A recent study by Statistics New Zealand, New Zealand Now: Incomes, provides the most up-to-date information available on this topic. It is a competent and useful piece of research. The purpose of this talk is to summarise and comment on some of the key findings of this report, and to examine whether equalising incomes or reducing poverty should be our prime concern in designing welfare policies.

Statistics New Zealand correctly acknowledges that a country's economic performance has a significant impact on the standards of living of its citizens and on the distribution of income. It is economic growth – not redistribution of income – that facilitates higher personal consumption and better education and health for all. A high-income, high-growth country also has greater scope to alleviate hardship among people on low incomes than a low-income, low-growth country.

The study reviews New Zealand's lacklustre economic performance since 1960. In that year New Zealand's gross domestic product (GDP) per capita was 32 percent higher than the average for OECD countries. By 1993 it had slipped to 11 percent below the OECD average and our ranking had fallen from fourth to 19th out of 24 OECD member countries.

There is no mystery about the policies that lead to high per capita incomes. Economic growth is fostered by upholding private property rights, maintaining stable prices, implementing prudent budgetary policies, adopting outward-looking trade policies and open markets, and keeping tax and regulatory burdens low.

The Labour government between 1984 and 1988 and the National government in 1990 and 1991 made commendable progress in implementing such policies. More recently, the quality of economic management has slumped. As a consequence, our annual rate of growth has been slowing since mid-1994 and was negative in 1998, and unemployment has climbed back to over 7 percent of the labour force.

Statistics New Zealand's study focused on the distribution of real income (that is, income adjusted for inflation) in the years ended March 1982, 1986, 1991 and 1996. Several categories of personal income (such as wages and salaries, market income and disposable income) and household income (adjusted for size and composition) were examined.

Income generated by personal effort is by far the most important source of individual and household income. In 1996 wages and salaries accounted for two-thirds of personal income from all sources, including welfare benefits. Income from capital is quite small.

The distribution of income was found to have become "substantially" more unequal between 1982 and 1996. The results were broadly similar for all categories of income because of the influence on the findings of changes in the distribution of wages and salary income. As expected, the study shows that disposable income is considerably more equally distributed than income gross of tax and benefits.

The findings differ, however, among three sub-periods. The distribution of income was broadly stable between 1982 and 1986, and again between 1991 and 1996. The change to a more unequal distribution of income mainly occurred between 1986 and 1991. The Labour Party was in government for most of this period.

The period from 1982 to 1986 covers the start of the reforms. The failure of the Labour government to reform the labour market accentuated the fall in employment, particularly of young males, and was a major factor in the sharp rise in unemployment that occurred between 1986 and 1991. Unemployment is almost always a dominant factor behind poverty and a less equal distribution of income. The number of welfare beneficiaries of working age grew from about 100,000 in 1982 to over 300,000 in 1991. More than 80 percent of that increase occurred between 1986 and 1991.

An article of 10 April 1999 in The Economist discussed the broad economic strategies adopted by several countries, including New Zealand. While The Economist claims to have been "a big fan" of the policies pursued by New Zealand, it states that their "bad point" is "a big increase in inequality". However, the failure to tackle impediments to employment and the problems of welfare dependency may well do more to explain the less even distribution of income between 1986 and 1991 than the reforms that were actually implemented.

The change in the distribution of equivalent household disposable income between 1986 and 1991 reflects a large growth in the share of income earned by households in the top two income deciles and a reduction in the relative income of middle-income households. The share of disposable income attributable to the bottom fifth of households, which are almost entirely dependent on government transfers, declined slightly between 1986 and 1991. However, their spending was much higher than their disposable income – an important and not uncommon finding of such studies, on account of factors such as unmeasured sources of income. The study confirms that the relative position of low-income households was largely protected by welfare arrangements. These results are consistent with an analysis by George Barker that was published by the Institute of Policy Studies in 1996.

Statistics New Zealand did not explain why the reported share of disposable income attributable to households in the top decile increased by a quarter between 1986 and 1991. A number of possible reasons suggest themselves. Most obviously, the margin for higher skills may have increased. Such an increase has been a worldwide phenomenon and helps explain the strong growth in enrolments in tertiary education. The opening of the economy to international competition may also have led to higher salaries for some people. Another possible factor is the growth in two-income households where both incomes are above the average. Despite these increases, New Zealand still stands out as having quite low incomes at the top end, relative to countries such as Australia and the United States.

