Part I tests the present value model in aggregate and disaggregate land markets, based on the testable restrictions derived from the representative agent dynamic asset pricing model. The hypothesis of cointegration between land prices and rents is formulated and tested by using data on Iowa land prices and rents over the 1950-1986 sample periods. I found mixed evidence in regard to implications of the model. At the aggregate level, there exists evidence of cointegration between land prices and cashrents but not between cropshare rents and land prices. When both types of contracts are considered, there is stronger evidence of a cointegration relationship between land prices and rents. There is less evidence of cointegration at the disaggregate level. One possible explanation of the conflicting results between the aggregate and disaggregate levels is that micro bubbles from local speculation in disaggregate markets are washed out in the aggregate market;Part II investigates spatial market efficiency of land markets. The spatial land markets efficiency hypothesis implies the existence of a long run relationship between land prices in different geographical areas. Long-run equilibrium in relative land prices suggests that a cointegration relationship should exist across markets. The empirical results supported cointegration in one-half of the tests. Stronger evidence that markets have a long run equilibrium relationship was obtained in adjacent markets;Part III examines the risk-return relationships in farmland markets to establish the existence or absence of effects corresponding to the equity puzzle found in financial markets. Two measures of risk, the market beta and the consumption beta, are estimated for two contracts used in farmland markets: the cropshare contract and cashrent contract. The results reveal no significant differences in betas between the two contracts. The low beta estimates indicate that farmland contributes little systematic risk to a well-diversified asset portfolio or a consumption path.