|
Field
Trial Status
|
Before
in_development
|
After
on_going
|
|
Field
Abstract
|
Before
In low-income contexts, particularly where state-capacity is low, informal taxation is a prevalent mechanism by which local leaders rely on social norms to raise contributions (money or labor) for public goods. In this paper, I compare two ways of measuring the incidence of informal taxation using data I collected in rural Sierra Leone. The first measure is based on past contributions to chiefs for multiple local public goods and the second is inferred from behavior in a labor supply field experiment. The latter has the advantage that captures distortions in behavior arising from informal taxation rather than past contributions that could have been partially voluntary. I then use these two measures to examine how the incidence of informal taxation is distributed among the rural population, particularly to shed light on the claim that informal taxation is regressive as previous work has emphasized.
|
After
This paper investigates the relative efficiency of traditional leaders when informally taxing citizens in low-income states and whether this comes at the expense of relatively poor households. I design a field experiment to measure whether citizens engage in costly actions to avoid contributing their labor to a public good. I randomize communities across different methods to select contributors to compare the status quo of chiefs selecting contributors to two alternatives: random lotteries and progressive selection based on household surveys. I use the random selection arm as a benchmark and estimate whether selection by chiefs or progressive selecting are relatively efficient by generating more or less costly behavior from citizens. I also study the heterogenous effects of these treatment arms by household wealth. This allows me to asses if chiefs appear regressive by mostly burdening poor households and whether this can be corrected by a simple policy instrument.
|
|
Field
Trial End Date
|
Before
July 31, 2023
|
After
December 31, 2023
|
|
Field
JEL Code(s)
|
Before
O17, H41, H71
|
After
C93, D91, H21, H23, H41, I32, O12, O17
|
|
Field
Last Published
|
Before
December 13, 2022 10:51 PM
|
After
October 04, 2023 03:49 PM
|
|
Field
Intervention (Public)
|
Before
The intervention happens within an experimental environment where participants will be offered a one day job in rural Sierra Leone. At random, for some of them their earnings in this job will be visible to their traditional leader (chief) as part of a report of activities. Therefore, this will create a random increase in the possibility of informal taxation.
|
After
The intervention happens within an experimental environment where participants will be offered a one day job in rural Sierra Leone. Some participants will be selected to work for themselves and others for their community. Community workers do the same task as others, but their earnings are donated to their local clinic.
|
|
Field
Intervention Start Date
|
Before
February 01, 2023
|
After
October 09, 2023
|
|
Field
Intervention End Date
|
Before
July 31, 2023
|
After
November 30, 2023
|
|
Field
Primary Outcomes (End Points)
|
Before
Labor supply decisions within the one-day offer (extensive and intensive margins)
|
After
Costly decisions by participants to avoid being selected.
|
|
Field
Primary Outcomes (Explanation)
|
Before
These outcomes will be constructed the following way: for the extensive margin decision participants will do a Multiple Price List to state for which wages they will take up the job under each visibility condition. For the intensive margin, participants will both state how much time they want to work plus how many units of work (completed tasks) they end up doing at the randomly offer piece-rate.
|
After
The experiment is done in two days and selection of community workers only takes place on the second day. I then measure costly behavior by participants to work on the first day. I do this by eliciting the largest wage-cut they are willing to accept in order to work earlier.
|
|
Field
Experimental Design (Public)
|
Before
The experimental design consists on recruiting a random sample of people in rural Sierra Leone for a one day job opportunity. I particularly want to sample heavily from the tails of the wealth/income distribution within each community. Then, once I have a sample of people in each community I will offer them a one day job. However, to 50% of them at random I will tell them their income in the job will be visible to their local chief as part of a reporting requirement. This will create the opportunity for informal taxation to take place. Then I will measure labor market outcomes in the one-day job (take-up and effort), which will allow me to infer an experimental informal tax rate.
|
After
I implement in multiple communities different methods to select community workers. In some places the local chief selects, in others it is decided by lottery, and in others it is decided by a progressive selection rule inspired by proxy-means-testing. These random variation across communities allows me to study the effects of each selection method on participant's behavior.
|
|
Field
Randomization Method
|
Before
The randomization will be done with each participant via a coin-flip.
|
After
The randomization will be done by the research team in a computer before sending the field team. Field teams will simply have instructions to implement them. Other randomizations within surveys are done via public lotteries with participants.
|
|
Field
Randomization Unit
|
Before
Participants in the experiment.
|
After
Communities in rural Sierra Leone and participants within these communities.
|
|
Field
Was the treatment clustered?
|
Before
No
|
After
Yes
|
|
Field
Planned Number of Clusters
|
Before
2000
|
After
88
|
|
Field
Planned Number of Observations
|
Before
2000
|
After
1300
|
|
Field
Sample size (or number of clusters) by treatment arms
|
Before
1000 in each arm
|
After
Between 28 and 32.
|
|
Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
Based on simulation that aim to capture an implied average rate of informal taxation of 6.6%, I have 95% power to detect a significant decrease in labor supply of about 8% on average and about 15% for the top half of the most "taxed" individuals. Splitting into more heterogeneity groups, I have an 80% to detect a difference in labor supply between the top tercile and the bottom tercile in the distribution of informal tax incidence, a 90% power to deter differences in labor supply between the top and bottom quartiles in the distribution of tax incidence, and a 70% power to detect differences in labor supply between the top quartile and the second quartile in the distribution of tax incidence.
|
After
Based on a clustered design, I have around 73% power to detect a 15% increase in my main outcome. The outcome is measured in monetary units (Leones) and captures wage-cuts people are willing to accept. A 15% increase in wage-cut implies accepting a daily salary 3% lower.
|
|
Field
Secondary Outcomes (End Points)
|
Before
|
After
Labor supply decisions within the experiment.
|
|
Field
Secondary Outcomes (Explanation)
|
Before
|
After
This is based on the experimental job people are offered which is a simple classification task. The outcome is the number of pages participants complete.
|