x

We are happy to announce that all trial registrations will now be issued DOIs (digital object identifiers). For more information, see here.
The Elasticity of Tax Compliance: Evidence from Randomized Property Tax Rates
Last registered on January 28, 2019

Pre-Trial

Trial Information
General Information
Title
The Elasticity of Tax Compliance: Evidence from Randomized Property Tax Rates
RCT ID
AEARCTR-0003818
Initial registration date
January 24, 2019
Last updated
January 28, 2019 8:52 AM EST
Location(s)
Primary Investigator
Affiliation
Harvard University
Other Primary Investigator(s)
PI Affiliation
Harvard University
PI Affiliation
London School of Economics
Additional Trial Information
Status
On going
Start date
2018-06-15
End date
2019-04-30
Secondary IDs
Abstract
How does tax compliance vary with the size of the tax burden when opportunities for evasion are high? This paper estimates the elasticity of property tax compliance in a field experiment in Kananga, the Democratic Republic of Congo, a setting where the status quo level of compliance is low. In collaboration with the provincial government, we randomly assign four fixed tax rates at the household level as part of a door-to-door city-wide tax collection campaign. Individuals face between 50 and 100% of their true liability. We study how compliance and total government revenues vary with the rate. We also examine the effects of randomized rates on bribe payment and contributions to informal taxes (in-kind labor payments). Our findings will contribute to knowledge about the determinants of tax compliance in weak states as well as the design of optimal liabilities and enforcement in such settings.
External Link(s)
Registration Citation
Citation
Bergeron, Augustin, Gabriel Tourek and Jonathan Weigel. 2019. "The Elasticity of Tax Compliance: Evidence from Randomized Property Tax Rates." AEA RCT Registry. January 28. https://doi.org/10.1257/rct.3818-2.0.
Former Citation
Bergeron, Augustin, Gabriel Tourek and Jonathan Weigel. 2019. "The Elasticity of Tax Compliance: Evidence from Randomized Property Tax Rates." AEA RCT Registry. January 28. https://www.socialscienceregistry.org/trials/3818/history/40744.
Experimental Details
Interventions
Intervention(s)
In collaboration with the provincial government, we randomly vary the tax rate faced by periphery and midrange property owners between 50 percent and 100 percent of the full property tax rate. For periphery properties, the property tax rate experimentally varies between 1,500 Congolese Francs (CF) and 3,000 CF, in increments of 500 CF. For midrange properties, the property tax rate experimentally varies between 6,600 CF and 13,200 CF, in increments of 2,200 CF. Citizens are not informed that they may be receiving discounts: they are simply randomly assigned the rate on the tax bill without mention of the full rate.

In addition, we randomize the 'bonuses' received by tax collectors for collecting taxes:
1. Half of the houses in the periphery category are assigned a fixed 750 CF bonus.
2. Half of the houses in the periphery category are assigned to a bonus equal to 30% of the tax rate corresponding to that household.
Holding the bonus constant—at a fixed amount or proportional share of the tax amount— will permit us to control for tax collector incentives across rates
Intervention Start Date
2018-06-15
Intervention End Date
2019-04-30
Primary Outcomes
Primary Outcomes (end points)
The primary outcome is tax compliance (i.e., whether citizens pay the property tax).
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary outcomes include 1) bribe payment to tax collectors, 2) households' contributions to informal taxes, 3) citizens' views of the local government, 4) measures of spending and consumption.
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
The unit of randomization is the property. There are approximately 46,000 properties in total in Kananga according to data from the 2016 property tax campaign described in Weigel (2018).

We employ a block-randomized design, stratifying on neighborhood, each of which was identified on a satellite map with boundaries like roads, ravines, or other natural features that would be easily identifiable from the ground. There are 364 neighborhoods in total in Kananga. We excluded the 8 neighborhoods that were part of the logistics pilot mentioned above.
Experimental Design Details
Randomization Method
Randomization was done by a computer.
Randomization Unit
The unit of randomization is the property / property owner.
Was the treatment clustered?
Yes
Experiment Characteristics
Sample size: planned number of clusters
356 neighborhoods (polygons defined on a satellite map of the city of Kananga using boundaries like roads, ravines, or other natural features that would be easily identifiable from the ground)
Sample size: planned number of observations
For our primary outcome tax payment we will use administrative data to evaluate the effect of the various treatment arms on compliance. For this analysis, we have the universe of compounds in Kananga, approximately 46,000 according to data from Weigel (2018). For other outcomes, we will rely on a midline survey conducted with all the property owners in the sample (N=46,000) and on an endline survey with a sample size at least as large as our baseline sample (N=4,343).
Sample size (or number of clusters) by treatment arms
50 Percent of the Full Tax Rate : 11,500 (10,350 periphery properties, 1150 midrange properties)
66 Percent of the Full Tax Rate: 11,500 (10,350 periphery properties, 1150 midrange properties)
88 Percent of the Full Tax Rate: 11,500 (10,350 periphery properties, 1150 midrange properties)
100 Percent of the Full Tax Rate: 11,500 (10,350 periphery properties, 1150 midrange properties)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Simple power calculations were conducted using data from the 2016 property tax campaign. According to Stata’s power command, we would need a sample of 838 individuals per tax-rate arm to detect an effect of 0.2 standard deviations (about 5 percentage points) on the mean from neighborhoods visited by tax collectors in 2016. Our projected sample size of 11,500 individuals per tax-rate arm allows us to detect effects of less than 0.1 standard deviations (less than 2.5 percentage points).
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Harvard Institutional Review Board
IRB Approval Date
2017-07-28
IRB Approval Number
IRB17-0724
Analysis Plan
Analysis Plan Documents
PAP_elasticity_AEA.pdf

MD5: de54ecd07b170b71f8947b9d146483c5

SHA1: 2911f77a721144b94bc87869f77c14766cd036e3

Uploaded At: January 24, 2019

Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
No
Is data collection complete?
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers