I have written before about oddities with government survey methods:
I was recently pointed to an even better example. Bolhius et. al show how future inflation can be predicted using real-time measures of housing and rental prices, especially when the economy experiences rent shocks. I would have taken Larry Summers’ confident early 2022 inflation prediction even more seriously, had I read this first (Summers is a co-author on the paper).
The key insight in the paper rests on how the BLS and BEA estimate prices for renters and homeowners. In short, consumers do not experience the same price volatility in housing costs as they do in say food (because they sign long-term leases or long-term fixed rate mortgages and change homes relatively rarely) but a spot market for the real-time cost of housing exists. Conventional inflation indices attempt to approximate the former, but eventually converge on the latter. So, if one has a reliable measure of the current price of new rentals or housing costs, the future housing component of inflation is mechanically predictable. I recommend reading the paper for a full explanation.
Bolhius et. al use an index of housing costs from Zillow and the S&P Case-Shiller index to show robustness, indeed, any faster measure of the current housing costs would do1.
The philosophical question that arises is. whether the BLS or BEA are incorrect in their measure of housing inflation. If the objective of the survey is to measure the current average consumption cost incurred by the population on housing, then the existing measure may well be more accurate as most homeowners are not paying market-clearing prices (especially in the United States, where 30-year fixed rate mortgages are common).
From a practical perspective, if the objective is to understand the trajectory of future inflation, predict future Federal Reserve actions, or just understand Summers’ confidence, then clearly these faster measures are more useful. More broadly, these types of mechanical mismatches are among the most lucrative sources of alpha for investors looking for alpha.