The reduction in the top marginal rates of income tax in October 1986 and October 1988 significantly increased the disposable income of middle- to high-income earners. However, it does not directly account for the rise in the gross income of high-income earners. One likely explanation is that the reduction in tax rates, coupled with other tax changes, encouraged people to earn more gross income and to reduce tax planning activities, such as the retention of income in private companies, aimed at avoiding or evading tax.

Another possible explanation for the increase in the gross income of high-income earners relates to the fact that following the introduction of fringe benefits tax in April 1985, many firms 'cashed out' such benefits. Fringe benefits were not included in the study. Thus the switch from fringe benefits to salaries will have boosted the reported gross income of many people, even though their effective income may not have changed.

These tax changes affected people on higher incomes the most for the simple reason that they face the highest marginal tax rates and pay the most tax. Households in the top decile by market income paid an average of $39,200 in personal income taxes in 1996. Households in the bottom four deciles received substantially more income by way of government transfers than they paid in taxes.

What does the study tell us about poverty, as opposed to changes in the distribution of incomes? Between 1982 and 1996, according to Statistics New Zealand, there was no increase in the proportion of individuals or households with an income of less than 50 or 60 percent of median disposable income. On the basis of these poverty benchmarks, between 6 percent and 12 percent of households were in poverty over the period.

These measures of poverty are similar to measures used by Stephens, Waldegrave and Frater in an article published in the Social Policy Journal of New Zealand in December 1995. Using a benchmark of 60 percent of median disposable income, the latter found that the percentage of households in poverty fell from 13.7 percent to 10.8 percent between 1983/84 and 1992/93. At a benchmark of 50 percent, poverty was stable at 4.3 percent of households. The period examined did not reflect fully the economic recovery that started during 1991. Nevertheless, the study suggests that fewer households were in poverty than reported by Statistics New Zealand. Neither study suggests a rise in reported poverty since the reforms began.

The Statistics New Zealand study contradicts three claims that have frequently been made. First, the claim that the rich are getting richer while the poor are getting poorer is simply not true. People on high incomes have increased their share of total disposable income while middle income earners suffered a significant loss of income share. However, there was no significant change in the share of income of low-income households. Had the momentum of the reform programme been maintained, it is highly likely that all incomes would have been higher by now and fewer people would be dependent on welfare benefits.

As an aside, it is worth noting that, although we do not have a long-term statistical series, incomes have probably become more evenly distributed over the long run. Brian Easton's analysis of individual tax data, for instance, suggests that incomes became more equal between 1953 and 1976. The same could also be expected to be true for the distribution of wealth.

Secondly, the study refutes the claim that people on low incomes bore a significant fall in their real disposable incomes between 1991 and 1996, as had been predicted when benefits were reduced in April 1991. The average real income of the bottom decile fell by 1.75 percent while that for the second decile was unchanged.

Thirdly, the claim that low-income households were badly affected by the introduction of GST in October 1986 is not borne out by the study. This is no surprise. Benefits were increased, and family support and the guaranteed minimum family income scheme were introduced to fully compensate low-income households for the effect on prices of GST.

In evaluating changes in the distribution of income, Statistics New Zealand relies mainly on the Gini coefficient. This measures the extent to which the actual distribution of income differs from an equal distribution of income – for example, when 40 percent of the population earns 40 percent of total income.

The Gini coefficient is, however, a deficient measure. For instance, a reduction in the income of the highest earners with no change in the income of any other group would be reported as a desirable change in the distribution of incomes.

The Gini coefficient only makes sense as a measure of inequality if envy about high incomes is an equally valid motivation for redistribution as compassion towards people facing hardship. Most people would balk at the notion that envy of the blind for the sighted ought to be rectified by destroying the vision of those who can see. Envy is an unacceptable basis for public policy, just as it is usually regarded as an unacceptable basis for ethical personal behaviour.

This leads to a broader criticism of many analyses of income distribution. Explicitly or implicitly, they tend to assume that the goal should be to equalise incomes. A move toward a more equal distribution of income is described as desirable, whereas a move in the opposite direction is said to be undesirable. The implicit goal is for all people to receive the same level of income regardless of their preferences, age, ability and work effort.

