What to expect of inflation

This week, the April update on CPI inflation will be released. I think we might be coming into an interesting time.

Rent inflation has been messing up monetary policy for more than 20 years. Since federal bean counters treat rising land rents (and mostly imputed rents of homeowners to boot) the same as rising nominal cost of production, rent inflation has been continually pushing up inflation numbers. (Don’t get me wrong. That’s a perfectly reasonable way to track cost of living. But elevated imputed rents shouldn’t be dominating cyclical decisions about currency growth.) And, it has especially been an important factor leading into recessions. Before each of the last 3 recessions, core CPI was above the 2% target, but core CPI excluding shelter was below it. And the gap keeps getting larger.

Figure 1

In Building from the Ground Up, I wrote about the Fed in 2006:

The Fed was purposefully slowing down housing construction, but they were still aiming for stability. Bernanke expected this investment to be reallocated to other sectors. The Fed saw the housing boom as inflationary. But how exactly was a decline in construction going to reduce inflation? Was a decline in construction going to reduce inflation in San Francisco? The inflation there was largely due to a shortage of housing. Would it reduce inflation in Dallas? In Dallas, home prices weren’t excessive, so home equity lines of credit weren’t a sizable source of funds for household consumption, and residential investment there had been pushing down rent inflation. Economic growth did slow in 2006 as the Fed had intended, but it was a decline in real economic growth without much effect on inflation—mostly because rents increased as home building cooled off.

The persistent inflationary pull of the housing shortage, together with the pro-cyclical inflationary pulse that happens when housing starts decline, was an important element early in the Fed’s transition from aiming for stability in 2005 to aiming for crisis in 2008.

The post-Covid period has been unusual. Since the CPI shelter component is measured with a lag, the transitory spike in inflation didn’t show up in the shelter component in a timely way. So, in 2022, aggregate CPI measures understated the true scale of the transitory inflation. If shelter had been measured with a timely component, the CPI would have shown annualized monthly inflation higher than 10% for much of 2021 and the first half of 2022.

In that case, the lag of the CPI shelter component probably helped the Fed maintain better policy because it muted the mistaken calls for monetary austerity, and may have given the Fed room to keep their target interest rate low until transitory inflation had passed.

But, it also has led to some confusion on the topic of housing. Since reported CPI inflation was understated, it appears that home prices had an unusual jump after Covid. But, with correctly measured inflation, home prices were not unusual. If you match home price trends to rent trends from 2015 to 2019, then the relationship continues during and after Covid, until today. Basically, rent inflation is always 2% to 3% above general inflation. Prices across the board added a one-time pop of about 7%, and except for a wiggle here and there, the persistent trends continued on top of that one-time pop.

So, in Figure 2, as of this month, home prices are exactly where you would expect them to be if all you knew were what home price and rent trends were before 2020 and what level rents were at today. If you want to attribute that little temporary 5% bump in 2022 to low interest rates, knock yourself out. But, in general, lesson number 1 at the EHT is, “Rents explain everything.” and lesson number 2 is “Interest rates don’t explain much.”

Figure 2

This also tosses some cold water on the Covid demand story. The conventional story is that there was a big demand spike coming out of Covid. Clearly there is some truth to that. But, the problem is that our pitifully low level of annual construction and increasingly inelastic supply condition makes the smallest little rise in demand seem like a really big deal. Even the 2005 boom represented a very mild increase in housing demand. It would have barely registered as a boom in the mid-20th century, based on construction activity. Of course things have only gotten worse since then. And, in addition to having much lower construction activity, we have harvested much of the vacant stock. So, there is nowhere left for housing supply to budge.

There is a short run vs. long run issue in housing supply. Some might point out that, even if long run supply is elastic, short run supply is inelastic because it takes months to build a unit.

Vacant supply is very elastic. Vacancies were regionally low before 2008, so there was a regional price boom. Vacancies are universally low today. Inelastic short run supply (because of a lack of vacancies) is the result of layers of failure over decades.

In total, the net demand spike in 2021 and 2022 was quite small. There was an uptick in home sales. But, also, I shouldn’t have to remind you that at the time, a million and a half Americans died in a plague.

So, again, comparing rent inflation to general inflation, the accumulating rent premium is basically where it would have been if you drew a straight line from the 2015-2019 trend. New home completions have been almost flat for this entire period. There was no additional quantity of new supply. Both construction and rent inflation are near their pre-Covid trends. There really is no sign, all told, of a national demand spike, in either quantities or rents. We’re just marching right along the same supply crisis we’ve been marching along for decades. There were just a couple wiggles in the trend during Covid, and a lot of trading among the existing stock and between locations.

Rising general inflation explains rising rents and rising rents explain rising home prices. Not only are factors like low interest rates poorly identified. They are poorly identified to explain phantom outcomes. There is nothing to explain.

Figure 3

Figure 4 compares Core CPI inflation excluding shelter, the CPI shelter component, and the Zillow estimate of US rent. (The Zillow measure tends to run about 0.8% higher than the quality adjusted CPI.) Here, I have added in a very simple, poor-man’s lagged rent estimate, using a mixture of lagged Zillow rent estimates to try to mimic the CPI shelter component (the dotted line). It’s not perfect, but for our purposes here, I think this basically can provide us with a way to forecast what to expect with CPI inflation moving forward.

Figure 4

I think understanding all the points I have discussed above would improve policy decisions, and I also think this can improve asset allocation and financial trading decisions. Analyzing recent trends, I think there may be a somewhat surprising turn in the CPI coming up. My discussion of that will be below the paywall.