02.06.2014 Views

Evaluation of the Australian Wage Subsidy Special Youth ...

Evaluation of the Australian Wage Subsidy Special Youth ...

Evaluation of the Australian Wage Subsidy Special Youth ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

xv<br />

Abstract<br />

The job subsidy <strong>Special</strong> <strong>Youth</strong> Employment and Training Program (SYETP) was<br />

introduced in Australia with <strong>the</strong> aim <strong>of</strong> improving <strong>the</strong> movement into work. In 1984, <strong>the</strong><br />

SYETP was a flat rate subsidy <strong>of</strong> A$75 a week paid to employers for 17 weeks,<br />

equivalent in value to half <strong>the</strong> average teenage wage, and was available to youths aged<br />

15-24 who had been claiming unemployment benefits and not studying full-time for at<br />

least 4 <strong>of</strong> preceding 12 months.<br />

A review <strong>of</strong> <strong>the</strong>oretical literature indicates <strong>the</strong>y can give no pro<strong>of</strong> <strong>of</strong> employment gains<br />

for wage subsidies. The empirical ambiguity <strong>of</strong> employment gains is concluded<br />

unresolved, in both recent overseas and <strong>Australian</strong> literature. A contributing factor is <strong>the</strong><br />

insufficiency <strong>of</strong> <strong>the</strong> evaluation methods. Appraisal <strong>of</strong> <strong>the</strong> micro-evaluation evidence for<br />

SYETP and o<strong>the</strong>r <strong>Australian</strong> wage subsidies is also found to suffer <strong>the</strong>se deficiencies.<br />

The inadequacies <strong>of</strong> past analyses <strong>of</strong> SYETP contribute three <strong>the</strong>mes to address: suitable<br />

modelling <strong>of</strong> selection to account for <strong>the</strong> influence <strong>of</strong> observables or unobservables,<br />

dealing with non-response in <strong>the</strong> observational data, and appropriate control for <strong>the</strong><br />

differences between <strong>the</strong> SYETP and comparison groups.<br />

Past evaluation by Richardson (1998), modelling <strong>the</strong> Heckman selection bivariate probit,<br />

using <strong>the</strong> <strong>Australian</strong> Longitudinal Survey <strong>of</strong> <strong>Youth</strong>s 1984-1987 found a very large<br />

positive employment effect for SYETP participants 26 months after taking part. A key<br />

issue with <strong>the</strong> results is that no account <strong>of</strong> sample attrition was made. Theory indicates<br />

bias to be a potentially serious problem with results. Two evaluation methods are<br />

explored – <strong>the</strong> Heckman selection bivariate probit model, and matching methods, in<br />

particular propensity score matching. Both identify a parameter corresponding to <strong>the</strong><br />

mean effect <strong>of</strong> treatment on <strong>the</strong> treated, which can be used to decide whe<strong>the</strong>r <strong>the</strong><br />

programme leads to employment gains. However each method uses different assumptions<br />

to achieve this. Selection on unobservables is assumed by <strong>the</strong> Heckman bivariate probit,<br />

while selection on observables is assumed by matching methods.<br />

A series <strong>of</strong> empirical studies assesses a number <strong>of</strong> questions – what happens to <strong>the</strong><br />

evaluation outcome if selection is assumed to be based on observables instead <strong>of</strong><br />

unobservables; what is <strong>the</strong> importance <strong>of</strong> sample reduction to <strong>the</strong> evaluation outcome;<br />

and how sensitive is <strong>the</strong> employment impact to variation in modelling. To provide a<br />

foundation for useful comparison Richardson (1998) is first replicated successfully. The<br />

more recently popular propensity score matching method (PSM) is <strong>the</strong>n applied. The<br />

PSM results reduce <strong>the</strong> size and significance <strong>of</strong> <strong>the</strong> employment effect found. The effects<br />

<strong>of</strong> attrition are examined and <strong>the</strong>n accounted for, and <strong>the</strong> impact on evaluation discussed.<br />

The results are found to be smaller and have lower statistical significance. Correctly<br />

accounting for weights is found to be important in applying PSM.<br />

The Heckman and <strong>the</strong> PSM method both make strong but very different assumptions<br />

about <strong>the</strong> selection into SYETP. A comparison <strong>of</strong> <strong>the</strong> employment impacts found under<br />

each method is undertaken, toge<strong>the</strong>r with a discussion <strong>of</strong> <strong>the</strong> most suitable assumptions<br />

for this evaluation. The value <strong>of</strong> replication is validated and advocated. The research<br />

confirms that careful accounting for data and modelling problems is important. External

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!