The Taxation of Microenterprises
How to cite
Overbeck, Daniel. 2023. "The Taxation of Microenterprises." AEA RCT Registry. AEARCTR-0012682. https://doi.org/10.1257/rct.12682-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
The project aims at understanding how SMEs approach taxes and how tax policy can be improved accordingly. In particular, the study tries to elicit explanations for why firms bunch in "dominated regions" above tax schedule discontinuities in Zambia. To this end, a survey (-experiment) on turnover taxpayers will be conducted. Information on the way firms file their taxes, their interaction with the Zambia Revenue Authority, and on how they keep records of business activities will be collected. This may yield relevant behavioral patterns which can help to understand the bunching behavior. In addition, randomly assigned information treatments on audit probabilities as well as tax clearance requirements to secure government contracts will allow to causally identify the determinants of the observed bunching behavior.
Experimental design
The experiment is included in a survey of owners (or employees with financial responsibilities) of firms under turnover tax. Respondents are asked what tax payment (in Zambian Kwacha) the would consider appropriate for their business. Then they are randomly assigned to one of three groups: 1. control group (no relevant information treatment), 2. audit treatment group (information about tax audits of SMEs - aimed at increasing perceived audit probability), 3. government contracts treatment group (information on the volume of business contracts between the Zambian government and SMEs together with the reminder that one needs a tax clearance certificate for such contracts - aimed at increasing the perceived benefit of paying taxes).
The respondents are then confronted with a "notch" i.e. they are being told that their payment cannot be made. Instead they have to choose to either pay 10% more or 10% less than what they have stated.
It is then possible to estimate the effect of the information on the "propensity to bunch above the threshold" i.e. the probability of stating to prefer a high payment.
Design details
Cited by
No citation data available for this registration yet.