RDP 2002-04: Labour Market Adjustment in Regional Australia Appendix B: OLS Results for Regional Employment Growth

Jeremy Lawson and Jacqueline Dwyer

June 2002

Results from an OLS Regression for Regional Employment Growth(a)
Variable Coefficient Significance
Coastal 4.58 **
Size −2.67  
Remoteness −1.11 ***
Structural change 0.93 ***
Diversity 1.38 *
Human capital 0.90 *
Agriculture −0.24 ***
Manufacturing 0.14  
Mining −0.28 **
Retail −0.91 ***
Accommodation 1.70 ***
Property 3.21 ***
Utilities −1.17 ***
Victoria −0.47  
Queensland 11.84 ***
South Australia −1.46  
Western Australia 11.29 ***
Northern Territory −1.24  
Tasmania −3.97  
Number of observations = 573
F (19,553) = 24.62
Probability > F = 0.00
Adjusted R2 = 0.439

Note: ***, **, and * represent significance at 1, 5 and 10 per cent levels.
(a) Note that the sample size in the OLS regression (573) differs from the sample size in the binomial logit model (637). We trimmed 5 per cent of observations from each tail of the distribution for the model estimated using OLS because we were concerned that outlying growth rates would unduly influence the regression results. This is not an issue in the binomial logit model because each region receives an equal weight.