Dividend Pricing Models and REITs
Kallberg, Jarl G., Liu, Crocker H., Srinivasan, Anand, Real Estate Economics
Dividend pricing/present value models relate current stock prices to expectations of future dividends. In this study we apply the West and Campbell Shiller tests of the dividend pricing relation to an index of real estate investment trusts (REITs). REITs provide a unique test of these models since, during our study period, REITs were mandated to pay out at least 95% of taxable income as dividends. While our results complement previous research which finds that the dividend pricing model cannot be rejected if share repurchase is included as part of dividends, our data contain a much less significant amount of share repurchase, so that our approach to the issue of the viability of dividend pricing models offers an alternative insight. Our research suggests that, for our REIT population, dividend pricing models cannot be rejected.
Empirical research has found that dividend pricing models, or present value models, when tested on market indexes, are poor predictors of true prices. The literature (1) initially cited this lack of fit as evidence of "excess volatility," the notion that prices move too much to be explained by changes in fundamental factors such as dividends and discount rates. More recently, however, this research has been recast as a rejection of the pricing model, in particular, the failure of the assumption of a constant dividend rate to hold. The viability of dividend pricing models has important implications for market efficiency, market microstructure and asset pricing.
The majority of the research uses the following simple present value relation, which forms the basis of the valuation of financial and real assets:
(1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII],
where [P.sup.*.sub.t] is the "ex post rational" price at time t, [[beta].sup.i] is the reciprocal of the discount rate for cash flows i periods in the future and [d.sub.i+t] is the payoff i periods from time t. In these models, payoff is usually synonymous with future dividend payment, although a few authors have taken a broader definition. The intuition of Equation (1) is that in an efficient market a firm's stock price is the discounted value of its future dividend stream.
This article assesses the validity of dividend pricing/present value models by focusing on real estate investment trusts (REITs). While the characteristics of REIT dividend payments are clarified below, it is important to emphasize a fundamental distinction between REITs and regular corporations. REITs are unique in that they typically pay out almost all of their funds from operations as dividends. (2) This dividend payout regime stems from their legal obligation to pay out at least 90% of their taxable income as dividends; this figure was 95% prior to 2000. This removes much (but not all) of the firm's discretion over its dividend payout and should make the dividend pricing/present value relation more plausible than for typical equities or equity indexes. While West (1988a), Campbell and Shiller (1987, 1988a, 1988b) and others reject models based on Equation (1), Ackert and Smith (1993), by broadening the definition of dividends to include share repurchases and other distributions, cannot reject the dividend pricing model.
Our study is much in the spirit of Ackert and Smith's (1993) work. However, in our data set, share repurchases are not nearly as significant a factor. Rather, the driving force is the difference in the manner in which REITs versus typical equities determine dividend payouts. Like Ackert and Smith, we find support for dividend pricing models. Knowing that these models fit REITs well provides important insight into the pricing dynamics of REITs versus other equities.
The outline of this study is as follows. The second section briefly reviews the literature. The third section describes the relevant characteristics of REIT dividend policies and our data set. The section following describes the econometric results, and the final section presents our conclusions. …