Archive for June, 2010
In a recent post, Rajiv Sethi questions the tendency to find behavioural explanations for financial crises and argues for an ecological approach instead – a sentiment that I agree with and have touched upon in previous posts on this blog. This post expands upon some of these themes.
A More Realistic View of Rationality and Human Cognition, Not Irrationality
Much of the debate on rationality in economics focuses on whether we as human beings are rational in the “homo economicus” sense. The “heuristics and biases” program pioneered by Daniel Kahneman and Amos Tversky argues that we are not “rational” – however, it does not question whether the definition of rationality implicit in “rational choice theory” is valid or not. Many researchers in the neural and cognitive sciences now believe that the conventional definition of rationality needs to be radically overhauled.
Most heuristics/biases are not a sign of irrationality but an entirely rational form of decision-making when faced with uncertainty. In an earlier post, I explained how Ronald Heiner’s framework can explain our neglect of tail events as a logical response to an uncertain environment, but the best exposition of this viewpoint can be seen in Gerd Gigerenzer’s work which itself is inspired by Herbert Simon’s ideas on “bounded rationality”. In his aptly named book “Rationality for Mortals: How People Cope with Uncertainty”, Gigerenzer explains the two key building blocks of “the science of heuristics”:
- The Adaptive Toolbox: “the building blocks for fast and frugal heuristics that work in real-world environments of natural complexity, where an optimal strategy is often unknown or computationally intractable”
- Ecological Rationality: “the environmental structures in which a given heuristic is successful” and the “coevolution between heuristics and environments”
The irony of course is that many classical economists had a more accurate definition of rationality than the one implicit in “rational choice theory” (See Brian Loasby’s book which I discussed here). Much of the work done in the neural sciences confirms the more nuanced view of human cognition espoused in Hayek’s “The Sensory Order” or Ken Boulding’s “The Image” (See Joaquin Fuster on Hayek or the similarities between Ken Boulding’s views and V.S. Ramachandran’s work discussed here).
Macro-Rationality is consistent with Micro-Irrationality
Even a more realistic definition of rationality doesn’t preclude individual irrationality. However, as Michael Mauboussin pointed out: “markets can still be rational when investors are individually irrational. Sufficient investor diversity is the essential feature in efficient price formation. Provided the decision rules of investors are diverse—even if they are suboptimal—errors tend to cancel out and markets arrive at appropriate prices. Similarly, if these decision rules lose diversity, markets become fragile and susceptible to inefficiency. So the issue is not whether individuals are irrational (they are) but whether they are irrational in the same way at the same time. So while understanding individual behavioral pitfalls may improve your own decision making, appreciation of the dynamics of the collective is key to outperforming the market.”
Economies as Complex Adaptive Systems: Behavioural Heterogeneity, Selection Pressures and Emphasis on System Dynamics
In my view, the ecological approach to macroeconomics is essentially a systems approach with the emphasis on the “adaptive” nature of the system i.e. incentives matter and the actors in a system tend to find ways to work around imposed rules that try to fight the impact of misaligned incentives. David Merkel explained it well when he noted: “People hate having their freedom restrained, and so when arbitrary rules are imposed, even smart rules, they look for means of escape.” And many of the posts on this blog have focused on how rules can be subverted even when economic agents don’t actively intend to do so.
The ecological approach emphasises the diversity of behavioural preferences and the role of incentives/institutions/rules in “selecting” from this pool of possible agent behaviours or causing agent behaviour to adapt in reaction to these incentives. When a behaviourally homogeneous pool of agents is observed, the ecological approach focuses on the selection pressures and incentives that could have caused this loss of diversity rather than attempting to lay the blame on some immutable behavioural trait. Again, as Rajiv Sethi puts it here: “human behavior differs substantially across career paths because of selection both into and within occupations….[Regularities] identified in controlled laboratory experiments with standard subject pools have limited application to environments in which the distribution of behavioral propensities is both endogenous and psychologically rare. This is the case in financial markets, which are subject to selection at a number of levels. Those who enter the profession are unlikely to be psychologically typical, and market conditions determine which behavioral propensities survive and thrive at any point in historical time.”
In a previous post, I quoted Richard Fisher’s views on how bailouts cause business cycles and financial crises: “The system has become slanted not only toward bigness but also high risk…..if the central bank and regulators view any losses to big bank creditors as systemically disruptive, big bank debt will effectively reign on high in the capital structure. Big banks would love leverage even more, making regulatory attempts to mandate lower leverage in boom times all the more difficult…..It is not difficult to see where this dynamic leads—to more pronounced financial cycles and repeated crises.”
