The largest piece of “unfinished business” from the financial crisis, housing finance reform, slipped back into “not yet” status this month. The Senate Committee on Banking, Housing, and Urban Affairs passed, on a bipartisan 13-9 vote, S. 1217, a bill that would wind down Fannie Mae and Freddie Mac, and replace them with a system featuring an explicit government guaranty. The narrow majority and the absence of support from key Democrats on the committee means that the bill has little chance of being scheduled for a vote by the full Democrat-led Senate. With that, the prospects for housing finance reform moving ahead in the dwindling days of this session of Congress are close to zero.
This pause in the housing finance reform process provides us the opportunity to ponder a fundamental question about the role of the government in our housing finance system. When statistics like “the government is backing 90% of mortgages originated” are used, they are describing the breadth of the market that has a federal guaranty. There is also, however, the depth of that guaranty. The taxpayers are on the hook for 100% of loans guaranteed by FHA and, for loans without private mortgage insurance, by Fannie Mae and Freddie Mac. On loans with private mortgage insurance, the taxpayers are in a remote loss position, with the private insurers covering losses up to 25-30% of the loan amount before the GSE exposure begins. Loans insured by VA turn that around, with the taxpayers on the hook for the first losses up to around 25% of the loan amount, after which the Ginnie Mae Issuer/Servicer must absorb the remaining losses. In a Ginnie Mae structure, the taxpayers then have a second exposure, to the failure of the Issuer/Servicer, which is the risk covered by Ginnie Mae.
So, the question to be pondered is, should taxpayers be more concerned about the breadth or the depth? Keep reading, if you like, but I think the answer is that taxpayers (and their elected representatives) should be more concerned about the depth. In fact, there is substantial theoretical and historical evidence that, if there is a guarantee for some of the market, there should be a guarantee for all of it, but only at a very remote level.
Depth of exposure to mortgage loss is complicated by the nature of secured lending. The risk to a mortgage lender (and to the guarantor) is two-fold: that the borrower will stop making payments; and that the lender won’t be able to recover the amount due by selling the house. The uncertainty over how much can be recovered from the property creates the possibility of splitting loss exposure into layers. The borrower is always in the first layer, losing his equity in the home if he defaults and there is not sufficient value to pay off the loan balance. Mortgage insurance typically provides the next layer, absorbing further losses as the home value declines. The likelihood of a home being worth zero is extremely small, so the last dollars in the layers are in a very safe position.
Thus, mortgage insurance that covers 25% of the loan amount will generally cover most of the losses under normal housing market conditions. The uncertainty of home prices makes it hard to say how much of the risk is covered by guarantees layered in this way. It is much easier to answer that question if the losses are shared proportionally, in what insurers call a “quota share” arrangement. In that structure, each participant covers as predetermined percentage of the losses, regardless of how much is lost. Someone with a 25% quota share would always bear 25% of the risk, regardless of what happens to home prices.
Simpler is usually better, but in this case, complicated is worth the effort. The goal of providing an explicit government guarantee on mortgages should be to make it unlikely that the guarantee is needed. An obvious way to accomplish this is to have private capital in the first loss position on every loan, with the depth of exposure sufficient to cover the losses in all but the most severe housing downturn.
The success of Ginnie Mae, which has never had a losing year, illustrates the benefits of this approach. As described earlier, VA insurance typically covers the first 25% of losses on loans that typically have no down payment. Why hasn’t Ginnie Mae had to cover huge losses on VA insured loans? In between the VA insurance and Ginnie Mae is the obligation of the servicer to cover the losses. The combination of VA insurance, the ongoing servicing fees, and the servicer’s capital protects Ginnie Mae (and the taxpayers) from losses. The taxpayer risk position on Ginnie Mae is sufficiently remote that it was not called upon, even with a worse home price decline than in the Great Depression.
A structure similar to Ginnie Mae, but featuring private mortgage insurance instead of government, would reduce taxpayer exposure to an extremely remote level. By making that risk so remote, the question of breadth of exposure can be approached differently. Politically, reducing the breadth of exposure is extremely difficult, as evidenced by the lack of consensus in Congress to reduce loan limits back to pre-crisis levels. From an economic perspective, the distortions created by guaranteeing only a portion of the market are a good reason to explore alternatives.
What are your thoughts on the role of the government in our housing finance system? Please share in the comments.Tags: GSE, GSEs, Housing Finance Reform, Housing Market, Mortgage Industry, Top Content