I submit that a goal of equal incomes, either as a means to an end or as an end itself, is a misguided one for any society to pursue. As Aristotle noted long ago, equalising the amount of income people possess will not equalise their well-being because income is a very imperfect measure of welfare. Observed income reflects not only innate abilities but also the myriad choices that people make. Some people voluntarily forgo income to have more time for leisure activities, to rear children, to further their education and training or to work in their preferred occupations.

Redistribution ought not to be about achieving equality but about displaying compassion for fellow human beings who are in hardship and cannot help themselves. A compassionate concern for people who are unable to afford even the minimum material essentials of life should make us very concerned about deviations of income – or, more importantly, spending power – below some minimum acceptable standard. It should not make us concerned about the distribution of incomes or spending per se and, in particular, about variations in incomes among people whose income is well above the minimum. A danger we face as a nation in having an unhealthy preoccupation with high incomes is the loss of talented and entrepreneurial people seeking better opportunities elsewhere. An increase in the top personal tax rate would exacerbate this problem.

Relative poverty can only be reduced by raising the income of those who are judged to be in poverty at a faster rate than that of other groups. A doubling in everyone's income, for example, would have no effect on the reported level of poverty according to a relative poverty standard.

An absolute income or expenditure threshold, on the other hand, focuses debate on the explicitly identified commodities and expenditure patterns that are necessary to avoid hardship. It is likely to show substantially less poverty than that reported in the studies discussed earlier. An absolute standard recognises that an increase in real income reduces poverty. This is simply common sense.

The 1972 Royal Commission on Social Security was responsible for a shift from an absolute to a relative standard in setting benefit levels. This was a step in the wrong direction. The average New Zealander can relate to the idea of a safety net which ensures that fellow citizens are not deprived of food and shelter. I believe people are less comfortable with a concept of redistribution based on 'keeping up with the Joneses' in owning, say, one or more telephones, microwave ovens, television sets, videos or cars.

The poor will always be with us, especially if we measure poverty on a relative scale, but the same people are not always poor. People's incomes change markedly over time as they move between education, part- or full-time work, different jobs and so forth. One recent New Zealand study showed that almost half those on welfare benefits in June 1992 were off benefits a year later. Another showed that a quarter of the taxpayers in the bottom income quintile moved to a higher quintile one year later.

Greater emphasis should be placed on individual and voluntary community responsibility to assist people facing hardship, with the government providing a safety net of last resort. The government usually cannot compete with friends, families and voluntary organisations as a means of expressing compassion effectively. As David Green observed in his book From Welfare to Civil Society, published by the New Zealand Business Roundtable in 1996, government welfare programmes "have tended to impair human character, above all because they have undermined the older ethos of 'community without politics'."

In a follow-up study Towards Personal Independence and Prosperity, also published by the Business Roundtable, Jim Cox showed that New Zealand's welfare arrangements are more generous and place fewer obligations on beneficiaries than Australian arrangements. As a consequence we have proportionately more people dependent on welfare.

If we are serious about helping the less fortunate, we must look beyond static analyses of income distribution and shift the focus to broader policies that raise incomes, create jobs, promote individual responsibility, and reflect compassion for those facing hardship. Foremost among these are:

• a return to a coherent medium-term economic growth strategy to raise incomes: only a rising tide can lift all boats;
• the removal of obstacles to free contracting in the labour market, including the Employment Court and statutory minimum wages. There is no reason why New Zealand should tolerate less than full employment. A well-functioning labour market operates like a vacuum cleaner, sucking low-skilled and other disadvantaged people into the workforce. Very few people of working age who work full time, even on low wages, are among the poor;
• a reduction in the levels of government spending relative to GDP, allowing tax reductions which will enable more people to become self-sufficient. Middle-class welfare should be substantially cut back;
• a concerted attack on low education standards, principally by eliminating the virtual state monopoly on education;
• allowing the private insurance market to play the main role in covering insurable risks such as accidents and sickness;
• placing stronger reciprocal obligations on able-bodied welfare recipients to try to become self-supporting, and introducing time limits on benefits as in the United States;
• insisting on greater parental responsibility to provide for children in the event of marriage breakdowns; and
• a larger role for the voluntary sector and private charity to assist those in hardship, with the government providing a back-up safety net.

The impact of such a programme on income differentials cannot be predicted with confidence, although in the long run I believe the historical trend towards a more even distribution of income would reassert itself. What can be said with confidence, however, is that the programme would lift many more people out of poverty, and that, in my view, should be much the more important priority.