Dividend Aristocrats are listed stocks that have increased their dividend payouts for the past 25 years. Does it make sense to use a similar filter for real estate? Rogier van Aart  and Ruud Ramaekers investigate 

Some stock market sectors are known for their high dividends, such as telecoms and utilities. Real estate stocks, a subsector of financials, are also known for their yield. Since 1972 the average dividend yield on the FTSE EPRA/NAREIT US index has been 6.9%, while the S&P 500 has averaged 2.9%. 

Almost 60% of total return originates from dividends and reinvested dividends: $100 (€88) invested in the Global Property Research GPR250 United States price index in 1989 grew to $290 in October 2015, which is 4.2% annually. Including dividends, the investment would have grown to $1,256 (10.3% annually). Due to a short period, the importance of the dividend may be overstated.

Investors are attracted by the high and stable dividend of real estate. If this is a buying argument, we wonder if a portfolio that uniquely focuses on long and stable dividends will outperform the market index. This strategy is inspired by a common equity strategy: the S&P Dividend Aristocrats index. The stocks in this index have increased their dividends every year in the past 25 years. This index has beaten the S&P 500 with 2.3% since 1989. We adopt our strategy the ‘REIT Dividend Elite’. To study this strategy and some other strategies, we have set up a research project in co-operation with Global Property Research (GPR). 

The calculation of the REIT Dividend Elite strategy has been performed in three regions and the world: North America (AME), Europe-Middle East & Africa (EMEA), Asia Pacific (APAC) and the Global universe (GLOB). We look at the return of those strategies in local currency terms. We adopted a regional approach, to avoid any bias to a specific region. Earlier research, Van Aart (2015), showed that the strategy would yield only two to three names in Japan, so Japan was replaced by Asia Pacific.

On 31 December of each year, a selection is applied to the real estate universe of every region. The strategy looks at the cash dividends per share of the preceding years and selects those stocks that have been able to keep the cash dividends constant or to grow the cash dividends. The effects of a survivor bias are limited due to the method applied. A company that would go bankrupt right after selection, would remain in the basket in the rest of the year.

The REIT dividend elite strategy

Another element that is relaxed compared with the S&P Dividend Aristocrats strategy is the length of the dividend track record. Using a term of 25 years to review the dividend track record for real estate stocks would result in a portfolio too concentrated and too undiversified. Dependent on the region, the term is changed to keep a well-diversified portfolio. For North America, the strategy reviews the dividend track record for 15 years, for Europe 10 years and for Asia Pacific seven years. The stocks are equally weighted in a basket and rebalanced every quarter.

The strategies all have the same base year, to make the strategies’ performance comparable. Excess annual performance is 2.5% North America, 1.9% in Europe and 1.7% in Asia. The global strategy generated an annual outperformance of 2.5%. What is striking is the difference between average and median outperformance, which suggests the distribution of returns is influenced by outliers. Measuring the significance of the strategies we test for excess returns between the strategy and the regular benchmark to be positive, it shows that the p-value for the global strategy is lowest at 0.125. The other regional strategies have p-values that are a lot higher: 0.399 for the AME, 0.358 for EMEA and 0.382 for APAC. As the real estate sector has a much smaller universe than common equities, these portfolios are not always well diversified.

During some periods the number of stocks that qualify to be included in one portfolio drops significantly (for example during the global financial crisis in 2008/09), which erodes diversification. 

Exploring four alternative strategies

The results on the REIT Dividend Elite strategy are weaker than for the S&P Dividend Aristocrats index. There are several explanations. One is that the measurement period differs. Another is that the strategy employed by the original Dividend Aristocrats index is just an accidental finding, which is a typical Type I error (an incorrect rejection of the true null hypothesis). But, it may also have to do with the universe or sample size. Where the real estate is a specific equity sector with specific fundamentals, rules and regulations, the dividend strategy may not be applicable to this sector. In addition, diversification is worse. Where the S&P 500 index contains a universe of 500 stocks, the GPR consists of 250, making the starting universe a lot smaller. In addition, deeper analysis of the real estate data reveals that the number of selected stocks is low early in the sample.

To explore if other common equity strategies are applicable to real estate securities, we explore four other strategies. 

