5  Discussion

The starting point in this paper, looking at final household incomes in New Zealand, is to consider market incomes, and then to see how the tax and benefit system redistributes this income. Market incomes do not, however, represent a “natural state” of income which would exist if there were no taxes or government benefits. The existence of taxes and benefits clearly affects people’s participation in the workforce and the amount of money they are prepared to earn. Market incomes differ significantly across households, for a variety of reasons. Income differences are to a large extent due to the changing patterns of earnings over the life-cycle. Individuals tend to earn more as they get older and gain more work experience. Couples who both work can bring in a large annual income, although this is significantly reduced if one leaves work to look after young children. On retirement, market income drops considerably although many older people still have income from investments. Some people inherit wealth, from which they can gain income, and inherit characteristics, such as intelligence, dedication and perseverance, which make them more productive and better-paid workers. Others choose to develop their skills, work long hours, take risks, or do unpleasant work for greater pay. Some people suffer ill-fortune and others in contrast are just plain lucky.

This study found, as other New Zealand studies have also found, that market income increased, on average, for households in the top deciles, especially in decile 10, between 1987/88 and 1997/98. Market income decreased for households in most of the lower deciles. As discussed in Section 4.2, these trends are not easy to explain, and the role of government policy in influencing them is unclear. Since household market income varies with age, marital status and family formation, for example, demographic and social trends will influence the distribution of income in New Zealand. Some of these trends, such as the ageing of the population, are clearly independent of any policy reforms of the time. Others, such as the number of sole parent families, or trends in unemployment, might be directly or indirectly influenced by changes in government policy. The government might also have an influence on the returns to some population characteristics, for example by stipulating a minimum wage, but no influence on many others. Changes in tax policy, particularly the difference between the top income tax rate and the company tax rate, might have an effect on how much market income was reported by households, especially households at the upper end of the income distribution.

The key finding which this study reaches, however, is that government intervention—through taxes and social expenditure—has acted to protect less well-off households by increasing their net government benefits to counteract declining market incomes. Between 1987/88 and 1997/98, redistribution became more favourable toward households in the low income deciles. A part of this increase in redistribution is automatic—households earning more income pay more tax, for example, but are unlikely to receive more in government benefits—but a part of the increase is also likely to be a result of specific government policy changes. As a result of this change in redistribution, real incomes were maintained for all income deciles between 1987/88 and 1997/98. While there will undoubtedly have been individual households which suffered financial hardship and declining circumstances, all deciles were, on average, doing at least as well in 1997/98 as they were ten years previously, and in most cases were doing better.

As measured by Gini coefficients, final household income was more unequally distributed in 1997/98 than it was in 1987/88. This growth in inequality of final incomes was relatively small, however, and considerably less than the growth in inequality of market incomes or disposable incomes. This paper makes no judgement about whether, or to what extent, increases in income inequality are desirable, tolerable or otherwise.

Other New Zealand studies have examined changes in household income during the 1980s and 1990s, although none has used final income as a measure. What does the estimation of final income add to what is already known? Estimating final income, rather than just disposable income, reveals further redistribution from more well-off households to less well-off households. This redistribution is especially evident in 1997/98. In all deciles, the average final income was at least maintained between 1987/88 and 1997/98, but in some deciles the average disposable income fell (see Appendix Table 2). Final incomes are more equally distributed than disposable incomes. Using final income as the measure of household income therefore provides a richer picture of incomes and income distribution in New Zealand. It is arguably a less intuitive measure than disposable income, however, not least because households which are heavy users of publicly-funded education and health services are not obviously better off than households which are not (household members are likely to be sicker, for a start).[17]

More generally, this study is in the broad tradition of studies which measure incomes, by decile, over a period of time and attempt to make sense of the observed changes. It is worth considering what these types of studies do and do not tell us. Firstly, households in each income decile are not homogeneous with regard to their circumstances, and are likely to vary widely in terms of household expenses and household wealth. In the lower deciles of income, for example, retired people who have paid off their home and accumulated savings are likely to live much more comfortably than sole parent families with young children. Incomes are only one element of people’s material standard of living, which in turn is only one element of their overall standard of living or quality of life.

Secondly, studies compare two or more cross-sectional “snapshots” of incomes and therefore cannot show how the incomes of particular individuals or households change over time. The distribution of people’s lifetime income is more equal than the distribution of personal income at a given point in time (Creedy 1997), and this is likely to be the case for household income as well. Since the current study only compares a snapshot of incomes in 1987/88 with another snapshot in 1997/98 it is not possible to say whether individual households were better or worse off in 1997/98 than they were in 1987/88. Over this ten-year period, individual households may have moved between income deciles several times, as their circumstances change. At one point in time a household may be a net “winner” from fiscal policy and at another, a net “loser”. How these benefits and losses accumulate over the life cycle is only loosely covered in this study through the analysis of households by life-cycle stages which are presented in Appendix 2.[18]

Thirdly, studies of income changes between two points in time will show different results depending on exactly which two points are chosen. The New Zealand economy was deep in recession in 1992 (Figure 1) and if the second reference point had been 1992/93, rather than 1997/98, the results may well have shown a reduction in final incomes for all deciles. On the other hand, if the 2000/01 HES had been used as the second year of the study,[19] the results may well have shown a strong increase in final incomes for all deciles, given the strong economic growth which has occurred in New Zealand since late 1999.

Finally, it is easy to speculate (as this paper does) on the reasons for observed trends in household income and to link these trends to particular government policies. It is quite another matter, though, to demonstrate that these really are major contributing factors. Little rigorous analysis on the reasons for income changes has been undertaken in New Zealand and what has been done does not always accord with commonly-held beliefs. Hyslop and Maré (2001) show, for example, that job losses in the late 1980s and early 1990s, although substantial, had only a small net effect on overall income inequality. Income studies by decile also do not reveal what would happen if tax or expenditure policies changed in the future, although they may allow some “first-round” predictions to be made. That is, such studies do not predict longer-term behavioural responses to government fiscal policy. For example, if the government reduced the level of New Zealand Superannuation paid to elderly people we would not simply find that the incomes of older people declined, but rather that, after a period of adjustment, older people might stay in the workforce longer, earning more income, or save more during their working lives to give themselves more investment income in their retirement.[20]

These considerations should be borne in mind when interpreting the results of this study, and others like it. For example, it may or may not be true, as some commentators assert, that “the rich are getting richer”, but this study cannot prove or disprove this assertion because it does not track individual households over time. Many of the households who were in the top decile in 1987/88 will be in lower deciles a decade on (as people retire, for example) and, similarly, some households which were in low income deciles will be in deciles 9 or 10 in 1997/98. What the study does show is that the top 10% of households in 1997/98 had higher final incomes, on average, than the top 10% of households in 1987/88, but this is a slightly different finding. Similarly, the study shows that, on average, final household incomes in the lowest income deciles were maintained over the decade, but this does not reveal whether the proportion of the population facing genuine hardship has grown or diminished. The contribution of this study is to show that redistribution through taxes, cash benefits and social services has protected the position of less well-off households in New Zealand over a decade of extensive policy reforms and significant economic fluctuations.