Implied transaction costs in agricultural futures markets

The estimation of costs associated with operating in futures markets is challenging as portions of the costs are implied by prices and many times are not observed. This dissertation deals with measurement of the costs associated with volatility in the market, the risk premium, and the cost of immediate purchase or sale, the liquidity cost. Failure to understand these real transaction costs when liquidating positions may corrupt marketing strategies aimed at managing price risk.;Research has provided mixed results regarding the presence of risk premium in agricultural futures markets. In the first paper we test for the presence of a time-varying risk premium focusing on the properties of the underlying data. Our results show that accounting for the structural break in the 1970s plays a key role in the findings. In contrast to recent research, we find only limited evidence of time-varying risk premium. For a two-month horizon the corn, soybean meal, and hog markets show no signs of a risk premium, while very weak support emerges in live cattle. For the four-month horizon, no evidence of a time-varying risk premium appears in any of the markets.;Estimating liquidity costs in agricultural futures markets is challenging because bid-ask spreads are usually not observed. In the second paper we develop a modified Bayesian estimator to measure liquidity costs. Based on an ability to reflect simulated data from Roll's spread model, the modified estimator is precise and works well under conditions found in agricultural futures markets. Conventional serial covariance and absolute price change spread estimators are biased. Using data from live cattle and lean hog contracts, similar performance patterns emerge which result in economically meaningful liquidity cost differences. Results from the modified estimator identify a highly liquid and efficient environment, consistent with the openness, transparency, and volume traded in these markets.;Understanding the determinants of liquidity costs in agricultural futures markets has been hampered by the need to use proxies for the bid-ask spread which are often biased, and by a failure to account for the effect of new information on jointly determined micro-market variables. In the third paper we assess the determinants of liquidity costs using the modified Bayesian bid-ask spread measure that has been shown to be more accurate than traditional measures, and a GMM IV approach to account for joint determination in the micro-structure process. Our results show that total daily volume, volume per transaction, and price volatility are the main determinants of liquidity costs in both hogs and cattle markets. Higher levels of electronic trading consistently lead to lower pit bid-ask spreads particularly in the cattle market, implying the existence of competitive pressure.;The combined research leads to a more comprehensive understanding of market behavior. We show that implied transactions costs do exist, at times are large, and are not easy to estimate, but careful examination of the data and selection of procedures can shed light on their magnitude and patterns of behavior. Accurately measured and understood transaction costs can then be used by traders, exchanges, researchers, and policy makers to develop a better understanding of markets and to make appropriate decisions.