A Note on the Nerlove Estimate of Supply Elasticity

Abstract

N 'ERLOVE has proposed an ingenŸ method of estimating the response of acreage of a crop to the anticipated price for the crop. 1 Anticipated prices are not observable, but he hypothesizes that they change from year t --1 to year t by some fraction of the difference between the actual and anticipated price in year t -1:(1) Pt* -P't-1 = 5(Pt-1 -P't-l) 0 < ~ < 1 Pt* and P*t-a are anticipated prices, Pt-1 Ÿ an actual market price, and is a "coefl~cient of expectation."Nerlove shows that (1) implies (2): (2) Pi* = 5Pt-~ q-(I -5)SPt-2 + (1 -5)=SPt-8 q-" ""Acreage, xt, is assumed to be a linear function of expected price and an error term, ut (together with trend where appropriate):(3) xi = a0 + alPt* -+-uf A supply elasticity coefficient can be computed from al. Equations (1) and (3) lead to (4) xt = r0 + r~Pt-~ + r~xt_l + vt ~2 equals 1 --~, ~1 equals al~ and T0 equals ao~. All variables of (4) are observable. ~~ and ~1 can be statistically estimated, and they lead to an estimate of al. Using this approach, Nerlove computed supply elasticities for cotton, wheat and corn over the period 1909-32.The approach assumes that acreage in any year represents the influence of 15ast market prices, a smooth trend, and random factors. Intuitively, one 1Marc Nerlove, "

Regional Supply of Hired Labor to Agriculture

“…; 3 In the present research the demand relation remains latent except for purposes of identification. The justification for this particular demand model can be found in Schuh and Leeds [25].…”

Economic factors associated with variation in aggregate wheat acreage in New Zealand (1945–1965)

“…Indeed it can be plausibly argued that the mis-specification of the price variables is the basic reason for the appearance of autocorrelation in Candler's results, since in other respects his final model used in statistical estimation is not significantly different from the one used by Nerlove and Addison [20]. They state that "the most significant finding of the paper is that evidence of positive serial correlation in the 2 In other work not reported in this article specific allowance was made for some aspects of technological change, and particularly for change in the level of physical output associated with an enterprise. The results obtained provided a somewhat inferior explanation of wheat acreage changes.…”

“…Such an assumption seems to be consistent with the available evidence. 2 Candler experienced considerable difficulty with multi-collinearity and in consequence, when he applied the conventional significance tests to his results derived from data relating to harvests in the years 1924 to 1953 he obtained the regression relationship…”

“…Ai,=the long term equilibrium acreage which would be devoted to the i-th enterprise if the expectations about the relative prices of products which were held in time period t were to continue through all future periods, and other non-price factors affecting acreages were also allowed for along the lines suggested by Brandow [2] and Nerlove [19]; together with the identity m (2) L = 2 Ai.…”

Econometric simultaneous equation models of the cattle cycle in the United States and three selected regions

“…Since data on the inventory of heifers more than two years of age are available in Switzerland (p. 37) (they are not available in the U.S.A.), one could specify and estimate (by, e.g., 2SLS) an equation with two (or more) endogenous variables. The number of young heifers kept as heifers over two years of age could be taken as the left-hand dependent variable, ^See especially Kierlove (1958aKierlove ( , 1958bKierlove ( , 1958c, Brandow (1958), andLadd (1959).…”