A Randomized Evaluation of a Skills-Diagnostic and Personalized Reskilling Program for Underemployed Young Graduates in Chile

Last registered on September 09, 2026

Pre-Trial

General Information

Title

RCT ID

AEARCTR-0019552

Initial registration date

September 02, 2026

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

September 09, 2026, 4:08 PM EDT

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

Locations

Country

Chile

Region

Primary Investigator

Name

Pablo Egana-delSol

Affiliation

UAI

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Sebastian Gallegos

PI Affiliation

UAI

Contact Investigator

PI Name

Constanza Gonzalez

PI Affiliation

UAI

Contact Investigator

Additional Trial Information

Status

On going

Start date

2026-08-26

End date

2029-09-01

Keywords

Education, Labor

Additional Keywords

reskilling; upskilling; skills diagnostic; skill mismatch; overeducation; underemployment; unemployment; active labor market policy; vocational training; job training; digital skills; employability; job matching; employer linkage; formal employment; labor informality; social security contributions; higher education graduates; early-career workers; Chile; Latin America

JEL code(s)

C93, J24, J64, J68, I26, J46

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Chile has experienced a rapid expansion in the tertiary-educated workforce — from 21.9% of the labor force with completed higher education in 2010 to 41.6% in 2025 — without a commensurate expansion in demand for skilled labor. Roughly 250,000 people aged 25–35 are underemployed, close to 40% of all underemployment in the country, while about 60% of firms report persistent difficulty finding talent. Only 44% of affiliates aged 20–35 actively contribute to the pension system.
This study evaluates Despega tu Pega (DtP), a national program combining (i) an individualized skills diagnostic delivered through the Relink platform using the MapHa skills taxonomy, across employability, digital and productivity dimensions; (ii) personalized training pathways ("skillpacks") designed to close the diagnosed gaps; and (iii) job linkage with participating firms, including the use of the Chilean training pre-contract (precontrato de capacitación) as a bridge into formal employment.
Approximately 2,200 eligible individuals aged 25–35, with a completed higher-education qualification and currently underemployed or unemployed, are randomly assigned at the individual level, in waves, to a treatment group that receives DtP, a pure control group, or a waitlist group that may be offered the program if places become available. Assignment is stratified by sex, region, time since graduation, assigned training track and detailed labor market status, and is nested within two program modalities (precontrato and becas laborales). Enrollment is continuous: eligible individuals are randomized in sequential waves as they complete registration, consent and the baseline survey, and waves continue until the 2,200 target is reached. The first wave, executed on 26 August 2026, randomized 724 individuals — approximately one third of the target — of whom 161 were underemployed and 563 were unemployed or searching for work for the first time. At the time of registration the assignment had been generated and its balance verified, but the intervention had not begun: no participant had been contacted or informed of their assignment, and no outcome had been observed.
Primary outcomes are formal employment, employment stability, monthly labor income and pension contributions, measured at 6, 12 and 24 months through participant surveys and, where access is obtained, administrative records from unemployment insurance, the pension supervisor and the tax authority. We estimate intention-to-treat effects and, instrumenting participation with random assignment, treatment-on-the-treated effects, together with a pre-specified set of conditional average treatment effects. The study provides causal evidence on a scalable, diagnostic-driven reskilling model and on the role of the private sector in reducing graduate underemployment.

External Link(s)

Registration Citation

Citation

Egana-delSol, Pablo, Sebastian Gallegos and Constanza Gonzalez. 2026. "A Randomized Evaluation of a Skills-Diagnostic and Personalized Reskilling Program for Underemployed Young Graduates in Chile." AEA RCT Registry. September 09. https://doi.org/10.1257/rct.19552-1.0

Sponsors & Partners

Partner

Name

SOFOFA Capital Humano

Type

ngo

URL

https://www.sofofa.cl/ecosistema-sofofa/capital-humano/

Experimental Details

Interventions

Intervention(s)

Intervention(s)
Participants assigned to treatment receive the full Despega tu Pega program over approximately six months, comprising three components:
(1) Skills diagnostic. An individualized assessment delivered on the Relink platform using the MapHa skills taxonomy, measuring existing capabilities and gaps across three dimensions: employability, digital skills, and productivity.
(2) Personalized training pathway. Assignment to a training track ("skillpack") built to close the gaps identified in the diagnostic. Courses run 50–80 hours, are delivered in three levels with roughly 30% asynchronous and 70% synchronous content, over about two months. Tracks cover employability (job search, job retention and progression), digital skills (digital management tools, applied artificial intelligence, content creation and editing, data handling and analysis, digital security), and productivity (self-management for sustainable work performance).
(3) Job linkage. Articulation with participating firms, early linkage opportunities and, where applicable, use of the training pre-contract (precontrato de capacitación) as a bridge into hiring. Training is financed through two modalities: precontrato and becas laborales (training grants under the Franquicia Tributaria scheme).

Intervention Start Date

2026-09-22

Intervention End Date

2029-08-31

Primary Outcomes

Primary Outcomes (end points)

Formal employment; employment stability; monthly labor income; pension system contributions.

