Ooh look a chart:
That's from this quite punchy paper by Patrick Slovik of the economics department at the OECD. It shows you that, in 1992, big banks had risk-weighted assets, which determine how much capital they're required to have, of over 65% of their total assets, which measures how much lending and investing and trading and financing-their-governments they actually do. In 2008, the ratio was under 35%. What's special about those dates?
When the first Basel accord was implemented in 1992, risk-weighted assets represented close to 70% of bank total assets, which means that bank regulatory capital was calculated based on a large share of bank total exposures. In the years following the introduction of the Basel accord, the ratio of riskweighted assets to total assets (RWA/TA ratio) gradually decreased and reached about 35% in the immediate pre-crisis period, which means that the regulatory capital of systemically important banks was calculated based on only a small fraction of their total exposures. ... [T]he drop in the RWA/TA ratio has been very smooth since the implementation of the Basel accords without any significant deviations from the trend line until the crisis. This trend suggests that innovative engineering of regulatory risks and the move to unconventional business practices by systemically important banks has been a consistent trend for almost two decades and was not limited to a few years preceding the financial crisis. The trend reached its lowest point at the onset of the financial crisis when the capital requirements of systemically important banks were determined based on the historically lowest amounts of risk-weighted assets (relative to total assets). Risk-weighted regulation leads to unintended consequences as it encourages innovation designed to bypass the regulatory regime rather than to serve non-financial enterprises and households. Strengthening capital requirements based on risk-weighted assets may further contribute to these skewed incentives and their profitability.
This is a theory that has been sloshing around the internet for a good long while but this is neatly expressed and - that chart! It is - pretty stark, no?
That chart is worth the price of admission, but Slovik goes on to argue that risk-weighted capital requirements lead to a decline in lending activity: "One of the main reasons why non-loan-related activities have become so important for banks is the relatively high regulatory risk weights on loans relative to other types of assets, which puts them at a comparative disadvantage in the profit-seeking strategies of banks." The numbers look ... not totally unlike the risk-weighted-to-total-assets chart:
From this, Slovik draws the simple conclusion that banks should have their capital requirements set as a percentage of total assets, not of risk-weighted assets, because that would reduce incentives for banks to game the system by putting all their money in AAA CDOs and Greek government bonds. And, lest you worry that raising capital requirements (from call it 7% of RWAs to 7% of total assets) would be pretty expensive for banks, Slovik argues that "For a one percentage point increase in the equity ratio, a non-risk-weighted regulatory framework would reduce medium-term annual GDP growth on average by only 0.02 percentage point more than risk-weighted regulation." Specifically, increasing capital requirements by 1% of RWAs would reduce GDP growth by 4bps in the US and 6bps in Europe, while increasing capital requirements by 1% of total assets would reduce GDP growth by 5 and 11bps respectively.
That section's a little hand-wavy, and you can find things to object to. Like: that wasn't the question you were asked! If you think the platonic ideal of capital is % of total assets, and we're currently at % of RWAs, then moving to 10% of total assets is a much bigger percentage-point move than moving to 10% of RWAs. Also: even on these numbers, doubling the impact on GDP is kind of a big deal!
But you don't need to lose much sleep over the specifics, because the interesting claim here is not the proper capital level but the right way of measuring it. Here, of course, there's a reason that Basel relies on risk weighting rather than total assets, and that reason is that some assets really are safer than others. The pro-risk-weighting argument is that it prevents banks from chasing yield into reckless risky-but-profitable products, reduces systemic risk, and pushes banking activity into safe high-quality investments rather than unduly risky ones.
Except that that wasn't exactly true in the build-up to the financial crisis. Relatedly, Tracy Alloway and Robin Wigglesworth had a great thing in the FT yesterday evening about the demise of collateral loan obligations in Europe, and how that's going to cause trouble for the real economy:
Bankers are worried about how a wall of corporate debt set to mature in 2012 will be refinanced as the bulk of outstanding collateralised loan obligations – structured investment vehicles that buy loans made to private equity firms to finance acquisitions – goes into run-down mode.
Such CLOs have a finite life span after which they are not allowed to trade new loans for existing ones, or reinvest money received from repayments or interest on existing loans they hold.
By the end of next year, the majority of CLOs will have gone “static”. By 2014, more than 98 per cent of European CLOs will have have a reached the same point, according to a report by Standard & Poor’s. ...
“European banks have lots of this on their balance sheets already and are more keen to sell it to improve their capital ratios than anything else,” said Michael Hampden-Turner at Citigroup.
And why are these loans the first thing banks will get rid of in improving capital ratios? Because "capital ratios" means essentially "RWAs to equity," and leveraged loans get a higher risk weighting than, say, I'm again gonna go with Greek government bonds. Change the rules to regulate bank capital as a percentage of total assets, and higher-yielding loans to real-economy companies become a far more attractive place for banks to spend their scarce capital.
Or, don't change the rules, continue to encourage banks to stick to "higher quality assets," and see what you get. Maybe you get a more stable banking system at the cost of a negative effect on the real economy (as argued by this weirdly partially prescient paper from 2006). Or maybe you get further financial innovation that hides the destabilizing effect of risky assets and shifts them to the shadow banking system - while leaving banks less capitalized to withstand the risks that they retain. The dormancy of European CLOs suggests that, in the short term, the former answer is right. The experience of the last twenty years maybe points the other way in the long term.
If you're an Occupy Wall Streeter or a Sheila Bair or someone who just wants to get rid of mortgage tradeability, a magical unicorn wonderful thing to want is a thing that slows scary financial innovation without hurting the real economy - that cuts down on high-frequency-structured-whatnot voodoo without cutting down on lending and investing in worthwhile non-financial businesses. If you're one of those people, you have to have mixed-to-positive feelings about headlines like "Europe's CLO market in danger of extinction," since CLOs might seem to you like what got us into this mess in the first place. But if the result is real companies with real leveraged loans going bankrupt, that's not exactly what you want either. "Require banks to have a capital level based on total assets rather than risk-weighted assets," while not quite the perfect magic unicorn, is the sort of step that maybe accomplishes both goals.