Do notifications affect consumers willingness to incur power outages? Evidence from Public Safety Power Shutoffs in California

Download data and study materials from OSF

Principal investigators:

Duncan Callaway

University of California, Berkeley



Fiona Burlig

University of Chicago



Will Gorman

Lawrence Berkeley National lab



Sample size: 2120

Field period: 06/16/2021-01/10/2022

Electric power outages are costly and negatively impact individuals, businesses, and governments. For individuals, outages result in the loss of food, access to leisure activities, and potentially vital health services such as breathing machines. For businesses, electricity is often critical to serving customers and earning revenues while governments rely on electricity to provide vital public services. But, how much should soceity pay to maintain a reliable electricity system? This project involves a randomized experiment that aims to generate estimates for how advanced notification impacts household willingness-to-pay (WTP) to avoid outages in the context of Public Safety Power Shutoffs that have become more common in California due to wildfire risk. Though prior work measuring the VoLL has indicated that advanced warning likely has a mitigating effect on the cost of power outages, there have been no studies that quantify the cost reduction potential of advanced warning. We find positive and statistically significant WTP to avoid power outages of $11/kWh, consistent with the expectation that outages are costly to the residential sector. We do not find statistically significant evidence that advanced notification reduces WTP, though we find directional evidence that it does. There is limited evidence that these results vary by income and wealth levels. Back-up power ownership is positively correlated with respondents’ WTP to avoid outages.

Respondents who are notified of hypothetical upcoming power outages will have lower stated willingness-to-pay to avoid power outages.

Additional Research Questions:
1) How does prior ownership of back-up power impact respondents willingness to pay to avoid power outages?
2) Are respondents with higher-income and/or weather levels willing-to-pay more to avoid power outages?

Experimental Manipulations
We conducted a stated-preference experiment where respondents were randomly assigned to one of 3 study groups who were all told they would hypothetically experience 2 outages, each lasting 40 hours, between the months of September and November, in line with the typical timing of Public Safety Power Shutoffs in California. Group 1 was a no exposure control group (received no notification of power outage). Group 2 was provided the same description of a power outage event, but told they would be notified of the power outage one day in advance. Group 3 was told they would be notified of the power outage one week in advance. We stratified our sampling by whether or not customers owned back-up power generators.
Participants answered one question after receiving the experimental manipulations: how much would they be willing to pay per month for the entirety of a year to avoid power outages?
Summary of Results

The first questions of the survey focused on respondent experiences and future expectations for sustained power outage events. We draw three key conclusions from the responses to these questions. First, we find that almost the entirety of the sample is aware of wildfire-related power outages, though only 25% experienced one. Second, we find that at the time of the survey the annual mean expectations for future power outages are relatively low at 12 hours per year per respondent, with that number roughly doubling if respondents reported experiencing a sustained power outage over the last three years (i.e. 2018-2021). Finally, we find that while backup power purchases have increased in recent years, low cost mitigation options like using flashlights and candles remain more popular (75% of respondents report using these low-cost strategies compared to 20% for other mitigation options).

For the experimental question, we find that respondents had positive and statistically significant stated WTP to avoid power outages ($85/month std. error $11/month). The WTP did not depend significantly on the advanced notification of the power outage ( - $16/month std. error $16/month for 1-day notice; -$6/month std. error $16/month for 1-week notice). Respondents who indicated ownership of back-up power had a higher WTP to avoid power outages than non-back-up power owners ($83/month increase s.e. $13/month). An exploration of heterogeneity suggested that household wealth (proxied using estimated home value) is negatively correlated with WTP. This relationship, however, is not present with household income, which was not a strong determinant of respondents’ WTP.