Senior Lecturer in Public Policy
Victoria University of Wellington
Social Policy Consultant
Lower Hutt Family Centre
In Issue Four of this journal, Krishnan (1995) published a statistical analysis of trends in poverty, using six different poverty thresholds. Her analysis showed an increase in the incidence of poverty for all of the different categories of household types which can be developed from unit records of the Household Expenditure and Income Survey data base (Frater, Stephens and Waldegrave forthcoming, Stephens 1994a). However, use of the Benefit Datum Level, the 1994 married couple Invalid or Unemployed Benefit rate, or various percentages of mean and median equivalent household disposable income, by themselves, do not constitute a measure of poverty or income inadequacy due to lack of independent argument on the standard of living achieved at each poverty level.
Church, welfare and community organisations have undertaken a variety of small-scale studies, documenting the living conditions of the poor, and the difficulties that these households have in making ends meet (Craig et al. 1992, NZCCSS 1995). But these studies cannot provide an estimate of the income required to avoid poverty and hardship, nor the statistical details required by policy-makers on the extent and severity of that hardship, nor information on which kinds of households are likely to be poor. Thus, it is necessary to establish an independent estimate as to what constitutes an adequate income to avoid poverty and hardship, and then relate that poverty measure to an independent statistical source.
This paper sets out the New Zealand Poverty Measurement Project's method of establishing a poverty line (Frater, Stephens and Waldegrave forthcoming). Focus groups are used to establish that poverty measure, by asking households to determine an income level which will give a standard of living which provides for a minimum adequate household expenditure. A benchmark standard of living is thus developed which is both absolute in that it represents a standard of living below which households should not fall, and is relative in that it is set in relation to economic conditions within New Zealand. Changes in either economic conditions or policy parameters, e.g. charging for health services or lowering GST rates, would alter the income level required to avoid poverty. Thus the policy measure is time-specific to a given set of economic or policy parameters.
The same time-specific argument indicates why the Benefit Datum Level (BDL) used by many analysts as a poverty measure (Easton 1976, 1986, 1994, Krishnan 1995, Rochford and Pudney 1984),  is no longer appropriate. Use of the BDL was initially a logical choice as the married couple benefit level was based on the Royal Commission on Social Security's (1972) expert deliberations and research on income adequacy as to the benefit level required for beneficiaries to "be able to belong to and participate in the wider community". However, economic conditions, social attitudes and public policies have significantly altered over the intervening twenty years, so that continued use of the BDL as a poverty measure no longer relates to its original conceptualisation. Even the Royal Commission (1972) argued that "poverty, need, and benefit adequacy are relative concepts", and that "the relationship between the benefit level and the selected wage level will not be fixed and immutable". Updating the 1972 benefit level by use of the Consumer Price Index (Easton 1994, Krishnan 1995) implies an absolute poverty standard, or one where the economic and social reforms from 1984 have had no impact on the required standard of living of a beneficiary.
Equally, the Royal Commission relationship between benefit rates and average earnings in the 1972 formulation of benefit levels is no