The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
The R e v i e w is published 10 times per year by the Research Department of the Federal Reserve Bank of St. Louis. Single-copy subscriptions are available to the public free of charge. Mail requests for subscriptions, back issues, or address changes to: Research Department, Federal Reserve Bank of St. Louis, P.O. Box 442, St. Louis, Missouri 63166. Articles herein may be reprinted provided the source is credited. Please provide the Bank’s Re search Department with a copy of reprinted material. http://fraser.stlouisfed.org/ 2 Federal Reserve Bank of St. Louis T he Lag From Money To Prices K E IT H M . C A R L S O N ^ E c o n o m i s t s generally agree that money affects prices with a lag. Research conducted at this Bank suggests that a change in the growth rate of money is fully reflected in the inflation rate in about five years. This conclusion was based on a statistical anal ysis of the relation between money and prices in the U.S. from 1955 through the 1960s.1 The length of the lag between money and prices represents important information that must be con sidered in the policy formulation process. The policy maker must allow for such lags when developing a policy to control inflation; he must also consider pos sible future impacts of short-run policies designed to combat recession. Given the historical presence of lags between money and prices, a policy designed to control inflation will not have immediate effects. The possible short-run costs (benefits) of a restrictive (stimulative) policy in terms of employment and output must be assessed against its long-run benefits (costs) in terms of the price level. The nature of the lag enters importantly into the decision to adopt a specific policy, whether it be short- or long-run in character. The purpose of this article is to examine the rela tion between money and prices in light of the U.S. 1Denis S. Kamosky, “ The Link Between Money and Prices,” this Review (June 1976), pp. 17-23. Also see Albert E. Burger, “ Is Inflation All Due to Money?” this Review (Decem ber 1978), pp. 8-12. economic experience of the 1970s. Statistical results are summarized first and the economics of informa tion and search are then summarized to provide a theoretical rationale for the results. Statistical Results Kamosky’s original estimate of the price equation was based on the sample period from 1955 through mid-1971 and used what is now known as “old M l” for the money variable. A version of this equation, estimated by using the “new” M1B definition of money, is summarized in table l .2 Compared to the original results, using a different definition of money and modifying the sample period affects the pattern of the coefficients very little. The sum of the coeffi cients is one, as would be expected from economic theory.3 The mean lag is estimated at 10.96 quarters for the 1955-69 sample period.4 2In this article, money is defined as M1B ( currency plus check able deposits at financial institutions). See R. W . Hafer, “ The New Monetary Aggregates,” this Review (February 1980), pp. 25-32. The sample period differs slightly from Karnosky’s for puiposes of balancing degrees of freedom, so that the 1970s can be compared with the “pre-1970s” . 3For a discussion of the theory, see Leonall C. Andersen and Denis S. Kamosky, “ The Appropriate Time Frame for Con trolling Monetary Aggregates: The St. Louis Evidence,” in Controlling Monetary Aggregates II: The Implementation, Conference Series No. 9, Federal Reserve Bank o f Boston (September 1972), pp. 147-77. 4The mean lag serves as a summary measure of the speed with which prices respond to money. It is calculated as a sum of roducts ( where each product is the coefficient times the numer of the lag) divided by the sum of the coefficients. E 3 F E D E R A L R E S E R V E B A N K O F ST . L O U I S OCTOBER 1980 Table 1 Estimate of Money-Price Equation: Original Specification Sam ple period: I/5 5 -IW 6 9 : m2i = 20 0 . P = -.1 4 6 + Z m , M , (.3 9 5 ) i = 0 Coeff. |t| m0 .041 1.276 mi .034 1.538 Coeff. |t| Coeff. mg .048 3.249 m 16 .069 3.943 m9 .054 3.783 m 17 .062 3.712 4.305 m is .053 3.511 4.673 mig .039 3.338 |t| m2 .030 1.903 mio .059 m3 .029 2.171 mn .065 m4 .030 2.235 m 12 .069 4.795 m2o .022 3.191 m5 .033 2.294 ^13 .072 4.694 2m i 1.031 7.870 m6 .037 2.475 m 14 .073 4.468 Mean lag 10.959 5.634 m7 .042 2.798 mis .072 4.202 R2 .525 S.E. 1.066 D.W. 2.00 • Notation: P • com pounded annual rate of change of G NP deflator; M To examine the nature of the money-price lag in light of the experience of the 1970s, it is necessary to consider other factors that influenced the price level during this period. From August 1971 to April 1974, a government program of wage and price con trols disrupted the money-price level link. In addi tion, in late 1973 and early 1974, substantial increases in energy prices occurred. At various times during the 1970s, agricultural conditions also appeared to affect movements in the price level or, more properly, in the indexes that are used to measure changes in aggregate prices. Because of these factors, the basic price equation in this article has been respecified to include prices of food and energy relative to overall prices and dummy variables to capture nonmonetary effects of wage and price controls. Table 2 summarizes the re sults for the 1970-79 period (center columns) and, for comparison purposes, also summarizes the results of this same specification for the 1955-69 period (lefthand column). As implied in Karnosky’s specification, food and energy prices did not play a statistically http://fraser.stlouisfed.org/ 4 Federal Reserve Bank of St. Louis com pounded annual rate of change of M1B. significant role in explaining inflation during the 195569 period.® The results for 1970-79 indicate a number of changes relative to those for 1955-69. The price con trol dummy is significant with a negative sign, and the post-control dummy has the expected (positive) sign but is not significant. The sum effect of energy prices, as measured by the producer price of fuels and related products and power, is positive and signifi cant. The food price variable has the expected sign and is just short of being significant. More impor tantly, however, the pattern of coefficients on money is substantially different from that estimated for 195569. No coefficients are significant after the eighth lag, and the mean lag is 5.05 quarters. The sum of the coefficients, although close to one in value, is not significantly different from zero. These results suggest that the 20-quarter lag struc ture is no longer appropriate for data from the 1970s. 5Throughout this article, “ statistical significance” refers to a two-tailed test conducted at the 5 percent level. For large samples, the critical “t” is ±1 .9 6 . F E D E R A L R E S E R V E BAN K O F ST. L O U IS OCTOBER 1980 Table 2 Estimates of