Labor supply responses to tax changes are a core issue in public economics, reflected by numerous estimates from different methodological approaches. This dissertation contributes to the literature along two main avenues. First, I demonstrate how ex-ante simulations from a static discrete choice model can be validated ex-post by panel data methods estimating the elasticity of taxable earnings (Chapter 2). Second, I analyze the dynamics of labor supply responses to tax changes over time both with regards to panel data approaches (Chapter 3) and incorporated in a structural labor supply framework (Chapter 4). I use Norwegian register and survey data provided by Statistics Norway and focus on labor supply responses in working hours and wage earnings. In Chapter 2, I demonstrate how the standard elasticity of taxable income (ETI) methodology can be used to validate predictions from a structural labor supply model, by analyzing the Norwegian tax reform in 2006. To facilitate comparison, I use the discrete choice model to simulate the effects of the 2006 tax reform on hours of work, and use predicted income levels to obtain a comparable estimate for income elasticity with respect to the net-of-tax rate. In Chapter 3, I analyze the dynamics of earnings responses to tax changes by exploiting substantial exogenous variation in the two-tier surtax schedule for labor income over a period of 14 years (1995-2008). I adopt the dynamic panel data framework by Holmlund and Söderström (2011), and compare with estimates from the conventional static panel approach by Gruber and Saez (2002). I find that the long run responses to tax rate changes in labor supply are about twice as large as the short run responses due to a strong autoregressive effect in earnings. Supplementary conducted simulations suggest that up to 40 percent of the revenue loss due to tax rate cuts in the surtax schedule can be self- financed in the long run by inducing additional generated income. In Chapter 4, I examine the effect of state dependence to policy change responses over time within the framework of an intertemporal structural discrete choice model for women in couples. I allow for state dependence both in preferences and in the set of job opportunities. This causes sluggish labor supply responses to various changes in the wage rate and the tax schedule. Within the first year of a permanent policy change the simulation results suggest that about 50 percent of the full labor supply response can be expected, whereas 90 percent of the response is reached after about 4 years.
Diese Dissertation trägt zur vorhandenen Literatur anhand zweier Vorgehensweisen bei. Im ersten Schritt zeige ich, wie Simulationen eines statischen, diskreten Wahlmodells anhand von Methoden für Paneldaten validiert werden können (Kapitel 2). Im zweiten Schritt analysiere ich die Dynamik des Arbeitsangebots bei steuerlichen Änderungen über Paneldatenansätze (Kapitel 3), sowie über ein strukturelles Arbeitsangebotsmodell (Kapitel 4). Ich konzentriere mich auf Auswirkungen auf die Arbeitszeit und Arbeitseinkommen von Lohnempfängern, bei Verwendung norwegischer Registrierungs- und Befragungsdaten.
