Bachelor Thesis Finance ANR: Real Estate Securities as an Inflation Hedge Study program: Pre-master Finance Date:

Transcription

1 Bachelor Thesis Finance Name: Hein Huiting ANR: 097 Topic: Real Estate Securities as an Inflation Hedge Study program: Pre-master Finance Date: Abstract In this study, I reexamine the research of Glascock, Lu and So (00). I also have examined the relationship between REIT (Real Estate Investment Trust) returns and inflation. This I do by testing the causal relationship between the variables REIT returns, monetary policy (Federal Funds Rate), real activity and inflation through a Vector Error Correction Model (VECM). Subsequent to Glascock, Lu and So (00), the results provided from these test imply that the observations of REIT returns as perverse inflation hedges are spurious. Monetary policies (Federal Funds Rate) influence the relation between inflation and REIT returns in a negative way. These results and findings match with the results and findings of Glascock, Lu and So (00) and Darrat and Glascock (989). Darrat and Glascock (989) have provided evidence of monetary effects on REIT returns. Bachelor Thesis Finance Hein Huiting 8-0-0

2 . Introduction Several studies are dedicated to Real Estate Investment Trusts (REITs) as an inflation hedge. Glascock, Lu and So (00) find inflation hedging a major concern to real estate investors (e.g. insurance companies, pension fund managers). Reason for this phenomenon is that investors are interested in the real return. Glascock, Lu and So (00) suggest that common stocks represent claims on future consumption. Since stockholders give up present consumption for returns in the future, intuitively, rational investors will require higher returns during periods of high inflation in order to maintain the level of real payoffs. Moreover, Glascock, Lu and So (00) state inflation as a significant determinant in asset returns and it is also a key consideration in investment decisions. Liu, Hartzell and Hoesli (997) provide evidence that returns on the underlying real estate are positively related to both anticipated and unanticipated inflation. Geske and Roll (98) and Titman and Warga (989) find that stock returns are crucial to changes in fiscal and monetary policy. This explanation causes an opposite change in the rate of inflation. Furthermore, Liu, Hartzell and Hoesli (997) imply that a relationship should exist between current stock returns and future changes in the rate of inflation. Indeed, stock returns inclusive of real estate securities can provide inflation forecasts. Inflation hedging is a major concern to real estate investors, thus several studies are dedicated to the fact whether unsecuritized real estate can serve as an inflation hedge. Real estate investors find this important due to the fact that real estate investors have long-term investment holding periods (Glascock, Lu and So, 00). Gyourko and Linneman (988), Sirmans and Sirmans (987) and Hoag (980) find that there has been a consensus that unsecuritized real estate can serve as an inflation hedge. The reason for this assumption is that the underlying assets of REITs are primarily real estate and REITs are expected to be inflation hedges as well (Glascock, Lu and So, 00). Though, several studies have provided empirical results regarding REITs abilities. These results concerning hedge inflation are mixed. On top of that, studies tend to suggest that REIT returns have a negative relationship with inflation (Gyourko and Linneman, 988; Goebel and Kim, 989; Titman and Warga, 989; Park, Mullineaux and Chew, 990; Chen, Hendershott and Sanders, 990; Liu, Hartzell and Hoesli, 997). Glascock, Lu and So (00) suggest two implications regarding this conclusion of a negative relationship between REIT returns and inflation. REITs behave more like common stocks as perverse inflation hedges and secondly, REITs behaviors deviate from those of traditional real estate. Glascock, Lu and So (00) suggest several possible explanations to the fact that REITs are perverse inflation hedges. They find this counter-intuitive due to the fact that unsecuritized real estate is a good inflation hedge, while REITs are securitized forms of real estate. Firstly, results from spurious regressions in previous studies reverse the causal relationships between inflation and REITs returns. Secondly, Glascock, Lu and So (00) suggest that REITs are more efficient in information processing. They state that REITs reflect future economic prospects more efficiently and instantaneously than the general real estate sector. Since information occurs first in the REITs market, it is possible that REITs are more efficient in information processing. Moreover, Glascock Lu and So (00) suggest that the observed positive relationship between unsecuritized real estate and inflation could be caused by the usage of appraisal or transaction data that contain more recent rather than future information. Purpose of this research is to re-examine the relationship between REITs (real estate investment trusts) and inflation. Furthermore, this study investigates the link between fundamental economic activities (e.g. monetary policies and industrial production) and perverse inflation hedge (Glascock, Lu and So, 00). Following Glascock, Lu and So (00), I investigate whether securitized real estate can function as inflation hedges. Bachelor Thesis Finance Hein Huiting 8-0-0

