The perils of capital buffers

I have of course been wittering on for months about how the banks are really insolvent; indeed, I have been doing so for nearly as long as the regulators have been trying to work out new ways and means to stop it happening again.

But I was thinking this morning that perhaps bank collapses are a bit like forest fires. Much though we might try to "freeze" things as they are, it might be better in the long run to actually have a boom-bust cycle.

This is very much a "thought in progress" and I'm not sure where it might lead me. Coincidentally, a new report out this morning on Globalization covered a similar theory but from a different (risk-avoidance) aspect. It said that globalization brought a great many benefits, but it also created what had hitherto been unappreciated increases in systemic risk.

My own original thoughts had started along the lines of "if you created a new bank from scratch, what should you do?"

Well, I said to myself, I'd make sure that I didn't go broke by lending long and borrowing short.

This would imply matching loans to deposits not in absolute amounts, but also in time-scale. Savers, in other words, would have to lock up their money for longer, because borrowers locked up the bank's money for longer.

But then I thought to myself, whoah, that would have a negative impact on economic activity, simply because it would, of necessity, reduce the velocity of money. Getting a little bit of money to do a lot of work is one of the tricks of stimulating economic activity (we'll leave to one side whether this really is desirable. Let's just assume for the sake of this argument that it is.)

Then it struck me that banks borrowing short and lending long is actually a necessary part of modern economic actvity. And if the system works reasonably well (i.e., volatility is low), everything is fine.

Once again, there's a brilliant poker comparison. Look at Clarkatroid and Isildur. The former has run a very conservative bankroll strategy and has made a reasonably large amount of money fairly safely. Isildur, meanwhile, ran an insane bankroll strategy and went skint.

If we take this in theoretical terms, suppose you have a 40% chance of losing 5 buy ins and a 60% chance of winning 5 buy ins in any one session. If you play with a bankroll of 100 buy-ins, you are very unlikely to go broke in your lifetime. If you play with a bankroll of 5 buy-ins (i.e., at 20 times higher stakes), then there's a 40% chance that you will go broke on day one.

But is this necessarily the wrong strategy? After all, there's also a 60% chance that you will win 5 buy-ins, 20 times more than our original gambler. If you get through the early luck-related patch, and if you don't keep increasing your stakes then you will end up far far richer than the player who stuck to the conservative strategy.

In the banking world, this isn't so much a desireable option, as a necessary one. If you and four others start off as small conservative players, while five others start off with the risky strategy (effectively lending a higher proportion of their capital at the beginning), then two of the riskier players will go bust straightaway, one more will go bust soon after, and two will succeed. These will then get much bigger than the smaller players, and the power of size will see the smaller players either bought or crowded out.

Equally inevitably, as far as I can see, is that this process will continue. The "lucky" risky lenders will continue to grow faster, and risky strategies become endemic.

Eventually, it all blows up.

Now, the plan, as far as I can see, is to introduce capital regulations that will force the banks to be the conservative types of lenders rather than the risky types. That, the governments say, will stop the recent disasters.

But is this for the best? Let's look at it in terms of overall economic activity.

In our old scenario, we have 10 banks and, say, $10m in capitalisation.

Five of them lend out their capital at 100% at 10% interest with a 98% payback
Five of them lend out their capital at 1,000% (which entails choosing riskier loans) at 10% with a 95% payback.

This generates $490k in returns for group 1, and $2.5m in returns for group 2, even though two of them go bust. More importantly, it creates $55m worth of loans, which can be put to work in the economy.

In our new, "safer" scenario, only $10m of economic activity is created.

Implicitly, therefore, more stringent stress testing will result in lower levels of economic activity. Of course, this is obvious when you think about it for a second, but what we seem to have been saying is that the economic activity from 2002 through to 2008 in the US, based as it was on lending which should never have taken place (because the people doing the borrowing were never going to pay it back), was "fake". But it wasn't fake at all. It was very real. As the Youngster observed -- these people got five years in a house that they would not have had otherwise. The activity wasn't "fake", but it was unsustainable.

But, thinking about it again, is this bad? If you get volatility in living standards, but an overall higher average (i.e., if boom and bust has better boom times than it has worse bust times) then is this better or worse than a relatively stable period of slow growth in living standards?

Well, I suppose it depends on your point of view, because the "volatile" system will create relative winners and losers. But it will also generate faster wealth creation, which means that, eventually, the people at the "bottom" of the cycle (e.g. those who are just the "wrong age" at every point in the economic cycle) are still better off than they would have been if 200 years earlier a policy of slow-but-steady growth had been put in place.

And there's no absolute proof to my theory (although it feels right). Perhaps slow-but-steady has equal periods above and below the "volatile" line, meaning that the loss of voatility has no impact on overall economic growth. Perhaps the very downside of the volatility outweighs the happiness generated by the overall greater wealth.

But I think that it's a factor that the governments need to consider when they carry on with their mantra of "stability, stability".