The effect of first language and education on literacy, employment and income: An analysis from the Adult Literacy and Life Skills survey

6. Income

This section looks at the effect on income of having English as an additional language. The section looks at direct earnings as expressed in hourly wages, overall total income per person and equivalent household income.

6.1: Hourly wages

Overall, the ALL data shows that people with English as an additional language were more likely to be at the lower end of the wage distribution. They were about as twice as likely to be in the lowest 20 percent of the overall wage distribution and half as likely to be in the highest 20 percent of the overall distribution than people with English as a first language.

Figure 16: Distribution of hourly wages by first language for people aged 25 to 65

Note: Each quintile represents 20 percent of the total population. The income boundaries for each quintile are set out in Table 3 in Appendix A.

Hourly wages varied by qualification level. Figure 17 shows difference in median hourly wages by first language and qualification level. The chart on the left shows the observed values from the ALL survey, without taking into account differences in literacy or numeracy, age, gender and skill level of work. The chart on the right is the predicted values from a model of hourly wages which controls for prose literacy and other variables.

The observed values show a substantial gain in hourly wages by qualification level for all people. However, there was a consistent gap at each level between those with English as a first language and those with English as an additional language. The predicted values show a similar gain in wages by qualification level for people with English as a first language. However, even once prose literacy in English was controlled for, people with English as an additional language still earned less and derived less benefit from higher-level qualifications. Furthermore, among people with English as an additional language, the earnings of those with a degree or above were only marginally higher than those with lower qualifications.

The predicted values when controlling for numeracy were almost the same as those when controlling for prose literacy.

Figure 17: Median hourly wages by highest qualification and first language for people aged 25 to 65

 

Note: Predicted values are for a respondent with mean prose literacy, who is female, aged 40 and uses a computer at work.

Country of education did not have a statistically significant effect on hourly wages, once literacy or numeracy was accounted for.

The region of the first language spoken did have a significant impact on hourly wages, even after literacy or numeracy and qualifications were accounted for. People who spoke an East or South East Asian language as their first language earned on average 22 percent lower hourly wages than people not born in New Zealand who spoke English as their first language. People who spoke an Eastern European language as their first language earned 20 percent less on average and people who spoke a Middle Eastern or South Asian language earned 17 percent less.

The differences were similar when taking account of literacy or numeracy, with the exception of speakers of East and South East Asian languages. When numeracy was taken into account, their wages were 26 percent less than people not born in New Zealand with English as a first language. This indicates that the relatively higher average numeracy skill of people with these languages as their first language is not fully recognised in the workplace, due to their lower prose literacy skills.

6.2: Total personal income

Total personal income provides a figure for the total annual income received by the individual interviewed in the ALL survey from all sources, including wages, benefits and investments. It effectively adjusts for whether the person was employed or not and how many hours they worked during the year. It covers both labour force participation and earning effects.

Overall, the ALL data showed that people with English as an additional language were more likely to be in the lower 40 percent of the income distribution and less than half as likely to be in the top 20 percent.

Figure 18: Distribution of total personal income by first language for people aged 25 to 65

Note: Each quintile represents 20 percent of the total population. The income boundaries for each quintile are set out in Table 4 in Appendix A.

Total personal income varied by qualification level. Figure 19 shows difference in median personal income by first language and qualification level. The chart on the left is the observed values from the ALL survey, without taking into account differences in literacy or numeracy, age, gender and labour force status. The chart on the right is the predicted values from a model of total personal income which controls for prose literacy and other variables.

The observed values show a substantial gain in personal income by qualification level for people with English as a first language. For people with English as an additional language there was little gain for having a qualification up to degree level and only modest gain for holding a postgraduate qualification.

The predicted values show that once literacy and other factors were controlled for, people with English as an additional language did not have any additional income benefit from higher level qualifications. The difference between people with and without English as an additional language showed up most for people with postgraduate qualifications. The predicted values when controlling for numeracy were almost the same as those when controlling for prose literacy.

Figure 19: Median total personal income by highest qualification and first language for people aged 25 to 65

 

Note: Predicted values are for a respondent with mean prose literacy, who is female, aged 40 and employed full-time.

Country of education did not have a statistically significant effect on total personal income, once literacy or numeracy was accounted for.

There was some evidence that the region of first language spoken had an impact on total personal income after English-based literacy or numeracy, qualifications and labour force status were accounted for. When prose literacy was controlled for, people whose first language was a Pacific language had 19 percent lower personal income than people not born in New Zealand who spoke English as a first language. When numeracy was controlled for, people whose first language was a South East or East Asian language had 22 percent lower income. The rest of the differences were not statistically significant.

6.3: Household income

Equivalent household income provides a measure of the income of households adjusted by the size of the household. Overall, people with English as an additional language were more likely to be living in households in the lowest income quintile for equivalent household income and less likely to be living in the highest income quintile.

Figure 20: Distribution of equivalent household income by first language for people aged 25 to 65

Note: Each quintile represents 20 percent of the total population. The income boundaries for each quintile are set out in Table 5 in Appendix A.

Equivalent household income varied by the qualification level of the respondent. Figure 21 shows difference in median equivalent household income by first language and qualification level. The chart on the left is the observed values from the ALL survey, without taking into account differences in literacy or numeracy, age, family background and household composition. The chart on the right is the predicted values from a model of equivalent household income which controls for prose literacy and other variables.

The observed values show a substantial gain in household income by qualification level for people with English as a first language and a much more moderate gain for people with English as an additional language. The predicted values show no gain in income with level of qualification for people with English as an additional language, once prose literacy is controlled for. The predicted values when controlling for numeracy were almost the same.

Figure 21: Median equivalent household income by highest qualification and first language for people aged 25 to 65

 

Notes: Incomes are equivalised to a household with two adults and no children, using the Revised Jensen scale (see Appendix A). First language and highest qualification refer to the member of the household selected for the ALL survey. Predicted values are for a household with 2 adults and no children, where the respondent has mean prose literacy, is full-time employed, aged 40 and both parents had secondary school or tertiary qualifications.

The models show that country of education had a significant effect on the household incomes of people with English as an additional language, even when literacy or numeracy, qualifications and other factors are accounted for. Holding all other factors constant, a household where the respondent had English as an additional language and was entirely educated in New Zealand on average had a 30 percent higher income than a household where the respondent had English as an additional language and had all or some education overseas. This effect may be a result of those with a New Zealand education being more likely to partner with people with English as a first language, resulting in a higher household income. Unfortunately, there is insufficient household data in the survey to explore this further.

The region of the first language spoken by the respondent had a significant impact on equivalent household income, even after literacy or numeracy, qualifications and labour force status were accounted for. People who spoke an East or South East Asian language as their first language had household incomes that were on average 42 percent lower than people not born in New Zealand who spoke English as their first language, having controlled for prose literacy and other factors. For other language groups the differences were: 34 percent for Eastern European languages; 26 percent for Middle Eastern and South Asian languages; 24 percent for Pacific languages and 18 percent for Western European languages. The results were similar when numeracy was controlled for.