Money-Price Equation : Modified Specification 6 . . k • • • P = const. + Z mi M .i + d i D i + d2 D2 + Z e, (P b - P ) - i + f (P f - P ) i = 0 I= 0 1 /5 5 -IV /69: m2i = 0 I/7 0 -IV /7 9 : m2j = 0 I/7 0 -IV /7 9 : m 13 = Coeff. Coeff |t| Coeff. |t| .440 1.454 |t| 0 m0 .048 1.215 .035 .680 .038 mi .041 1.484 .070 1.881 .065 ITI2 .036 1.875 .096 3.168 .087 2.800 m3 .034 2.217 .115 3.821 .104 2.963 m4 .034 2.307 .125 3.777 .116 3.005 m5 m6 .036 2.310 .129 3.458 .123 3.222 .039 2.407 .128 3.789 .124 3.443 m7 .044 2.641 .121 2.692 .121 3.324 m8 .049 3.007 .111 2.308 .113 2.804 mg .054 3.470 .097 1.926 .100 2.213 mio .059 3.946 .081 1.549 .082 1.744 mu m i2 .063 4.293 .064 1.182 .056 1.404 .067 4.396 .046 .833 .032 1.159 ™13 nrii4 .069 4.268 .028 .508 — — .070 4.014 .011 .210 — — m is .069 3.728 -.0 0 3 .058 — — rtiie .065 3.459 -.0 1 5 .296 — — m 17 .059 3.224 -.0 2 3 .506 — — m 18 mi9 .050 3.026 -.0 2 6 .690 — — .037 2.858 -.0 2 4 .850 — — m2o Zm i .020 2.717 -.0 1 6 .990 — — 1.044 7.457 1.150 1.631 1.164 3.297 10.542 5.201 5.047 .756 5.908 3.279 e0 .002 .089 .001 .076 .003 .314 ei .004 .216 .013 1.652 .014 1.755 e2 e3 e4 .007 .010 .370 .018 2.097 .017 2.089 2.465 .018 .574 .017 2.602 .013 .706 .013 2.233 .013 2.892 e5 .013 .674 .008 1.123 .007 1.386 e6 Zei .009 .595 .002 .407 .002 .444 .058 .580 .073 2.086 .075 2.586 -.1 0 9 .281 -.6 8 8 .174 -.7 7 0 .356 dx — — -1 .7 2 4 3.010 -1 .7 3 5 2.801 d2 — — 1.619 1.134 1.772 1.168 -.0 3 2 .536 .131 1.941 .129 1.969 Mean lag Const. f R2 .495 • Notation: P = = .728 1.264 1.294 1.94 D.W. Di = .741 1.099 S.E. 2.27 • com pounded annual rate of change of GNP deflator; M = price control dummy ( III /7 1 - I/ 7 4 = 1 , 0 elsew h ere); D2 = 2.18 com pounded annual rate of change of M1B; decontrol dummy ( II/7 4 - IV /7 4 = 1 , 0 elsew h ere); PE com pounded annual rate of change of producer price index of fuels and related products and power; and PF = com pounded annual rate of change of food deflator. 5 F E D E R A L R E S E R V E BAN K O F ST. L O U IS OCTOBER 1980 C hart 1 C um ulative Effect on Rate of Price Increase of a Perm anent Increase in the Rate of G row th of M oney Percent of Total Effect Perceat Percent 100 100 \ \ \ 75 75 1 7 -7 90 9 50 50 ^ 1 5 -6 95 9 25 25 ................... i i i i i 10 Q enrters The results of shortening the lag structure to 12 quar ters (with the thirteenth constrained to equal zero) are shown in the right-hand columns of table 2. With this specification, the effect of money on prices equals slightly more than one after 12 quarters. The mean lag for the specification is 5.91 quarters, which is significantly different from the 10.54-quarter mean lag obtained for the 1955-69 period. Chart 1 portrays the results from the left- and right-hand columns of table 2 and indicates that prices apparently responded more rapidly to changes in money during the 1970s than previously. Why did this happen and what does it imply in terms of for mulating a policy to combat inflation? Theoretical Rationale for Lags Questions about the lag between prices and money can be analyzed within the framework of information and search theory.6 To facilitate an understanding of lags and of the reasons they change, this article de velops a theory of lag determination.7 6Most of the literature on information and search theory is in the context of labor markets. For a survey, see Steven A. Lippman and John J. McCall, “ The Economics of Job Search: A Survey,” Economic Inquiry (June and September 1976), pp. 155-89, 347-68. For a discussion of the moneyprice lag within the context of rational expectations, see Bennett T. McCallum, “ Price Level Adjustments and the Rational Expectations Approach to Macroeconomic Stabiliza- i i i i 15 i i i 20 Consider a typical firm that is a price-setter in an economic environment in which information regarding equilibrium prices and quantities is costly to ob tain on both sides of the market.8 Firms do not pos sess full information about the prices or the quality of their competitors’ products. Similarly, customers do not possess full information about the prices that all sellers are charging. Firms must determine whether a change in demand for their products is caused by a switch in consumer preferences or by a general shift in aggregate demand. Moreover, they have to decide whether such a shift is permanent or temporary. For purposes of illustration, assume that the typical firm obtains information about the demand for its product by observing its sales at the current “posted” price. Given the fact that the firm possesses accumu lated information on quantities sold at a given price and assuming that the firm knows its own cost struc ture, it will eventually learn which price is optimal for its operations. tion Policy,” Journal of M oney, Credit and Banking (N ovem ber 1978), pp. 418-36. 7Although the economics of information and search is not neces sarily a theory of lag determination, this is a common implica tion of the analysis, as it is usually applied. 8The classic article which develops this point is Kenneth J. Arrow, “ Toward a Theory of Price Adjustment,” in Moses Abramovitz, ed., The Allocation of Economic Resources (Stan ford: Stanford University Press, 1959), pp. 41-51. F E D E R A L R E S E R V E BAN K O F ST. L O U IS Figure 1 E cono m ics of the Firm F a cin g U n ce rtain Dem and OCTOBER 1980 riod, the firm realizes sales of Q, at price P0; will it change its price? If demand shifts so that Qi is the mean of the new distribution, the profit-maximizing price would be Pi (and the distribution as drawn with respect to Pi will be slightly to the right of Q0).9 However, the firm will not change its price to Pi unless its subjective assessment of the distribution has shifted accordingly; that is, the firm will change its price to Pi only if the solid line shifts to coincide with the Qi distribution (drawn with respect to P0). In the absence of other information, it is reasonable to assume that the firm’s subjective distribution will shift only slightly with a single observation, depending on past experience. Continued sales around Qi for a number of periods, however, would eventually shift the subjective distribution so that it would be cen tered over Qj. Furthermore, the speed with which the firm will move to Pi depends on the nature of the distribution around Q,. If sales are distributed narrowly around Qi, the firm will have greater confi dence in the new distribution than if sales are dis tributed broadly. Over an extended period of time, the magnitude of price response will be the same but the speed of response will vary. firm. If costs and demand are known perfectly, P0 and Q0 represent the profit-maximizing price and quantity. More realistically, perhaps, the demand curve can be viewed as the average rate of sales for given prices based on experience, with some “normally expected” variation around this average. For the sake of exposi