3 Besides, this study investigates REIT returns and REIT movements regarding changes in inflation. Finally, I provide evidence for their relation with inflation taking monetary policy or real activities in consideration. These issues contribute to find whether or not REITs are perverse inflation hedges. For this investigation, the vector error correction model (VECM) of Glascock, Lu and So (00) will be used. With this model I will be able to study the potential causal relationships among REIT returns, inflation, industrial production and monetary policies. In summary, the results of this investigation show that price movements of REITs are greatly influenced by monetary policies (FFR). In addition, the relationship between REIT returns and inflation is negative and is the embodiment of the monetary policies (its effects of changes). Moreover, the outcomes of this investigation show that the phenomenon of REITs being perverse inflation are doubtful. This paper is organized as follows. First, the next section contains information regarding perverse inflation hedge and asset returns. In section the data and methodology is explained further. Section contains the empirical findings and discussions. Finally, section contains a conclusion. Bachelor Thesis Finance Hein Huiting 8-0-0

4 . Literature review In the past, several studies have been dedicated to the perverse hedge inflation of common stock returns. Glascock, Lu and So (00) have documented a number of studies in their research whether REITs are perverse inflation hedges or not. Since these studies have documented different results, it is necessary to enumerate and summarize their findings to come up with a clear view of the literature. The negative relationship between stock returns and inflation describes only a part of the relationships between real estate activities and stock returns (Fama, 98). In contrast to Fama (98), Geske and Roll (98) suggest that this relationship is notional and is induced by reversed adaptive inflation expectation. Furthermore, James, Koreisha and Partch (98) find that stock signals are significantly important to changes in real activities and in the monetary base. On top of those findings, Lee (99) finds that real economic activities can be explained by stock returns. However, Lee (99) also states that stock returns cannot explain inflation. Finally, Thorbecke (997) provided evidence that future cash flows or the discount rates are influenced by expansionary monetary policies. Glascock, Lu and So (00) state that stock returns will be affected, because future cash flows of the discount rates concern the present value of future earnings. In their paper Glascock, Lu and So (00) conclude that the observed perverse inflation hedge of common stocks actually reflects the relationships between inflation and other fundamental economic activities. In REIT market research, perverse inflation also occurs. Darrat and Glascock (989) investigated the effect of monetary and fiscal policies on investors perceptions. Darrat and Glascock (989) find that there is a relationship between the mortgage markets and the government bond markets. Since these markets have a relationship, impact of federal deficits have an effect on the mortgage market through the government bond markets. Furthermore, real estate and mortgage markets are linked as well. This means that real estate returns will be affected (Glascock, Lu and So, 00). Darrat and Glascock (989) provided evidence that market returns and base money have significant lagged effects on current real estate returns. Nevertheless, Glascock, Lu and So (00) find that inflation does not have a significant effect on real estate returns. Subsequent to real estate returns, inflation does not appear to have an effect on REIT returns. Geske and Roll (98) suggest that stock returns will be influenced by economic shocks, which affect real economic variables. Therefore, tax income, budget deficit, monetary base and inflation will be affected (Glascock, Lu and So, 00). Chen and Tzang (988) state that REITs returns are closely related to interest rates and Darrat and Glascock (989) find that federal deficits have important wealth effects on returns of securitized real estates. Therefore, Glascock, Lu and So (00) argue that economic shocks will have great impacts on REIT markets. Consequently, information about changes in monetary policies and inflation should be provided by REITs returns. Glascock, Lu and So (00) suggest that changes in industrial production will have impacts on the real estate sectors due to the fact that real activities reflect future economic prospect. Vacancy rates of real estates are expected to decrease when real output increases and business expands (Glascock, Lu and So, 00). Since the supply for real estate grows slower than the demand for real estate, prices and rental incomes will be pushed higher (Glascock, Lu and So, 00). Therefore, REITs returns will also increase. Fama (98) states that a decrease in inflation is caused by real output increases. Hence, Glascock, Lu and So (00) hypothesize that real activities should be able to forecast REITs returns and that real activity is negatively correlated with inflation. Besides, Glascock, Lu and So (00) suggest that the relationship between REITs returns and inflation should become weaker when real activity is taken into account. Bachelor Thesis Finance Hein Huiting 8-0-0

