The Vanishing January Effect: Evidence from REITs

Fahad Almudhaf
REAE 5311 Blog Post

The Vanishing January Effect: Evidence from REITs

Introduction:
Time patterns in stock returns are reported in a number of studies. Rozeff and Kinney (1976) find that the monthly return is the greatest during the month of January compared to other months of the year in NYSE. This is early evidence of the existence of seasonality in monthly stock returns. This anomaly is known as the “January effect” in the literature. Seyhun (1993) lists potential explanations for the January effect and classify them into two categories. The first group of explanations which is consistent with efficient markets includes omitted risk factors, seasonalities in the risk-return trade-off, insider trading hypothesis, risk and econometric mismeasurement problems. The second group of possible explanations is inconsistent with the efficient market and includes portfolio rebalancing/window dressing and tax-loss selling pressure.

Previous research has well documented many seasonal effects which exist in the stock market returns in the US and other international countries. Day of the week effect, holiday effect, January effect, and turn of the year effect are examples of such calendar effect anomalies. Such anomalies cast doubts to the “efficient market hypothesis”.

Our study examines seasonality in Real Estate Investment Trusts (REITs) returns in the US using more recent data collected from NAREIT. Our study covers the period from January 1972 to September 2008. Our empirical results support the existence of positive returns during the month of January relative to other months during the whole sample. However, the January effect disappears and turns into being statistically insignificant during the most recent period. The results of such a study are important for both investors and academicians. An investor can construct a trading strategy which uses observed seasonalities in real estate returns to earn excess returns. Such calendar anomalies contradict the “efficient market hypothesis” which is still debatable between financial economists.

Literature Review:
Few studies exist in the literature examining the existence of calendar anomalies in REITs compared with numerous studies covering calendar anomalies in stocks returns. Colwell and Park (1990) show that seasonality effects exist in real estate related investments as well. They document that average REIT returns on January are higher than the other months of the year in both equity and mortgage REITs. Their study covers the period from 1964 to 1986. They also find that January effect is much stronger in smaller REITs which is consistent with the inverse relation between size and January effect previously documented in stock returns by Banz (1981) and Reinganum (1983). Colwell and Park (1990) claim that the exact economic forces which drive this anomaly are undiscovered.

Redman, Manakyan and Liano (1996) extend the previous study by covering the period from 1986 through 1993. Their results show the existence of day-of-the-week effect, turn-of-the-month effect, and January effect in REITs. Average US REIT returns were higher in January and Friday through their sample. The results of their study provide evidence against the weak-form of efficient market hypothesis in REITs.

Friday and Peterson (1997) cover the period from 1974 to 1993. Consistent with previous studies, they observe January effect in all sizes of REITs and in all classification (Equity, Mortgage and Hybrid). They argue that the tax-loss selling hypothesis is more likely to cause the January effect in REITs than information effects. Their empirical results which support the Tax-loss selling hypothesis are consistent with inefficient REIT markets.

Cromwell, Friday and Yoder (2000) use stochastic dominance methodology to study return seasonality in equity REITs. They reconfirm the existence of the January effect in equity REITs. They also find that the January effect is stronger in small REITs than in large REITs. Their findings are inconsistent with the efficient market hypothesis.

More recently, Lee and Lee (2003) provide evidence that the January premiums decreased after the increase of institutional investment in the REIT market after the Revenue Reconciliation act of 1993. The decline in January premiums was lower in Mortgage REITs compared to Equity REITs. Their evidence is consistent with the tax-loss selling hypothesis and inconsistent with window-dressing hypothesis. They cover the period from 1972 to 2002. However, contrary to prior research Hardin et al. (2005) shows that the January effect turned to be statistically insignificant during the period 1994-2002.

Our study extends the previous literature by using more recent data from NAREIT all REIT. This will help us to know whether or not seasonalities in REIT returns still exist or if they vanished after being presented in the literature. We cover the period of 1972 through 2008. We then study sub-periods of the whole sample as a robustness check. In addition, we use the S&P REITs composite index to confirm our findings.

Methodology:
This study uses OLS as the estimation method. We run the regression for the following equation to test for REITs seasonal effects:

Rt = a1*D1t + a2*D2t + ... + a12*D12t (1)
Where Rt is the REIT return at time t. D(it) is the seasonal dummy variable which equals 1 if the return at time t corresponds to month i, and 0 otherwise. is the average monthly return on month i. D1 corresponds to January and D12 corresponds to December.

We then use the following regression equation to test for the January effect in US REITs:
Rt = C + a2*D2t + ... + a12*D12t (2)
Where the intercept C represents the average return for January and the coefficients (ai) indicates the difference in returns between the return of January and month i. A negative value of the dummy coefficients in equation (2) would be a proof of a January effect (Higher returns on the month of January relative to other months of the year).

