[Published] Bridging the gap: Unveiling the potential of Tanzania’s SMMEs through VAT insights, with Amina Ebrahim, Ezekiel Swema, Oswald Haule, and Vincent Leyaro. CESifo Economic Studies (2024) (Technical report - UNU WIDER)
The collection of value added tax (VAT) is essential for achieving domestic revenue objectives, yet VAT gap estimation is rarely performed in developing countries. This study utilizes innovative tax declaration and audit data to estimate VAT misreporting in Tanzania, applying a machine learning approach to predict evasion in unaudited firms and periods. We measure the underreporting component of the compliance gap, quantifying potential revenue losses due to inaccurate reporting. Our findings indicate that firms often avoid excessive audits, leading to increased evasion, with those firms showing the largest VAT gaps. We estimate a 62% VAT gap among small and medium-sized enterprises in Tanzania. Finally, we present a cost-benefit ratio, suggesting that while auditing sectors with large VAT gaps is cost-effective, those contributing significantly to the overall VAT gap should also be prioritized for revenue generation.
Presented at the IIPF Annual Conference 2024*, Wider Development conference 2023, FIT seminar (Tampere) 2023.
[Published] Tax policy design in a hierarchical model with occupational decisions. International Tax Public Finance, 31, 1295–1341 (2024). (FIT WP),
This research examines the impact of occupational choices and tax evasion on the tax administration policy in a hierarchical tax model. The economy has two sectors, wage-earners and self-employment, with evasion only possible in the latter. Incorporating occupational decisions produces a smaller marginal tax rate and a larger budget for the IRS. However, the resources are still insufficient to audit all self-employed, resulting in distortions in occupational choices favoring self-employment. These distortions prevent production efficiency from achieving the optimum level, indicating that the Diamond-Mirrlees theorem is not applicable in this context. Finally, applying differential taxation represents a Pareto improvement, but it results in higher taxes for self-employment.
Presented at Universidad Alberto Hurtado (Chile), IIPF Anual Conference 2021, SECHi Annual Meeting 2021, and LACEA/LAMES 2021.
[Published] Early Impacts of College Aid, with Eugenio Giolito and Julio Cáceres-Delpiano. Economics of Education Review, Elsevier, vol. 63(C), pages 154-166 (2018).
We analyze the impact of an expansion in government aid for higher education in Chile in a sample of elementary and high school students. Using students who had an alternative source of funding as a control group, and administrative records before and after the reform, we present evidence that students are affected in different ways. First, we show that parents of students who ex ante were more likely to be credit restricted are more likely after the reform to report that their child ends up completing college. Second, we find that students in the same groups that increase their college expectations, obtain a score in high-stakes examination that actually qualified them for college aid. Third, we find that lifting future credit restrictions reduces the probability of dropping out of high school.
This paper studies the causal effect of income tax evasion opportunities on the self-employment decision. Two peculiarities of the Chilean scheme enable us to identify this effect. First, in the Chilean tax design, self-employed and wage-earners are levied with equal marginal taxes, eliminating the differential tax effect. We disentangle two channels through an occupational choice model: taxable income and evasion. Second, we exploit a tax reform that exogenously affects agents' evasion decisions. Following a consumption-based approach, we obtain a tax evasion measure, and estimate two behavioral parameters: (i) the evasion elasticity to marginal tax rate equals 1.4; (ii) an increase of 1 percentage point in the evasion opportunity makes being self-employed 6.1 percentage points more likely. The evasion opportunity is a crucial determinant of the self-employment response to the policy change, mainly driven by agents' behavior near the first income bracket. While women-headed households' evasion behavior is less sensitive to a tax change, having higher education primarily drives this behavior.
Presented at the RIDGE Labor Workshop 2021*, SECHi Annual Meeting 2022*, Pontificia Universidad Católica de Chile, LACEA/LAMES 2021, Universitat de les Illes Balears, RIDGE Workshop on Public Economics 2022, IIPF Anual Conference 2022, Helsinki GSE, FIT Workshop 2022, SECHi Annual Meeting 2023* , Economics of Informality Conference 2022*, LACEA/LAMES 2023*, and M-NEW Annual Workshop*.
*Presented by Romina
Redistributive policies are crucial in tax design but are often overlooked when examining optimal tax administrative policies to fight tax evasion. This paper extends the Keen and Slemrod (2017) framework to analyse how redistributive concerns and inequality aversion affect tax administration. Optimal tax enforcement strategies depend on the connection between private concealment costs and inequality aversion. With social marginal welfare weights falling in income, society chooses less strict enforcement and a larger implied optimal tax gap if concealment costs are relatively high among low-income individuals. Income-dependent enforcement policies also need to account for the impact of these policies on the selection of income-earners to different regimes. When income is taxed in a non-linear fashion, the optimal income tax includes additional considerations arising from concealment costs.
