NYU Wagner Seminar - Fall 2017

All presentations take place from 12:30-2:00PM in the Rudin conference room on the 2nd floor of the Puck Building (295 Lafayette St.). A light lunch will be served.


September 14, 2017

Donald Rubin

John L. Loeb Professor of Statistics

Department of Statistics, Harvard University



Embedding the Analysis of Observational Data for Causal Effects within a Hypothetical Randomized Experiment

Abstract: Consider a statistical analysis that draws causal inferences using an observational data set, inferences that are presented as being valid in the standard frequentist senses; that is an analysis that produces (a) point estimates, which are presented as being approximately unbiased for their estimands, (b) p-values, which are presented as being valid in the sense of rejecting true null hypotheses at the nominal level or less often, and/or (c) confidence intervals, which are presented as having at least their nominal coverage for their estimands.  For the hypothetical validity of these statements (that is, if certain explicit assumptions were true, then the validity of the statements would follow), the analysis must embed the observational study in a hypothetical randomized experiment that created the observed data, or a subset of that data set.  This effort is a multistage effort with thought-provoking tasks, especially in the first stage, which is purely conceptual.  Other stages may often rely on modern computing to implement efficiently, but the first stage demands careful scientific argumentation to make the embedding plausible to thoughtful readers of the proffered statistical analysis.  Otherwise, the resulting analysis is vulnerable to criticism for being simply a presentation of scientifically meaningless arithmetic calculations. In current practice, this perspective is rarely implemented with any rigor, for example, completely eschewing the first stage.  Instead, often analyses appear to be conducted using computer programs run with limited consideration of the assumptions of the methods being used, producing tables of  numbers with recondite interpretations, and presented using jargon, which may be familiar but  also may be scientifically impenetrable.  Somewhat paradoxically, the conceptual tasks, which are usually omitted in publications, often would be the most interesting to consumers of the analyses. These points will be illustrated using the analysis of an observational data set addressing the causal effects of parental smoking on their children’s lung function.  This presentation may appear provocative, but it is intended to encourage applied researchers, especially those  working on problems with policy implications, to focus on important conceptual issues rather than on minor technical ones.


October 5, 2017

Julia Carboni

Assistant Professor, Public Administration and International Affairs

The Maxwell School of Citizenship and Public Affairs, Syracuse University



Systemic Risk in Networked Services: Implications for Governance

Abstract: Government increasingly relies on complex arrangements of contracted providers to implement public policy but does not consider the possibility of systemic risk- the risk the contract system will collapse. Most public management research on contract management focuses on dyadic interactions between funders and contractors while interorganizational network scholarship overlooks the potential for network failure. This study examines systemic risk in complex, networked services funded by government and produced by a mix of public, nonprofit, and for-profit actors. I employ affiliation network methods to understand network level systemic risk and to develop an index ranking the importance of individual actors in maintaining system stability. The empirical context for this study is state funded juvenile justice services over a five year period. Findings indicate that though the state has a well-designed contract management system, it does not consider systemic risk when awarding individual contracts.  Over time, the system created by individual contracts concentrated service production into handful of providers, increasing risks of catastrophic network failure if one of the “big” producers fails due to limited absorptive capacity in the system.  This raises questions about whether and how government funders should attend to systemic risk in contracted services and has implications for governments’ ability to meet statutory obligations to provide services.  It also raises questions about conditions under which networks of providers can withstand shocks such as organizational failure. This study makes several contributions.  First, it addresses an overlooked issue in governance- systemic risk of contracted services.  Though the issue is understudied, potential for systemic failures are present in all cases of networked services. Second, it demonstrates the utility of using secondary affiliation network data to understand complex structures used to implement public policy.


October 12, 2017

Hilary Hoynes

Professor of Public Policy and Economics, Haas Distinguished Chair in Economic Disparities

Goldman School of Public Policy, University of California Berkeley



Local Food Prices, SNAP Purchasing Power, and Child Health

Abstract: The Supplemental Nutrition Assistance Program (SNAP, formerly food stamps) is one of the most important elements of the social safety net. Unlike most other safety net programs, SNAP varies little across states and over time, which creates challenges for quasi-experimental evaluation. Notably, SNAP benefits are fixed across 48 states; but local food prices vary, leading to geographic variation in the real value of SNAP benefits. In this study, we provide the first estimates that leverage variation in the real value of SNAP benefits across markets to examine effects of SNAP on child health. We link panel data on regional food prices to National Health Interview Survey data and use a fixed effects framework to estimate the relationship between local purchasing power of SNAP and children’s health and health care utilization. We find that children in market regions with lower SNAP purchasing power utilize less preventive health care. Lower real SNAP benefits also lead to an increase in school absences. We find no effect on reported health status.