tion, this demand curve is shown in the top half of fig ure 1 as a band rather than a line, with the additional assumption being made that quantity sold at a par ticular price is distributed normally about the mean. The bottom half of figure 1 summarizes the nature of this demand curve in terms of a probability distri bution. The solid line in the bottom half of figure 1 is a subjectively determined distribution that is based on sales experience, as well as other informational factors, when the price of the product is equal to P0. Each additional observation of quantity sold at P0 will affect the firm’s assessment of the nature of the distribution it faces. Suppose that, in a particular pe Even with a new subjective distribution, the firm will not immediately change its price. The fact that the process of adjusting price is costly will influence the firm’s decision to change price. Changing price tags, making up new price lists, notifying salesmen, and/or reprogramming computers all involve costs. In addition, because firms do not know precisely what their competitors will do, a premature decision to in crease price could result in a loss of customers. There is also a possible loss of customer goodwill if a firm changes price frequently, thereby shifting additional search costs to consumers. The change in sales must be both sufficiently large and perceived as relatively permanent before the firm will adjust its price. The Money-Price Lag in an Aggregate Context In a growing economy, firms will experience in creasing sales over time and/or the number of firms will change. However, expansion of quantities sold need not imply rising prices. Prices will rise only if aggregate demand is shifting outward more rapidly 9Note that nominal resource costs are assumed to be unchanged. In a general inflation, resource costs will also rise. The focus here is on the firm’s response to a shift in aggregate demand. Recognition of such a shift before resource costs increase represents exploitable profit opportunities for the firm. 7 F E D E R A L R E S E R V E BAN K O F ST. L O U IS than aggregate supply (which can be interpreted as a “suitable” aggregation of individual firms’ marginal cost curves). If the position of the demand curve is dominated by movements in the stock of money, firms’ assessments of demand will depend on their expectations of monetary growth.1 0 This reinterpretation of the money-price lag in an aggregate context can be illustrated in terms of figure 1. Q0 represents an average level of sales for a given planning period and is associated with a given rate of monetary growth. The price will equal P0 if this expected monetary growth is realized.1 However, if 1 the rate of monetary growth is higher than expected, sales averaging Qi ( at price P0) could be consistently realized. Firms will have to determine whether this change in monetary growth is permanent or tempo rary. Ultimately, firms must decide whether a price change is necessary. As explained earlier, a change in monetary growth will not necessarily lead to an immediate pricing response by firms. Profit-maximiz ing considerations will still determine whether the decision to change price should be made immediately or postponed until further information is obtained. Reinterpretation of the analysis demonstrates how additional information influences the pricing process. Firms derive information about the state of demand by observing their sales. In an aggregate context, however, some connection between monetary growth and firms’ sales will also apply. For example, if firms have observed a close relation between sales and monetary growth, their subjective distributions might shift significantly in anticipation of a change in mone tary growth. In other words, firms’ prices might be changed in anticipation of an increase in sales.1 In 2 formation about changes in monetary growth will reduce the lagged impact of money on prices. The whole process of determining price changes involves both sides of the market. If firms’ customers have similar perceptions about monetary growth, they will expect prices to change, and firms’ loss of customer goodwill, as a result of changing prices sooner, will be reduced. 10See Leonall C. Andersen and Jerry L. Jordan, “ Monetary and Fiscal Actions: A Test of Their Relative Importance in Economic Stabilization,” this Review (November 1968), pp. 11-24. 11Different rates of expected monetary growth will, of course, imply different P0 but costs will also be different so that s, Qo need not differ. The emphasis here, however, is on the decision to change price. 12See Charles Pigott, “ Expectations, Money, and the Forecast ing of Inflation,” Federal Reserve Bank of San Francisco Economic Review (Spring 1980), pp. 30-49. http://fraser.stlouisfed.org/ 8 Federal Reserve Bank of St. Louis OCTOBER 1980 A Suggested Explanation of the Statistical Results The analysis suggests that when sales deviate from expected levels, price changes will eventually result. The length of the time interval between sales devia tions and price changes will depend on the firm’s per ception that demand has shifted. This perception is conditioned by (1 ) the past history of inflation and monetary growth and (2 ) the distribution of re cently observed deviations. First, as shown by the results summarized in table 2, firms and their customers have developed a greater sensitivity to inflation and monetary growth during the 1970s. During the 1955-69 period, the response of prices to money was quite slow because the early part of the period was dominated by relatively slow inflation. As a result, during the latter part of the period, there was a tendency to consider higher rates of inflation and monetary growth as temporary.1 3 During the 1970s, however, economic units began placing more emphasis on recent experience when forming their expectations; they learned from their past errors. Second, during the 1970s, the nature of the distribu tion of deviations of money growth from what was expected ( and, consequently, deviations of sales from what was expected) appears to have changed con siderably. Some summary statistics are presented in table 3. Quarter-to-quarter rates of change are ex amined for 20- and 12-quarter periods during the full 1955-79 period. These measures are examined to de termine if the pattern of monetary growth has changed from the pattern observed for 1955-69. The summary statistics that appear at the bottom of table 3 indicate that the results are mixed. The mean standard deviation, for example, changed little between the 1955-69 and the 1970-79 periods. How ever, the standard deviation of the means dropped sharply in the latter period, suggesting that the vari ation in monetary growth became more regular in the 1970s. This movement toward regularity suggests — although it does not necessarily imply — greater predictability. Nonetheless, tentative evidence appears to support the notion that monetary growth became more predictable in the 1970s.1 4 13This has been labeled the “ retum-to-normality” hypothesis. For discussion and statistical results, see David H. Resler, “ The Formation of Inflation Expectations,” this Review (April 1980), pp. 2-12. 