Bookblogging: After the EMH, what next?

Writing a critique of the Efficient Markets Hypothesis in terms rigorous enough to stand up to scrutiny, but comprehensible to the average reader hasn’t been easy, and I still have a lot more work to do. But thanks to the help I’ve had from commenters here and at my blog, and from other readers, I hope to make a go of it. Now comes the hard bit: suggesting some alternatives, both in theory and policy. I’m not by any means satisfied with this draft. In particular, I need to go back and get a better linkage to the question “if the market price for assets is not the “right” price, what is?”. But, I thought I’d do better getting some help and criticisms now, rather than trying for some more polish first.

What next?

Realistic theories of financial markets 

The theoretical analysis underlying the EMH shows that perfectly rational investors, operating in perfectly efficient financial markets, will produce the best possible estimate of the future value of any asset. The catastrophic failure of the EMH in reality suggests the need to re-examine not only the theoretical premises of individual rationality and market efficiency and the whole concept of “best possible estimate”.

The first of these tasks is well under way. For the past twenty-five years or so, economists have been seeking to replace assumptions of perfect rationality with more realistic models of how individuals make choices under uncertainty and over time. Much of this work goes under the banner of ‘behavioral economics’ or ‘behavioral finance’ (in true academic fashion, there is some dispute about the ownership of this term, with some economists trying to tie it to a specific research program, and others preferring a broader view that encompasses any work based on actual behavior rather than a priori rationality assumptions). 

Many of the crucial ideas of behavioral economics are derived from the work of psychologist Daniel Kahneman and his longtime collaborator, the late Amos Tversky. In 2002, Kahneman became the first, and so far only psychologist to be awarded the Nobel prize in economics, while Tversky, almost uniquely in the history of the award, received a posthumous mention. Among other crucial ideas, Kahneman and Tversky showed that people have difficulty in handling probability judgements and, in particular, tend to overweight certain kinds of low-probability events, such as the chance of winning the lottery or dying in an airplane crash. 

Moreover, while people are mostly risk-averse, they tend to “chase losses”, taking additional risks in the hope of recouping losses from an original reference point. Far from being the reliable calculating machines assumed in the standard theory, people rely on ‘heuristics’ such as ‘availability’, which means that they tend to overestimate the probability of events of which examples are readily available. 

Another collaborator of Kahneman and Tversky, Richard Thaler has focused on how people make decisions over time. The standard model requires people to value future flows of income using a moderate, constant discount rate, such as the rate of interest on bonds. So, if the annual rate of interest on bonds is 5 per cent, a sum of $100 invested now will be worth $105 in a year’s time and (because of compound interest) about $110.25 in two years’ time. Turning this argument around, a sum of $105 received in a year’s time, or $110.25 in two years’ time, should be worth $100 today.

Observing what people actually do in day-to-day decisions reveals a quite different pattern, called hyperbolic discounting. People greatly prefer to receive benefits immediately rather than, say, in a year’s time. They are similarly keen to defer costs from the present into the relatively near future, even when facing high interest costs for doing so. But if asked to choose between a benefit (or cost) in one year’s time, and a larger benefit or cost in two year’s time, they are fairly patient. 

Unawareness

Unawareness

The study of behavioral economics shows that people often fail to follow the precepts of rational decision making. And the economics of asymmetric information literature gives reasons why even when participants in financial markets are entirely rational, market outcomes may not be efficient. But there is a more fundamental challenge which economists are only now beginning to address. This is the fact that, since the number of possible contingencies that may affect economic outcomes is effectively infinite, no decisionmaker, no matter how well-informed and sophisticated, can possibly take them all into account.

This point has arisen in popular discussion, for example with Donald Rumsfeld’s famous observation ‘There are known knowns. There are things we know that we know. There are known unknowns. That is to say, there are things that we now know we don’t know. But there are also unknown unknowns. There are things we do not know we don’t know.’ Although Rumsfeld was much derided for this statement, is both valid and important. The real problem was that, having made the point, Rumsfeld did not consider that, since launching a war exposes a nation to a host of ‘unknown unknowns’, decisions to do so should be made with extreme caution.

