Robert Samuelson's column today is devoted to explaining why housing has not recovered. According to Samuelson the problem is lack of credit. This in turn is the result of the fact that lenders are feeling so beaten up that they are scared to make loans. The moral is that if we don't stop beating up on the banks then no one will be able to buy a house.

This is a nice story that unfortunately does not fit the data. At the most basic level the problem is that people are actually buying just about as many homes as we should expect. Samuelson focuses on the rate of housing starts, which is below trend, but the relevent measure for a discussion of homebuying and credit is the number of homes that people are buying.

Currently people are buying existing homes at close to a 5 million annual rate. They are buying new homes at close to a 500,000 annual rate for a total rate of home purchases of 5.5 million a year. If we go back to the mid-1990s, after the recession but before irrational exuberance began to dominate the housing market, existing home sales averaged around 3.5 million a year (1993-1995). New home sales averaged just under 700,000 a year for total sales of around 4.2 million a year.

The population is roughly 20 percent larger in 2013 than it was in 1994, which means that we should be seeing around 5.2 million home purchases a year if we are even with the pre-bubble pace. That's about 5 percent fewer sales than we are actually seeing. This means that if we compare current sales levels to the pre-bubble period we are seeing somewhat more sales than we should expect, not less.

We are seeing considerably less construction than trend levels, but this really should not be any surprise to anyone familiar with housing data. Vacancy rates remain well above normal levels. With a large backlog of vacant homes it is hardly surprising that builders would be reluctant to undertake large amounts of new building.

If anyone wanted to check the credit story that Samuelson tells, they could also look at the Mortgage Bankers Association mortgage application index (sorry, no link). If homebuyers were having trouble getting mortgages then there should be a sharp rise in this index relative to sales, as homebuyers have to put in multiple applications to secure a mortgage and some may not even get a mortgage after many applications. The index actually tracked sales fairly closely through the downturn, which suggests that the percentage of people being denied mortgages had not changed much.

In short, Samuelson has a good story about how we are hurting the housing market by holding bankers responsible for reckless and/or fraudelent mortgage issuance, but it doesn't fit the data. Nice try, though.