America has not had a nationwide housing crash since the 1930s. At one point during that calamity, an estimated 60 percent of all mortgages were in technical default. The rather primitive housing credit system of the time, which relied on five-year balloon mortgages, certainly exacerbated the problem, but the bulk of the problem was related to the general economic downturn. There have been some regional housing crashes that were short and relatively mild, most notably in California, Texas, and New England in the late 1980s and early 1990s. Most of those were caused by declines in key local industries: oil in Texas, aerospace and defense in Southern California and Massachusetts.

The current downturn, by contrast, is due almost exclusively to a change in the housing credit cycle from excessively easy to modestly restrictive. Housing turned down before the economy, and even now, nearly 18 months into the housing recession, the national unemployment rate is still at what economists consider full employment. That is unlikely to last as credit problems spread into the consumer sector, layoffs spread, and the resulting rise in unemployment makes the consumer credit situation still worse.

It is the uniqueness of the current housing crash that adds to its intractability. Policymakers haven't been here before, so they're not certain of the way out. Many of the institutions that underpin the industry are relatively new--actually created since the last downturn in the early 1990s--and untested. We also know that many of those institutions were far from transparent, and some were fraudulent. As a result, everyone needs to be suitably modest about predicting how the housing crash will end and remain flexible about the policy actions that may be needed to augment the normal functioning of the market. Some basic facts about supply and demand offer a good, if sobering, place to start.

There are 129 million housing units in the United States, comprising owner-occupied, rented, and vacant units. Of these, 18.5 million are empty. This vacancy rate is 2.5 percentage points higher than it has been at any point in the half century the data have been tracked, translating into at least 3 million too many empty housing units in the country. This number, moreover, is rising. This is the most intractable part of the real estate bubble, for we cannot find a true bottom to home prices until this inventory of empty units starts to clear, and we cannot find a bottom to the mortgage finance market until home prices bottom out.

The worst type of inventory is an empty house, which people in the industry like to say has about the same half-life as a head of cabbage. As the former chairman of the Neighborhood Investment Corporation, I've seen the damage done to neighborhoods by vacant homes. They are never maintained adequately, depress surrounding property values, and can quickly become temporary retail space for drug lords and a playground for juvenile delinquents. They are also the homes whose owner has the least incentive, and usually the least ability, to service the mortgage or pay the property taxes. So whittling down the inventory of empty houses should be the first economic, social, financial, and political objective.