Working Paper

Assessing the evidence on neighborhood effects from moving to opportunity

Abstract: Trying to learn about neighborhood effects from the Moving to Opportunity (MTO) housing mobility experiment by focusing on its program effects obfuscates the evidence on neighborhood effects from MTO. This paper shows that using Intent-to-Treat (ITT) and Treatment-on-the-Treated (TOT) program effects from MTO to indirectly draw conclusions about neighborhood effects (1) offers no advantage for learning about neighborhood effects over directly estimating neighborhood effects, and (2) answers an ill-posed question as a result of allowing central identifying assumptions to be made implicitly. Focusing attention on directly specifying and estimating models of neighborhood effects, the paper presents empirical evidence that MTO only identifies effects from moves between neighborhoods of low quality. These results have broad implications for the way program effects are used to learn about parameters of other models, and they have not been sufficiently addressed in the literature to interpret the evidence from MTO as a test of Wilson (1987).

JEL Classification: C30; H50; I38; J10;

https://doi.org/10.26509/frbc-wp-201233r

Bibliographic Information

Provider: Federal Reserve Bank of Cleveland

Part of Series: Working Papers (Old Series)

Publication Date: 2014-09-01

Number: 12-33R

Note: This paper replaces earlier versions published as working papers 11-01 and 11-22. This version was posted in November 2012. It was revised in May 2013