http://www.iser.essex.ac.uk/publications/working-papers/iser/2009-18

This paper uses the Linked Income Supplement (LIS) of the New Zealand Household Labour Force Survey (HLFS) to investigate the annual transitions in hourly earnings of working age individuals over the years 1997 to 2004. The primary aim is to estimate the determinants of annual changes in hourly earnings, for those who have positive hourly earnings in each year. A first issue that needs to be addressed is the significant attrition in the LIS mainly due to the fact that the HLFS only follows individuals who remain at the same address between years. A second more important issue is the possibility of biased estimates due to focusing only on people who have positive hourly earnings in each year. Because the LIS provides only two observations at a one-year interval on the earnings of individuals, it is not possible to use sophisticated statistical methods to account for these possible sources of bias. Instead, the paper shows that, in practice there are only small differences in the observed characteristics of people in this study, compared with people of the same age represented in the HLFS who have positive earnings in any year. The shape of the earnings distribution is also quite similar for both groups. The paper shows that the probabilities of moving across earnings quintiles between years in the LIS data are quite similar to those found in earlier analyses using tax data and show a moderate degree of earnings mobility in New Zealand. Differences can potentially be accounted for by known differences in the earnings measure used in the two sets of analyses. The paper finds that differences in educational qualifications are associated with differences in earnings growth between years. Better educated people are likely to have better learning ability and better opportunities to learn while in employment. However, because of data limitations, the results are best interpreted as associations rather than as necessarily showing cause and effect.

This paper uses the New Zealand Linked Income Supplement (LIS) to investigate the annual transitions in hourly earnings of working age individuals over the years 1997 to 2004. I first construct transition matrices for annual changes in weekly and hourly earnings, to enable comparison with previous analyses using New Zealand tax data. I then estimate the determinants of annual changes in hourly earnings using OLS and quantile regressions. Differences in human capital are associated with differences in the rate of earnings growth. The results were broadly similar across the sub-periods 1997-2001 and 2001-2004. Working Paper 2009-18 2009-06-11 Ron Crawford Independent economic consultant