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.

232<br />

<strong>the</strong> employment equation, and <strong>the</strong> final panel considers <strong>the</strong> effect <strong>of</strong> including both age<br />

and CEP referrals in <strong>the</strong> employment equation as well as <strong>the</strong> participation equation.<br />

Table A2.6 in <strong>the</strong> appendix shows <strong>the</strong> effect on <strong>the</strong> weighted model <strong>of</strong> changing <strong>the</strong><br />

specification by including age in <strong>the</strong> employment equation. The first column in <strong>the</strong> first<br />

new results panels in Table 7.2 gives <strong>the</strong> lack <strong>of</strong> results for <strong>the</strong> unweighted equation – it<br />

proved impossible to estimate <strong>the</strong> unweighted equation with age in <strong>the</strong> employment<br />

equation. This is interpreted as indicative that because <strong>the</strong> selection arising from data loss<br />

is unaccounted for, <strong>the</strong>re is misspecification when age is included in <strong>the</strong> unweighted<br />

equation. In <strong>the</strong> second column <strong>of</strong> this panel is shown <strong>the</strong> weighted Heckman bivariate<br />

probit results where age is in <strong>the</strong> employment equation and participation equation, but<br />

CEP referrals remains as an exclusion restriction. The weighted specification is estimable.<br />

However, age does not have a reasonably sized coefficient nor is it statistically significant<br />

in <strong>the</strong> employment equation. The size <strong>of</strong> <strong>the</strong> SYETP coefficient is slightly affected by<br />

this alteration to <strong>the</strong> specification – with <strong>the</strong> size increased, although remaining<br />

statistically insignificant at conventional test levels.<br />

Now, <strong>the</strong> same form <strong>of</strong> specification change is shown in panel 2 <strong>of</strong> Table 7.2, with CEP<br />

referrals dropped from <strong>the</strong> exclusion restriction and included in <strong>the</strong> employment equation<br />

as well as <strong>the</strong> participation equation. Again, Table A2.7 in <strong>the</strong> appendix shows <strong>the</strong> effect<br />

on <strong>the</strong> weighted model <strong>of</strong> changing <strong>the</strong> specification by including CEP referrals in <strong>the</strong><br />

employment equation. The first column <strong>of</strong> <strong>the</strong> second panel again shows no results for<br />

<strong>the</strong> unweighted model, as again this specification proved impossible to estimate for <strong>the</strong><br />

unweighted data. In <strong>the</strong> second column is shown <strong>the</strong> weighted Heckman results where<br />

CEP referrals are in <strong>the</strong> employment equation and participation equation, but age remains<br />

as an exclusion restriction entered only in <strong>the</strong> participation equation. The consequences<br />

are similar to those found for age – <strong>the</strong> equation is estimable, and <strong>the</strong> coefficient <strong>of</strong> CEP<br />

referrals is not <strong>of</strong> reasonable size and it is not statistically significant, but <strong>the</strong> change to<br />

specification raises <strong>the</strong> size <strong>of</strong> <strong>the</strong> SYETP coefficient although <strong>the</strong>re is no gain in<br />

statistical significance.

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

Saved successfully!

Ooh no, something went wrong!