Fisher utilises the “incentives” argument but the same argument could also be made via the language of natural selection and Hannan and Freeman did exactly that in their seminal paper that launched the field of “Organizational Ecology”. Hannan and Freeman wrote the below in the context of the bailout of Lockheed in 1971 but it is as relevant today as it has ever been: “we must consider what one anonymous reader, caught up in the spirit of our paper, called the anti-eugenic actions of the state in saving firms such as Lockheed from failure. This is a dramatic instance of the way in which large dominant organizations can create linkages with other large and powerful ones so as to reduce selection pressures. If such moves are effective, they alter the pattern of selection. In our view, the selection pressure is bumped up to a higher level. So instead of individual organizations failing, entire networks fail. The general consequence of a large number of linkages of this sort is an increase in the instability of the entire system and therefore we should see boom and bust cycles of organizational outcomes.”
Richard Fisher of the Dallas Fed delivered a speech last week( h/t Zerohedge) on the topic of financial reform, which delivered some of the most brutally honest analysis of the problem at hand that I’ve seen from anyone at the Fed. It also made a few points that I felt deserved further analysis and elaboration.
The Dynamics of the TBTF Problem
In Fisher’s words: “Big banks that took on high risks and generated unsustainable losses received a public benefit: TBTF support. As a result, more conservative banks were denied the market share that would have been theirs if mismanaged big banks had been allowed to go out of business. In essence, conservative banks faced publicly backed competition…..It is my view that, by propping up deeply troubled big banks, authorities have eroded market discipline in the financial system.
The system has become slanted not only toward bigness but also high risk…..if the central bank and regulators view any losses to big bank creditors as systemically disruptive, big bank debt will effectively reign on high in the capital structure. Big banks would love leverage even more, making regulatory attempts to mandate lower leverage in boom times all the more difficult…..
It is not difficult to see where this dynamic leads—to more pronounced financial cycles and repeated crises.”
Fisher correctly notes that TBTF support damages system resilience not only by encouraging higher leverage amongst large banks, but by disadvantaging conservative banks that would otherwise have gained market share during the crisis. As I have noted many times on this blog, the dynamic, evolutionary view of moral hazard focuses not only on the protection provided to destabilising positive feedback forces, but on how stabilising negative feedback forces that might have flourished in the absence of the stabilising actions are selected against and progressively weeded out of the system.
Regulatory Discretion and the Time Consistency Problem
Fisher: “Language that includes a desire to minimize moral hazard—and directs the FDIC as receiver to consider “the potential for serious adverse effects”—provides wiggle room to perpetuate TBTF.” Fisher notes that it’s difficult to credibly commit ex-ante not to bail out TBTF creditors – as long as the regulator retains any amount of discretion with the purpose of maintaining systemic stability, they will be tempted to use it.
On the Ineffectiveness of Regulation Alone
Fisher: “While it is certainly true that ineffective regulation of systemically important institutions—like big commercial banking companies—contributed to the crisis, I find it highly unlikely that such institutions can be effectively regulated, even after reform…Simple regulatory changes in most cases represent a too-late attempt to catch up with the tricks of the regulated—the trickiest of whom tend to be large. In the U.S. financial system, what passed as “innovation” was in large part circumvention, as financial engineers invented ways to get around the rules of the road. There is little evidence that new regulations, involving capital and liquidity rules, could ever contain the circumvention instinct.”
This is a sentiment I don’t often hear expressed by a regulator – As I have opined before on this blog, regulations alone just don’t work. The history of banking is one of repeated circumvention of regulations by banks, a process that has only accelerated with the increased completeness of markets. The question is not whether deregulation accelerated the process of banks’ maximising the moral hazard subsidy – it almost certainly did and this was understood even by the Fed as early as 1983. As John Kareken noted, “Deregulation Is the Cart, Not the Horse”. The question is whether re-regulation has any chance of succeeding without fixing the incentives guiding the actors in the system – it does not.
Bailouts Come in Many Shapes and Sizes
Fisher: “Even if an effective resolution regime can be written down, chances are it might not be used. There are myriad ways for regulators to forbear. Accounting forbearance, for example, could artificially boost regulatory capital levels at troubled big banks. Special liquidity facilities could provide funding relief. In this and similar manners, crisis-related events that might trigger the need for resolution could be avoided, making resolution a moot issue.”
A watertight resolution regime may only encourage regulators to aggressively utilise other forbearance mechanisms. Fisher mentions accounting and liquidity relief but fails to mention the most important “alternative bailout mechanism” – the “Greenspan Put” variant of monetary policy.
