Taxation and Development: Evidence from Pakistani Administrative Tax Records

点击翻译按钮获取中文摘要

Understanding behavioural responses by individuals and firms to tax incentives is essential for the optimal design of tax systems. The empirical analysis of these behavioural responses plays a central role in economic research, but researchers have mainly focused on the behavioural effects of personal income taxation in developed countries due to data restrictions or the lack of exogenous tax variation. In comparison with developed countries, developing countries face enormous difficulties raising a desirable level of revenues, which makes the empirical investigation of the impact of tax incentives in developing countries imperative for the design of tax policies. One particular difficulty that hinders development is the significance of the informal sector. Hence, to achieve sustainable growth, a better understanding of the decisions of firms and small businesses to leave informality seems crucial.This project benefits from unique access to administrative micro data from the Federal Board of Revenue (FBR) for the three main federal taxes in Pakistan: personal income taxation (imposed on individuals and unincorporated firms), corporate income taxation, and value-added taxation (labelled the general sales tax). Access to such exhaustive administrative data is unique even for developed-country research and in combination with compelling identifying tax variation in Pakistan allows for estimating behavioural responses of both individuals and firms, including real effects, tax avoidance and tax evasion. Some responses have not been explored before and we will develop and apply new empirical methods for these. First, our analysis of the personal income tax will exploit the presence of notches-discrete jumps in the average tax rate-creating large and compelling identifying variation. We will provide a structural analysis to estimate the elasticity of taxable income with respect to the marginal tax rate and shed light on the importance of optimization frictions. We will compare these novel elasticity estimates for a developing country with existing estimates for developed countries. Second, our analysis of the corporate income tax will exploit unique tax variation coming from notches as well as an unusual cohort-based tax reform implemented in 2005. This variation allows us to analyse the effect of corporate income taxation on the decision to incorporate and on corporate taxable earnings, akin the elasticity of taxable income for individuals. Finally, our analysis of the value-added tax will exploit the gradual implementation of the system and various tax reforms during the implementation including tax exemptions, zero-rating and changes in tax rates. This variation allows us to analyse the effect of the VAT on the decision to register (formalize) and on taxable value added conditional on registration. Invoice data allows us to analyse evasion and the presence of compliance spill-overs, crucial to the desirability of a VAT system, but previously untested.Even though our research uses exclusively Pakistani tax records, the uniqueness of the data and the compelling tax variation will make the project interesting to academics, policy makers and tax administrators in other developing and developed countries, including the U.K.