The Ledger

From Professor John McInnis
Does Fair Value Accounting Contribute to “Contagion” in the Banking Industry?

Yes, according to a new study by Urooj Khan, a PhD student at Washington (and soon-to-be Assistant Professor at Columbia I’m told).  Let’s set the stage for why you should care.

Fair value is the price an asset would fetch in an orderly sale in a liquid market, or how much it would take to offload a liability on another party. These fair values can either be observed in active markets or estimated in some fashion. Current GAAP in the U.S. requires banks to record many of their financial assets (stocks, debt securities, derivatives) at fair value, with changes in the fair value of these assets flowing through owners’ equity.

Critics, mostly in the banking industry, contend this practice exacerbated – or perhaps even partially caused – the credit crises of 2007-2009. The argument goes something like this.  Government regulations require banks to maintain a minimum level of “capital,” which is roughly owners’ equity as a percentage of risky assets. When asset values are marked down to fair value, regulatory capital gets too low, forcing banks to convert some of their “risky” financial assets into “riskless” cash to boost the ratio described above. This leads to a “fire sale” of financial assets, which further reduces market prices for these assets, leading to further mark-to-market write downs by other banks, and the process starts anew. This is an example of the “pro-cyclical’ effect of fair value accounting you may have heard about in the popular press. 

Until Khan’s work, such a story was just theory, the subject of two recently-published academic articles laying out how this could happen in a stylized economic model. Khan’s contribution is to bring much needed empirical evidence to the table. Consistent with prior work, he finds that all banks tend to do poorly (in terms of stock price) when conglomerate, “money center” banks do poorly.  This is not shocking since banks hold correlated positions in an inter-connected economy. More importantly, he finds this relation has gotten stronger as the percentage of assets in the banking industry recorded at fair value has increased. In addition, individual banks with a greater proportion of their assets marked to fair value face a greater exposure to this “contagion” effect.

Of course, Khan’s evidence is suggestive, not causal and definitive (see also the disclaimer below).  Nor does the evidence necessarily indicate fair value accounting should be scrapped. Less drastic measures may be appropriate.  The FASB has altered GAAP recently to reduce the influence of “fire sales” on the measurement of fair value. Amendments to capital requirement regulations have also been considered as a way to reduce the pro-cyclical effects of fair value write downs. 

You can access the research paper here

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1327596

DISCLAIMER: I offer links to unpublished working papers that I find interesting and well-executed. I do not necessarily agree with the conclusions of the linked papers, nor do I offer any assurances the papers have or will be vetted by the appropriate peer-review process. All scientific research is subject to criticism, but I offer none here primarily because authors have no chance to respond. The lack of criticism should not be interpreted as an endorsement of any kind, and I leave it to the reader to form their own conclusions and critiques.