Distrust in Financial Professionals and Robo Advisor Adoption

How to cite

Ding, Wenzhi; An, Jiafu; Wang, Xincheng. 2023. "Distrust in Financial Professionals and Robo Advisor Adoption." AEA RCT Registry. AEARCTR-0011886. https://doi.org/10.1257/rct.11886-1.0

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

Wenzhi Ding* , Jiafu An , Xincheng Wang

* Primary investigator

Registry ID
AEARCTR-0011886
DOI
https://doi.org/10.1257/rct.11886-1.0
First registered
Last updated
Status
In development
Trial dates
2023-08-07 – 2023-08-21
Countries
China ()
JEL codes
D14, C93, G23
Registry record
http://www.socialscienceregistry.org/trials/11886

Abstract

Cases of financial professionals misappropriating client assets for personal gain are rampant. In personal finance, financial institutions have introduced AI-assisted robo-advisors to optimize investments for investors using AI technology. We are trying to understand if investors will convert their distrust in financial professionals into trust in objective algorithm-based investments. We plan to collaborate with a Chinese mobile financial app company, disseminating information to users regarding occupational misappropriation by financial professionals and observing whether users increase their preference for robo-advisor products in their choice of financial products.

We plan to replace the splash screen of the experimental group's mobile app with slogans related to losses caused by human errors of financial experts. On the other hand, the control group sample will use a splash screen with sayings unrelated to human errors of financial experts. By comparing the experimental and control groups, we can determine whether AI's objectivity or financial experts' subjectivity plays a positive role in user adoption of robo-advisor portfolios.

Intervention

When the user opens the mobile application, we show the treatment on the splash screen.

The treatment consists of two parts: (1) Slogan; (2) loss amounts. Since our treated users are Chinese, we will describe the treatment in Chinese and English.

We have three types of slogans:
S1. No slogan
S2. "Research shows that financial markets have generated at least X billion human errors in the past ten years" (研究表明仅10年金融市场产生至少X0亿元人为错误)
S3. "Research shows that robo-advisors can avoid at least 10 billion yuan in human errors in the past ten years" (研究表明近10年智能投顾可避免至少X0亿元人为错误)

The treatment combinations to be tested are:

- S1 (1 group)
- S2 (2 groups, where X=1, 10)
- S3 (2 groups, where X=1, 10)

Therefore, there will be five groups in total.

Experimental design

We plan to show the users the intervented splash screen throughout the 14-day experimental period. We will randomly draw 5,000 users who logged into the mobile application within four weeks to one week before the experiment. Then we will randomize them into 5 groups. We repeat the randomization for several rounds to ensure the users are balanced regarding several pre-trial traits, such as age, gender, and asset amount, register date.

The treatments are described in the "Intervention" part.

Design details

Randomization method
Pre-treatment trait balanced sampling. The randomization is done in the office by a computer, and the group of users is pre-determined based on the randomization outcome.
Randomization unit
Individual user-level.
Clusters
2500 active users.
Observations
2500 active users.
Arms
500 active users for each treatment arm.
Primary outcomes
The net inflow investment amount into the robo-advisor portfolios.

Cited by

No citation data available for this registration yet.

Keywords