The Value of Public Employment: A Survey Experiment

How to cite

Ferreira, Pedro de Souza. 2026. "The Value of Public Employment: A Survey Experiment." AEA RCT Registry. AEARCTR-0019202. https://doi.org/10.1257/rct.19202-1.1

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

Pedro de Souza Ferreira*

* Primary investigator

Registry ID
AEARCTR-0019202
DOI
https://doi.org/10.1257/rct.19202-1.1
First registered
Last updated
Status
In development
Trial dates
2026-07-27 – 2027-07-31
Countries
Private
JEL codes
J45, J31, J64, D83, C93
Registry record
http://www.socialscienceregistry.org/trials/19202

Abstract

This study uses a survey experiment to measure how students and workers in Brazil perceive and value public employment relative to private employment. Participants report baseline beliefs about wages, job stability, and probabilities of entering public and private employment under different search strategies. They are then randomly assigned to receive aggregate labor-market information about one of these topics or to a no-information control condition. After the information module, the survey elicits expected wages, reservation wages, cross-sector switching thresholds, career-path probabilities, willingness to pay for additional information through a Becker-DeGroot-Marschak (BDM) mechanism, and choices over hypothetical jobs in a discrete choice experiment (DCE).

The main hypotheses are that public employment is valued because of a combination of perceived wages, perceived stability, perceived entry probabilities, and non-wage amenities; that information changes beliefs, downstream choices, and behavioral measures of demand for career-related information and tools; that demand for information depends on what participants already know and what information they have already received; and that the DCE can recover willingness to pay for public-sector employment, lower dismissal risk, remote work, teamwork, regular hours, and contract type. The analysis will estimate intention-to-treat effects of information on beliefs, reservation wages, crossing thresholds, BDM willingness to pay, and DCE choices, with pre-specified heterogeneity by family public-sector exposure, state, academic program or field, previous experience with concursos publicos, and selected exploratory moderators such as risk tolerance, family norms, mobility constraints, and salary-privacy norms.

The survey also includes two behavioral measures of downstream demand: whether participants click a link to sign up to receive forthcoming material related to public-sector exams, and whether they choose to enter a lottery for a three-month subscription to either QConcursos, a platform for public-sector exam preparation, or LinkedIn Premium Career, a private-sector-oriented professional platform.

Intervention

The intervention is a randomized information-provision experiment embedded in an online survey about public and private employment in Brazil. After reporting baseline beliefs about public- and private-sector wages, job stability, and entry probabilities, participants are randomly assigned to receive one short module of objective, aggregate labor-market information on one of three topics — wages, job stability, or probability of entry — or to a no-information control arm. The information is presented as group-level statistics (drawn from Brazil's PNAD Continua for comparable groups) and states explicitly that the figures describe groups and do not determine what will happen to any individual. After the information module, participants complete an incentivized Becker-DeGroot-Marschak (BDM) exercise eliciting willingness to pay for one additional information topic (which is drawn randomly among one of the three topics not yet displayed), and a discrete choice experiment (DCE) in which they choose between pairs of hypothetical jobs that vary in wage, sector, dismissal risk, remote work, teamwork, work pace, and contract type.

At the end of the survey, participants are also offered two optional behavioral choices: they may click a link to sign up to receive forthcoming material related to public-sector exams prepared by ENAP (Escola Nacional de Administração Pública), and they may choose whether to enter a lottery for a three-month subscription to either QConcursos or LinkedIn Premium Career. These are outcomes used to measure downstream demand for public-sector-related information and for career-search tools -- not additional randomizations.

Experimental design

This is an individual-level online survey experiment. The main randomization (R1) assigns each participant to one of four arms: wage information, stability information, entry-probability information, or a no-information control. A second randomization (R2) determines which not-yet-shown information topic is offered in an incentivized BDM exercise and the random price that determines whether it is delivered. A third randomization (R3) generates six pairs of hypothetical job profiles per participant, randomizing job attributes subject to logical restrictions that ensure the job profiles are credible. The main analysis estimates intention-to-treat effects of R1 assignment (relative to control) on beliefs, reservation wages and crossing thresholds, BDM willingness to pay, DCE choices, and the pre-specified behavioral outcomes measured at the end of the survey, with pre-specified heterogeneity by parental public-sector employment, state, program/field, prior concurso experience, and baseline beliefs. Risk aversion will also be considered as a heterogeneity variable. Treatment is not clustered. Respondent-level standard errors are robust to heteroskedasticity and DCE standard errors are clustered by respondent. The study is fielded to two samples — UnB undergraduates and an online sample of Brazilian workers/job-seekers — analyzed separately and, secondarily, pooled with sample fixed effects and treatment-by-sample interactions.