In the financial literature, writer and investor Nassim Taleb has popularised the term ‘black swans’ to describe such unforeseen contingencies. For Europeans, the proposition that “all swans are white” was confirmed by all experience. It seems unlikely that Europeans ever contemplated the possibility of a black swan, until they came to Australia and found them. Fortunately, there was not a large financial system built on the whiteness of swans. However, history is full of examples of careful planning brought undone by unconsidered possibilities.

It is not hard to point out that we are regularly surprised by ‘unknown unknowns’ and ‘black swans’. A much harder problem is to describe a system of reasoning and decisionmaking that takes account of events that are, by definition, unforeseen by those reasoning and making decisions. Economists, philosophers and decision theorists have been wrestling with this problem for a long time, and at last seem to be making some progress. It turns out that it is possible to develop formal models of bounded rationality in which decisionmakers are unaware of some possibilities and unable to fully articulate and communicate all the possibilities of which they are aware.

The implications are profound. One is that in environments where surprises are likely to be unfavorable, it makes sense to apply a precautionary principle to decisionmaking, preferring simple and easily understood choices to those that are complex and poorly understood, even when the complex option appears to offer greater net benefits. A similar point relates to contracts. In joint work with Simon Grant and Jeff Kline, I have shown that contracts between boundedly rational parties always involve some element of ambiguity. For this reason, simple contracts with terms that are understood by both parties may be preferred to the complex arrangements indicated as optimal by standard economic theory.

Trust and crises

But individual deviations from rationality aren’t the only problem. In perfectly efficient markets, even a small number of of hardnosed and rational speculators would be enough to ensure the outcomes predicted by EMH theory. Such speculators could take advantage of the irrational behavior of the majority of investors, turning them into “money pumps”. As Keynes observed though, successful speculation depends on the market getting things right, not just eventually, but before the speculators run out of money. A realistic theory of financial markets must  explain how bubbles can persist long enough to deter speculators from betting on a return to market equilibrium.

We also need a deeper understanding of financial collapses like the current crisis. Although the textbooks represent financial markets as involving impersonal exchanges of precisely defined assets, the actual operation of the system relies crucially on trust and more generally, on understanding the amount of trust that should be placed in particular kinds of promises. In the last few decades, economists have spent a lot of time studying trust, and particularly the problem of when one party to a contract should trust the other to tell the truth and keep faith. This problem has been analysed  in terms of asymmetric information (when one party knows something the other does not, and both parties know this). But the problems go deeper than this, to situations where it is impossible to calculate all the possible outcomes, and individuals must decide whether to rely on the judgement and good faith of others.

Trust breaks down in crises. All institutions, both public and private, rely to some extent on trust, and when trust breaks down it is often hard to rebuild. In the crisis of the 1970s, the failure of governments to deliver on their commitments to manage the economy and maintain full employment led to a loss of public trust, which was transferred (more or less by default) to markets and particularly financial markets. This loss of trust made it difficult, if not impossible, to implement policies that might have made a difference, such as agreements to stabilise wages. By the time such agreements were feasible, in the 1980s, the balance of economic power had already shifted to financial markets.

Even more than governments (which have, after all, the direct power of the state behind them) financial markets depend on trust. The central financial institution of modern capitalism is the fractional reserve banking system, whereby banks lend out most of the money that is deposited with them, keeping only a fraction to meet calls for withdrawals. In an unregulated system, a failure of trust in a given institutions leads to a ‘run on the bank’ as depositors scramble to get money out while they can.

Systems of deposit insurance and bank guarantees now ensure that depositors’ trust in banks is backed up by their trust in the ability of governments to protect them in the event of default. But other kinds of trust in the financial system are not so easily maintained. Banks are sustainable if and only if they can accurately assess the willingness and ability of borrowers to repay their debts. In normal conditions, this is not an exceptionally difficult task. Banks can look at standard measures of ability to repay, credit histories and so on to distinguish good risks from bad and, in any case, the first group are in the overwhelming majority. 

