The Moving to Opportunity (MTO) residential-mobility experiment found little effect of moving from high-poverty to low-poverty neighborhoods on the economic outcomes of low-income adults, while observational studies using surveys such as the Panel Study of Income Dynamics report robust associations between neighborhood disadvantage and adult economic outcomes. A within-study comparison by Harding and colleagues ruled out many candidate explanations for this divergence, leaving two surviving accounts: selection on unobservables inflating nonexperimental estimates, and unusually disruptive features of constrained MTO voucher moves, concentrated among adults who were employed at baseline. The open problem is to determine whether neighborhood environments causally affect adults' economic outcomes and which of the two surviving explanations accounts for the experimental–nonexperimental discrepancy.

Background

A long observational literature in sociology, economics, and public health reports robust associations between living in a disadvantaged neighborhood and residents' economic outcomes, but the causal interpretation of these associations has been persistently contested. The key identification problem is selection: families are not randomly assigned to neighborhoods, and unmeasured personal characteristics (skills, motivation, preferences, health, social networks) may simultaneously place people into poorer neighborhoods and depress their later earnings, biasing observational estimates.

The U.S. Department of Housing and Urban Development's Moving to Opportunity (MTO) experiment, launched in 1994 in five cities, offered families in high-poverty public housing randomized vouchers: an "experimental" voucher usable only in low-poverty census tracts (a census tract being a small, fairly homogeneous statistical subdivision of a county), a traditional unrestricted Section 8 housing voucher (the standard federal rental-assistance voucher, usable in any participating unit), or no assistance (control). Randomization makes comparisons of intent-to-treat effects (effects of being offered the voucher) credible. Kling, Liebman, and Katz (2007) showed that MTO produced large improvements in adults' mental health and safety but essentially no gains in adult employment, earnings, or welfare receipt — a stark contradiction with the observational literature. Ludwig and coauthors (2012) later confirmed, 10–15 years out, substantial gains in adult subjective well-being but not in economic self-sufficiency. In contrast, moves to better neighborhoods do improve children's long-run outcomes (Chetty, Hendren, and Katz 2016; Chyn 2018, using public-housing demolitions in Chicago), so the puzzle is specific to adults' economic outcomes.

Several quasi-experimental studies of adults, which exploit assignment-like variation in neighborhood placement, largely side with the experimental nulls: Oreopoulos (2003) found no effects on adult earnings from quasi-random assignment to Toronto public housing projects of different quality, and Jacob and Ludwig (2012), using a Chicago housing-voucher wait-list lottery, found no evidence that neighborhood quality or residential stability affects adults' labor supply. Damm (2014) contributed a related quasi-random neighborhood assignment design for immigrants in Denmark. However, Aliprantis and Richter (2020) re-analyzed MTO with an instrumental-variables strategy (using the randomized voucher offer as a source of variation that shifts neighborhoods but has no direct effect on economic outcomes, so that its association with outcomes reveals the neighborhood effect) that models neighborhood choice and estimated local average treatment effects (LATEs — effects for the subpopulation whose neighborhood was changed by the instrument) of moving between the first and second deciles of national neighborhood quality; these LATEs are large but pertain to only about nine percent of program participants, suggesting that part of the disagreement may stem from how the "treatment" (neighborhood change) is parameterized. Bergman and coauthors (2024) showed in the Creating Moves to Opportunity experiment in Seattle that customized search assistance dramatically raises low-income families' moves to high-opportunity areas, indicating that constrained, unsupported housing searches — exactly the kind of friction MTO movers faced — are a mutable policy feature rather than a preference.