14The pattern of mohetary growth would have to be examined more thoroughly, and probably in conjunction with a formal F E D E R A L R E S E R V E BAN K O F ST. L O U IS OCTOBER 1980 Table 3 Summary of Monetary Growth: 1955-1979 12-quarter periods 20-quarter periods Mean Standard deviation Mean Standard deviation 3.14 1.88 2.23 1.54 IV /5 6 2.56 1.95 1.88 1.50 IV /5 7 1.54 1.66 1.31 1.78 I V / 58 1.74 1.97 1.33 2.09 IV /5 9 1.85 2.36 1.74 2.77 IV /6 0 1.34 2.41 1.68 3.01 IV /61 1.71 2.47 2.06 2.59 I V / 62 1.92 2.47 1.34 2.20 I V / 63 2.38 2.24 2.65 1.45 IV /6 4 2.42 2.34 3.43 1.65 End of period IV /5 5 I V / 65 3.31 1.81 4.11 1.55 IV /6 6 3.49 2.17 3.86 2.47 IV /6 7 4.28 2.38 4.55 2.87 5.65 3.06 IV /6 8 5.01 2.66 IV /6 9 4.95 2.79 5.48 2.77 IV /7 0 5.15 2.71 5.42 2.15 IV /7 1 5.59 2.51 5.38 2.52 IV /7 2 6.25 2.38 6.37 2.45 IV /7 3 6.05 2.45 6.92 2.15 IV /7 4 5.94 2.24 6.07 2.31 IV /7 5 6.12 2.24 5.50 2.18 IV /7 6 5.85 2.12 5.18 1.71 IV /7 7 6.09 2.14 6.25 2.05 I V / 78 6.31 2.09 7.11 1.92 IV /7 9 6.89 2.20 7.97 1.80 M ean (19 5 5 -6 9) 2.78 2.24 2.89 2.22 Mean (19 7 0 -7 9) 6.02 2.31 6.22 2.12 Standard deviation (19 5 5 -6 9) 1.21 1.51 Standard deviation (19 7 0 -7 9) .46 .90 prediction model, before more definitive conclusions could be developed. Expectations formation is a complex process and the modeling o f this process is probably still in its in fancy. More refined explanations of the shortening of the lag await further research. See, however, Robert J. Barro, “ Unanticipated Money, Output, and the Price Level in the United States,” Journal o f Political Economy (August 1978), pp. 549-80. 9 F E D E R A L R E S E R V E BANK O F ST. L O U IS Implications of the Analysis The lag in the effect of money on prices appears to have shortened during the 1970s, but the reasons for this contraction are unclear. One interpretation is that recent experience now receives more weight in the formation of expectations. Such a situation would be beneficial for the policymaker, because it indicates that there is less inertia to be overcome in designing a policy to reduce inflation. On the other hand, a policy of short-run economic stimulus can give rise to a burst of inflation expectations, with little realized positive effect on output. A second interpretation of the shortened lag be Digitized for 10 FRASER OCTOBER 1980 tween money and prices is that it occurred because of the pattern of monetary growth. Although conclusions about the nature of the distribution of realized mone tary growth are not definite, this interpretation implies that a steady reduction in monetary growth will re sult in less output loss than an erratic reduction. If both expected and actual monetary growth can be reduced simultaneously, the effect on output need not be severe or prolonged.15 15Past relationships based on an environment of uncertainty and continuing deviation of expected and realized monetary growth are misleading in assessing the costs of reducing inflation. See Laurence H. Meyer and Robert H. Rasche, “ On the Costs and Renefits of Anti-Inflation Policies,” this Review (February 1980), pp. 3-14. Our “Shrinking” Farmland: Mirage or Potential Crisis? C L IF T O N B . L U T T R E L L K a CH year, more American farmland is being con verted to nonfarm uses such as highways, houses, air ports, and shopping centers. This development has engendered fear that the decline in farmland will eventually produce a severe crisis for U.S. food production. A recent study, in which 11 U.S. government agen cies participated, stated: “Every day in the United States, four square miles of our nation’s prime farm lands are shifted to uses other than agriculture. The thief is urban sprawl. . . . Ten years from now, Americans could be as concerned over the loss of the nation’s prime and important farm lands as they are today over shortages of oil and gasoline.” 1 Leading proponents of the shrinking farmland thesis contend that decisions to convert agricultural land to 1W here Have the Farmlands Gone? (Washington, D .C.: Na tional Agricultural Land Studies, September 1979), pp. 1-2. Similar views were expressed in The Farm and The City (The American Assembly, Columbia University, April 10-13, 1980), and in Erick P. Eckholm, Losing Ground: Environmental Stress and W orld Food Prospects (N ew York: W . W . Norton and Company, Inc., 1976), pp. 183-86. nonagricultural uses should be transferred from the private to the public sector. Michael Brewer states: “Each choice [by individual farmers to sell farmland to developers] may be sensible in its own context. But, collectively, they reduce the country’s capacity to produce food, fiber, and wood.” He argues: “The first step is to ‘find out’ . . . what tools are available to local, state and Federal governments to deal with it.”2 Lester Brown concludes: “ . . . it [cropland] can be protected from competing nonfarm demands only through land use planning.”3 In contrast to these views, this article asserts that the arguments for social planning of land use are erroneous. First, there is no evidence that the quan tity of cropland is shrinking or that shortages of food are imminent. Furthermore, even if the alleged prob lem did exist, there is no evidence that it could be solved more efficiently by social planning than by market participants. 2W here Have the Farmlands Gone?, p. 6. 3lbid., p. 14. 11 F E D E R A L R E S E R V E BAN K O F ST. L O U IS Some Arguments for Social Land Use Planning A num ber o f individuals and groups have expressed con cern about the quantity o f prim e farm land that is bein g diverted from agricultural uses. Secretary o f A griculture B ob Bergland stated: “ Failure to protect our agriculture and the natural resources on w h ich it depen ds w ill pu t us on a collision course w ith dis aster.” 1 F orm er Secretary o f Agriculture Earl Butz w arned the nation that the loss o f farm land to urban ization uses cou ld spell trouble fo r our fo o d supplies: “ I f U S D A ’s projections h old true, the consequences o f failure to stem the shrinkage o f U.S. farm land w ill b e om inous for the Am erican econ om y.” 2 T h e A m eri can Lan d Forum , in calling for action to protect farm lands from further loss, stated: . . . decisions about agricultural lands are actually bein g m ade now , at a tim e w hen the crucial im portance o f the resource is practically invisible to the average citizen.” 