> Bail-in will only be useful when the banks have built up enough bail-inable capital, and depending on what is the definition of bail-inable capital, most bank

> Bail-in will only be useful when the banks have built up enough bail-inable capital, and depending on what is the definition of bail-inable capital, most banks are not there yet I'm glad to hear you say this. There is definitely a wide gap between the reality of available bail-in capital today and the promise of the theory if it were available. > However all the draft regulations that are being prepared now point to relatively high requirements, in the region of 25-30% of the Risk Weighted Assets, which should be ample to absorb a very large loss. Theoretically. But the problem is the "Risk Weighted Assets". How do you do the weighting? A great deal of work has gone into this (as you know), but RWA calcs existed pre-crisis, as did specific RWA tiers for securitized instruments. We failed to properly weight the risks before and nothing says we won't do so again. The Fed is still wrestling with "advanced approaches" to RWA and the last time I checked finalizing the requisite approach was on "indefinite delay". Point being... we don't have any assurance here. And we don't want to repeat the mistakes of overconfidence in our prowess of risk-assessment that we made last time around. > So if the crisis happens this week, it won't help, if it happens in 5 years or after it will certainly help a lot, and be a first line of defence before contemplating a bail out. I agree with you fully here. We also need to address what happens in the more extreme cases (and we need to go further in preventing the likelihood of more extreme cases). > On the separation of commercial and investment banks, I am not convinced it actually helps. Reproducing another of my comments on this article: You were replying to me in that other comment as well. :) As I mentioned there, I also think depository banks should be more regulated in a) the total risk they can take on, and b) what sorts of investments they can make. (So the total quantity of risk and the type of risk.) The depository banks in the UK were not regulated enough, clearly. I don't see how combining poorly regulated depository banking risk with poorly regulated investment banking risk would possibly help. Imagine Lehman directly combined with RBS... it's an even bigger disaster. Besides, my contention is not that retail banking = safe while investment banking = risky. Both are risky. It's that contagion is bad. Increased correlation is bad. Combining retail banks and investment banks is a bad idea both theoretically and as proven in practice. We should contain risk. Let's allow some institutions (investment banks) to create complex derivatives, advanced securitizations, make markets, participate in diverse investments, trade fairly liberally and generally do what investment banks do. Let's put that type of risk in one bucket, and still regulate the total risk they can take on, the means by which they are unwound in crises, etc. Let's have a separate bucket of risk for depository/retail banks, which is as separated as possible (in an interconnected and fast-moving economy and financial system) from that bucket of risk. The only possible reason not to separate these two buckets of risk is if you think they diversify each other. But that's not right even theoretically and it definitely has not been the case in practice.