Getting Unemployed Youth into Jobs: Lessons for New Zealand from - - PowerPoint PPT Presentation

Getting Unemployed Youth into Jobs: Lessons for New Zealand from Jobs: Lessons for New Zealand from experiments in middle income p countries

David McKenzie World Bank David McKenzie, World Bank

The problem The problem

30

The Problem: Unemployment Rates Around the World in 2011

25 30 20 10 15 Youth 5 10 Youth Adult

Source ILO Global Employment Trends 2012

Why might youth have such trouble finding employment?

k f k i d d

• Lack of work experience and untested – so

employers find it difficult to assess quality.

• Lack of soft skills – employers complain

graduates lack key soft skills like how to work g y well in teams, dress and behave in professional manner, etc. p ,

• Skills mismatch – may have dropped out of

school or trained in areas with low labor school, or trained in areas with low labor demand

Policy response Policy response

• Whole range of different active labor market

policies designed to try and help youth find jobs

– Training programs – Employment centres p y – Internship/Job experience programs – Hiring subsidies etc – Hiring subsidies, etc.

(and of course real solution may lie in policies to i l l b d d i stimulate labor demand – private sector development policies).

But do any of these work? But do any of these work?

• Consider a training program run by a

Government for unemployed youth. p y y

• The program notes that 50% of all people

trained were employed a year later trained were employed a year later.

• Is this good or bad?

The problem of knowing whether our programs worked?

Th l d f h th h i t

• The revealed preference approach: youth are choosing to

take these programs, they tell us afterwards they thought they were good – so they must be

– But might just reflect low opportunity cost of time/desperation.

• The before‐after approach: none of them were employed

before 50% are afterwards so program is big success before, 50% are afterwards, so program is big success.

– But maybe they would have found jobs anyway

• The difference‐in‐differences approach: compare them to

unemployed youth who didn’t go through the program:

– Suppose we find only 30% of those who didn’t go through program are employed afterwards, vs 50% for those trained p g p y , % – This might just reflect that more motivated people are more likely to take the course, and also to find jobs afterwards.

Randomized experiments Randomized experiments

i id k l d

• Basic idea: take 1000 unemployed

– Use a random number generator to randomly choose 500 of them to be the Treatment group (offered training) – The other 500 are the control group (not offered training) – Two groups should be comparable, with only difference due to chance

h l bl h

• So comparing outcomes should enable us to see what

would have happened.

Illustration 1: Jordan Illustration 1: Jordan

• At the time of graduation, 93% of female

community college graduates in Jordan say y g g y they want to work

• But 16 months later only 23% are employed
• But 16 months later, only 23% are employed

Request to try out different policies to help them find jobs. Jordan NOW (New Opportunities for Women) Jordan NOW (New Opportunities for Women) pilot program.

Obtaining the Experimental Sample Obtaining the Experimental Sample

8 i bli i ll

• 8 main public community colleges
• Baseline surveys taken in July 2010 of all second‐year students

in these colleges giving data on 1755 female students before in these colleges, giving data on 1755 female students, before students had taken their final examinations.

• In August 2010 this was then merged with administrative data
• n examination results gave 1395 who had passed
• Randomly selected 1350 of these to be experimental sample

Who are these graduates? Who are these graduates?

• Typical age 20‐22
• 13% married at baseline; only 16% ever

13% married at baseline; only 16% ever worked

The Interventions The Interventions

W S b id

• Wage Subsidy

‐ Graduate given a job voucher they could take to firm when looking for work looking for work ‐ Voucher would pay the firm 150 JD ($225) per month for up to 6 months if they hired worker (= minimum wage) Fi h d t b l ll i t d h b k t d ‐ Firm had to be legally registered, have a bank account, and give job offer in writing ‐ Monthly monitoring to ensure worker still employed. y g p y ‐ If worker leaves firm before 6 months up, they take the voucher with them – can apply remaining months at another firm another firm. ‐ Valid for max of 6 months in 9 month window (Oct 2010‐ Aug 2011).

Why might wage subsidies work? Why might wage subsidies work?

h b idi h l ff b

• short‐term subsidies may have long‐term effects by

raising the productivity of youth through work (Bell et al 1999 et al., 1999

• may encourage employers to take a chance on hiring

inexperienced untested workers (World Bank 2006) inexperienced, untested workers (World Bank, 2006).

• May provide youth with the crucial experience

needed to find other jobs needed to find other jobs

• Might give youth confidence to approach employers

(Galasso et al 2004) (Galasso et al, 2004).

