Sunday, February 17, 2008

Scrap heap for financial models

(Prudent Bear) It is now clear that the credit crunch was not due simply to bull market over-optimism, but resulted very largely from the failures of a number of the financial models that have been a staple of the last generation. As the crunch spreads its malign tentacles ever wider into every corner of global economic life, the dust of collapse after collapse isn’t even beginning to clear. However there are now coming to be things one can usefully say about those models, and about the assumptions on which they were based.

From what appeared to be a modest glitch in the mortgage market, the damage to the world of financial modeling is ever-extending, and has now come to be surprisingly widespread, involving huge swathes of modern financial theory:

  • Subprime mortgages turned out to be correlated with each other, so that securities apparently rated AAA were in reality dangerously concentrated in a particular sector of the market that could and did collapse.
  • That also blew out the theory surrounding monoline insurance, that a well capitalized insurance vehicle could insure debt representing a large multiple of its capital base, without its bond rating or profitability coming into question.
  • Then there was asset backed commercial paper, under which commercial paper of short term maturity was issued by a shell company against the backing of financial assets of a long term maturity, and through this means removed from a bank’s balance sheet – it turned out that in a financial crisis the funding for these vehicles dried up.
  • Finally, credit default swaps are showing signs of strain, and may have turned out to have concentrated risk in unsuitable hands rather than spreading it as had been promised for them. In particular the counterparty problem in the CDS market becomes acute when defaults rise to a substantial level and declared debt ratings turn out to be unreliable.

As well as the instruments themselves, their risk management turns out to have been flawed. Value at Risk, the paradigm of risk management systems, recognized by the Basel II system of bank regulation and incorporated into it, has proved to be almost entirely useless – like rain-proofing that works well in a light shower, but falls apart completely in a heavy storm. The central assumption of VAR, that if you have measured and capped the moderate risks, then extreme risks will also fall into line at only a modest multiple of the moderate ones, has been proved wrong. In reality, if a particularly risk class goes wrong, it is capable of destroying an arbitrarily large amount of value.

The hapless David Viniar, Chief Financial Officer of no less an institution than Goldman Sachs, who announced last August that he was “seeing things that were 25 standard deviation events, several days in a row” had in reality announced to the market that Goldman Sachs’ risk management systems were so much waste paper, or, more likely, junk software. 25 standard deviation events should happen once every 100,000 years; if you think you are seeing them several days in a row, you are merely proving that in reality you have no idea of the characteristics of the risks you are supposedly managing.

Modern financial theory rests on two fundamental axioms: that markets are efficient, in the sense that there are no profits to be made legally by superior analysis or better information, and that price movements are in some sense random, so that risks can be assessed using standard well-understood Gaussian distribution analysis. In reality, neither assumption is true: superior analysis can indeed allow you to earn superior returns, and markets can from time to time behave in a highly non-random manner, jumping in price far more than would be suggested by analysis of past price patterns.

Whether or not modern finance rested on false assumptions, an enormous amount of money has been made by betting that they were true. If the market is thought to price all risks automatically, then even the doziest mortgage broker can originate subprime mortgages for even the least creditworthy borrowers. The fact that the borrowers are incapable of making payments on the mortgage will magically be priced into the mortgage by the securitization process, which will bundle the mortgage with other mortgages originated by a similarly lax process and sell the lot to an unsuspecting German Landesbank attracted by the high initial yield. Everybody will make fees on the deal, everybody will be happy, and the Landesbank and the homeowner will have nobody legally to blame when the homeowner is unable to make payments and the Landesbank finds a shortfall in its investment income. By making the market responsible, modern finance has removed the responsibility of all the market’s participants.

Looked at in this way, the subprime mortgage is simply a scam, and the market a giant Ponzi scheme that could survive only as long as more people entered into subprime mortgage contracts, keeping house prices high and mortgage brokers active. Once interest rates began to rise, the demise of the market became inevitable, and it also became clear that the market was not simply entering a downturn but would disappear altogether. In 10 years’ time, only Fannie Mae and Freddie Mac will be making subprime mortgages, and they will exist only because they have been bailed out by the taxpayer through the generosity of their friends in Congress, possibly several times.