5 This paper is a re-examination of Glascock, Lu and So (00). Their investigation is based on Darrat and Glascock (989) and is extended by excluding real estate management and service related firms This study re-examines the issue of perverse inflation in the REIT markets. Darrat and Glascock (989) and Glascock, Lu and So (00) combine REITs, builders, real estate investment and management firms in their analysis. However, Glascock, Lu and So (00) state that the relationships among monetary policies, inflation and returns on securitized real estates are not clearly identified. Therefore, Glascock, Lu and So (00) include only REITs in their sample. Subsequently, this study investigates whether REIT returns signal changes in anticipated or unanticipated inflation or REIT movements are caused by expected or unexpected inflation. Finally, this study explains that when monetary policy or real activities are considered, how REIT returns, expected and unexpected inflation relates to one another. By answering these research questions, this study provides evidence whether or not REITs are perverse inflation hedges. Bachelor Thesis Finance Hein Huiting 8-0-0

6 . Data and methodology.. Data sources Subsequent to Glascock, Lu and So (00), data is used from the European Public Real Estate Association (EPRA) and concerns EPRA/NAREIT indices from several countries, namely REITs Total Return Index. Additionally, data is added from the Federal Reserve Bank (FRB) at St. Louis. This includes seasonally adjusted Consumer Price Index, Federal Funds Rate and seasonally adjusted Total Industrial Production Index. These indices are used as proxies for several performances. REITs Total Return Index for the performance of the REITs market, seasonally adjusted Consumer Price Index as a proxy for the performance of the inflation, Federal Funds Rate as a proxy for the performance of the monetary policy and seasonally adjusted Total Industrial Production Index as a proxy for the performance of real activities. Finally, as a proxy for the risk free rate, One-month Treasury Bill (TB) rates are obtained from the database of Fama and French. Table contains the specific data description. Table. Data description Content Source Sample Period Consumer Price Index FRB - St. Louis Dec. 970 Dec. 99 REITs Total Return Index NAREIT Jan. 97 Dec. 99 Federal Funds Rate FRB - St. Louis Jan. 97 Dec. 99 Total Industrial Production Index FRB - St. Louis Jan 97 Dec. 99 One Month Treasury Bill Rate Fama/French Dec. 970 Dec A preview of the data To clarify, the terms REIT, CPI, FFR, IP are, respectively, the rates of changes (first differences of natural logarithms) of REITs (Real Estate Investment Trusts), Consumer Price Index, Federal Funds Rate and Total Industrial Production Index. Table contains summary statistics and contains a correlation matrix. From this table, in panel A, evidence is provided that the Federal Funds Reserve (FFR) and the returns on REITs have the highest standard deviation. Ross (989) and Glascock, Lu and So (00) state that variance itself is a source of information. Therefore, the finding of high variances in changes of monetary policies and REIT returns imply that monetary policies and the REIT market have greater information content than other economic variables. Correlations among the variables REIT, CPI, FFR and IP are shown in panel B of Table. The information shown in this table explains a negative relation between REIT and FFR and CPI. This result is consistent with the perverse inflation hedge and the observation that high interest rate leads to lower stock returns. Furthermore, FFR is positively correlated with IP. The positive relation between FFR and IP indicates that when the economy is expanding, there will be pressures for interest rate increases. In contrast to Glascock, Lu and So (00), FFR is positively correlated with CPI. However, CPI is not significantly correlated with either FFR of IP. Subsequent to Glascock, Lu and So (00), the Fama and Gibbons (98) model is used to estimate expected and then unexpected inflation. This is in order to calculate inflation. Glascock, Lu and So (00) find that the Fama and Gibbons (98) model is commonly used NAREIT: National Association of Real Estate Investment Trust s. Bachelor Thesis Finance Hein Huiting

