Investing has been at the heart of EV since its inception, and we boast a long pedigree of investment solutions that have delivered outstanding performance through booms and busts.
Throughout it all, a recurring question has been what we mean by long term investing with numerous follow-ups about how we account for it, and what we can do to help in concrete, practical terms.
In this two-part blog series, we gather our answers and some of our thinking. We also point out what is missing in the industry in general and in the propositions of our competitors in particular.
In a word, yes.
Of course, this only matters if the term does make a difference. We strongly believe it does and you don't have to look far for evidence. For a prime example, look no further than the main tools for modifying risk for personal investors: cash and bonds. Both of these show completely different behaviour over short and long terms.
Cash is highly predictable in the short term but, over the medium term or longer, interest rates are highly unpredictable and results are often a long way from anything a simple extrapolation from current rates and short term volatility would lead you to expect.
Bonds are just the reverse. In the short term, they can be quite volatile but, if you hold one and the issuer doesn't go bust, you know precisely what you are getting.
These asset classes, the basic building blocks of portfolios, show materially different behaviour over short and long terms. It just can't not make a difference.
Now that we’ve established that term matters, how can it be incorporated into investment decisions? Models are an obvious approach. But not any model.
Our investors have long-term goals and anything that offers them a solution is doing long-term modelling. Many models out there may come to similar conclusions but most feature important components that are either oversimplified, arbitrary or inconsistent. A model should be as simple as possible but no simpler.
The way our approach is different is grounded in our core philosophy of making our modelling explicit, transparent and integrated. This feeds into our portfolio optimisation which is driven purely by model returns, and that means that our allocations address those issues automatically and systematically. This has implications for accumulation and for decumulation, and we have both bases covered.
We believe that careful, consistent and uncompromising analysis of market data provides the best possible basis for modelling from available data. We believe that careful, consistent and uncompromising analysis of market data provides the best possible basis for modelling from available data. We achieve this in the EV asset model, our global stochastic asset model, which is a scenario generator that produces realistic forecasts of the distribution of future market conditions and asset class returns. We use the EV asset model's projections to determine optimal allocation weights which, for accumulation portfolios, are typically determined by optimising for the best average growth rate subject to annual rebalancing and volatility constraints.
Let’s return briefly to the example of bonds above. In the short term, their price can be very volatile. If this level of volatility is assumed to persist over the long term, bonds would offer a very different proposition altogether. In fact, they would leave little room for diversification with higher volatility assets like equities, contrary to what is observed in practice. By the same token, if the high volatility associated with equities is expected to last forever, then the prospect of high growth portfolios would remain elusive for most investors, as having March 2020-like volatility all the time is not going to find many takers. What is required is a model that is able to correctly identify the evolution of risk with the term and uses this to find the most suitable asset allocation for a given term (and other considerations). This is what we have with the EV Asset Model.
What is clear is that assuming constant volatility in asset returns into the future is both unrealistic and in contrast with the historical record. MVC (mean-variance-covariance) models, for instance, fall short by this count. A tool that is unable to account for term in the investment choice is also unlikely to be able to offer a realistic picture of the possible outcomes. There is a good analogy with bicycles and gears here: at any moment, there is probably a fixed gear that would be appropriate but, the longer the journey, the more likely it is that more than one gear is needed. It makes the bike more complicated but nobody wins the yellow jersey with a single gear!
So, investment decisions should be based on as realistic assumptions as possible, and that should necessarily incorporate the realistic behaviour of key asset classes. For us, that includes reflecting the dynamics of the term structure of interest rates in fixed income investments and modelling the volatility clustering observed in equities to allow for the possibility of events like seen most recently in March 2020. These are two big direct features of our approach and have implications for long-term investment modelling. An indirect feature that results from these is that our model makes a proportionate response to market conditions which is crucial: static asset allocations have a limited shelf life. In our next blog, we will go into more detail about what this means for asset allocation, and how this impacts our approach.
EV’s range of strategic asset allocations is grounded in a set of robust and academically-tested investment beliefs that deliver outperformance by design. Our model has outperformed the market and other funds consistently for over a decade. This outperformance and resilience were especially visible during the Global Financial Crisis and during the Market Dislocation experienced earlier this year.
Download your copy of our Capital Markets Assumptions report now.