Risk & Portfolio Construction: Time to diversify
Investors diversify across asset classes, risk factors and investment styles. Maha Khan Phillips asks if trading time horizon brings anything extra to the process
The decision about when to buy an asset and when to sell it can be driven by a wide range of considerations. A pension fund’s decision to buy might be driven by the fact that a cash contribution has just come in from the sponsor, or a view that the asset is under-valued, or that its price is enjoying a rising trend. The decision to sell might be driven by the need to realise cash, or a view that the asset has reached fair or excessive valuation, or that its price is falling. Those might all seem sensible reasons to buy or sell.
But does it also make sense to buy an asset with a view to holding it for a certain amount of time – minutes, days, weeks, months, years? And does it make sense to try and diversify portfolios between these different time horizon-determined trading and investment strategies? That might seem perfectly arbitrary.
But pension funds may benefit from at least thinking in these terms. First, because they are forced into these trading strategies anyway: they experience regular cash flows, rather than receiving one lump sum one day and paying out another lump sum 40 years later; and while most institutional investors will take a strategic view of their portfolio, wanting to match assets to long-term liabilities, they also need to manage risk by looking to their short and medium-term asset allocations.
Second, perhaps more importantly, those time horizons might not be as arbitrary as they seem – some well-known investment ‘styles’ seem to fit naturally with certain broadly-defined time horizons.
Many investors already use investment styles in portfolio construction. Momentum, which works over the shorter term (from a few days to several months), and value, which works over a longer period (several months to several years), offers time-horizon diversification.
Time horizon also occurs naturally in portfolios, with some funds using high-frequency trading strategies and, at the other end of the spectrum, many adopting private equity or other long-term investments. And, note, these are both equity strategies – this is not asset-class diversification.
A 2009 research paper by Albert Desclée, Jay Hyman, and Simon Polbennikov at Barclays Capital argues that combining alpha strategies based on independent time signals can help reduce portfolio risk, even when the returns of the underlying assets are correlated. So a portfolio manager may run two active strategies on the same underlying asset but with different investment horizons. The firm believes that including strategies with different rebalancing frequencies in the same portfolio can be seen as a natural way to diversify risk.
“Intuitively, a lot of the style investing is looking at time horizons,” says Antti Illmanen, managing director of AQR Capital Management. “Value is about looking at things for the long run, whereas momentum looks at shorter time horizons. I think sometimes time horizon diversification is just a characteristic that comes with the type of allocations we make. For instance, some style characteristics are slow moving, whereas others are fast moving, with shorter time horizons. It is helpful to use both, and important to diversify. But using value and momentum are the best time-horizon diversifiers.”
For some industry participants, the issue is about different attitudes to liquidity. Martin Dietz, fund manager within the multi-asset group at L&G Investment Management, points out that short-term views are driven by liquidity considerations because they depend upon liquidity being available to trade, whereas medium-term strategies are more about buying a series of cash flows or a claim on some economic activity at a certain present value.
“[Short-term traders] look at who is investing in these markets, what are the supply-and-demand pressures, and what other market participants and hedge funds are doing,” he says. “Medium-term views are more valuation-based.”
Others put the liquidity argument a different way. “I think that the vast majority of academics in the financial profession think that time diversification is a myth, a fallacy,” says Kerrin Rosenberg, UK CEO of Cardano. “That aside, having different types of investments which have different levels of liquidity, for example a private equity instrument versus a more liquid instrument, is a separate argument. There is a liquidity-risk premium for less liquid assets.”
It also comes down to whether you believe that the statistical effect of mean reversion can trump an economic reality. At very short time horizons – intra-day – mean reversion seems to be quite powerful and is at the heart of most high-frequency and statistical arbitrage strategies. But what about the other end of the spectrum?
“If you look at the Japanese market, we are still waiting for it to come back,” observes Eduardo Repetto, CIO and co-CEO of Dimensional Fund Advisors. “If you go back to the 1900s, the Buenos Aires Stock Exchange was one of the top 10 in the world. Argentina today is not even considered an emerging market. So if you are waiting for time diversification based on the concept of mean reversion, you are still waiting.”
If mean reversion were true, he suggests, an investor could buy a 20-year equity put option cheaper than he could buy a one-year put. But in reality, the longer-dated the put, the more expensive it becomes.
For CTAs and other trend-following strategies that use short horizons, mean reversion is very much a reality, and the issue also circles around what kind of market view you are taking, and the risk profile it has.
“Trend-following strategies and fundamental strategies are different from a risk management perspective,” explains Matthias Hagmann, head of equities at Man Group’s $14.4bn managed futures strategy AHL. “As a trend-follower you buy an asset because it recently went up. Let’s imagine now the price goes down and you lose money. Then the system automatically cuts the position as the trend looks less attractive. For a fundamental system, the situation is the opposite. If you buy an asset because it is cheap, you lose money if the price falls further. But now it looks even cheaper and, ideally, you would like to buy more. That’s a very different risk profile, in one case you cut your position when you lose money, in the other you are in danger of catching a falling knife.”
In his trading context, Hagmann sees the logic in making decisions based on time horizons, and feels that combining different time horizons improves diversification.
“It may be that for a while, short-term trend following doesn’t work but long-term trend following does,” he says. “This is why it is beneficial to diversify across different frequencies.”
Outside this trading context, does the idea still hold? Possibly – but only in a given set of circumstances. It may go under different names – like ‘the illiquidity premium’ or ‘the value premium’. But as pension funds find themselves buying and selling assets, it might pay to think more deeply about what effects the timing of those decisions has on diversification across their portfolios.