Pushing on a String? Price Interventions to Reduce Theft in Pakistan’s Power Sector

Last registered on September 05, 2023

Pre-Trial

Trial Information

General Information

Title
Pushing on a String? Price Interventions to Reduce Theft in Pakistan’s Power Sector
RCT ID
AEARCTR-0010777
Initial registration date
March 01, 2023

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
March 13, 2023, 8:38 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
September 05, 2023, 1:39 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
Tufts University

Other Primary Investigator(s)

PI Affiliation
London School of Economics and Political Science
PI Affiliation
The University of Chicago
PI Affiliation
The University of Chicago
PI Affiliation
London School of Economics and Political Science

Additional Trial Information

Status
On going
Start date
2022-11-01
End date
2024-06-30
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Our study tests how prices and the availability of outside options like off-grid solar alter theft decisions in rural Pakistan. We begin by documenting Pakistan’s highly sophisticated electricity theft landscape. Despite growing government enforcement efforts, households in many areas still steal liberally. They demonstrate complex evasive behaviours such as running assets on different meters, bunching at key tariff thresholds while using illegal lines for excess consumption, or frequently switching between solar and illegal
consumption. The experiment explores these dynamics directly by introducing block-by-block subsidies to measure the price elasticities of demand and theft (unpaid bills and illegal connections). The introduction of block-by-block subsidies enables for more accurate estimation of demand elasticities compared to prevailing estimates that ignore the non-linear tariff structure. We also estimate a novel elasticity of electricity theft with respect to the price. By introducing subsidies, we change the set of relative prices between grid, solar, and illegal connections, enabling us to study the substitution dynamics between these competing technologies for accessing electricity. This study is a product of sustained collaboration with the Government of Pakistan over the past four years. They remain the primary stakeholder and facilitator of this research.
External Link(s)

Registration Citation

Citation
Burgess, Robin et al. 2023. "Pushing on a String? Price Interventions to Reduce Theft in Pakistan’s Power Sector." AEA RCT Registry. September 05. https://doi.org/10.1257/rct.10777-2.0
Experimental Details

Interventions

Intervention(s)
This experiment takes place in Khyber Pakhtunkhwa, a rural and poor province of Pakistan with extremely high levels of theft. A recently conducted village census of 2500 households reveals stark facts. Nearly 15% of households never paid any amount of their bill in the past year. Conditional on paying something, average bill recovery remains staggeringly low at 59%. 30% of metered households also have an illegal connection, while 35% also have a solar panel. Substantial bunching is observed at key price thresholds, indicating consumers are salient to price changes and make deliberate adjustments to lower their bills. The electricity theft landscape is therefore highly sophisticated.

Our first goal is to understand how sensitive demand and theft is to the price of electricity. Our experiment will cover 1700 households across 75 villages. We will introduce a 30% per unit subsidy at three different blocks in the non-linear tariff schedule (or the supply curve). Providing within-bin subsidies shifts different segments of the supply curve, which allows us to estimate elasticities for those who consume in the different segments. The subsidy will last for six months. Mean metered consumption varies between 120-140 units/month depending on the season.

A. 0-100 units [340 households]
B. 101-200 units [340 households]
C. 201-300 units [340 households]
D. 0-300 units [340 households]
E. Control [340 households]

This setup will enable us to capture the non-linear budget set that exists in Pakistan’s power sector. When calculating bills, households get the benefit of the previous block only, generating discrete jumps in both the average price and marginal price schedule. A household in group B that consumes 50 units will receive no subsidy, while a household which consumes 250 units will receive a 30% discount on its consumption between 101-200 units.
Intervention (Hidden)
This experiment takes place in Khyber Pakhtunkhwa, a rural and poor province of Pakistan with extremely high levels of theft. A recently conducted village census of 2500 households reveals stark facts. Nearly 15% of households never paid any amount of their bill in the past year. Conditional on paying something, average bill recovery remains staggeringly low at 59%. 30% of metered households also have an illegal connection, while 35% also have a solar panel. Substantial bunching is observed at key price thresholds, indicating consumers are salient to price changes and make deliberate adjustments to lower their bills. The electricity theft landscape is therefore highly sophisticated.

Our first goal is to understand how sensitive demand and theft is to the price of electricity. Our experiment will cover 1700 households across 75 villages. We will introduce a 30% per unit subsidy at three different blocks in the non-linear tariff schedule (or the supply curve). Providing within-bin subsidies shifts different segments of the supply curve, which allows us to estimate elasticities for those who consume in the different segments. The subsidy will last for six months. Mean metered consumption varies between 120-140 units/month depending on the season.

A. 0-100 units [340 households]
B. 101-200 units [340 households]
C. 201-300 units [340 households]
D. 0-300 units [340 households]
E. Control [340 households]

This setup will enable us to capture the non-linear budget set that exists in Pakistan’s power sector. When calculating bills, households get the benefit of the previous block only, generating discrete jumps in both the average price and marginal price schedule. A household in group B that consumes 50 units will receive no subsidy, while a household which consumes 250 units will receive a 30% discount on its consumption between 101-200 units.
Intervention Start Date
2023-01-01
Intervention End Date
2024-06-30

Primary Outcomes

Primary Outcomes (end points)
Monthly household electricity consumption and payments will be provided by the government. Our primary outcomes of interest will be measured through changes in consumption and changes in the percent of electricity bills paid following the subsidy.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
From our completed census of 2500 households, we will first conduct a baseline survey of 1700 households. Prior to baseline a household will be assigned into one of the four treatment groups or into control. We will stratify at the distribution feeder level. We notify a household about their treatment status at the end of the survey and verify that they understand precisely how their subsidy will be calculated. We will notify households that this subsidy will last for six months. Payments to households are delivered each month in person in cash along with an explanation that spells out how the end subsidy amount was calculated.
Experimental Design Details
Randomization Method
Done in office by a computer.
Randomization Unit
Household.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
1700 households.
Sample size: planned number of observations
1700 households.
Sample size (or number of clusters) by treatment arms
340 households in each of the five groups: four treatment and one control.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Given that we have panel data, using the methodology of Burlig, Preonas & Woerman (2020), we require roughly a 10% (13 kWh) average treatment effect from our 30% subsidy to be powered with an optimal sample size of 340 households. Using administrative data on the universe of electricity bills in the province and exploiting the presence of bunching at key price thresholds, we estimate a price elasticity of demand in excess of -0.5. This gives us confidence that our design is powered.
IRB

Institutional Review Boards (IRBs)

IRB Name
LSE Research Ethics Committee
IRB Approval Date
2021-03-08
IRB Approval Number
21425

Post-Trial

Post Trial Information

Study Withdrawal

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

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials