97. Lying to Ourselves

Episode Summary

Executive Summary: The segment argues that people are poor predictors of their own future behavior, especially around taxes, gas prices, and voting. Stephen Dubner uses polling examples and behavioral economics to show that survey responses often reflect wishful thinking or social pressure rather than actual actions; the practical lesson is to trust observed behavior over stated intentions.

Main Topics: Self-deception in forecasting personal behavior (Priority: 5/5): Dubner’s central point is that people routinely overstate what they will do in hypothetical future scenarios, such as leaving the country over taxes or changing driving habits after gas price increases. Taxes, the Laffer curve, and labor supply (Priority: 5/5): The discussion uses tax rates as a concrete example: people may say they would leave or work less if taxes rose, but those claims are speculative and often inconsistent with real-world behavior. Polling and survey reliability (Priority: 5/5): The segment questions how much trust should be placed in surveys, since polls depend on strangers accurately predicting future behavior and may be distorted by suggestion, vanity, or forgetfulness. Behavior versus stated intention (Priority: 4/5): Dubner and Levitt emphasize that actual actions are more reliable than verbal claims, applying this to driving habits, voting, and everyday moral behaviors people claim to follow. Political polling and turnout uncertainty (Priority: 4/5): Political polls are especially shaky because even basic turnout intent is hard to measure; people may say they will vote but fail to do so, affecting election forecasts. Media framing of debate and election anxieties (Priority: 3/5): The conversation is framed by the presidential debate and campaign rhetoric, but quickly pivots from partisan claims to a broader behavioral-economics critique of how citizens talk about policy choices.

Key Arguments: People often lie to themselves unintentionally when predicting how they will react to higher taxes, higher gas prices, or other policy changes. Survey answers can be unreliable because respondents are influenced by the question itself and may respond aspirationally rather than truthfully. The Laffer-curve logic is theoretically interesting, but individuals usually cannot accurately predict their own labor-supply response to tax changes. Observed behavior is a better guide than self-reported intentions; what people do matters more than what they say they will do. Political poll results can be distorted because even the simple question of whether someone will vote is hard to answer truthfully and accurately ahead of time. Claims such as 'I’ll move to Canada' or 'I’ll drive less' are often not borne out when checked later.

Data Points: Yahoo Finance poll response: About one-third said they would never leave the U.S. over taxes - Used to illustrate how some respondents reject the idea of tax-driven relocation, while others entertain it. Yahoo Finance poll response: Another one-third would leave the U.S. if taxes rose above 40% - Shows how many people claim they would relocate at a tax level not far beyond current norms. Tax threshold example: 50% - Dubner poses this as a hypothetical tax rate to ask how behavior might change. Tax threshold example: 70% or 80% - Steve Levitt says he would likely work less hard at these higher rates.

Pivotal Quotes: "We lie to ourselves all the time." — Stephen Dubner: Dubner’s central thesis about why people mispredict future behavior in surveys and polls. "The cardinal rule should be don't listen to what people say, right? Watch what they do." — Stephen Dubner: His concluding advice for interpreting survey responses, political promises, and everyday claims. "What every poll relies on is one stranger telling the truth to another stranger about the future." — Stephen Dubner: Explanation of why polling is inherently fragile and often unreliable.

Implications: Listeners should treat survey answers and political pledges cautiously and prioritize real-world behavior when judging forecasts, policy effects, and election outcomes.