The ratio of the home price to median household income has nearly doubled since 2000. It's counterfactual to think that would have occurred if the BLS (Bureau of Labor Statistics) had not changed in 1999 its survey of owner-occupied housing to measure implicit owner-rents. Unbeknownst to many, the change has led to a significant undercount of housing and overall CPI inflation for decades.
In 1999, BLS made a significant change by 'simplifying the housing survey' to measure owners-rents in the CPI. It no longer tracked two types of housing units, owners and renters, with different characteristics and, in many cases, various locations. BLS argued that this change was necessary because it could no longer find an adequate sample of owner-occupied units for rent. BLS stated that the change was made to ensure the accuracy of the data, but the actual outcome was an understatement of housing inflation and overall inflation.
During the housing boom of the early 2000s, a crucial period that underscored the importance of accurate inflation data, I debated with BLS at an annual economic conference in 2004. Our research found that the rental survey underestimated actual or experienced inflation by several hundred basis points each year in the owner-occupied housing market. My main point was that the owner and rental markets were two distinct housing markets with different vacancy rates, and the latter was a primary determinant of rent changes.
BLS countered, stating, "There is no clear theoretical relationship between the homeowner vacancy rate and owners' equivalent rent." I emphasized that the practical relationship states otherwise. When housing prices are rising rapidly, and there is a short supply of housing units, why would owner-rents track the rental market and not the owner-housing market? Different markets yield different inflation.
This practical relationship, often overlooked, is crucial to understanding the actual inflation rates in the owner housing market. During the housing boom of the early 2000s, owners' rents increased between 2% and 3% per year, as they were linked statistically to rents of primary residences, while house prices rose at double-digit rates. Even today, owners' rents are not tracking house price inflation.
Many factors are behind house price inflation. Yet, if consumer price inflation tracked actual house price inflation versus a "hypothetical measure," it is hard to argue against the view the fact that overall inflation would not be higher, as official and market interest rates would be as well, resulting in a median house price much below current levels.
The change in the housing rent survey is a much more significant BLS blunder than the recent controversy over the overstatement of payroll job growth, as it affects reported inflation, inflation adjustments to entitlement programs, and, in some cases, workers' wages and official and market interest rates.
Comments