NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Paying for urban services: utility bills, prepaid metering and spending patterns of the poor
Last registered on September 03, 2018


Trial Information
General Information
Paying for urban services: utility bills, prepaid metering and spending patterns of the poor
Initial registration date
January 20, 2017
Last updated
September 03, 2018 5:15 PM EDT
Primary Investigator
UC Santa Barbara
Other Primary Investigator(s)
PI Affiliation
PI Affiliation
University of Cape Town and J-PAL
Additional Trial Information
On going
Start date
End date
Secondary IDs
Revenue recovery is a challenge for urban service providers in developing countries. Poor customers often struggle to pay monthly bills and providers face both high costs and political economy barriers to enforcing payment. Prepayment is an increasingly popular solution to this problem in the electricity and water sectors. In our study setting, low income households purchase prepaid electricity every 3 days, on average, and use 12 percent less electricity than when they are billed monthly. We will investigate the preferences and constraints that give rise to these spending and consumption patterns on prepaid metering and is organized around two core questions:
a) Is the reduction in consumption associated with prepaid metering a choice by the household, or is it driven by constraints such as liquidity constraints and higher transaction costs?
b) Do high frequency purchasing patterns indicate demand for self-control, cash-on-hand liquidity constraints, or demand for savings?
Our findings will both shed light on factors that contribute to current high rates of payment delinquency on monthly billing and will inform the design of interventions to improve revenue recovery for urban services while simultaneously meeting the needs of low income households.
External Link(s)
Registration Citation
Jack, Kelsey, Kathryn McDermott and Anja Sautmann. 2018. "Paying for urban services: utility bills, prepaid metering and spending patterns of the poor." AEA RCT Registry. September 03. https://doi.org/10.1257/rct.1886-3.0.
Former Citation
Jack, Kelsey et al. 2018. "Paying for urban services: utility bills, prepaid metering and spending patterns of the poor." AEA RCT Registry. September 03. http://www.socialscienceregistry.org/trials/1886/history/33840.
Sponsors & Partners

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
Experimental Details
We will offer transfers of electricity credit to prepaid electricity customers in conjunction with two rounds of surveying. Specifically, a first survey will be accompanied by surprise transfers of either cash or electricity. The total value of the transfers will be the same but some households will receive the transfer in cash, some in a single prepaid electricity token and some in two prepaid electricity token (one of which will be provided at the time of the survey and one delivered by text message after the survey). A first control group will receive the survey but no transfer. A second control group will receive neither surveys nor a transfer.

In a second survey round, with the same households, participants will be randomly assigned to choose between two of the three transfer options offered in the first survey round. These choices will be administered using multiple price lists to measure preferences across different transfer types.

See the pre-analysis plan for further detail on the design.
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Our analysis will include three main outcomes:
1) purchase quantity and timing, constructed from the prepaid vending data
2) consumption, collected from meter readings during the two survey rounds
3) choices among transfer options in survey round 2
Primary Outcomes (explanation)
See pre-analysis plan for further detail on outcome variable construction.
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
Round 1 treatment – along with the first survey round, households will be randomly assigned to one of the following four groups.
1) 100 Rand in cash handed over at the end of the survey.
2) One electricity token with the same value as the cash transfer. The token will be uploaded on to the household meter by the subject, with assistance from the survey enumerator.
3) Two electricity tokens with the same total value as the cash transfer. One of these tokens will be provided to the household at the time of the survey and uploaded by the subject onto the household meter with assistance by the survey enumerator. The second token will be sent by text message approximately three days later.
4) Survey-only control.

All tokens are sent to the subject directly by text message. All households in the first survey will receive R20 of electricity as compensation for participating in the survey. This amount will also be uploaded onto the household meter by the subject in the presence of the survey enumerator. Thus, survey-only households also upload a token with the surveyor.

Round 2 choices – Households will be partially randomly assigned (see below) to receive two of the following choice sets as part of the second survey.
a) Cash versus one electricity token
b) Cash versus two electricity tokens (with one token sent 3 days later)
c) One versus two electricity tokens (with one token sent 3 days later)

Round 2 choice sets will be presented as multiple price lists. Respondents are asked to make a series of choices between two possible options in each choice set. The highest possible value that can be obtained remains constant, while the value of the other option increases or decreases. We randomized both the order within each choice (left or right choice on the screen) and the order from first to last within each choice set. One choice within one of the choice sets is drawn for implementation. The choice sets are depicted in the Appendix.

In addition, a pure control group will be followed in the administrative data throughout the study. The pure control consists of all eligible households according to the first stage sampling rules (excluding eligibility criteria for the survey respondent), who were never selected for surveying (anyone selected for surveying, regardless if the survey took place or not, is excluded from the pure control).
Experimental Design Details
Randomization Method
Randomization done by computer
Randomization Unit
Randomization will be stratified by survey team and geographic area. Surveys will be collected using handheld devices. Surveyors are not aware of a household’s treatment at the start of the survey and will not be able to manipulate treatment assignment.

In the second round, treatments will be based on round 1 treatment assignment. Specifically, households will be assigned to two out of the four possible round 2 choice conditions. Following the numbering above, assignment will be as follows:
- Round 1 treatment 1 will receive round 2 choices c and randomly drawn a or b
- Round 1 treatment 2 will receive round 2 choices b and randomly drawn a or c
- Round 1 treatment 3 will receive round 2 choices a and randomly drawn b or c
- Round 1 control (4) will receive a random set of two choices

In other words, respondents who received a round 1 treatment will always see the choice set involving the two transfers that they did not experience in round 1, plus one of the two choice sets involving the transfer they did experience in round 1. The sample of households will be assigned evenly across the four round 1 groups.
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
Sample size: planned number of observations
Target is 800 households for round 1 survey
Sample size (or number of clusters) by treatment arms
200 per round 1 arm (plus pure control)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
UC Santa Barbara
IRB Approval Date
IRB Approval Number
IRB Name
University of Cape Town
IRB Approval Date
IRB Approval Number
IRB Name
Tufts University
IRB Approval Date
IRB Approval Number
Analysis Plan
Analysis Plan Documents
Pre-analysis plan

MD5: 793c20c6083d59d0f1e82318898ea279

SHA1: aa6037b55ce12be94558a0124aef0c159e6459b4

Uploaded At: September 03, 2018

Post Trial Information
Study Withdrawal
Is the intervention completed?
Is data collection complete?
Data Publication
Data Publication
Is public data available?
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)