Is the credit crunch a lemons problem?

Joint post by Willem H. Buiter and Anne C. Sibert. Virtually the same post appeared on September 2 on Martin Wolf’s Economists’ Forum.

In his column ‘Central Banks should not rescue fools’, Martin Wolf, Chief economics commentator of the Financial Times, characterises the credit crunch/liquidity crisis of August 2007 as a ‘lemons’ problem due to asymmetric information. We argue that this is not the case.

Consider the securities for which the liquidity crunch was most serious: collateralised debt obligations backed by US subprime mortgages (CDOs) and commercial paper backed by (not necessarily subprime) mortgages and hire purchase receivables such as credit card debt (ABCP). If the market failure was due to a “lemons” problem, then potential buyers of the assets that became illiquid must have suddenly realised that there was an asymmetric information problem in the markets for these assets and that they were on the wrong side of it. That is, potential lenders realised not only that they knew a lot less about the creditworthiness of the securities being offered than they had thought they knew, but that this increase in uncertainty was not shared to the same degree by the sellers of these securities. This discovery caused the market to shrink and to cease to function. Was this what happened?

It is true that the asset-backed securities in the markets that became illiquid are not homogeneous; the quality of the underlying assets (for example, mortgages) is not uniform. If the would-be sellers of the securities have private information about the quality of the underlying assets and are unable to reveal this private information to the buyers, then a lemons problem can arise in the following way. Potential buyers of the securities are only willing to purchase the securities if the price is no higher than what they would pay for securities backed by underlying assets of average quality. If the dispersion in the quality of the underlying assets is large enough, the sellers of assets backed by the best-quality underlying assets will be unwilling to sell at that price and withdraw from the market. Potential buyers know this and realise that the average quality of the underlying assets is lower than it would have been if the sellers with the best-quality underlying assets had not departed the markets. Thus, the price at which they are willing to buy securities falls and additional sellers of good-quality assets may leave the market, causing the purchase price to fall further. A continuation of this process can cause the market to shrink further and perhaps even to collapse entirely.

Is this what happened? Consider the case of mortgage-backed securities. There is certainly asymmetric information in the primary market for individual mortgages. Typically, the applicant for a mortgage knows more about his creditworthiness than the loan officer of a bank. In some subprime mortgage markets, the asymmetric information problem may have been on the other side. Smooth-talking salesmen may have convinced uninformed or gullible mortgage applicants to take on loans that the mortgage sellers knew to be in excess of the homeowners’ capacity to service.

Regardless of who has the informational advantage in the market for the original individual mortgages, when these mortgages get securitised there is an asymmetric information problem between the banks originating the mortgages and the special purpose vehicle (SPV) that buys the mortgages from the banks and then issues the securities backed by these mortgages. The SPV is unlikely to know as much about the quality of the underlying mortgages as the originator. When the SPV sells its asset-backed securities, the purchasers have little idea about the quality of the underlying assets that were pooled.

When the current holders of asset-backed securities try to sell them in the market, neither the sellers nor the buyers have superior information. The only parties with private information – that is, the original mortgage borrowers lenders -- are not in the market or not in a position to make use of their superior, asymmetric information.

The recent illiquidity in financial markers is certainly an information problem, but it is not an asymmetric information problem. I think Martin and we agree that in July-August (perhaps even a bit earlier), there was a general realisation that the credit ratings granted by the main rating agencies to many asset-backed securities and structured financial products in general, were wildly generous. There therefore was an associated increase in the market’s perceived average probability of default for wide classes of securities. But the greater awareness of ignorance and the increased uncertainty are market-wide and symmetric, affecting would-be buyers and sellers equally. The private information that could have caused a lemons problem was destroyed by the process of securitisation and pooling.

It is not clear why the recent uncertainty caused such upheaval. Perhaps – for some reason -- uncertainty has become more Knightian in that it has become less possible to attach a subjective probability distribution to the range of possible outcomes or even to describe the set of possible outcomes. Pervasive, Knightian uncertainty could be associated with the fear, panic and herding behaviour that caused market failure.

Illiquid markets prevent financial firms from raising cash by selling the securities that become illiquid. These firms will have to fund themselves in other ways and this may lead to a lemons problem. If firms try to raise money through unsecured loans or debt or through loans and debt secured against assets about which do indeed have private information, then asymmetric information may play a role in causing these sources of credit to fail. It will be interesting to uncover, when the history of this crisis is written, to what extent asymmetric information caused certain sources of credit to vanish. It played no role, however, in the drying up of the asset-backed security markets.

©Willem H. Buiter and Anne C. Sibert 2007