But in a crisis all this breaks down. Formerly reliable formulas cease to work as borrowers realise they are better off walking away from their debts (through bankruptcy or foreclosure) than struggling to repay them and failing anyway. At this point, trust can only be restored through personal knowledge of particular borrowers, the kind that is built up through a long business relationship. But it is precisely in a crisis that such business relationships break down. Banks fail and their assets are taken over by others with no knowledge of the customers beyond what they can glean from formal records and remaining employees. In a complex and interlinked system like that built up over recent decades, the failures can cascade until the entire system ceases to function beyond a minimal level.

Financial regulation

The global financial crisis has been, above all, a failure of models of financial regulation based on the EMH.The approach to financial regulation developed in response to the Depression was highly restrictive. Financial institutions were confined to a limited range of services, and financial innovation was limited. Financial institutions seeking to create new assets had to satisfy regulators that these assets could be fitted into the existing regulatory framework, or else wait until a new set of regulatory structures was developed.

Financial deregulation in the 1970s put an end to these constraints. The term ‘deregulation’ is something of a misnomer, since no system in which the public is the ultimate guarantor can be regarded as unregulated. Rather a system of regulation focused on protecting the public and stabilising the economy was replaced by one in which the primary concern was to facilitate innovation and to manage risk in the most ‘light-handed’ possible fashion. The EMH played a crucial role in designing these regulatory systems, which went through a variety of forms before their final embodiment in the Basel Accords issued by the Basel Committee on Banking Supervision, which is made of up senior representatives of bank supervisory authorities and central banks from the G-10 countries.

The Basel Accords attempted to assess the riskiness of banks’ holdings using a combination of market prices and ratings from private agencies (such as Moody’s and Standard & Poors). According to the EMH, market values of classes of risky assets are the best possible estimate of their value. While the EMH does not have direct implications for the interpretation of agency ratings, the fact that such ratings are sought by bond issuers, implies, according to the EMH that the ratings contain valuable and reliable information, since they would otherwise be ignored.

Until 2007, the Basel system had never had to deal with a serious financial crisis in the developed world. It failed catastrophically at its first test. Not only did many banks fail, but the measures of capital adequacy required under the Basel system proved all but useless in assessing which banks were at risk. 

Radical changes in financial sector regulation have already taken place as a result of the financial crisis. Guarantees of bank deposits have been introduced or greatly expanded in all major economies. Partial or complete nationalisation of failing institutions, with the resulting assumption of risk by the public, has been widespread.

However, these policies have been introduced as emergency measures, with the implicit (and sometimes explicit) premise that they will be ended when normal (pre-crisis) conditions are restored. his premise is untenable. By the time the crisis is over, the financial sector will be radically transformed, and will require a radically different mode of regulation.

The starting point for a stable regulatory regime must be a reversal of the burden of proof in relation to financial innovation. The prevailing rule has been to allow, and indeed encourage, financial innovations unless they can be shown to represent a threat to financial stability. Given an unlimited public guarantee for the liabilities of these institutions such a rule is a guaranteed, and proven, recipe for disaster, offering huge rewards to any innovation that increases both risks (ultimately borne by the public) and returns (captured by the innovators).

What is needed is a system of ‘narrow banking’ where with a clearly defined set of institutions (such as banks and insurance companies) offering a set of well-tested financial instruments with explicit public guarantees for clients, and a public guarantee of solvency, with nationalisation as a last-resort option. Financial innovations must be treated with caution, and allowed only on the basis of a clear understanding of their effects on systemic risk.

It is important not to suppress the activity of those willing to take risks with their own capital. As Adam Smith observed, (CH X, Part I)

The chance of gain is by every man more or less overvalued, and the chance of loss is by most men undervalued, and by scarce any man, who is in tolerable health and spirits, valued more than it is worth.

Smith argues that this characteristic over-optimism is crucial in promoting investment and enterprise. Later writers such as Keynes described attitudes to risk in terms of ‘animal spirits’, and noted, in the light of experience, the occurrence of periodic panics and depressions, in which animal spirits could not be roused. But even so, there is a clear need to allow scope for those with an optimistic view to chance their arm, while ensuring that the costs of the inevitable failures are borne by those concerned in the speculative investment and not by the community as a whole.

