The jump in the price of shelter, by 0.5% month on month, accounted for more than the total increase in core CPI (less food and energy). The weighted contribution of shelter to core CPI was 0.16%, vs. a total increase in core CPI of 1.58%. The shelter series has behaved erratically during the past several months, and the big increase probably reflects some underestimation earlier in the year.
The implication of this last observation is remarkable: Net of shelter, we had a net decline in consumer prices ex-food and energy during August. Items that declined include used cars and trucks (driven by credit problems among low-income car buyers), medical care commodities, and airline fares. Commodities less food and energy showed a -0.1% decline for the month, or a 1.3% annual rate of decline. The overall inflation rate in the US excluding food, energy and shelter is only 0.5% year on year, the lowest since 2010.
This suggests that demand is exceptionally weak outside of the housing market.
It would be shocking if housing prices did not increase at a faster pace than the rest of the economy, considering that they only regained their 2007 peak recently, and that mortgage rates remain extremely low. For the subset of American households that can afford to buy a home, the availability of 30-year mortgages at less than 4% with deductible interest is a big incentive. At the present 6% rate of housing price increase, a home purchase with 20% down yields a 30% gross rate of return, financed at an after-tax cost of slightly over 2%.
Time series analysis shows that the shelter component of CPI responds to changes in home prices with a considerable lag. I surmise that the relative low prints for shelter CPI earlier this year reflected a pause in home price growth during 2016, and that this month’s jump gets the series back to trend.
Deflation in used car and truck prices, meanwhile, tells a different story about the state of household demand. The difficulty in measuring demand stems from the bifurcation of the economy (a growing disparity between mean and median incomes, for example), with growing distress in lower-income households masked by affluence among the upper tier of households.
The net effect is extremely low inflation in CPI net of food, energy and shelter, which may be the most meaningful measure of price change, given the special circumstances affecting shelter.