Employability skills training Employability skills training

• Training course of 45 hours (5 hours/day for 9 days) on key soft skills
• Training course of 45 hours (5 hours/day for 9 days) on key soft skills

employers want graduates to have

• Provided by BDC, local NGO with widespread local name recognition

and good reputation for skills training and good reputation for skills training.

• covered effective communication and business writing skills (e.g.

making a presentation, writing business reports, different types of correspondence) team‐building and team work skills (e g correspondence), team‐building and team work skills (e.g. characteristics of a successful team, how to work in different roles within a team), time management, positive thinking and how to use this in business situations, excellence in providing customer service, , p g , and C.V. and interviewing skills.

• Sessions were based on active participation and cooperative learning

rather than lectures, with games, visual learning experiences, group g g p g p exercises, and active demonstrations used to teach and illustrate concepts

Why might this work? Why might this work?

• Growing evidence that non‐cognitive or soft skills are

important for employment and a range of other life

• utcomes (e.g. Bowles et al, 2001; Heckman et al.,

2006).

• May enhance employment prospects by giving youth

better skills and confidence for looking for jobs and by making them more productive in their first months in the job by reducing the amount of time firms need to spend training them on the basics of working in a business environment.

Experimental Design Experimental Design

• 1350 students randomly allocated into one of

four groups: g p

– 450 into a Control group 300 offered just the wage subsidy – 300 offered just the wage subsidy – 300 offered just the soft skills training – 300 offered both the wage subsidy and soft skills training.

Isn’t this unfair? Isn t this unfair?

G t ti f l t d i ti i

• Gut reaction of some people to randomization is

that it is unfair that we are giving a program to some people and not others some people and not others.

• But

– Funding and training capacity limited – we did not – Funding and training capacity limited – we did not have funds to give to everyone – so this way everyone gets the same chance of being chosen. – Funding has an opportunity cost – before spending lots of public money on such a program, want to know whether or not it actually works – it is unfair to whether or not it actually works it is unfair to taxpayers to spend lots of their money on something that doesn’t work.

Randomization ensures similar looking groups

Voucher Training Voucher & Control Only Only Training Group Table 2. Comparison of Means of Baseline Characteristics by Treatment Group Only Only Training Group Stratifying Variables In Amman, Salt, or Zarwa 0.43 0.44 0.43 0.44 Tawjihi score above median 0.55 0.55 0.55 0.55 Low desire to work full time 0.41 0.41 0.41 0.41 Is allowed to travel to the market alone 0.51 0.51 0.51 0.51 Other Baseline Variables Age 21.2 21.1 21.1 21.3 Age 21.2 21.1 21.1 21.3 Married 0.14 0.16 0.12 0.13 Mother Currently Works 0.07 0.06 0.08 0.06 Father Currently Works 0.59 0.61 0.57 0.53 i l k d Has Previously Worked 0.15 0.18 0.16 0.16 Has a Job Set Up for After Graduation 0.05 0.08 0.10 0.08 Has Taken Specialized English Training 0.31 0.26 0.26 0.30 Household Owns Car 0.62 0.66 0.62 0.64 Household Owns Computer 0.72 0.75 0.74 0.70 Household Has Internet 0.28 0.18 0.26 0.26 Prefers Government Work to Private Sector 0.82 0.81 0.79 0.81 S l Si 299 300 299 449 Sample Size 299 300 299 449 Note: The only statistically significant difference across groups is internet access which is significant at the 10% level.

Timeline Timeline

li l

• Baseline – July 2010
• Graduation – August 2010

g

• Soft skills training: Sept‐Nov 2010
• Voucher period: Oct 2010 Aug 2011
• Voucher period: Oct 2010‐Aug 2011
• Midline survey: April 2011
• Endline survey: December 2011

Proportion of Female Graduates Employed in April 2011 (while voucher is still active) 0 57 0.58

0.60

April 2011 (while voucher is still active) 0.57 0.58

0.50 0 30 0.40 0 20 0.30

0.21 0.18

0.10 0.20 0.00 0.10 ‐0.10 Voucher Training Both Control

Proportion of Female Graduates Employed in December 2011 (three months after all vouchers

0.60

December 2011 (three months after all vouchers ended)

0.50 0.60 0.40 0 26 0.30 0.26 0.24 0.25 0.23 0 10 0.20 0.00 0.10 ‐0.10 Voucher Training Both Control

0.60

Proportion Employed in April vs. December 2011 0.57 0.58

0.50 0.40 0.30

0 21 0.26 0.24 0.25 0.23

0.20

0.21 0.18

0.10 0.00 ‐0.10 Voucher Training Both Control

Modest benefits Modest benefits

h did i l b k i (b

• Voucher did increase labor market experience (by

about 2 months), and labor force participation t (b b t 10% t dli ) b t h d rates (by about 10% at endline), but had no lasting impacts on employment over period we looked at looked at.

• Soft skills did improve mental health/subjective

llb i if l ff wellbeing, even if no employment effect

• Raises two key questions:

– Why not? – Is this a surprising result?

Why not more impact? Why not more impact?

S f h ld h f d

• Some of the young women would have found

jobs anyway – so subsidies for them at most d thi speed up this process.

• Main reason firms give for firing workers is that

they aren’t productive enough – high minimum wage means these workers don’t get employed.

• Firms that hired young women were from sectors

that already typically do so – didn’t succeed in breaking down stereotypes/gender segregation, so little learning by firms.

Is this a surprising result? Is this a surprising result?

• Expectations elicitation exercise

– While presenting these results at World Bank, p g , InterAmerican Development Bank, University of Virginia, Paris School of Economics, and in Jordan, g , , , solicited expectations about impacts. – Also did this through World Bank’s Development Also did this through World Bank s Development Impact blog.

60

Midline Voucher

40 50 ncy 10 20 30 Freque ‐12 ‐9 ‐4 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 More Expected ITT of Voucher 60

Midline Training

30 40 50 uency 10 20 30 Frequ ‐12 ‐9 ‐4 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 More Expected ITT of Training

60

Endline Voucher

40 50 ency 10 20 30 Freque 10 Expected ITT of Voucher 70

Endline Training

40 50 60 70 ency 10 20 30 Freque ‐12 ‐9 ‐4 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 More Expected ITT of Training

Example 2: Turkey Example 2: Turkey

T ki h E l A (ISKUR)’ i i

• Turkish Employment Agency (ISKUR)’s main active

labor market policy for the unemployed V ti l T i i

• Vocational Training courses
• Rapid expansion of courses – from 25,000 people

taking them in 2008 to 200 000+ in 2010 taking them in 2008, to 200,000+ in 2010.

• ISKUR contracts with private and public providers to

provide training courses in a range of occupations provide training courses in a range of occupations.

• Courses free to participants + they get small stipend
• ISKUR wants to know

do these work in getting people

• ISKUR wants to know – do these work in getting people

employed?

Details Details

di id l l ll d k i

• Individuals only allowed to take one course in

a 5 year period

• Courses are frequently over‐subscribed (many

more people apply for them than the courses p p pp y can take)

• This oversubscription provides a good basis

This oversubscription provides a good basis for randomized evaluation – just randomly choose who gets in the course from among choose who gets in the course from among all the qualified individuals who apply.

Design features Design features

• Evaluation sample of 5,700 individuals

applying for 130 courses in 23 provinces pp y g p

• Training courses began Oct‐Dec 2010.

M di i 320 h i l h

• Median course is 320 hours in length

The Sample: Oversubscribed but i i i Diverse Training Courses

Computer /Computer programming 24 2 3 2 2 2 2 1 1 1 1 Accounting professionalist/Computerized Accounting Babysitter Cashier Foreign Trade and Customs Professional 4 3 3 Fitter(natural gas)/Plumbery Old and Sick People Nurse Welder, Gas Arc Retailing and Merchandising/Salesperson 7 5 Retailing and Merchandising/Salesperson Cook Weaver, carpet‐ehram Modelist/Stilist C iff d H i /Ski d b t 38 5 3 Coiffeur and Hair care/Skin care and beauty Operators (forklift/sewing machine) Medical Secretary Human resources Management 6 9 6 Manufacturer, furniture Applied Basic Electronics/Electronic technicians Finalcut Waiter, service Moulder, grouting Mechanic, maintenance

83% of participants think course will raise their probability of employment

.025

Expected increase in probability of employement from course

.02 .015 ensity .01 D .005

• 100
• 50

50 100 Expected Change in prob. employed

But 1 year after courses ended But 1 year after courses ended

0.35 0.4 0.25 0.3 0.15 0.2

Treatment Control

0 05 0.1 0.05 Employed Formally Employed p y y p y

Modest impact Modest impact

i i i lik lih d f b i l d

• Training increases likelihood of being employed
• r formally employed by 1‐2 percentage points,

l ti t f 36% ( l d) 26% relative to mean of 36% (employed) or 26% (formally employed).

• Likewise see very little in the way of impact on

earnings conditional on being employed.