Presented at: IIPF Annual Conference 2025*, Coe-FIT workshop 2025
*Presented by Jukka
[Work in Progress] Welfare Effects of Income Tax Reform and Tax Evasion: Evidence from Chile, with Romina Safojan
Different tax evasion possibilities in occupations affect agents' employment decisions with welfare implications, being critical elements for accurately measure the welfare consequences of tax changes. In the context of progressive piecewise tax schemes and tax evasion, this paper studies the welfare effects of tax changes incorporating those elements previously omitted in the literature. We model an economy with occupational decisions and a piecewise progressive tax scheme, where self-employed can bunch at the income bracket threshold declaring less income to face a smaller marginal tax rate than they owe. Later, we characterize the welfare effect of tax change, allowing for evasion, occupational decisions, and introducing frictions originating from the tax system characteristics, particularly bunching. Those elements produce a divergence with the elasticity of taxable income or total earned income. We propose a new metric to estimate the marginal change in welfare based on a trapezoid, complementing the traditional Harberger's triangle measure. Considering a tax reform in Chile, we estimate the welfare change of this reform, demonstrating the relevance of incorporating occupational decisions and tax evasion for welfare measurement.
Presented at the IIPF Annual Conference 2024, Brown Bag University. of Chile (Econ), Alberto Hurtado University.
*Presented by Romina
Modern tax design and enforcement rely on classification data reported by taxpayers. Data improve the targeting of tax policy and audits but also open avenues for evasion. How can tax authorities use such data while accounting for their potential strategic distortion? We demonstrate that restoring deterrence generates an endogenous global verification frontier: every report below the frontier is verified through either a profit audit or data verification. Optimal hybrid tax enforcement requires selecting which data classes are subject to profit audits up to the global frontier and which are profit-audited only up to lower class-specific cutoffs, with data verification taking over beyond these cutoffs. This costly data verification requirement limits the benefit from differentiated tax enforcement. Optimal implementation is characterized by profit bunching but involves no data bunching.
Presented at the IIPF Annual Conference 2024, University of Chile (DCS).
*Presented by Saara or Eduardo
[Work in Progress] Screening, Not Revenue: The Shape of Optimal Auditing
Forty years of enforcement theory prescribes a cutoff rule: audit every taxpayer below a productivity threshold with certainty, and nobody above it. I show that this prescription is optimal if and only if three conditions hold jointly — the penalty is levied on unreported income rather than on evaded tax, labour supply is exogenous, and the cost per audit is constant — and that each of these conditions severs a channel through which an audit has consequences beyond revealing information. In a model where self-employed taxpayers choose both their sector and their hours, and a revenue-maximising authority commits to an incentive-compatible audit schedule, the optimal schedule is interior and declining in productivity. The reason is that auditing never pays on the taxpayer being audited: its direct revenue effect is weakly negative at every type. Audits pay instead by compressing the information rents of all higher types, which makes enforcement a screening instrument rather than a revenue instrument. All three cutoff conditions fail empirically, so the cutoff is not merely knife-edge but unattainable.
[Work in Progress] Prices, Incentives, and Attention in Retirement Saving: Evidence from Chile’s voluntary pension system
Chile’s voluntary pension scheme asks savers to choose between two subsidies for the same contribution: a flat 15 percent state match, or a deduction from taxable income at the saver’s own marginal rate. Because the income-tax base is narrow, almost no formal worker faces a marginal rate above 15 percent — 99.4 percent of taxpaying worker-years fall below it — so the match is the better instrument for virtually everyone. A substantial share of savers nonetheless choose the deduction. Using administrative pension records linked to a household panel, this paper documents that choice and bounds the implicit value a saver would need to place on the deduction to justify it. The bound is sharpest among savers who owe no income tax at all, for whom the deduction is worth nothing, and it rests on no identifying assumption. Movements between the two regimes are three to four times more common in one direction than the other — the reverse of what attention allocated by expected gain predicts — which locates the friction in how the subsidies are delivered rather than in how savers evaluate them. When a subsidy is paid in arrears, moving the payment forward is worth more than raising its rate.
[Work in Progress] Institutions and the Misallocation Cost of Weak Enforcement
Tax enforcement is delegated to an agency operating under a revenue mandate, and that mandate, not the delegation itself, is what costs welfare. I show that a benevolent agency facing the same institutional separation attains the joint optimum exactly, so the entire welfare cost of the arrangement is attributable to the agency’s objective rather than to the separation of instruments. A revenue-mandated agency over-audits every type, and the burden of that over-auditing falls on the least productive self-employed. Calibrating the model to US self-employment, I then ask whether such an agency is under-funded. The answer is two-sided: it is under-funded relative to its own revenue mandate at every degree of redistributive taste, but relative to welfare it is over-funded below a threshold and under-funded above it. The cost of weak enforcement is overwhelmingly misallocation rather than foregone revenue, since the occupational distortion is roughly two hundred times the institutional wedge. Audit design is therefore nearly irrelevant at current enforcement intensity and matters a great deal above it.