REITs and inflation: A long-run perspective

Abstract. We examine whether REITs provide an inflation hedge in the long run. We also investigate whether the apparent lack of a positive relationship between general prices and REIT returns in prior studies arises from the impact that stock market movements have on REITs. As in most prior research, regression analysis provides no evidence that REIT returns are positively related to temporary or permanent components of inflation measures. We rule out the possibility that a stock market-induced proxy effect is the cause for the apparent lack of relationship between REITs and inflation. On the other hand, we find some evidence that REITs provide a long-run inflation hedge. Johansen (1988) tests for cointegration isolate cointegrating vectors between alternate REIT indices and the CPI over the 1972-95 interval. However, the more standard residual-based cointegration techniques failed to provide similar evidence.

Introduction

Several researchers have suggested that real estate investment trusts (REITs) tend to behave like other equities with respect to their inflation-hedging characteristics. It is now well documented that stock returns in the United States and several other countries are either unrelated or negatively related to inflation, inconsistent with the Fisher (1930) hypothesis (e.g., Bodie, 1976; Jaffe and Mandelker, 1976; Fama and Schwert, 1977; Fama, 1981; Geske and Roll, 1983; Mandelker and Tandon, 1985; and Stulz, 1986). Most studies on the relationship between REIT returns and inflation arrive at similar conclusions (Murphy and Kleiman, 1989; Chan, Hendershott and Sanders, 1990; Park, Mullineaux and Chew, 1990; and Yobaccio, Rubens and Ketcham, 1995). Only a few studies, such as Gyourko and Linneman (1988) and Chen and Tzang (1988), indicate that REITs possess some inflation-hedging properties. Gyourko and Linneman document that REITs provide a partial hedge against the inflation rate derived from the Consumer Price Index (CPI) adjusted for the Home Purchase Price component. However, the authors also find that REITs act as a perverse hedge against unexpected inflation. Chen and Tzang find that REITs have some ability to hedge the expected component of inflation.1

The evidence from unsecuritized real estate has been far more favorable. For instance, Hartzell, Heckman and Miles (1987) document that a portfolio of commercial real estate provided an effective hedge against inflation over the 1973-83 interval.2 An earlier study by Fama and Schwert (1977) suggested that residential real estate was a complete hedge against the expected and unexpected components of inflation. More recently, Rubens, Bond and Webb (1989) find that residential, commercial and farmland real estate provide at least partial hedges against inflation. The authors also find improvements in the hedging effectiveness of portfolios once real estate is included. Similarly encouraging results are obtained in Brueggeman, Chen and Thibodeau (1984), Ibbotson and Siegel (1984) and Miles and Mahoney (1997). In sum, the evidence on the inflation-hedging potential of real estate is very different across studies that employ unsecuritized real estate and those that employ securitized real estate.3

These differences in the evidence for unsecuritized real estate and REITs give rise to two interesting, albeit related, questions. First, is it only smoothing bias that causes the divergence in the evidence, or could it also be the result of the behavior of REITs themselves? It is now well known that REITs have a substantial stock market component, and as widely documented, stock returns have tended to be negatively related to the rate of inflation. Second, is there a longer-run relationship between inflation and REITs that standard econometric techniques fail to capture? Real estate markets are prone to long boom-and-bust cycles that are known to be out of sync with other markets (Grenadier, 1995). Out of sync cycles and the market component in REITs may be obfuscating a long-run real estate-inflation relationship. This line of reasoning is bolstered by prior evidence that there exists a temporally unstable relationship between expected inflation and REIT returns (Chan, Hendershott and Sanders, 1990).

We address the above two questions in this article. We also extend prior studies by conducting standard regression tests on the short-run relationship between REIT returns and the permanent and temporary components of inflation. However, the inflation decomposition techniques employed here differ from prior efforts in that we allow for a stochastic trend component in the inflation proxies. Prior findings of a weak and even perverse relationship between REIT returns and inflation continue to be supported by the standard regression analysis. Further, we find no evidence that the apparent lack of relationship between REITs and inflation is attributable to the stock market component in REITs.

