Essays on economics of banks and credit unions

This dissertation is composed of three essays that study various aspects of proper econometric modeling of the operations by financial intermediaries. I focus on extending and implementing methods to address issues of endogenous selection, observed and unobserved heterogeneity, and inherent uncertainty in operations by credit unions and banks.;Credit unions differ in the types of financial services they offer to members. In the first chapter, I explicitly model this observed heterogeneity using a generalized model of endogenous ordered switching. My approach captures the endogenous choice that credit unions make when adding new products to their financial services mixes. Failure to do so yields biased and inconsistent estimates. The model I develop also allows for the dependence between unobserved effects and regressors in both the selection and outcome equations and can accommodate the presence of predetermined covariates in the model. I use the model to estimate returns to scale for U.S. retail credit unions in 1996-2011. I find strong evidence of persistent technological heterogeneity among credit unions offering different financial service mixes.;The next chapter considers a more flexible model of credit union technologies in which parameter heterogeneity is extended to institutions offering the same service menu. I do so by permitting the outcome equation of interest to take a semiparametric, varying coefficient form. I propose a consistent, asymptotically normal two-stage estimator. Applying this estimator to credit union data, I document significant distortions in the estimates of technological metrics if more restrictive conventional models are used.;The third chapter discusses how to properly incorporate risk into an empirical model of banks' production. The literature has largely accounted for it by using ex-post realizations of banks' uncertain outputs and risk. This is equivalent to estimating an ex-post realization of bank's technology which, however, may not reflect optimality conditions that banks seek to satisfy under uncertainty. I offer an alternative methodology based on the ex-ante cost function. I model credit uncertainty explicitly by recognizing that bank managers minimize costs subject to given expected outputs and risk. I use this framework to estimate production technology of U.S. commercial banks in 2001-2010.