Behavioral adjustments, income volatility and attention to the tax and benefits schedule

Tax- and benefit systems are distortive: The create incentives to work less. Understanding how people respond to such incentives is crucial to understanding efficiency- and equity effects of tax- and benefit systems. From simple economic theory we expect that individuals will bunch at thresholds where the incentives in tax- and benefits systems change, of which there are many in Norway. Examples include changes in marginal tax rates, reduced costs in childcare or child benefits that depend on income, disability insurance that is phased out, all at thresholds of annual income. Estimating the labor supply response to such incentives has a long tradition in economics, yet the consensus, if there even is one, suggests surprisingly little response. This project aims to understand why. The project has three parts. In the first part, we will do a comprehensive analysis of the variation in labor supply responses to tax- and benefit incentives across a range of benefits, thresholds, groups and years and investigate patterns in these responses to disentangle potential explanations. This is made possible by unique Norwegian registry data and modern microeconometric methods. In the second part of the project, we aim to understand the role of income volatility. With progressive taxes, individuals with variable income are taxed more heavily than individuals with constant income. Income volatility may also be a potential explanation for the limited response to tax incentives. This part of the project aims to understand the role of income volatility for efficiency- and equity effects of tax- and benefits system and its role in shaping responses to taxation. The final part of the project addresses the role of information directly, testing the role for information frictions by exploiting a unique experiment where a subset of disability insurance recipients was mailed direct information on upcoming changes to the incentives to combine work and disability insurance.

This project has three parts. First, we provide the most comprehensive evidence to date on heterogeneity in labor supply responses to tax incentives from a large range of tax incentives, subgroups and time periods. The key to our empirical contribution is the combination of large administrative datasets and unique features of the Norwegian tax and benefits schedules that generate and shift notches and kinks in the budget constraints of individuals. This allows us to estimate labor supply elasticities using state of the art bunching methods, difference in differences and regression kink designs. Motivated by economic theory, this part aims to explore the heterogeneity in our own labor supply elasticities and estimates from the literature by comparing elasticities estimated from different tax incentives, groups and years. Second, we examine the role of income volatility in shaping both the distribution of taxes and the response to tax incentives. Starting with the simple fact that households with unstable earnings pay on average more taxes than households with stable earnings when taxes are progressive, we will build a dynamic theoretical model of labor supply that incorporates earnings volatility and tax policy to understand how taxing at alternative frequencies may affect labor supply, taxes and inequality. This is highly policy relevant: While most countries tax at an annual frequency, some countries are allowing individuals to be taxed based on average earnings over multiple years. Third, we address the role of information directly by exploiting plausibly exogenous variation in the access to information about the size and position of a particular kink in the system for disability insurance, zooming in and providing causal evidence on the importance of information frictions. The key to the success of our proposal is a combination of high-quality administrative data, theory and credible identification strategies building on features of the Norwegian setting.