Real Estate Investment Trusts (REITs) with different property type focuses exhibit significant variation in capital structure. Average leverage ranges from a low of forty percent for industrial REITs to a high of sixty-one percent for hotel REITs. Debt maturity ranges from seventy-four months for office REITs to 105 months for hotel REITs. The extent of this variation is surprising because all REITS share common tax and governance attributes. In this paper, we explore the systematic variation in capital structure by REIT type. We first examine how well the traditional corporate finance explanations for variations in capital structure explain the observed variations within the REIT sector. Myers (1977) and Hart (1993) show that changes in leverage and the mix of long- and short-term debt can alter the incentives of managers and help resolve the over- and under-investment problems identified in the agency and corporate governance literature. Other theoretical papers have focused on the problems associated with asymmetric information between managers and investors (Diamond, 1991 and Titman, 1992) and the use of debt to signal the quality of the firm and its investment prospects. Several recent empirical studies have modeled a firmís capital structure choices (Datta, Iskandar-Datta and Raman, 2005, Johnson, 2003 and Barclay, Marx and Smith, 2003) that support these theories. We estimate a simultaneous equation model similar to Johnsonís (2003) for a panel data set of REITs and find that the ratio of market-to-book value (traditionally viewed as a proxy for firm growth opportunities) and firm size are significant in explaining both the leverage and debt maturity equations and have signs consistent with their hypothesized effects. We then extend the Johnson (2003) model to include measures of the liquidation value of the underlying collateral. Shleifer and Vishny (1992) argue that variations in the liquidation value of the assets across different industries can help explain the observed variations in capital structure by industry and time. Briefly, Shleifer and Vishny argue that firms in industries where assets have high liquidation values will use higher leverage and longer maturity debt. This liquidation value theory has been hard to test empirically because of difficulty in measuring liquidation values and consequently liquidation value has not been included in previous studies of firm capital structure. However, the REIT sector provides a unique opportunity to test this theory because REITs typically concentrate in a particular type of real estate asset and different types of real estate assets exhibit significant differences in liquidity and redeployability. We use two different quantitative measures of liquidity as well as a more subjective rank ordering of property types by liquidation value to extend the model of capital structure. First, because short-term leases provide the landlord an option to reorient the current property, we add the average lease maturity to the system of equations for leverage and debt maturity. Consistent with Shleifer and Vishnyís (1992) hypothesis, we find that REITs with shorter maturity lease structures use lower leverage and shorter maturity debt. Second, we measure liquidity by the average recovery on defaulted Commercial Mortgage-backed Securities (CMBS) loans collateralized by different property types. Highly liquid, easily redeployed assets should retain their value better than less liquid assets and thus result in lower average losses for lenders when a default occurs. When Johnsonís model is extended using this categorical variable, it provides further evidence that liquidation value influences capital structure. Finally, we extend the concept of liquidation value to include more qualitative measures of liquidity such as zoning flexibility and physical redeployability. We rank order the five major REIT property types using all four dimensions of liquidity and identify industrial and residential REITs as having the most liquid assets and office and retail REITs as having the least liquid assets. We re-estimate the model using property type indicators as proxies for liquidation value and find evidence consistent with the hypothesis that firms with highly liquid assets use more leverage and longer debt maturity. We conclude that the liquidity of the underlying assets significantly influences managementís choice of capital structure within the REIT sector.