October 19, 2017

Katherine W. Phillips

Paul Calello Professor of Leadership and Ethics

Columbia Business School, Columbia University



The doubleedged sword of diversity: Toward a dual pathway model

Abstract: Williams and O'Reilly (1998) published a seminal review of diversity research that has become a classic resource for researchers and practitioners alike. In the current review, we update the theoretical record by discussing traditional views of, as well as recent develop- ments to, the 3 prominent frameworks used to understand diversity: social categorization, similarity‐attraction, and information and decision‐making. Furthermore, we propose a dual pathway model of diversity's effects that integrates all 3 frameworks. In our model, both positive and detrimental effects of diversity stem from pro- cesses of social categorization. Whereas these processes disrupt group functioning when intergroup bias is activated, when bias is not activated, we propose that social categorization and reduced attraction to dissimilar others will allow individuals in diverse rather than homogeneous groups to focus more on the task, anticipate differences in task‐relevant opinions and perspectives, and engage in more effortful information processing. Finding the balance is key.


October 26, 2017

Andrew Breck and Davin Reed

NYU Wagner Doctoral Students

The Robert F. Wagner Graduate School of Public Service, New York University




The Effects of Gentrification on Original Neighborhood Residents: Evidence from Longitudinal Census Microdata

Authors: Quentin Brummet (U.S. Census Bureau) and Davin Reed (NYU Wagner)

Abstract: During the past two decades, college-educated and high-income households have increasingly located in urban neighborhoods. We use new longitudinal Census microdata to study the effects of this gentrification process on original neighborhood residents. We first develop a local labor markets model of gentrification, which yields two insights. First, gentrification's effect on original residents' observable well-being is approximated by its combined effects on their incomes, housing costs, commuting costs, and neighborhood amenities, regardless of whether they move or stay. Second, its effect on their unobservable well-being is proportional to its effect on out-migration. Uniquely, our data allow us to estimate these effects for tens of thousands of original residents of central city neighborhoods in all major metropolitan areas of the United States. We define gentrification as large increases in the number of college-educated individuals living in a neighborhood and employ various identification strategies to help address endogeneity from selection, omitted variables, and reverse causality. Our results suggest that gentrification may lead to small increases in original resident out-migration but that it has no effect on original resident employment, income, commuting distance, neighborhood quality, or rents paid and positive effects on original resident house values. Aggregating these estimates suggests that on average, gentrification benefits original resident homeowners and has little effect on the observed well-being of original resident renters. However, our results are consistent with well-being losses for original resident out-migrants if unobservable neighborhood preferences are strong.

Effect of the Supplemental Nutrition Assistance Program on Diagnosis of Diet-Related Diseases

Abstract: The Supplemental Nutrition Assistance Program (SNAP) administered near-cash benefits valued at almost $67 billion dollars to over 44 million Americans in 2016. Although SNAP is well studied, little is known about its effect on the prevalence of diet-related diseases, including heart disease, type 2 diabetes, and high blood pressure. I use restricted-use, longitudinal data from the National Center for Health Statistics to provide new evidence of the effects of SNAP on health. These data include multiple years of Medicaid claims linked with survey information for respondents to the nationally representative National Health Interview Survey. The analytic dataset includes household and individual level characteristics, including SNAP participation, self-reported health variables, and claims-level data for diagnoses, procedures, and Medicaid expenditures. I treat participation in SNAP as endogenous using several linear and non-linear instrumental variable estimation techniques, exploiting plausibly exogenous variation in state-level SNAP policies. I present results from each of the estimation techniques and discuss their relative strengths and weaknesses. Overall, results suggest participation in SNAP has no robust discernible effect on likelihood of diagnosis of hypertension, type 2 diabetes, or heart disease. Following from these findings, I consider a number of policy implications and weigh several opportunities to leverage SNAP to improve long-term health outcomes among income-eligible populations.  


November 9, 2017*

Jesse Shapiro

George S. and Nancy B. Parker Professor of Economics

Brown University

*Joint with NYU Behavioral Economics and Public Policy workshop;

*SPECIAL TIME: Thursday 4-5:30, NYU Law School


How are SNAP Benefits Spent? Evidence from a Retail Panel

Abstract: We use a novel retail panel with more than six years of detailed transaction records to study the effect of participation in the Supplemental Nutrition Assistance Program (SNAP) on household spending. We frame our approach using novel administrative data from the state of Rhode Island. The marginal propensity to consume SNAP-eligible food (MPCF) out of SNAP benefits is 0.5 to 0.6. The MPCF out of cash is much smaller. These patterns obtain even for households for whom SNAP benefits are economically equivalent to cash in the sense that benefits do not cover all food spending. We reject the hypothesis that households respect the fungibility of money in a semiparametric framework. A model with mental accounting can match the facts.

November 16, 2017

Brigitte Madrian

Aetna Professor of Public Policy and Corporate Management

John F. Kennedy School of Government, Harvard University



November 30, 2017

Lawrence P. Casalino

Professor of Healthcare Policy and Research

Weill Cornell Medical College, Cornell University