Faulty rental estimate skews US inflation data
‘Who are you going to believe, me, or your own eyes?’
“Who are you going to believe, me, or your own eyes?,” Groucho Marx once asked, and the same question is implicit in this morning’s release of closely-watched US consumer price data. The official data just don’t correspond to actual observations on the ground.
A jump in the reported cost of shelter pushed monthly core inflation up to 1.8% year-on-year vs. a prior level of 1.7% year-on-year, higher than the forecasters’ consensus anticipated. The small difference in numbers is actually a big difference, because the Federal Reserve has said that core inflation (excluding volatile food and energy prices) above 2% would motivate additional tightening.
Market reaction to the unexpectedly high number was muted. The 10-year Treasury yield rose by 4 basis points, somewhat less than what we would expect if inflation really were about to roar.
The decisive number in the inflation report is shelter, which makes up 34% of the total Consumer Price Index and more than 40% of the core index net of food and energy. The Bureau of Labor Statistics reported a 0.4% rise in December vs. a 0.2% rise in November, or an annual rate of just below 5%.
Eyes on the ground, though, see something entirely different. The largest US shelter website, Zillow, compiles its own national statistics on rents.
Evidently the BLS underestimated rental inflation during 2015, but overestimated it through most of 2016 and 2017—and by a very large margin.
Bond investors will need to see a few months’ more data before drawing conclusions from government data that deviate from private-sector observations.