Start with what seems like a good idea: assets and liabilities on a company’s balance sheet should be adjusted to reflect their fair value.  For instance for assets, fair value would not be the price the company paid when it bought the asset (or the cost it incurred in building the asset), but the value it could realize if it sold the asset.  A similar concept applies to liabilities.  For a long time the rules governing what constitutes fair value were different on a case by case basis resulting in complexity and leeway.  Along came FAS 157, with the goal of unifying how fair value should be determined across all the cases where that concept applies.

Even VC firms are affected by this rule.  We need to determine the fair value of the stakes we have in our portfolio companies.  That of course immediately illustrates a big problem with the whole idea.  There is no transparent market for early stage companies and companies are highly idiosyncratic and hard to compare.  So how does one determine fair value?  One way is to try to apply public and M&A comps where they might be available.  Usually, however, the size difference between the public companies and even the M&A transactions that have reported prices relative to early stage companies is one or two orders of magnitude.  So applying the same metrics is highly suspect.  Another way is to create a probability distribution of possible outcomes and analyze the value of the securities in each outcome. That approach of course is highly subjective.

More importantly, both approaches suffer from a pernicious and sytemic flaw that has come to haunt the entire financial system.  When things are going well, the comps or possible outcomes tend to go up, pushing up the fair value of assets.  Liabilities on the other hand remain more or less unchanged (since they mostly tend to be fixed amounts, e.g. bonds or loans).  This makes everyone feel good.  Too good in fact because these are really just accounting fictions.  The fair value history is ultimately of no relevance for the actual outcome, especially for VC firms which tend to be invested in the same company for many years. At least VC firms don’t use leverage but imagine everyone during an upswing applying fair value and then taking on more debt because their assets look so good.  It is easy to see how the fair value concept will dramatically amplify a bubble on the way up.

In the same way, fair value will amplify a contraction.  As soon as there is some sense that assets might be impaired they need to be written down – no matter what the ultimate recovery several years down the road might look like!  Liabilities on the other hand remain fairly constant.  That of course means companies are now overlevered which *forces* them to sell assets in a distressed state to try to pay down some of the debt.

So here we have a classic version of the Law of Unintended Consequences. A rule that was intended to make financials more transparent wound up significantly increasing volatility.  In fact, I believe that the only way to get ourselves out of the current mess it to suspend the rule, get all of the unregulated derivatives contracts into a database, and try to make all of it go away in one big settlement.