Pick Yer Poison, or, If It Ain’t Baroque, Don’t Fix It
The BofE’s Andrew Haldane is getting a lot of attention-and rightfully so-for his critique of complex financial regulation, most notably Basel Faulty. Yes, there is a lot to criticize in complex financial regulation-and lord knows, I’ve indulged-but it’s essential to emphasize that simplicity is no panacea. Indeed, complex regulations evolve precisely from the deficiencies of simpler ones.
That is, you have to ask: Why did the rules become complex? Was complexity valued for complexity’s sake? Or did complexity develop/evolve because of the recognized defects of simpler structure?
Most likely. Rules become more complex in an attempt to respond to attempts to exploit simpler rules. There is a regulatory dialectic. The regulators and the regulated interact in a way that leads to spiraling complexity.
Take bank capital requirements. Simple capital requirements specify, say, capital equals a certain percentage of assets. This is fine if assets are identically risky, or the regulated have no ability to tilt their portfolios towards risky assets.
But that’s exactly what happened in the old, simple world with simple capital requirements. Banks looked for riskier assets in order to expropriate their depositors, lenders, and deposit insurers. With capital requirements insensitive to risk, banks had a tremendous incentive to make their assets as risky as possible.
There are two responses to this. One is to limit the nature of the assets banks can hold. The other is to make capital requirements depend on the riskiness of assets.
Both are great in theory, devilish hard in practice. Risk-based capital requirements (or limitations on the assets that banks can hold) demand that regulators have the ability to make discriminating evaluations of the relative riskiness of assets. Even more crucially, they demand that regulators know at least as much as those they are regulating.
Neither condition is likely to hold in practice. And again, there is a regulatory dialectic, as banks sniff out-or create-assets that are “underpriced” from a capital requirements perspective. That is, they identify-or create-assets with risk prices implicit in capital charges that do not reflect true risks, and which satisfy the restrictions on the nature of assets that banks can hold. (Think of AAA ABS of various types. They’re AAA! What could go wrong?) Thus, portfolios end up being riskier than regulators had anticipated.
So they adjust the capital requirements. But as in any regime of price controls-and capital requirements are essentially risk price controls-the regulators don’t have the necessary information to set the prices properly, meaning that some risks are always underpriced. The capital requirements become more complex, setting off a new round of attempts to identify the underpriced and overpriced risks.
What starts simple becomes progressively more complex. Regulations and capital requirements end up being bizarrely baroque not because people like it that way, but because the less baroque rules were exploited.
Put differently: complexity is endogenous, and almost inevitable. Simple sounds nice, but simple is easy to exploit, and complexity blooms in order to counter this exploitation.
Haldane argues that during the last crisis institutions subject to simple rules were less likely to fail than those subject to more complex ones. This is superficially plausible, but again, remember that the rules are endogenous, complexity is endogenous. There was no experiment, natural or otherwise, that randomly assigned some institutions to simple rules and others to more complex ones. The rules and the institutions co-evolved, endogenously. I would bet that there was some other constraint on the complexity of the firms subject to the simple rules that made it unduly costly for them to outwit the simple rules. They stayed simple as a result, so the rules could stay simple too.
Maybe Haldane is right. Maybe there is a simple but robust way of regulating financial institutions. But I have my doubts. Complexity results from the failures of the simple.
Haldane is stating, in an ironically more complex way, the old Indian story of the village beset by mice that brought in cats to rid itself of the rats, then brought in dogs to rid itself of the infestation of cats, then brought in elephants to drive out the resulting packs of dogs-and responded by bringing back the mice to scare away the elephants. Seeing the effects of complexity (elephants) the simpler problems (mice) don’t look so bad.
But don’t be deceived. If Haldane were to get his way, the cycle would begin again. This is the fallen condition of homo economicus.
[…] That said, Streetwise Professor on why simplicity in regulation is no panacea. […]
Pingback by Links: What Pulp Fiction taught us about LIBOR « Rhymes With Cars & Girls — September 20, 2012 @ 8:27 am
Clearly, and maybe by chance, all the nasty stuff only happened once the really simple Basle 1 was modified to include trading books etc etc. The simple fact is that under 1992 Basle 1, all the stuff which caused the GFC would not be part of the banking system. It would be, as it was in the 90’s, done by investment banks who relied on ratings to gather liabilities and all counterparties would have have much lower limits to investment banks of a similar rating to a regulated bank with that rating. So the same mistakes would have been made, but they would not have been as significant
Comment by campbell dawson — September 20, 2012 @ 7:20 pm
Craig,I’ve rarely been able to disagree with you. But after reading your comment above, I thought about Dodd-Frank. Thousands of pages ordering thousands of regulations. I don’t know how that legislation was written, but I have a hard time believing all those pages were necessary to correct imperfections in previous legislation.
