Clearing the Air on "Too Big To Fail"
|Written by David Rosnick|
|Monday, 17 October 2011 16:23|
With Occupy Wall Street continuing to build steam, Cato's Mark Calabria chose to engage in a little friendly fire. I like to believe that libertarians and progressives could come together to rewrite the rules of a rigged system, but Calabria seems interested in thickening the fog of war rather than clearing the air.
Cato's Mark Calabria leveled a strange charge at Joseph Stiglitz, suggesting that in 2002 Stiglitz and his coauthors (Jonathan and Peter Orszag) "sold their work to the highest bidder defending the system" of socialized losses and privatized gains. The paper analyzed the taxpayer risk of guaranteeing the debt of Government-Sponsored Enterprises (GSEs—primarily Fannie Mae and Freddie Mac) and found that under one of the two capital standards, the expected costs to taxpayers would be very low.
Of course, in 2011 this appears laughably naïve, given the many billions of dollars of support to Fannie and Freddie in the wake of the housing bust. However, Calabria's charge is utterly bizarre for several reasons. First, Stiglitz, Orszag, and Orszag specifically address the importance of "too big to fail" in taxpayer risk. Second, Calabria ignores the fact that risks grew considerably since the publication of the paper. Third, Calabria abuses some math in order to make it appear that the authors downplayed the potential costs.
Far from avoiding the question of "socialized losses and privatized gains," Stiglitz, Orszag, and Orszag rightly point out that the risk to taxpayers is far from limited to GSEs. They write,
"In the absence of Fannie Mae and Freddie Mac, mortgage risk would likely be held by large banks and other types of financial institutions, which themselves benefit from the perception that they are 'too big to fail.' Fannie Mae and Freddie Mac are among the largest financial institutions in the country. Even in the absence of a GSE charter it is likely that they would continue to benefit from their size, since the government has intervened on behalf of other large institutions in the past."
The authors observed that the risk to taxpayers was system-wide, "regardless of the existence of the GSEs" and therefore the actual cost of any GSE-specific guarantee is low. This is a lamentation that the losses would be socialized regardless-- not a denial of that fact.
Furthermore, the housing bubble grew far worse from 2002 on. While the bubble was obvious at that time, the risk of complete economic meltdown was still very low. Without question, Fannie and Freddie were bad actors during the run-up in the housing market—and the existence of this bubble was clear as far back as 2002. However, the GSE share of outstanding mortgage debt peaked in 2002 (or 2003 if mortgage-backed securities are included.) Considering that the GSEs started with a large share of higher-quality mortgage debt and lost market share as the flow of new debt deteriorated, it is hard to argue that Fannie and Freddie led rather than followed.
Nor is it appropriate to attribute the GSE losses to their own risk-taking. The GSEs were called upon to unfreeze the market as the private actors tried to shed their losses. They almost certainly took on large amounts of the worst debt even after the government took them over. These losses were deliberately assumed as part of bailing out the private sector. Though one might argue that absent the GSEs, there could be no taxpayer losses as a result of the GSEs, this assumes the GSEs were the only vehicle available to bail out the market. In fact the Federal Reserve would most likely have taken a much larger role in bailing out private companies and the TARP program would have shown much larger losses. All of this is entirely consistent with the warning of Stiglitz, Orszag, and Orszag regarding the systemic nature of "too big to fail."
Finally, Calabria compares the $2 million estimate of Stiglitz, Orszag, and Orszag to the "actual cost so far of $160 billion." This comparison, however, is meaningless. In constructing the $2 million figure, the authors simply estimated the probability of an economic "nuclear winter" of 1 in 500,000. Thus, they reason, $1 trillion in GSE debt bore (in 2002) an expected cost of $1 trillion/500,000=$2 million. Even if one attributes all $160 billion to GSE failure, one cannot compare this to the 2002 estimate.
Suppose I made a bet with Mark Calabria in which I will roll a fair die, and if it comes up '6' then I will pay him $6. Otherwise, no money will change hands. Obviously, this is a losing proposition for me, on average. Five times out of six I will pay nothing, but one time out of six I will pay $6. I expect, therefore to lose $1 on the bet. Suppose I then roll the die, and in fact it comes up '6'. Now I owe Mark $6—six times what I expected to pay before I rolled the die.
Did I downplay the risk of the bet? Am I "shocked" at the eventual cost? Of course I am not, and neither are Stiglitz, et al. Whatever one thought of the odds, what looked to be a very rare and catastrophic event in 2002 came to pass and we expect to pay what we owe. For their part, Stiglitz, Orszag, and Orszag based their estimate on the assumption it would cost trillions if disaster struck, and we've paid "only" $160 billion.
Calabria may dispute the authors' assumptions and simulations, but cannot fairly charge them with "selling their work" based on a mathematically faulty comparison. Nor can Calabria fairly fold in the cost of a shadow bailout without considering the authors' express warnings of the greater "too big to fail" taxpayer risk.
In the end, even if Calabria's critique of Stiglitz' past work had been valid, it is truly unfortunate that he chose to deflect ire away from the "too big to fail" system we all abhor and toward potential allies who have in fact critiqued that system. I have respect for Calabria's aversion to privatizing gains while socializing losses, yet we must develop tools to ensure that this comes to an end. And yes, this means an end to Fannie and Freddie's grotesque public/private hybrid charters.