Automation and White Collar Employers

Last registered on November 15, 2024
View Trial History

Pre-Trial

Trial Information

General Information

Title
Automation and White Collar Employers
RCT ID
AEARCTR-0014788
Initial registration date
November 09, 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
November 15, 2024, 1:42 PM EST

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

Name
Davud Rostam-Afschar
Affiliation
University of Mannheim
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Eduard Brüll
PI Affiliation
ZEW
Contact Investigator
PI Name
Samuel Mäurer
PI Affiliation
University of Mannheim
Contact Investigator

Additional Trial Information

Status
In development
Start date
2024-11-11
End date
2026-05-31
Keywords
Behavior, Education, Firms & Productivity, Labor
Additional Keywords
Automation, Labor Demand, Tasks, Investment, White-Collar, Tax, Audit
JEL code(s)
E24, J23, J24, M40, O33, D83, C91
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
How do employers respond when they learn about an automation shock? We leverage randomized information provision that exogenously shifts employers’ prior beliefs about automation rates to investigate the impact of automation on employment strategies, wage structures, skill investment plans, and information acquisition. We use the tax consulting and auditing sectors as a case in point due to their clearly defined job titles and significant exposure to generative language models, which present credible automation potential in white-collar professions. We assign respondents to one of three treatment groups that receive a correction of their beliefs using official estimates of automation rates, alongside a control group. We then study to what extent an update in their beliefs about job-specific automation rates causes reassignment to new roles, wage adjustments, education investment, and information acquisition. We also assess effects on firm-level outcomes, such as profit, revenue, and cost growth, and interpret our results through the lens of a task-based automation model.
External Link(s)

Registration Citation

Citation
Brüll, Eduard, Samuel Mäurer and Davud Rostam-Afschar. 2024. "Automation and White Collar Employers." AEA RCT Registry. November 15. https://doi.org/10.1257/rct.14788-1.0
Sponsors & Partners

Sponsors

Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2024-11-18
Intervention End Date
2025-05-07

Primary Outcomes

Primary Outcomes (end points)
• Reinstatement effect (tasks and hiring plans)
• Displacement effect (tasks and separation plans)
• Productivity effect
• Factor-augmenting effect
• Employment plans
• Wage setting plans
• Investments in education and training
• Information acquisition about specific AI tools
• AI impact on profits, revenues, and costs
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
• Perceptions of AI
• Concerns about automation
• Potential for automation
• Frequency of use of AI
• Likelihood of job change/human capital investments
• Firm outcomes
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
In the tax advisory industry four job titles are common with different official automation rates from the Job-Futuromat of the Institute for Employment Research (IAB):

1. Tax consultant (Steuerberater, high-skilled, 62%)
2. Chartered accountant/auditor (Wirtschaftsprüfer, high-skilled, 57%)
3. Tax specialist (Steuerfachwirt, low-skilled, 80%)
4. Tax clerk (Steuerfachangestellter, low-skilled, 100%)

Participants are uniformly assigned to one of four groups and receive their own prior belief next to official automation rates for the following groups:
1. High-skilled job titles: own prior beliefs and 62%, 57%
2. Low-skilled job titles: own prior beliefs and 80%, 100%
3. High and low-skilled job titles: own prior beliefs and automation rates 62%, 57%, 80%, 100%
4. Control Screen: own prior beliefs

Next, respondents are asked whether they would like to update their prior beliefs.

Then, outcomes are elicited.
Experimental Design Details
Not available
Randomization Method
online computer software
Randomization Unit
tax advisor and/or auditor
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
95% of the observations are tax advisors, 5% auditors. 67% of tax advisors are self-employed.
Sample size: planned number of observations
>1,600 tax advisors and/or auditors
Sample size (or number of clusters) by treatment arms
>400 tax advisors and/or auditors
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
With 400 respondents per group, an significance level of 0.05 and a residual variance of 1, 2SLS can detect a treatment effect of 0.2 with power of 80%. With 800 respondents per group, an significance level of 0.05 and a residual variance of 1, 2SLS can detect a treatment effect of 0.14 with power of 80%.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Ethics Committee at the University of Mannheim
IRB Approval Date
2024-04-16
IRB Approval Number
EC Mannheim 8/2024
Analysis Plan