Date of Award
Doctor of Philosophy
Luther H. Keller
Charles Sappington, Larry Bauer, Keith Phillips
The purpose of this study was to estimate the investment behavior function of individual farm firms using the concept of rational dis-tributed lag functions developed by Jorgenson to estimate the time structure of the investment process. The fundamental flexible acceler-ator model was used to estimate the investment behavior function. The model assumed net investment to be a lagged function of all changes in desired capital. Desired capital was assumed to be proportional to net farm income and gross farm income, respectively, for the expected profits and accelerator theories of investment. Because of the importance of the proper specification of the lag distribution in determining the time structure of the investment process, eight lag functions were imposed on the estimating equation for the expected profits apd accelerator theories of investment. In addition to the mechanism for converting changes in desired capital into changes in actual capital, nonfarm income, size of the farm firm, age of the farmer, and equity of the farmer in the farm business were added to the estimating equation in a linear fashion to establish the effect of these variables on net investment. Observations on 180 Tennessee test demonstration farms for the four-year period, 1965-1968, were used to estimate the parameters of the lag distribution and the structural equation. The data included observations on Grade A dairy farms, farms producing manufacturing milk, swine farms, and beef farms. A dummy variable for each of these farm classifications was added to the estimating equation to account for the effect of farm classification on the investment expenditures of the cross section of farms. The expected profits theory of investment was rejected by this study as an explanation of the investment behavior of this group of farmers. The accelerator theory of investment appeared to be a better explanation of investment. The lag distribution relating changes in gross income to net investment involved only two periods. One distri-bution accounted for 83 percent of the total effect in the first year while a second accounted for 93 percent in the first period. The estimates of the structural parameters of the accelerator model varied considerably with the lag function being estimated; however, some weak relationships were revealed. It appeared that as changes in gross income increased by $1.00, net investment increased by approxi-mately $1.53 to $1.69 given time for the adjustment to occur. Each $1.00 increase in nonfarm income seemed to have a negative effect on net investment ranging from -$13.00 to -$11.00. Each $1.00 increase in the size of the farm business increased net investment by about $0.12. The effect of age and number of dependents of the farmer could not be determined by this study. A positive relationship was estimated between equity of the farmer in the farm business and net investment ranging from 0.08 to 0.12. A weak relationship was interpreted regarding the effect of farm classification on net investment. In general. Grade A dairy farmers were the least likely to invest; all other variables held constant. They were followed in order of unwillingness to invest, all other variables held constant, by swine farmers, farmers producing manu-facturing milk, and beef farmers. The distributed lag theory and estimating procedure used in this study are theoretically sound, but the results of this study were not encouraging because of the instability of the structural parameter estimates between lag function specifications and the problem of obtaining an acceptable lag distribution. The evidence indicated variables other than those accounted for in this study play an important role in the investment behavior of Tennessee farmers.
Trevena, Billy J., "Estimating the investment behavior of farm firms using the concept of national distributed lag functions. " PhD diss., University of Tennessee, 1972.