I have a good deal of respect for Gillian Tett, but I disagree with her analysis in a Financial Times piece today:
A decade ago, it was fashionable for Western consultants, bankers and business people to decry Japan’s domestic service industry. For Japanese business sectors, ranging from milk production to financial broking, have long been plagued by complex distribution chains and numerous middlemen.
So, Anglo-Saxon consultants – such as McKinsey – would regularly urge the Japanese to reform their distribution chains, and flourish data showing how much more “efficient” the US was than Japan in sectors such as retailing.
Back then, I was working in Tokyo as a reporter. So I dutifully reported those studies-cum-sermons on the evils of middlemen.
However, amid all that debate about American efficiency, one point that Western commentators almost never discussed was the proliferation of middlemen in America’s financial world.
If you were to sketch a map of how credit has been sliced and diced in 21st century banking, for example, there would be so many stages – and commission-hungry middlemen – in that process, the Japanese dairy industry might seem positively rational. Yet, for many years the apparent contradiction went almost entirely unnoticed, by Western politicians, bankers, and consultants alike. Middlemen were regarded as bad in Japan; but they were somehow overlooked in America’s financial world.
As incredible and counterintuitive as it may sound, Tett has the basic premise wrong. I worked for the aforementioned McKinsey in the 1980s and the firm had tons of analyses showing how securitization was fundamentally cheaper than on-balance sheet lending. The reason was the cost of bank equity and FDIC insurance.
Let us put it another way: if old fashioned banking was cheaper, there would be no obstacle to going back to that system. The economics would be obviously superior and the banks would be able to raise equity to reclaim the securitization business they ceded.
The problem is that this more efficient system entailed information loss. There was no free lunch. By reducing the role of the bank, you eroded credit screening, had no ability to monitor the borrower, and wound up losing the ability to do workouts (that was not inherent to the securitization model, you could in theory have servicers do workouts, but the focus was on streamlining processes as much as possible. The servicers became factories with standardized processes, and successful mods require work and tailored solutions. That is an underappreciated reason why so many mortgage mods are failing. We are using a vastly different template than the traditional one).
Another symptom of too much rather than too little efficiency is that the system is tightly coupled, meaning processes move so quickly that they cannot be interrupted. As Richard Bookstaber pointed out in his Demon of Our Own Design, tightly coupled systems are unstable.
Tett hits another possible wrong note:
And securitisation has produced a particularly curious – or absurd – paradox. A few years ago, it was widely assumed that the process of slicing and dicing credit would create a more “complete”, free-market financial system.
The concept of “completing markets” has nothing to do with structured credit. It refers to a 1953 paper by Kenneth Arrow and Gerard Debreu, which is arguably the most celebrated paper in economics, the Arrow-Debreu theorem (personally, I take a dim view of it, but my opinion on such matters is of no importance).
Arrow-Debreu establishes that a general equilibrium can be achieved, but under absurdly restrictive assumptions: there are hedging markets for everything conceivable (all places and all future periods) meaning from now till the end of time, AND everyone had perfect foreknowledge (everyone knows what will happen until the end of time, so they know what they might want to hedge). The idea of “completing markets” is the justification for creating more derivatives, since you need derivative of all sorts of crazy flavors to hedge all the contingencies the universe will offer from now until the end of time. As I understand it, the role of hedges in Arrow Debreu is strictly about hedging real economy exposures. My impression is that risk shifting within the financial realm is not a feature of that model although some may argue it is implied (but A-D would seem to call for it via a hedging instrument, rather than the less efficient mechanism and as we now know ineffective means of creating bundled products that in fact were poor means of reducing risk).
I’d be curious to get reader input, but Arrow Debreu as a justification for structured finance, as opposed to derivatives, seems a pretty big stretch. (financial markets tend to play a limited role in economic models, mainly to provide important inputs like the price of money),