We are happy to announce that all trial registrations will now be issued DOIs (digital object identifiers). For more information, see here.
Economics of a Light Bulb: Experimental Evidence on CFLs and End-User Behavior
Last registered on August 14, 2014


Trial Information
General Information
Economics of a Light Bulb: Experimental Evidence on CFLs and End-User Behavior
Initial registration date
August 14, 2014
Last updated
August 14, 2014 2:31 PM EDT
Primary Investigator
World Bank
Other Primary Investigator(s)
PI Affiliation
University of Michigan
Additional Trial Information
On going
Start date
End date
Secondary IDs
Residential access to modern energy and lighting is important for development in that it improves living standards and productivity. However, developing countries often face severe constraints on available electricity, which result in blackouts. In the developing world, where lighting is a major component of residential electricity consumption, energy efficient technologies, such as compact fluorescent lamps (CFLs), can reduce overall household electricity demand while permitting utilities to reach additional customers with existing supplies. Consumers’ decisions to purchase energy efficient durables require a tradeoff between upfront and ongoing operating costs; however, evidence suggests that individuals are not maximizing the net present value of their energy-related expenditures when making purchase decisions for electricity-consuming durables. To date, empirical evidence on this phenomenon - the “energy efficiency gap” – has focused primarily on the initial purchase decision. In contrast, we focus on the behavioral response following technology adoption, household learning with respect to energy savings, and how these impact future purchase decisions. In order to better understand the adoption, continued use and diffusion process of the CFL technology, we conduct randomized experiments with 1,000 households in Kyrgyzstan to measure: (1) the “rebound” or behavioral responses that boost consumption and offset CFLs’ technologically feasible electricity savings, (2) the differential impacts of energy efficient technologies received for free or a positive price, and (3) the role of information spillovers in spreading information on the energy efficient technology. Willingness to pay (WTP) was also experimentally elicited.
External Link(s)
Registration Citation
Carranza, Eliana and Robyn Meeks. 2014. "Economics of a Light Bulb: Experimental Evidence on CFLs and End-User Behavior." AEA RCT Registry. August 14. https://doi.org/10.1257/rct.461-1.0.
Former Citation
Carranza, Eliana and Robyn Meeks. 2014. "Economics of a Light Bulb: Experimental Evidence on CFLs and End-User Behavior." AEA RCT Registry. August 14. https://www.socialscienceregistry.org/trials/461/history/2466.
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
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Household-reported willingness to pay (WTP) measures, electricity usage reports from meter logs, ownership and usage of CFLs, measures of electricity-related behaviors.
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
The experiment was conducted in a district with over 40,000 residential electricity customers, which is situated close to Bishkek, the capital of Kyrgyzstan. Within this district, seven villages were selected due to their accessibility from Bishkek. Within the 7 villages, 110 transformers groups (TGs) were randomly assigned to one of three treatment intensities: control (zero percent of households treated), low (60% of surveyed households treated), and high (80% of surveyed households treated). Within each transformer group, approximately 20% of households were randomly selected to be surveyed, resulting in 1,000 households participating at baseline. Of the selected sample, households were randomly assigned to one of three groups: the control group (C), the group receiving the opportunity to purchase the light bulbs at a randomly drawn price (T1), and the group receiving the light bulbs for free (T2). Assignment occurred via a random number generator. Households assigned to receive the CFL bulbs either for free or for a positive price played a willingness to pay game. We measure WTP by randomly assigning a purchase price to each household, in an experiment based on Becker-DeGroot-Marschak (1964) (BDM). BDM is particularly useful in tracing the demand for CFLs at a number of price points and estimating WTP for products that are quite price sensitive. Households were able to bid on up to four CFL bulbs, as the average household had between 5 and 6 light bulbs in total at the time of the baseline survey. Participants were asked to state the total amount that they are willing to pay for one, two, three and four CFLs before a unit price is randomly drawn by the experimenter. If the amount they state for a number of CFLs is greater than or equal to their value at the randomly drawn price, they ought to purchase that number of bulbs at the random price. A practice round using an alternative familiar technology (pens) prior to the bidding on the product of interest helped to ensure that participants understand the method. This method is designed to create incentives for truthful value revelation (Lusk and Rousu, 2006).
Experimental Design Details
Randomization Method
Computer-based randomization.
Randomization Unit
Two levels of randomization: transformer groups and households
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
110 transformer groups.
Sample size: planned number of observations
1,000 households
Sample size (or number of clusters) by treatment arms
Transformer groups: 25 control, 45 low-intensity, 40 high intensity.
Households: 460 control, 262 T1, and 278 T2.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB Name
Harvard University Committee on the Use of Human Subjects in Research
IRB Approval Date
IRB Approval Number
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 and Papers
Preliminary Reports
Relevant Papers