Determinants of Farmland Values

Real-estate sale data for 1975 and 1976 in Riley, Geary, Morris and Chase counties, Kansas, were analyzed to determine effect of soil productivity on sale price. Both soil and nonsoil components of the sale tracts were regressed against sale price per acre for 57 sales within the study area. The study showed that bordering roads, year sold, soil capability class, grain sorghum yield and percent cropland had the greatest influence on sale price. The average per-acre value of farm land and buildings in Kansas was $590 February 1, 1981 (Anonymous, 1981). The average farm had 653 acres, an average real-estate value of well over a third of a million dollars. The value per acre varies widely, from a few dollars for wasteland to thousands of dollars for good irrigable land or land with minerals or near urban centers. The variation in agriculture-land values stems largely from factors that affect earnings such as soil, improvements, and roads. What factors and how they affect land values are questions in the minds of buyers and sellers, creditors, assessors, and others. Two basic approaches can be used to estimate values: 1) estimate the expected net earnings and 2) analyze realestate sales. This article is concerned with the second approach. We used multiple regression to analyze 57 sales in Riley, Geary, Morris, and Chase counties in 1975 and 1976 for an indication of the factors affecting prices and an estimate of each factor's effect. Judgement or experience and the availability of information determine factors to include in the analysis. A given factor, say roads, may affect 1 Contribution 80-23-J, Department of Agricultural Economics and Department of Agronomy, Kansas Agricultural Experiment Station, Kansas State University, Manhattan, Kansas 66506. 2 Former Graduate Assistant in Agronomy, Professor of Agricultural Economics Emeritus, and Soil Survey Research Scientist, Kansas Agricultural Experiment Station, respectively. This content downloaded from 207.46.13.129 on Fri, 01 Jul 2016 04:16:48 UTC All use subject to http://about.jstor.org/terms VOLUME 85, NUMBER 3 153 prices of tracts of land differently; that is, the effects vary among the tracts, so the average effect may not be reliable. Many factors may require many sales for reliable estimates of each's influence, and sales of farm real estate are relatively infrequent. Still we obtained information for nine characteristics of 57 tracts in the four counties and used them in the following multiple linear-regression model:3 Y = 3o + 81X10 + 12X20 + 33X21 +4?X3o + 35X40 + 7X51 + P 8X60 + P9X70 + 310X71 + /31X72 + 312X80 + 913X90 + E Where Y = Sale price per acre X70 = Geary County tract X10 = Total acres in sale tract X7, = Morris County tract X20 = Full improvements X72 = Chase County tract X21 = Partial improvements X80 = Percent cropland X3o = Miles to town X9o = Capability class 4 X40 = Bordering hard-surface road fs = Parameters of the model X,5 = Grain sorghum-yield estimate E = Random error X60 = Year sold Of the nine characteristics, each of five affected the price per acre at the 95 percent or higher level of confidence. Characteristics Estimated effect per acre Bordering hard-surface road compared with graveled road $85.91 Grain sorghum yield (bushels per acre) 7.13 Year sold (1976 over 1975) 62.48 Percent in cropland 1.40 Land capability class 4 compared with class 34 -54.41 For discrete variables such as improvements (none, partial, full) one was eliminated to avoid multicollinearity. 4 Class 3 not included to avoid multicollinearity. This content downloaded from 207.46.13.129 on Fri, 01 Jul 2016 04:16:48 UTC All use subject to http://about.jstor.org/terms 154 TRANSACTIONS OF THE KANSAS ACADEMY OF SCIENCE Those five characteristics accounted for 59 percent of the variation in prices paid for the 57 tracts, so 41 percent of the variation stemmed from other factors.5 The 59 percent is high compared with results of similar studies. The value of hard-surfaced road over graveled road is understandable but $85.91 per acre may seem high. Dividing $85.91 by the average sale price per acre of $375.45 gives 0.23. In a similar multiple-factor analysis of Morris, Chase, Pottawatomie, Wabaunsee, and Lyon counties for 1957 and 1958 (Pine and Johnson, 1978), the ratio ranged from 0.26 to 0.37, so 0.23 is reasonable. A $7.13 increase for each additional bushel yield per acre of grain sorghum indicated that grain sorghum yield had a high correlation with sale price.6 The variable for date of sale showed a $62.48 increase per acre from 1975 to 1976, which is in line with past increases in land values (Pine and Johnson, 1978). The $1.40 increase in price per acre for each additional percent of cropland is reasonable for cropland is valued higher than grassland. A tract of cropland sold on an average for $140 per acre more than pastureland. Having more management limitations, Class 4 soil being worth $54.41 less than Class 3 also seems plausible. Three characteristics did not significantly (95 percent level) affect the price per acre: improvements, miles to town, and location (county). The first two were significant in other studies (Edwards et al., 1964), so they may be significant in another area or period of years. Likewise, the five significant factors in this study may not be significant in another situation. Effects will vary among individual tracts. Estimating the value of a tract of land requires judgement. The many variables (characteristics of tracts and of buyers and sellers), difficulty in measuring influential factors, and infrequent sales of farm real estate complicate the task. One should use all reasonable information and procedures in evaluating land.