Ok Tim I’ve come across this idea about banking regulation I’ve been wanting to run past someone so here it is:
There are two related problems that need addressing. Firstly that the systemic risk in the banking system is largely born by governments and secondly that this systemic risk is very hard for regulators to identify in advance. Now in the UK we have a bank levy that is supposed to address the first problem by getting the government back some of the cost of bearing the systemic risk, but I think both problems could be addressed if we changed the bank levy slightly.
The obvious problem with the bank levy in dealing with the first problem is that it is levied on debt and not on actual risky behaviour. Now you might say that taxing risky behaviour would be impossible because, as stated in the second problem, it is very hard to identify. But what if we decided to simply do the best we could, simply take a few of HMRC’s finest and get them to make a best guess at which banks pose the most risk and tax them accordingly. Presumably this best guess would be pretty rubbish, but the effect of this would be to give every bank a big incentive to point out the dodgy practices of their rivals in order to convince HMRC to tax the other bank rather than them. Also anyone trying to address overly risky practices at their own bank would have something very valuable to take to a rival bank if they were not listened to.
Basically I think you could address the problem of identifying risk by getting each of the banks to do a bit of the regulator’s job for them. You would have created an open forum where everyone in the industry would come to discuss (argue about) who exactly was going to screw things up next. Which at the very least would be better than the wall of dismissively imperious smugness that has traditionally greeted anyone suggesting the banks were doing anything wrong.
I greatly enjoyed the talk. Looking forward to more
There is a big difference between designed systems (nuclear reactor) and emergent institutions (economy in general, financial system for example). In one case, the blame is centralized (owners and designers and operators) and in the other it is distributed.
That said, in both case who we hold accountable is key to creating a reasonably good solution.
I would say the first and possibly only solution (a meta-solution really) is to bundle control and accountability. In short, owners are responsible and accountable. This means relatively aligned incentives for starters.
Unfortunately, it is not so in the worlds of finance or insurance or nuclear energy.
If we keep incentives aligned, the risk of loss brings about prudence and creativity in managing risks responsibly. This may mean asking more transparency from systems you depend on (to avoid the A->B->C->A circular dependency of re-insurance, understanding exposure chains), creating buffers (such as reducing leverage), coming up with better indicators, committing to third-party audits or voluntary standards, and allowing the emergence of diversification (avoid a mono-culture and widerspread exposure) and experimentation (competition and innovation in how companies manage risks).
Things that broke this important rule and contribute to irresponsible risk-taking: cartelization of banks around the central bank, government deposit insurance (depositors become less prudent), bailouts (less shareholder skin in play) and generally regulations (control and accountability are separate). Regulations are numerous from standard way of evaluating risk (forcing a dependence on privileged rating agencies), forcing government bonds as “safe assets” (see Greek pension funds), and prohibiting insider trading (fired whistleblowers could make a fortune). Those create a mono-culture which is uniformly vulnerable rather than resilient. They socialize responsibility and dilute accountability instead of keeping it with the owners.
You mention regulators a few times in your talk. In the first part, regulators are involved in creating the incidents you recount. Yet in the later part, you invoke them as solutions. Please consider that central planning is the worse of the two available solutions to complex social problems in this imperfect world. All safety involves trade-offs and complex distributed knowledge. Regulators not only have misaligned incentives, there is little reason to think they could access or deal with such complex knowledge, or that their powers of regulations would not be captured.
As an oil industry risk engineer I can say you present a fascinatingly good analogy of process system risk identification and control to financial systems. Well done.
-Mark Fitzgerald
4 Comments
clarke ching says:
Very nice Tim! This might sound strange but you’ve provoked many thoughst about how we build and test software-based systems. So thank you.
18th of July, 2012Eric_A_Blair says:
Ok Tim I’ve come across this idea about banking regulation I’ve been wanting to run past someone so here it is:
There are two related problems that need addressing. Firstly that the systemic risk in the banking system is largely born by governments and secondly that this systemic risk is very hard for regulators to identify in advance. Now in the UK we have a bank levy that is supposed to address the first problem by getting the government back some of the cost of bearing the systemic risk, but I think both problems could be addressed if we changed the bank levy slightly.
The obvious problem with the bank levy in dealing with the first problem is that it is levied on debt and not on actual risky behaviour. Now you might say that taxing risky behaviour would be impossible because, as stated in the second problem, it is very hard to identify. But what if we decided to simply do the best we could, simply take a few of HMRC’s finest and get them to make a best guess at which banks pose the most risk and tax them accordingly. Presumably this best guess would be pretty rubbish, but the effect of this would be to give every bank a big incentive to point out the dodgy practices of their rivals in order to convince HMRC to tax the other bank rather than them. Also anyone trying to address overly risky practices at their own bank would have something very valuable to take to a rival bank if they were not listened to.
Basically I think you could address the problem of identifying risk by getting each of the banks to do a bit of the regulator’s job for them. You would have created an open forum where everyone in the industry would come to discuss (argue about) who exactly was going to screw things up next. Which at the very least would be better than the wall of dismissively imperious smugness that has traditionally greeted anyone suggesting the banks were doing anything wrong.
18th of July, 2012Julien Couvreur says:
I greatly enjoyed the talk. Looking forward to more
There is a big difference between designed systems (nuclear reactor) and emergent institutions (economy in general, financial system for example). In one case, the blame is centralized (owners and designers and operators) and in the other it is distributed.
That said, in both case who we hold accountable is key to creating a reasonably good solution.
I would say the first and possibly only solution (a meta-solution really) is to bundle control and accountability. In short, owners are responsible and accountable. This means relatively aligned incentives for starters.
Unfortunately, it is not so in the worlds of finance or insurance or nuclear energy.
If we keep incentives aligned, the risk of loss brings about prudence and creativity in managing risks responsibly. This may mean asking more transparency from systems you depend on (to avoid the A->B->C->A circular dependency of re-insurance, understanding exposure chains), creating buffers (such as reducing leverage), coming up with better indicators, committing to third-party audits or voluntary standards, and allowing the emergence of diversification (avoid a mono-culture and widerspread exposure) and experimentation (competition and innovation in how companies manage risks).
Things that broke this important rule and contribute to irresponsible risk-taking: cartelization of banks around the central bank, government deposit insurance (depositors become less prudent), bailouts (less shareholder skin in play) and generally regulations (control and accountability are separate). Regulations are numerous from standard way of evaluating risk (forcing a dependence on privileged rating agencies), forcing government bonds as “safe assets” (see Greek pension funds), and prohibiting insider trading (fired whistleblowers could make a fortune). Those create a mono-culture which is uniformly vulnerable rather than resilient. They socialize responsibility and dilute accountability instead of keeping it with the owners.
You mention regulators a few times in your talk. In the first part, regulators are involved in creating the incidents you recount. Yet in the later part, you invoke them as solutions. Please consider that central planning is the worse of the two available solutions to complex social problems in this imperfect world. All safety involves trade-offs and complex distributed knowledge. Regulators not only have misaligned incentives, there is little reason to think they could access or deal with such complex knowledge, or that their powers of regulations would not be captured.
Cheers,
20th of July, 2012Julien
Mark Fitzgerald says:
As an oil industry risk engineer I can say you present a fascinatingly good analogy of process system risk identification and control to financial systems. Well done.
22nd of July, 2012-Mark Fitzgerald