Staff Paper P73-34 December 1973 OOCIAL COST OF THE DAIRY PRICE SUPPORT PROGRAM

Transcription

1 " Staff Paper P73-34 Deceber 1973 OOCIAL COST OF THE DAIRY PRICE SUPPORT PROGRAM Boyd M. Buxton and Jeroe W. Haond* ANDREW M. NOVAKOVIC AGRICUl rural ECONOMICS *Staff papers are published without foral review within the Departent of Agricultural and Applied Econoics.

2 SOCIAL COST OF THE DAIRY PRICE SUPPORT PROGRAM* by Boyd M. Buxton and Jeroe W. Haond** Decisions to change the level of ilk price support or even to continue the progra should include soe consideration of the associated cost to society. This paper describes a ethod to estiate the net social cost of the price support progra and estiates the change in net social cost resulting fro changing the level of support price under the existing progra. Net social cost obviously is not the sole criterion for policy decisions but it should be considered along with ~ore conventional criteria such as governent progra cost, far incoe, ilk supply and utilization and consuer prices. Estiating the social cost of the dairy price support progra is ore coplex than it would be for any other agricultural products because of the ultiplicity of price regulations. There are three basic prices deterining supply and utilization of ilk: the "anufacturing" ilk price paid by processors of anufactured ilk products; the "fluid bottling" price paid by processors of fluid ilk products (usually priced at a fixed differential above the anufacturing use price); and the weighted "all wholesale ilk" price received by producers that reflects the proportion of all ilk used for both fluid and anufactured products. Although 75 percent of all ilk arketed in the United States is Grade A or eligible for fluid use, only 44 percent of all ilk is actually used *Paper presented at the Contributed Papers Section on Econoic Policy of the Aerican Agricultural Econoic Association Annual Meeting at Edonton, Alberta, Canada, August 9, **Boyd r~. Buxton, Dairy Progra Leader, Coodity Econoics Division, ERS, Stationed at the University of Minnesota; and Jeroe W. Haond, Professor, Departent of Agricultural and Applied Econoics, University of Minnesota, respectively.

3 for fluid ilk products (see figure 1). The reaining Grade A ilk is sold on the anufacturing ilk arket. This diverted Grade A ilk along with all Grade B ilk is then the available supply to eet anufactured ilk deand. It is the supply and deand of t~is ilk for anufacturing that under the present pricing policies deterines the arket value of ilk. The price paid by processors of fluid ilk products is calculated fro the anufacturing arket price and can be closely approxiated by using a $2.17 differential above the anufacturing price (see Appendix table 1). Also, the all wholesale price received by producers can be closely approxiated by using a $1.00 differential above the anufacturing price (see Appendix table 1).11 These prices tend to ove together because they are often tied by forula to anufacturing ilk prices and they tend to aintain or hold at fixed differentials. Further analysis suggests that changes in the anufacturing ilk price will result in approxiately equal changes in the bottling ilk price and in the all wholesale ilk price received by farers (see figure 2). It is the anufacturing ilk price that the governent ust support between 75 and 90 percent of parity. Under the current price support progra, the governent stands ready to purchase butter, non-fat dry ilk powder and cheese at prices sufficient to support the anufacturing ilk price at the level established under the support progra. 1IHowever, increasing the anufactured ilk price through the support progra would raise the fluid eligible price and decrease the proportion of ilk used in fluid products, thereby resulting in a relatively saller increase in the "all wholesale ilk" price. The exact relationship between changes in the anufacturing ilk price and the all wholesale ilk price would depend, in part, on the relative elasticities of deand for fluid eligible and anufacturing ilk. The year-to-year changes in prices plotted in figure 2 suggest that a change in anufacturing price will result in about equal changes in fluid eligible and in all wholesale ilk prices. This enables us to ake soe useful siplifying assuptions about price changes which aids in estiating the social cost of the price support progra.

4 .. 3._ TOT. lk ~,1' KETED *.~ Fl.. --" ' ".. BOTTER- CHEESE D Y LK R - -- ".. -- '." _- _.-. _., *Abuut 3 percent of the ilk supp~ is used on fanns where' (iodjjced. Fig. 1. Utilization of total ilk arketed in the United States, 1970.

