The German statesman Konrad Adenauer was quoted as saying “We all live under the same sky, but we don’t all have the same horizon.”
That’s certainly true of investors: We all live among the same set of assets, but how we combine them into an investment portfolio depends on our own circumstances—including our investment horizons.
To simplify, we might say there are two types of investors. The first type—call them tactical—pays attention to short-term fluctuations, hoping to identify opportunities to buy undervalued assets or to sell overvalued ones. Tactical investors play an active role in identifying mispricings and trading in a way that tends to eliminate them. For tactical investors, correlations among asset classes are less important than current valuations—and volatility may actually be good, because it creates trading opportunities.
The second type of investor—call them strategic—doesn’t have time to look for short-term mispricings; instead, they structure a well-diversified portfolio and take advantage of the interactions among their assets—that is, the power of diversification—to reduce the likelihood of a hit that would affect the entire portfolio. Strategic investors have a long investment horizon: Current valuations are less important than low long-term correlations and strong long-term risk-adjusted returns.
It’s not surprising that the investment characteristics of particular assets change over time: For example, return volatility for publicly traded real estate investment trusts, along with most or all other assets, spiked during the credit crisis of October 2008–March 2009. (REIT volatility rapidly declined to normal levels after the crisis ended, however.) What is more surprising, though, is that investment characteristics may also change over different investment horizons. If that’s the case, then you can potentially make better portfolio decisions by understanding investment characteristics at your own most relevant horizon.
As an example, Figure 1 shows correlations between the telecommunication services sector of the U.S. stock market and two measures of returns to the broad stock market, the Wilshire Total Market index and the S&P 500 index of large-cap stocks. [Correlation measures how closely two variables tend to move in relation to each other.] All of the correlations are computed over the same historical period, from January 1990 through November 2011—but they’re computed over differing investment horizons ranging from one-month returns to 60-month returns. [All sectors are represented by the S&P Global Industry Classification Standard series of sector total return indexes.]
As Figure 1 shows, the return correlations over longer investment horizons are much greater than the return correlations over shorter horizons—in fact, the correlation of the telecommunication services sector with the Wilshire Total Market increases from just 63% using a one-month horizon to 94% using a 60-month horizon. That kind of a change can make a huge difference in portfolio strategy: It suggests that telecom stocks provide little diversification benefit relative to the broad market for investors with relatively long investment horizons.
Why might long-horizon correlations be higher than short-horizon correlations? One explanation may be that, while telecom stock prices may fluctuate in the short term in response to conditions specific to the telecom sector, longer-term returns depend more on the state of the overall economy—which is reflected by returns to the broad stock market, rather than just one sector.
That explanation is supported by other sectors of the stock market, where a similar pattern holds: Return correlations between the sector and the broad market increase as the investment horizon lengthens, as shown in Figure 2. That makes sense if stock market returns over longer horizons are driven by factors common to all companies—such as growth in the gross domestic product or changes in interest rates—while returns over shorter horizons may be affected by factors specific to a given sector of the economy. Notably, the materials sector, which encompasses many commodity producers, differs from other sectors of the stock market, with correlations declining as the investment horizon lengthens (not shown in Figure 2).
There is another segment of the stock market, however, in which this pattern is reversed: publicly traded equity REITs. Equity REITs are the stocks of companies that own commercial real estate assets, such as office buildings and shopping centers. REITs are required to pay out almost all of their taxable income to their shareholders annually as dividends. The effect of the large dividend requirement is to channel the bulk of rent income to shareholders, providing them with the kind of real estate income stream they would receive if they owned and managed the properties directly.
As Figure 3 shows, equity REIT returns actually have a lower correlation with broad stock market returns over longer investment horizons. In fact, for investment horizons greater than six months, the decline in REIT-stock correlations is both steady and dramatic. For example, the correlation between equity REITs and the Wilshire Total Market is 67% over six-month horizons, but only 14% over 60-month horizons. (Equity REIT returns are represented by the FTSE NAREIT All Equity REITs total return index.)
Perhaps even more surprising, the same pattern holds even when REIT returns are compared with the financials sector of the stock market—ironically, the sector in which equity REITs are classified. For example, the REIT-financials correlation is measured at 79% over six-month investment horizons, but only 27% over 60-month investment horizons.