Comment by enronal — September 20, 2012 @ 7:54 pm
Question: Does your analysis apply to private contractual clauses such as bond covenants? Aren’t bond covenants equally the sort of thing smart operators would try to circumvent, leading to increasingly complex language? And if so, how do those markets work, especially when the assets are widely traded on secondary markets?
I’m just trying to get a handle on the range of applicability one should place on this endogenous complexity argument.
Comment by srp — September 20, 2012 @ 8:50 pm
@enronal-1. Thanks re rare disagreement. 2. I agree Frankendodd was not an example of the regulatory dialectic I describe. Put differently, complexity can arise in different ways.
Perhaps I can summarize the post in a different way. It seems that many are nostalgic for a simpler time, with simpler rules. That is very attractive, but even when you try to go back, it is very hard to stay there. The simple rules are exploitable, leading to increasingly complex rules in attempts to mitigate the exploitation.
Another example of this process in action was derivatives regulation pre-CFMA. The rules were simple: futures had to be traded on exchange. So firms devised products that had futures-like features, but weren’t called futures. So there was a stream of ad hoc rulings and no action letters, etc. that attempted to deal with these things. By the mid-90s the regulation was a confusing thicket, and there was little legal certainty.
Frankendodd is an example of another kind of phenomenon. Post-crisis regulation.
But maybe there is a similarity after all. The basic ideas in Dodd-Frank were established early. Take just the derivatives title. The goals were simple. Clearing more derivatives, increased capital and collateral on non-cleared derivatives. Trading swaps on centralized markets. When the legislation was being written, however, the difficulties became apparent, and the dialectic worked in real time. The simple concepts rapidly morphed into a massively complicated piece of legislation which spawned an even more massive, and more complicated set of rules.
@srp. Yes, I think endogenous complexity occurs in private settings too. To reiterate what I just wrote to @enronal, my point in the post was meant to be, er, simple. Simplicity seems appealing, but it is not an equilibrium, not a steady state. That’s all. Nothing more, nothing less.
OK, so we can get endogenous rule complexity in private settings. I agree. Next logical question: Do private parties do a better job than regulators of trading off Haldane’s drawbacks of complexity with your drawbacks of exploitable simplicity? And is there a stable (not necessarily optimal) interior equilibrium in your theory for either public or private actors?
This is a pretty fundamental topic. It’s related to the vertical relations, agency in corporate governance, and formal institutional design literature, but none of those do a bang-up job of characterizing complexity.
Comment by srp — September 21, 2012 @ 5:12 pm
@srp. Yes, fundamental. Complexity is inherently difficult to model formally. Indeed, modeling is all about simplifying . . . and the very idea of simple representations of complexity is paradoxical.
Transactions cost economics and property rights economics and incomplete contracting models get at some of these issues.
Re equilibrium. I don’t know what the equilibrium is. The main point of my post is that when it comes to things like banking regulation, simple rules are not the equilibrium-or are unlikely to be so. So when Haldane and others preach the virtues of simplicity, they are advocating something that’s unsustainable. It’s not really in our choice set.
The main virtue of private arrangements is that profit signals and survival mechanisms permit more discriminating trade-offs in the complexity-simplicity dimension. They also are less informationally demanding than centrally imposed rules. Or put differently, there is a closer joining up between those that have the information and those who make the decisions. Also, the one-size-fits-all nature of regulation means that mistakes tend to be systemic in nature.
Looks like we have at least 4 more years of regulatory repression thanks to Romney’s blunders.
Comment by Surya — September 22, 2012 @ 7:11 pm
SWP, I will jump on the rare wagon of disagreeing with you also. The simple-complex axis you contemplate is more of a dance than a dialetic. For example, the universe is must more complex today than right after the big bang, but we do not need have an increasingly complex regulator (god) to guide the evolutionary path of life. Deep complexity exists where it clearly does not have to, and often simple systems exist which would operate much better with a more complex algorithm- (many examples available). The degree of complexity is not a conditional variable on the health of a (financial) system. The important variable instead is an environment of trial and error. Good decisions win, and poor decisions fail. The ISDA contract is fantastically complex due to the trial and error of private negotiations and many dollars lost/won with earlier versions. FrankenDood, Basel and all the their forebearers hopelessly lack any pass/fail mechanism. Regulators do not go to jail or lose money on poor regulations, generally they only get lucrative jobs by creating a loophole for a favored constituent (the dance). Instead, an environment of creative-destruction for financial decision-making is important. In some areas the environment will be complex and others simple… but suffering fair and unprivileged consequences for all counterparties is what is important. It is not a measure of their complexity.
Complex derivatives are fantastic instruments for a company to transfer risk. For example a “Cat Bond” is complex due the inherent nature of the risk. It does not have complexity for complexity sake, over an over-engineered structure; but clearly a regulator could never determine the correct amount of risk-based collateral for a Cat bond. Let the few companies who understand Cat bonds trade them, and eliminate regulators and complex collateral rules entirely. Buyer beware.