5 " [cents} line assues 60 4p = ~pw = l. pf = ~ P 50 -C) --..., Go) c.,) 'S:::: / CL / Go) 40 ~ B 30 '.~ ~, 3r: =iij VA 0 / 00 0 / i. -c c::: 20 "'.. ~ / :E / 10 / ~ 'S:::: CL /,,. :"S! 0-10 ~ 1- Go),. b.o ~ C-:) "-,.' v / Q ' I. / / / / 70 Change In anufacturing price [cents).6. P Fig~ 2. Rel~tionship between changes in anufacturing ilk prices and in fluid bottling and in all wholesale ilk prices, 1960 to 1972.

6 Our ethod and estiates of social costs do not include considerations of the soci.al benefits and costs of the discriinatory price systes. Rather, they look at the price support progra. That is, the federal ilk arketing order progra and other institutional and classified pricing arrangeents and their associated social costs and benefits are assued to be given. Theoretical Background Equilibriu price and quantity established in free arket, unrestricted by production controls and price supports, axiizes the net social gain, given the initial resource distribution.~ ( A net social cost is iplied whenever the support price for anufacturing ilk exceeds the free arket price. Consuers ust pay higher prices for fluid ilk and anufactured dairy products and producers are encouraged to produce ore ilk than is deanded. The nature of this social cost of supporting ilk prices ay be best understood by considering separately and also jointly the consuers and producers of ilk. First consider the consuer. A consuer's preference ay be represented on his indifference ap. Let this ordinal easure of utility show the consuer's preferences between ilk and all other goods (nueraire). The consuer's utility will be at a axiu where the highest indifference Curve is tangent to his budget constraint. Here the arginal rate of substitution between ilk and all other goods is just equal to slope of the budget constraint. The slope of the budget constraint will equal lifor a further discussion of this see: Luther G. Tweeten and Fred H. Tyner, "The Utility Concept of Net Social Cost--A Criterion for Public Policy," AER, Econ. Res. Ser., USDA, Vol. XVIII, No.2, April 1966.

7 6 the' price of ilk since the price of the nueraire is set equal to unity. Assuing that all people have identical preferences and initial endowents, a counity indifference ap ay be constructed, changing only the scale on the two axes. Counity utility will be axiu when the highest indifference curve between ilk and all other goods is just tangent to the counity budget constraint, point A in figure 3. Also, at this point the arginal rate of substitution between ilk and all other goods for the counity as a whole is just equal to the price of ilk. A copensated deand curve for ilk can be derived for the counity as a whole by 'changing the price of ilk, copensating the counity incoe so that the sae level of satisfaction is aintained, and by observing the change in the quantity of ilk deanded, e.g., point C in figure 3. The curve will be downward sloping assuing ordinary convex indifference curves. Now consider the producer. With a fixed aount of resources available to the counity, the quantity of all other goods and services will decrease as ore resources are used for ilk production. The various quantities of ilk and all other goods that can be produced in the counity ay be represented by a production possibility curve. Here the optial cobination of ilk and all other goods is attained when the arginal rate of product transforation is equal to the price of ilk. Assuing a curved production possibility Curve, the counity will have to give up ore and ore of all other goods to obtain larger outputs of ilk. The supply curve derived fro this possibility Curve will have a positive slope since the cost to the counity in ters of all other,goods increases as larger quantities of ilk are produced. The supply curve then reflects the opportunity cost to the counity,. in ters of all other goods, for additional quantities of ilk production. The counity indifference ap and production possibility curve are cobined in figure 3.

8 7 All OTHER GOODS ($) PRODUL11ON POSSIBIUTY WM \.... Y f II---~----"'k:-- o xo x [ x, \1.----"1 ~---'/ v. (govene purchases) ~ price ~atio e -.; slope = PE. new price ratil ine \, --r slope = pl MLK Fig. 3. Production possibility and indifference ap for a hypothetical counity_

