Mental Accounting for Credit: Experimental Evidence from Venezuela

Last registered on October 06, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

Mental Accounting for Credit: Experimental Evidence from Venezuela

RCT ID

AEARCTR-0016361

Initial registration date

September 30, 2025

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

October 06, 2025, 11:32 AM EDT

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

Locations

Country

Venezuela (Bolivarian Republic of)

Region

Primary Investigator

Name

Sean Higgins

Affiliation

Northwestern University

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Ray Kluender

PI Affiliation

Harvard Business School

Contact Investigator

Additional Trial Information

Status

In development

Start date

2025-09-22

End date

2026-10-03

Keywords

Behavior, Finance & Microfinance

Additional Keywords

mental accounting, credit, marginal propensity to borrow, financial technology, buy now pay later

JEL code(s)

D14, G4, G5

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Recent evidence indicates that consumers increase borrowing when their credit limits rise, even when they are far from fully utilizing their existing credit lines. This behavior remains a puzzle in household finance: why do unconstrained consumers respond to limit increases? Existing experiments find evidence for various mechanisms, including precautionary motives (Aydin, 2022; Aydin and Kim, 2024) or interpreting the limit increase as a signal about future income (Yin, 2025). Another potential mechanism is mental accounting: the idea that consumers separate financial decisions based on how money is categorized (Thaler, 1995; Hastings and Shapiro, 2018), but this hypothesis has not been tested experimentally.

We partner with Cashea, a Venezuelan buy now, pay later (BNPL) provider with two distinct credit lines: one for food and drugstore purchases, and one for clothing, electronics, furniture, and other retail goods. We will conduct a randomized controlled trial (RCT) to study the effect of increasing one or both credit lines. By comparing borrowing responses across credit lines, we will provide an experimental test of mental accounting for credit. Using administrative contract-level data, we will estimate the marginal propensity to borrow (MPB) overall and on each line in response to the randomized credit line increases.

External Link(s)

Registration Citation

Citation

Higgins, Sean and Ray Kluender. 2025. "Mental Accounting for Credit: Experimental Evidence from Venezuela." AEA RCT Registry. October 06. https://doi.org/10.1257/rct.16361-1.0

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)

Cashea users have two credit lines: a Línea Cotidiana that can be spent on food and drugstore purchases, and a Línea Principal that can be spent on clothing, furniture, electronics, and other retail products.

Our experiment will include 210,000 “active” Cashea users as of August 1, 2025. We define active users as users who (1) have made a purchase on both credit lines in the 40 days preceding the sample definition date and (2) made their first purchase using Cashea at least 62 days before the sample definition date.

We will randomly assign users to receive a credit limit increase on the Línea Cotidiana, the Línea Principal, both lines, or neither line (control). Among treated users, we will randomize whether the limit increases by 50% or 100%. All limit increases will be permanent, though credit limits will otherwise evolve (from the new levels) as normal based on user behavior. Treated users will be notified of changes to their credit limit(s) via push notifications and emails. To isolate the effect of limit increases from the effect of the messaging, users in the control group will receive push notifications and emails reminding them of their credit limits at the same time as the treatment groups.

Intervention Start Date

2025-10-03

Intervention End Date

2026-10-03

Primary Outcomes

Primary Outcomes (end points)

Purchases (total across both lines, and separately by line)
Balance (total across both lines, and separately by line)

Primary Outcomes (explanation)

Measured in USD. The outcome variables will be in either levels or changes depending on the specification (see attached pre-analysis plan for more details).

Secondary Outcomes

Secondary Outcomes (end points)

We will also test repayment behavior, measuring installments past due and in default (total across both lines, and separately by line).

Secondary Outcomes (explanation)

Measured both in USD and using a binary variable equal to 1 if the amount is positive.

Experimental Design

Cashea will automatically adjust the credit lines in the Cashea app and notify users of treatment. Treated users will receive a credit limit increase of 50% or 100%, and this increase will apply to the Línea Principal only, the Línea Cotidiana only, or both lines. The sample will be evenly split across the treatment arms and control group.
o Treatment 1: 50% limit increase to the Línea Cotidiana only
o Treatment 2: 100% limit increase to the Línea Cotidiana only
o Treatment 3: 50% limit increase to the Línea Principal only
o Treatment 4: 100% limit increase to the Línea Principal only
o Treatment 5: 50% limit increase to both lines
o Treatment 6: 100% limit increase to both lines
o Control: No limit increases

Experimental Design Details

Not available

Randomization Method

Randomization done by a computer (Stata script)

Randomization Unit

Individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

210,000 users

Sample size: planned number of observations

210,000 users

Sample size (or number of clusters) by treatment arms

Approximately 30,000 in each treatment arm and 30,000 in control

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

We used baseline administrative data from Cashea to estimate sample means and standard deviations that we use to calculate minimum detectable effects (MDEs) for our primary outcomes for pairwise comparisons. Table of MDEs: Variable / Mean (USD) / Standard deviation (USD) / MDE (USD) / MDE as % of baseline mean Total purchases / 224.3 / 193.3 / 4.4 / 2.0 Purchases on Línea Cotidiana / 113.8 / 114.9 / 2.6 / 2.3 Purchases on Línea Principal / 110.5 / 136.1 / 3.1 / 2.8 Total balance / 79.6 / 74.6 / 1.7 / 2.1 Balance on Línea Cotidiana / 29.3 / 29.8 / 0.7 / 2.3 Balance on Línea Principal / 50.3 / 60.6 / 1.4 / 2.8

Supporting Documents and Materials

IRB