3 In ad di tion to their con cern for fo o d production, critics o f cro p land conversion to urban uses see other social costs, including a degraded environm ent, im paired w ater quality, lost w ildlife habitat, and dim inished beauty o f landscapes.4 D av id Berry and Thom as Plaut likew ise consider the loss o f scen ic qualities an ad di tional cost o f urbanization o f farm land.5 Bupert C u d er stated: “ M any otherw ise politically aggressive Am ericans seem to ‘clam u p,’ look the other w ay, or change the subject w henever it’s suggested that the p u b lic’s stake in private land use decisions has b een inadequately protected. H asn’t the tim e com e for a com prehensive effort b y local governm ents, aided b y state and Federal agen cies, to preserve som e o f these traditions, in a d em o cratic w ay, through the use o f local land use plans ap proved b y local p eop le?” 6 1W here Have the Farmlands Gone? (Washington, D.C.: National Agricultural Land Studies, September 1979), p. 3. 2Earl Butz, “ An Economic Analysis: U.S. Farmland Shrinking,” N ew York Journal of Commerce (O ctober 16, 1979). 3W here Have the Farmlands Gone? p. 14. *lbid., p. 10. 5David Berry and Thomas Plaut, “Retaining Agricultural Activities Under Urban Pressure: A Review of Land Use Conflicts and Policies,” in Policy Sciences ( Amsterdam, Elseview Publishing Company, 1978), p. 160. OCTOBER 1980 offset by other factors of production, a constant or rising demand for food coupled with a declining quan tity of prime cropland would lead to declining farm production and rising farm commodity and food prices relative to prices of other products. Finally, if food is becoming more scarce relative to nonfood products, given a relatively inelastic demand for food (a one percent change in the supply of farm products results in a larger than one percent change in price), a ris ing proportion of disposable personal income (per sonal income after taxes) would be spent on food. In other words, with a fixed relationship between land and farm production, a reduction in the real quantity of cropland with a constant or rising demand for food leads to rising farm product and food prices, higher real food costs, and a smaller percent of personal income available for nonfood purchases. Although, all of the above would be implied if a shrinking farmland crisis actually existed, none of these events is consis tent with the data. Quantity of Cropland Difficult to Measure As Theodore W. Schultz noted, economic analysis of land is not a simple matter. “Land as an economic variable is exceedingly hard to get at. . . . The fact that land is open and aboveboard, physical and con crete, and legally divided into neat, carefully de scribed parcels or lots . . . does not help one deter mine the supply of land.”4 In the early 1800s, economists such as Thomas Malthus and David Bicardo considered the contribution of land to food production to be relatively fixed and concluded that the real value of food would inevitably rise along with population growth, eventually neces sitating the use of poorer land, more machines, and more labor to produce additional food. Consequently, food prices and rent would rise relative to other prices.5 While this view recognized that cropland did not refer to a fixed number of acres, the potential real output of the land was assumed to be predeter mined and relatively fixed.8 It is now recognized that eW here Have the Farmlands Gone? p. 20. Implications of “Shrinking Farmland Problem” Several implications are immediately suggested by claims that there is a shrinking farmland crisis. First, and most obvious, is the contention that the quantity of farmland is declining. Second, if the amount of such land actually is declining, this fact should be reflected in the relative price of farm products and food. Unless Digitized for 12 FRASER 4Theodore W . Schultz, The Economic Organization of Agri culture (N ew York: McGraw Hill Book Company, Inc., 1953), p. 145. 5David Ricardo, The Principles of Political Economy and Tax ation, ed. Ernest Rhys (N ew York: E. P. Dutton and Co., Inc., 1948), p. 280; and Thomas Robert Malthus, On Popu lation (N ew York: The Modern Libraiy, published by Ran dom House, 1960), pp. 12, 13, 32, and 33. 6David Ricardo, The Principles of Political Economy and Tax ation, pp. 80-81. For a discussion of classical views, see Harry G. Johnson, Theory of Income Distribution (London: Gray-Mills Publishing Ltd., 1973), p. 74. F E D E R A L R E S E R V E BANK O F ST. LO U IS OCTOBER 1980 Table 1 Total U.S. Land Area, Farmland, Cropland, and Crop Yields Millions of acres* Date Total land Land in farms Cropland harvested Yield per acre 1967 = 100 1910 1,934 879 317 56 1920 1,934 956 351 61 1930 1,934 990 360 53 1940 1,934 1,065 331 62 1950 1,934 1,161 336 69 1959 2,314 1,124 317 85 1964 2,314 1,110 292 95 1969 2,314 1,063 286 106 1974 2,316 1,017 322 104 1979 2,316 1,049 337 130 "Includes Alaska and Hawaii, beginning with 1959. Source: Statistical Abstract of the United States (Washington, D .C.: U.S. Department of Commerce, Bureau of the Census, 1979) 100th ed., p. 6; Agricultural Statistics (Washington, D .C.: U.S. Department of Agriculture), 1978, p. 419; 1979, p. 417; Changes in Farm Production and Efficiency (Washington, D .C.: U.S. Department of Agriculture, 1978), p. 19; Agricultural Outlook (Washington, D .C.; U.S. Depart ment of Agriculture); Crop Production, 1979 Annual Survey (Washington, D.C.: U.S. Department of Agriculture, January 1980). the surface area that can be used for crops is more variable than Malthus and Ricardo thought and that output per unit of surface area is likewise more variable. Acres of Cropland Variable but Increasing Despite the difficulty of estimating the amount of cropland, it is now evident that the amount is not fixed. Rather, the quantity actually in use at any given time depends on a number of factors. Because new technology reduces land development costs and/or increases crop yields, thus providing favorable returns on the investment, land areas cur rently used for other purposes can be developed into profitable cropland. As Martin Bailey noted: “ . . . mountainous land good only for grazing could be leveled and made arable, and marshy lands, lake bottom and the fringes of the ocean could be filled to make arable land.”7 Examples of such conversion in the U.S. include the Imperial Valley in California, which was irrigated and developed into cropland, 7Martin J. Bailey, National Income and The Price Level ( New York: McGraw Hill Book Company, Inc., 1962), p. 111. and the Obion River Valley in Tennessee, much of which was developed into cropland through con struction of a drainage system. Further, although there is a vast amount of acreage (such as grazing, range, or forest land) that is not currently used for cropland at present prices, this acreage could be con verted to crop production within a short period of time if it was profitable to do so. If the net return on an acre of land is greater when used for crop produc tion than when used for grazing, it will be used for crops. Conversely, if the expected net return on land is