Do changes in consumers' home buying attitudes predict directional change in home sales?

Abstract

The Survey of Consumers probes consumer sentiment on personal finances, business conditions and buying conditions. We focus on the latter category by examining the surveys question ‘Generally speaking, do you think now is a good time or a bad time to buy a house?’ and the follow-up question ‘Why do you say so?’ The responses to these questions provide us with several measures of consumers' home buying attitudes. We show that changes in survey data on home buying attitudes (measured by consumers' assessments of current and future mortgage interest rates and house prices) accurately predict the direction of change in home sales 3 months later.

Keywords:

• Surveys of Consumers
• consumer sentiment
• US housing market
• mortgage rates

JEL Classification:

• E2
• E4
• R2

Notes

¹ The percentage of those who chose ‘uncertain’ is around 4%, on average, for our sample period.

² Survey respondents are allowed to select more than one option as the reason for their answers. As such, the answers for the options may add to more than 100%. For simplicity, however, we have subtracted Ph, Rh, Ca and Uf from 100.

³ The MacKinnon p-value of the calculated Augmented Dickey–Fuller (ADF) test statistic is 0.309 (0.618) for January 1984 to December 1995 (January 1996 to June 2008). The number of augmented lags in the ADF test equation is determined using the Schwarz information criterion. Our conclusion that home sales (in logarithms) has a unit root remains unchanged when we employ other commonly used unit root tests.

⁴ The p-value of the calculated F-statistic is 0.056 (<0.10).

⁵ As noted in the text, the actual data on home sales for month t − 1 are released near the end of month t. This is right before when the survey data for month t become available, enabling a user to form a prediction about the direction of change in home sales 3 months later (t + 3).