Saturday, March 17, 2007

Mortgages and Houses

A few years ago, Alan Greenspan suggested that many households would have been better off financing their houses with adjustable rate mortgages. This is because for many years payments on adjustable rate mortgages were lower than on fixed rate mortgages--so after the fact, Greenspan was correct.

The problem with advising people to use adjustable rate mortgages, however, is that ARMs give households liabilities that have short duration--that is, liabilities whose market value remains close to face value at all times. This is because the rates on ARMs by definition change to meet market rates on a regular basis. Houses, on the other hand, are assets with lots of duration. The services they give to homeowners (shelter and a set of amenities) is pretty much invariant to market conditions. Consequently, house values change with market conditions, such as changing interest rates.

Good financial management practice suggests that to minimize risk, the duration of of assets and liabilities for any institution, including households, should be matched. In the case of houses, this means that households looking to minimize risk should use a fixed rate mortgage to finance their house. There are exceptions--if one buys a house and expects to sell it in five years, a five year ARM makes lots of sense, because the duration of the asset (housing services over five years) and the liability would match.

This is not to say there is anything wrong per se with people getting ARMS, so long as they explicitly understand the risk embedded in them. But a principle I have been pushing for years is that if people can't afford a house with a fixed-rate mortgage, they probably shouldn't buy a house. It is one thing to have the option of the FRM, and then decide to take the risk of the ARM anyway. One of the nice things about the United States is that FRMs are easy to come by--this is not true in most countries around the world. It is something else to be forced into taking a risk in order to buy. Under these circumstances, buying probably isn't worth it.

That said, there is probably too much ink being spilled on the downside of home-owning. The Times this morning had a piece that would make one wonder why anyone should own a house. But it is important to remember that renting is risky too--leases are usually only a year long, which means renters are subject to increases in rent or may even be forced to move every year.

Notice that I have not even mentioned the subprime market in this post. That will be for another day (probably tomorrow).

2 comments:

Anonymous said...

Richard,

Tim Harford recently wrote a column on Oswald's hypothesis in Slate, which has been picked up by the major blogs (Marginal Revolution, Greg Mankiw, Economist's View) but without mention of any recent research. Is there still debate among the housing researchers over the validity of the hypothesis?

Richard Green said...

So far as I know, Oswald's work was never published. He doesn't simple correlations and controls for very little.

A paper I published with Hendershott in Urban Studies suggests there isn't much to the Oswald Hypothesis. We have a better, unpublished paper that I think refutes Oswald, but by the time we wrote it, the controversy had died down. Maybe we need to dust it off.