Against this backdrop, Harding, Sanbonmatsu, Duncan, Gennetian, Katz, Kessler, Kling, Sciandra, and Ludwig (2021) conducted a within-study comparison (WSC: a design in which a nonexperimental estimator is benchmarked against a randomized experiment in the same target population) between MTO's experimental estimates and nonexperimental estimates computed in parallel on MTO and on the Panel Study of Income Dynamics (PSID), a long-running U.S. household panel survey. Nonexperimental estimates from the two data sources were strikingly similar to each other and different from the experimental estimates. The study found no clear evidence that the divergence is explained by measurement differences, sample composition, duration of adult exposure to disadvantaged neighborhoods, nonlinearities in neighborhood effects, the magnitude of neighborhood change, frequency of moves, observed treatment-effect heterogeneity, or the choice of neighborhood measure — while acknowledging limited statistical power. Two explanations survive: (i) selection on unobservables, whereby nonexperimental estimates of neighborhood effects on adult economic outcomes are upward-biased; and (ii) MTO voucher moves being unusually disruptive — program time pressure and locational constraints potentially offsetting neighborhood gains — with suggestive evidence that negative experimental effects are concentrated among adults employed at baseline. The authors call for data collection targeting treatment-selection variables, and for natural experiments (program lotteries, public-housing demolitions, displacement events) that generate exogenous variation in neighborhood conditions, to adjudicate. Methodological tools for quantifying unobservable selection have since advanced — coefficient-stability bounds, which use how much an estimate moves when observed controls are added to bound how strong selection on unobservables must be (Oster 2019), and bounding omitted-variable bias with auxiliary proxy data (Hwang 2026) — but have not been applied to close this gap, and the WSC methodology literature (Steiner and Wong 2018) has produced no new neighborhood WSC for adult economic outcomes.

Problem Statement

Determine whether neighborhood environments causally affect the economic outcomes of low-income adults, by adjudicating between the two surviving explanations for the divergence between experimental and nonexperimental estimates identified by Harding et al. (2021): (1) selection on unobservables, under which observational estimates of neighborhood effects on adult economic outcomes are upward-biased; and (2) unusually disruptive effects of constrained voucher moves in MTO, concentrated among adults employed at baseline. Resolving this requires an identification strategy that isolates the causal effect of adult neighborhood conditions on economic outcomes and discriminates between these explanations, and the explanation offered must also be consistent with the observed pattern that experimental and observational estimates diverge for adult economic outcomes but not for adult health or child outcomes, including the concentration of the divergence among adults employed at baseline. The identification strategy must define and defend the target estimand it isolates — the treatment contrast (how the neighborhood treatment is parameterized) and the target population it speaks to. The unifying objective is the resolution of the experimental–nonexperimental discrepancy itself: determining whether selection bias, move disruption, or both account for the gap.

