« Paul Krugman: The Politics of Spite | Main | links for 2009-10-05 »

Monday, October 05, 2009

"Capital Market Theory after the Efficient Market Hypothesis"

Dimitri Vayanos and Paul Woolley argue that we do not need to abandon the assumption of rational expectations in order to better understand how asset markets function, but we do need to incorporate principal agent problems into asset market models.

Setting aside whether agents are rational, I am fully sold on the idea that we need to do a better job of incorporating principal agent problems into these models, I have pointed to them myself and believe they pervade just about every step in the mortgage process (real estate agents, banks, appraisers, mortgage brokers, financial managers, etc.). However, in the past I have thought of these mechanisms as allowing bubbles to inflate rather than being the cause of them. That is, I have argued that two things that must happen for bubbles to occur. First, there must be a source of liquidity that can blow the bubble up, e.g. there must be something like a low interest rate policy from central banks or high savings rates in some countries of the world. Second, the protections that markets and regulations provide must fail so that all of the liquidity can pass through the regulatory "baffles and checkpoints" that would  prevent the liquidity from over inflating some sector in the economy such as housing or stocks. Thus, within this framework, I see the ideas below as explaining how excess liquidity might cause a bubble to develop, but it does not tell us why excess liquidity might be present.

However, I now wonder if that separation is valid. Is it possible for a bubble to occur simply through reallocation of existing investments, i.e. without an external source of liquidity to drive it? The story below is one where asset price "momentum" is generated from principal agent issues driven by incomplete information on the quality of investment fund managers. People see high returns in a particular sector, and they cannot tell whether the lower returns they are receiving are due to their fund manager's proper avoidance of risk, or incompetent management. As they increasingly conclude that incompetence is to blame, funds shift to the new sector and this creates a self-reinforcing process where prices are driven above their fundamental values, i.e. a bubble occurs. It seems like such reallocation of investment funds could, if driven by a strong enough incentive, be enough on its own to drive a bubble even without an external source of liquidity.

What does seem to be key is there must be some reason to believe that the higher returns are due to something other than taking more risk. In the present crisis, it was financial innovation coupled with the idea that policymakers and the market could maintain the Great Moderation that led to the idea that higher returns could be generated without a corresponding increase in risk. In the dot.com case, it was the promised higher productivity from the internet that provided the necessary story, and in the Great Depression it was electricity and the internal combustion engine (among other technological advances) that convinced everyone that we had entered a new, unprecedented era of higher productivity.

So I would now amend the list a bit and say that (at least) three things are needed to generate a bubble. First, an idea that makes people believe that higher returns are available without assuming more risk needs to be present. Second, there must be a source of liquidity to inflate the bubble. This can come from external sources such as high saving or low interest rate policy, or it can come from reallocation of existing investments (e.g. when people in the U.S. stopped loaning to foreign governments prior to the Great Depression so that they could chase the higher returns at home). And third, there must be regulatory and/or market failures that allow the bubble to inflate with little or no resistance.

In any case, here's the argument on how "momentum" might be created:

Capital market theory after the efficient market hypothesis, by Dimitri Vayanos and Paul Woolley, Vox EU: Forty years have passed since the principles of classical economics were first applied formally to finance through the contributions of Eugene Fama (1970) and his now-renowned fellow academics. Over the intervening years, capital market theory and the efficient market hypothesis have been developed and modified to form an elegant and comprehensive framework for understanding asset pricing and risk.

But events have dealt a cruel blow to these theories, as John Authers argued in his recent FT column. Capital market booms and crashes, culminating in the latest sorry and socially costly crisis, have discredited the idea that markets are efficient and that prices reflect fair value.

Some economists still insist these events are simply the lively interplay of broadly efficient markets and see no cause to abandon the prevailing wisdom. Other commentators, including a number of leading economists, have proclaimed the death of mainstream finance theory and all that goes with it, especially the efficient market hypothesis, rational expectations, and mathematical modelling. The way forward, they argue, is to understand finance based on behavioural models on the grounds that psychological biases and irrational urges better explain the erratic performance of asset prices and capital markets. Presented this way, the choice seems stark and unsettling, and there is no doubt that the academic interpretation of finance is at a critical juncture.

The need for a science-based, unified theory of finance

At stake is the need for a scientifically based, unified theory of finance that is rigorous and tractable; one that retains as much as possible of the existing analytical framework and simultaneously produces credible explanations and predictions. This is no storm in an academic teacup. On the contrary, the implications for growth, wealth and society cannot be overstated. The efficient market hypothesis has beguiled policymakers into believing that market prices could be trusted and that bubbles either did not exist, were positively beneficial for growth, or could not be spotted. Intervention was therefore unnecessary, and regulation could be light-touch. By contrast, a theory of asset pricing that did a good job of explaining mispricing would provide policymakers with a stronger rationale for intervention and more scepticism about mark-to-market, index-tracking, and derivative pricing, to name but a few examples.

Principal-agent investment problems: Mispricing with rationality

We believe that a first step in the search for a new paradigm is to avoid the mistake of jumping from observing that prices are inefficient to believing that investors must be irrational, or that it is impossible to construct a valid theory of asset pricing based on rational behaviour. Finance theory has combined rationality with other assumptions, and it is one of these other assumptions that has proved unfit for purpose. The crucial flaw has been to assume that prices are set by the army of private investors, the "representative household" as the jargon has it. Households are assumed to invest directly in equities and bonds and across the spectrum of the derivatives markets. Theory has ignored the real world complication that investors delegate virtually all their involvement in financial matters to professional intermediaries – banks, fund managers, brokers – who dominate the pricing process.

