Job Market Paper
"Using the Alternative Minimum Tax to Estimate the Elasticity of Taxable Income for High Earners"
Abstract: Personal Income tax revenue in the United States draws heavily from high-income taxpayers. How high earners respond to tax changes has repercussions for tax revenue, the efficiency costs of taxation and the optimal progressivity of the tax schedule. Prior research that uses bunching methods to estimate the taxable income response of taxpayers has presented no evidence of high-income bunching at the top kink in the regular, federal income tax schedule. I argue that the regular schedule does not identify the actual tax-related incentives that apply to high-income individuals. At the federal level, high earners are subject to a combination of the regular income tax and the Alternative Minimum Tax. I use annual tax codes and publicly available samples of Internal Revenue Service individual income tax return data from 1993-2011 to characterize the combined schedule for each taxpayer. I discover previously undetected bunching at the top kink in this schedule and use it to estimate the elasticity of taxable income with respect to the net-of-tax rate for high earners to be between 0.15 to 0.28. This estimate implies a lower bound on the optimal top marginal tax rate of 70 percent, suggesting an optimal rate that is higher than prevailing top rates. I also use this setting to make a unique methodological contribution: I show that the location of the top kink in the combined schedule for each taxpayer varies across the distribution of taxable income. This generates novel variation in marginal tax rates that is separable from variation in taxable income, allowing me to mitigate a key endogeneity concern associated with the use of bunching estimators on fixed kink points.
"An Alternative Approach to Property Valuation and Taxation in a Developing Country Context" (with Ali Cheema and in collaboration with the Government of Punjab, Pakistan)
Abstract: Developing economies routinely lag behind developed countries in generating revenue from property taxation, which is a critical source of own-source revenue for financing municipal public goods. This is often due to low tax rates, narrow tax bases, taxpayer non-compliance, lack of administrative capacity, and broad-based homestead exemptions. Understudied in the literature is the effect of the choice of the valuation system on the tax base which can place arbitrary bounds on tax revenue. This paper directly tests the impact of varying tax base valuation systems. We exploit novel, administrative property tax data as well as specially constructed satellite data from the 110-million strong Punjab province in Pakistan to examine two valuation methods: annual rental value (ARV) that is currently in use, and a capital value system which is similar to locationally-adjusted area-based methods. We unpack the effect of switching from one to the other on distributional fairness, revenue potential and buoyancy. Preliminary results indicate that capital value-based valuation methods generate more progressive taxation even under revenue-neutrality; use of satellite imagery for assessing properties unleashes significant tax potential even after retaining basic exemptions; and there exists substantial potential for spatially expanding the tax net when switching to a capital value-based assessment method.
"Multidimensional Measurement of Sectoral Performance: Evidence from Public Schools in Pakistan"
For the most recent version of this working paper, please click here
Abstract: In this paper, I develop a tool for measuring multidimensionally, the state of the public sector in the spirit of multidimensional measures of poverty and apply it to the case of public education in Pakistan. The framework allows fiscally constrained policymakers and relevant development practitioners to measure a sector’s resource base, follow it over time, and optimize targeting of resources. The measure's decompositional properties provide for easy identification of the sources of deprivation along various dimensions and across subgroups, such as geographical areas and subsectors. In an application to the public education sector in Sindh province, Pakistan, I show that 27 percent of public schools are multidimensionally deprived and the weakest dimensions are physical infrastructure and facilities. Single-sex, rural schools, where instruction is in the native Sindhi language contribute the most to the overall deprivation measurement. Such identification permits efficient allocation of policy attention. Targeting public resources to these weak links can generate the biggest bang for the buck. This is especially valuable in resource-constrained, developing countries. The measure allows policymakers to glean critical sectoral information from the din of administrative and survey data.
Work in Progress
Project summary: We research taxpayer responses along the voting margin to changes in tax burdens in the United States. We unpack the idea posited in Thomas Frank’s “What’s the Matter with Kansas?”: that voters increasingly vote on cultural and social issues, rather than their economic self-interest. The project uses an innovative simulated-instruments strategy embedded in a fixed effects model at the county level to identify the impact of taxes on voting patterns in presidential, senate, house, and gubernatorial elections.
"The Impact of Inequality on Federal-State Transfers: The Case of the Federal Medicaid Assistance Percentage (FMAP)"
Project summary: I consider a central piece in the allocation mechanism for tax revenues: intergovernmental transfers. One of the key connective tissues of federal-to-state transfers in the United States is the Federal Medicaid Assistance Percentage (FMAP), which determines the amount of matching funds received by states from the federal government. Statutorily, the FMAP is a function of a state’s per capita income relative to the national per capita income but is silent on the variation around the state’s mean income per capita. Variation in state-level income distributions without commensurate changes in income per capita creates a wedge between actual and potential federal-to-state transfers. Balanced-budget constraints and an increase in the demand for public services driven by rising inequality can plausibly crowd out allocations to other services such as education and public safety. I study the impact of this differential funding on the provision of public services at the state level in the United States.
Project summary: We are conducting an experiment in the Khyber Pakhtunkhwa province of Pakistan to test the impact of institutional reforms and tax policy changes on tax morale, tax compliance, and the quality of public services at the local level. The interventions aims to improve taxpayer perceptions of property tax fairness by making the valuation regime fairer, and to strengthen the tax-benefit linkage. Multiple arms of the experiment will target taxpayer behavior on the intensive margin for those who are currently being taxed and on the extensive margin, related to property owners who currently fall outside of existing property tax areas. This project is being executed in collaboration with the Government of Khyber Pakhtunkhwa and is currently in the design phase.
“Reforming the Property Tax System in Khyber Pakhtunkhwa” (with Ali Cheema). Nov 2020. Consortium for Development Policy Research (CDPR)
“Pilot Study on Effect of Political Influence on Tax Payment Compliance” (with Adnan Khan, Asim Khwaja & Monica Singhal). Mar 2017. International Growth Centre
“Community Engagement Mechanisms: Field Experiment in Pakistan” (with Salman Asim and Mariam Adil). Sep 2015. Washington, D.C.: World Bank Group