https://orcid.org/0000-0003-3928-0516
![Sample our Built Environment journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5926/TOC-Subject-Sample-2021-Built-Environment.jpg"})
Abstract
This paper investigates the impact of property type and geographical diversification on the cost of public debt for U.S. real estate investment trusts (REITs). We measure diversification using the negative of the Herfindahl–Hirschman Index (HHI) and represent the debt cost by the yield spread of the debt issue. We find that property-type diversification has a cost-decreasing effect on public debt. For example, our estimations illustrate that a one standard deviation increase in property type diversification decreases the yield spread by 10.97 basis points on average. In contrast, geographical diversification has a cost-increasing effect. For example, we estimate that a one standard deviation increase in regional diversification increases the yield spread by 9.95 basis points on average. Moreover, we find that controlling for the S&P’s credit rating of the debt issue does not entirely absorb the impact of diversification on the yield spread, suggesting a difference in the credit-risk assessment of diversification’s effects between public lenders and credit rating agencies.
Notes
1 There is also literature that discusses the impact of diversification on the value of equity (Campbell et al., Citation2003; Eichholtz et al., Citation2000; Feng et al., Citation2021; Hartzell et al., Citation2014); furthermore there are studies that investigate the impact of diversification on the firm’ performance (Capozza & Seguin, Citation1999; Ambrose et al., Citation2000).
2 Glancy et al. (Citation2021) argue the importance of recourse for real estate lenders.
3 An alternative measure of diversification is the entropy index (Jacquemin & Berry, Citation1979). The entropy index uses a weighing scale that is more sensitive to types and regions of small weights, providing more variation than the Herfindahl–Hirschman index (HHI). We obtain the same findings using the entropy index. See Appendix A, .
4 Better weights are the market value of the properties, however we couldn’t observe this data. The literature uses the book value of properties instead (e.g., Hartzell et al., Citation2014). In a robustness test we re-estimate the effect of diversification using the book value of the properties as weights. The results are consistent with our findings, however, we only have 215 observations for this alternative weighting method. See Appendix A, .
5 We recognize that the book value of the total assets does not generally reflect the market value of the asset portfolio. Hence, in a robustness check we define total capitalization as the market capitalization plus the book value of debt and the book value of any preferred issued. The results lead to the same conclusions. See Appendix, .
6 The market value of the assets is defined as the book value of the assets + market capitalization − the book value of the equity.
7 See, for example, Rajan and Zingales (Citation1995), and Teixeira (Citation2007).
8 In an unreported check, we also test an alternative measure of profitability EBITDA/Assets as in Hartzell et al. (Citation2014). The coefficients on the diversification measures maintain their direction of effect and the statistical significance.
9 We recognize that the book value of the properties does not reflect their market value, but we cannot observe the individual property’s market value.
10 See, for example, Titman et al. (Citation2005) and Huang and Huang (Citation2012).
11 It is worth mentioning that the lifecycle hypothesis of REITs suggests that REITs with access to public debt (arm’s length debt) are more advanced and well-established than their counterparts that depend solely on private channels (Jiménez et al., Citation2006; Colla et al., Citation2013; Letdin, Citation2017). Hence, the sample is expected to represent REITs that are well established and willing to pay a quasi-entrance tax fee for the public debt market.
12 The NCREIF divides the US commercial real estate market into eight regions: Northeast, Mideast, Southeast, East North Central, West North Central, Southwest, Mountain, and Pacific.
13 We merge the debt with firm- and property-level data using ultimate parent CUSIP.
14 The –HHI is bounded by 0, perfectly diversified, and –1, perfectly concentrated.
15 We do acknowledge that geographical diversification doesn’t perfectly correlate with long distance. Some geographical locations can be in a close proximity.
16 We test (F-test) whether the sum of the coefficients on both dimensions of diversification is equal to 0 and fail to reject this hypothesis.
17 Since the performance (betas) of markets is cyclical, a more general specification would allow these region and property- type fixed effects to interact along with a year indicator rather than assuming that the fixed effects are additive. Unfortunately, we do not have enough degrees of freedom to estimate such a model.
18 We obtain the same results if we measure diversification across five property types
19 We obtain the same results if we measure diversification across the US states
20 The credit rating is not available for all debt issues in our sample.