Alternative exposures can be the 'secret sauce' for a multi-asset portfolio: in the search for diversification, they offer low correlation with other assets and the potential to improve portfolio outcomes. Many alternatives are illiquid real assets such as property, infrastructure, forestry or farmland. Because these can be expensive to purchase and sell, they are less well-suited to multi-asset portfolios that require a high level of liquidity.
Alternatives include investment trusts as well as regular companies that own and manage a real asset, such as Real Estate Investment Trusts (REITs), listed airports and utility or railroad companies. While the underlying assets remain illiquid, investors trade in shares of listed companies directly with each other, without buying or selling the underlying asset. Trading costs are thus low and liquidity can be high.
Some investors question the use of listed alternatives: how can we expect an illiquidity premium if we are holding liquid instruments? And how can we reconcile the volatility of the listed alternatives with the smooth pricing of direct real assets? Intuitively, listed and direct exposure have the same underlying economic exposures and so they naturally should provide the same outcomes over the long-term. Unfortunately, there isn't much research to support that claim.
Listed real estate
This post will ignore a few aspects of REITs: leverage, sector bias and market depth. I'll focus on the price returns, and while there is no reason to believe that rents would differ between listed and direct structures, there may well be differences in fees and management style.
Let's look at some figures from the US, where we have good data for commercial property as well as for REITs.
Commercial property shows the smooth performance that many investors expect, while REITs are clearly much more volatile. A correlation analysis of their returns versus those of the S&P 500 index – importantly, over different time horizons – establishes how closely the two series are linked:
Correlations over three time horizons
1M 1Y 3Y REITs vs direct 0.01 0.22 0.37 REITs vs equity 0.60 0.57 0.47 Direct vs equity -0.02 0.25 0.42
Source: Bloomberg, CoStar, BIS, LGIM calculations, as at 31 January, 2019
There is virtually no link between listed and direct property over the short-term (0.01 over one month); listed alternatives instead trade more closely in line with equities (0.60). This result changes a lot if I extend the time horizon. Over a three-year horizon, the correlation between REITs and direct real estate increases to 0.37, while the correlation to equities is almost identical for both property series (0.47 and 0.42).
1M 1Y 3Y REITs 0.28 0.26 0.25 Direct 0.75 0.38 0.29
The next question is how the two series compare in terms of returns: do the REITs offer lower returns (miss out on the liquidity premium) or suffer from higher risks (for the same returns)? The table above suggests the annualised risk-adjusted returns for REITs don't change much with the time horizon. The most striking result is how the risk-adjusted returns of the commercial property series fall as the time horizon increases. This is purely driven by the underlying risk (annualised volatility) which is very low when calculated using short-term (one-month) returns, but grows using longer time horizons.
To understand the changing volatility of direct property, let me briefly dive into its valuation process. With no obvious market price, valuations are linked to historic transactions of comparable properties. But transactions take a while to complete, so when deal prices become public they are already old. And sometimes there aren't many transactions taking place so pricing remains stale by default. This valuation approach thus creates gradual price movements over the short term (an econometrician calls this autocorrelation) and only longer-horizon data shows the full change in underlying valuations.
So the initial data analysis confirms that prices of listed and direct real estate converge over time – which is what we would expect, as they own similar assets whose underlying performance will determine long-term returns. What is still a bit unclear from this is how the dynamics work: Do REIT prices revert towards the direct valuations, or are they actually a leading indicator?
Some further fancy econometric analysis – a "vector-error correction model" (VECM) – describes the dynamic behavior between the two markets. The charts on the left (impulse response functions) highlight how a shock to one series impacts on the other. There is strong evidence (the grey lines show a 95% confidence interval) that a REIT price move is gradually priced into the direct property (top left). Similarly, a change in commercial property is incorporated into REITs (bottom left), though the result isn't statistically significant. The analysis also incorporates equity returns as an external variable. Direct property doesn't budge if equities move (the beta is literally 0.0), but the REITs do, and have a beta of around 0.8.
Now, the slow-moving nature of commercial property prices means there aren't many shocks in that series that can feed into REITs. I'll highlight how much of the volatility in one series is explained by the other (using a FEVD) – which is effectively telling us which series is leading the other: almost 70% of the volatility of the commercial property over five years is pre-empted by REIT pricing (blue bars, top right). Price changes in commercial property explain very little of the REIT prices (grey bars, bottom right).
So because their prices can move easily, REITs quickly reflect the short-term risk-on/risk-off feature of markets as well as market-moving news. The valuations of the direct real assets simply move too slowly – but they would reflect these factors gradually if they became a permanent feature (top row above).
Other real assets – infrastructure, farmland and forestry
The same pattern is evident across other listed alternatives, which is a good cross-check of the results: correlations and risk-adjusted returns for infrastructure and farmland/forestry show the same difference between short and long-term analysis. And the impulse response functions look comparable (see chart below) with similar equity betas (0.0 for direct infrastructure vs 0.6 listed, 0.0 for direct farmland/forestry vs 1.3 listed).
Investors should expect similar long-term outcomes when investing in direct and listed property alternatives. Over the short term, listed alternatives respond more quickly to changes in their underlying assets and as a result, unfortunately, trade in line with the changing risk-on/risk-off pattern of other risky assets.
The stability of the accounting valuations of real assets will be an important benefit for many investors. But for those who require short-term liquidity, listed alternatives can be a suitable way of getting the diversification benefits offered by reals asset exposure, as long as they take into account the higher short-term volatility.