9 8 At the free arket price the production possibility curve is just tangent to indifference curve II in figure 3. The slope of both the indifference curve II' and the production possibility curve is equal to the price of ilk (slope of price line) at point A. the free arket equilibriu price. Now consider a support price higher than Dairy producers, attepting to axiize total revenue with available resources, will want to produce on the production possibility curve where the slope of the new price line is just tangent to the production possibility curve. This would be point B in figure 3 and for the entire counity OX I ilk, and OY lof all other goods would be produced. The consuers on the other hand would want to consue less ilk at the relatively higher ilk price to iniize the cost of aintaining the sae level of sat isfaction.,;y This. is equivalent to consuing, where the slope of the new price line is just tangent to the sae indifference curve, point C in figure 3. The entire counity would want to consue OX o ilk and OYO for all other goods at the higher ilk price. The difference between what the farers produce and what the consuers consue at the higher price (XOX l ) is the excess supply purchased by the governent. Only axo ilk and ay l of all other goods are now available to the counity, aking 12 the highest indifference curve obtainable if the governent donates abroad or "destroys" the ilk purchased under the support progra. The corresponding deand and supply curves for fluid and anufacturing ilk are presented in figure 4. This figure is a odified version of the ore s conventional two price plan. The fluid deand curve labeled D f, and the supply curve labeled Ss show the aount of ilk deanded as fluid and the total ilk,;ythe change in quantity qf ilk de:nanded so that total satisfaction is unchanged in the counity due to a higher ilk price is equivalent to a copensated deand curve. The copensated deand Curve is approxiately equal to the own deand curve when the incoe elasticity of deand for ilk is sall, i.e., the proportion of total expenditure for ilk is sall.

10 9 supplied, respectively,.t each anufacturing ilk price. The actual fluid ilk price corresponding to any anufacturing price for any ~antity can be s. deterined by reading the value directly above D f on the Df curve which is the actual deand curve for fluid ilk~ This vertical distance between D S f and D f is the differential between anufacturing ilk and fluid eligible ilk prices ($2.17). The all wholesale ilk price nece ssary to call forth alternative quantities fro farers and corresponding to any anufacturing price is the price directly above SS on the S curve which is the actual supply curve for all ilk. This vertical distance between the SS and S curves is the differential between anu~ facturing ilk and all wholesale ilk prices ($1.00). \ The area under the fluid deand curve between 0 and X f and under the anufacturing deand curve between 0 and ~ represents the total value of fluid ilk and anufacturing ilk to the counity in ters of all other goods.~ The area under the supply curve between 0 and Xs represents the cost to society of the resources used to produce ilk in ters of all other goods. Assuing the governent donates abroad or destroys its, purchases, the. social cost of higher support prices would be approxiated by the three shaded areas in figure 4. The shaded area under the deand curve for fluid ilk and the deand for anufacturing ilk would be the loss of consuer surplus while the shaded area under the supply Curve would be the cost of the added resources used to produce the additional ilk at the higher all wholesale ilk price. This area or loss would be approxiately equal to: ~ Assues that both deand curves intersect the vertical price axis.

11 PRICE PER 100 POUJ~OS ~ l~j.j IJY<r- Y ~ s. G ~., ~r \~? :..rr. ~/~ ~~J.. r:; Il> D S S. Df SS. S- ~ vtt~- GJG-.. G t I> "'...; ~ ~~ ~(\ ~Pfn \ \ U~ (\ I \. ). I \ Of I. ~~ { /?AK'"( ~ AP W [I. (. P p.{ I I I I I I I. o i Xf. X Xs POUNDS OF MILK y governent purchases. Fig. 4. Ipact of increasing support price for anufacturing ilk on net '::'

12 11 where: II qf = change in fluid consuption A~ = change in anufactured ilk consuption Aq s = change in quantity supplied Pf P P w = fluid eligible ilk,price = anufacturing ilk price = all wholesale ilk price. This social cost can be approxi.ted in ters of elasticity of deand for fluid and anufacturing ilk; elasticity of supply, the initial prices and quantities and the change in ilk price as follows: First, approxiate the fluid deand (qf), anufacturing deand (fl)' and supply (c) elasticities with I'\.f = f Aq P f flpf%, fl= d1~ A.P~ s (;; Ilq Pw llp'w~ where: qf o = equilibriu or initial quantity of ilk sold for fluid use ~ = equilibriu or initial quantity of ilk sold for o anufacturing use.

