What is the value of a United States government guarantee?
It’s not often you get to answer a question with such import or significance. But for economists Robert McDonald and Deborah Lucas, it’s an academic matter. One of their latest collaborations has produced an estimate of the until-recently-implicit government guarantee for Fannie Mae and Freddie Mac, which amounts to between 0.2 to 0.3 percent annually. Their results—compiled before the two entities entered conservatorship—largely reflect what the government has sunk into the two companies.
Prior to conservatorship, Fannie and Freddie were essentially private companies with the backing of the federal government. Though the government was not technically responsible—“The actual legal guarantee was trivial; it’s in the single billions,” McDonald says—the two enjoyed what he calls a “sort of limbo status where everybody believed there was a government guarantee for Fannie and Freddie debt.”
As a result, “Fannie and Freddie are enormous obligations for the U.S. government.”
“The actual legal guarantee was trivial; it’s in the single billions.” — Robert McDonald
Investors happily purchased Fannie and Freddie debt, assuming the government would intervene if the sky fell. And that is just what happened in September 2008 when the two mortgage companies became wards of the government. But the two companies’ ties to the government are decades old. Fannie Mae—the oldest of the two—is a product of the New Deal that was intended to prop up the sagging housing market at the time. As a government agency, it bought up mortgages from banks, freeing them to make more loans. In the late 1960s, President Lyndon Johnson privatized Fannie to lighten the government’s balance sheet during the Vietnam War. Freddie Mac was created a few years later to provide some competition.
More Ways than One
A handful of economists over the years have tried to calculate the value of Fannie and Freddie’s implicit government guarantee, and their estimates have varied widely, from as low as $200 million to as high as $182 billion. One method, called the spread-based approach, compares the borrowing cost for the government-sponsored enterprises (GSEs) with a comparable private company. Spread-based proponents say this difference values the implicit government guarantee.
But McDonald and Lucas say the method does not account for Fannie and Freddie’s unique bankruptcy threshold—investors will be hesitant to declare bankruptcy if doing so will cost the GSEs their special status. Thus, Fannie and Freddie will likely stay solvent in circumstances that would cause a private company’s shareholders to walk away.
McDonald and Lucas argue that the implicit guarantee can be more accurately valued by estimating how much it would cost to insure Fannie and Freddie against bankruptcy. This has essentially been the government’s position—it insures Fannie and Freddie, albeit with a one-time payment rather than more traditional regular installments. Options-based estimates are nothing new, but previous attempts were relatively rigid. They did not allow the modeled firms to respond to events nor did they forecast far into the future. McDonald and Lucas’s model also allows for a reasonably chaotic world, and they say that this flexibility corrects for some of the low-balling of more simplistic options-based approaches.
McDonald and Lucas valued the insurance as a put option on each GSE’s assets, where if the assets drop in value, the government is responsible for a portion of the difference. But Fannie and Freddie’s asset volatilities are obscured by their insured debt, so the researchers used equity volatility as a proxy. McDonald and Lucas began by simulating a fictitious GSE and examined how the company’s value changed over time as its assets fluctuated in value. As long as the company avoided bankruptcy, it could sell more guaranteed debt at a low rate.
A comparable uninsured company could undertake similar levels of debt, but at higher rates. When the simulation had finished, it was clear that the guarantee added substantial value to the company. The debt and equity values of the insured company exceeded its assets—the difference was the guarantee. An uninsured company would have had no such discrepancy.
The researchers then calibrated the model based on Fannie and Freddie’s actual finances for 1995 to 2005. They took the fine-tuned model and ran it 50,000 times with an equal number of permutations to determine how different management and regulatory policies can affect the value of the guarantee. Their projections forecast to 2015 and 2025.
A Range of Estimates
In stable times, McDonald and Lucas estimate the government guarantee to be an annual premium of about 20 to 30 basis points over a Treasury bill, or a lump sum of $35 billion for Fannie and $30 billion for Freddie that would cover twenty years. That is not too bad considering bonds for the companies combined are worth about $1.5 trillion.
But when things get rocky, premiums—both annual and lump sums—start to climb. A small annual chance of a 5 percent reduction in assets adds between $2 billion and $3 billion to the cost of the guarantee. But if Fannie’s assets, for example, were certain to lose five percent of its value, the value of the government guarantee for Fannie would jump to over $50 billion, or about 45 basis points. The next 5 percent drop would be even harsher, pushing the cost to $72.5 billion or about 81 basis points.
When and if Fannie and Freddie are released back on to the market, McDonald thinks the government needs to more explicitly acknowledge the government guarantee. “We should be thinking about it as a real cost to the government,” he says. That could be tricky, given that the guarantee is all but invisible in the best of times. But, as McDonald points out, “a Fannie and Freddie meltdown happens at the worst time.”