The effect of land quality on agricultural land values

Abstract

The average land price in Kansas has recently been through a period of large growth and decay, nearly doubling from 2010 to its peak in 2014, but falling from 2014 to 2017 in both real and nominal terms. However, there is anecdotal evidence that not all land prices are dropping at the same rate. Lower quality land prices seem to be dropping at a higher rate than the higher quality land prices. The goal of this analysis is to give analytical evidence to support the belief of different rates of price changes for different qualities of land. The hypothesis is that once the farm economy entered its period of negative growth, the producers that over-leveraged themselves needed to sell some of their assets to correct their balance sheets and that low quality land is the primary asset liquidated. The producers that did not over-leverage themselves would still be looking to purchase the right piece of land. This creates a surplus of less preferred low quality land on the market, while the supply and demand for the high quality land stays strong. This analysis was completed using 56,291 observations on land sales from 33 years starting at the beginning of 1985 and continuing on through the middle of 2017. A real price per acre for the land weighted by the number of acres in each parcel was calculated for each quarter, as well as a variable with the price of land lagged one quarter. Data on real net farm income, the S&P 500, and 30-year fixed-rate mortgage interest rates were also collected and used to create averages for each of the 131 quarters analyzed in this work. Finally, a variable representing the percent of all sales in each quarter in the bottom 25 percent in quality of all land sales was created. Quarterly dummy variables were included to control for seasonality. Two regressions were run with the only difference being the exclusion of the variable representing the bottom quality sales in the first in order to compare the results. Analysis of the first regression shows positive relationships between the dependent variable of the logged real price per acre and the independent variables of the logged lagged real price per acre, real net farm income, the S&P 500, and land sold during the third quarter of the year compared to the first quarter. There is a negative relationship suggested between the logged land price and the 30-year fixed-rate mortgage interest rate. Inclusion of the variable representing sales of land in the bottom quartile of quality suggest results consistent with the first regression, with some of the variables becoming more statistically significant. More importantly, this analysis shows a negative relationship between average land price and the variable representing land quality. This shows that the average land price is affected by the quality of the land sold at a statistically significant level.