greater when it is used for purposes such as grazing, forestry, etc., the land will be used for these purposes. As shown in table 1, only a small portion of the U.S. land area is currently used for crops. Of the 2.3 billion acres of land in the nation, only about 40 per cent is farmland and less than one third of this farm land has been actually used for crop production. Hence, a large amount of land area is available for conversion to or from crop production. Data on cropland harvested indicate that sizable changes have occurred during the past 80 years in terms of the land area used for crop production. Acres 13 F E D E R A L R E S E R V E BANK O F ST. L O U IS OCTOBER This relative increase in farm prices provided farmers with sufficient incentive to convert additional land to crop production. Table 2 Changes in Cropland Harvested and Selected Prices (Annual Rates) 1950-69 1969-79 -0 .8 1.7 Prices received by farmers 0.2 8.5 Price of industrial commodities 1.6 8.3 G NP price deflator 2.6 6.7 Consumer price index (all items) 2.2 7.1 Producer price index (finished goods) 1.6 7.3 Acres harvested Source: Economic Report of the President (Washington, D.C.: United States Government Printing Office, 1980), pp. 208, 259, 265, 268, 312; Changes in Farm Production and Efficiency, (Washington, D .C.: U.S. Department of Agriculture, 1978), p. 19. harvested rose from 317 million in 1910 to a peak of 360 million in 1930. By 1969, acres harvested had de clined to 286 million but increased again in the 1970s and rose to 337 million in 1979. Although government production control and crop diversion programs re duced the acreage of some crops harvested from 1934 through 1974, the effectiveness of these programs in terms of total crops harvested can be overemphasized since production of uncontrolled crops on diverted acres was permitted in most years. Furthermore, the impact of these production controls has been sharply reduced since 1969. The change in acres of crops harvested has been positively correlated with the change in farm product prices relative to other prices. For example, when the cropland acreage was declining (0.8 percent per year during the two decades, 1950-69), the index of prices received by farmers declined relative to other prices (table 2 ). During this same period, farm commodity prices rose only 0.2 percent per year — 1.4 percentage points less per year than the prices of industrial com modities or the producers price index, 2 percentage points less than the consumer price index, and 2.4 percentage points less than the GNP price deflator. During 1969-79, however, when the number of acres of crops harvested was rising, prices received by farmers rose at a slightly faster rate than most other prices. For example, during the 1969-79 decade, farm prices rose at an annual rate of 8.5 percent per year, compared with 8.3 percent for industrial commodities and less than 8 percent for each of the other series. http://fraser.stlouisfed.org/ 14 Federal Reserve Bank of St. Louis 1980 The increase in farm product prices relative to other prices during 1969-79 is not an indication of potential famine either in the U.S. or abroad. Bather, it repre sents a rise in export demand for U.S. farm products, attributable primarily to a gradual reduction in for eign trade restrictions ( that began prior to W orld War II) and to the large volume of U.S. currency accumu lations abroad (resulting from U.S. petroleum imports following the sharp increase in petroleum prices).8 The U.S. farm sector, having a comparative advantage in production of farm products (it is relatively cheaper in terms of resources used to produce farm products in the U.S. than in other countries), exported an in creasing proportion of total farm output. Yields Per Acre Rising While output per acre of cropland varies from year to year as a result of weather and other short-run factors, the sharp increases in crop yields over the longer run indicate that yields are sensitive to other factors of production such as technology and prices. The development of new technology and/or a change in the price of crops relative to the cost of farm inputs leads to a change in output per acre. For example, new technology that reduces the real cost of ferti lizers, improves insect and plant disease control, and provides improved seeds increases output per acre and, consequently, has the same impact on output as an increase in the acreage of farmland. In essence, the increase in farmland “quality” produces the same result as an increase in quantity. Similarly, an increase in the price of crops relative to the returns on land from alternative uses provides farmers with incentive for using more yield-increasing factors (e.g., greater quantities of fertilizer) per acre, as well as for using more acres of land for crop production. With the increase in the value of farm products in the early 1970s as a result of rising foreign demand, greater quantities of yield-increasing inputs were added to cropland; consequently, yields increased as did the number of acres harvested. The use of yield-increasing factors caused average com yields to rise from 70 bushels per acre in the three years, 1964-66, to 100 bushels per acre in 197779, despite the increase in acres harvested during the 8See Clifton B. Luttrell, “ Rising Farm Exports and Interna tional Trade Policies,” this Review (July 1979), pp. 3-10. F E D E R A L R E S E R V E BANK O F ST. LO U IS OCTOBER 1980 Table 3 Rates of Change of Selected Prices, Per Capita Personal Income, Percent of Personal Income Spent on Food, and Percent of Farm Commodity Sales Exported and Imported 1950-60 1960-70 1970-79 1950-79 Prices — rate of change: Received by farm ers (U S D A )1 -0 .8 1.5 9.1 3.0 Food ( C P I) i 1.7 2.7 8.2 4.0 All com modities less food ( C P I) 1 1.4 1.9 6.3 3.1 All services ( C P I) 1 3.6 3.8 7.6 4.9 Industrial commodities ( P P I) 1 2.0 1.4 8.9 3.9 Disposable personal income per capita (rate of c h a n g e )2 3.6 5.6 9.2 6.0 Percent spent on: total food2 22.4 — 20.2 20.2 — 17.3 17.3 — 16.6 22.4 — 16.6 17.8 — 16.1 1 6 .1 — 13.4 13.4 — 12.5 17.8 — 12.5 Exported3 10.1 — 14.1 14.1 — 14.6 14.6 — 24.8 10.1 — 24.8 Imported3 14.0 — 11.1 11.1 — 11.4 11.4 — 12.6 14.0 — 12.6 food at home2 Percent of total farm commodity sales: iEconomic Report of the President (Washington, D.C.: U.S. Government Printing Office, 1979), pp. 240, 248, 290; Eco nomic Indicators (Washington, D .C.: U.S. Department of Agriculture, 1980); CPI (Consumer Price Index); and PPI (Producer Price Index). 2National Food Review (Washington, D .C.: U.S. Department of Agriculture, Winter 1980), pp. 6, 56; and Economic Indicators. 