Primary Outcomes (explanation)

Formal employment. Binary indicator equal to one if the individual holds a dependent job with a written contract and social security contributions in the reference month. Survey measure: situacion_laboral = 1 and contrato = 1. Administrative measure, where access is obtained: a positive taxable-earnings record in the unemployment insurance (Seguro de Cesantía) database in the reference month.
Employment stability. Months of continuous employment with the same employer over the period since randomization, and a binary indicator for being employed in the same job at both the 6- and 12-month follow-ups. Administrative measure: number of consecutive months with a positive contribution record with the same employer.
Monthly labor income. Gross monthly labor earnings in Chilean pesos, in real terms (deflated to the month of randomization). Analyzed in levels, winsorized at the 99th percentile, and in inverse hyperbolic sine transformation. Administrative measure: taxable earnings from the unemployment insurance records, complemented with fee-based income (boletas de honorarios) from the tax authority (SII).
Pension contributions. Binary indicator for having contributed to the pension system in the reference month, and the count of months contributed over the follow-up window. Administrative measure: Superintendencia de Pensiones contribution records.
All four primary outcomes are measured at 6, 12 and 24 months after randomization. The 12-month measurement is the primary endpoint; the 6- and 24-month measurements are reported as pre-specified secondary time horizons.

Secondary Outcomes

Secondary Outcomes (end points)

Digital skills; employability skills; productivity skills; underemployment status; alignment between education and job; program take-up and completion.

Secondary Outcomes (explanation)

Skills outcomes. Scores on the three MapHa dimensions (employability, digital, productivity), measured by re-administering the Relink diagnostic at follow-up. Reported both as standardized dimension scores and as a pre-specified summary index constructed following Anderson (2008), with inverse-covariance weighting within dimension.
Underemployment status. Binary indicator for remaining underemployed, defined following the Chilean statistical office (INE) as either (i) involuntary part-time work — usual hours below 30 per week, wanting and being available to work more hours — or (ii) qualification mismatch, that is holding an occupation whose skill requirements fall below the worker’s completed level of education.
Education–job alignment. The five-item baseline alignment module (alineacion_1 to alineacion_5), covering required education level relative to own, subjective alignment between field of study and job, use of skills acquired in study, perceived progress toward a desired job, and satisfaction with the fit between education and current work. Items 2–5 are 0–10 scales; a standardized index is the pre-specified summary measure.
Take-up and completion. Share of assigned participants who begin the diagnostic, who are assigned a track, and who complete their track. These serve as the first-stage measures for the treatment-on-the-treated estimates and are reported descriptively by arm and modality.

Experimental Design

Individual-level randomized controlled trial with three arms — Treatment, Control and Waitlist — implemented in sequential waves as participants complete registration, consent and the baseline survey.
Eligible individuals are split into two mutually exclusive lists by baseline labor market status. Those who are underemployed are randomized within the becas laborales modality. Those who are unemployed (or searching for work for the first time) are randomized jointly across two modalities, precontrato and becas laborales. Within every list-by-modality cell the assignment shares are 30% Control, 50% Treatment and 20% Waitlist; within the unemployment list, 75% of participants are allocated to the precontrato modality and 25% to becas laborales.
Assignment is stratified. Within each arm and training track, participants also receive a random priority number that determines the order in which they are contacted, so that operational contact sequencing is itself randomized and does not correlate with participant characteristics.

Experimental Design Details

Not available

Randomization Method

Randomization done in office by a computer, using Stata with a recorded and archived seed. Within each stratum, participants are assigned a uniform pseudo-random draw, sorted, and allocated to arms according to the modality-specific shares. A second independent draw assigns the contact-priority order within each arm-by-track cell. The eligibility database is frozen before the draw, the do-files and seed are archived for each wave, and balance is verified after each wave using both classical F-tests and randomization inference with 2,000 within-stratum permutations.

Randomization Unit

individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

2,200 individuals (treatment is not clustered; the unit of randomization is the individual).

Sample size: planned number of observations

2,200 individuals (treatment is not clustered; the unit of randomization is the individual).

Sample size (or number of clusters) by treatment arms

Planned, across all waves: 1,100 treatment; 660 control; 440 waitlist.
Within the unemployment list these are further split across modalities in a 75/25 ratio: 825 precontrato (413 treatment / 248 control / 165 waitlist) and 275 becas laborales (138 treatment / 83 control / 55 waitlist), for an unemployment list of 1,100. The remaining 1,100 correspond to the underemployment list under becas laborales (550 treatment / 330 control / 220 waitlist).
Randomized to date, wave 1 (26 August 2026): 724 individuals — 367 treatment, 221 control, 136 waitlist. This is enrollment in progress, not the final sample.
Composition target. The per-arm figures above assume the 2,200 target divides evenly between the two eligibility lists. In wave 1 the realized split was 161 underemployed against 563 unemployed — an underemployment share of 22%, not 50%. Because underemployment is the population this study is designed around, recruitment in later waves will be redirected toward it, with a target of 1,100 underemployed participants across all waves. Section 6.6 of the analysis plan sets out the power arithmetic behind that target.

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Power curves were constructed for sample sizes between 400 and 1,200 individuals at power levels of 0.70, 0.80 and 0.90, with α = 0.05 and two-sided tests. At 80% power the minimum detectable effects are: Formal employment: 6–8 percentage points. Employment stability at 6 months: 5–7 percentage points. Monthly labor income: 5–7% in relative terms. Pension contributions: 0.05–0.07 of the baseline mean. A complementary simulation exercise drew 100 bootstrap samples of sizes 300, 500, 800 and 1,000 from parameters observed in comparable training programs, and found power above 0.80 in all four primary outcomes for effect sizes in the range reported by recent skills-based employability interventions. These values are conservative: they do not model the residual variance reduction produced by stratification on sex, region, time since graduation, track and labor market status, so the realized MDE is expected to be somewhat smaller.

Supporting Documents and Materials

IRB

Institutional Review Boards (IRBs)

IRB Name

Universidad Adolfo Ibanez

IRB Approval Date

2026-09-30

IRB Approval Number

N/A

Analysis Plan

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

Request Information