Inflation is complicated

As expected, U.S. inflation – both headline and core – moderated in July, according to the latest Consumer Price Index data. And assuming global food and energy commodity prices continue to ease, June likely marked the peak in the year-over-year (y/y) rate of headline inflation.

The y/y rate of core inflation also appears poised to peak, although it may not happen until September, due to a combination of easing supply chain bottlenecks, the stronger U.S. dollar, and the pass-through of lower commodities prices onto core CPI components (more notable in the core transportation services category, which includes airfares). 

The fact that inflation is peaking and will moderate isn’t debated. In a survey of 75 economists taken by the Wall Street Journal, all 75 expect inflation to moderate from its current level. Where there is more disagreement is the destination after the moderation. In the same WSJ survey, the range of 2023 CPI forecasts spans 0.0% to 5.3%, while 2024 isn’t much better with a forecast range of 0.3% to 4.1%. Elevated disagreement can also be seen in the University of Michigan Survey, where the 25th and 75th percentile of longer-term inflation expectations is the widest it’s been since the late 1980s. The New York Fed consumer survey of inflation expectations is realizing a similar phenomenon. 

What’s driving the divergence in forecasts? While there are a multitude of factors that contribute to the outlook, we think the key question can be boiled down to this: How sticky is the underlying trend in inflation? This question is important because, as researchers at the Atlanta Fed show, inflation “stickiness” is related to the inflation expectations of the average household. And while we also try to measure inflation expectations through various surveys, surveys can be biased, noisy indicators, which can be heavily influenced by volatile food and energy prices. 

Since households’ purchasing behavior is influenced by near-term expectations, and households are more inclined to consume today if they expect prices to rise in the future, rising inflation in stickier categories can signal that inflation expectations are rising, even if other survey-based measures are well contained. This divergence in signals appears to be happening today, likely contributing to the disagreement among forecasts for where inflation ultimately lands. Looking across a range of inflation expectations surveys suggests that inflation expectations have risen since the start of the pandemic, but the rise largely retraced a decline that occurred around 2016 (for more on inflation expectations surveys, see the September 2020 Fed Note, Index of Common Inflation Expectations).

Meanwhile, measures of sticky prices have exhibited a noticeable trend higher, according to the Atlanta Fed, and at 5.6% y/y are currently sitting at their highest level since the 1990s (from 2008–2019, y/y inflation in the CPI-based sticky price measure was at or close to 2.0%). 

Because shelter inflation makes up a large portion of the sticky price index, and rental unit prices are likely to continue to rise, sticky price inflation appears poised to accelerate further in the coming months, and remain elevated after other flexible components (those impacted by pandemic-related supply/demand imbalances) moderate. We expect Owners’ Equivalent Rents (the CPI measure of rental inflation) to reach 8% by the end of this year, vs. the current rate of 5.8% and a pre-pandemic average of 3%–3.5%.This would put rental inflation well past the peak realized in the 1980s of 6.7% y/y. To understand why, consider two facts: 

First, the CPI’s use of lease current costs instead of a market cost causes the CPI measure to lag broader rental and housing market trends by around 6 to 12 months. The CPI estimates shelter costs through a survey of renters, who are asked about the cost of rent they currently pay. Because rental lease contracts tend to lock in rents for 6 to 12 months, the measure will 1) only slowly track current market rents and 2) usually exhibit less volatility if rollover leases don’t adjust fully to market rates. Today, various alternative data sources suggest that rental market prices have adjusted very quickly, and although rollover leases historically tend to adjust less than the current market value, the magnitude of the recent market value adjustments has given landlords a big incentive to hold out for much higher rents from current lessees with expiring contracts. 

Second, and perhaps counterintuitively, when the Federal Reserve raises rates, as it has recently, rental inflation tends to risein response, at least at first. This is because rising rates make owning a home less affordable, pushing would-be buyers into the rental market, which in turn further drives up rents. Usually it’s not until housing price inflation starts to cool – as a result of rising unemployment and falling aggregate income growth – that rental market inflation also decelerates. Today, the sequential monthly pace of national housing price indexes hasn’t clearly moderated from the pandemic-induced frenzy, while the Federal Reserve is expected to continue to increase interest rates, at least over the next several quarters, as labor markets and aggregate income growth remain strong. 

Why does any of this matter? The moderation in headline and core inflation measures that are likely in the coming months will take pressure off of the Fed, and we expect the central bank will slow the pace of rate hikes in the second half of the year. However, with the underlying trend in CPI inflation now appearing to be closer to 3.5%–4% (well above the Fed’s long-term goal), we doubt Fed officials will declare victory. Furthermore, because the government’s lease current cost measure of rental inflation tends to lag current market trends, it may well be 6 to 12 months before any rise in unemployment and a deceleration in market rents is reflected in the government data. This also raises the risk of a hard landing as monetary policy reacts to lag trends in the inflation data.