Academic journal article Economic Commentary (Cleveland)

The Slowdown in Residential Investment and Future Prospects

Academic journal article Economic Commentary (Cleveland)

The Slowdown in Residential Investment and Future Prospects

Article excerpt

Housing has historically been a key driver of the U.S. economy. Many past recessions have either coincided with or been caused by downturns in the housing market, and turnarounds in residential investment have helped propel past economic recoveries.

Today, the housing market has improved markedly compared with where it was during the depths of the financial crisis. But concerns linger over the state of residential investment, and activity generally remains at low levels. Contributing to the concern is the fact that residential investment contracted in the past two quarters-the first consecutive quarterly declines since the end of the recession (figure 1).

We disentangle the causes of the recent slowdown in residential investment using a statistical model, and we examine the future prospects for this important sector of the economy. Our model points to three primary factors behind the recent weakness : the increase in mortgage rates since early 2013, the unusually cold winter, and a modest tightening of lending standards in the residential mortgage market. The model suggests that with normal weather and ongoing improvements in labor markets and the broader economy, growth in residential investment should rebound soon. But the experiences of the past year highlight the key role of mortgage rates and the strong interest rate sensitivity of the housing sector. Our forecasts suggest that even moderate increases in mortgage rates through the end of next year could pose a headwind to residential investment.

Put differently, mortgage rates that remain low by historical standards are likely to be an important factor underlying ongoing recovery in the housing market and, by extension, the economy overall.

A Model of Residential Investment

To empirically capture the relationship between residential investment and its key determinants, we use a statistical model known as a Bayesian vector autoregression (BVAR). Our model includes variables that are significant for the U.S. economy and that are expected to have a strong influence on residential investment. Our medium-scale BVAR model includes eleven variables, in the following order: a measure of weather conditions (defined below), real gross domestic product (GDP), real personal consumption expenditures (PCE), a measure of lending standards, a survey-based indicator of consumers' perceptions of home-buying conditions, real residential investment, the unemployment rate, core PCE inflation, PCE inflation, CoreLogic home prices, and the 30-year mortgage rate.1

A few variables merit explanation. Our weather measure captures unseasonably cold temperatures: a large positive weather reading, as occurred in 2014:Q1, is consistent with colder-than-usual temperatures.2 To capture lending standards, we look to the Federal Reserve Board's Senior Loan Officer Opinion Survey on Bank Lending Practices (SLOOS) and use the net percentage of domestic respondents who report tightening standards for residential mortgages.3 Having been negative through mid-2013, SLOOS-based readings have been positive for the last two quarters, and positive readings indicate a net tightening of lending standards. Finally, consumers' perceptions of homebuying conditions are measured by a diffusion index that subtracts the percentage of consumers reporting that it is a bad time to buy a home from the percentage reporting that it is a good time to buy a home in the Thomson Reuters/ University of Michigan Survey of Consumers.

Decomposing the Recent Past

While our model is relatively simple, it can effectively explain the recent weakness in residential investment. Formally, we employ a decomposition to see what the model would have predicted at some previous point in time and then identify what caused the data to diverge from that forecast.4 To do so, we estimate the model using quarterly data from 1990 :Q4 through 2012:Q4 and then generate a forecast for residential investment growth through 2014 :Q1. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.