Forty-three percent of US homeowners with an adjustable-rate mortgage reported regretting their choice in a December 2022 survey of 1,203 ARM holders conducted by PureSpectrum for U.S. News & World Report, the larger and earlier-fielded of the two ARM-regret surveys cited here. The primary driver, cited by 44% of regretters, was that their interest rate had adjusted to a higher level than they anticipated. A follow-up survey by Point in January 2024, drawing on 332 ARM holders, placed the regret rate at 70%, consistent with the pattern of escalating regret as further Federal Reserve rate increases pushed variable payments well above their initial levels. By contrast, the fixed-rate regret rate — proxied from Bankrate’s 2025 general homeowner survey (n=1,363), where roughly 90% of respondents hold fixed-rate loans — runs substantially lower: approximately 12%, reflecting owners who locked in at elevated 2022-2024 rates and carry payment regret without regretting the choice of a fixed product over a variable one.
The evidence base is asymmetric. Both ARM surveys asked a direct regret question to ARM holders specifically; the fixed-rate figure is a cross-survey proxy, not a head-to-head comparison. No published study has asked matched samples of ARM and fixed-rate borrowers the same regret question in the same instrument. The direction of the asymmetry is nonetheless robust: the ARM regret rate climbed from 43% to 70% between the two survey waves as the rate environment deteriorated, while fixed-rate holders’ payment complaints remained bounded by the absence of unexpected rate changes. In the US context, the fixed-rate mortgage is the dominant default (over 90% of originations since 2010), so choosing an ARM represents an active deviation from that default — a bet that rates would stay low. Borrowers who took that bet during 2019-2022 and then faced resets into a 7%-rate environment bear the hallmarks of action regret: a counterfactual (the fixed product they passed up) that is concrete, specific, and visible in every mortgage statement.
The survey timing matters for interpretation. Both ARM surveys were fielded during or shortly after a historically rapid rate-tightening cycle (2022-2023), creating a period of maximum regret exposure. ARM regret in a falling-rate environment would likely be far lower, since resets would land at or below the initial rate rather than well above it, and fixed-rate holders might then experience their own form of regret for not capturing the lower variable rates. The 43% figure (the more conservative primary estimate) is a rate-cycle-specific snapshot, not a structural truth about the product. What is structural is the uncertainty asymmetry: fixed-rate borrowers know their payment in advance; ARM borrowers do not. That uncertainty is the proximate cause of regret in both surveys, not the product category per se.