The Economic Sentiment Gap: Why Voters Feel Worse Than the Numbers Say

Consumer confidence and economic reality have diverged so thoroughly that they now appear to describe different countries. GDP growth is positive. Unemployment is historically low. Inflation has declined from its 2022 peak. And approximately 62% of Americans describe the economy as "poor" or "not so good." Both the numbers and the sentiment are accurate. Neither tells the whole story.

WHAT HAPPENED

• University of Michigan Consumer Sentiment Index (February 2025): 64.7, below the 50-year average of 85.4
• GDP growth Q4 2024: 2.3% annualized
• Unemployment: 4.1% (historically low by post-1960 standards)
• Real median household income: declined 2.1% from 2019 peak after adjusting for cumulative inflation
• Consumer price index: up 21.8% from January 2021 to January 2025

THE NUMBERS BEHIND THE NUMBER

The disconnect between economic indicators and economic sentiment has a straightforward explanation: macroeconomic indicators measure flows (GDP growth, monthly job creation) while household economic experience is dominated by levels (the price of groceries, the cost of rent, the monthly mortgage payment). A family whose grocery bill rose 25% between 2021 and 2024 does not feel better because the rate of increase slowed from 9% to 2.5%. The bill is still 25% higher.

This is not irrational. It is the difference between acceleration and velocity. Economists celebrate deceleration. Households experience velocity. The two perspectives are both correct and fundamentally incompatible.

THE POLLING IMPLICATION

For electoral forecasting, the sentiment gap creates a modeling challenge. Traditional models use economic indicators (GDP, unemployment, inflation rate) as predictors. But if voters respond to price levels rather than rates of change, the models need recalibration. A 2.3% GDP growth rate that would historically correspond to an incumbent advantage of 1.5–2 points may correspond to zero advantage or a slight disadvantage if voters are responding to cumulative price increases.

Our model currently weights consumer sentiment at 0.30 and traditional economic indicators at 0.25. The gap between them is the widest since the University of Michigan began the survey.

POLLERBULL SIGNAL

• What moves odds: Economic sentiment is a leading indicator for midterm outcomes. The current sentiment level (64.7) has historically corresponded to incumbent-party losses of 25–35 House seats.
• What would falsify this: If sentiment recovers above 75 by mid-2026 despite price levels remaining elevated, voters are adjusting to the new price baseline and the lag effect is shorter than historical norms.