3Economic Report of the President (Washington, D .C.: U.S. Government Printing Office, 1979), pp. 287 and 296; U.S. Foreign Agricultural Trade Statistical Report (Washington, D .C.: U.S. Department of Agriculture, 1970), p. 2; Agricultural Outlook (Washington, D .C.: U.S. Department of Agriculture). latter period.9 As shown in table 1, the trend rise in crop yields is not limited to com . Yields of all crop land harvested rose from an average index of 83 in 1957-59 to 123 in 1977-79, an increase of 48 percent. Furthermore, there is no indication that a slowing has occurred in the trend growth of crop yields. From 1967-69 to 1977-79, crop production per acre rose at a 1.7 percent rate, well above the 1.1 percent rate of increase from 1910 to 1969.1 0 9Agricultural Statistics (Washington, D .C.: U.S. Department of Agriculture, 1979), p. 30; and Agricultural Outlook (W ash ington, D .C .: U.S. Department o f Agriculture, May 1980), p. 33. 10A rate of yield growth higher than that of 1969-79 was realized only in the decade of 1950-60, when output per acre rose at a 2.6 percent rate. During the 1950-60 decade, however, the number of acres harvested declined sharply indicating that less fertile acres were taken out of crop production. A number of recent studies point to a possible decline in the rate of growth in crop yields in the years ahead. Agrtcultural Production Efficiency (Washington, D .C .: National Academy of Sciences, 1975), p. 195. This study concluded that biological realities suggest a slowing of the rate of in crease in productivity for most crops. Yoa-chi Lu, Philip Cline, and Leroy Quance, Prospects for Productivity Growth in U.S. Agriculture (Washington, D .C.: U.S. Department of Agri- Shrinkage Not Indicated b y Relative Prices of Food During the period, 1950-79, farm product prices rose at a slower rate than other major price series and only half as fast as disposable personal income (table 3 ). Consequently, the proportion of disposable personal income spent on food declined from 22.4 per cent in 1950 to 16.6 percent in 1979. During the more recent decade, 1970-79, farm prod uct prices rose somewhat faster than prices of most nonfarm products. Farm prices rose at a 9.1 percent rate, slightly faster than the 8.9 percent rate for in culture, September 1979). The authors expressed doubt that agricultural productivity growth through the year 2000 will equal the historical rate unless research and extension invest ment increase and unprecedented technologies develop. On the other hand, Glen L. Johnson contends that agri culture has a high long-run supply elasticity in The Over production Trap in Agriculture, ed. Glen L. Johnson and Leroy Quance (Baltimore: The John Hopkins University Press, 1972), pp. 20 and 183. Furthermore, he argues that if demand (fo r farm products) was doubled or tripled, we would have so much money invested in land that earnings would not cover acquisition costs. 15 F E D E R A L R E S E R V E BANK O F ST. L O U IS OCTOBER 1980 Table 4 Rates of Change of Specified Prices, Per Capita Personal Income, Percent of Personal Income Spent on Food, and Percent of Farm Commodity Sales Exported and Imported in the First and Second Half of the 1970s 1970-75 1975-79 Prices — rate of change: Received by farmers (U S D A ) 11.0 6.8 Food (C P I) 8.8 7.5 All com modities less food (C P I) 5.8 7.0 All services (C P I) 6.5 8.9 Industrial com modities (P P I) 9.3 8.3 8.7 9.7 17.3 — 17.0 17.0 — 16.6 13.4 — 13.0 13.0 — 12.5 Disposable personal income per capita (rate of change) Percent spent on: total food food at home Percent of total farm commodity sales: Exported 14.6 — 25.2 25.2 — 24.8 Imported 11.4 — 10.6 10.6 — 12.6 dustrial commodities, and well above the rate of in crease for all commodities (less food ) and for all services. The relative increase in farm product prices during this decade, however, was related to a sharp increase in demand for U.S. farm products, primarily for export, rather than to a shrinkage in cropland. (There is no evidence that farmland conversion to urban uses was greater in 1970-79 than in any other post-World War II decade.) Exports started rising in the 1950s, rose moderately in the 1960s, and acceler ated sharply in the 1970s. For instance the rate of increase was relatively low in the 1950s and the 1960s, and exports totaled only 14.6 percent of sales in 1970. However, exports accelerated from 14.6 percent of total sales in 1970 to 25.2 percent in 1975 (table 4 ). Furthermore, sharp increases in farm commodity ex ports were not offset by rising imports of farm com modities. Farm commodity imports declined from 11.4 to 10.6 percent of sales of farm products during the period. By 1975, farm commodity exports as a percent of sales had leveled off, and farm commodity prices be gan to decline again relative to other prices (table 4 ). From 1975 to 1979, farm commodity prices rose at the rate of 6.8 percent, 2.1 percentage points less than the rate of increase in the price of all services and 1.5 percentage points less than the price of in dustrial commodities. The price of food, which had 16 increased at about the same rate as disposable per sonal income in the first half of the 1970s, rose 2.2 percentage points slower in the second half of the decade than did disposable personal income. As indicated earlier, much of the increase in farm exports since the mid-1950s can be attributed to a gradual reduction in foreign trade restrictions, which had been almost prohibitive following the SmootHawley Tariff Act of 1930. A number of major reduc tions in average ad valorem rates have been nego tiated since the 1947 General Agreement on Tariffs and Trade; numerous studies show that these reduc tions have a major impact on trade.1 However, re 1 ductions in tariff duties do not increase trade imme diately, as evidenced by the gradual rise in exports during the 1950s and 1960s. Part of the sharp increase in exports during the early 1970s may be attributed to the implementation of monopolistic petroleum pol icies by the OPEC nations, which resulted in a sharp increase in dollar accumulations abroad and the dol lar’s reduced value in terms of foreign currencies. Despite the accelerating export demand for U.S. farm products, however, the farm sector concurrently produced enough food to maintain relatively stable 11See Clifton B. Luttrell, “ Rising Farm Exports and Interna tional Trade Policies,” pp. 6-7. F E D E R A L R E S E R V E BANK O F ST. L O U IS real prices for food in the U.S., further reducing the proportion of disposable personal income spent on food. Little Basis for Cropland Preservation Plans Considering the facts that cropland acreage is not shrinking, crop yields have increased, and food costs as a percent of personal income have declined, alle gations of a “shrinking” farmland appear to be un founded. Consequently, the arguments for develop ing comprehensive social plans to convert cropland to urban uses have little validity. Moreover, it is impor tant to note that, even if the claims of reduced farm land had been substantiated, proponents of social cropland control have not made a convincing case for such action. There are certain