by Calculated Risk on 1/02/2012 12:15:00 PM

This morning Dean Baker wrote about the Housing Vacancies and Homeownership Survey: Robert Samuelson Oversells the Case for Economic Optimism (ht Joe)

[W]e are still far from making up for the overbuilding of the bubble years as indicated by the fact that the vacancy rate remains at near record levels.

(There have been some questions raised about the accuracy of the Census Department's data, claiming that it overstates the number of housing units in the country. Those raising the issue fail to note that measures of housing starts do not include housing units that were created by conversion of commercial or industrial property, such as an old warehouse being turned into condos. The rehabilitation of dilapidated units would also not be included in housing start numbers. There were many cases of both ways of adding to the housing stock during the bubble years. Also, it is important to note that the Census data is giving the percentage of units that are vacant. The critics of this measure must show how the Census methodology would lead it to overstate the share of units that are vacant.)

First, the main criticism of the HVS is it doesn't match the decennial Census results. The Census Bureau has acknowledged this and promised to investigate the differences. Here are some recent comments from the Census Bureau:

The most recent research has shown that the CPS/HVS and the 2010 census produced significant differences for vacancy characteristics. The rental vacancy rate from the April 2010 census was 9.2 percent, whereas the CPS/HVS reported the rental vacancy rate of 10.6 percent for the first half of 2010. The April 2010 census had a homeowner vacancy rate of 2.4 percent, while the CPS/HVS had a vacancy rate of approximately 2.6 percent for the first half of 2010. For occupied housing, the April 2010 census produced a homeownership rate of 65.1 percent, while for the first half of 2010, the CPS/HVS produced a rate of 67.0 percent.

It is important to note that the HVS is benchmarked to the decennial Census, so the most recent vintage for housing inventory was benchmarked to the 2010 Census. So clearly the Census Bureau thinks that is a better estimate of the total housing inventory.

Although the HVS is probably useful in showing the trends for the vacancy and homeownership rates, I wouldn't rely on the absolute numbers - and I look forward to the investigation by the Census Bureau on the differences. Unfortunately this report is commonly used by analysts to estimate the excess vacant supply for housing, but - because the vacancy rates do not match the Census data (or the much larger ACS data) - it doesn't appear to be useful for that purpose.

Here are some previous posts about some of the HVS issues by economist Tom Lawler:
• Lawler to Census on Housing Data: "Splainin" Needed Not Just on Vacancy Rate
• Census Bureau on Homeownership Rate: We've got “Some 'Splainin' to Do”
• Be careful with the Housing Vacancies and Homeownership report
• Lawler: Census 2010 and the US Homeownership Rate
• Lawler: Census 2010 Demographic Profile: Highlights, Excess Housing Supply Estimate, and Comparison to HVS
• Lawler: The “Excess Supply of Housing” War
• Lawler: Census Releases Demographic Profile of 12 States and DC: Confirms Bias of HVS
• Lawler: Census 2010 and Excess Vacant Housing Units
• Lawler: On Census Housing Stock/Household Data
• Lawler: Housing Vacancy Survey appears to massively overstate number of vacant housing units
• Lawler: US Households: Why Researchers / Analysts are “Confused”