Back to History

Fields Changed

Trial
Field Before After
Trial Title The Origins of Tax Compliance and Accountable Governance The Origins of Tax Compliance and State Capacity
Abstract What are the barriers to raising revenues in developing countries? Governments struggle to build tax bases—with the goal of reducing reliance on foreign aid and natural resource sales—among citizens who operate predominantly within the formal sector and feel underserved in access to fundamental public goods like property rights. Recent research has documented the difficulty of encouraging citizens in these contexts to enter the formal sector and regularly pay taxes. Individuals and firms do appear responsive, on the margin, to changes in the likelihood of punishment for non-compliance or the lowering of administrative barriers to formal participation and tax compliance. However, we argue that there are important understudied determinants of formalization and tax compliance, namely the costs of obtaining formal recognition, the method of tax collection, and the amount of tax liability. We plan to test these hypotheses through a field experiment to be undertaken in Kananga, Democratic Republic of Congo, between August 2017 and December 2018 by randomizing (1) subsidized access to formal land titles, (2) direct vs. indirect tax collection, and (3) discounted tax rates. What are the barriers to raising revenues in developing countries? With the goal of reducing reliance on foreign aid and natural resources, many governments seek to build tax bases among citizens who operate predominantly within the informal sector. Recent research has documented the difficulty of encouraging citizens in these contexts to enter the formal sector and regularly pay taxes. This study explores three understudied determinants of low tax compliance. First, low-capacity states are typically weak at protecting property rights. Second, inefficient methods of tax collection may constrain the effectiveness of enforcement. Third, when citizens face high liquidity constrains, compliance may be elastic to the tax burden. We will assess the effect of these factors on tax compliance with a randomized field experiment that implements interventions affecting each of these margins through a collaboration with the Provincial Government of Kasaï Central in Kananga, Democratic Republic of the Congo, from August 2017 to December 2018. We will vary (1) subsidized access to formal land titles, (2) the method of tax collection---comparing centralized and local collection---and (3) the property tax rates faced by households. This Pre-Analysis Plan presents the research design and introduces the hypotheses pertaining to the land titling intervention. It will be updated after a pilot of the tax intervention in early 2018 to include hypotheses pertaining to taxation before endline data collection.
Trial End Date September 22, 2018 December 1, 2018
Last Published August 01, 2017 04:04 PM March 14, 2018 10:01 PM
Intervention End Date September 22, 2017 November 1, 2018
Primary Outcomes (End Points) The main outcome is tax compliance, i.e., whether an individual has payed the property tax. Other outcomes include 1) rental tax payment, 2) bribe payment to tax collectors, 3) people's view on the local government, 4) views on traditional authorities, 5) civic engagement, 6) access to public goods, 7) investment in land plots, 8) level of confidence in security of property rights, and 9) formalization. The main outcome is tax compliance, i.e., whether an individual has payed the property tax. Other outcomes include 1) rental tax payment, 2) bribe payment to tax collectors, 3) people's view on the local government, 4) views on traditional authorities, 5) civic engagement, 6) access to public goods, 7) investment in land plots, 8) level of confidence in security of property rights, 9) formalization, and 10) social networks.
Experimental Design (Public) The unit of randomization for these taxation treatments is the polygon. There are 361 polygons in the map of Kananga, which will be randomly assigned to the direct tax collection arm, in which tax ministry collectors will be charged with property tax collection; the indirect tax collection arm, in which local avenue chiefs will be charged with this task; or the control arm, in which citizens are expected to visit a local bank to pay the taxes due on their property. In contrast to these three treatment arms, which vary on the polygon level, the subsidized access to formal land titling and tax rate coupons will be cross-randomized at the household level. We will first administer a baseline survey at selected households, and households will be randomly selected to a land titling intervention group within each polygon. Individual households selected to receive the land titling intervention will receive on-site appraisals and encouragements to obtain formal titles. Treatment assignment for the discounted tax rate intervention will occur during baseline survey visits — randomly selected households will receive tax coupons based on algorithm implemented within the surveying software on tablets used by enumerators. To access the tax rate discounts, selected individuals will need to present their coupon to avoid paying the full rate. Coupons will be individualized using the name and identification code written on the property owner’s state identity card. Failure to present the coupon and/or the accompanying identity card disqualifies households from the discount The unit of randomization for these taxation treatments is the polygon. There are 361 polygons in the map of Kananga, which will be randomly assigned to the direct tax collection arm, in which tax ministry collectors will be charged with property tax collection; the indirect tax collection arm, in which local avenue chiefs will be charged with this task; or the control arm, in which citizens are expected to visit a local bank to pay the taxes due on their property. In contrast to these three treatment arms, which vary on the polygon level, the subsidized access to formal land titling and tax rate coupons will be cross-randomized at the household level. We will first administer a baseline survey at selected households, and households will be randomly selected to a land titling intervention group within each polygon. Individual households selected to receive the land titling intervention will receive on-site appraisals and encouragements to obtain formal titles. Treatment assignment for the discounted tax rate intervention will occur during baseline survey visits — randomly selected households will receive tax coupons based on algorithm implemented within the surveying software on tablets used by enumerators. To access the tax rate discounts, selected individuals will need to present their coupon to avoid paying the full rate. Coupons will be individualized using the name and identification code written on the property owner’s state identity card. Failure to present the coupon and/or the accompanying identity card disqualifies households from the discount.
Randomization Unit Th city of Kananga has been divided in 361 polygons. Randomization of the taxation arms will be implemented at the polygon level. Randomization of the property titling intervention will be implemented at the individual level within polygons. The city of Kananga has been divided in 361 polygons. randomized. Polygons were defined for a recent tax collection campaign studied by one of the researchers using a satellite map of Kananga and dividing the urban areas into 361 polygons of similar size along naturally occurring boundaries, such as avenues and ravines. Randomization of the taxation arms will be implemented at the polygon level. Randomization of the property titling intervention will be implemented at the individual level within polygons.
Sample size (or number of clusters) by treatment arms The study comprises 361 polygons. The direct taxation arm comprises 120 polygons. The indirect taxation arm comprises 120 polygons. The control arm (neither direct nor indirect taxation) comprises 121 polygons. Within these categories of polygons, 12 people will be sampled per polygon, so the total sample size will be 4,332. See page 39 of the PAP for further information. The study comprises 361 polygons. The central taxation arm comprises 181 polygons. The local taxation arm comprises 180 polygons. Within these categories of polygons, 12 people will be sampled per polygon, so the total sample size will be 4,332. See page section 2.1 of the PAP for further information.
Power calculation: Minimum Detectable Effect Size for Main Outcomes See power.xls attached.
Public Analysis Plan No Yes
Back to top
Analysis Plans
Field Before After

Title

The Origins of Tax Compliance and State Capacity: Pre-Analysis Plan

Document

Back to top