• Impacts are much lower than participants

expected (suggesting revealed preference argument may be misleading) and than ISKUR staff expected.

Looking beyond overall impact Looking beyond overall impact

• Evaluation can help us understand which

types of courses seem to have most effect and yp types of people course has most impact for.

0 35

Proportion Formally Employed by Subgroup

0.3 0.35 0.25 0.2 T t t 0.15 Treatment Control 0 05 0.1 0.05 Men Women High School + <High School Private Provider Public Provider Accounting

0 35

Proportion Formally Employed by Subgroup

0.3 0.35 0.25 0.2 T t t 0.15 Treatment Control 0 05 0.1 0.05 Men Women High School + <High School Private Provider Public Provider Accounting

Illustration 3: Using experiments to test not just whether program works but which components of whether program works, but which components of program work best in the Dominican Republic

• Target Population:
• Target Population:

– 16‐29 years old – Not completed secondary school Unemployed under employed or inactive – Unemployed, under‐employed or inactive – From poorest 40% of households

• Objective: improve employment opportunities of

at‐risk youth by building:

– technical skills (TS) technical skills (TS) – life‐skills (LS) – work experience (WE)

• Training provided by private institutes contracted

by Ministry of Labor

Program Components

• Technical/vocational skills (VS)

– 150 hours

g p

– $160 USD per student – Heterogeneous curriculum: Beauty, sales, tourism & hospitality, carpentry, electricity, etc

• Life skills (LS)

– 75 hours $80 USD t d t – $80 USD per student – Standardized curriculum: Self‐esteem and self‐realization, communication, conflict resolution, life planning, time management, team work, decision making, hygiene and health, etc ,

• Work Experience (WE)

– Apprenticeship in private company pp p p p y – 240 hours

• Daily stipend of US$2

y p

Questions: Opening the “Black Box” Questions: Opening the Black Box

• Should youth employment programs

emphasize “hard” or “soft” skills? p A ll d d?

• Are all components needed?
• Are employment outcomes the main area we

p y should expect to see impacts?

Experimental Evaluation

18,270 eligible applicants for 10,400 slots in 520 courses

Experimental Evaluation

Random Assignment of 35+ applicants per course to:

Random Assignment

VS +LS+WE (20) LS+WE (5) 341 courses Wait List (5+) Control (5)

Random Sample

T1 = 1560 youth T2 = 1560 youth C = 1560 youth

balanced

Impacts after 1.5 years:

Women Men

Employment

+ ‐ (VS) ( ) 0 (LS)

Active job search

+

j

+

Work hours Wages

+

Job Satisfaction

+ +

Future Expectations

+ +

Pregnancy reduction

+ NA

Pregnancy reduction

+ NA

Cost of LS+WE = 2/3 VT

Lessons for Youth Employment Policies in New Zealand

R d i d i id f ibl d

• Randomized experiments provide a feasible and

credible way of telling whether policies are having the desired impacts desired impacts

• Impacts of youth employment programs often less

than policymakers and youth themselves expect than policymakers and youth themselves expect

– Many youth will find jobs anyway – Impacts can be short‐lasting, disappearing soon after Impacts can be short lasting, disappearing soon after program ends – Underlying issue may be lack of demand for labor – so policies which try and find youth jobs without increasing private sector demand for workers will struggle.

Not just work/not work Not just work/not work

P li id

• Policy side

– Should think of these experiments as way of testing

• Which components work best, whether you need multiple

Which components work best, whether you need multiple components or just some of them

• For whom program works best – so can better target program and

think about constraints/alternatives for those who program not working

• Enabling cost‐benefit comparisons among competing policies
• Youth side: employment is only one of the impacts we

Youth side: employment is only one of the impacts we should care about, may be other important benefits we should measure

d l h l h f f k ll d – E.g. improved mental health from soft skills in Jordan – Lower risk of pregnancies among youth in DR.

Other key issues to think about in evaluating these programs

Di l t?

• Displacement?

– Are you creating more jobs, or just changing who gets them?

• Timing of effects

g

– Need measurement over longer term horizons if youth move in and

• ut of labor market a lot, and it takes a while for effects to occur.
• Learn from what you are doing

Learn from what you are doing

– Policymakers always say they have no time to wait, want to do something now – so instead of massive large‐scale program where results uncertain, why not a range of pilots, which you rigorously , y g p , y g y evaluate.

• May be easier to implement in NZ than developing countries

– More administrative data available from companies, tax records, More administrative data available from companies, tax records, unemployment records, etc. – so can in principle track outcomes over long periods at relatively low cost.