Asset backed commercial paper, too, is unlikely to recover from its drubbing in the market. Investors will no longer believe that commercial paper can automatically be renewed, whatever the illiquidity of the assets and regulators will no longer believe that setting up assets in a separate vehicle funded by the commercial paper market automatically allows those assets to be removed from a bank’s leverage calculations.

Monoline insurance is an interesting hybrid case. There’s no question that insuring pools of real estate debt or other assets to bring them to a AAA rating is like the subprime mortgage business themselves and ABCP, an economic activity that should be lost to the mists of history. However, there is a rather more worthwhile business, that originally undertaken by monoline insurers, that provided credit assurance to small municipalities, who by their size were unable to tap the public markets effectively on their own. By assembling a substantial pool of funding requirements from a number of municipalities, and putting a single guarantee over the entire pool, a careful monoline insurer is not assuming any excessive credit risk, but simply allowing small borrowing needs to be aggregated efficiently into larger ones.

There is some risk that a real estate downturn such as we are witnessing could remove the funding base for a high proportion of the nation’s municipalities, by dragging down property tax valuations, but generally that risk is concentrated in the larger, higher taxing jurisdictions for which monoline insurance is not particularly economic. Thus, though the monoline insurers may disappear because of their non-municipal business, it is likely that other monoline insurers will take their place, funded by the deep pockets of the likes of Warren Buffett, and that these new insurers will continue to have a stable if not very exciting business —the quintessence of a good insurance business, in short.

Credit default swaps, on the other hand, would appear almost pure scam, with very little reason for their existence. Credit assessment is a difficult task, better undertaken by specialist entities such as banks with a deep knowledge of the borrower. That’s why rating agencies were invented, to allow bond investors to purchase credit products without the need for detailed credit assessment. However, credit default swaps allow the banks that best know a credit to sub-underwrite it not among other banks who might be supposed to have sophistication in credit assessment, but among institutional investors who generally do not have such sophistication. Further, the amount of credit default swaps written need bear no relationship to the size of the credit itself; thus the original lender may “go short” in the credit risk by writing CDS for more than the amount of the loan.

Needless to say, the opportunities for chicanery and malfeasance in such a business are legion, and made more manifold by bonus systems which reward bank officers and brokers for the business done in a particular year, without regard to the losses that business may produce in later years. Risk assessment in this business is a joke; the VAR models that fail in assessing the risk of a broad based portfolio fail even more spectacularly in assessing a narrow based portfolio of credit risks, in which correlations between different assets are not properly explored and for which the experience is at most a few years. In spite of their convenience to loan originating banks, it is thus likely that the market for credit default swaps will in future be very limited indeed.

When all these products are taken into account, it becomes obvious that the financial system of the future will look very different from that of the recent past. Shareholders will pay much more attention to the conflicts of interest between traders’, brokers’ and bank officers’ bonus schemes and their own returns. Opportunities to make large amounts of money by pure salesmanship, without regard as to the quality of the underlying assets, will disappear. Risk management will become much more conservative, and will treat exotic and little-understood assets with the utmost suspicion; that in itself will greatly limit the market for profitable “financial engineering” creativity.

The percentage of finance’s value added in the US and world economy will shrink once again, close to the levels of the 1970s and 1980s, around half those of today, and remuneration for bankers, traders and salesmen will be correspondingly more restricted. Since new career opportunities on “Wall Street” will be few and far between, there will be an aging in place of existing staff, which will itself increase those institutions’ conservatism, probably replacing it with gerontocracy.

Eventually, perhaps not before 2030, another financial revolution, immensely profitable to its participants, will begin. It is undoubtedly the case however that the new revolution will involve products and sales methodologies far different from those of recent decades.

No comments: