The Cost of Debt for REITs: The Mortgage Puzzle

Abstract

Established, low-leverage equity REITs with access to the public debt market rely on both non-recourse mortgages and full recourse bonds/notes as sources of long-term debt. Interest rates on secured, non-recourse debt (mortgages) include a costly strategic default option premium and do not benefit from a firm’s overall financial capacity. We find that use of non-recourse, mortgage debt is more likely for longer-term, smaller borrowings, and during recessionary periods, consistent with REITs valuing financial flexibility in their capital structure. The higher rates for property-level debt suggest a benefit to REITs versus single asset investors in terms of cost of capital. Since REITs also access debt at the corporate level, the spread between long-term non-recourse debt and long-term recourse debt implies a benefit to the REIT structure.

Keywords:

• Strategic default
• recourse debt
• REIT
• financial flexibility
• interest rates
• debt overhang
• secured debt
• portfolio risk

Acknowledgments

We’d like to thank William Hardin, III (Editor), anonymous referees, Shaun Bond, Andrey Pavlov, Eli Beracha, Jiajin Chen, Ken Mischel, Miriam Krausz-Cohen, Joseph Aharony, Meni Abudy, Alon Raviv, Guy Kaplanski, Jacob Weisberg and participants of the Ashkelon College School of Business seminar, the Bar Ilan seminar, the UNSW Seminar, the Monash University Seminar, the Midwest Finance Association Meeting 2016, the American Real Estate Society Meeting 2016, the Eastern Finance Association Meeting 2016, and the American Real Estate and Urban Economics Association National Meeting 2016 for their helpful feedback and comments. We gratefully acknowledge that this paper is the winner of the 2016 ARES Manuscript Prize for research on Real Estate Investment Trusts sponsored by NAREIT.

Notes

1 The REIT literature on capital structure has included studies by Boudry et al. (2010), Hardin and Hill (2011), Harrison et al. (2011), Giambona et al. (2008) and Giambona et al. (2012), among others.

2 One unobserved source of financing is debt taken on by REITs in Joint Ventures. Since these are off balance sheet transactions, we cannot control for those borrowings. It is likely that those borrowings are also mortgage loans, and that our mortgage borrowings are understated.

3 Contrary to residential real estate and small business loans (Giambona et al., 2013), large commercial real estate loans are typically non-recourse. There are a few exceptions to this in the case of development projects and other speculative ventures, perhaps in the case of smaller banks and borrowers. However, in the institutional space and, for example, in the CMBS world, most cash flow dependent loans are non-recourse.

4 While a REIT may choose to continue to service a non-recourse mortgage, the debtholder only has recourse in default to the property including associated cash flows.

5 Another indication that REITs’ unsecured, recourse debt issues do not have strategic default risk premiums emanates from conversations with REIT bond buyers at insurance companies, who have revealed that asset market values are not used to conduct covenant tests. A Morningstar REIT Credit Rating Methodology Report reveals a “market value” adjustment to historical acquisition prices as a 125% across the entire asset base. Thus, it would appear that the bond LTV covenant verification, if conducted, is primarily based on historical values, and not current appraisal values, thereby mitigating the risk of strategic default based on property market values.

6 At least 90% of taxable income must be distributed to shareholders annually in the form of dividends. Source: http://www.sec.gov/answers/reits.htm. While recent studies show that the dividend to FFO distribution ratio is closer to 70%-80% (Case et al., 2012), a considerable proportion of operating cash flows is nevertheless distributed.

7 Feng et al. (2007) have shown that REITs are more likely to have high leverage ratios when they have high market valuation and growth opportunities.

8 Sun et al. (2014) have shown that REITs with higher leverage had inferior performance subsequent to the financial crisis.

9 Debentures for firms without market access consist primarily of private placement debt instruments, with the majority issued by one REIT, Monmouth Real Estate Investment Corp (NYSE: MNR).

10 A portion of the mortgages are provided by CMBS lenders. While the loans are securitized, the performance of the property is still monitored by the servicer via lease approvals, springing cash flow sweeps, various lock box structures, DSCR covenants and other monitoring mechanisms.

11 The data shows that at least 83% of the observations are fixed rate, however the index and spread information is not available. By controlling for various index rates (Libor, 10-Year Treasuries, 30-year Mortgage rates) we attempt to address the rate variance due to index fluctuations.

12 The higher interest rates on mortgage debt may reflect an illiquidity premium included in the rate. Thus, the differential rate may include both a premium for monitoring costs and for illiquidity.

13 Revolving debt typically also includes non-usage fees and other fees not available in the data.

14 Hardin and Hill (2011) have shown that REIT lines of credit have a significant relationship with access to public debt markets and provide substantial liquidity.

15 Riddiough and Wu (2009) find that lines of credit are used to preserve debt capacity.

16 This is previously shown to be of importance for REITs by Hardin et al. (2017), Han (2006), Capozza and Seguin (2003), and others.

17 Chui et al. (2003) find that size is a persistent factor in REIT returns and Ambrose et al. (2005) show that larger REITs have lower costs of capital.

18 Alcock et al. (2014) have shown that debt maturity is an important factor in determining REIT leverage.

19 Downs and Guner (2000) show that analyst coverage has a significant relationship with REIT liquidity.

20 Letdin et al. (2018) find that NAV Dispersion is a significant indicator of firm value, where greater analyst dispersion has a negative and significant relationship with firm value. Earnings forecast dispersion is found to not be significant in comparison.