Public EV Charging Price Plunge Trial: Supporting Grid Stability through Dynamic Discounting

How to cite

Schein, Andrew; Bernard, Louise; Panzone, Luca; Metcalfe, Robert. 2024. "Public EV Charging Price Plunge Trial: Supporting Grid Stability through Dynamic Discounting." AEA RCT Registry. AEARCTR-0015061. https://doi.org/10.1257/rct.15061-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Andrew Schein* , Louise Bernard , Luca Panzone , Robert Metcalfe

* Primary investigator

Registry ID
AEARCTR-0015061
DOI
https://doi.org/10.1257/rct.15061-1.0
First registered
Last updated
Status
In development
Trial dates
2024-12-19 – 2025-03-31
Countries
Private
JEL codes
Q41, C93, D12
Registry record
http://www.socialscienceregistry.org/trials/15061

Abstract

This study will run a randomised control trial on electric vehicle users in the UK. We will have access to a pool of approximately 450,000 electroverse customers. Of these customers, approximately our sample will be the approximately 120,0000 users resident in the UK (where the trial takes place) who (as of 19 December 2024) have their device language set to English, have push notifications enabled, and have not opted out of receiving price plunge notifications. These customers will be exposed to one of four treatments:
1. Control: will receive no notifications about price plunge events;
2. Moral suasion: will receive a notification encouraging them to take advantage of hours with green electricity to charge their EV;
3. Low discount group: will receive notifications informing them that they will receive a discount of 15% during price plunge event hours; and
4. High discount group: will receive notifications informing them that they will receive a discount of 40% during price plunge event hours.
The specific times specified in the push notifications (i.e., “9pm-11pm”) will be exactly the same for all groups 2, 3, and 4.
Customers will be randomised into groups, each of which receives a sequence of treatments in a partial “Latin Squares” structure as part of a crossover randomisation design. There will be wash-out periods of 1 week between crossovers.

Intervention

Customers will be exposed to one of four treatments:
1. Control: will receive no notifications about price plunge events;
2. Moral suasion: will receive a notification encouraging them to take advantage of hours with green electricity to charge their EV;
3. Low discount group: will receive notifications informing them that they will receive a discount of 15% during price plunge event hours; and
4. High discount group: will receive notifications informing them that they will receive a discount of 40% during price plunge event hours.

Experimental design

See analysis plan.

Design details

Randomization method
Randomisation will take the form of a “natural” randomisation. Specifically, Electroverse customers are assigned a unique user ID when they first join the platform; this ID is a positive integer. While these IDs are not explicitly random, they are sequential with respect to the join date of the customer. Customers will be assigned to one of four groups based on the modulo operation applied to their user ID: user_id mod 4 This calculation assigns each customer to a group based on the remainder when their user ID is divided by 4 (i.e., 0, 1, 2, or 3). This randomisation method naturally creates balanced groups across customers’ join dates.
Randomization unit
user id
Clusters
The cluster is the user id, because there will be ~15 events (and thus up to 15 outcomes per user). We have approximately 120,000 users.
Observations
15 events * 120,000 users = 1,800,000 user*events
Arms
Discussed above: ~15 events per user (see analysis plan for details)
Primary outcomes
Charging: We will analyse whether the customer has charged during Price Plunge Events. This is a binary variable (1 = charged; 0 otherwise). Electricity Consumption in kWh: We will analyse the electricity consumption (kWh) of participating customers during and around Price Plunge Events. This is a continuous variable, truncated at 0.
Secondary outcomes
Charging outside of events (both as a binary outcome and in terms of kWh)

Cited by

No citation data available for this registration yet.

Keywords