House Prices Push US Inflation Higher

Summary

• A new NBER working paper co-authored by economist Larry Summers looks at how rising housing and rental prices will affect US inflation in 2022 and 2023.
• The results are staggering. Housing inflation is set to contribute 2.6pp to YoY core CPI by December 2022 – the most since 1986.
• The predicted rise could completely offset a normalisation of used car and truck prices from 40% to 2%. And, if all other components were at the target rate of 2% YoY, it would keep headline CPI at 3.4% and core CPI at 3.8%.

Introduction

For the first time this century, UK households are paying 40% of their pre-tax incomes on rent. It is a symptom of ballooning prices in the residential real estate market. Prices in the UK and US are both up over 10% YoY and, with incomes struggling to catch up, renters and homeowners must now put more aside than ever to pay for accommodation.

Despite this, we are yet to see the full impact of higher residential real estate prices on inflation, according to a new NBER working paper. In the US, residential services inflation, which accounts for a large proportion of common inflation measures, was around 4% YoY in January 2022. Based on the paper’s analysis, this YoY figure is likely to increase to 6.5-7% in 2022, adding as much as 2.6pp to core CPI.

How House Prices Feed Into Inflation

The US measures inflation with two main price indices – the Personal Consumption Expenditure (PCE) Price Index, and the Consumer Price Index (CPI). They each assign very different weights to housing. For example, the two largest components of residential services, owners’ equivalent rent of residence (OER) and rent of primary residence (Rent), account for roughly 40% of core CPI but just 17% of core PCE (Chart 1).

The reason behind the different weights is data collection. For CPI, OER comes from the Current Expenditure Survey of households. This asks, ‘if someone were to rent your home today, how much do you think it would rent for monthly, unfurnished and without utilities?’ In the PCE, the share of income used on housing for rent is estimated from PCE data in the National Income and Product Accounts.

But, while the housing weights differ across PCE and CPI, how each index tracks changes in housing prices over time does not. They both use the BLS’s CPI price measure for inflation. The reason is that assessing rent changes, especially changes in OER, is a difficult problem. And the BLS methodology creates an accurate but, crucially, backward-looking measure.

Why backward-looking? The prices of most items in the PCE and CPI indexes are collected monthly or bimonthly. Rental data, though, is collected every six months because households will often be tied into fixed long-term leases. Then, to circumvent the problem of only getting data every six months from one area, the BLS measure creates six different sub-groups (called panels) within each area. The OER and rent data for Panel 1 is collected in January and July, Panel 2 in February and August, etc.

Many have studied the impact of calculating monthly OER and Rent using this method. The consensus is that it over-smooths rental growth and reflects spot rental rates with a long lag. One author documents how the BLS measure of the 12-month change in rental prices correlates more strongly with the 12-month change in housing prices observed 16 months earlier than any more recent reading. Another finds evidence that we can use house price changes five quarters earlier to forecast inflation.

Forecasting Inflation

Using the above information, the authors use private-sector data on rental and housing prices to forecast the impact on OER and Rent, and therefore also the impact on CPI and PCE inflation. For rental prices, they use Zillow (ZORI) and CoreLogic, which are available from 2014 and 2005, respectively. For home prices, they also use Zillow (ZHVI), alongside the S&P/Case-Shiller Home Price Index.

These private sector measures of prices have historically been a powerful leading indicator of OER and rent inflation. For example, YoY changes in the CPI’s OER component correlate most strongly with the 16-month lag of Zillow’s Home Value Index (Chart 2). Meanwhile, YoY changes in the CPI’s rent component correlate most strongly with a 13-month lag of CoreLogic’s Single-Family Rent Index.

To make the forecasts, the authors use the ‘elastic net regression’ method. This is a machine-learning algorithm that performs regularisation and variable selection on linear models. It is a popular forecasting method for macroeconomic series, but it yields similar results to a simple OLS regression in this case.

Forecasts of OER in 2022 and 2023

Using monthly data from 2005 to January 2022, the authors begin by forecasting OER inflation in 2022 and 2023. Their forecasts show a steady increase in the first half of 2022, peaking in December at 6.3% and 6.4% for CPI and PCE inflation, respectively (Chart 3). This will be the highest YoY rise since the 1980s. OER is then expected to remain elevated in 2023, remaining above 6% for much of the year. The YoY growth in rent inflation follows a similar dynamic, rising above 7% in December 2022 (Chart 4).

As for the accuracy of these forecasts, the authors test how well the model performs on a rolling basis between 2014 and 2021. The one-month-ahead predictions differ from the actual rate of OER inflation by only 0.06pp, on average. The 12-month-ahead forecasts miss by 0.34pp on average.

Contributions to Headline and Core Inflation

The forecasts imply that housing will form a significant part of overall inflation in 2022. Using the weights in Chart 1, the authors can forecast the contribution of OER and rent to overall inflation. For US core CPI, which currently stands at 6%, they expect the projected rise in housing inflation to add an additional 1pp by December 2022.

In other words, if all other components remain unchanged, core CPI would rise to 7% by the end of the year because of the recent jump in housing and rental prices. In another scenario, if all other components return to the target of 2%, the increasing contribution of housing inflation would still keep core CPI at 3.8% and headline CPI at 3.4%.

In total, the authors project rent and OER to add a total of 2.6pp to core CPI by the end of the year (Chart 5). This would be the largest contribution of housing to core CPI since 1986. And this contribution would be high enough to offset a complete reduction of YoY changes in used car and truck prices back to 2% (currently over 30%). For core PCE, the contribution is set to increase from 0.6 to 1.1pp.

Bottom Line

Housing and rental prices are growing at their fastest pace in over 40 years. The rise has outlasted pandemic-induced demand shifts, and a supply shortage is exacerbating it. In January 2022, both rental and ownership vacancies were at their lowest recorded levels, and one of the largest US homebuilders recently stated these supply bottlenecks are unlikely to abate soon.

The impacts of these supply and demand dynamics have not impacted headline and core CPI significantly yet. But they will soon. If the forecasts from the paper hold true, housing inflation will keep inflation elevated throughout 2022 and 2023. Strikingly, the results show that even if the prices of all other components normalise, core CPI will not fall below 3.8% this year.