Investor Confidence in Robo Advice

Last registered on August 14, 2024

Pre-Trial

Trial Information

General Information

Title
Investor Confidence in Robo Advice
RCT ID
AEARCTR-0013760
Initial registration date
August 08, 2024

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 14, 2024, 2:25 PM EDT

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

Locations

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

Request Information

Primary Investigator

Affiliation

Other Primary Investigator(s)

PI Affiliation

Additional Trial Information

Status
In development
Start date
2024-08-08
End date
2025-08-08
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The emergence of algorithm-generated financial advice challenges the traditional market for financial delegation, which strongly relies on human-generated advice. Despite the fact that such robo-advisors may provide opportunities for a large number of household investors, they may refrain from utilizing such advisors due to a phenomenon commonly referred to as algorithm aversion: Individuals typically favor a human advisor when choosing between algorithmic and human forecasters, even when they know that algorithms exhibit a better forecasting accuracy. This study investigates algorithm aversion in the context of financial delegation using a lab experiment. We exogenously manipulate the advisor type (human vs. algorithm) and measure individuals’ belief formation about the fundamental quality of recommended assets and the quality of the assigned advisor (i.e., the amount of information available to the advisor). We then observe subsequent investor decisions. Under the algorithm aversion hypothesis, we expect participants who receive advice from an algorithm to be less confident about the fundamental quality of the asset and about the quality of the advisor. Additionally, we expect this belief to affect subsequent investment decisions, such as a lower willingness to take risk and a higher propensity to switch advisors when provided with the opportunity among those in the algorithm treatment.
External Link(s)

Registration Citation

Citation
Birk, Simon and Marten Laudi. 2024. "Investor Confidence in Robo Advice." AEA RCT Registry. August 14. https://doi.org/10.1257/rct.13760-1.0
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2024-08-08
Intervention End Date
2025-08-08

Primary Outcomes

Primary Outcomes (end points)
Participants’ beliefs about the fundamental quality of stocks that they hold.
Primary Outcomes (explanation)
Does delegation influence belief formation about stock quality? We calculate the difference between participants’ estimates of the fundamental quality of their stock and the normative Bayesian posterior. We compare beliefs and belief errors of those who base their stock selection on the advice of a human and those who receive algorithmic advice. Under the algorithm aversion hypothesis we expect participants' reduced confidence about the advisors' quality to influence the perceived fundamental quality of the asset. Thus, beliefs are expected to deviate when compared between treatments. Specifically, the hypothesis entails that those who receive advice by an algorithm form more pessimistic beliefs about their stocks, relative to those who receive advice by a human.

Secondary Outcomes

Secondary Outcomes (end points)
Participants' beliefs about information available to advisors; Subsequent behavior.
Secondary Outcomes (explanation)
1. Participants' beliefs about information available to advisors: Participant are informed that the recommendation given by the advisor (either human or robo) was based on between 5 and 0 previous price levels. We elicit beliefs about how many previous price levels their advisor had access to. Under the algorithm aversion hypothesis, we expect participants to predict a lower number of observed price levels for robo advisors, relative to human advisors.

2. Subsequent behavior:
2.1. How do investor beliefs drive risk taking? The first behavioral outcome variable is the level of risk taking. After the first block, participants have the option to invest again in the same asset they were invested in before. They are invested for the next 10 investment periods of the second block. However, this time they can decide how much (from 0% to 100%) of their experimental endowment they want to allocate to the asset. The amount not invested in the risky asset will be invested in a risk-free asset with a small but certain return after 10 periods.

2.2. How do investor beliefs drive switching? The second behavioral outcome variable is switching behavior. After the first block, participants invest again for the next 10 investment periods of the second block. Depending on the treatment group, participants can switch their manager or switch the asset they were invested in. Participants can decide to either keep their advisor of block 1 or to switch and get a new advisor, which can either be a robo advisor or a human advisor.

Under the algorithm aversion hypothesis, we expect participants who receive algorithmic advice to take less risk and to switch more often relative to participants who receive human advice.

Experimental Design

Experimental Design
Participants take part in an experimental stock market that lasts 10 time periods t. In each period t, the price level of all stocks in this market either increases or decreases; a price increase is always +6% and a price decrease is always -5%. Each stock has a pre-determined fundamental quality, which is a stock’s probability of a price increase each period (randomly drawn with replacement from the set {0.1, 0.15, 0.2, 0.25, …, 0.9}), and which remains constant throughout the experiment. Participants are not informed about the true fundamental quality of the stocks.

Prior to the main experiment, we collect recommendation data. Both human and algorithm advisors receive a set of stocks and additional price information for prior periods to give recommendations on. A buy recommendation should be given for any asset with an estimated fundamental quality at or above 0.5. Thereby, we harmonize the instructions presented to both human advisors and the algorithm. The algorithm advice is based on a Bayesian posterior. We collect human advice data from real financial advisors, who are employed at a German bank. Since we expect variation in the quality of recommendations between the human and algorithm advisors, we match algorithm recommendations with equivalent human recommendations prior to the main experiment to mitigate noise.

In the main study, participants choose a stock based on advisor recommendations and remain invested in this stock for all 10 periods. They are randomly allocated to one of two main treatments; The HUMAN ADVISOR treatment and the ROBO-ADVISOR TREATMENT. In the beginning, participants select one of the recommended stocks. They solely decide based on their advisor’s recommendations, knowing their advisor type; here, they are not informed about any previous price levels. To incentivize participants of the main survey, they are informed about receiving a payment according to a fraction of the value of their assets at the end of the experiment.

The main experiment is divided into two blocks: (i) belief elicitation and (ii) changes in behavior. In block i, participants iterate through a sequence of ten belief elicitation rounds. In each round, we present them with additional price information and ask about their current belief of the fundamental quality of the asset.

Additionally, we vary how many previous price levels human and algorithmic advisors saw when selecting a stock on behalf of the client. This number of previous price levels is unknown to participants. We ask clients how many previous price periods they believe their human/algorithmic advisor saw when selecting the stock three times in the experiment (at t=0, t=5, and t=10). This gives us an approximation of the perceived decision quality of the advisor.

In block ii we collect data on subsequent investor behavior conditional on the advisor type. Participants across both treatments are allocated to one of two tasks: (1) risk taking and (2) advisor switching. Participants in the risk-taking task are requested to distribute a portion of a fictional portfolio between the current (risky) stock and another risk-free asset. Both assets exhibit different expected outcomes. Participants in the advisor switching task choose between sticking to their initial stock (as recommended in block i) and investing in a new stock out of a recommendation list including both human and algorithm advice.

In a post-survey questionnaire, we collect data on participants’ self-perceived risk preferences (Dohmen et al., 2011), financial literacy (van Rooij, Lusardi & Alessie, 2011), and socio-demographics.
Experimental Design Details
Not available
Randomization Method
Block randomization
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
No Clusters
Sample size: planned number of observations
At least 100, depending on budget constraints.
Sample size (or number of clusters) by treatment arms
At least 50 per treatment arm, depending on budget constraints.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
German Association for Experimental Economic Research e.V.
IRB Approval Date
2024-04-15
IRB Approval Number
njE9xrvK