Paul Woolley is digging at the foundations of modern portfolio theory

17 Mar, 2014

One of the curiosities of the long and unresolved financial crisis is how little damage it has done to the theoretical foundations of investment management. The irrelevance of the mixture of equity and debt in a capital structure, the trade-off between risk and return and the Capital Asset Pricing Model (CAPM) that rests upon it, the efficient market hypothesis that holds stock prices reflect all available information and no investor can earn excess returns, the Black-Scholes-Merton options pricing theory that underpins derivative pricing and even Value at Risk (VaR) as the principal tool of risk management are all still in day-to-day use in the investment management industry.

But there is one man who thinks the persistence of flawed theories of finance is putting capitalism itself at risk. He is Paul Woolley, the former stockbroker, academic economist, IMF official, merchant banker and co-founder of the investment management firm GMO Woolley, from which he retired in 2006. A year later, puzzled by one episode in particular – the withdrawal by clients at the turn of the century of 40 per cent of the assets they held at GMO Woolley to board the ill-fated TMT bandwagon, from which the firm had (largely) abstained – he returned to the academy to see if he could find a theory that fitted the experience.

Since 2007, the Paul Woolley Centre for the Study of Capital Market Dysfunctionality at the London School of Economics has undertaken the task of addressing the obvious shortcomings of financial economics, by using its own methods. Amid all the moralising about the evils of bloated finance and financiers, this has the virtue of being unhysterical, while losing nothing on terms of revolutionary potential. At last, Paul Woolley himself has started to put into the marketplace of ideas some of the conclusions from the work which the Centre has undertaken.

At a recent CFA Institute conference in London, he offered a glimpse at what he calls a “new paradigm” to replace the “fallibility” of present financial theory. As Woolley rightly pointed out, modern financiers are applying the theory of the special case – so useful in benchmarking a finding in the laboratory – to practical conditions. It is like engineers designing motor cars for roads without friction. No wonder the results are so often disastrous.

In fact, Modigliani-Miller, whose initial formulation assumed zero transaction costs, zero taxes, no difference in the rates at which individuals and corporations can borrow and no agent-principal conflicts, could be described as the bad science metaphor brought to life. Ironically, once the tax-deductibility of interest was added to the model, Modigliani –Miller became even more dangerous, since it argued in effect for an infinite amount of leverage. Banks that entered the crisis with $1 of equity for every $100 of loans on the asset side of the balance sheet were the perfect expression of the theory in practice.

But the greatest flaws in Modigliani-Miller were ignoring transaction costs and principal-agent problems. It is these which Woolley had in mind when he told his audience that modern financial theory describes an illusory world in which “rational investors invest directly in securities in the absence of frictions or limits to arbitrage.” The conflict between management and shareholders in listed companies is well-known and well-understood. Paul Woolley is a lone voice in extending that principal-agent conflict to the management of money by fund managers on behalf of investors.
“One of the core assumptions of standard theory is that households do their own investing, yet the bulk of investors now delegate this responsibility to financial intermediaries,” he explained. “Delegation raises the prospect of principal/agent problems, nothing new to economists in corporate finance or banking, but not widely explored in their implications for asset pricing.  Delegation creates information asymmetry; agents have better information and different objectives, but the main problem is that principals are unsure of the competence and diligence of agents. Our model incorporates delegation and is conducted in a rational framework with principals and agents acting optimally to maximise their risk-adjusted returns. The basic model is able to explain many of the common examples of systematic mispricing that have proved difficult to square with standard theory.”

To illustrate how agents (fund managers) and principals (investors) can “act optimally to maximise their risk-adjusted returns” and still lose their shorts, Woolley alluded to his own experience at GMO Woolley during the TMT Bubble at the turn of the century. Like Tony Dye, who lost his job as chief investment officer at Phillips & Drew for scorning TMT stocks in the late 1990s, GMO Woolley lost clients which withdrew money from the firm to chase the higher-performing managers that were pushing TMT stocks way above their fundamental value. They did not return to GMO Woolley until the firm had already seen its sceptical strategy succeed after the TMT Bubble was pricked.

In other words, clients of GMO Woolley missed the TMT boom, invested in the TMT bust, and missed the return to value stocks that followed. Yet they thought they were behaving rationally. The lessons of that episode, explained Woolley, were two-fold. First, that investors could behave rationally, and still lose a lot of money by failing to understand what was going on. Secondly, that momentum (the sheer weight of money flowing into an asset class) is as important of a determinant of stock prices as the real source of value in investment (future cash flows received).
Nothing new there, you might think, and in a sense you would be right. Professional investors have always understood that there is even dumber money just behind them. What is novel about the work at the Paul Woolley Centre for the Study of Capital Market Dysfunctionality is that it provides a theoretical model that explains how asset prices are determined by flows of funds rather than flows of cash. Importantly, Woolley says the theory has so far survived tests using 70 years of US stock price data.

It is easy to under-estimate how catastrophic this “new paradigm” will be, when fully unveiled, for modern portfolio theory and its twin foundations: the CAPM and the efficient market hypothesis. If it proves robust, it will be equally disastrous for the preferred method of managing an investment management relationship by measuring performance against a benchmark, with a margin built in for tracking error. To understand why, it is worth recalling how modern financial theory proceeds.

Portfolio theory is based on the simple insight that ice cream sells well on hot days and umbrellas sell well on cold days. All of the jargon about diversification of risk, non-correlated assets and volatility essentially reduces to a search for a portfolio of assets that, in aggregate, delivers returns commensurate with the level of risk assumed. In other words, there is a trade-off between risk and return: extra return entails additional risk.

The CAPM is really no more than these ideas reduced to a practicable form. By assuming that every investor is holding a market portfolio plus a risk-free asset, the CAPM enables investment managers to calculate the impact of adding any new asset on the overall riskiness of the portfolio. That additional risk is beta and, if you can work out the beta for any and every single asset, you can work out if adding any asset you care to name to your portfolio adds a lot or a little to existing risks (beta) or a lot or a little in terms of return above the portfolio risk (alpha).

What the new paradigm created by Paul Woolley and his colleagues suggests is that the combination of momentum investing and benchmarking systematically distorts prices in ways that make a nonsense of the CAPM and the efficient markets hypothesis. They encourage investment in higher beta stocks (i.e. volatile stocks that add a lot to existing risks in a portfolio) and discourage investment in lower beta stocks (i.e. less volatile stocks that adds little to existing risks in a portfolio).

By pushing higher beta stocks far beyond their fundamental value, and lower beta stocks far below their fundamental value, momentum investing and benchmarking guarantee that the riskiest investments end up with lower returns than the less risky ones. This obviously contradicts the central tent of modern portfolio theory, as expressed in the CAPM and the efficient markets hypothesis, that there is a trade-off between risk and return. If the new paradigm stands up, it is the first of many intellectual defeats for modern financial theory. This is useful and necessary work, because catastrophic practical failure is clearly not enough.

Dominic Hobson

Paul WoolleyVaRCapital asset pricing model