Preventing Systemic Risk perpetuates the Too-Big-To-Fail Problem
Fisher: “Consider the idea of limiting any and all financial support strictly to the system as a whole, thus preventing any one firm from receiving individual assistance….If authorities wanted to support a big bank in trouble, they would need only institute a systemwide program. Big banks could then avail themselves of the program, even if nobody else needed it. Systemwide programs are unfortunately a perfect back door through which to channel big bank bailouts.”
“System-wide” programs by definition get activated only when big banks and non-banking financial institutions such as GE Capital are in trouble. Apart from perpetuating TBTF, they encourage smaller banks to mimic big banks and take on similar tail risk thus reducing system diversity.
Shrink the TBTF Banks?
Fisher clearly prefers that the big banks be shrunk as a “second-best” solution to the incentive problems that both regulators and banks face in our current system. Although I’m not convinced that shrinking the banks is a sufficient response, even a “free market” solution to the crisis will almost certainly imply a more dispersed banking sector, due to the removal of the TBTF subsidy. The gist of the problem is not size but insufficient diversity. Fisher argues “there is considerable diversity in strategy and performance among banks that are not TBTF.” This is the strongest and possibly even the only valid argument for breaking up the big banks. My concern is that even a more dispersed banking sector will evolve towards a tightly coupled and homogenous outcome due to the protection against systemic risk provided by the “alternative bailout mechanisms”, particularly the Greenspan Put.
The fact that Richard Fisher’s comments echo themes popular with both left-wing and right-wing commentators is not a coincidence. In the fitness landscape of our financial system, our current choice is not so much a local peak as a deep valley – tinkering will get us nowhere and a significant move either to the left or to the right is likely to be an improvement.
The Franken Amendment draws upon Richardson and White’s idea of a centralised clearing platform which I had criticised earlier. This proposal is based upon a flawed understanding of the structured products’ ratings process and the incentives guiding the agencies during this process and arises from a false extrapolation of the corporate and sovereign bond ratings process into the realm of structured products.
The fatal flaw in our ratings regime is not the issuer-pays model but the fact that ratings agencies only get paid if the bond is issued. In the structured products space, the difference between a potential AAA rating and a AA rating is not just that a higher spread is paid to the investor on the bond. The lower rating usually means that the bond will not be issued at all, which means that the ratings agency will not earn any fees. This problem cannot be solved even if we have a single monopolistic ratings agency paid by the SEC, so long as the fees are payable only upon issuance of the bond. As I have discussed earlier in more detail, ratings agencies are incentivised not only to expand market share but to expand the size of the market for rateable securities.
Let me explain the logic with a simple example. A pension fund approaches a bank for a bespoke AAA tranche on a portfolio of mortgage-backed securities. The bank constructs an appropriate tranche paying Libor + 100 bps and asks for a rating, upon which the clearing platform allocates it an agency. The agency comes back with a AA rating instead – so what does the bank do in this instance? It cannot change the tranching without damaging its own economics and the client will not accept a AA tranche paying the same coupon. So the deal just does not get done and the ratings agency is left without any fee for its opinion.
Let us go a little further along this chain of thought – all competing agencies are similarly stringent in their ratings and discover after six months that their earnings and dealflow have collapsed! At this point, they will of course gradually start easing their ratings requirements and sooner or later we will end up in the same position we were in before the crisis hit us. Its worth noting that this outcome does not change if someone other than the issuer pays the agency or even if we have a monopolistic ratings agency. Provided that the agency is a profit-maximising entity, the removal of direct competition may slow the process of easing of ratings criteria, but it will not change the end result.
In fact, the above example is too generous as it ignores the ease with which the centralised platform process can be gamed by banks. The central problem here is the fact that there are a multitude number of structured bonds that can fulfill a typical client request, such as the one above. For example, let us assume that the bank above constructs a tranche from a portfolio of MBS and applies to the platform which allocates it to Moody’s. If Moody’s comes back with an unsatisfactory rating, it cancels the issuance, makes a small modification to the portfolio and tranching and tries its luck again. The process can continue until the bank gets allocated to a more friendly ratings agency and the desired rating is achieved.
The fundamental issue here is that tinkering with the system in this manner is futile – the problems inherent in our current financial system are too fundamental and we have only two choices as I hinted at in an earlier post. We can either put in place blunt and almost certainly efficiency-reducing regulations or we can move towards a free-market system where the implicit and explicit protection provided to the banking sector is removed in a credible and time-consistent manner. To give you a simple example of a blunt regulation that will reduce the potential for ratings arbitrage, we could legislate that if a portfolio of sub investment-grade assets cannot be tranched to produce a AAA tranche. The price we pay for such regulations is that we eliminate a significant proportion of legitimate tranching, but this trade-off is unavoidable.