6.2  Short and long regressions: Two-way split in youth ethnicity

There are two broad issues that are addressed in the regression analysis in this section. First, although we know that Maori in the CHDS have fewer qualifications and less work experience than non-Maori, these same observable differences may not exist among the subsample of workers. In other words, the gap in measured productivity characteristics between the ethnic groups might be narrowed (and even reversed) once we focus on only youth who were employed at the time of the interview. Second, there may be differences in the ‘potential wages’ that face all individuals in the sample. It is possible that the sample selection process that underlies the employment outcome systematically overstates the wages that face Maori relative non-Maori. We already know from our sample that only 55.0% of Maori were working at the interview, compared to 70.7% of non-Maori. We want to control for some aspects of this sample selection process in order to estimate the extent to which the wages that face all individuals vary between observationally equivalent Maori and non-Maori.

Table 16 reports the results from the short and long regressions on wages using the two-way split in youth ethnicity. The dependent variable is the natural logarithm of hourly earnings at the time of the survey.

** Significantly different from zero at 1% level.

* Significantly different from zero at 10% level.

Notes: These data are taken from respondents in the CHDS who provided valid information for the purposes of this study. The dependent variable is the natural logarithm of hourly earnings for the 671 individuals who were working at age 21. Youth are defined as ‘Maori’ in these regressions if they identify Maori as at least one of their ethnicities at age 21, and had at least one parental figure claiming Maori ethnicity by age 14 of the CHDS child. All other youth are considered to be ‘non-Maori’. Standard errors are in parentheses.

The sample consists of the 671 youth (69.0% of the individuals interviewed at age 21) who were working at the time of the survey. The covariates include indicator variables on ethnicity, gender, all school and post-school qualifications, and whether or not the individual was working less than 30 hours per week and working while enrolled in tertiary education. The last two variables are included to capture any penalties associated with part-time employment and flexible work arrangements intended to fit around tertiary study. We also include the actual work experience accumulated by age 21. Formal qualifications and work experience should proxy for individual differences in human capital. The long regression specification includes the same measures of personal and family background characteristics used in the earlier work experience regressions.

Consistent with log wage regressions in other studies, the explanatory powers of these regressions are relatively low. The R² statistics are both 0.064 and 0.080 in the short and long regressions, respectively. The adjusted R² statistic actually declines slightly from 0.047 to 0.043 when these personal and family background factors are added to the model. Dropping the Maori indicator variable from these specifications causes the R² statistics to decline from 0.064 to 0.060 in the short regression and from 0.080 to 0.077 in the long regression.

The estimated coefficients on actual years of work experience are both positive and significant at better than a 1% level in columns 1 and 2. They are equivalent to incremental changes in hourly earnings (or rates of return) of 6.7% and 6.8%, respectively. There is no evidence in these regressions that females face systematically lower wages than males, once other factors are held constant. The estimated coefficients are negative, but insignificant. The only qualifications that have positive and significant incremental effects on hourly earnings are School Certificate and University Bursary. We estimate that a School Certificate leads to average increases in wages of 8.2% to 8.3%, while University Bursary leads to an additional average increase in wages of between 12.3% and 14.0%.[23]

The lack of any significant effects from a University Degree or Diploma is not surprising given that very few individuals in this sample have obtained these qualifications, and most of these university graduates would have just entered the work force by age 21. No evidence is found of any ‘wage penalty’ associated with either part-time employment or employment while studying. If anything, there appears to be a ‘wage premium’ associated with part-time work. Yet, the estimated coefficients on these two indicator variables are statistically insignificant.

Once all of these factors are held constant, Maori are found to face systematically higher wage rates than non-Maori. The estimated coefficient on this ethnic indicator (0.088) is significantly different from zero at an 8% level in the short specification. It is slightly smaller (0.083) and significantly different from zero at only a 10.5% level in the long specification. These estimated parameters translate into marginal effects on wages of between 9.2% and 8.7%, respectively. Recall that the overall hourly earnings of Maori and non-Maori were not statistically different from one another. Yet, Maori workers have less work experience and poorer qualifications than their non-Maori counterparts. The result is that observationally equivalent Maori workers receive higher hourly earnings than non-Maori workers.

We can next ask whether this same result holds for all youth, and not just those currently employed. Is there any evidence that Maori face higher wages than non-Maori? One way to answer this question is to use the results from the wage regressions to predict the hourly earnings of all individuals (regardless of their work status). This controls for sample selection in terms of these observable factors. Using actual productivity and other characteristics, we predict the mean wages of all non-Maori and Maori if they were to work full-time while not enrolled in tertiary education. Using the short specification, non-Maori and Maori face average wages of $8.36 and $8.74, respectively. Using the long regression, non-Maori and Maori face average wages of $8.37 and $8.75, respectively. In both cases, Maori youth would receive hourly earnings that are approximately 4.5% higher than their non-Maori counterparts. This can be compared to a 5.8% higher predicted wage for Maori relative to non-Maori among those working at age 21. Although the ethnic gap in hourly earnings narrows when we focus on the potential wages facing all individuals, there is no evidence in this study that Maori youth face systematically lower wages than non-Maori youth.