The Most Important New Disinflation Indicator

Updated Data Shows Newly-Signed Lease Prices are Cooling Rapidly—and Housing Inflation Will Follow

Hi All!

First, I wanted to say that last week Fast Company recognized me as one of this year’s most creative people in business for my work on Apricitas—thank you all for the support you’ve given me and the newsletter over the last two years. I couldn’t have done it without you.

Also, welcome to the 2,000 new readers who joined in the wake of the profile! I hope you enjoy what you read here—if you’re looking for where to start, some of my favorite posts from this year are on how the financial system has changed in the wake of this year’s banking crisis, Russia’s shifting post-invasion trade relationships, and how higher mortgage rates are hitting the US housing market.

At the start of this year, I highlighted a new data series from researchers at the Bureau of Labor Statistics and Cleveland Fed, calling it “the most important new inflation indicator.” That is a title I do not bestow lightly.

Their research addressed a massive concern with official measurements of rent prices, which alone make up roughly one-third of the Consumer Price Index (CPI). To accurately calculate housing inflation, it's necessary to measure contract rents—the amount tenants are currently paying—instead of the rents for homes that are currently for lease. That in turn requires repeatedly surveying a large sample of housing units—most of which will have longer-term leases that last for a year or longer and are often just re-signed by the preexisting tenant upon expiry. The end result is an extremely accurate way of measuring average prices, but a thorny problem for policymakers—official inflation always reflects shifts in average rents, not new rents, and therefore lags current macroeconomic conditions significantly. Since housing inflation is the single most cyclical component of overall inflation and also the most heavily influenced by monetary policy, it’s especially important to get an accurate hold on rent prices—making this lag a major problem.

This is where the New Tenant Repeat Rent (NTRR) Index comes to the rescue—developed by Brian Adams, Lara Loewenstein, Hugh Montag, and Randal J. Verbrugge, it uses the same underlying microdata as the CPI to look at price changes for only the subset of units where new tenants have signed leases. That gives a much better picture of where the housing market is right now, and by proxy where official rent inflation is headed. Critically, the NTRR tends to lead the official CPI rent components by one year—and right now, it is saying we should expect significant disinflation over the next three quarters.

Inflation is Now Mostly a Housing Issue

That’s good news because inflation is now mostly a housing phenomenon—rising rents are the single-largest contributor to the CPI by a wide margin, and outside of shelter headline inflation has completely flatlined over the last year. To be fair, just aggregating all non-housing costs as “everything else” masks the fact that rising prices for core goods and nonhousing services have been counteracted by a fall in volatile energy prices, and it’s also worth noting that the Federal Reserve’s preferred Personal Consumption Expenditures Price Index (PCEPI) places a much smaller weight on housing than the CPI. However, even when looking at excess core trend PCE inflation the largest individual contribution comes from sector-specific housing price dynamics—making the future path of rents especially critical to normalizing inflation.

The good news is that the NTRR has decelerated substantially over the last year—and that its deceleration is strongly corroborated by slowdowns in comparable private-sector data. Growth rates in the Zillow Observed Rent Index have decelerated to normal levels, while the NTRR and ApartmentList Median New Lease estimates show functionally zero growth over the last year. That deceleration is already starting to pass through to official CPI inflation—though with a longer lag than originally expected.

Indeed, prior to the publication of the NTRR, accurately estimating the precise lags between new tenant rent prices and official CPI data proved difficult. In normal times, both private sector new lease data and official rent indices move at a relatively constant pace, so there is not much of a noticeable discrepancy caused by lags. The initial collapse in nationwide asking rents caused by the pandemic showed up almost immediately in private-sector new lease indices, and a corresponding slowdown hit the CPI just 6-8 months later, leading many people (including myself) to expect the subsequent surge and deceleration of the Zillow/ApartmentList data to show up in CPI within a similar timeframe. Instead, the lags of the price acceleration were significantly longer.

