Sapra on Mark-to-market accounting

Here's a paragraph from the Sapra piece:

"Allen and Carletti identify an interesting and novel endogenous cost of marking-to-market: whenever a crack appears in some part of a financial system, a mark-to-market measurement regime may serve as an important catalyst in propagating this crack to other parts of the system. Such propagation, in turn, may lead to financial contagion which is detrimental to welfare. Put differently, marking-to-market has financial stability implications because it may amplify the nefarious effects of liquidity pricing.

The entire paper, along with the Allen and Carletti paper that this paper is building off of, is worth reading. It's basically a model of how this particular kind of accounting rule ("mark to market") can exacerbate small vibrations in an economy and turn them into massive ones via the feedback effect between demand for assets and their present market value. What's new to it is its focus on non-liquid assets, like houses, where I think the point being made is that they are more susceptible because they trade so irregularly (maybe once every 5 years, unlike bonds which exchange hands constantly everyday) that marking them to current market values is not going to reveal their underlying value as well as marking to a model's estimates, which are based on longterm valuations.

Of course, if that's true, then it would make sense that no one on Wall Street said a word during the run-up. Mark-to-market would've been a boom, since the feedback was creating a feeding frenzy on homes. But the same thing driving it up drives it down, and I think the theory is that mark-to-market is creating a kind of massive volatility and variance in the asset prices themselves, and thus the same variability among those institutions leveraged with them.