Depository institutions must properly manage their liquidity to meet daily cash needs, to deal with a decline in the relative size of the industry, and to avoid the possibility of a bank run. Members of the Federal Home Loan Bank System can use the Advance Program to enhance their liquidity and simultaneously limit some of the interest rate risk associated with mortgage lending and mortgage related securities. This study investigates which factors influence the decision to use the Advance program and which factors influence the volume of Advances held by participating thrifts. A logistic regression analysis is used to evaluate which variables significantly impact a thrift's decision to use the FHLB Advance Program. A multiple linear regression model of thrifts with FHLB Advances measures which variables significantly affect the level of Advances held by thrift institutions. Fed Funds Borrowed .and Repo Agreements Sold was found to have a negative impact on the decision of a thrift to have FHLB Advances. Total Assets ( + ), Equity (-), Deposits Less Than $100,000 (-), and Deposits Greater Than $100,000 (-) were all found to have a significant influence on both the decision to have Advances and the volume of Advances held.

This dissertation examines several aspects of U.S. financial institutions’ real estate-related activity. The first two essays examine the impact of Federal Home Loan Bank (FHLB) membership and funding on bank and thrift holding company (BHC and THC) risk and returns. The first essay uses risk measures derived from BHC and THC stock prices, while the second essay uses risk measures based upon BHC and THC bond prices. The third essay studies the impact of BHC investment in real estate on risk and returns using measures based on stock prices. In the first essay, BHC and THC stock portfolios are formed along several dimensions. Bivariate generalized autoregressive conditional heteroskedasticity (GARCH) models are estimated to produce measures of total risk, market risk, and interest rate risk for the time period from the beginning of 2001 through 2009. Two sets of results related to FHLB activity are obtained. First, FHLB membership is found to be associated with lower total risk and market risk while having no association with interest rate risk. Second, and similarly, greater reliance on FHLB advances is associated with lower total risk and market risk but is not associated with interest rate risk. These results are consistent with the view that the risks created by government backing of the FHLB system and some of the system’s policies are mitigated by FHLB policies and products that reduce risk. In addition, THC stocks are found to have lower total and market risk than the portfolio of BHC stocks. The second essay investigates the relationship of both FHLB membership and funding with BHC and THC risk by using the cost of uninsured debt as a measure of risk. These relationships are analyzed in a simultaneous equation regression framework using data from the start of the third quarter of 2002 through the end of the first quarter of 2009. The cost of uninsured debt is proxied by yield spreads calculated from trading data on holding company (HC) bonds. Several interesting results are obtained. Reliance on advances is found to have a negative effect on the cost of debt throughout the sample period (the third quarter of 2002 through the first quarter of 2009). Cost of debt has a significant effect on the level of advances only during the recent financial crisis (the third quarter of 2007 through the first quarter of 2009), when the effect is negative. The negative association between cost of debt and the level of advances suggests that BHCs and THCs, on the whole, do not use FHLB advances to make unusually risky loans and supports the argument that FHLB policies and services have some risk-reducing effects. FHLB membership, independent of advances, is found to have no influence on HC cost of debt. Additional analysis indicates that THC status is associated with higher cost of debt than BHC status. The third essay examines the influence of real estate investment by BHCs from the third quarter of 1990 through the fourth quarter of 2010 on their risks and returns. Portfolios are formed of BHC stocks according to BHCs’ ratio of real estate investment to total assets and according to the type of regulation - lenient or strict - under which they invest in real estate. Tests of differences in median portfolio returns between these portfolios are performed. In addition, the effects of real estate investment on risk and return are estimated using univariate GARCH models of portfolio returns. The main results are as follows: 1) BHCs that invest in real estate have greater total risk and lower risk-adjusted returns than those that do not; 2) greater real estate investment is associated with lower returns and greater market risk for some types of BHCs while it is not associated with significant differences in total risk or risk-adjusted returns; and 3) BHCs that invest in real estate under relatively lenient rules have lower returns, greater total risk, and lower risk-adjusted returns than those that invest in real estate under relatively strict rules. The results indicate that benefits from real estate investment by banks - such as diversification of cash flows, economies of scale and scope, and increased charter value - are outweighed by greater variability of returns and lower returns due to BHCs’ lack of expertise in the field. The findings also provide evidence that rules granting banks greater freedom to invest in real estate result in increased risk but not increased returns. / Economics