Friday, September 2, 2022

Apartment inflation

 


This beautiful graph comes from calculatedriskblog.com. (Courtesy Andy Atkeson who used it in a nice discussion of a great paper by Ivan Werning at the Minneapolis Fed Foundations of Monetary Policy conference.) 

The central lines that don't move so much are the average rent. This is the quantity used by the Bureau of Labor Statistics to compute the consumer price index. The blue and yellow lines are the rent of new leases. 

The first thing this informs is the economic theory of "sticky prices." Apartment rents are a classic "sticky price;" the rent is fixed in dollar terms for a year. So, landlords deciding how much rent to charge, and people deciding how much they're willing to pay,  balance rents now vs. higher rents in the future. If everyone believes that inflation will be 10% over the next year, then it makes sense to raise the rent 5% now, and to pay the 5% higher rent, because  the savings at the end of the year balance the cost in the beginning. (Obviously, the economics are much more subtle than this, but you get the idea.) And Voila', you see it. 

The graph also says there is some predictability and nomentum to inflation. Inflation should not be a surprise to forecasters. If you see rents on new leases much above average rents, it's a pretty good bet that average rents will be rising in the future! This kind of phenomenon may be under exploited in formal inflation forecasting. 

And, on the continuing speculation whether inflation will go away with interest rates still substantially below current inflation, the graph does seem a leading indicator that the rational expectations model is winning.  

Clarification: Of course, the graph says nothing about causality; did new leases rise sharply because people expected inflation in average leases, or did new leases rise for other reasons, and we're just seeing the old theorem that marginal  > average when average is rising. But it is consistent with the expectations story, and illuminates that story nicely. 


6 comments:

  1. And perhaps the covid rent suspension policy may have affected landlord calculations as to future profitability?

    ReplyDelete
  2. Surprising graph. Unaffordable housing has contributed to the highest GINI coefficient in 50 years. Houses built in the 50's averaged 1000 sq. feet (1950: 983). Today that's increased to 2500 sq. feet (2014: 2,657). That's nearly tripled since 1950. And apartment sizes are now on the downswing.

    ReplyDelete
  3. Construction of the Zillow index of rental unit valuation is described by Zillow here:
    https://www.zillow.com/research/methodology-zori-repeat-rent-27092/

    The chart shown in the article suggests that the private sector indices have the appearance of a rapid cycle effect, while the government sector indices appear to show a slow cycle effect. Construction of the government rental price indices probably differs significantly from the private-sector produced indices. Zillow's rental price index projects rental price changes from a few transactions to the entire block of rental units without regard to the actual change in rental price of those units that lack a rental transaction change in the period. Effectively, Zillow is constructing a rental block price index, assuming that the entire block changes hands in the period. This will tend to exaggerate the actual rental price change. The government sector indices probably reflect reported price/rental change in a sample. Since real estate transactions are typically few and far between for specific units, the government indices demonstrate a slow cycle effect

    Anonymous's remark that the 2021-22 year-over-year price index %-change probably reflects the effect of landlord efforts to recoup losses from the 2020 pandemic period rental freezes and related losses of income, may have some validity. The year-over-year %-change measure will exaggerate price level change relative to a baseline trend measure. For example, plotting the Zillow rental unit price index values for the City of Seattle and the City of New York showed declines of -5.25% and -9.35% in March 2021 from year earlier rental unit prices, respectively. A year later, the year-over-year %-change in the rental unit price indices for those two cities are +17.85% and +17.05%, respectively.

    Zillow's rental unit price indices for the City of Seattle and the City of New York exhibit a linear trend from 1/15/2014 through 1/15/2020. Extrapolating the trend lines and deducting the Zillow index value for 3/15/22 from the trendline extrapolated index value for 3/15/2022 shows the variance from trend to 3.3% for Seattle, and 6.6% for New York. Seattle's trend line is given by 0.2463 X - 8833, r^2 = 0.929, NYC's trend line is given by 0.1434 X-3548, r^2 = 0.581. X = numerical date value, e.g. calendar date format 3/15/22 = 44635 in Excel's general numeric format. Zillow index values for the U.S. and 100 large metro census units (downloaded from Zillow) were imported into an Excel worksheet and the index values for the U.S. average and City of Seattle and City of New York were plotted in X-Y charts, and linear trendlines automatically computed by the application.

    Year-over-year changes are dramatic, given the pandemic effects on the economic activity in 2020-2022. Headline numbers are eye-catching, to be sure, but fail to reveal the entire story. No doubt the government rental price index components will reflect the year-over-year changes of the fast index values with a lag (12-18 months?). Is this sufficient to cause concern, viz. the policy choices to be made by the FOMC going forward?

    Zillow data series are available here:
    https://www.zillow.com/research/data/
    The data is freely downloadable, in CSV format suitable for most spreadsheet and database manipulation.

    ReplyDelete
  4. The issue about the lag in "official" shelter inflation has been written upon extensively by Summers and colleagues. This inefficiency in shelter cost tracking will create headwinds for "official" inflation for some time to come even if other aspects of inflation show some signs of easing.

    ReplyDelete
  5. I remember Zillow’ algorithms appraising the value of my house, it was a dumb percentage change off of price paid, no knowledge of building permit or 100% increase in square footage. Would not recommend trusting their data.

    ReplyDelete
  6. So how do we interpret the trajectory that appears to have the lines cross sometime this fall? I’d rather look at the new vs average rent level data than the derivative, but still.

    ReplyDelete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.