The Effect of Financial and Non-Financial Incentives on Willingness to Conserve Energy During Peak Periods

Last registered on August 28, 2017

Pre-Trial

Trial Information

General Information

Title
The Effect of Financial and Non-Financial Incentives on Willingness to Conserve Energy During Peak Periods
RCT ID
AEARCTR-0002297
Initial registration date
August 27, 2017

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
August 28, 2017, 5:41 PM EDT

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

Locations

Primary Investigator

Affiliation
UCSD

Other Primary Investigator(s)

PI Affiliation
UCLA
PI Affiliation
Stanford

Additional Trial Information

Status
In development
Start date
2017-08-28
End date
2018-12-31
Secondary IDs
Abstract
This study has three primary goals. First, we plan to estimate how willingness to reduce electricity use during "peak periods" varies based on weather conditions (i.e. heat), individual characteristics (i.e. baseline income) and household characteristics (i.e. having solar or a PEV).
Second, we plan to examine how the interaction between financial and non-financial framing affect consumers willingness to reduce electricity during peak periods. This part of the project has two parts. First, we randomly assign individuals to receive one of two frames when being introduced to the project: a financial frame (emphasizing financial savings from conservation) and an environmental frame. Individuals in the environmental frame are then further randomized to receive one of two framings when receiving demand events: a moral tax frame that emphasizes the negative consequences of not participating and a moral subsidy frame that emphasizes the benefits of participating. All individuals are then further randomized into either receiving a financial reward for participating in the demand event or no reward. Using the interactions between these treatment cohorts, we will test: a) how introductory framing affects willingness to reduce energy , b) whether framing the consequences of (not) participating in a positive or negative light affects willingness to reduce energy and c) whether any effects of the environmental framing are "crowded in" or "crowded out" from receiving a financial reward.
Finally, we will examine how various features of demand events affects consumer's willingness to reduce energy. Specifically, we examine: a) whether consumers respond differently to a constant vs event-varying price and whether this response depends upon if the price increased or decreased relative to the previous event and b) if consumers willingness to reduce is increased or decreased by concentrating events (i.e. 3 events/week vs 1 event/week) and if a greater concentration of events leads to longer-term energy conservation habit formation. Our sample will consist of more than 12,000 customers from Chai Energy, an app-based energy demand system. We will attempt to recruit customers from diverse socio-economic backgrounds as well as from a wide geographic range (within California) in order to test heterogeneity in response.
External Link(s)

Registration Citation

Citation
, , KC Hirsch and Sam Krumholz. 2017. "The Effect of Financial and Non-Financial Incentives on Willingness to Conserve Energy During Peak Periods." AEA RCT Registry. August 28. https://doi.org/10.1257/rct.2297-1.0
Former Citation
, , KC Hirsch and Sam Krumholz. 2017. "The Effect of Financial and Non-Financial Incentives on Willingness to Conserve Energy During Peak Periods." AEA RCT Registry. August 28. https://www.socialscienceregistry.org/trials/2297/history/20928
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
The primary intervention in this study is a demand response event, in which households will be given an incentive to cut electricity use during certain (randomly-determined) periods. The types of incentives and amount of information contained about the purpose of the event will vary across treatment cohorts in a manner that is described in the experimental design section. A secondary intervention is the provision of Chai Gateway Devices, which provide a read-out of instantaneous energy consumption direct to a consumer's phone. These devices will be randomly distributed to half of our sample.
Intervention Start Date
2017-08-28
Intervention End Date
2017-11-03

Primary Outcomes

Primary Outcomes (end points)
1) Energy consumption during, demand event periods, periods surrounding demand events and after the end of the experiment
2) Changes in energy use including air conditioning "set points" (derived from a proprietary Chai algorithm) and solar adoption
3) User activity on the Chai site (frequency of log-in, amount of time spent on the site, etc)
4) Willingness to pay to receive an event
5) Attitudes on conservation from a household survey
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Our study sample is the user base of Chai Energy, an app-based energy demand management company. Individuals in our sample will be randomized into treatment or control. Control individuals will receive no information about the study. 2,600 individuals will be randomized into the control group. Treatment individuals will receive a notification that they are part of a new Chai program aimed at saving energy during peak periods. Individuals in this group will receive "demand events" during randomly selected days. There are 13 different variations of treatment (sample size included in parentheses):

Financial: Individuals in the financial group will receive only messaging about the financial benefits from cutting energy use during demand events (or for info-only subjects, a neutral message informing them about the demand event). Members in this group will receive the following financial rewards for each KwH of energy reduced relative to baseline (trailing average of previous 10 days use) during the event period:
Treatment 1: Info-only (600 hh)
Treatment 2: $.05/KwH (600 hh)
Treatment 3: $.5/KwH (600 hh)
Treatment 4: $1/KwH (600 hh)
Treatment 5: $2/KwH (600 hh)
Treatment 6: $5/KwH (300 hh)
Treatment 7: Randomly selected amount-equal prob of (.05,.5,1,2,5) (1,200 hh)

Moral Tax and Moral Subsidy: Individuals in the moral tax and moral subsidy groups will receive an identical introductory message describing how cutting electricity use has large health and environmental benefits. Individuals in the moral tax group will receive event notifications emphasizing the negative consequences of not participating, while individuals in the moral subsidy group will receive event notifications emphasizing the positive consequences of participating. In both groups a third of participants will receive no financial reward, a third will receive a “moderate” financial reward and a third will receive a “large” financial reward for participating in the event. Accordingly, this creates the following 6 treatments:
Treatment 8: Moral Subsidy-Info Only (900 hh)
Treatment 9: Moral Subsidy-$.5/KwH (900 hh)
Treatment 10: Moral Subsidy-$2/KwH (900 hh)
Treatment 11: Moral Tax-Info Only (900 hh)
Treatment 12: Moral Tax-$.5/KwH (900 hh)
Treatment 13: Moral Tax-$2/KwH (900 hh)


In addition to the 13 treatments described above, all participants will be randomized into the following cross-cutting treatments:
1) Event Frequency: Half of participants will receive one event/week for ten weeks and half will receive three events/week for the first 2 weeks and then 1 event/2 weeks for the following 8 weeks.
2) Chai Gateway: Half of participants will receive a Chai Gateway device, which will provide households with instaneous information about their electricity use.
Experimental Design Details
Randomization Method
Randomized in office by a computer
Randomization Unit
Chai account (household)
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
12,500 households
Sample size: planned number of observations
12,500 households
Sample size (or number of clusters) by treatment arms
2,600 Control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The mininum detectable effect size for cohorts with 900 individuals is .19 standard deviation units of energy consumption, for 600 individuals is .23 standard deviation units of energy consumption and for 300 individuals is .33 standard deviation units of energy consumption . This implies power to detect a 11% change, 13% change and 20% change respectively. However, this power calculation is extremely conservative because it assumes a perfect ICC within individuals across events. Because each individual will have ten events, in reality we will likely have significantly greater statistical power than the estimates presented here.
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

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

Request Information

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