What the financial sector does: an intuition
[Note: in all these recent posts, I’ve deliberately ignored the money-creation aspect of banking in a fractional-reserve system. In all honesty, I’ve done this because I haven’t quite figured out how to fit it into the story yet. I am aware of the issue and am thinking about how to work it in.]
The thing I find most simultaneously infuriating and rewarding about this whole blogging exercise is that writing down what I’m thinking in public forces me (purely for reasons of avoiding intellectual embarrassment) to spend a lot of time thinking about the more speculative and “out-there” things I write and what the truth in them may or may not be. This post is a continuation of my recent theme, and please accept it in the speculative and open spirit in which it is offered.
In structured finance – say, a mortgage-backed security – you take the total cashflows of the underlying mortgages and assign them to different ‘tranches’. The super-safe debt at the top has first right to any cashflow (thus are safer), and the equity tranches at the bottom have the last right (thus are risky, but get a high return if there is any remaining cash).
One way of thinking about what the financial sector does as a whole is it structures the cashflows of the entire economy according to our preferences for financial assets. Term deposits and things like that are kind of like the super-senior tranche of an MBS – low return, very low risk. And at the other end, owning equities is like, well, the equity tranche of an MBS. A well-functioning financial sector should co-ordinate the desires for financing (mortgages, business loans, IPOs etc.) with preferences for different kinds of financial assets (equity, safe bonds, junk bonds etc.), and at a system-level it does this by essentially pooling together and divying up cash.
The wider the divergence in the non-financial sector between the preferences for risk in real investments and preferences for risk in owning financial assets, the more structuring of the underlying cashflows you have to do. But the ability to structure risky things into safe things is highly, highly dependent on correlation. Owning one sub-prime mortgage is risky, but lots of them put together aren’t – assuming there’s no correlation in default.
Now, in the case of MBS, one thing that was correlated within particular mortgage pools was often underwriting. Or, strictly speaking,
shitty non-existent poor underwriting. Default risk due to the fact that the underwriting wasn’t done properly was (to my understanding) common across particular securitizations as the mortgage pools often came from the same originators. Defaults were correlated (as well as being higher than expected), and so the supposedly safe senior tranches turned out to be not so safe.
This is what happens at a system-wide macro level when you introduce correlated risks in the real economy. Furthermore, insofar as different banks are not the same in terms of their risk exposure towards different sectors of the economy, in that starts mattering which sectors (for example, those most vulnerable to flagging aggregate demand) individual banks are more exposed to. One reason the MBS market crashed was because there were a lot of defaults, but another reason was because the underlying mortgage pools in different securitizations had regional and other biases that then made it matter a lot what the nature of those mortgages were. But no one really knew this, other than the bankers who put the deal together. Even the ratings agencies who rated it wouldn’t have seen the details all the underlying mortgages! So, the market tanked as you had all these securities which you essentially couldn’t value, because the theory that led to their prior valuation had broken down and there wasn’t the information available to conduct that valuation yourself (and besides it would be phenomenally complicated to do so even if you had it).
This seems to me like a mirror of what happens in the financial sector as a whole when the entire economy is hit by a negative AD shock. Most of the time, no one needs to really know exactly what each bank is exposed to, because the theory that turns risk into safety is working. Introduce correlation and then the nature of the underlying risks starts mattering to the safety of all these supposedly low-risk assets.
On an economy-wide level, the maturity and risk structure of our financial assets does not match that of our non-financial assets. We have loads of houses and office buildings and factories etc. – that is to say, illiquid high-risk non-financial assets – and lots of liquid, low-risk financial assets. In order to achieve this outcome, the financial sector has to structure the cashflows of the entire economy. And it seems to me that the greater that structuring is geared towards the production of low risk, safe and liquid financial assets, the more vulnerable the system is to correlated risks and crises of confidence.
So whilst structured finance was in some sense a creature of the late 20th and early 21st century, it seems to me there’s also a very important sense in which it has been around forever. When I first started reading about all this stuff 2-3 years ago, I initially thought about securitization (something I didn’t understand at all) as basically like banking (which I understood a little more, though not much). But now I wonder whether thinking about it the other way round may be helpful. Maybe the problems we’ve seen played out most clearly in securitization and structured credit could potentially be illuminating for thinking about the vulnerabilities inherent in the basic functions of the financial sector as a whole.