13 12 Q'; -0 P f - equilibri.u or initial quantity of ilk supplied ; equilibriu or initial fluid eligible ilk price per one hundred pounds P P w = equilibriu or initial anufacturing ilk price = equilibriu or initial all wholesale price of ilk received by farers. Assuing constant price differentials the following also hold: = where: a = fluid eligible price differential above the anufacturing price b = all wholesale ilk price differential above the anufacturing price, and therefore, price. ~ P = A P f = ~ P w = ~ P = the increase in anufacturing support The social cost can then be expressed as: The ore elastic the supply and deand curves the greater the social cost of a support price increase. Also the greater the

14 13 increase in support price (or the aount the support price exceeds the free arket price) the greater the social cost. In fact the social 'cost increases at an increasing rate for successively higher support prices above free arket equilibriu prices. Reducing the Dead Weight Loss In reality, uch of the governent purchases is distributed back to the counitythiough various progras such as the school lunch progra and other doestic donations. Assuing all the governent purchases were equally distributed back to the U.S. counity as, anufactured products, the counity would be able to achieve indi fference curve 13 through point B in figure 3. However, 13 is a lower level of satisfaction than is 1 1, which would be attainable under free arket conditions. Any progra distributing the governent purchase of ilk back to the counity would reduce the dead weight social cost and allow the counity to attain soe higher indifference curve between 12 and 13 in figure 3. A gift of part of the governent surplus to foreign countries would be equivalent to destroying the surplus as far as the United States counity is concerned, even though world welfare w0uld clearly be iproved. The 60cial cost would be reduced by the area under the anufacturing curve labeled "a, b, c, d," in figure 4 if governent purchases were distributed back to the U.S. counity as anufactured products.

15 14 The net social cost then would be approxiated as follows: where: P' = iplied price necessar~ for consuers to purchase as r- anufacturing products the anufacturing ilk diverted fro fluid use (A qs). (d q~) and additional ilk supplied b,p' = P' - P P' < P The net social cost could then be approxiated as follows: Net loss-- +f'\f b.p~ - I (-~), P +b J tt II P s qo P + b where: ti P' = -t. P P 11 ~ % ] [ -P--+--b pi = P + Il pi pi ( P

16 15 Epirical Estiates Two kinds of related policy questions can be considered within the fraework presented. First, the annual social cost of the progra itself can be estiated by estiating free arket prices and quantities and the aount any selected support price exceeds the free arket anufacturing ilk price. Second, the social cost of increasing the support levels over any current level can be estiated by using current prices and quantities and the intended increase in support price. These questions are part of the sae broader proble. The social costs estiated in this paper assue a deand elasticity for fluid ilk of -0.35, for anufacturing ilk of -0.5 and the elasticity of aggrega~e supply of 0.15.~ These estiates are fro arket data. The elasticity of deand would actually represent the own deand curve for ilk rather than a copensated deand Curve. The own and copensated deand curves are identical when the incoe elasticity is zero. Since total expenditure for ilk products is a relatively sall part of total expenditures, we assue incoe elasticity is sall and no great distortion will result fro using the epirically estiated own deand curve. In early March 1973, the Departent of Agriculture announced that the new support price for anufacturing ilk would be increased fro $4.93 to $5.29 the iniu 75 percent of parity as of April 1, The new support price becae effective on March 15, To estiate the social cost of this decision to raise support price to 90 percent of parity ($6.35) requires estiates of the free arket price and quantities that would otherwise exist in the absence of the progra. The anufacturing ilk price for 1973 is ~Estiates reported in an unpublished report to ASCS by Econ. Res. Ser., USDA entitled, "Ipacts of Alternative Price Support Levels," Janu~ry 1973.

17 16 expected to be about $5.50 which would exceed the announced 75 percent of parity support price of $5.29. Therefore, the social cost would be zero. If the support price had been raised to 85 percent of parity (about $6.00) the support price would have exceeded the free arket price by 50 cents. The estiated annual social cost for 1973 without distributing governent purchases back to the counity would have been an estiated $340 illion (table 1). 1 A further increase in the support level to 90 percent of parity ($6.35) would have increased the social cost in 1973 another $107 illion to a total of $447 illion. The annual social costs of the progra itself would, therefore, be those in table 1. The social cost of increasing the support level would be the difference between the totals in table 1. For i exaple, the annual social cost of the progra would have been $340 illion if the support price had been set at 85 percent of parity and $447 illion if the support price had been, set at 90 percent of parity. The increase in social cost of increasing the support price fro 85 to 90 percent of parity is the difference between $447 and $340 illion or $107 illion (see table 1). The annual social cost when the increased governent purchases due to the higher support price are distributed back to the counity as anufactured products is uch less at all support levels than without redistribution (see table 1). At 85 percent of parity the estiated annual social cost would decrease fro $340 to $65 illion if all governent purchases were distributed back to the counity rather than being donated abroad or destroyed. The social cost of a use. lsocial cost is based on 1972 production, fluid and anufacturing

18 17 Table 1. Estiated annual net social cost of alternative price support levels, 1973 Percent of parity Support price1! (dollars) Estiated free arket price (dollars) Aount the support price exceeds the free arket price (dollars) Social cost~/.65 without distribution back to the counity (illion dollars) ', with distribution back to the counity (illion dollars) -" liestiated 100 percent of parity price as of April 1, 1973 was $7.05. This actually turned out to be $7.01. ~Excludes adinistrative and storage costs of the progra.

