0-10 12.33000 -1 1 12.23440.12 12.13880.13 12-04320-14 1 1.94760.15 J 1.85200-16 11-75640.17 1 1.66080-18 1 1 .56520.19 1 1.46960-20 11 3740

Make AllowanceWhey, $/lb

0.10 11.89120.1 f 1 1.82850.12 1 1 .76580.13 1 1.70310.14 1 1.64040.15 1,1.5/770.16 1 1 .51500.17 11.45230.18 1 1.3896

Table 4 Sensdivlty Analysls - NASS Cheese and Whey PricesUsing March 1999 Prlce Data and Current Default Assumptkons

Class llICheese Pnce ($/cwt) atPrice, $/1b 3 5% fat

1.00 8.79291.10 9 78361.20 10.77431 -3Q 11 .76591.40 12.75571.50 1374641-60 14.73711 .70 15.72781-80 16.71851.90 1/.70922.(0 IB 6999

WheyPrlce, $/1b

0.15 11 .5669Q.16 11.62960.17 11 69230.18 11 75500 19 11 81770.20 1 1 .88040.21 1 1 94310 22 12 00580 23 12 06:5() 24 12 13120-25 12.1939

REFERENCES

1. USDA. March 1999. Federal M ilk M arketing Order Reform. New England, et a1. FinalDecision. USDA, M arketing and Regulatory Programs, Agriculmral M arketing Sezvice,Dairy Programs.

2. Barbano, D.M . 1996. M ezzarella cheese yield: factors to consider. Pages 29-38 inProceedhgs of the seminar on M aximizing Cheese Yield. Center for Dairy Research,University of W isconsin, Madistm, W 1.

3. Barbano, D.M . 1990. Seasonal and Regional Variation i.n M ilk Composition in the U.S.Pages 96 to l05 in the Proceedings of the 52nd CornellNutrition Conference. Rochester, NY.Department of Animal Sciences, Com ell Univeristy, Ithaca, NY.

4. Barbano, D.M . 1999. Trends ill M ilk composition and analysis in New York. Pages 32 to 39h4 the Proceedings of the 61St Comell Nutrition Conference. Department of Animal Sciences,Cornell University, Ithaca, NY.

5. Barbano,DmM., J.L. Clark, C.E. Dtmham, and J.R. Fleming. 1990. Kjeldahl method fordetermination of total nitrogen content of milk: collaborative smdy. J. Assoc. Off. Anal.(E1e1r1.73:849-859.

6. Barbano, D.M ., J.M . Lynch, and J.R. Fleming. 1991. Direct and indirect determination oftnze protein content of milk by Kjeldahl analysis: Collaborative sttzdy. J. Assoc. OfE M al.Chem. 74:281-288.

7. Lynch, J.M ., D.M . Barbano, and J.R. Fleming. 1998. lndlrect and direct determination of the

casein content of milk by Kjeldahl nitrogen analysis: collaborative study. J. Assoc. Off. Anal.Chem. lnt. 81 :763-774.

25