Effective regulation comes from fear of losing money, not from a bureaucrat.
Comment by scott — September 23, 2012 @ 4:43 pm
Good points, Scott. I’ve long thought laws and regulations were complex because politicians like them that way. Complexity makes exceptions, exemptions, and other loopholes valuable.
Comment by emerich — September 23, 2012 @ 8:50 pm
@scott-I don’t disagree with your comment, and I really don’t think you disagree with me, particularly if you look at my previous comment in response to @srp.
Again-the point of my post is that those who claim that there is a possibility of having a stable, simple regulatory system are wrong. You may start trying to KISS, but you end up complicated. (You can also start complicated, e.g., Frankendodd, Basel Faulty, and get even more complicated.) That is, the post focuses on government regulation, in direct response to Haldane and others who have called for simplicity in that regulation.
Your point really relates to private orderings, which in turn relates to my response to @srp’s question: “Does your analysis apply to private contractual clauses such as bond covenants?” You give a couple of additional examples-the ISDA master and cat bonds. You say “The important variable instead is an environment of trial and error. Good decisions win, and poor decisions fail. . . . In some areas the environment will be complex and others simple… but suffering fair and unprivileged consequences for all counterparties is what is important. It is not a measure of their complexity.”
Compare that to my response to @srp: “The main virtue of private arrangements is that profit signals and survival mechanisms permit more discriminating trade-offs in the complexity-simplicity dimension. They also are less informationally demanding than centrally imposed rules. Or put differently, there is a closer joining up between those that have the information and those who make the decisions. Also, the one-size-fits-all nature of regulation means that mistakes tend to be systemic in nature.”
I really don’t see any daylight between your thinking and mine. We both emphasize evolutionary mechanisms: I call them “survival mechanisms” and you call them “an environment of trial and error” and “creative destruction.” I mention information, which you don’t. You mention incentives (“regulators do not go to jail or lose money on poor regulations”)-I don’t here, but I have elsewhere.
So I think we are in a complete agreement that private orderings are more likely to achieve the right amount of complexity based on the union of our justifications: evolutionary/creative destruction elimination of maladapted arrangements; superior information held by the private negotiators; and superior incentives by private agents as opposed to public ones. Or at least, we can agree that for these reasons private orderings are likely to be superior to regulatory/public ones.
To which @emerich’s public choice argument can be added: yes, the incentives of regulators and legislators are in fact perverse because they can sell indulgences.
To summarize: 1. The post is about whether simple public regulation can survive. 2. Private orderings can be complex. 3. Private orderings are far more likely to achieve the Goldilocks level of complexity than public ones.
Everybody happy now?
Good thread, and I hope the Professor is not too exasperated with people reacting to such a fundamental and interesting topic. HIs last comment provides a clear summary.
There is another school of thought about regulation that tries to escape from the simplicity-complexity dance by moving orthogonally away from the rule-of-law spectrum where it takes place. That school advocates “expert” or “prosecutorial” discretion as a substitute for codified rules of any kind. (Our friends at the SEC have traditionally preferred this mode, resisting precise definitions for “insider trading,” for example.) The theory here is that by letting the regulator use an “I know it when I see it” approach, efforts at clever circumvention of loopholes are thwarted. Their response to the obvious objections regarding lack of ex ante certainty, arbitrariness, and abuse of power is that the regulator should be subjected to a regime of accountability where the public or its representatives can likewise evaluate its performance on a discretionary basis. This idea seems terrible to me as it casts every transaction into a political free-for-all, but it does short-cut the simplicity-complexity dynamic.
Comment by srp — September 26, 2012 @ 5:02 pm
I am happy… and enjoy your eloquent prose. excellent writing- thank you. We do agree comprehensively on the 3 issues in your summary. Where we may disagree still is regarding the simple/complexity axis contained in the original post. I contest that “…Complexity results from the failures of the simple”. Instead I think complexity results for myriad reasons, often unknown, like the Cambrian Explosion of complexity in life-forms. Financial complexity can arise during a immense growth phase of innovation, or from a crisis and heavy discipline for conditional rulemaking. Most often complexity has nothing to do with the underlying complexity of the phenomenon, and more to do with historical developments, random winners or losers.
The Simple regulation promulgated by Haldane is a populist ruse for the attention of rule-makers, and his discourse only distracts the debate from the necessity of a market-mechanism to restore a “survival mechanism”, borrowing your words. It is ttue that private ordering is better than public, but the critical axis is not public/private or simple/complex, but life/death axis for financial players (non-privileged survival mechanisms). Until financial firms are allowed to go bust, absolutely no genius can come up with a better framework for regulation. Haldane the Brilliant is nothing more than royal bluster.
Comment by scott — September 26, 2012 @ 5:20 pm