146. Fighting Poverty With Actual Evidence

Episode Summary

Executive Summary: The episode examines whether large cash windfalls meaningfully improve poor families’ long-term prospects, using evidence from the Georgia land lottery and modern cash-transfer experiments. The discussion with Dean Karlan and Richard Thaler argues that direct cash often improves living conditions and income generation, but impacts on children’s human capital and long-term mobility are mixed, so policy should be driven by rigorous evidence rather than ideology.

Main Topics: Georgia land lottery as a natural experiment (Priority: 5/5): The episode opens with a 19th-century Georgia lottery study showing winners became wealthier, but their children and grandchildren did not show better schooling, literacy, occupations, or wealth in adulthood. Evidence-based poverty policy (Priority: 5/5): Stephen Dubner frames the broader question of how to fight poverty and why evidence, not intuition, should guide policy decisions. Cash transfers versus in-kind aid (Priority: 5/5): Karlan and Thaler debate whether giving cash directly is better than providing targeted assets like goats, fertilizer, or training, emphasizing that effectiveness depends on context. GiveDirectly Kenya experiment (Priority: 5/5): A randomized trial of unconditional cash transfers in Kenya found spending largely on housing and income-generating investments rather than alcohol or gambling, with reduced hunger and better farm outcomes. Behavioral economics and nudges (Priority: 4/5): Thaler explains how small behavioral interventions—like norms-based tax letters or automatic enrollment—can produce large changes at low cost. Measuring outcomes rigorously (Priority: 4/5): The discussion stresses control groups, independent surveyors, list randomization, and administrative data to overcome biases in self-reported behavior. Scaling effective interventions (Priority: 4/5): Audience questions focus on how to move successful pilots into government or NGO adoption, with the panel emphasizing simplicity and clear evidence presentation.

Key Arguments: The Georgia lottery suggests windfalls can raise wealth without improving the next generation’s human capital or mobility. Evidence-based policy requires humility: many policy beliefs are wrong, so data must replace hunches and ideology. Cash transfers can be efficient and often do not lead to obvious wasteful spending, but context matters and in-kind support may sometimes be superior. Household dynamics are a black box; giving aid to women often does not outperform giving to men as much as expected, though sometimes women invest more in children. Self-reported spending is unreliable, so experiments should use control groups, independent enumerators, list randomization, and objective measures when possible. Behavioral nudges can produce substantial policy gains at minimal cost, such as improving tax compliance or retirement saving. Scaling interventions is easier when results are simple, transparent, and easy for policymakers to understand.

Data Points: Innovations for Poverty Action scale: 900+ people in 51 countries - Described as the organization’s current size and global reach. Kenya cash-transfer study sample: about 500 households - GiveDirectly experiment with nearby controls. Unconditional cash transfer amounts: about $300 and about $1,100 - Different grant sizes in the Kenya trial. Grant timing variation: lump sum vs installments - Experimental design to compare payment structures. Treatment effect in UK tax letters: up to 5% increase in take-up rate - Using local social norms and a park-funding message to improve tax compliance. Default enrollment effect: 90% follow along - Automatic enrollment and escalation in savings plans led most participants to accept higher saving rates. Georgia lottery finding: no difference in school attendance, literacy, adult wealth, or occupation across winners’ descendants - Summary of the intergenerational effects from the land lottery study. Behavioral savings study setting: Denmark - Used wealth tax/admin data to test whether 401(k)-style savings increases were new saving.

Pivotal Quotes: "We see a really huge change in the wealth of the individuals, but we don't see any difference in human capital." — Hoyt Blakely: Describing the Georgia land lottery study’s intergenerational results. "We can't do evidence-based policy without evidence." — Richard Thaler: Thaler’s core argument for data-driven poverty policy. "The jury is not yet in on whether giving cash works better than giving stuff." — Stephen Dubner: Closing takeaway that the best anti-poverty strategy remains uncertain.

Implications: The episode suggests cash aid can help, but there is no universal anti-poverty fix. Policymakers should test interventions rigorously, compare cash to in-kind programs, and favor simple, scalable designs backed by evidence.