Stabilising financial markets does not mean that it is necessary to prohibit risky investments, or even to prevent speculators from developing and trading in risky new financial assets. What is crucial is that these operations should not threaten the stability of the system as a whole. Publicly regulated (and guaranteed) banks and other financial institutions  should be prohibited from engaging in speculative trade on their own account and from extending any form of credit to institutions engaged in such speculation, as they did with LTCM and its successors. Governments should commit themselves not to allow any bailout if speculators get into trouble, again as occurred with LTCM. After that speculators can safely be left to sink or swim. 

The starting point for a stable regulatory regime must be a reversal of the burden of proof in relation to financial innovation. The prevailing rule has been to allow, and indeed encourage, financial innovations unless they can be shown to represent a threat to financial stability. Given an unlimited public guarantee for the liabilities of these institutions such a rule is a guaranteed, and proven, recipe for disaster, offering huge rewards to any innovation that increases both risks (ultimately borne by the public) and returns (captured by the innovators).

Post-crisis financial regulation must begin with a clearly defined set of institutions (such as banks and insurance companies) offering a set of well-tested financial instruments with explicit public guarantees for clients, and a public guarantee of solvency, with nationalisation as a last-resort option. Financial innovations must be treated with caution, and allowed only on the basis of a clear understanding of their effects on systemic risk.

In this context, it is crucial to maintain sharp boundaries between publicly guaranteed institutions and unprotected financial institutions such as hedge funds, finance companies, stockbroking firms and mutual funds. Institutions in the latter category must not be allowed to present a threat of systemic failure that might precipitate a public sector rescue, whether direct (as in the recent crisis) or indirect (as in the 1998 bailout of Long Term Capital Management). A number of measures are required to ensure this.

First, ownership links between protected and unprotected financial institutions must be absolutely prohibited, to avoid the risk that failure of an unregulated subsidiary will necessitate a rescue of the parent, or that an unregulated parent could seek to expose a bank subsidiary to excessive risk. Long before the current crisis, these dangers were illustrated by Australian experience with bank-owned finance companies, most notably the rescue, by the Reserve Bank, of the Bank of Adelaide in the 1970s.

Second, banks should not market unregulated financial products such as share investments and hedge funds.

Third, the provision of bank credit to unregulated financial enterprises should be limited to levels that ensure that even large-scale failure in this sector cannot threaten the solvency of the regulated system.

The state and the market

The EMH implies that governments can never outperform (well-informed) financial markets in making investment decisions. The failure of the EMH does not imply the converse claim that governments will always do better. Rather, the evidence suggests at markets will do better than governments in planning investments in some cases (those where a good judgement of consumer demand is important, for example) and worse in others (those requiring long-term planning, for example).

This inference from capital market outcomes is consistent with the general experience of the 20th century, and particularly the decades after World War II. In these decades, governments took a central role in the development of a wide range of infrastructure services including transport and telecommunications networks, and the provision of electricity, gas and water. These investments were not, in general, motivated by doctrinaire socialism, but by a belief that the development of market economy would be promoted by the reliable supply of infrastructure services.

The rise of economic liberalism saw a substantial, though far from complete, shift of responsibility to the private sector. Reform of electricity, telecommunications and other infrastructure services … The results have been mixed, to put it as charitably as possible. In some cases need some examples, reliance on private capital has led to new and innovative investment strategies. In others, such as electricity transmission in the US, failure to take account of the public good character of infrastructure has led to inadequate investment by all parties, and to a gradual deterioration in the quality of the network. In still other cases, as in the creation of supposedly competitive electricity markets in California, financial engineering and market manipulation have produced catastrophic failures.

The experience of the 20th century suggests that a mixed economy will outperform both central planning and laissez faire. The economic doctrines derived from the EMH seemed to contradict that suggestion. It is now clear,however, that it is the EMH and not the mixed economy that has failed the test of experience.