What Households Notice When Rates Move

The question here is what households’ own reports reveal about mortgage affordability when the nominal rate changes. The evidence comes from the Michigan Survey of Consumers, where respondents say whether it is a good time to buy a home and, in some cases, whether low interest rates are a reason. Reading those answers alongside the nominal mortgage rate and a respondent-specific real-rate measure helps distinguish a household-attitude association from a direct estimate of who actually receives a mortgage.

What the survey measures

The survey outcome is deliberately simple: a respondent is coded as saying it is a good time to buy a home or not. The underlying question asks, in general terms, whether now is a good time or a bad time to buy a home. It does not ask whether the respondent has submitted an application, qualified for a loan, signed a contract, or completed a purchase. That distinction matters. A person can judge that conditions are unfavorable while still buying because a job, family need, or unusual opportunity makes the purchase worthwhile. Another person can report favorable conditions but be unable to finance the desired home.

The simplicity of the outcome is an advantage for the question the paper asks first. It captures a broad measure of home-buying sentiment across a long period. The authors pair it with the actual national 30-year fixed mortgage rate, rather than with a respondent’s belief about what rate is available. They also construct a respondent-specific real mortgage rate by subtracting the respondent’s expected annual inflation over the next five to ten years from that nominal mortgage rate. In rough notation:

expected real mortgage rate ≈ nominal mortgage rate − expected inflation.

The two rates are related but do different conceptual work. The nominal rate is the quoted rate that helps determine the dollar payment for a new fixed-rate mortgage. The real rate is meant to capture the purchasing-power cost after expected inflation. If household sentiment responds to the nominal rate even after the real rate is included, that pattern is consistent with the contract-level affordability channel: people may be reacting to the payment they see in dollars, not only to an inflation-adjusted intertemporal price.

The baseline regression uses the respondent’s home-buying attitude as the outcome and includes the national nominal mortgage rate, the respondent-specific real mortgage rate, and a monthly time trend. The sample runs from February 1979 through June 2025 and contains 228,425 observations in the baseline table. The time trend matters because housing attitudes can change over decades for reasons that have nothing to do with a particular month’s mortgage rate. A trend does not solve every problem in a long time series, but it prevents the rate coefficient from simply absorbing an undifferentiated secular movement in sentiment.

The main nominal-rate coefficient is approximately -0.11. In the paper’s linear probability specification, that means a one-percentage-point increase in the nominal mortgage rate is associated with an 11-percentage-point lower probability that a respondent says it is a good time to buy a home, holding the included real-rate measure and time trend constant. The number is large, but its interpretation must stay attached to the outcome and comparison.