In defence of the leverage ratio

Regular readers know that I blame risk-based capital requirements for many of the ills of our current banking system. Before the introduction of Basel regulations, banks’ capital level used to be assessed using more standard and simple leverage ratios (equity or capital/total assets). Those ratios have mostly disappeared since the end of the 1980s but Basel 3 is now re-introducing its own version (Tier 1 capital/total exposures).

While I believe there should not be any regulatory minimum capital requirement, I also do believe that, if regulators had to pick one main measure of capitalisation, it should be a standard leverage ratio. All RWA-based ratios should be scrapped.

A new study just added to the growing body of evidence that leverage ratios perform better than RWA-based ones as predictor of banks’ riskiness. Andrew Haldane, from the BoE, has been a long-time supporter of leverage ratios. Admati and Hellwig also backed non-risk weighted ratios. Another paper recently suggested that there was nothing in the literature that justified the level of risk-weights.

Still, most economists, central bankers and regulators consider leverage ratios as mere backstops to complement the more ‘scientific’ (read, more complex, as there is no science behind risk-weights) Basel RWA-based ratios. See this speech from Andreas Dombret, which sums up most criticisms towards simple leverage ratios:

Yet a leverage ratio would also create the wrong incentives. If banks had to hold the same percentage of capital against all assets, any institution wanting to maximise its profits would probably invest in high-risk assets, as they produce particularly high returns. This would eradicate the corrective influence of capital cover in reducing risk.

Unfortunately, Mr. Dombret and many others are very misinformed.

A leverage ratio would not incentivise banks to leverage up to the allowed limit. Under Basel’s RWA framework, no banks operated with the bare regulatory minimum. Critics forget that different banks have different risk aversion and different risk/reward profiles. Some banks generate relatively low RoEs in return for lower level of risk. Others are willing to take on more risks to generate higher margins and higher RoEs. Banks are not uniform.

Banks would not necessarily pile into the riskiest assets under a leverage ratio either. The answer to this is the same as above. Banks have different cultures and different risk/return profiles to offer to investors. There is no reason why all banks would suddenly lend to the riskiest borrowers to improve their earnings. Such criticism could also easily apply to Basel capital ratios: why didn’t all banks follow the same investment strategy? Critics forget that banks do not try to maximise their profits. They try to maximise their risk-adjusted profits. Finally, such argument only demonstrates its proponents’ ignorance of banking history, as if all banks had always been investing in the riskiest assets in the 300 years before Basel introduced those risk-weights.

RWA-based capital ratios are very patronising: because the riskiness of the assets is already embedded within the ratio, banks are effectively telling markets how risky they are. This became overly sarcastic with Basel 2, which allowed large banks to calculate their own risk-weights (i.e. the so-called ‘internal rating based’ method). It has been proved that, for a given portfolio of assets, risk-weights were considerably varying across banks (see here and here). Given the same balance sheet, one bank could, say, report a 10% Tier 1 ratio, and another one, 14%, implying massive variations in RWA density (RWAs/total assets). A given bank could also change its RWA calculation model (and hence its RWA density) in between two reporting periods, making a mockery of period to period comparison. Of course, all this is approved by regulators. Consequently RWA-based capital ratios became essentially meaningless.

As a result, a leverage ratio would provide a ‘purer’ measure of capitalisation that markets could then compare with their own assessment of banks’ balance sheet riskiness.

Scrapping RWA-based capital ratios would also provide major economic benefits. As regularly argued on this blog, RWA-incentivised regulatory arbitrage has been hugely damaging for the economy and is in large part responsible for the recent internationally-coordinated housing bubbles and ‘secular’ low level of business lending. Getting rid of such regulatory ratio would benefit us all by removing an indicator that has big distortionary effects on the economy.

Of course, there are still a few issues, though they remain relatively minor. The main one is that differences in accounting standards across jurisdictions do not lead to the same leverage ratios (i.e. US GAAP banks have much less restrictions to net their derivative positions than IFRS ones). But those accounting issues can easily be corrected if necessary for international comparison. The second one is what definition of capital to use: common equity? Tier 1 capital? Another problem is the fact that very low risk banks, which don’t need much capital, would also get penalised. In the end, it’s likely that any regulatory ratio will prove distortive in a way or another. Why not scrap them all and let the market do its job?

Chart: Guardian