By looking at a metro-area level, comparing the available BLS rent data from America’s 21-largest cities to their Zillow/ApartmentList counterparts, we can better appreciate how the data lags have manifested. The average metro-area level lag between the lowest point of Zillow/ApartmentList rent growth and CPI rent growth was a meager 6.3 months, but the average metro-area level lag between the highest growth point of both data series was more than double that at 13.3 months. That 6.3-month lag headfaked me and others into originally expecting swift rent disinflation, but the actual 13.3-month lag lines up fairly well with the 1-year lag estimated by the NTRR researchers. Critically, it also lines up with what we’re seeing in their newly-release data, lending more credibility to the idea that rent decelerations will continue from Q3 of this year through Q1 of 2024.

Also, the updated NTRR data answers a key lingering question from the initial release—whether the price of all rentals would eventually catch up with the significant growth in new rental prices, or if new rentals had seen disproportionately high price growth that would not be mimicked by the rest of the market. As of Q1, the CPI had fully caught up to the ATRR while the NTRR had declined to converge with both, meaning there is now substantially less risk that CPI rent inflation will remain high in order to “catch up” even as NTRR cools. When I spoke to the BLS and Cleveland Fed research teams this Thursday, they stressed that they do not assign any special significance to the three indices converging, and I likewise believe it is the NTRR growth rate, not the index level, that is relevant for CPI forecasting—but it is at least good to now be able to put this risk to bed.

However, the NTRR research team likewise cautioned me about the volatility of recent point estimates, in particular, the 0% year-on-year growth recorded in Q1 2023. Initial readings from the index comprise a much smaller sub-sample of new lease signings than what becomes fully available several months later since, for example, respondents to the June CPI can indicate that their unit signed a new lease in February. That’s why confidence intervals for the most recent quarter are always so high, and why we should be skeptical that new lease prices are collapsing even as we can be confident they have slowed significantly.

A Victory For the Speed Limit Theory of Inflation?

When thinking about inflation, most economists and central bankers have an implicit or explicit model containing a Phillips curve and a “natural rate of unemployment”—the idea being that when unemployment rates get below a “normal” level the economy becomes too hot and inflation begins accelerating. The corollary to that is when inflation starts getting out of hand the “responsible” thing to do is immediately induce a rise in unemployment—failure to do so will result in inflation becoming entrenched and requiring an even deeper recession in the future to bring price growth back to normal.

However, this has gone against much of the pre-pandemic research suggesting a low and stable Phillips curve, to the extent one exists at all. Plus, it invites one really obvious line of questioning: what could have changed such that the 3.5% unemployment rate of 2019 coexisted with substantially below-target US inflation but that the 3.5% unemployment rate of 2023 is supposedly going to cause entrenched, runaway inflation?

Perhaps the core mistake comes from thinking about the relationship between the labor market and inflation as one of levels and not one of growth rates. When I first wrote about the NTRR last year, I pointed out that it has an extremely tight relationship with growth in the aggregate wages and salaries of workers, something called Gross Labor Income. That makes intuitive sense—renters tend to spend a relatively fixed portion of their income on housing, and new workers often require new housing, so a rise in aggregate wages entails some proportional rise in headline rents—and it helps explain why shelter costs are the most cyclical and monetary policy driven components of inflation. However, critically, it is not the level of labor market strength per se that is the primary driver of rent inflation but the growth in the labor market—which is how a 3.5% unemployment rate can produce low rent inflation in 2019 and extremely high rent inflation during the pandemic.

Hence, the “speed limit” theory of the pandemic-era inflation: the massive pace of job and wage gains in 2021/2022 was always going to induce price increases as the economy recovered, but as growth rates slowed naturally and in combination with tighter monetary policy inflation would likewise cool off—without an outright contraction in employment being necessary. That’s how employment rates have reached the highest levels in 20 years even as inflation cools.

Indeed, far from a wage-price spiral, it looks as though quarterly growth in Gross Labor Income has completely normalized over the last nine months, growing at just over a 6% rate in Q4 2022 and Q1 2023 before slowing down to just over a 4% annualized rate in the last quarter. It’s still too early to declare a win—much of the official disinflation that has occurred thus far has been thanks to supply-side, not demand-side renormalization—but if the NTRR is any indication, unemployment may not yet need to rise in order to stop today’s inflation.