7 in the literature. Liu, Hartzell and Hoesli (997) find that this model dominates other inflationary proxies. Therefore, the Fama and Gibbons (98) model is used in this research. The difference between the TB rate and the expected real rate is defined as expected inflation. The expected real rate is the equally weighted moving average of the ex post real rates in the past twelve months. The ex post real rate is defined as the difference between TB rate and rate of change of the Consumer Price Index. Mathematically, expected inflation at time t,, is defined as: Furthermore, unanticipated inflation at time t,, is defined as the difference between actual inflation (rate of change in CPI) and expected inflation. Mathematically, unanticipated inflation at time t,, is defined as: In imitation to Glascock, Lu and So (00), the relationship among REIT returns, expected, unexpected and nominal inflation in the context of the Fama and Schwert (977) model is checked to ascertain that the data which is used are consistent with prior studies. Mathematically, the relationship among REIT returns, expected, unexpected and nominal inflation in the context of the Fama and Schwert (977) model is defined as: Table. Summary statistics of the rates of changes of the variables. Panel A: Summary statistics Variable Mean ( x ) Standard Deviation ( x ) Minimum Maximum REIT CPI FFR IP Panel B: Correlation Matrix REIT CPI FFR CPI ** FFR ** IP ** Notes: REIT, CPI, FFR, and IP denote, respectively, for the rates of change of the REITs market, Consumer Price Index, Federal Fund Rates and Industrial Production. ** Significant at the percent level. Bachelor Thesis Finance Hein Huiting

8 Consistent to Glascock, Lu and So (00), the previously defined equations (a and b) are estimated by ordinary least squares (OLS) with White s (980) correction for heteroskedasticity and the results are presented in Table. Panel A of Table provides information concerning REIT and contemporaneous inflation. The results in Panel A state that REITs are not complete inflation hedges as the coefficient is significantly different from one. Besides, the estimated value of is negative. This is consistent with the observed perverse inflation phenomenon. Findings concerning the contemporaneous relationship between REIT returns, expected and unexpected inflation are shown in Panel B of Table. Perverse inflation hedge occurs between anticipated and unanticipated inflation. These findings are similar to the findings in Panel A. Overall, these findings are also consistent with the evidence documented in the literature, e.g., Glascock, Lu and So (00) and Liu, Hartzell and Hoesli (997). Table. Results of the contemporaneous relationship between inflation and REITs returns. Panel A: REIT and Contemporaneous Inflation Parameter Estimated Value Standard Error t-statistic for = 0 Adjusted = ** ** Panel B: Fama and Schwert (977) Model Parameter Estimated Value Standard Error t-statistic for = 0 Adjusted = ** ** ** Notes: The equations are estimated by OLS with White s (980) correction for heteroskedasticity. * Significant at the percent level, ** Significant at the percent level.. Vector autoregressive model and causal relationships In their research, Glascock, Lu and So (00) state that before a test on causal relationship among variables can be conducted, the variables have to be checked whether they are stationary. Glascock, Lu and So (00) find that this is necessary, as the Vector Autoregressive (VAR) model requires the variables to be stationary. Furthermore, error correction terms have to be included in the VAR model if the variables are cointegrated. In imitation to Glascock, Lu and So (00), indices of expected and unexpected inflation are constructed and tests are performed on the stationarity and cointegration of the variables. Results of these tests are shown in Table. Results of unit root test for the variables are shown in Panel A of Table. The Phillips-Perron (PP) test is used for each series, due to that this test is robust to autocorrelation and heteroskedasticity. Bachelor Thesis Finance Hein Huiting

