Measuring Worker Value Added

How to cite

Falco, Paolo. 2026. "Measuring Worker Value Added." AEA RCT Registry. AEARCTR-0018428. https://doi.org/10.1257/rct.18428-1.0

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

Paolo Falco*

* Primary investigator

Registry ID
AEARCTR-0018428
DOI
https://doi.org/10.1257/rct.18428-1.0
First registered
Last updated
Status
In development
Trial dates
2026-04-22 – 2026-07-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18428

Abstract

This study measures the value of workers to firms in a developing country. The project focuses on small and medium-sized garment manufacturing firms in Addis Ababa, Ethiopia.

We directly elicit managers’ willingness to accept (WTA) compensation to release specific workers for one day using an incentive-compatible Becker–DeGroot–Marschak (BDM) mechanism. This approach identifies the shadow cost of a one-day absence.

The design experimentally varies the following dimensions: (i) whether a specific or generic replacement worker is offered to the firm, and (ii) whether the absence is anticipated or unanticipated. These variations allow us to test for production complementarities, firm-specific human capital, and planning failures.

Experimental design

We interview managers in garment SMEs in Addis Ababa, Ethiopia. Each manager evaluates 9 non-managerial workers.

Managers state their WTA to release each worker. A price is randomly drawn; if price ≥ WTA, the worker is released and the firm is compensated.

Treatments:

(1) Known vs generic vs no replacement worker;
(2) Advance notice vs unanticipated absence.

For Treatment 1, randomization occurs at the worker-level within the firm. For Treatment 2, randomization occurs at the firm level.

Design details

Randomization method
Randomization is carried out by the survey software.
Randomization unit
For Treatment 1, randomization occurs at the worker-level within the firm. For Treatment 2, randomization occurs at the firm level.
Clusters
We plan to survey 150 firms and elicit WTA for 9 workers per firm.
Observations
1350 (150 firms x 9 workers)
Arms
Timing (between-firm, firm-level randomization): - 75 firms tomorrow - 75 firms in 8 days (expected, 50/50 randomization) Replacement condition (within-firm, worker-level randomization): - 450 workers with no replacement (150 firms × 3 workers) - 450 workers with a generic replacement (150 firms × 3 workers) - 450 workers with a sector-skilled replacement (150 firms × 3 workers) Combined cells (6 total, worker-level, expected): - 225 workers tomorrow × no replacement (75 firms × 3) - 225 workers tomorrow × generic replacement - 225 workers tomorrow × sector-skilled replacement - 225 workers in 8 days × no replacement - 225 workers in 8 days × generic replacement - 225 workers in 8 days × sector-skilled replacement
Primary outcomes
Worker-level WTA for a one-day absence We'll study heterogeneity by key firm and worker characteristics, including size, profitability, gender, age, education, ethnicity, and tenure.

Cited by

No citation data available for this registration yet.

Keywords