Using two different approaches, the relationship between a firm's investment and its financial variables is examined. Imperfections in the credit market such as asymmetric information have led researchers to explore these relationships. This study incorporates the 5 Cs of lending (character, capacity, collateral, credit rating, and capital) into the farmer's investment decision and explores the impacts of these variables on a basic data set. The data set is composed of 590 Iowa farms that are members of the Iowa Farm Business Association and have reported farm level financial and production data from 1991 to 1995;The first approach consists of a composite regression model constructed from various elements of traditional investment models and variables representing the 5 Cs. The second approach derives an investment equation from the firm's optimization problem, an Euler equation approach. The 5 Cs of lending are incorporated into the problem through a borrowing constraint;The composite regression approach is conducted under a Bayesian framework with variable selection and outlier detection components. The results imply strong support for the accelerator model of investment and the inclusion of other relevant variables, among them the value of short-term assets, one of the proxies for the 5 Cs. Another of the proxies, operator age, receives less support. The Bayesian framework with the variable selection and outlier detection components works extremely well;The Euler equation approach is more problematic. Under the original specification looking at net investment, all models are rejected and the most preferred model is also the most restrictive with symmetric adjustment costs and no financial constraint. Within the financial constraint, only the value of short-term assets and net worth are ever found to be statistically significant. Estimated adjustment costs are either negative or positive but extremely small. The shadow value of external finance is estimated to be around 100 percent. Other formulations, extensions, and reduced form models are explored and similar results are found. Given the mostly negative results from the Euler equation framework possible reasons for them are reviewed.
Available at: http://works.bepress.com/chad-hart/135/