Comments on: The biggest weakness of Basel III A slice of lime in the soda Sun, 26 Oct 2014 19:05:02 +0000 hourly 1 By: tariqscherer Wed, 15 Sep 2010 16:55:07 +0000 Okay, I might be throwing a spanner in the works here, but aren’t there more effective risk management methodologies than just straight historical VaR?

The guys at RiskMetrics have been warning us for ages, well before the current crisis, of the risk underlying purely historical measures (and yes, this is even before Taleb’s black swan hit the public lexicon). And yet, from a regulatory point of view, we still try and define our risk on a calendar, five year, ten year, line in the sand kind of way, when we know that the reason markets are good is because they can react on information immediately: not with a particular timeframe in sight, but with a ‘real’ profit mindset.

Certainly, when it comes to futures and forwards, we’ve understood that margining is an ongoing process that needs to be constantly attuned so why is it that banks can not be given the same review/monitoring/risk management outlook? (I wrote an article on this recently at

Thank you Felix for highlighting the issue of sovereign risk but even sovereign risk need not become systemic. If it is acknowledged that even your own country’s securities still require some level of hedging then there should be some scope to aver the worse. There should be a level of freedom in the market to split the political operational risk from the real intrinsic economic risk – if markets are not allowed to incentivise this kind of rational behaviour then we are leaving ourselves open to ongoing confrontation between markets and local spendthrifts government: we should be able to work together on the long run, not just against each other.

Tariq Scherer

By: drewiepe Wed, 15 Sep 2010 14:51:44 +0000 A few things. First I’m pretty sure nothing should be 0% risk weighted. Second, yes, risk weightings are going to continue to be a big part of the real problem. Third Basel III fails to address the issue of the hegemony of the ratings agency cartel.

On another note, I remain fairly convinced that there isn’t any way to completely remove the chance that banks blow up. Banks are completely unlike any other corporate structure – they are inherently levered and can therefore be wiped out by a nasty run on the bank, even if perfectly solvent. So what’s the answer? The minute we admit that TBTF exists, we have to charge for the privilege of being TBTF rather than try to regulate it out. Insurance premiums payable to a pre-funded disaster recovery fund, that are heavily geared to (true) leverage and spiral exponentially above 30x leverage sounds like a good start. Throw in additional premia based on the % of revenues from trading, the type of funding that props up the balance sheet and the liquidity of underlying assets, and Bob’s your uncle, we might have the beginnings of something that works.

By: MJGW Wed, 15 Sep 2010 14:41:35 +0000 Re the RW discussion, it seems to me BCBS fully realises its limitations hence the introduction of the Leverage Ratio which sort of aims to backstop any model/model input error. Interestingly to see how the off B/S items (incl. netting) will be factored into this measure. Off course, too low weightings will allow for gaming the system.

Re the Lehman, BCBS pushes for Central Counterparties (allowing low RW) which should limit losses when derivatives are closed-out through adequate collateral mngt processes and multilateral netting. Note BCBS has also proposed to take longer time windows and increased haircuts for bilateral agreements.

BTW I’m surprised by all the complexity BCBS raises about the Trading Book. Why couldn’t they have just introduced Holding Period Limits (for Equity and Debt like instruments), breach of which leads to capital surcharges or alternatively -and rigourously- full deduction? In this way, banks would take on illiquid assets more prudently and if not willing to hold them at punitative capital levels, sell them onwards. It would seem to me that the market would do its work on valuing such illiquid assets.

By: Greycap Wed, 15 Sep 2010 13:50:42 +0000 I think that by “mean variance”, Nick Gogerty is referring to mean-variance; that is, a statistical model parameterized by mean, variance, and correlation. I’m not exactly sure what Nick’s beef was because he wasn’t very specific.

What I would say: the problem is *not* the lack of “fat tails.” It is easy to construct mathematical models with fat tails, but impossible to calibrate them. This should be obvious given that we can’t reliably calibrate the parameters we already have (I think this might have been Nick’s point, in which case we agree.)

The underlying difficulty is that the distributions of asset returns are not endogenous properties; they are functions of what people think about the assets, and opinions are highly unstable, non-linear, full of feedback loops, etc. One could say that assets don’t really have return distributions. The consequence of this is that any model of return distributions must be continually updated, if modeled returns are to resemble observable returns. And the consequence of updating is that risk measures derived from these models are strongly pro-cyclical. You just have to look at historical time series of asset vols and returns to see what I mean.

So there is still a powerful pro-cyclicality baked into the foundations of B3. The ad hoc measures like variable capital ratios duct-taped onto the structure won’t be enough to mitigate this; not by an order of magnitude.

By: Greycap Wed, 15 Sep 2010 13:24:48 +0000 Felix, it is a bit rich for you to be saying that you have only just noticed this little problem with B3. Your own blog commenters have brought up the points you mention and more besides. This issue of risk weights has long been a hobby horse of Per Kurowski, and he left a comment on your blog just the other day, did he not?

By: NickGogerty Wed, 15 Sep 2010 11:59:48 +0000 Basel ii and Basel iii are both inherently based on a mean variance assumptions of risk. This is deeply flawed. Mean variance is nothing more than naive trend following which tends to severely over estimate component co-variance stability.

The more complex the system and bucketing of risk, the more fun people will have gaming it. Basel iii will most likely make the CDO act of stuffing 10 lbs of crap into a 5 lb bag look like amateur hour. The bag will of course be tiered capital requirements.

The only thing we can hope is that Basel gets implemented and people have so little faith in it that they force each other to be more transparent and not just really on a single metric such as tiered capital ratios.

I am now convinced the regulatory solution has 2 answers. No international bank regulation in which case all participants are wary of each other and tread more carefully or a reporting regime so full of holes that all participants are wary of each other and tread more carefully knowing the system is bogus and to be gamed.

By: GingerYellow Wed, 15 Sep 2010 08:47:30 +0000 “Since it did not change this risk-weighting, Basel III effectively doubles down on Basel II. ”

Well, that depends what you mean. Basel III certainly keeps the backward looking risk weighting approach, but it did dramatically increase the risk weighting on the assets that brought down Citi et al (CDOs of ABS). And, probably more importantly in the long run, it removed the major source of arbitrage under Basel II, the disparity in treatment between securitisations in the trading book and the banking book. For the most part, the problem wasn’t just banks loading up on CDOs. It was banks loading up on illiquid CDOs in the trading book, where they didn’t even have to hold the small amount of capital they would have had to hold if the assets were in the banking book, only a small market risk charge. Now, securitisation positions in the trading book get effectively the same treatment as the banking book, and the market risk framework has been significantly tightened up. As a result, the average bank is going to be holding about 2-3 times as much capital against its total trading book as in the past, even before you take into account the basic capital ratios.

By: Polycapitalist Wed, 15 Sep 2010 05:09:40 +0000 Felix,

I haven’t seen any mention of off-balance sheet assets and Basel III. Would appreciate it if in a future article you can address this issue. Thanks.