9 Table. Stationarity and cointegration test of REITs, inflation, monetary effects and real activities. Panel A: Stationary Test Level Series Return Series REIT Return Index Consumer Price Index Unexpected Inflation Index Expected Inflation Index Federal Fund Rate Industrial Production Index Panel B: Cointegration Results : Number of Variables in the System Cointegrating Vectors Trace REIT and Nominal Inflation 0 REIT, Expected and 0 Unexpected Inflation REIT, Expected Inflation, 0 Unexpected Inflation, Federal Fund Rate and Industrial Production Notes: The Phillips and Perron (PP) test with trend is used in testing for stationary of the variables. ** Significant at the percent level. Bachelor Thesis Finance Hein Huiting

10 . Empirical results and discussions. Causality relationships among REITs returns and other economic activities In order to measure the impact of inflation on REITs returns, a Vector Error Correction Model (VECM) is estimated among inflation variables and REITs returns. This is to isolate the influence of other economic variables. Thereafter, I incorporate effects of monetary policies and real activities. Two VECMs are estimated. Firstly, one VECM includes REIT returns and nominal inflation. Secondly, the other includes REIT returns, expected inflation and unexpected inflation. Mathematically, the VECM between REIT returns and nominal inflation is defined as: In these formulas, the error correction terms from the cointegrating regressions with REIT and CPI, respectively, are and. Bachelor Thesis Finance Hein Huiting

11 Table. VECM results for REITs returns, nominal inflation, Expected Inflation and Unexpected inflation. Panel A: REITs and Nominal Inflation Estimated Values of Independent Variables Dependent Variables Panel B: REITs Returns, Expected Inflation and Unexpected Inflation Estimated Values of Independent Variables Dependent Variables Notes: is the return of REITs at time t, CPI, is the rate of change in Consumer Price Index at time t, standard errors are in parentheses. * Significant at the percent level, ** Significant at the percent level. Fama (98) and Darrat and Glascock (989) state that monetary policies and real output have effects on the returns of stocks. Therefore, monetary policies and real output are included in the analysis in order to measure a possible change in the causal relationship among REIT returns and inflation, both expected and unexpected. Federal Funds Rate (FFR) is used as a proxy for the performances of monetary policies. Glascock, Lu and So (00) state that expansionary monetary policies will lead to higher price levels and that thereby interest rates and inflation are affected. Furthermore, Industrial Production is included to investigate whether real economic activities have influence in the causal relationship. Bachelor Thesis Finance Hein Huiting 8-0-0

12 Table 6. VECM results for REITs returns, Federal Funds Rate, Industrial Production, Expected and Unexpected Inflation Dependent Variables Coefficient Bachelor Thesis Finance Hein Huiting 8-0-0

13 . Impulse response analysis and innovation accounting Subsequent to Glascock, Lu and So (00), this investigation contains VAR innovation accounting to provide more insight in the speed and sources of transmissions of information. This, in order to clarify the forecast error variance decomposition (FEVD). The FEVD is explained as the percentage of forecast error variance explained by shocks, both own shocks and shocks in other markets. As mentioned earlier, the FFR, and thus monetary policies, have large impacts on the lead-lag relationship of FFR, EI, UEI, REIT and IP. In imitation of Glascock, Lu and So (00), the results are qualitatively the same of different ordering is used in the Choleski decomposition of the impulse response function. Table 7. Accounting innovations in REIT returns, Expected Inflation, Unexpected Inflation, changes in monetary effects and Industrial Production. By Innovations In Bachelor Thesis Finance Hein Huiting 8-0-0