We conduct Johansen (Johansen and Juselius, 1990) cointegration tests to examine the long-run relationship between REITs and inflation. We find evidence of cointegration when employing the CPI as the proxy for inflation. We find much weaker evidence of cointegration between the REIT index levels and the three-month or the one-year Treasury bill rate, the other proxies for inflation. Moreover, there is no evidence of cointegration between REITs and inflation, however measured, when we employ the more traditional residual-based cointegration methods. We suggest that the mixed evidence from the cointegration tests are symptomatic of fractional cointegration. In this scenario, REITs provide only a partial hedge against inflation in the long run.

Data

The investigation employs monthly and quarterly indices for all REITs (AREIT), equity REITs (EREIT), mortgage REITs (MREIT) and hybrid REITs (HYREIT), along with inflation proxies over the interval January 1972 through December 1995.(4) Following prior research, we employ the CPI and Treasury bill rates to derive proxies for inflation (see Park, Mullineaux and Chew, 1990; and Yobaccio, Rubens and Ketcham, 1995).

Exhibit 1 provides correlations and unit root tests for the series examined. Panel A presents the Pearsons coefficients for correlation between monthly REIT returns and alternate measures of inflation. The REIT returns and inflation rates are given by ((index^sub i,t^ - index^sub i,t-1^)l index^sub i,t-1^)*100), where the REIT and CPI indices are benchmarked 1972/01 = 100. The trailing one-year inflation rate (annualized) is employed to provide indications on the relationship of REIT returns and a slower-- adjusting measure of inflation. The correlations suggest that the return and inflation rate proxies are negatively related or, at best, unrelated. The coefficients for the CPI rate are consistently negative, notably so for EREIT returns.

Panel B reports the Phillips-Perron (Phillips and Perron, 1986; and Phillips, 1987) t-- Statistics that test for the null of nonstationarity, with the alternate that the series has no unit root or that the series is I(0).5 We cannot reject the null for any of the series considered. On the other hand, the statistics suggest the series are stationary in their first differences. The t-Statistics consistently reject the null that the differenced series are nonstationary at the .01 level. In other words, the level series may be appropriately described as integrated of order 1 or I(1), and therefore, are appropriate for the deployment in standard cointegration tests.6 The characteristics of If ) and I(0) processes are discussed further later in the article.

However, the results are weaker for the other REIT series. For instance, the statistics for the quarterly MREIT and HYREIT series with quarterly seasonal dummies are not significant at the .10 level. Moreover, the tests fail to provide any evidence of a longrun relationship between the REIT indices and the one year T-bill rate. In Panel B, the trace statistics and the maximal eigenvalues fail to reject the null of at most one cointegrating vectors.

Exhibit 5 reports results from the Johansen procedures from the model with trend. The test statistics and their significance levels are conspicuously different from those in Exhibit 4. For instance, in contrast to the results in Exhibit 4, the trace statistic for the monthly AREITs-CPI series is barely statistically significant, and that for the EREIT-CPI series is not significant. Interestingly, Exhibit 5 provides some evidence of cointegration between REITs and the one-year T-bill rate. In sum, the cointegration results are fairly sensitive to the trend term in Equation (5).

While the Johansen tests provide some encouraging results (of cointegration), it should be noted that the more standard, residual-based Dickey-Fuller and Phillips tests for cointegration failed to provide any indication of cointegration between either the REITs and CPI, or between the REITs and the one year T-bill rate. For the sake of brevity we do not present these test results.14 However, to highlight the differences across the procedures, we provide in Exhibit 6 a summary of the evidence from the regression and cointegration tests presented so far.

