![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
Real estate investment trusts (REITs) provide a unique setting to study the impact of aggregate funding on momentum. REITs provide many of the same features of common stocks, but also include a dividend provision mandate, which may mitigate the illiquidity concerns that drive momentum patterns in common stocks. Overall, we found that REIT momentum was associated only with restrictive funding conditions. However, unlike in common stock, this effect can be explained by particularly strong performance of winner REITs during restrictive periods. Furthermore, we found that the strong performance of winner REITs in restrictive states was related to strong dividend performance. These results suggest that the unique income provision feature of REITs generates momentum patterns as aggregate funding availability is diminished.
Acknowledgments
We thank John Crespi, Anil Kumar, session participants at the 2019 American Real Estate Society National Conference in Paradise Valley, and three referees for valuable comments. We thank Editor Hardin and the editorial team for their insight and guidance.
Notes
1 For example, Chui et al., Citation2003a, Citation2003b; Derwall et al., Citation2009; Goebel et al., 2013; Hung & Glascock, Citation2008; Ling et al., Citation2019, though it is not the focus of their research. Parhizgari and Pavlova (Citation2010) examined momentum trading strategies in a multicountry setting.
2 Recent evidence suggests that REIT dividend payouts are high and stable, in part due to use of dividend reinvestment plans (Bond et al., Citation2019), even in the presence of retained cash shortages (Riddiough, Citation2019). Furthermore, Kallberg et al. (Citation2003) suggest that high past dividends provide positive signals to investors about future income provision.
3 Though their setting differs from ours, DiBartolomeo et al. (Citation2021) provide support for the notion that investors value REITS for their high mandated dividend payouts to enable increased liquidity.
4 The strong income provision (dividend) pattern in REITs creates a potential misspecification problem when evaluating momentum in REITs. Specifically, it is possible that prior returns influence future returns explicitly due to dividends if total return measures are used to assess both prior and future returns. We thank Editor Hardin and an editorial team for inquiring if REIT momentum may be tautological when the return series includes dividends.
5 NAREIT estimates that 40% of U.S. households hold REITs in their investment portfolios either directly or indirectly though other investment vehicles such as mutual funds; estimates are based on several data sources, including the 2016 Survey of Consumer Finances. See https://www.reit.com/data-research/research/nareit-research/80-million-americans-own-reit-stocks.
6 The early 1990s is viewed as the start of the “Modern REIT era” due to the legislation and industry structural changes that occurred in that time frame and altered REIT ownership structure, management style, and tax treatment (as discussed in Ambrose and Linneman (Citation2001), Chui et al. (Citation2003a), and Hung and Glascock (Citation2008)) and increased valuation uncertainty (Ling & Ryngaert, Citation1997).
7 However, a large literature examines monetary policy and the level of REIT returns. Chou and Chen (Citation2014) and Goebel et al. (Citation2013) provide excellent reviews.
8 For stock returns, short-term reversals are commonly observed in cross-sectional monthly return regressions and may reflect a correction to an overreaction in the prior month.
9 We also ran all specifications using the market index measured as the value-weighted return on the CRSP universe of stocks. As we detail in the data section, the key results were robust to the measure used. In addition, we utilized equal-weighted versions of both the CRSP market index and the Ziman REIT index as controls and found qualitatively similar results. Results are available from the authors on request.
10 To construct the Amihud (Citation2002) illiquidity measure, we computed the ratio of the daily absolute REIT return to its dollar volume and generated a monthly average ratio.
11 This procedure provides a proxy for most recent dividends that investors are likely to be aware of by the month of the REIT return. Since most dividends are issued on a quarterly basis, averaging over three months should capture only the most current dividend. Additionally, since dividends are typically announced at least a month prior to payment, any dividends paid in the subsequent month will likely be known by investors by the end of the current month.
12 There did also appear to be a positive momentum pattern in relatively small REITs in size quintile 2, which was driven by a marginally significant WML premium in expansive conditions. However, this pattern stems from exceedingly poor performance of relatively small REITs during expansive conditions (−0.014), a point estimate which is inconsistent with any prior explanation for momentum.
13 There is research examining REIT dividend policy variations related to the low-liquidity and volatile cash flow environment of the 2009 financial crisis (e.g., see Hardin & Hill, Citation2011; Sahin et al., Citation2020). As a robustness check, we therefore examined the time period post-financial crisis, from January 2010 to December 2015. Our findings were robust: momentum was present only during restrictive time periods; its coefficient estimate was sizable at 0.0254, yielding a marginal effect of .74 percentage points per month (per1 standard deviation increase in past returns). These results are reported in in the Appendix.
14 Given the similarity in these results, in the forthcoming tests we only included the results using the REIT index control. However, the results were qualitatively similar using the stock market return as a control and are discussed later in this section. These additional results are available from the authors on request.
15 As an additional robustness check, we estimated the specifications of Panel A with controls for REIT property type. The results are similar and available on request. We found that momentum was present only during restrictive funding conditions.
16 As noted in the literature review, Hung and Glascock (Citation2008) found momentum returns were higher in “Up markets.” Our findings could be interpreted as complementary in the sense that episodes of “Up markets” tended to be correlated with episodes economic expansion, and business cycles were correlated with interest rate regimes. Episodes of tight economic markets (high level of economic activity and tight labor markets) tended to be characterized by “Up markets” (average stock market returns over some time period are non-negative), and tightening funding conditions. We address this further in the robustness checks.
17 For example, differentiating Equation (2) with respect to winner status, applying the coefficient estimates reported for Panel A Model 5 and setting ILLIQ and dividend yield to their respective sample means yields: 0 or .7 percentage points per month, which is a large effect and statistically significant at the 99% level of confidence.
18 When stock returns were the dependent variable, the loser momentum was only statistically significant when dividend yield was near zero (10th percentile, ) under restrictive funding conditions. For the 10th percentile of dividend yield, the point estimate was –0.6 percentage points. Thus, at low dividend yield and restrictive funding conditions, past losers still lost, but that effect was gone for higher dividend levels.