The Effect of Smart Metering on Revenue Collection, Electricity Access and Supply

Last registered on December 09, 2020

Pre-Trial

Trial Information

General Information

Title
The Effect of Smart Metering on Revenue Collection, Electricity Access and Supply
RCT ID
AEARCTR-0006022
Initial registration date
December 08, 2020
Last updated
December 09, 2020, 10:54 AM EST

Locations

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

Request Information

Primary Investigator

Affiliation
University of Chicago

Other Primary Investigator(s)

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

Additional Trial Information

Status
On going
Start date
2020-07-01
End date
2021-11-30
Secondary IDs
Abstract
Over a billion people in the developing world are not connected to the grid at all, and hundreds of millions more, once connected to the grid, get only erratic supply. A fundamental reason for this gap is that in many developing countries theft and non-payment are widespread and state-run electricity distribution companies lose money on each kilowatt-hour sold. These distribution companies are not given unlimited public subsidies, so their response is to limit the supply of electricity to those on the grid and make it hard for those off the grid to get connected.

This project will evaluate whether using technology, in the form of smart metering, can lead to increased consumer payments and/or concomitant improvements in service quality. Smart meters can lower both monitoring costs and the transaction costs associated with enforcement actions such as disconnections. They may also mitigate the role of liquidity constraints by enabling new payment models such as pre-paid metering and flexible top-ups of a running balance used to pay for power. Smart meters also enable real-time pricing, improving the utilities ability to charge more (less) for electricity at times when it is more (less) expensive to supply.

For all these benefits, there is very little evidence on whether this technology can change payment behaviors or improve reliability and access, in a developing country context with modest state capacity and high levels of theft. Our study involves conducting a large-scale randomized-control trial to measure the effect of enabling online billing and remote disconnections, via smart-meters, on consumer payment behavior, consumption levels, and supply quality. We will also seek to measure spillovers by varying the number of consumers in a neighbourhood for whom these smart meter enabled services are enabled. Finally, we will measure outcomes relating to service quality including billing complaints and consumer satisfaction measures.

The experimental setting is the state of Haryana in India and the experiment is in partnership with the Department of Energy and the state electricity utilities.
External Link(s)

Registration Citation

Citation
Burgess, Robin et al. 2020. "The Effect of Smart Metering on Revenue Collection, Electricity Access and Supply." AEA RCT Registry. December 09. https://doi.org/10.1257/rct.6022-1.0
Sponsors & Partners

Sponsors

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

Request Information
Experimental Details

Interventions

Intervention(s)
The main intervention involves the random assignment of consumers to two different regimes (not mutually exclusive) that use the advanced features of smart meters, plus a control condition.

(i) Control: status-quo billing where meters are manually read by utility staff and bills generated thereafter

(ii) Treatment 1: online billing where smart meters transmit information remotely to the utility based on which bills can be generated immediately

(iii) Treatment 2: remote disconnections based on specified utility criteria linked to non-payment of bills
Intervention Start Date
2020-07-01
Intervention End Date
2021-06-30

Primary Outcomes

Primary Outcomes (end points)
Revenue collection, electricity consumption, billing complaints, billing efficiency, electricity supply
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experiment design involves a two stage randomized assignment of consumers and 'binders' to different treatments. A binder is a group of households for whom a single meter reader records consumption each billing cycle. The treatment assignment involves a assigning randomly chosen fraction of consumers from within each binder to one of the intervention arms described above. The binders in the experiment are in the town of Karnal in North Haryana.

Randomization proceeds as follows. 125 binders in Karnal are part of the experiment. Binders are made up of many consumers - roughly 200 household consumers on average. Binders are first assigned to three groups.

(i) 25 Controls Binders - No consumers switched to online billing
(ii) 75 Treatment Binders - 80 percent of eligible consumers (based on presence of appropriate hardware) will be switched to online billing
(iii) 25 High Treatment Binders - 100 percent of eligible consumers switched to online billing

In addition, a fraction of eligible households in each binder group will be cross randomized to have remote disconnection facilities enabled. From within the three binder groups described above, 20, 60, and 20 binders respectively will be assigned to a "high disconnection" condition where 90 percent of eligible consumers in those binders will have remote disconnection facilities enabled. Consumers in the remaining 5, 15, 5 binders will be assigned to a "low disconnection" condition where only 50 percent of eligible consumers will have remote disconnection facilities enabled.

Overall the design generates random variation in both the treatment assignment at the consumer level and the intensity (saturation) of treatments at the neighbourhood (binder) level.
Experimental Design Details
Not available
Randomization Method
Randomization done through code written on STATA
Randomization Unit
Binder and consumers within binders
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
125 binders
Sample size: planned number of observations
23,541 consumers (across all binders)
Sample size (or number of clusters) by treatment arms
AMI billing treatment arm: 25 binders control, 25 binders treatment with 100% consumers switched to AMI billing, 75 binders with 80% consumers shifted to AMI billing
Remote disconnection treatment arm: 25 binders with 50% consumers in treatment and 50% consumers in control and 100 binders with 90% consumers in treatment and 10% consumers in control.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number