The inconsistencies among the cointegration results do raise other questions regarding the degree of comovements between the REIT and CPI series. The residual based tests for cointegration (Dickey-Fuller and Phillips tests) are predicated on the assumption that the cointegrating vector is I(0). Among others, Baillie and Bollerslev (1994) consider the possibility of a form of cointegration that exists with the cointegrating vector being I(d), d

We plot the autocorrelations for error correction terms, z^sub 1,t^ = (AREITs index^sub t^ - beta(CPI index^sub t^)) and z^sub 2,t^= (AREITs index - beta(T-bill Rate^sub t^)), where z^sub 1,t^ (z^sub 2,t^) is obtained from the monthly ordinary least squares regression of the AREITs index on a constant term and the CPI (one-year T-bill rate). These plots are presented in Exhibit 7. For comparative purposes, we also plot the functions for the AREIT series and for the first difference in the AREIT series.

The autocorrelation function for the AREIT series has a slow decay process that is typical of an I(1) series. On the other hand, the first differenced function is obviously I(0). The z^sub 1,t^ autocorrelations has a slow decay relative to the function representing the first differenced series. However, these autocorrelations exhibit long-term cycles, similar to function demonstrated by Baillie and Bollerslev (1994) in their study of exchange rates. The z^sub 1,t^ function is also in marked contrast to the I(1) AREIT series, making a case for further investigation into the possibility of fractional cointegration between REITs and general price levels. The z^sub 2,t^ function makes a weaker case for cointegration or fractional cointegration between the AREIT index and T-bill rate, consistent with the results in Exhibits 4 and 5.

Conclusion

This article examines the inflation-hedging characteristics of REITs over the 1972/ 01-1995/12 interval. Regression estimations in the more traditional mold fail to provide evidence that REITs are effective inflation hedges. We also rule out the possibility that a stock-market induced `proxy effect' causes the apparent lack of regression relationship between REITs and inflation.

We find some evidence of a long-run relationship between the CPI and alternate REIT indices when employing Johansen cointegration tests. However, while the Johansen tests provide some encouraging results with respect to the long-run relationship between the CPI and the REIT indices, the overall evidence of cointegration between REITs and inflation is tenuous: very weak evidence of cointegration is noted between REITs and the T-bill rate when employing the Johansen procedures, and more traditional tests for cointegration (e.g., Phillips and Dickey-Fuller) fail to indicate any evidence of cointegration between REITs and inflation, however measured. We make a case for an investigation into the possibility of fractional cointegration between REITs and general price levels. Fractional cointegration would imply that REITs provide only a partial hedge of inflation in the long run.

It is widely recognized that any asset class in a portfolio should not provide a negative real return over the long run. However, portfolio managers seeking to hedge inflation risk must seek out assets that not only provide positive real returns on average, but also those that are positively related to general price levels. Convincing evidence of a long-term general price component in the REIT indices would thus have had important implications to such portfolio managers. The implications of a weaker form of cointegration between real estate and general prices remains to be examined.

Notes

1 For a review on the evidence of the REITs-inflation relationship and the models employed to examine this relationship, see Yobbaccio, Rubens and Ketcham (1995).

2 The authors also provide an excellent review of literature on the inflation-hedging capacity of commercial real estate.

3 It should be noted, however, that the inflation hedging effectiveness evidenced in the group of studies that employ unsecuritized real estate data has often been attributed to appraisal-- smoothed biases common to such data series (Geltner, 1989; and Giaccotto and Clapp, 1992).

4 Monthly REIT indices are from the 1996 NAREIT Fact Book, Washington, DC. Other data series are from the Pinnacle Data Corporation, Webster, NY.

5 The Phillips-Perron test statistics are obtained from employing the equations (without trend and with trend (t)):

(Delta)Y^sub t^ = alpha^sub 0^ + alpha^sub 1^Y^sub t-1^ + epsilon^sub 1^

(Delta)Y^sub t^ = alpha^sub 0^ + alpha^sub 1^Y^sub t-1^ + alpha(Delta)Y^sub t^ = alpha^sub 0^ + alpha^sub 1^Y^sub t-2^ + epsilon(Delta)Y^sub t^ = alpha^sub 0^ + alpha^sub 1^Y^sub t-1^.

The test statistics (on al) are transformed to remove the effects of serial correlation on the statistics' asymptotic distribution as proposed by Phillips and Perron (1986) and Perron (1988: 308-9). The critical values are: with trend: -3.13 (10%), -3.41 (5%), -3.96 (1%), without trend: -2.57 (10%), -2.86 (5%), -3.43 (1%).

6 For a review of the concepts of integration/cointegration, see Engle and Granger (1987).

7 See, for example, Kydland and Prescott (1990) and Backus and Kehoe (1992).

8 The deployment of the three-month T-bill rate provided very similar results.

9 The results from the alternate decompositions are available from the author.

10 To clarify, each regression of forty-eight sets of observations produced a regression coefficient that was employed as the hedge ratio pertaining to the last month in that subsample.

ll Similar results are obtained for the other REITs in the sample.

12 Johansen and Juselius term II the 'long-run impact matrix.'

13 The sensitivity of the Johansen procedures across models with and without trend has also been noted by Baillie and Bollerslev (1994) and Diebold, Gardeazabal and Yilmaz (1994).

14 The results from the Phillips and Dickey-Fuller cointegration techniques are available from the author.

15 Standard unit root tests (e.g., Phillips-Perron) are known to have very low power against fractional alternatives (Diebold and Rudebusch, 1991).

References

Backus, D. K. and P. J. Kehoe, International Evidence on the Historical Properties of Business Cycles, American Economic Review, 1992, 82:4, 864-88.

Baillie, R. T. and T. Bollerslev, Cointegration, Fractional Cointegration, and Exchange Rate Dynamics, Journal of Finance, 1994, 49, 737-45.

Brueggeman, W., A. Chen and T. Thibodeau, Real Estate Investment Funds: Performance and Portfolio Considerations, The Journal of the American Real Estate and Urban Economics Association, 1984, 13, 333-54.

Bodie, Z., Common Stocks as a Hedge Against Inflation, Journal of Finance, 1976, 31, 45970.

Chan, K. H. Hendershott and A. Tzang, Interest-Rate Sanders, Risk and Return on Real Estate: Investment Trusts, Journal of from Equity REITs, The Journal of the American Real Estate and Urban Economics Journal of Finance, 1990,18, 431-52.

Chen, K. and D. Rudebusch, Long Memory and Persistence in Aggregate Sensitivity of Real Estate Investment Trusts, Journal of Journal Estate Monetary Economics, 1988, 3, 13-22.

Diebold, F. and C. W. J. Gardeazabal and K. Yilmaz, On Cointegration and Exchanger, Long Run Economic Relationships, Econome Dynamics, Fama, E. F., Stock Returns, Real Activity, Inflation, and Money, American Economic Review,

Diebold, F. X. and G. D. Rudebusch, Long Memory and Persistence in Aggregation, Journal of Financial Economics, Fisher, I., The Theory of Interest, New York: Monetary Economics, 1989, 24, 189-209.

Engle, R. F. and C. W. J. Granger, D., Bias in Appraisal-Based Returns, The Journal of the American Relationships, Econometrica, and Urban Economics Association, 1989, 251-76.

Geske, R. and R. Roll, The Fiscal and Monetary Linkage between Stock Returns, Real Activity, Inflation, and Money, American Economic Review, Journal of Finance, 1981, 77, 545-65.

Fama, E. F., and G. W. Schwert, Appraisal-Based Real Estate Returns and Alternative Journal of Financial Economics, 1977, 5,115-46.

Fisher, I., The Theory of Interest, New York: MacMillan, 1930. Geltner, D., Bias in Appraisal-Based Returns, The Journal of the American Real Estate and Urban Economics Association,1989, 17, 338-52.

Geske, R. and R. Roll, The Fiscal and Monetary Linkage between Stock Returns and Inflation, Journal of Finance,1983, 38, 49-65.

Giaccotto, C. and J. Clapp, Appraisal-Based Real Estate Returns Under Alternative Market Regimes, The Journal of the American Real Estate and Urban Economics Association, 1992,

20, 1-25.

