Ever since Gary Brinson’s 1986 publication of a paper demonstrating that asset allocation is the dominant factor in determining the return of a portfolio, strategic asset allocation has been accepted as important. When it comes, however, to including alternative asset classes in an overall asset allocation framework, the industry is still foundering.
Too often, clients and advisers rely on what seems like a rather arbitrary decision-making process to decide on the weighting of alternative asset classes in investor portfolios. How else can we explain the typical 5% or 10% allocations that we see, which seem like good numbers not only because they are rounded, but also because they tend to be in line with what other market participants have done?
As just one example, let’s consider private equity. Technically speaking, private equity appears to resemble public equity in a number of ways, but there are also significant exceptions. Investors must be qualified to participate in these investments. Furthermore, no liquid and public markets exist to enable qualified investors to trade in and out of private equity vehicles. Because no public market exists, it does seem somewhat unfair for managers and investors to come up with a rather arbitrary weighting for private versus public equity investments. For example, typically, pension funds would hold up to 10 times as much public versus private equity.
In another instance, private property too is often treated as an alternative asset class, and it too is frequently assigned a portfolio weighting derived from what seems like a rule of thumb. Real estate as well does possess a publicly quoted liquid alternative: the real estate investment trust, (REIT). Typically, pension funds hold almost no REITS other than a negligible percentage within their global equity portfolios.
So, if our analogy holds and private equity indeed relates to public equity in the same way as private property relates to public property, then how can we explain why investors typically hold 10 times more public equity than private equity – but their ratio of public to private real estate is almost zero? We definitely see some inconsistency in this situation, which may boil down to the fact that our industry hasn’t thought carefully enough about the role of liquidity.
At least to some degree, most asset allocation decisions are derived from models that trade off three factors, either implicitly or explicitly: expected return, expected volatility and expected correlations. In these models, we see a major risk of over- or under-allocating to illiquid assets for three reasons. First is an inconsistent measurement of risk. Due to stale pricing in illiquid markets, volatility estimates that are based on historical volatility should be corrected so that they can be compared on an apples-to-apples basis to estimates in the liquid markets. Chart 1 illustrates what we mean when we speak of stale pricing. As the chart indicates, the volatility of the marked to market asset is measured as being higher due to the simple reason that the non-traded asset is not re-valued frequently. That is why its pricing is said to be stale, as opposed to the pricing of a REIT, which is revalued on a daily basis.
Our group has created modern asset allocation techniques that seek to provide more useful estimates of risk. Chart 2, for example, shows adjusted volatility and correlation numbers for venture capital as well as leveraged buy-outs. As you can see, when the numbers are adjusted for serial correlation and stale pricing, they must be revised significantly upward.
The second reason is the fact the conventional asset allocation models do not provide for the challenges created by rebalancing portfolios that contain substantial allocations of illiquid investments. Imagine that two investments – one illiquid, the other liquid, provided an investor with the same expected risk and return. Given the similarity of risk and return, an investor would be prudent to favour the liquid over the illiquid investment because the liquid investment also enables the investor to rebalance his or her portfolio efficiently or to withdraw altogether, as one can from a liquid asset class. The advantage afforded by this ability to liquidate a position is, basically, the investor’s potential ability to switch completely into a new opportunity that presents itself rather than be, essentially, ‘trapped’ within an illiquid asset.
Clearly, then, for an investor to sink money into an illiquid investment, he or she must be compensated for locking up that capital. In short, there should be an illiquidity premium. The question becomes, of course, what type of illiquidity premium an investor would require in order to feel fairly compensated for participating in the illiquid investment opportunity. Here is where modern allocation techniques come into play by optimising portfolios so that investors gain an appropriate understanding of how to trade off choices in three dimensions: risk, return and liquidity.
A third reason why conventional allocation models are challenged when asked to take illiquid asset classes into account has to do with how an investor’s personal liquidity preferences are treated.
Let us return to the example of the liquid and illiquid asset choices. Now, let us assume that the illiquid asset has an expected return that compensates exactly for the fact that the asset does not trade. In other words, it provides an illiquidity premium.
How do you decide the relative allocations for these two investments, which appear – at least in terms of numbers – to be perfectly comparable? Your decision, in fact, rests not on numbers, but on your need for liquidity. In other words, how soon do you think you’re going to need your money? The less you need liquidity, the more readily you can target illiquid investments that expect to pay a premium of which you can avail yourself because you do not need to pull money out in the near future.
As a result, our traditional efficient frontier diagrams actually should possess a third component: the more an investor can tolerate his money to be locked in, the higher the efficient frontier he can choose a portfolio from. Chart 3 illustrates that exact point. As the maximum percentage of assets invested in illiquid holdings increases, the efficient frontier shifts upward.
Now that we have dealt with the liquidity issue, I would like to look into another asset allocation topic, namely the role of global diversification within asset classes. There largely are two key benefits from broadening the allocation within an asset class to include global investments. The first is that widening the opportunity set should allow a talented manager to, on average, seek to achieve a better return. The second benefit is the risk reduction benefit, thanks to the fact that the different securities within an asset class are not perfectly correlated. Thanks to Harry Markowitz, we all understand that the lower the correlations, the more we can potentially benefit from this phenomenon of risk reduction.
What do the correlations tell us? Empirically, we can demonstrate that equities across the globe have a reasonably high correlation. In other words, to an increasing degree, their performances are coming to resemble one another. Intuitively, this can be explained by the fact that most quoted stocks are globally oriented companies. Hence, they are exposed to many of the same economic factors.
This tends to be truer for publicly traded securities than private investments: cross-regional correlations in property and private equity are much lower. Here, the investments tend to have more local flavours. As Chart 4 indicates, the correlations for properties are distinctly different by region, which should make it possible for a global real estate investor to diversify investments within this particular asset class.
As Chart 4 indicates, the historical correlations are quite low and intuitively this can be explained by the fact that the property markets have very local characteristics. In private equity as well the cross-regional correlations tend to be lower than in the public equity markets. Here the intuitive explanation is that the companies tend to be smaller than public, listed ones and they have more of regional, if not local characteristics. You can see that this line of reasoning leads us to a paradox: most investors are embracing global investing more in terms of public equities versus private equities or private property even though benefits of diversification appear higher in the latter two categories.
Furthermore, even though most of the necessary technology and know-how to incorporate all the new asset classes in a total asset allocation exercise exists, most of today’s allocations still look like the result of arbitrary decisions. Pension funds, consultants and alternative investment specialists may desire to cooperate more closely to give thought to global investment portfolios and, in particular, to how they treat some of the alternative asset classes and their specific features because all investment opportunities, alternative or traditional, merit a deeply professional evaluation about whether they can help pension funds to achieve their objectives.
Owen Thiers is a director and head of institutional sales at Citigroup Alternative Investments