14 Table 8. Individual impulse response functions to a unit shock in one market. Impulse Response In (cumulative) 0 (cumulative) (cumulative) (cumulative) 0 (cumulative) (cumulative) (cumulative) 0 (cumulative) (cumulative) (cumulative) 0 (cumulative) (cumulative) (cumulative) 0 (cumulative) (cumulative) Bachelor Thesis Finance Hein Huiting 8-0-0

15 . Conclusions Acknowledgments I am grateful to the helpful and constructive comments from MSc P.F.A. Tuijp from Tilburg University School of Economics and Management. All remaining errors, however, are my own responsibilities. Bachelor Thesis Finance Hein Huiting 8-0-0

16 References K. C. Chen, P. H. Hendershott, and A. B. Sanders, 990, Risk and Return on Real Estate: Evidence from Equity REITs, AREUEA Journal 8: -. K. C. Chen, and D. D. Tzang, 988, Interest-rate Sensitivity of Real Estate Investment Trusts, Journal of real Estate Research : -. A. F. Darrat, and J. L. Glascock, 989, Real Estate Returns, Money and Fiscal Deficits: Is the Real Estate Market Efficient?, Journal of Real Estate Finance and Economics : E.F. Fama, 98, Stock Returns, Real Activity, Inflation and Money, American Economic Review 7: -6. E.F. Fama, and M.R. Gibbons, 98, Inflation, Real Returns and Capital Investment, Journal of Monetary Economics 9, 97-. E. F. Fama, and G. W. Schwert, 977, Asset Returns and Inflation, Journal of Financial Economics : -6. R. Geske, and R. Roll, 98, The Fiscal and Monetary Linkage Between Stock Returns and Inflation, Journal of Finance 8: -. J. L. Glascock, C. Lu, and R. W. So, 00, REIT Returns and Inflation: Perverse or Reverse Causality Effects?, Journal of Real Estate Finance and Economics :, 0-7. P. R. Goebel, and K. S. Kim, 989, Performance Evaluation of Finite-life Real Estate Investment Trusts, Journal of Real Estate Research : J. Gyourko, and P. Linneman, 988, Owner-Occupied Homes, Income-Producing Properties, and REITs as Inflation Hedges: Empirical Findings, Journal of Real Estate Finance and Economics : 7-7. J. W. Hoag, 980, Towards Indices of Real Estate Value and Return, Journal of Finance : C. James, S. Koreisha, and M. Partch, 98, A VARMA Analysis of the Causal Relations Among Stock Returns, Real Output, and Nominal Interest Rates, Journal of Finance 0: 7-8. B.S. Lee, 99, Causal Relations Among Stock Returns, Interest Rates, Real Activity, and Inflation, Journal of Finance 7: C. H. Liu, David J. Hartzell, and M. E. Hoesli, 997, International Evidence on Real Estate Securities as an Inflation Hedge, Real Estate Economics : 9-. J. Y. Park, D. J. Mullineaux, and I. K. Chew, 990, Are REITs inflation hedges?, Journal of Real Estate Finance and Economics : 9-0. S. Ross, 989, Information and Vlatility: The No-Arbitrage Martingale Approach to Timing and Resolution Irrelevancy, Journal of Finance : -7. G. S. Sirmans, and C.F. Sirmans, 987, The Historical Perspective of Real Estate Returns, Journal of Portfolio Management : -. W. Thorbecke, 997, On Stock Market Returns and Monetary Policy, Journal of Finance : 6-6. S. Titman, and A. Warga, 989, Stock Returns As Predictors of Interest Rates and Inflation, Journal of Financial and Quantitative Analysis : 7-8. D. C. Quan, and S. Titman, 999, Do Real Estate Prices and Stock Prices Move Together? An International Analysis, Real Estate Economics 7: H. White, 980, A Heteroskedasticity-consistent Covariance Matrix Estimator and A Direct Test for Heteroskedasticity, Econometrica 8: I. de Wit, and R. van Dijk, 00, The Global Determinants of Direct Office Real Estate Returns, Journal of Real Estate Finance and Economics 6:, 7-. Bachelor Thesis Finance Hein Huiting