What Drives Stock Return Ignorance?

How to cite

Ozyilmaz, Hakan; Merkoulova, Yulia; Veld, Chris; Zhang, Guangli. 2025. "What Drives Stock Return Ignorance?." AEA RCT Registry. AEARCTR-0015224. https://doi.org/10.1257/rct.15224-2.0

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

Hakan Ozyilmaz* , Yulia Merkoulova , Chris Veld , Guangli Zhang

* Primary investigator

Registry ID
AEARCTR-0015224
DOI
https://doi.org/10.1257/rct.15224-2.0
First registered
Last updated
Status
In development
Trial dates
2025-01-17 – 2025-05-01
Countries
Private
JEL codes
D14, G11, G41, G51, G53
Registry record
http://www.socialscienceregistry.org/trials/15224

Abstract

The neoclassic models in finance often assume that investors are capable of forming accurate expectations about stock market returns. However, recent literature (Merkoulova and Veld, 2022a) has shown that many individuals appear to lack the necessary knowledge or understanding to form reasonable expectations, and in turn, make informed financial decisions. To develop effective policy interventions and financial education programs, it is important to understand the fundamental mechanisms behind these failures. The project aims to conduct an assessment of alternative hypotheses about what drives stock return ignorance with the ultimate goal of offering valuable insights on how to address and remedy the issue. Specifically, we test the role of access to information and statistical sophistication. Additionally, we test if the forecast horizon influences stock return ignorance and estimate a simple model of recency and salience in attention allocation.

Experimental design

The design consists of a main experiment and a follow-up survey.

The main experiment consists of four treatments: Baseline, Information, Statistical Reasoning, Statistical Reasoning-Alternative. These four treatments are assigned between subjects.

In each treatment, subjects report their subjective return beliefs on the stock market return over the next year and the next 10 years. In particular, subjects report the mean, 10th, and 90th percentiles of their subjective stock market return beliefs.

In Baseline, subjects do not receive any historical stock return information before reporting their beliefs.
In Information, subjects receive the annual return on S&P500 in the last 10 years before reporting their beliefs.
In Statistical Reasoning, subjects receive the annual return on S&P500 in the last 10 years, and to help with the processing of this information, they are trained on the basic statistical properties of the provided return information before reporting their beliefs.
Statistical Reasoning-Alternative is identical to the Statistical Reasoning treatment except that the provided return information on S&P500 is calculated accounting for inflation and without assuming dividend reinvestment.

After reporting their beliefs, subjects answer a set of questions on their stock ownership in and outside of their retirement accounts, their knowledge of stock market returns in the last 10 years, their financial literacy and statistical literacy, trust, reasons for participating in the stock market.

A follow-up survey is conducted after two weeks to test if the interventions are effective in reducing stock return ignorance and influencing investment behavior. For this purpose, we asked all participants their subjective return beliefs over the next year and next ten years without providing any historical stock return information irrespective of their treatment assignment in the main experiment.

Participants receive a fixed payment of 4 USD for the main experiment and 2 USD for the follow-up survey corresponding to an hourly rate of 24 USD.

The sample is representative of the US population and collected on Prolific.

Edit April 2025:
An additional follow-up survey is conducted in April 2025 to explore the effect of the tariff related market turmoil on stock return ignorance. This additional follow-up survey is identical to the original follow-up survey.

Design details

Randomization method
Randomization is done by the survey tool (Qualtrics)
Randomization unit
Individual
Clusters
1300 individuals
Observations
1300 individuals
Arms
The sample is split evenly across 4 treatments
Primary outcomes
expected return ignorance, risk ignorance, overoptimism, stock return expectations over the next year and next 10 years

Cited by

No citation data available for this registration yet.

Keywords