In spite of great expectations, the Financial Stability Plan that Treasury secretary Tim Geithner presented last Tuesday was short on details. Its biggest innovation was a way for the government to persuade private investors to buy toxic assets from banks. The investors would have started buying those assets at the banks prices already if they thought it profitable; instead, they held back. How much will it cost for the government to change their minds? Potentially, an amount so enormous as to risk the credit of the U.S. government itself.
Judging by Geithners past behavior as chairman of the Federal Reserve Bank of New Yorkwhere he helped lead bailout deals for Bear Stearns, Citigroup, and others last yearits likely that the Treasury will try to attract investors by using government guarantees to cap their possible losses. On the face of it, this seems like a smart way for the government to stop the financial industrys meltdown without incurring astronomical costs. By covering some potential losses, the thinking goes, the government can calm investors fears and lure them back into buying mortgage-related and similar securities from banks. If the government guarantee is large enough, investors can even pay for the securities at close to the value on the banks bookssparing banks the burden of recognizing additional losses, which could push many of them into official insolvency. Last but not least, the plan minimizes the amount of money that the government must request from Congress.
To understand the problems lurking beneath this idea, though, lets analyze a similar deal: the guarantee that the federal government provided to Citigroup in November 2008. For a $306 billion pool of Citigroup assets consisting of loans and securities backed by residential real estate and commercial real estate, the government committed to providing something like an insurance policy with a deductible. Citigroup would absorb the first $39.5 billion of losses on the loans and other securities, with the government picking up 90 percent of the additional losses and Citigroup just 10 percent. The government ingeniously divided its responsibility among the Treasury (the first $5 billion), the Federal Deposit Insurance Corporation (the next $10 billion), and the Federal Reserve (all the rest). In this way, the Treasury committed only $5 billion of the finite TARP bailout money to the deal.
The real value of the guarantee (and thus its potential cost to taxpayers) should include not only the TARP funds but also these other commitmentsand it is massive. We estimate that if the government were held to the same accounting standards as private companies, the expected liability it would have to report for the Citigroup guarantee would be $66 billion. Nobody knows the true value of the volatile loans and securities underlying the dealmeaning that nobody knows the true extent of potential losses. And even the $66 billion estimate assumes that the assets comprise an average pool of residential-backed securities. Since Citigroup has a strong incentive to put the worst, most overpriced assets in the pool, the governments actual liability could easily be $78 billion or more.
Geithners plan suggests that the government might be applying similar sleight of hand to the entire financial system. We estimate that to restore the solvency of the top 10 banks to their pre-crisis level, the banking system needs at least $4.5 trillion in purchases of its toxic securities. (Thats why former Treasury secretary Henry Paulson abandoned the Bush administrations idea simply to buy up all the toxic assetshe knew that the government couldnt afford it.) Based on the Citigroup example, we calculate that Geithner would have to commit $75 billion of TARP money to attract enough private capital for the plan. But just as with the Citigroup case, that initial commitment wouldnt tell the whole story. Under proper accounting standards, and taking the entire governments commitments into consideration, the strategy would actually impose a cost of around $1.2 trillion on taxpayers.
Geithner is contemplating a Public-Private Investment Fund of $1 trillion and a similar Consumer and Business Lending Initiative of another $1 trillionabout $2 trillion in all, rather than the full $4.5 trillion. But the initial cost of the plan for the government would still come to a staggering $510 billion. And whats even scarier is the potential losses for which the government could be liable. With $2 trillion in principal at stake, we calculate that the government would face a 10 percent chance of losses greater than $1.3 trillion and a 5 percent chance of losses larger than $1.4 trillion. Such losses, unlikely though they are, could be large enough to undermine faith in the governments solvency, leading to a currency crisis.
Some might dismiss these worst-case scenarios as purely hypothetical. After all, the crisis could still recede; the U.S. government could get lucky and make a profit on the overall transaction. But such reasoning is myopic. Underwriting insurance against rare events always seems a profitable businessuntil the rare event occurs. Until June 2007, insuring Lehman Brothers against default looked like a great deal. Back then, the market attributed only a 0.6 percent probability to a Lehman default (over two times lower than the probability the market attributes today to a default of the U.S. government). Fifteen months later, after Lehman had collapsed, the hedge funds and other investors that had insured the firm had to pay 92 cents on the dollar to every insured debt.
For that matter, ignoring the real risk of insuring rare events is also what brought down AIG, the largest insurance company in the world, which had insured many mortgage bonds and other securities similar to the ones that the financial industry is stuck with now. Do we want to bring down the worlds most powerful government in the same way?