19 . 18 cobined redistribution and foreign donation policy by the governent Would be soewhere between the above extrees depending on the proportion redistributed and donated abroad. Suary and Iplications We developed a ethod to estiate net social cost of the dairy price support progra using consuer and producer surplus concepts. The ethod is intended to be another criterion for policy decisions regarding whether there should be a progra or at what level the ~rice support should be set. It is not intended to replace other iportant criteria such as governent progra costs, far incoe levels, ilk supply and utilization, and consuer prices. Results showed that the social cost of the March 1973 decision to set the support price at the iniu 75 percent of parity was zero. Raising the support prices to 85 percent of parity would have resulted in a net scial c6st of $340 illion if increased governent purchases due to higher support price were donated abroad or destroyed. The net social cost of this sae decision could be reduced to $65 illion by redistributing the increased governent purchases back to the United States counity as anufactured products. The ethod developed here has the sae weakness as ost tools of policy evaluation and forulation. It assues static deand and supply curves for fluid and anufacturing grade ilk. That is, elasticities are constant and there are no shifts in the curves. Shifts in one or ore of these curves will lead to different social cost estiates. To the extent that shifts and changes can be forecasted they can be

20 19 incorporated into the estiating odel. Generally, it is difficult to forecast these changes. Had we ade the 1973 social cost estiate without anticipating the shifts in deand and supply that raised anufacturing ilk prices above the support level in late 1972 and early 1973, estiates would have been uch higher because the free arket price probably would have been underestiated.

21 (, :-- Appendix Table 1. Selected historical price and quantity data. U.S.Y Prices $1 cwt. Manufac. Milk Fluid All Milidi All Milk suppor~milky Used For Grade Bottling 11 Diff. Wholesale Diff. Price Prod. P P f. (PeP) P w (pw-p ) Billion Pounds FarY Milk~ Used in~ Used in~ Use Marketed Fluid Manufac. Gov'tJi Products Products Purchase i LOI I\) 0

22 . " Appendix Table 1. (continued) Selected historical price and quantity data. U.S;1I Prices ~l cwt. Billion Pounds Manufac. Milk Fluid All MilkY All Milk support21 MilkY FarrrY MilkY Used in3/ Used in~ Grade Used FOil. &ling Diff. Wholesale Diff Price Prod. Use Marketed Fluid Manufac. Gov1tY P P f (PrP).. Pw (pw-p ) Products Products Purchase Y March 1973 Dairy Situation. V April 1973 Fluid Milk and Crea (sales of fluid ilk and crea-ilk equiv.) p. 24. Y April "Milk-Production, Disposition and Incoe ," p. 4 (ilk utilized includes ilk sold to plants and dealers as whole ilk and as far separated crea and ilk sold directly to consuers). 51 Residual - ilk utilized inus governent purchase inus used in fluid products. 21 For arketing years source May 1973 Dairy Situation.

23 .. 22 REFERENCES Currie, J.M., Murphy, J.A. and Schitz, A., "The Concept of Econoic Surplus and Its Use in Econoic Analysis," The Econoic Journal, Deceber' 1971, pp Johnson, Paul R., "The Social Cost of the Tobacco Progra," Journal of Far Econoics, I~ay 1965, pp Mishan, E. J., "A Survey of Welfare Econoics, ," The Econoic J?urnal, June 1960, pp Nerlove~ Marc~, The Dynaics of Supply-Estiation of Farer's Responses to Prices, The John Hopkins Press, Baltiore, Tweeten, Luther G., and Tyner, Fred H., The Utility Concept of Net Social Cost--A Criterion for Public Policy, AER, ERS, USDA, Vol. XVIII, No. 2, APril 1966~ Wallace, I.D., "Measures of Social Costs of Agricultural Progras," Journal of Far Econoics, May 1962, pp