How do Firms Hire Youth? A Demand-Side Experiment to Reduce Matching Frictions in Uganda Labour Markets

This study implements an intervention to support firms in Uganda hiring higher educated youth. All participating firms receive free access to online job posting services. Treatment groups receive additional support through a Job Description tool - a structured platform that helps firms:

1) Define key position requirements and responsibilities
2) Identify essential and preferred qualifications
3) Specify required technical and soft skills
4) Create clear, comprehensive job descriptions

The intervention varies the timing of this support:

Treatment 1 firms receive support before posting jobs
Treatment 2 firms receive support after posting jobs but before reviewing candidates

This design helps understand whether clearer job requirements can improve hiring outcomes through better applicant pools or through better candidate evaluation. The intervention targets firms hiring for positions requiring university degrees or diplomas.

2024-11-25

2025-12-01

There are three domains of outcomes of interest: (i) match quality; (ii) hiring process; and (iii) firm learning. The first domain contains the primary outcomes of interest while (ii) and (iii) are intermediate outcomes. Primary outcomes will be excluded from analysis if: (1) more than 10 percent of observations are missing (not including attrition and skips) and (2) if more than 90 percent of indicator variable responses take a single value.

Match Quality Performance:
1) Employer perception of worker skill fit with the job (scale 0-10)
2) Employer performance evaluations of worker (scale 1-5)
3) Employee job satisfaction (1-5)
4) Employee perception of skill fit (0-10)
5) Wage levels

Secondary domains include:
Hiring process metrics (time-to-hire, applicant pool quality)
o Hiring process duration
o Number of applications
o Number/share of eligible candidates
o Applicant characteristics (gender, age, education etc)

Firm adaptation (changes in hiring practices, job description quality, and assessment methods)
o Job Advert characteristics (length, present sections, detail level)
o Advertising methods (ex. online, professional networks, personal networks, news papers)

The study will measure heterogeneous effects across several key dimensions:
Firm hiring practices: comparing effects between firms that already use modern recruitment methods (online job boards, formal HR processes) versus traditional methods

Position characteristics: analyzing whether effects vary by job complexity or required qualifications

Prior hiring experience: investigating whether firms/managers with more versus less experience hiring higher educated workers show different responses to each treatment

Hiring process duration: analyzing whether structured job analysis leads to longer but potentially more effective hiring processes, and whether this varies by firm type

This heterogeneity analysis helps identify which firms benefit most from pre-posting versus post-posting support, and allow us to test mechanisms behind any treatment effects.

This study uses a 3-arm randomized controlled trial with private sector firms in Uganda that employ higher educated youth. All participating firms receive free access to online job posting services.

Treatment 1: Firms receive support through a Job Description Designer tool before posting their job openings. This helps firms clearly define and communicate job requirements in their initial job postings.

Treatment 2: Firms post their usual job descriptions, then receive the Job Description Designer support before reviewing candidates. This helps firms clarify job requirements for candidate assessment.

Control: Firms post jobs using their standard practices without additional support.

Our evaluation uses a three-arm randomized controlled trial with stratification across four key dimensions, creating 16 distinct strata (2 × 2 × 2 × 2 = 16). Within each stratum, firms are initially randomized in a 1:1:2 ratio into T1 (25%), T2 (25%), or Control (50%).

Stratification dimensions:

Sector Type (2 categories):
High Tech/New Skills Sectors (ex. Information and Communication, Financial Services, Professional/Scientific/Technical Activities)
Traditional Sectors (ex. Wholesale/Retail, Transportation, Accommodation, Real Estate, Traditional Administrative Services)

Firm Recruitment Channel (2 categories):
Direct contact outreach: Firms contacted through cold calling or physical visits
Firm solicited participation: Firms responding to online promotions or referrals

Previous Match Quality (2 categories): Based on index of employer baseline values of match quality (excluding wage measure) using the median as a cutoff
Posting History (2 categories): Based on firms' experience with online job postings

Given rolling enrollment, we implement a dynamic stratification process:
1. Initial randomization follows the 1:1:2 ratio within each stratum
2. After every 50 posts, we evaluate balance across treatment arms within strata
3. Assignment probabilities are adjusted as needed to maintain balance across the four stratification dimensions
4. This process continues until reaching the target sample of 615 posts

Not available

Randomization will be done by a computer

Firm

Yes

- Number of clusters depends on the number of job posts per firm.

615 job posts

Treatment 1: 154 posts
Treatment 2: 154 posts
Control: 307 posts