Raj Chetty on Economic Mobility

Episode Summary

Executive Summary: Raj Chetty discusses his Nature papers on social capital and mobility, arguing that economic connectedness—especially cross-class friendships—is a strong predictor of upward mobility across U.S. neighborhoods. He explains how Facebook and tax data are combined to measure connectedness and mobility, stresses that the evidence is correlational at the neighborhood level but supported by causal-moving studies, and reflects on policy, segregation, and the limits of empirical economics.

Main Topics: Economic connectedness as a driver of mobility (Priority: 5/5): Chetty’s central finding is that neighborhoods where low- and high-income people interact more tend to produce better long-run outcomes for poor children. The key factor is class-bridging social capital, not just general trust or tight-knit community structure. Measurement using Facebook and administrative data (Priority: 5/5): The discussion explains how Facebook friendship networks are used as a proxy for offline social ties, validated against survey data, and combined with tax-return-based mobility measures to compare neighborhoods. Causality, selection, and neighborhood effects (Priority: 5/5): Roberts presses on whether neighborhood composition reflects causation or sorting. Chetty argues there is both selection and causal effects, citing sibling comparisons, moving studies, and randomized experiments as support. Magnitude and interpretation of the 20% estimate (Priority: 4/5): Chetty says equalizing connectedness between low- and high-income kids would raise low-income kids’ incomes by about 20%, while cautioning that this is not a silver bullet and likely works partly through education and aspiration channels. Policy implications: housing, counseling, and desegregation (Priority: 4/5): The conversation turns to housing vouchers, zoning, mobility counseling, and programs that pair resources with social support. Chetty emphasizes that access alone is often insufficient without guidance and social capital. The fading American Dream paper and measurement sensitivity (Priority: 4/5): They revisit Chetty’s work showing declining absolute upward mobility over time, with debate about inflation, family-size adjustments, and whether the headline figure is closer to 40%, 50%, or 70%. Chetty argues the decline itself is robust. Role of theory versus big-data empiricism (Priority: 3/5): Chetty defends empirical work but argues theory remains essential for generating hypotheses, extrapolating to new settings, and analyzing equilibrium effects that standard treatment-control methods miss.

Key Arguments: Cross-class social ties are the specific type of social capital most strongly associated with upward mobility; general measures like trust or tight-knit communities matter less. The Facebook-based connectedness metric is validated by comparisons with survey data and with close-friend subsets, suggesting it proxies real-world interaction patterns. Neighborhood-level results are not deterministic: place explains only a modest share of outcomes, and many children do well or poorly regardless of neighborhood. Evidence for causality comes from movers studies, sibling comparisons, Moving to Opportunity, and demolition-induced relocations, all showing better outcomes for younger children exposed longer to higher-opportunity places. A 20% income gain for low-income kids is large relative to many social interventions, but it likely works through multiple mechanisms including schooling, aspirations, and college attendance. Policy should not rely only on moving people or raising vouchers; it should pair resources with counseling, navigation help, and community-building interventions. The decline in absolute upward mobility is robust across reasonable assumptions, even if exact levels vary with inflation and family-size adjustments. Theory still matters in big-data economics because it guides hypothesis formation, extrapolation, and thinking about general equilibrium effects.

Data Points: Age range in Facebook data: 25 to 44 - Chetty says Facebook usage is high in this age band, making it a good sample for measuring social connectedness. Facebook coverage: 85% - Approximate share of U.S. people in the chosen age range on Facebook. Sample size: 72 million people - Size of the Facebook-based network data used to build neighborhood-level measures. Friendship links: 21 billion friendships - Total friendship ties in the Facebook dataset. Correlation between economic connectedness and mobility: About 0.7 - Univariate association between connectedness and upward mobility across places. Estimated income gain from closing connectedness gap: About 20% - Predicted improvement for low-income kids if they grew up in neighborhoods with connectedness like high-income kids. Typical low-income adult income example: $30,000 to $36,000 - Roberts translates the 20% gain into a concrete example. Return to education: 7% to 10% per year - Roberts compares the connectedness effect to estimated returns from additional years of schooling. Share of variation due to causal effects at census tract level: About 60% - Chetty states that roughly 60% of neighborhood variation in mobility appears causal, with 40% due to selection. Share due to selection: About 40% - Complement to the causal-effect estimate at census tract level. Current absolute upward mobility (baseline): About 50% - Chetty’s headline estimate for kids born in the current era using baseline assumptions. Historical absolute upward mobility: About 90% - For kids born in the 1940s and 1950s under baseline assumptions. Alternative current estimate: About 40% - If comparing boys only to fathers’ earnings. Alternative adjusted estimate: About 60% - If inflation is assumed to be 1 percentage point lower than CPI. Alternative family-size-adjusted estimate: As high as 70% - Using per-capita or square-root family-size adjustments. Housing voucher assistance: About $1,500/month in Seattle - Value of rental assistance in the housing mobility discussion. Randomized moving program result: 60% vs 15% - Share of treatment vs control families moving to high upward mobility places in the Seattle experiment. Program cost: $2,500 per family - Approximate cost of the mobility counseling intervention. Affordable housing spending: $25 billion per year - U.S. spending on affordable housing programs cited in the discussion.

Pivotal Quotes: "“if you grow up in a community where low and high-income people are interacting more, you are more likely to rise up in the income distribution”" — Raj Chetty: Chetty summarizes the paper’s main correlational finding on economic connectedness and mobility. "“we’re not per se interested or able to really isolate the impacts of online interactions themselves”" — Raj Chetty: He clarifies that Facebook is used as a measurement proxy for offline social ties, not as a claim about social media effects. "“if we close the gap in connectedness between low- and high-income kids ... the incomes of low-income kids would increase by about 20%”" — Raj Chetty: Chetty states the estimated magnitude of the mobility effect tied to connectedness.

Implications: The episode suggests that reducing economic segregation and building cross-class ties may boost mobility, especially when paired with housing and counseling supports. It also underscores that big-data economics can be powerful yet must be interpreted cautiously.