References

  1. Harding, David J., Sanbonmatsu, Lisa, Duncan, Greg J., Gennetian, Lisa A., Katz, Lawrence F., Kessler, Ronald C., Kling, Jeffrey R., Sciandra, Matthew, and Ludwig, Jens (2021). Evaluating Contradictory Experimental and Nonexperimental Estimates of Neighborhood Effects on Economic Outcomes for Adults. Housing Policy Debate 33(2):453–486, print issue March 2023, first published online 21 April 2021. doi:10.1080/10511482.2021.1881985, https://doi.org/10.1080/10511482.2021.1881985 (author manuscript: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10281691/)
  2. Kling, Jeffrey R., Liebman, Jeffrey B., and Katz, Lawrence F. (2007). Experimental Analysis of Neighborhood Effects. Econometrica 75(1):83–119. doi:10.1111/j.1468-0262.2007.00733.x, https://doi.org/10.1111/j.1468-0262.2007.00733.x
  3. Oreopoulos, Philip (2003). The Long-Run Consequences of Living in a Poor Neighborhood. The Quarterly Journal of Economics 118(4):1533–1575. doi:10.1162/003355303322552865, https://doi.org/10.1162/003355303322552865
  4. Jacob, Brian A., and Ludwig, Jens (2012). The Effects of Housing Assistance on Labor Supply: Evidence from a Voucher Lottery. American Economic Review 102(1):272–304. doi:10.1257/aer.102.1.272, https://doi.org/10.1257/aer.102.1.272
  5. Ludwig, Jens, Duncan, Greg J., Gennetian, Lisa A., Katz, Lawrence F., Kessler, Ronald C., Kling, Jeffrey R., and Sanbonmatsu, Lisa (2012). Neighborhood Effects on the Long-Term Well-Being of Low-Income Adults. Science 337(6101):1505–1510. doi:10.1126/science.1224648, https://doi.org/10.1126/science.1224648
  6. Chetty, Raj, Hendren, Nathaniel, and Katz, Lawrence F. (2016). The Effects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment. American Economic Review 106(4):855–902. doi:10.1257/aer.20150572, https://doi.org/10.1257/aer.20150572
  7. Chyn, Eric (2018). Moved to Opportunity: The Long-Run Effects of Public Housing Demolition on Children. American Economic Review 108(10):3028–3056. doi:10.1257/aer.20161352, https://doi.org/10.1257/aer.20161352
  8. Aliprantis, Dionissi, and Richter, Francisca G.-C. (2020). Evidence of Neighborhood Effects from Moving to Opportunity: LATEs of Neighborhood Quality. The Review of Economics and Statistics 102(4):633–647. doi:10.1162/rest_a_00933, https://doi.org/10.1162/rest_a_00933
  9. Bergman, Peter, Chetty, Raj, DeLuca, Stefanie, Hendren, Nathaniel, Katz, Lawrence F., and Palmer, Christopher (2024). Creating Moves to Opportunity: Experimental Evidence on Barriers to Neighborhood Choice. American Economic Review 114(5):1281–1337. doi:10.1257/aer.20200407, https://doi.org/10.1257/aer.20200407
  10. Damm, Anna Piil (2014). Neighborhood Quality and Labor Market Outcomes: Evidence from Quasi-Random Neighborhood Assignment of Immigrants. Journal of Urban Economics 79:139–166. doi:10.1016/j.jue.2013.08.004, https://doi.org/10.1016/j.jue.2013.08.004
  11. Oster, Emily (2019). Unobservable Selection and Coefficient Stability: Theory and Evidence. Journal of Business & Economic Statistics 37(2):187–204. doi:10.1080/07350015.2016.1227711, https://doi.org/10.1080/07350015.2016.1227711
  12. Hwang, Yujung (2026). Bounding Omitted Variable Bias Using Auxiliary Data: With an Application to Estimate Neighborhood Effects. Journal of Business & Economic Statistics (online first 6 March 2026, pp. 1–14). doi:10.1080/07350015.2025.2606063, https://doi.org/10.1080/07350015.2025.2606063
  13. Steiner, Peter M., and Wong, Vivian C. (2018). Assessing Correspondence Between Experimental and Nonexperimental Estimates in Within-Study Comparisons. Evaluation Review 42(2):214–247. doi:10.1177/0193841X18773807, https://doi.org/10.1177/0193841X18773807
  14. Leventhal, Tama, and Dupéré, Veronique (2019). Neighborhood Effects on Children's Development in Experimental and Nonexperimental Research. Annual Review of Developmental Psychology 1:149–176. doi:10.1146/annurev-devpsych-121318-085221, https://doi.org/10.1146/annurev-devpsych-121318-085221
  15. Alvarado, Steven Elías, and Cooperstock, Alexandra (2023). The Echo of Neighborhood Disadvantage: Multigenerational Contextual Hardship and Adult Income for Whites, Blacks, and Latinos. City & Community 23(1):47–80, print issue March 2024, first published online 22 June 2023. doi:10.1177/15356841231179436, https://doi.org/10.1177/15356841231179436
  16. Hasager, Linea, and Jørgensen, Mia (2026). Sick of Your Poor Neighborhood? Quasi-Experimental Evidence on Neighborhood Effects on Health. Journal of Public Economics 256:105599. doi:10.1016/j.jpubeco.2026.105599, https://doi.org/10.1016/j.jpubeco.2026.105599