|Title||Three Essays in Labor Economics|
|Year of Publication||2018|
|Number of Pages||152|
|Keywords||Causal effects, College quality, Informal caregiving, Labor economics, Labor market outcomes, Pension plans, Returns to schooling, Social Sciences|
Advances in computing power over the past 30 years have allowed economists to apply econometric techniques to larger and more complex data sets, to identify cause and effect relationships. This is certainly true in the field of labor economics, where labor economists must often evaluate the impact of policy interventions on labor market outcomes, without the benefit of randomized trials–the ideal method for drawing inferences about policy effects (Angrist & Pischke, 2015). Instead, labor economists must draw inferences from observational data, which present various challenges due to the presence of confounding factors (Athey & Imbens, 2017). In order to identify causal effects in the presence of confoundedness, economists use various econometric tools and strategies, such as propensity score matching, instrumental variables, and regression discontinuity design. Propensity score matching and instrumental variables use the Rubin Causal Model (RCM) framework. RCM interprets causal effects as comparisons of potential outcomes: pairs of outcomes for the same individual, conditional on different levels of exposure to a treatment (Imbens & Wooldridge, 2009). Researchers can only observe one of the two outcomes, so they can never directly observe causal effects (Athey & Imbens, 2017). Propensity score matching is a technique that allows the researcher to create the unobserved counterfactual outcome, and to then use the potential outcomes to estimate the causal effect of the treatment. The instrumental variables approach is designed to address the confoundedness created by unobserved factors, such as self-selection. For example, individuals who choose to receive treatment may be different than those who choose not to receive treatment, in a way that the differences are unobservable. These unobserved differences will influence response to the treatment, and thus, will void estimates of causal effects (Imbens & Wooldridge, 2009). In the instrumental variables approach, researchers use instruments–variables that are correlated with the treatment variable but not with the outcome variable, and thus, does not affect the potential outcome. These instruments allow the researcher to control for unobserved factors, like self-selection, and to estimate causal effects. This dissertation uses two of these approaches–propensity score matching and instrumental variables–to identify causal effects on three important current policy issues: the transition away from defined benefit pensions, the returns to education, and the provision of informal care by grandparents.
Copyright - Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works; Last updated - 2018-08-26