Granger, C. W. J. and P. Newbold, Spurious Regressions in Econometrics, Journal of Econometrics, 1974, 2, 111-20.

Grenadier, S. R., The Persistence of Real Estate Cycles, Journal of Real Estate Finance and Economics, 1995, 10, 95-119.

Gyourko, J. and P. Linneman, Owner-Occupied Homes, Income-Producing Real Property, and REITs as Inflation Hedges, Journal of Real Estate Finance and Economics, 1988, 1, 347-72. Hartzell, D., J. S. Hekman and M. E. Miles, Real Estate Returns and Inflation, The Journal of the American Real Estate and Urban Economics Association, 1987, 15, 617-37. Hodrick, R. J. and E. C. Prescott, Postwar U.S. Business Cycles: An Empirical Investigation, Discussion Paper 451, Carnegie-Mellon, 1980.

Ibbotson, R. and L. Siegel, Real Estate Returns: A Comparison with Other Investments, The Journal of the American Real Estate and Urban Economics Association, 1984, 12, 219-42. Jaffe, J. and G. Mandelker, The Fisher Effect for Risky Assets: An Empirical Investigation, Journal of Finance, 1976, 31, 447-58.

Johansen, S., Statistical Analysis of Cointegration Vectors, Journal of Economics Dynamics and Control, 1988, 12, 231-54.

Johansen, S. and K. Juselius, Maximum Likelihood Estimation and Inference on CointegrationWith Applications to the Demand for Money, Oxford Bulletin of Economics and Statistics, 1990, 52, 169-210.

Kydland, F. E. and E. C. Prescott, A Fortran Subroutine for Efficiently Computing HP-filtered Time Series, Federal Reserve Bank of Minneapolis Research Memorandum, April, 1989. -, Business Cycles: Real Facts and Monetary Myth, Federal Reserve Bank of Minneapolis Quarterly Review, 1990, Spring, 3-18.

Liang, Y., A. Chatrath and W. McIntosh, Apartment REITs and Apartment Real Estate, Journal of Real Estate Research, 1996, 11, 277-89.

Mandelker, G. and K. Tandon, Common Stock Returns, Real Activity, Money and Inflation: Some International Evidence, Journal of International Money and Finance, 1985, 4, 267-86. Miles, M. and J. Mahoney, Is Commercial Real Estate an Inflation Hedge, Real Estate Finance, 1997, 13:4, 31-45.

Murphy, J. and R. Kleiman, The Inflation-Hedging Characteristics of Equity REITs: An Empirical Study, Quarterly Review of Economics and Business, 1989, 29, 95-101. Newey, W. and K. West, A Simple, Positive Semi-definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix, Econometrica, 1987, 55, 703-8. Park, J. Y., D. J. Mullineaux and I. Chew, Are REITs Inflation Hedges? Journal of Real Estate Finance and Economics, 1990, 3, 91-103.

Phillips, P. C. B., Time Series Regression with Unit Roots, Econometrica, 1987, 55, 277-302. Phillips, P. C. B. and P. Perron, Testing for Unit Roots in Time Series Regression, Discussion Paper, New Haven, CT: Yale University, 1986.

Rubens, J., M. Bond and J. Webb, The Inflation-Hedging Effectiveness of Real Estate, Journal of Real Estate Research, 1989, 4, 45-56.

Stulz, R., Asset Pricing and Expected Inflation, Journal of Finance, 1986, 41, 209-24. Yobaccio, E., J. Rubens and D. C. Ketcham, The Inflation-Hedging Properties of Risk Assets: The Case of REITs, Journal of Real Estate Research, 1995, 10, 279-96.

*School of Business Administration, University of Portland, Portland, OR 97203 or chatrath@uofport.edu.

**Prudential Investments, Parsippany, NJ 07054 or youguo.liang@prudential.com.

Copyright American Real Estate Society 1998
Provided by ProQuest Information and Learning Company. All rights Reserved