Modern Portfolio Theory, with its origins in the 1950s and 1960s, provided investors with a coherent framework for making risk and return choices for the first time. Models, though, are always imperfect abstractions of reality, and a lack of computing power back then made it impossible at the time to test the theory’s embedded assumptions.
Fortunately, Moore’s Law (the theory that computer speeds and memory double every two years while prices halve) kept running in the background. By the 1980s, we had enough computing power and data to test the assumptions of Modern Portfolio Theory: and they were found to be critically flawed. Return distributions were neither normally distributed nor independent, and risk (defined as historical volatility) was not linearly correlated with return.
But Moore’s Law continued to compound, and we now have the computing power to build a better model—a better abstraction of reality. It is time to move on.
Investors should recognize that risk defined as historical volatility is not the whole story. They also need to recognize there are future events that they might reasonably anticipate—ageing demographics and climate change, for example. At the same time, investors shouldn’t be surprised by unforeseen events. They are a part of life, and disruptive technologies are the order of the day. Indeed, the pace of technological innovation has accelerated beyond most of our imaginations.
Yesterday’s winners are frequently overtaken by new business models and technologies. Consequently, cap-weighted benchmarks based on historical success are a poor place to start building portfolios. The Stone Age did not end because the world ran out of stones: it ended because we found something better (and you may wonder whether the same fate will befall the fossil fuel industry, leaving a lot of stranded “assets” to be written off).
Now we need to distinguish between different types of investors: asset accumulators and decumulators. Falling valuations are good news for one but not for the other.
Investors must not over simplify the problem by using time-weighted returns and treating almost everything as a single period optimisation. Fortunately, all of this is possible with currently available technology—with all the computing power on the typical desk today, investors can use money-weighted returns, non-parametric statistics and Monte Carlo simulations, none of which were options in the past.
Replacing market-based benchmarks with real world benchmarks that are directly linked to a pension fund’s purpose will require plan sponsors to re-visit governance structures. They will also have to engage with every one of the important stakeholders at the table: all important decisions need to be collective, if plan sponsors are to withstand the pressures of following an investment path different to the herd.
In this brave new world, the key benchmark becomes a fund’s actual liabilities, not a strategic, market-based benchmark. The move away from static, market benchmarks liberates asset allocation, and investors are now free to respond to shifting valuations in markets and shifting fund wealth.
The two together allow risk appetite to change in a way that acknowledges changes in rational return expectations, and changes in wealth in a fundspecific manner.
This is a brave new world, with the potential to be liberated from the rigidities of today’s best practice model.
Alan Brown is retired chief investment officer, Schroders