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.

142<br />

Table 4.8 Employment effects <strong>of</strong> Heckman versus PSM<br />

Heckman<br />

PSM<br />

selection bivariate probit<br />

Employment effect 0.264 93 0.18<br />

T statistic (2.85)** (2.74)**<br />

PSM: one-to-one nearest-neighbour within-caliper (0.001) matching with replacement.<br />

A number <strong>of</strong> factors are thought to be behind <strong>the</strong> difference in <strong>the</strong> two estimates. The<br />

PSM argues that it controls better for differences in individual characteristics that arise in<br />

<strong>the</strong> SYETP and comparison group. The differences between <strong>the</strong> groups were found to be<br />

large, and so <strong>the</strong> difference in part may be due to this.<br />

The participation equation was identical for both modelling methods. The importance <strong>of</strong><br />

this is now examined.<br />

In <strong>the</strong> Heckman method, <strong>the</strong> participation equation is explicitly modelled to provide a<br />

variable that is <strong>the</strong>n used to control for <strong>the</strong> part <strong>of</strong> unobserved variation <strong>of</strong> <strong>the</strong><br />

employment equation that is correlated with <strong>the</strong> unobserved variation in <strong>the</strong> participation<br />

decision. There is some evidence to support <strong>the</strong> need to model accounting for selection<br />

bias in <strong>the</strong> SYETP process. For example, Stretton and Chapman (1990) p42 in discussing<br />

programme entry conclude that it “…does involve some selection bias as <strong>the</strong> CES selects<br />

those eligible clients who are judged to be ‘job ready with assistance´ for referral ...<br />

Employers <strong>the</strong>n select <strong>the</strong>ir subsidised employee from among a number <strong>of</strong> referrals made<br />

by <strong>the</strong> CES.” Although Stretton and Chapman (1990) were describing <strong>the</strong> selection for<br />

<strong>the</strong> later Jobstart programme, a very similar process existed for SYETP, as can be seen<br />

described in <strong>the</strong> earlier Chapter 2. In <strong>the</strong> SYETP administrative data it was found that<br />

gaining referrals and gaining employment were closely linked in a selection process<br />

[Wielgosz (1984), Aungles and Stewart (1986)]. As <strong>the</strong>re is no variable indicating <strong>the</strong><br />

administrative value <strong>of</strong> whe<strong>the</strong>r individuals were ‘job- ready-with-assistance’, <strong>the</strong>n this is<br />

93 For SYETP in <strong>the</strong> employment equation: dy/dx =0.0325919 and mean for SYETP is 0.081060. The mean<br />

SYETP translates to 8.1 per cent. The marginal effect calculated at <strong>the</strong> mean is <strong>the</strong>n 0.26418. This is<br />

interpreted as a 26 per cent increase in employment. Estimated using <strong>the</strong> mfx command in STATA7.0

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

Saved successfully!

Ooh no, something went wrong!