Results:
Table 1 reports the results of tests from Equation 1 for calendar seasonal effects of US REITs during the whole sample. Results in Table 1 show evidence of statistically significant calendar seasonal effects in the months of January, March, July and December. We have positive statistically significant effects occurring in these four months.

Table 2 shows results of the estimated regression in equation 2 to test the existence of the January effect for the whole sample. Based on results shown in Table 2, average January returns exceeds average returns in all of the other months except March and December where the excess returns is statistically insignificant. This is inconsistent with previous results of researchers who found higher REITs returns in the month of January compared to all other months.

Tables 3 and 4 provide the results of equations 1 and 2 using sub-samples of our whole period. Results in these tables are much more interesting. More recent empirical evidence shows that January effect does not longer exist in US REITs. Returns in January are not significantly higher than other months anymore and that is consistent with Hardin et al. (2005). We see that it is statistically insignificant in the period 1993-2008.


Moreover, we have used another index to provide further support for our results. Using S&P REITs composite index covering the period from January 1990 to January 2008, we got similar results showing statistically insignificant positive returns in January. Table 5 shows results of equation 1 using the S&P REIT composite index. This confirms that US REITs return are not anymore higher in January compared to other months. Average REITs returns are positive and statistically significant in both May and October during our sample period and using S&P REITs composite index. These results are also consistent with the results of Hardin et al. (2005) who show that the January effect became statistically insignificant lately.

Using the Dow Jones Industrial Average (DJIA), Moosa (2007) finds that the January effect disappeared for stocks during the 1990-2005 period. Our results are consistent with his study. However, our study uses REITs indices and not stock indices to confirm the fact that the positive January effect in REITs has become statistically insignificant recently. This could happen because traders become more aware of such anomalies and begin exploiting them which cause the January anomaly to vanish.

Conclusion:
This study has tested for seasonality of monthly REITs returns in the US. It also investigated the existence of the January effect in US REITs. Our results show evidence of seasonality in US REITs returns. Results obtained indicate significant presence of the January effect (higher returns in January) for the whole period 1972-2008. Our whole period results are consistent with Colwell and Park (1990), Redman et al. (1996), Friday and Peterson (1997), Cromwell et al. (2000), Bley and Olson (2003), and Lee and Lee (2003).
However, our study confirms that more recently the January effect does not exist anymore which is consistent with Hardin et al. (2005).The publicity of such an anomaly in both academic literature and financial media may have caused the January premium in US REITs to vanish recently.

References:
Banz, R. (1981), “The Relationship between Return and Market Value of Common Stocks”, Journal of Financial Economics, 9, pp.3-18

Bley, J. and Olson D. (2003),”An analysis of Relative Return Behavior: REITs vs. Stocks” SSRN Working paper.

Colwell, P. and Park H. (1990), “Seasonality and Size Effects: The Case of Real Estate Related Investments” Journal of Real Estate Finance and Economics, 3, pp.251-259

Compton, W., Johnson, D. , and Kunkel , R. (2006) “ The turn-of-the-month effect in real estate investment Trusts (REITS)”Managerial Finance, Vol.32, No.12, pp.969-980

Cromwell, N., Friday, H., and Yoder J. (2000), “Equity REITs and the January Effect”, Journal of Alternative Investments, Spring 2000, V.2, I.4, pp.62-68

Fountas, S. and Segredakis K. (2002) “Emerging Stock market return anomalies: the January effect and the tax-loss selling hypothesis” Applied Financial Economics, Vol. 12, pp.291-299

Friday, H. and Peterson D. (1997), “January Return Seasonality in Real Estate Investment Trusts: Information vs. Tax-Loss Selling Effects”, The Journal of Financial Research, Vol. XX, No.1, pp.33-51

Hardin, W., Liano, K., and Huang G. (2005),”Real Estate Investment Trusts and Calendar Anomalies: Revisited”, International Real estate Review, V.8, No.1, pp.83-94

Lee M. and Lee M. (2003), “Institutional Involvement and the REIT January Effect over time”, Journal of Property Investment and Finance, Vol. 21, No.6, pp.435-449

Moosa, I. (2007) “The Vanishing January Effect”, International Research Journal of Finance and Economics”, Issue 7, pp.92-103

Redman, A., Manakyan, H., and Liano K. (1996), “Real Estate Investment Trusts and Calendar Anomalies” Journal of Real Estate Research, Vol. 14, pp.19-28

Reinganum, M. (1983) “The Anomalous Stock Market Behavior of Small Firms in January: Empirical Tests for Tax-Loss Selling Effects”, Journal of Financial Economics, 12, pp.89-104

Rozeff, M. S. and Kinney, W. R. (1976) “Capital seasonality: The case of stock returns”, Journal of Financial Economics 3, pp.379-402

Seyhun, H. (1993) “Can Omitted Risk Factors Explain the January Effect? A Stochastic Dominance Approach” Journal of Financial and Quantitative Analysis, 28, pp.195-212