Evidence-based Cooling Strategies for a Warming World: Assessing Demand for Efficient Fans in India

Last registered on June 15, 2022


Trial Information

General Information

Evidence-based Cooling Strategies for a Warming World: Assessing Demand for Efficient Fans in India
Initial registration date
June 06, 2022

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
June 15, 2022, 10:05 AM EDT

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


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

Request Information

Primary Investigator

UC Berkeley

Other Primary Investigator(s)

PI Affiliation
UC Santa Barbara
PI Affiliation
IIT Bombay

Additional Trial Information

On going
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
As global temperatures rise, more people are exposed to heat-related risks. Coping with heat extremes is particularly challenging for people with lower socioeconomic status who cannot afford air conditioning. A recent study estimates that over 300 million people in India are at high risk of extreme heat due to a lack of access to cooling (SE4ALL, 2022). Warmer temperatures are also straining India’s aging power system; electricity demand surges during hot weather events have triggered cascading outages across the country in recent years.

Delivery of sustainable and accessible cooling solutions in India will require innovation and intervention. Fans are the most common cooling appliance used by low-income households in India. Energy-efficient fans, which consume less than half the energy of standard induction fans, are rarely available in rural and peri-urban areas of India where a majority of low-income households live. Demand for energy-efficient appliances in these markets is perceived to be low, and the costs of building out distribution networks, marketing, and sales teams to support new product offerings are high.

We investigate the conditions under which rural households in India are willing and able to purchase energy-efficient BLDC fans. Working closely with India’s Energy Efficiency Services Limited (EESL) and a network of Self-Help Groups (SHGs) in Bihar, we combine a randomized control trial with a supply-side intervention that provides training for local shop owners in procurement, distribution, marketing, and sales support for energy-efficient fans. Fan purchase prices are randomized across customers and local warranty support is randomized across clusters of local shops.

The aims of this study are threefold (1) to determine the level of consumer demand for energy-efficient fans at different price points; (2) to determine whether demand at a given price point increases with investments in local supply chain/warranty support and (3) to assess the potential role of a well-established SHG network in accelerating the diffusion of energy-efficient fans to under-served, low-income market segments.

External Link(s)

Registration Citation

Deshmukh, Ranjit, Meredith Fowlie and Jayendran Venkateswaran. 2022. "Evidence-based Cooling Strategies for a Warming World: Assessing Demand for Efficient Fans in India ." AEA RCT Registry. June 15. https://doi.org/10.1257/rct.9524
Sponsors & Partners

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

Request Information
Experimental Details


Randomization of efficient BLDC fan purchase prices
Randomization of local warranty support
Supply-side support to facilitate local procurement/distribution/marketing/sales
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Households' willingness to pay for a BLDC fan.
Purchase/take-up of fans across a range of offered prices.
Impact of local warranty support on WTP.

Primary Outcomes (explanation)
We offer BLDC fans for sale to households who have previously expressed interest in purchasing a fan. Willingness to pay (WTP) for BLDC fans is inferred from an incentive-compatible BDM style mechanism. This experiment is conducted in the context of a real fan purchase at participating local shops. The value of local warranty support is also inferred from the difference in WTP, conditional on observable household attributes (e.g. income) and participant characteristics (e.g. risk aversion).

Secondary Outcomes

Secondary Outcomes (end points)
To shed more light on what generates heterogeneity in WTP, we will conduct additional exploratory analysis to test whether WTP varies systematically with:

• Baseline electricity expenditures/consumption levels
• Baseline energy literacy/prior beliefs about energy intensity of end uses.
• Baseline demographics (including income and savings)
• Baseline risk aversion
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
First, we elicit the willingness-to-pay (WTP) for BLDC fans in a sales trial using an incentive-compatible BDM style mechanism. This demand elicitation activity takes place at local shops that are participating in our study. Households interested in purchasing a new fan make sales appointments. This sales trial design will allow us to determine consumer demand for energy-efficient fans at different price points.

Second, for half the of the participating shops, we provide additional training and warranty support. Households are informed about this local warranty support when they are being educated about the study. Communications provide information about how any problems with the fan can be locally addressed. EESL provides a 2.5-year warranty for BLDC fans. But the web-based warranty support system is ill-equipped to support low-income customers in far-flung markets. To augment the manufacturer's warranty, we provided a 3-month local warranty whereby households could get a free replacement or repair supported by the local shop if technical problems arise. Random assignment of warranty support will allow us to determine whether investments in local supply chain support increase consumer demand.

In addition, we will assess the potential role of local SHGs and associated local shops and distribution networks in bringing energy-efficient appliances to under-served, low-income markets.
Experimental Design Details
Not available
Randomization Method
Fan purchase price/subsidy randomization was done using scratch cards. Participants scratched off a purchase price after stating their willingness to pay for a fan.

Local warranty support was randomly assigned across clusters.
Randomization Unit
Fan purchase price was randomized across individuals.
Local warranty support was randomly assigned across village clusters/solar shops.
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
Sample size: planned number of observations
~1500 households 8 solar shops
Sample size (or number of clusters) by treatment arms
The distribution of BDM prices was designed to fairly allocate prices across participants. There were 13 distinct price points. In expectation, we should have 115 participants assigned to each price point. The actual assignment will depend on which scratch card cells participants ultimately choose.
Warranty support treatment is assigned to 4 randomly selected shops. The other 4 shops serve as controls.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
To estimate how a change in the offer price impacts demand for energy efficient fans, we will be comparing adoption rates across experimental groups who received different price offers. In expectation, 115 participants will be assigned to each price point. For power purposes, we can aggregate price points once we learn (a) how many participants are allocated to each price point and (b) the baseline take-up at the highest purchase price (Rs 2390) which is close to the current procurement price and almost certainly exceeds future procurement prices. We are well-powered (using the standard threshold of 0.8) to detect an increase of 10% across groups of 250 if the adoption rates at the procurement price are low (this assumes a one-sided test, alpha =0.05). At higher baseline adoption rates (e.g. 30%), we will need to aggregate more price points. We will also be interested in the interactions between our interventions. Before knowing what baseline adoption rates look like, it is difficult to predict whether we will be able to estimate these interactions with precision.

Institutional Review Boards (IRBs)

IRB Name
UC Berkeley Committee for the Protection of Human Subjects
IRB Approval Date
IRB Approval Number