Essays on Measuring and Explaining Commodities Returns

This dissertation consists of three essays on measuring and explaining commodities returns. While the first two essays directly address the measurement and explanations for the returns to commodities futures, the final essay advances the econometrics tools required for studying commodities.;The first essay shows that adding durable consumption growth to a multi- factor consumption-based asset pricing model can explain high commodities returns. Commodities returns covary strongly with durable consumption growth, and systematic durables risk can explain the cross section of returns to commodity portfolios. Commodities' high durables risk may be explained by the business cycle properties of commodities and durable consumption growth.;The second essay uses a model to describe the inventory decisions of a commodity-producing firm to derive an inverse relationship between the inventory/sales ratio and the returns and convenience yield. Using data on commodities inventories, sales, and returns, this essay provides empirical support for the model. It also demonstrates the procyclicality of the excess returns and convenience yield and the countercyclicality of the inventory/sales ratio.;The final essay addresses methods for inference testing when using serially correlated data with missing observations. Current practices for working with missing data, including imputation or ignoring gaps in the data, are far from standardized and generally lack theoretical justification. This essay formalizes these existing methods and presents two new methods for heteroskedasticity and autocorrelation robust inference. It demonstrates that methods that use the observed values without imputing the missings are preferred to imputation methods.