circumstances that might call for social land use planning: (1 ) if farmers are not price conscious, i.e., they are not responsive to current or expected future crop prices since they do not recognize the real value of prime cropland; (2 ) if farmland prices do not reflect the true value of the product of the land; and (3 ) if social planners’ knowledge about future land values is superior to that of current landowners and developers. Existing evidence does not corroborate the validity of these circumstances. Research clearly indicates that farmers are highly responsive to current and expected future prices.12 They sell their land to urban devel opers because its value is greater if used for urbaniza tion purposes than for cropland ( cropland value being determined by the current and expected future prices of the crops grown on it). When the value of land converted to urban use exceeds the value obtained from farming, the farm owner, land developer, and the general public will profit from conversion.1 In 3 the absence of harmful neighborhood effects (hidden costs), the costs and benefits of such shifts are care fully assessed by the transacting parties. In other words, the cost to the individual and to society is the foregone value of the land’s contribution to farm out put. Unless the gain in the new use exceeds the loss, the individuals involved would have no incentive for making the change. 12See Holbrook Working, “ The Theory of Price of Storage,” in Selected Writing of Holbrook Working, ed. Dana Kellerman (Board of Trade of the City of Chicago, 1977), pp. 28-30; Marc Nerlove, The Dynamics of Supply: Estimation of Farm ers’ Response to Price (Baltimore: The John Hopkins Press, 1958), pp. 186-235; and Zvi Griliches, “ Estimates of the Aggregate U.S. Farm Supply Function,” Journal of Farm Economics (M ay 1960), p. 282-93. 13For a discussion, see Neil A. Stevens, “ Rising Farmland Prices and Falling Farm Earnings: Is Agriculture in Trouble?” this Review (M ay 1978), p. 16. OCTOBER 1980 The second argument for social planning — that prices alone do not reflect the true value of the prod uct — implies that neighborhood effects are an impor tant factor. Some external costs, such as reduced water quality and impaired landscapes, have been mentioned by the proponents of social control over cropland. However, this argument is subjective because one can easily visualize rural scenes that are quite the oppo site of the beautiful landscape ideally depicted by advocates of social control. Cattle feeding pens, swine producing areas, and other livestock facilities are often sources of pollution. In addition, other “unsightly” views associated with farming communities include dilapidated buildings, fences, and equipment dumped along the roadside. In regard to water quality, most authorities contend that both rural and urban uses may result in water pollution. Those types of pollution that result from farming activities include runoff from livestock hab itats and chemicals used for controlling crop diseases, insects, and weeds. Allen Kneese contends that agri cultural chemicals present a special ( pollutants) prob lem “as they are delivered to streams in storm runoff from the land and bypass waste treatment plants.”1 4 The third argument for social planning — that social planners possess superior knowledge compared to that of private individuals — implies that individual farmers and urban land users distribute their resources between the present and the future on a relatively uninformed (o f true value) basis. In other words, indi vidual landowners are perceived to be somewhat myopic in assessing the future value of cropland, whereas social land use planners can clearly foresee the “correct” future value of land in its various alter native uses. This argument fails to consider that such vision would provide social planners with amazing opportunities for personal investment gains so that they would not be likely to remain “planners” when they could become wealthy as “doers.” Of even greater importance for the public welfare, however, is the assumption by proponents of social planning that such programs operate in the “public interest” rather than in self-interest. There is little justification for the view that self-interest is eliminated when social groups are given monopoly power over economic functions. As pointed out so cogently by George Stigler, alleged market failures are not evi dence that social planners can provide more services 14Allen V. Kneese, The Economics of Regional W ater Quality Management (Baltimore: The John Hopkins Press, 1964), p. 11. 17 F E D E R A L R E S E R V E BAN K O F ST. L O U IS at reduced costs. “W e may tell the society to jump out of the market frying pan, but we have no basis for predicting whether it will land in the fire or a luxur ious bed.”15 Any impediment to the transfer of crop land to urban use will increase the cost of land for housing, factories, hospitals, parking, and other uses vital to the public well-being. There is no evidence that social groups can more equitably resolve the conflict between costs and benefits of land use than can private markets. SUMMARY As prime farmland is converted into streets, shop ping centers, and residential areas, observers conclude that the quantity of farmland is declining sharply and that this decline should be controlled by social action. Unobserved, however, are the less noticeable but dra 15George F. Stigler, The Citizen and the State (Chicago: The University o f Chicago Press, 1975), p. 113. For a further discussion of this problem, see R. H. Coase, “ The Problem of Social Cost,” The Journal of Law and Economics (O c tober 1960), pp. 1-49; and Roger Leroy Miller, Economics Today (San Francisco: Canfield Press, 1976), pp. 615-23. http://fraser.stlouisfed.org/ 18 Federal Reserve Bank of St. Louis OCTOBER 1980 matic increases in acres of cropland and in produc tion per acre. The number of acres from which crops were harvested rose from the 1969 low point of 286 million to 337 million acres in 1979. Yields per acre of cropland rose at a 1.1 percent rate from 1910 to 1969 and at a 1.7 percent rate during the period from 1967-69 to 1977-79. As a consequence of the increase in acres harvested and in yields per acre, farm product and food prices have consistently declined relative to other prices, ex cept during the first half of the 1970s when export demand rose sharply. Since 1950, consumers have spent a declining proportion of their disposable per sonal income on food, even while a larger proportion of domestic farm output was being exported. Consequently, there is no justification for using social action to preserve cropland as proposed by critics of the current land market system. Further more, even if there was some shrinkage in cropland, there is no evidence that the problem can be solved more efficiently by social action than it can be in the market place.