Policy Evaluation with Multiple Instrumental Variables

Mogstad, Magne; Torgovitsky, Alexander; Walters, Christopher

doi:10.3386/w27546

Cited by 13 publications

(8 citation statements)

References 50 publications

(58 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While our results provide theoretical and empirical guidance for researchers who wish to use 2SLS to combine multiple IVs, it is important to recognize that positivelyweighted averages of LATEs do not necessarily answer interesting scientific or policy counterfactuals. This point motivates our companion paper (Mogstad et al, 2020), in which we develop a framework for aggregating multiple IVs to conduct inference about specific target parameters that do answer well-posed counterfactual questions. The framework generalizes the approach of to replace IA monotonic-ity with partial monotonicity.…”

Section: Discussionmentioning

confidence: 86%

“…of Assumption IAM (or the equivalent ( 11)) is that the MTE schedule is invariant to which instrument is used to calculate these derivatives. In contrast, Assumption PM allows for the possibility that each instrument is associated with a different unobserved cost of treatment and its own MTE function (Mogstad et al, 2020). Comparing MTEs generated by different instruments therefore provides a test of Assumption IAM.…”

Section: Assessing Ia Monotonicitymentioning

confidence: 99%

“…This observation is based onMogstad et al (2020), who develop MTE methods for multiple IVs that enable testing Assumption IAM and aggregating information across instruments under Assumption PM.…”

mentioning

confidence: 99%

See 2 more Smart Citations

The Causal Interpretation of Two-Stage Least Squares with Multiple Instrumental Variables

Mogstad

Torgovitsky

Walters

2021

American Economic Review

Self Cite

View full text Add to dashboard Cite

Empirical researchers often combine multiple instrumental variables (IVs) for a single treatment using two-stage least squares (2SLS). When treatment effects are heterogeneous, a common justification for including multiple IVs is that the 2SLS estimand can be given a causal interpretation as a positively weighted average of local average treatment effects (LATEs). This justification requires the well-known monotonicity condition. However, we show that with more than one instrument, this condition can only be satisfied if choice behavior is effectively homogeneous. Based on this finding, we consider the use of multiple IVs under a weaker, partial monotonicity condition. We characterize empirically verifiable sufficient and necessary conditions for the 2SLS estimand to be a positively weighted average of LATEs under partial monotonicity. We apply these results to an empirical analysis of the returns to college with multiple instruments. We show that the standard monotonicity condition is at odds with the data. Nevertheless, our empirical checks reveal that the 2SLS estimate retains a causal interpretation as a positively weighted average of the effects of college attendance among complier groups. (JEL C26, I23, I26, J24, J31, R23)

show abstract

Section: Discussionmentioning

confidence: 86%

Section: Assessing Ia Monotonicitymentioning

confidence: 99%

See 1 more Smart Citation

The Causal Interpretation of Two-Stage Least Squares with Multiple Instrumental Variables

Mogstad

Torgovitsky

Walters

2021

American Economic Review

Self Cite

View full text Add to dashboard Cite

show abstract

“…The main parameter of interest in the framework, namely the marginal treatment effects, can be identified by the method of local instrumental variables and can later be used for predicting the effects of hypothetical policies. Recent works have proposed new approaches to its identification and estimation (Carneiro and Lee, 2009;Brinch et al, 2017;Mogstad et al, 2018Mogstad et al, , 2020Sasaki and Ura, 2021) and to apply MTE framework to various research topics such as unconditional quantile effects (Martínez-Iriarte and Sun, 2020) and external validity (Kowalski, 2018). Among these, our paper is most closely related to Sasaki and Ura (2020), which also ap-plies MTE to statistical decision rules.…”

Section: Connection To the Literaturementioning

confidence: 94%

Personalized Subsidy Rules

Chen¹,

Xie²

2022

Preprint

View full text Add to dashboard Cite

Subsidies are commonly used to encourage behaviors that can lead to short-term or long-term benefits. Typical examples include subsidized job training programs and provisions of preventive health products, in which both the behavioral responses and the associated gains can exhibit heterogeneity. This paper adopts the marginal treatment effect (MTE) framework to study personalized assignments of subsidies based on individual characteristics. First, we derive the optimality condition for a subsidy rule to be welfare-maximizing by showing that the welfare of an assignment rule can be represented as a functional of the MTE. Next, we show that assigning subsidies generally achieves higher welfare than directly mandating the encouraged behavior since subsidy rules implicitly target individuals via the unobserved heterogeneity in the behavioral response. When there is positive selection, that is, when individuals with higher returns are more likely to select the encouraged behavior, we show that the optimal subsidy rule can achieve the first-best welfare, the optimal welfare if the policy-maker can observe individuals' private information. We then provide methods to (partially) identify the optimal subsidy rule both when the MTE is identified and when it is not. In particular, positive selection allows for the point identification of the optimal subsidy rule even when the MTE curve is not point identified. As an empirical application, we study the optimal wage subsidy by applying our method to the experimental data from the Jordan New Opportunities for Women pilot study.

show abstract

“…Multiple dimensions of unobserved heterogeneity arise naturally in instrumental variable models with ordered and unordered treatments (e.g. Heckman and Vytlacil, 2007;Kirkeboen et al, 2016;Heckman and Pinto, 2018;Lee and Salanié, 2018;Mountjoy, 2021), as well as in settings with multiple instruments (Mogstad et al, 2020). While related, our multidimensional selection model is distinctly tailored to survey settings.…”

Section: Introductionmentioning

confidence: 99%

Selection in Surveys: Using Randomized Incentives to Detect and Account for Nonresponse Bias

Dutz¹,

Huitfeldt²,

Lacouture³

et al. 2021

Self Cite

View full text Add to dashboard Cite

We evaluate how nonresponse affects conclusions drawn from survey data and consider how researchers can reliably test and correct for nonresponse bias. To do so, we examine a survey on labor market conditions during the COVID-19 pandemic that used randomly assigned financial incentives to encourage participation. We link the survey data to administrative data sources, allowing us to observe a ground truth for participants and nonparticipants. We find evidence of large nonresponse bias, even after correcting for observable differences between participants and nonparticipants. We apply a range of existing methods that account for nonresponse bias due to unobserved differences, including worst-case bounds, bounds that incorporate monotonicity assumptions, and approaches based on parametric and nonparametric selection models. These methods produce bounds (or point estimates) that are either too wide to be useful or far from the ground truth. We show how these shortcomings can be addressed by modeling how nonparticipation can be both active (declining to participate) and passive (not seeing the survey invitation). The model makes use of variation from the randomly assigned financial incentives, as well as the timing of reminder emails. Applying the model to our data produces bounds (or point estimates) that are narrower and closer to the ground truth than the other methods.

show abstract

Policy Evaluation with Multiple Instrumental Variables

Cited by 13 publications

References 50 publications

The Causal Interpretation of Two-Stage Least Squares with Multiple Instrumental Variables

The Causal Interpretation of Two-Stage Least Squares with Multiple Instrumental Variables

Personalized Subsidy Rules

Selection in Surveys: Using Randomized Incentives to Detect and Account for Nonresponse Bias

Contact Info

Product

Resources

About