What might explain this pattern, in which long-horizon correlations are actually lower than short-horizon correlations? One explanation may lie in mutual funds and exchange-traded funds, which enable investors to trade broad categories of stocks rather than individual companies.
For example, investors who are bullish or bearish on bank stocks may trade their predictions by buying or selling a financial sector mutual fund or ETF—which would have the unintended consequence of driving up or down the stocks of non-bank companies encompassed within that sector, including REITs. Over longer investment horizons, though, REIT returns are driven by conditions in the real estate economy, rather than in the banking industry—or, indeed, in the broad economy as reflected by the broad stock market. Over longer horizons, then, any short-term mispricings of REIT stocks that are caused by non-REIT market factors (such as trading of bank stocks through a financial sector mutual fund or ETF) may be corrected in a way that brings down longer-horizon correlations.
If correlations can differ sharply over different investment horizons, then what about volatility? Figure 4 shows the annualized volatility of monthly returns measured over the same historical period—January 1990 through November 2011—but over different investment horizons extending from one to 12 months. As the figure shows, volatility declines sharply as the investment horizon lengthens. For example, the annualized volatility of equity REITs is 19.7% when measured using one-month investment horizons, but just 6.6% when measured using 12-month investment horizons.
(While correlations can be computed using overlapping periods of different lengths, volatilities must be computed using non-overlapping periods. As a result, the historical period January 1990 through November 2011 encompasses just 21 non-overlapping 12-month investment periods.)
How can an investor make use of these results? As noted, for tactical investors, correlations are less important than current valuations, and volatility may actually be good. In fact, it may be trades by tactical investors that explain the results, by driving mispriced securities back toward their fundamental values.
For strategic investors, however, these results are potentially very important in constructing optimal asset portfolios. If long-horizon correlations among sectors of the stock market are higher than short-horizon correlations, then strategic investors with long investment horizons won’t find good opportunities for portfolio diversification by holding assets in different sectors. In other words, investment characteristics based on short investment horizons (such as monthly data) overstate the diversification opportunities within the non-REIT parts of the stock market.
On the other hand, if long-horizon correlations between REITs and stocks are lower than short-horizon correlations, then strategic investors with long investment horizons can take advantage of greater diversification opportunities with larger holdings of REITs within their stock portfolios. Moreover, if volatility declines with the investment horizon, then risk-adjusted returns measured over long investment horizons are better than risk-adjusted returns over short horizons. In this case, investment characteristics based on short investment horizons (such as monthly data) understate the benefits of diversification through a mixed portfolio of REITs and non-REIT stocks, as well as the risk-adjusted returns of both asset classes.
Investors should consider which style best defines them: the tactical approach, in which investors look for short-term mispricings, or the strategic approach, in which investors take advantage of the power of diversification and long-term risk-adjusted returns. Long-horizon investing may well have greater benefits than we ever realized.
A REIT is a company that mainly owns, and in most cases operates, income-producing real estate such as apartments, shopping centers, offices, hotels and warehouses. Some REITs also engage in financing real estate.
To qualify as a REIT, a company must have most of its assets and income tied to real estate investment and must distribute at least 90% of its taxable income to its shareholders annually. A company that qualifies as a REIT is permitted to deduct dividends paid to its shareholders from its corporate taxable income. As a result, most REITs remit at least 100% of their taxable income to their shareholders and therefore owe no corporate tax. Taxes are paid by shareholders on the dividends received and any capital gains. Most states honor this federal treatment and also do not require REITs to pay state income tax. Like other businesses, but unlike partnerships, a REIT cannot pass any tax losses through to its investors.
For REITs, dividend distributions for tax purposes are allocated to ordinary income, capital gains and return of capital, each of which may be taxed at a different rate. All public companies, including REITs, are required early in the year to provide their shareholders with information clarifying how the prior year’s dividends should be allocated for tax purposes. This information is distributed by each company to its list of shareholders on IRS Form 1099-DIV. A historical record of the allocation of REIT distributions between ordinary income, return of capital and capital gains can be found in the Industry Data section of www.reit.com.
A return of capital distribution is defined as that part of the dividend that exceeds the REIT’s taxable income. A return of capital distribution is not taxed as ordinary income. Rather, the investor’s cost basis in the stock is reduced by the amount of the distribution. When shares are sold, the excess of the net sales price over the reduced tax basis is treated as a capital gain for tax purposes.