The value of land dominates the financial structure of most American agricultural production firms, and land values are an important factor in long-term agricultural planning and risk management. As the primary source of collateral for farm loans, farmland values have significant implications for both producers as well as bankers financing agricultural loans. The Federal Reserve Bank of Kansas City’s Survey of Agricultural Credit Conditions is an expert opinion survey in which agricultural bankers provide land value forecasts. As the survey has drawn increased attention, the survey has drawn criticism regarding its use qualitative data to forecast land values. Our research examines the value of the survey data with respect to its ability to forecast movement in land values. Three techniques are used in the analysis. Interpreting the aggregate forecasts as probability estimates, Brier’s probability scores are used to evaluate aggregate bankers’ predictions. Next, turning points are evaluated using contingency tables. Finally, Granger causality tests are used to determine the dynamic relationship between land value predictions and actual land value changes reported by bankers. Bankers’ forecasts predict land values for irrigated and ranchland well, but non-irrigated forecasts were only marginally helpful in prediction non-irrigated farmland values. Forecasts provided in the survey may be beneficial, especially considering the scarcity of other publicly available data.