Delegation creates an agency problem. Agents have more and better information than the investors who appoint them, and the interests of the two are rarely aligned. For their part, principals cannot be certain of the competence or diligence of their appointed agents. The agency problem has been acknowledged in corporate finance and banking but hardly at all in asset pricing. Introducing agents brings greater realism to asset-pricing models and can be shown to transform the analysis and output. Importantly, this is achieved whilst maintaining the assumption of fully rational behaviour on the part of all concerned. Such models have more working parts and therefore a higher level of complexity, but the effort is richly rewarded by the scope and relevance of the predictions.

By doing this in our recent paper (Vayanos and Woolley, 2008), we have been able to explain momentum, the commonly observed propensity for trending in prices, which in extreme form causes bubbles and crashes. Momentum is incompatible with an efficient market and has proved difficult to explain in the traditional framework. Indeed, it has been described by Fama and French (1993) as the “premier unexplained anomaly” in asset pricing. Central to the analysis is that investors have imperfect knowledge of the ability of the fund managers they invest with. They are uncertain whether underperformance against the benchmark arises from the manager's prudent avoidance of over-priced stocks or is a sign of incompetence. As shortfalls grow, investors conclude incompetence and react by transferring funds to the outperforming managers, thereby amplifying the price changes that led to the initial underperformance and generating momentum.1

The dot-com boom

The technology bubble ten years ago illustrates this well. Technology stocks received an initial boost from fanciful expectations of future profits from scientific advance. Meanwhile, funds invested in the unglamorous, value sectors languished, prompting investors to lose confidence in the ability of their underperforming value managers and switch funds to the newly successful growth managers, a response which gave a further boost to growth stocks. The same thing happened as value managers themselves began switching from value to growth stocks to avoid being fired.

Through this conceptually simple mechanism, the model explains asset pricing in terms of a battle between fair value and momentum. It shows how rational profit seeking by agents and the investors who appoint them gives rise to mispricing and volatility. Once momentum becomes embedded in markets, agents then logically respond by adopting strategies that are likely to reinforce the trends. Explaining the formation of asset pricing in this way seems to provide a clearer understanding of how and why investors and prices behave as they do. For example, it throws fresh light on why value stocks generally outperform growth stocks despite offering seemingly poorer earnings prospects. The new approach offers a more convincing interpretation of the way stock prices react to earnings announcements or other news. It also shows how short-term incentives, such as annual performance fees, cause fund managers to concentrate on high-turnover, trend-following strategies that add to the distortions in markets, which are then profitably exploited by long-horizon investors. At the level of national markets and entire asset classes, it will no longer be acceptable to say that competition delivers the right price or that the market exerts self-discipline.

More micro modelling of the financial sector

It seems self-evident that the way forward must be to stop treating the finance sector as a pass-through that has no impact on asset pricing and risk. Incorporating delegation and agency into financial models is bound to lead to a better understanding of phenomena that have so far been poorly understood or unaddressed. Because the new approach maintains the rationality assumption, it makes it possible to retain much of the economist's existing toolbox, such as mathematical modelling, utility maximisation and general equilibrium reasoning. The insights, elegance, and tractability that these tools provide will be used to study more complex phenomena with very different economic assumptions. The new general theory of asset pricing that eventually emerges should relegate the efficient market hypothesis to the status of special and limiting case.

Concluding remarks

Of course, investors may not always behave in a perfectly rational way. But that is beside the point. If this new approach meets the criteria of relevance, validity, and universality required of any new theory, then it provides a valuable starting point in understanding markets. Models based on irrational behaviour can always be helpful in offering supplementary or more detailed insights. 

The impact of the new general theory will extend well beyond explaining asset prices and investors' actions.

  • Corporate finance and banking theory have both been developed under the pro-forma assumption of price efficiency and will now need to accommodate systematic mispricing.
  • Macroeconomics has also treated finance as a pass-through and would benefit from changing the economic emphasis and focusing more on the impact of agency and incentives in the savings and investment process. 
  • In the context of the recent crisis, governments and regulators can only rebuild and re-regulate banking and finance successfully if they have a better idea of how crises form.
  • Finally, economists may start to ask questions about the social value of the finance sector, its size, and complexity – questions that could be conveniently brushed under the carpet given the prevailing paradigm of efficiency.

References 

Fama, Eugene F. (1970), “Efficient Capital Markets: A Review of Theory and Empirical Work”, The Journal of Finance, Vol. 25, No. 2, Papers and Proceedings of the Twenty-Eighth Annual Meeting of the American Finance Association

Fama, Eugene F. and Kenneth R. French,Common Risk Factors in the Returns on Stocks and Bonds,” Journal of Financial Economics, 1993.

Vayanos and Woolley (2008), “An Institutional Theory of Momentum and Reversal,” The Paul Woolley Centre for the Study of Capital Market Dysfunctionality, Working Paper Series No.1.

Footnotes

1. We show that as long as fund flows are gradual, as in the real world, price changes are also gradual. Intuitively, rational long-term investors are eager to buy an undervalued stock even when the stock is expected to become more undervalued in the future because of the risk that undervaluation might instead disappear. We term this the “bird in the hand” effect.

This article may be reproduced with appropriate attribution. See Copyright (below).

    Posted by on Monday, October 5, 2009 at 03:26 PM Permalink  Comments (41)


    Comments

    Feed You can follow this conversation by subscribing to the comment feed for this post.

    -->