Mismeasuring inflationPosted: November 22, 2009
When home prices plunged from 2007 to early 2009, some bloggers noted that the Consumer Price Index had done a lousy job of incorporating the decline. The problem, these bloggers pointed out, was the Bureau of Labor Statistics’ use of “owners’ equivalent rent”–the estimated costs that homeowners would assume if they rented their homes instead of owning them–to represent home prices.
Tim Iacono wrote back in 2007: “OER is one of the poorest proxies the world has ever seen as demonstrated by the comparison below with the Case Shiller Home Price Index.”
Source: Tim Iacono
Andrew Jeffrey made a similar point earlier this year:
The statistical alchemists, err, experts, at the Bureau of Labor Statistics use something called “owners equivalent rent,” OER, to measure consumer housing expenses. OER tries to approximate the cost to rent the country’s typical home, and according to the Wall Street Journal makes up 24% of the CPI and 31% of the core CPI, which backs out food and energy costs.
And since even as property values have slid in record-breaking fashion rents remained buoyant, OER has vastly understated the drop in home prices. This means the CPI–were it to reflect some sort of economic reality–would have fallen more than it actually has.
Of course, times change. Although home prices are still well below their peak, and some observers continue to talk of an ongoing crash in home prices, the housing market has been on the mend since the spring of 2009–at least if you believe the S&P/Case-Shiller Home Price Index. As was the case in 2007-08, there is a discrepancy between the Case-Shiller index and the owners’ equivalent rent component of CPI, but this time the mismatch causes inflation to be under-reported rather than over-reported.
Let’s look at the numbers. Between April 2009 and August 2009 (the latest month available for the Case-Shiller index), the Case-Shiller index rose from 139.2 to 146.0–an increase of 4.9 percent. By comparison, the owners’ equivalent rent component of the CPI rose from 256.6 to 257.2 during the same period–an increase of just 0.2 percent. In other words, the rate of home price inflation was nearly 25 times higher than that shown by the CPI.
The CPI as a whole increased from 213.2 to 215.8 — a rise of just 0.1 percent. But if we replace owners’ equivalent rent with the Case Shiller index, CPI would have increased 2.4 percent during this four-month period–an annualized rate of 7 percent.
For simplicity, these calculations are based on seasonally-unadjusted numbers. The results might be slightly different if they were based on seasonally-adjusted figures, but the basic point would not change, i.e., reported inflation since April 2009 would be much higher if not for the CPI’s use of owners’ equivalent rent.
Admittedly, the picture is very different if one looks at year-over-year inflation. The Case-Shiller index declined 11.3 percent between August 2008 and August 2009, whereas owners’ equivalent rent increased 1.7 percent. So CPI significantly overstates year-over-year housing inflation. This overstatement will gradually get smaller before disappearing entirely in mid-2010, unless housing prices begin to fall again. (Recent reports indicate that housing prices are continuing to rise.)