Essays on Fiscal Policy

Abstract

This dissertation arises from my studies in the doctoral program "Decision Sciences" at the Graduate School of Decision Sciences and my position as a research assistant at the Chair of Macroeconomics at the University of Konstanz. It comprises three independent chapters with a focus on trade-offs facing fiscal policy in the context of population aging and automation. In the first chapter, "Population aging, inequality and public policy", I examine the implications and quantify the consequences of population aging for the political economy underlying the allocation of public resources between pensions and education, two emblematic government items for the trade-off in the public resource allocation between elderly and young individuals in an economy. I develop a quantitative overlapping generations model with heterogeneous households, calibrated to the German economy in 2020. The model features a government that finances pensions and education investments. Public policy choices are determined through probabilistic voting, ensuring that shifts in the age structure endogenously and gradually affect the allocation of public resources. The central finding is that population aging increases the political power of older voters, leading to a long-run reallocation of public spending: pension expenditures rise at the expense of education investment. This shift has significant macroeconomic consequences. Production drops markedly, total income inequality declines, and wealth inequality increases. The increasing voting power of the elderly keeps the average pension replacement rate — which falls significantly in a counterfactual economy where public policy is independent of the age structure — almost constant over time. Higher individual pensions for retirees reduce incentives for capital accumulation throughout the life cycle, resulting in a smaller aggregate capital stock. Reduced investment in education reduces the average level of human capital and, consequently, the effectiveness of labor. In addition, maintaining the average pension replacement rate at a constant level requires a significant increase in the social security tax rate, as population aging leads to a rising share of retirees. Although the government cuts public education investment, overall tax liabilities for working individuals increase, distorting labor supply, and further reducing effective labor. Both the decline in capital and the reduction in effective labor lead to a long-run drop in production. Total income inequality declines, primarily due to the redistributive effect of pensions among retirees, while wealth inequality rises as higher pensions particularly disincentivize saving for low-income households. In my probabilistic voting setup, the policymaker maximizes a weighted average welfare function of all living individuals, where the weights reflect age- and education-type-specific voter turnout rates. These turnout rates, estimated from survey data for Germany, are positively associated with educational attainment and exhibit a hump-shaped relationship with age. In a counterfactual scenario, I replace German turnout rates with those estimated for Belgium, where compulsory voting is in place. It shifts political power toward younger and less educated individuals, reducing, though not eliminating, the bias toward pension spending and mitigating the macroeconomic consequences. Overall, the results highlight the severe economic consequences posed by the political economy underlying population aging, but also illustrate that an institutional reform can partially counteract the imposed economic challenges. The second chapter, "Fiscal policy and human capital in the race against the machine", is joint work with Daniele Angelini (University of Vienna) and Stefan Niemann (University of Konstanz). We study the trade-offs facing fiscal policy in a dynamic growth model that incorporates automation, educational choice, and human capital formation. We find that when human capital formation can be affected by government spending, fiscal policy can enhance welfare through a coordinated increase in labor and robot taxes. The composition of taxes used to finance transfers and education spending is key in determining their effects on growth and inequality, as the robot tax is more redistributive than the linear labor income tax. We calibrate our model to the US economy and determine the welfare-maximizing tax policy. We find that in future years, the government should initially reduce the robot tax substantially to foster automation-driven growth and compensate for the revenue loss with a higher labor tax. In later periods, the government should progressively raise the robot tax while reducing the labor tax. This dynamic tax pattern initially provides incentives for increased R&D and automation. As machine productivity increases and the skill premium widens over time, the government finds it optimal to increase the robot tax and reduce the labor tax to contain inequality. We also analyse the role of education subsidies. Welfare gains arise when per capita spending on basic education adjusts in response to the policy, and the labor tax finances the subsidies without negatively affecting spending on higher education. Finally, we extend the model to incorporate private investment in higher education. In this setting, changes in public education spending crowd out private contributions, thereby weakening the human capital channel. The importance of this mechanism depends strongly on the underlying funding mix for higher education. In a European setting, with mainly publicly financed tertiary education, optimal financing for the government’s redistribution and education policy involves a positive robot tax. Tertiary-educated individuals, who are typically the owners of capital, do not bear the direct cost of their higher education but contribute indirectly through the robot tax. For the US setting, where tertiary-educated individuals already bear the cost of their higher education, the optimal robot tax is zero, with public spending on redistribution and education financed solely by the labor tax. The third chapter, "Population aging and fiscal multipliers", studies the consequences of population aging for the effectiveness of fiscal policy, as measured by the size of the fiscal multiplier. The fiscal multiplier is a central concept for understanding how government spending translates into economic output, and the literature increasingly recognizes that its magnitude varies across countries and over time, depending on both structural and cyclical conditions. A critical but underexplored structural factor in this context is the age composition of the population. My empirical analysis is based on a structural vector autoregression (SVAR) framework applied to quarterly data for 35 countries over the period 1995–2019. I estimate impulse response functions to a positive government spending shock and derive corresponding fiscal multipliers. My empirical findings are summarized in three stylized facts: (i) government spending shocks generally lead to positive and significant output responses, consistent with much of the existing literature; (ii) these output responses are substantially larger in countries with relatively younger populations; and (iii) there is a significant negative relationship across countries between the size of the fiscal multiplier and the old-age dependency ratio — a key indicator of a population's age structure. To explain the stylized facts, quantify the impact of demographic change on fiscal multipliers, and assess how future population aging may affect the effectiveness of fiscal policy, I develop a medium-scale overlapping generations model with heterogeneous households and endogenous labor supply. The model incorporates idiosyncratic income risk and, importantly, country-specific demographic structures. Fiscal expansions in the form of additional government expenditures are financed via non-distortionary lump-sum taxation. The model, which successfully replicates the stylized facts, is calibrated for the United States in 2020 and extended to seventeen additional countries, allowing for a rich cross-country comparison. Fiscal multipliers in the model are driven primarily by the labor supply response to positive government spending shocks, which in turn depends on savings behaviour and tax sensitivity of the households. In aging societies, households anticipate longer retirement periods and save relatively more over the life cycle, which reduces their sensitivity to transitory wealth shocks such as temporary lump-sum tax increases. This, in turn, weakens the labor supply response and leads to smaller output effects from fiscal expansions. Quantitatively, I find that a one standard deviation increase in the old-age dependency ratio — equivalent to a 6.3 percentage point rise — reduces the fiscal multiplier by 17.7 percent on average. To disentangle demographic effects from other structural differences, I simulate counterfactual scenarios in which each country's age structure is varied while holding other model parameters constant. These within-country exercises show that even after accounting for non-demographic factors, the pure effect of population aging reduces the multiplier for a one standard deviation increase in the old-age dependency ratio by an average of 3.6 percent across countries. To explore the forward-looking implications of these findings, I project the effects of population aging on impact multipliers through 2070. This analysis predicts that population aging will reduce fiscal multipliers by an average of 11.3 percent across the model sample. The results highlight that policymakers aiming to stimulate output through additional government spending need to account more explicitly for the country-specific demographic structure in the design and timing of fiscal interventions.