• High valuation versus low valuation (FFO)
• Large capitalisation versus small capitalisation
• Reverse price momentum
• High dividend versus low dividend (and corrected for payout)

Most strategies are well documented for common equities in academic research, but less so for listed real estate. Only for price momentum, Eichholtz & Meijer (2006) and Smith, Malin & Nair (2010) found that listed real estate displays a strong reversion pattern. Fama & French described in 1992 how two effects – the small cap effect and value effect – help to explain stock returns. Since then many adjustments and improvements have been made, but the conclusion has been very robust over time: small cap stocks and cheap stocks outperform.

The REIT dividend elite strategy figures 2 & 3

To explore alternative strategies, aside from the dividend elite strategies, we will explore how these strategies perform for the US listed real estate sector only. For every test the US universe is ranked into four quartiles (buckets 1-4). This region is chosen because it has long history, is large and has the best availability. The strategies rebalance on a quarterly basis, just like the GPR-indices do. Trading costs are excluded. 

High valuation versus low valuation (FFO)

The most commonly used valuation ratio is the P/FFO-ratio which is an EBITDA-type ratio. The funds from operations (FFO) takes the cash flow from operations and adds back depreciation and amortisation. The US universe is ranked according to their P/FFO value. Bucket 1 contains the stocks with the highest P/FFO-value.

This analysis shows that expensive REITs outperform cheap REITs using the P/FFO valuation metric by 1.36% per month. The p-value of this outperformance is 9.9%. This may sound counterintuitive, but can be explained by seeing the FFO-metric as a quality measure. The best real estate is usually the most expensive (lowest FFO-yield). Research from GreenStreet (2013) confirms this.

Large vs small capitalisation

The US universe is ranked according to market capitalisation. Bucket 1 would contain companies referred to as large caps, while bucket 4 would contain small caps. We find that small caps outperform large caps by almost 1% annually. The risk for small caps is higher, but only marginally. Looking at the risk-adjusted return, the ratio doubles from 0.13 to 0.26. It is interesting to note that the median return of large caps is almost as high as the median return of small caps, suggesting there is a strong skew in the returns. Looking at the 2008 global financial crises small cap stocks fell almost as much as large caps, but their recovery was stronger.

Looking at the excess return and the standard deviation of this return, we can calculate the z-score and the corresponding p-value. The p-value is 0.3%.

Reverse price momentum

Following the research of Eichholtz & Meijer (2006), and Smith, Malin & Nair (2010), we investigate whether the reverse price momentum effect is still intact. In fact, we investigate several momentum effects: one-month, three-month, six-month and 12-month price momentum. Bucket 1 contains the stocks with the best performance, while bucket 4 has the worse performers. 

The REIT dividend elite strategy figures 4 & 5

We are interested in the excess performance of 4 over 1 (which describes a reversal effect). The one-month momentum strategy generates an outperformance of 1.21% per month, while the 12-month momentum generates an outperformance of 0.72% per month. It should be remarked that the average excess return is a lot higher than the median excess return, which indicates that the return distribution is skewed. The z-score of the one-month momentum strategy is 0.6%. 

The statistics do not reveal the underperformance of this strategy during the global financial crisis in 2008. Bucket 1 underperforms bucket 4 by a large amount and in a matter of months, but all of it is recouped after the 2009 turnaround.

High vs low dividend

Now we investigate how a selection strategy of REITs purely on the back of dividend yield would perform. 

Bucket 1 contains the REITs with the highest dividends, bucket 4 the lowest dividends (REITs that pay no dividends are excluded). It is interesting to note there is no excess performance of high dividend versus low dividend stocks. The low dividend stocks show substantially lower standard deviation. 

Following the Credit Suisse Global Investment Returns Yearbook 2011, this dividend strategy may be adjusted for the pay-out ratio, and we are interested to see whether this holds for REITs as well. There is a difference with common equities in the sense that REITs have to pay out dividends and have little flexibility. The results (not displayed here) do not show consistent positive excess performance. 

Conclusion

Buying REITs with long and stable dividend track records gives positive excess returns in most cases, but p-values are high. Selecting REITs on the back of valuation (P/FFO), market capitalisation and the one-month reverse momentum yield large positive excess returns displaying low p-values. Selecting REITs on the back of dividend yield does not provide positive excess returns. The p-value of this strategy is high.

Ruud Ramaekers is portfolio manager at Aegon Asset Management, and Rogier van Aart is investment strategist at Univest Company