Join our community of smart investors
OPINION

Resilience vs optimisation

Resilience vs optimisation
August 11, 2015
Resilience vs optimisation

What I mean is that the Aussies are not complete mugs: their 566 at Lord’s proves that. Instead, they are flat-track bullies. They do very well when conditions are right, but cannot cope with a ball that’s swinging or seaming; they don’t have the ability to make ugly 50s when the bowlers are on top. Their batting line-up is optimised for perfect conditions, but lacks the resilience to deal with imperfect ones.

In this sense, the Aussies are just like banks before the financial crisis. Their high gearing and dependence on wholesale finance meant they did very well in good times, but left them incapable of coping when things changed. Banks were optimised for a particular environment, but not resilient when that environment changed.

The same trade-off can apply to portfolios. What looks like an optimal portfolio might only be so because it is based upon an unrepresentative sample of data. If so, it will lack the resilience to survive when things change. The most famous example of this came in 2007 when Goldman Sachs’ David Viniar blamed losses on his companies’ hedge funds on “25 standard deviation moves.” However, the price moves only looked like 25 standard deviation events because Mr Viniar’s measures of risk were based upon data that over-sampled low-volatility and under-sampled high. His funds, like Aussie batsmen, were optimised for good times but not resilient.

It’s not just in finance that this trade-off exists. It’s also a problem for non-financial firms. Companies can thrive for years only to collapse when new technology makes their products obsolete: think of Nokia being undermined by the rise of smartphones, or Kodak and Polaroid by digital photography. Boyan Jovanovic and Peter Rousseau have shown that firms often embody “vintage organizational capital” - they have the technology that was cutting-edge when they were founded - whilst Paul Ormerod and Bridget Rosewell have shown that they often lack the ability to learn and adapt to changing conditions. Put these two together and we have firms that resemble Aussie batting - they can do well in good times, only to struggle when things change.

I should stress that what we have here is a trade-off; often, firms and portfolios cannot be both optimised and resilient. Resilience requires that companies have spare capacity that can be used when demand or technology shifts - that they have a plan B (or C, D and E). But that is incompatible with maximizing efficiency.

Think of markets as being like eco-systems; there are many points of similarity. If a species is dependent upon only one food source it would go extinct if that source disappears whilst the species that has many such sources will survive. Survival - resilience - thus requires a mixed strategy. But mixed strategies, by definition, are not optimal in an unchanging environment.

To see my point, take the “rock paper scissors” game. If your opponent always plays “rock”, your optimal strategy is to play “paper”. That has a 100 per cent success rate. If, however, your opponent plays at random, your best strategy is also the mixed one – to randomize. But this has a success rate of only 33 per cent.

There are, I think, at least two implications for investors in all this.

One is that stock-pickers, especially those who think they’re investing for the long-term, must be sceptical about companies. They must ask not just: is this firm doing well? But also: could it cope if the economy changes? If the firm has strong barriers to entry and low gearing, the answer might be yes. If not, it might be another Nokia or RBS.

Investors seem to under-appreciate this. Researchers at AQR Capital Management have shown that “quality” stocks - those that are profitable, growing and have low leverage - tend to out-perform around the world. This suggests that investors under-rate the importance of strong balance sheets and barriers to entry that protect profits from competition. As Warren Buffett has said, “a truly great business must have an enduring “moat” that protects excellent returns on invested capital.”

Secondly, all this strengthens the case for satisficing rather than optimising. What looks like an optimal portfolio might not be so if things change so that previously uncorrelated assets fall together - which is a distinct risk for bonds and equities. Given that we cannot predict future returns, volatilities or correlations, this argues for a portfolio that’s good enough - one with a spread of assets (including cash) that protects us from multiple risks even if it doesn’t maximize risk-adjusted returns in an environment which might not last.