Betsey Stevenson and Justin Wolfers on Happiness, Growth, and the Reinhart-Rogoff Controversy

Episode Summary

Executive Summary: Russ Roberts hosts Betsy Stevenson and Justin Wolfers to reassess the happiness-income literature and the Reinhart-Rogoff debt-growth controversy. They argue Easterlin’s “paradox” is overstated: richer people and richer countries are generally happier, income effects follow a log pattern, and economic growth still matters. They also stress the need for replication, transparency, and humility in economics.

Main Topics: The Easterlin paradox and happiness research (Priority: 5/5): The guests explain Easterlin’s original claims: richer individuals report higher happiness, but richer countries and economic growth supposedly did not raise happiness. They challenge the paradox as a data and inference problem rather than a deep economic truth. Measuring happiness and validating survey data (Priority: 4/5): Roberts presses on the reliability of single-question happiness scales. Stevenson and Wolfers defend them using psychological validation studies, correlations with life events, and consistency across respondents and observers. Income, well-being, and diminishing returns (Priority: 5/5): The discussion shifts to their finding that happiness rises with income, but roughly with the log of income. They emphasize that income buys capabilities, choice, respect, and better daily experiences, not happiness mechanically. Reinhart-Rogoff and the debt-growth debate (Priority: 5/5): They revisit the famous controversy over public debt and growth, distinguishing a spreadsheet error from the larger substantive debate. The main lesson is that high debt correlates with lower growth, but the causal story and any 90% threshold are far from settled. Replication, transparency, and the culture of economics (Priority: 5/5): The guests argue economics needs stronger replication norms, better data/code sharing, and more journal support for replications. They contrast their field’s weaknesses with emerging practices in psychology and discuss pre-specification in experiments. Policy interpretation and humility (Priority: 4/5): They caution against turning empirical correlations into simplistic policy slogans. Growth, redistribution, and debt decisions require nuance, and economists should communicate uncertainty more carefully to the public and policymakers.

Key Arguments: Easterlin’s original ‘paradox’ was not robust: later work across 160 countries finds a strong positive correlation between GDP per capita and life satisfaction. The relevant theoretical test is whether the within-country income-happiness relationship is steeper than the cross-country one; their reanalysis found no evidence for that claim. Happiness survey questions are imperfect but meaningful; they correlate with life events, outside judgments, and broader life conditions in expected ways. Income affects well-being mainly through the opportunities and capabilities it provides—choice, respect, food quality, and control over one’s day—rather than through money itself. The happiness-income relationship appears logarithmic: gains to well-being diminish at higher income levels, but more income still generally increases well-being. There is no evidence of a universal income threshold after which higher income stops mattering for happiness. The Reinhart-Rogoff spreadsheet error changed the magnitude of the results but not the broad finding that high-debt periods are associated with lower growth. The controversial 90% debt-to-GDP threshold looks fragile and should not be treated as a hard policy cliff. Economists should distinguish correlation from causation, especially in macro/debt debates where reverse causality and omitted variables are plausible. Replication and data/code sharing are essential for credibility; journals should create outlets for replication studies and reward original data collection. Empirical economics needs humility because noisy data, researcher degrees of freedom, and publication incentives can all distort conclusions.

Data Points: Countries in cross-country analysis: 160 - Wolfers says their updated happiness work uses data from 160 countries. Correlation between GDP per capita and life satisfaction: 0.8 - Used to argue rich countries are, on average, happier than poor countries. Happiness scale: 0–10 - Common survey format discussed as the standard well-being measure. Alternative happiness scale: 3 categories - Roberts refers to a chart using categories like very happy, fairly happy, and not very happy. Income-happiness functional form: log of income - Stevenson says well-being rises with income in a logarithmic pattern. Debt threshold criticized in public debate: 90% of GDP - Reinhart-Rogoff’s alleged cliff beyond which growth supposedly falls sharply. Timeframe of debt analysis: post-war period / past 200 years / past 50 years - They describe multiple debt-growth datasets from different periods and country groups. First assistant professor salary mentioned by Roberts: under $19,000 - Used as a personal example of how job meaning outweighed income. Volunteer work during graduate-aspiration period: 20 hours/week - Stevenson says she volunteered in spare time after choosing a lower-paying job. News test significance level mentioned in replication discussion: 5% - Wolfers notes that if researchers run enough studies, some will appear significant by chance.

Pivotal Quotes: "Absence of evidence is not evidence of absence." — Justin Wolfers: On Easterlin’s failure to find a cross-country happiness effect with small samples. "There is no evidence that there’s such a threshold where increases in income no longer make you happier." — Betsy Stevenson: On the supposed claim that income stops mattering after some point. "What we do matters. It’s the reason I became an economist." — Betsy Stevenson: On the responsibility economists have when their research influences policy debates.

Implications: Listeners should treat sweeping claims about happiness, debt, and growth with caution. The episode argues for better data, replication, and careful communication, while reinforcing that income, growth, and policy still matter—just in more nuanced ways than headlines suggest.