The Economist just ran a long piece about the unintended, and largely unforeseen, consequences of collateral and clearing mandates.* And they aren’t all good. Anything but.
I say largely unforeseen because I’ve been jumping up and down about these consequences here on SWP, in some academic papers (including one to be published in the next issue of the Journal of Applied Corporate Finance), and numerous presentations. I have repeatedly emphasized that it was a fallacy that requiring more collateral would necessarily reduce the amount of leverage in the system, make the leverage in the system less fragile, or reduce interconnectedness among financial institutions. I have consistently argued that market participants would substitute alternative forms of leverage and utilize other financing techniques (e.g., collateral transformation and liquidity swaps) as fragile or more fragile than the leverage embedded in derivatives, and that these new forms of financing would just change the topology of the network of connections among firms in the financial markets, and that the new topology would be as or more complex than the pre-mandate topology.
It was always particularly maddening to see regulators and others present pictures of the OTC derivatives market with a spaghetti bowl of interconnections between market participants, and contrast it to a neat and tidy picture of a cleared market with a few neat lines connecting market participants with the CCP. There was no consideration of how the necessity of funding collateral and making variation margin payments would lead to the creation of financing arrangements among firms that would resemble the spaghetti bowl picture. No consideration of how the very complex financial system would adjust on the myriad margins available to it.
In some respects, I can make some allowances for the regulators and legislators who imposed this on the world. Those allowances are mainly a function of low expectations. I am less forgiving of the academic community. Many people smarter than I whom I respect have bought into the neat and tidy picture, and have dismissed concerns that the various Frankendodd mandates (and their European cousins) will create a new set of problems that may be worse than the ones that were supposedly fixed.
I think that in large part, this failure to anticipate the real effects of mandates, and the blithe confidence in the efficacy of collateral in particular, reflects the standard partial equilibrium, ceteris paribus way that economists use to approach problems. This method is very powerful, and I utilize it regularly, but it has its limits. In particular, it blinds one to the systemic responses to big changes like Frankendodd, to the fact that market actors will respond to these shocks on many dimensions. As I’ve said repeatedly, there has been a systematic failure to consider how the financial system as a whole would respond to massive interventions designed to reduce systemic risk.
This is a particularly hard problem to analyze mathematically and formally due to its inherent complexity. The daunting nature of the task, and the fact that formal mathematical modeling is almost required by top tier journals, has, in my opinion, made it unattractive for academics to pursue these issues. The constraints of the accepted toolset have led to the slighting of an extremely important and interesting issue.
But this is not to say that the problem is immune to analysis that can generate meaningful insights and falsifiable predictions. Indeed, largely disdained non-formal (“literary”) approaches can do so-and have done so. I started from a very simple premise, that there were fundamental economic considerations that had led market participants to choose financing arrangements and contracts with certain amounts of leverage and fragility. Constraining the ability to use those sorts of financing arrangements would induce actors to substitute towards arrangements with similar properties. Non-formal analysis also shed light on how these mechanisms would operate during periods of systemic stress.
Based on these non-formal approaches, I made predictions about how the markets would evolve under clearing and collateral mandates. Those predictions can be tested by empirical observation (though again, more descriptive than formally econometric). That’s as “scientific” (in the Friedman Methodology of Positive Economics sense) as the densest, or most elegant, mathematical model.
And as it turns out, in this instance, some of those predictions are being borne out. In ways that make me appreciate Cassandra.
I think it is more important to focus on important problems and find the best tools for the job at hand, rather than let narrow conceptions about the appropriate tools define and the problems you work on. The latter mindset has limited academic economists’ influence on big things like Frankendodd.
The one-eyed man is king in the land of the blind. Non-formal methods provide sight-and insight-on issues to which formal methods are often quite blind. Given the incentive structure in academia, however, I am skeptical that the lesson will be learned.
* Link fixed. H/T Phil R.