From the ground up
Unexpected results occur regularly in our experiments, but we try to understand why. The lower-yielding REITs obviously are that way because their prices are higher relative to their dividends than the higher-yielding REITs. And since all REITs have the same profit-distribution requirement, there can be no individual payout policies affecting the results.
In looking at the yields through the years, we saw that the differential between the top and bottom halves averaged 2.59 percentage points. That meant the price appreciation part of total return had to be especially strong to overcome the yield gap.
By the way, the "lower" yielding REITs still boasted an average annual yield over the 8 3/4 years of 5.15%. The upper half had an average yield of 7.74%. In comparison, the Dow Jones Select Dividend Index, which tracks big dividend-paying companies but no REITs, averaged 4.03%. No wonder REITs appeal to many investors.
What explains why higher-priced REITs climbed faster than lower-priced REITs?
1. Momentum. Many investors when looking at a market segment for the first time gravitate toward the securities that seem to be leading the parade. And the period we examined saw lots of new investors and new capital move into the real estate space. But momentum, which doesn't usually work in reverse, can't account for lower-yielding REITs falling less so far this year than higher-yielding ones.
2. Size. We divided the index's components into "large" -- those comprising 70% of the overall market cap -- and "small." The track record of the large REITs nearly matched the lower-yielding REITs. On an annualized basis, large REITs returned 12.28% versus 11.18% for small.
3. Property type. Apartment buildings, shopping malls and strip centers weighed more heavily in the large category, while offices, mixed industrial-office and hotels were the heavyweights in the small grouping.
Assessing value
We'd discovered ways to divide REITs into segments that perform differently. Was that enough to justify building indexes to track them?
There was one more item to check -- how much the components of the two indexes switched back and forth between the segments. This data, which we call turnover, matters because if mutual funds or exchange-traded funds were based on such indexes, the trading costs could be prohibitive with high turnover.
Sadly, turnover was terrible. The higher-yielding REIT segment had average annual turnover by market cap during the period of 56.22%, double the 28.15% for the lower-yielding segment. By contrast, turnover for the DJ Wilshire U.S. REIT index -- measuring just those REITs added and dropped regardless of yield -- averaged 5.09%.
It was somewhat better for the size segmentation. The grouping of large REITs averaged turnover of 9.39% a year, versus 22.48% for small.
In the end, though, we concluded there wasn't a compelling reason for an index on this theme. But for those investing in individual REITs, our research suggests buying them big and relatively lower-yielding.
John Prestbo is editor and executive director of Dow Jones Indexes, a unit of Dow Jones & Co., Inc., publisher of MarketWatch. Clifton Dy and Cynthia Lambert contributed research to this article.
