Right now, Fannie Mae and Freddie Mac are in government hands. Everyone agrees they should be “wound down” and in some sense replaced with a new system. And everyone agrees on two things about this system. One is that it should somehow involve less big government, and the other is that it should somehow deliver the same amount of subsidy to home-buyers. This is a nutty consensus. The subsidy strikes me as undesirable, but insofar as the subsidy is going to be delivered the way to deliver it is through a direct big government provision of subsidies.
The first question to ask yourself is this: “Should there be a major government program whose purpose is to subsidize leveraged home-purchasing by middle class people, thus making the homeownership rate marginally higher and the average size of owner occupied houses marginally larger than it would otherwise be?” The correct answer to this question is obviously “no.” At the same time, it’s clear that all Democratic Party politicians and all Republican Party politicians agree that the answer is “yes.” So then the question becomes, what is an intelligent way to design such a program?
Start by asking: “what is the particular form of mortgage product you want to promote?” The answer, clearly, is the 30-year fixed rate self-amortizing loan.
Follow up by asking: “how cheap should these loans be?” Since the idea is to construct a subsidy the government can afford to offer, let’s say 1 percentage point above the government’s cost of funds for issuing a 30-year bond.
Another question: “What’s a reasonable amount of leverage for a middle class homeowner to have?” The conventional answer is 20 percent downpayment.
Another question: “How big and expensive does a house have to get until it stops being a middle class house?” According to the Census Bureau, the median owner-occupied house in the United States costs $186,200. Let’s say double that and we get $372,400.
So now we get to the shape of a reasonably designed program. Here’s how it works. If you want to buy a house with borrowed money, then the United States government will lend you up to $372,400 to do so at an interest rate that is 1 percentage point higher than the 30-year U.S. Treasury bond rate (today that would be 3.8% + 1% = 4.8%) as long as you make a downpayment of at least 20 percent. That doesn’t mean you couldn’t buy a house that cost more than $465,500 or that you couldn’t buy a house with less than a 20% downpayment, but it means that if you want a bigger or more leveraged loan you’ll need to get a bank to lend you the money. But as long as you stick within the lines, you’ll get a nice discounted loan from Uncle Sam. If you end up defaulting on your mortgage, then Uncle Sam takes the house, and you can’t get a new Uncle Sam loan unless you pay back your arrears.
In ordinary times this program would be “profitable” for the government (a somewhat meaningless concept, but it is what it is) and periodic episodes in which it generated large fiscal losses would coincide almost perfectly with severe economic downturns in which large budget deficits are prescribed.
Now is this program I’ve outlined a good idea? Not especially. There’s no good reason for the government to push marginally more people into borrowing money to buy houses, and many people into buying marginally larger houses. But insofar as there’s a political consensus around the idea that I’m wrong about this, that would be the way to do it. Leave it all hanging out there, as a slightly bizarre big government program. So why don’t we do it that way? In part the political influence of private sector actors. But more broadly, I think a key part of the politics here is that everyone agrees the subsidies should exist but nobody wants to admit they exist. Hence the bipartisan love affair over finding incredibly complicated and confusing ways of delivering the subsidy.