Interest rates and agricultural debt: estimating the marginal interest rate

Purpose - While the average cost of debt capital can be calculated from historical financial statement data by dividing the interest paid each year by the total level of debt, this average cost of debt provides little information regarding the true cost of acquiring additional debt capital, and hence, its use is potentially problematic in financial decision-making. This study focuses on the linkage between observed changes in the average interest rates calculated from financial statements (balance sheet and income statement) and the marginal cost of borrowing or the cost of acquiring new debt. Motivated by the capital asset pricing model (CAPM), the marginal cost of capital is modeled as a function of a risk-free interest rate (the return on Moody’s Aaa bonds), returns on the S\&P stock index capturing overall market returns and a portfolio of agricultural stocks to represent farm sector-specific risks. Design/methodology/approach - Using a unique dataset constructed from United States Department of Agriculture (USDA) state-level Financial Performance of the Farm Sector data for the years 1960 through 2003 and state-level Agricultural Resource Management Survey (ARMS) data for the years 2003–2014 and Bayesian methods, we model the observed interest rate as an autoregressive function controlling for changes in debt and key rates of return in the general economy. Findings - The results indicate that the marginal interest rate is a function of the Aaa corporate bond rate and the stock market. We also find evidence of a negative relationship between returns to a portfolio of agricultural stocks and the marginal interest rate. Overall, the findings suggest that the imputed interest rate frequently misrepresents the marginal cost of debt capital. Originality/value - Most farm financial datasets allow for the analysis of the farm firm’s average interest rate. However, farmers make decisions based on the marginal cost of credit – the interest rate on a newly issued note. This study estimates this marginal interest rate for the 15 states for which the ARMS data are representative for the years 1960 through 2014 and compares the estimated marginal interest rate with the imputed average interest rate.