Tax Administration, Incentives, and Causal Inference in Developing Economies

This lecture delves into the critical role of tax administration in developing economies, contrasting its impact with traditional tax rate adjustments. The speaker introduces two primary approaches to addressing information challenges in tax structures: designing tax codes for information revelation and improving data collection through tax administration systems. A key focus is on a study conducted in Indonesia, analyzing the effects of a corporate taxpayer administration reform that established special offices for medium-sized taxpayers. The findings reveal a significantly large increase in tax revenue, comparable to substantial tax rate hikes, underscoring the potent impact of administrative improvements.

The lecture then transitions into the methodological challenges of evaluating such reforms, particularly when ideal identification strategies like Regression Discontinuity (RD) are not feasible due to data limitations (e.g., lost assignment formulas). Instead, the speaker details the application of Propensity Score Matching combined with Difference-in-Differences (Diff-in-Diffs). This involves restricting the sample to areas of common support, re-weighting treatment and control groups based on pre-period data to ensure comparability, and then estimating the Diff-in-Diffs model. The discussion highlights the benefits of non-parametric matching methods, such as those by Heinmiller, for their ability to balance characteristics effectively when the exact propensity score functional form is unknown.

Beyond the immediate revenue effects, the lecture explores the underlying mechanisms driving the observed impacts, particularly the concept of size-dependent taxation. It suggests that improved tax administration can alter the relationship between firm size and the probability of enforcement, thereby reducing distortions that might otherwise discourage firm growth. This nuanced understanding helps explain why the revenue gains from administrative reform might persist and even grow over time, rather than diminishing as firms adapt to new enforcement regimes.

Finally, the lecture addresses the controversial topic of high-powered incentives for tax collectors, drawing parallels to historical examples like Roman tax farmers and pre-revolution French tax farmers. Using a randomized experiment in Pakistan focusing on property taxes, the speaker outlines a Nash bargaining model to analyze the trade-off between increased tax collection effort and the risk of collusive evasion between tax collectors and taxpayers. This highlights the complex behavioral responses and ethical considerations involved in designing incentive schemes for public officials in low-information environments, where the potential for corruption is high and taxpayer rights must be balanced against revenue generation goals.

is tax administration um sort of a a really big deal or not

can we compare just a straightforward change in Improvement in Tax Administration to one of these more traditional things like changing the tax rates

the Excel sheets were lost to the recycle bin of time

instead we're going to use propensity matching match differences and differences

you have to restrict you take your sample and you um you want to restrict to the area of which where there's called common support

The Next Step is you use the pre-period data to re-weight the treaty and control group so that the way the distributions look the same

the substantive point is that tax administration can make a really big difference

giving high-powered incentives to task collectors is a very very old idea right in fact it goes back to like the the Roman Empire

if the tax inspector were to come to my house and offer to you know adjust my taxes in return for a cash payment the probability that would end up in detection that you know they would end up going and going to jail is probably reasonably high

if they're potentially corrupt and kind of bargaining with the taxpayer this could just all this could do is kind of change the underlying bargaining game