Design details

Randomization method
Randomization done by a computer. Assignment is generated within the online survey instrument (Qualtrics) using its embedded-data random-number generator, which draws pseudo-random integers at the start of each respondent's session. The main information arm is a random integer 0–3 (information_group); the BDM offered topic and price and the DCE profile attributes are set by further within-survey random draws. The randomization uses simple (unstratified) individual random assignment.
Randomization unit
Individual participant (respondent). All treatment randomizations — the main information arm, the BDM topic and price, and the DCE profile attributes — are at the individual level, in both the student and worker samples.
Clusters
Not clustered — the unit of randomization is the individual, so there is effectively one individual per "cluster". In cluster terms: up to 500 undergraduate students at the University of Brasilia (student sample) plus a planned online sample of approx. 500 adult workers / job-seekers in Brazil = up to approx. 1,000 individuals.
Observations
Up to approx. 1,000 individuals: <=500 undergraduates at the University of Brasilia and <=500 online Brazilian workers/job-seekers. Randomization and the primary respondent-level analyses are at the individual level. For the profile-level DCE analysis only, each respondent contributes 6 tasks x 2 profiles = 12 job-profile records (up to ~12,000 profile observations), but the unit of randomization and of the main outcomes is the individual.
Arms
Four arms with equal assignment probability. Under the simple (unstratified) randomization implemented in Qualtrics, realized cell sizes are approximate rather than exactly: Control (no information): 125 students, 125 workers; Wage information: 125 students, 125 workers; Stability information: 125 students, 125 workers; Entry-probability information: 125 students, 125 workers. The BDM topic/price randomization (R2) and the DCE profile randomization (R3) are nested within every arm and are not separate allocation arms. The BDM offered topic is balanced across not-yet-shown topics.
Primary outcomes
Primary outcomes fall into four families, all measured within the same survey after the R1 information module: 1) Beliefs and belief updating: own expected prospects about wages, one-year job-loss probability, and entry probability under different search strategies. 2) Reservation wages, crossing thresholds, and subjective probabilities: reservation wage in the public sector (RW_P), reservation wage in the private sector (RW_R), the gap RW_R − RW_P and its log version, the minimum public salary required to leave private jobs anchored at R$15,000 and R$20,000, and the subjective probabilities of future career paths (from zero to 100), in particular: private sector jobs, public sector jobs, entrepreneurship, pursuing further education, moving to a different city, and moving abroad). 3) Willingness to pay for information (BDM): the bid (maximum amount given up) for additional wage, stability, or entry-probability information. 4) DCE choices and implied willingness to pay: the chosen job profile in each of six tasks, and implied monthly willingness to pay for public/statutory employment, lower dismissal risk, remote work, teamwork, respecting regular hours, and contract type (CLT vs PJ, and the public/statutory bundle). Treatment effects are estimated as intention-to-treat effects of information assignment relative to control, and as topic-specific effects on the corresponding belief family (assigned a topic vs not).
Secondary outcomes
1) Behavioral demand outcomes: (i) whether the participant clicks the optional link to sign up to receive forthcoming ENAP material related to public-sector exams; (ii) whether the participant chooses to participate in the subscription lottery; (iii) whether the participant chooses QConcursos; and (iv) whether the participant chooses LinkedIn Premium Career. The latter two outcomes are defined on the full randomized sample, conding as zero participants that chose not to participate in the lottery. 2) Open-text career descriptions and information evaluations: imagined private job, imagined public/statutory job, and ideal-job descriptions; perceived clarity, usefulness, surprise, and specificity of the information shown.

Cited by

No citation data available for this registration yet.

Keywords