Between 1999 and 2006, there were two episodes during which inflation in the Rent index in the CPI diverged markedly from inflation in the index for Owner's Equivalent Rent (OER); early in 2007, these series began to diverge again. Such divergence often prompts many to question CPI methods. A key difference between these two series is that OER indexes are based upon rents which have received a utilities adjustment — an adjustment which is necessary because the OER index is intended to track pure rent-of-shelter, not shelter-plus-utilities. Critics have claimed that the Rent-OER inflation divergences stem from an inappropriate utilities adjustment.
This claim is false. In this paper, we decompose the Rent-OER inflation differential into its various determinants, and explore the multiple causes of this divergence over time. There is only one divergence episode — of only six months duration — which is primarily attributable to the utilities adjustment procedure. Indeed, the utilities adjustment sometimes reduced potential divergence between the two series.
Instead, the main culprit is rental market segmentation; that is, different rent inflation rates were experienced by different parts of the rental market. Before 2003, the Rent-OER inflation divergence mainly resulted from divergent rental inflation rates within metropolitan areas: areas with a higher proportion of renters experienced higher rental inflation. After 2004, similar divergent inflation across metropolitan areas resulted in higher Rent inflation. Compared to other units, rent control units experienced higher inflation in 2004 (and, to a lesser extent, before mid-2001 and in 2006), which increased Rent inflation but not OER inflation. Finally, in early 2007, there was a sizable divergence between OER and Rent inflation, driven mostly by divergent rental inflation rates within metropolitan areas; the extent of the divergence only becomes evident once the effect of the utilities adjustment is accounted for.