Statistics UK definition of family:

A married or opposite sex cohabiting couple on their own, or

A married or opposite sex couple cohabiting couple/ lone parent and their never-married children, provided these children have no children of their own.

Persons who cannot be allocated to a family as defined above are said to be persons not in a family.  In general, families cannot span more than two generations.  (Statistics UK, 1998, p.204).

US Census Bureau definition of family:

A family is a group of two or more (one of whom is the householder) related by birth, marriage, or adoption and residing together; all such people (including related subfamily members) are considered as members of one family. 

(US Census Bureau, 2000, p.4).

Table 1.2.1:  Comparisons Between Selected International Statistical Definitions

As can be seen in table 1.2.1, all of these definitions of family are household based, that is family members must usually reside in the same household.  The US definition is the only one that does not include de facto relationships.  The Australian definition includes other blood relationships such as sibling relationships.  Neither the US nor the UK includes same sex relationships[1].  None of the definitions allow for parent types to be identified.  It can be seen that these definitions contain some quite significant differences that may bring into question the relevance of any comparison of family data between these nations.  The Statistics
New Zealand and Australian definitions are the most similar.

2.      Income Distribution

Recently there has been a great deal of research focusing on income distribution in
New Zealand.  Some of the studies conducted in recent years are: Podder and
Chatterjee 1998; O’Dea 2000; Martin 1995 and 1997; Barker 1996; Mowbray 1993; Creedy 1997; and Statistics New Zealand 1998 and 1999(a).  These studies use a variety of time periods, income measures (gross, disposable or market) and some equivalence scales.  All of these studies found that income inequality has increased over recent years.  The results however, tend to vary as to the magnitude of the change.  Over recent years the proportion of households in the $30,000 to $100,000 income range has fallen substantially, while the proportions in both the upper ($100,000 and above) and lower (below $30,000) ranges increased (Baker, 1996).  All of the studies that use the Gini Coefficient show that it has increased, indicating an increase in income inequality.  Figure 2.1 shows the Gini Coefficients calculated by some of the studies over different time periods and using differing income measures and concepts.

Figure 2.1:  Gini Coefficients

Each of the aforementioned studies source their data from Statistics New Zealand, either from the Household Economic Survey (HES) (formerly known as the Household Expenditure and Income Survey) or from the Census of Population and Dwellings.  Both of these surveys use the NZSCHF definitions.  As these studies use data from these two sources they are using data that are gathered using the Statistics New Zealand definitions.

In their studies O’Dea, Martin and Statistics New Zealand also focused on the family as the basic income unit.  All these studies found that the difference in average incomes between the family types is quite large.  They all used data from Statistics New Zealand that, as stated previously, use the household as the basic unit.  The data on family types gathered by Statistics New Zealand only differentiate between three family types; couple without children, couple with child(ren) and one parent with child(ren).  Of the three family types studied one parent with child(ren) families have the lowest median incomes followed by couple-only families.  Couple with child(ren) families have the highest median incomes (Statistics
New Zealand, 1998, p.68).  Statistics New Zealand also found that incomes of sole-parent families have only increased very slightly over the last decade while incomes of other family types have increased significantly.  Also single person households are not included in these studies.

2.1    Distortions in the Studies

The above studies use a variety of assumptions about the income unit.  For example, most use the household as the unit, but some use the family.  Podder and Chatterjee discuss their reasons for choosing the household as their unit of analysis:

The next issue to consider is the income unit.  The main contenders here are: the individual, the family and the household.  Often, the design of the survey does not give much of a choice in this respect.  In the New Zealand case, however, there are some choices.  We chose to use the household as the unit of analysis.   The rationale is simply that the members of a household usually pool their incomes, and spend it for the collective welfare of the household, and, therefore, in studies such as the present one, the household is the natural unit to study.  While a household may contain multiple families, the number of such cases is miniscule in the surveys being used.  As a result, the terms household and family can, for all practical purposes, be used interchangeably. (Podder and Chatterjee, 1998, p.11)

The assumption made here, that members of a household usually pool their incomes and spend it for the collective welfare of the household, is a simplistic one.  This may usually be correct for the traditional nuclear family but nothing is said of different types of households, such as blended families or flatting situations, where the distribution of income amongst household members can be quite different from the assumption of pooling for the welfare of the household.  Fleming (1997 and 1999) found that an unequal distribution of income amongst householders could occur in blended families where the ‘new partner’ is the principle income earner and feels no responsibility or is not required to provide for or support the children in the household.  Also the principle earner may not be able to support them because they have to support their own children who live in another household.

Podder and Chatterjee also say that the terms household and family can, for all practical purposes be used interchangeably.  This assumption is not valid.  The terms household and family are not interchangeable in many situations.  Although Podder and Chatterjee mention that this assumption may not be correct when multiple families occupy a household, they then say that the number of households in this situation is miniscule.  There are however, other situations prevalent in New Zealand society that may contradict this assumption.  For example, a household may contain a number of unrelated individuals in a flatting situation.  The household may also include unrelated people in addition to a family.  Therefore household and family are two distinct concepts that describe different populations.

In his study O’Dea (2000, p.18) argues that household based analysis of families may miss important inter-household transfers, such as child support paid to the custodial parent in case of separation.  Also intra-household transfers may be assumed that do not in fact take place, for instance when the principle earner’s income is not shared equitably between all household members.  Although O’Dea identifies these problems with the definitions and data, he still uses the original data in his analysis in a working paper for the Treasury.

2.2    Distortions in the Data

Many distortions exist in the currently available data that are used in the study of income distribution.  Some of the distortions caused by the definitions are:

Question five of the 2001 Census asks:  ‘Where do you usually live?’  In reference to this question the help notes state:  ‘children in joint custody should give the address where they spend most nights.  If children spend equal amounts of time at different addresses, they should give one of those addresses’  (Statistics New Zealand, 2001, p.5).  The definitions used in the Census do not reflect the reality of parents and children in shared parenting situations.  Neglecting this essentially results in false data, as the children are assumed to be living full time at the parent’s house where they were staying on census night.  An example is if two parents split up and remain single, but have an equal shared parenting arrangement, the parent with the child staying at their house on Census night will be classed as the sole parent.  The other will be assumed to be a single person household.  This approach will therefore overstate the membership of one household and understate the membership of another.  Overstating and understating the size of households can result in the expectation of costs for households being inaccurate.  The bias created in household composition data has implications for income distribution.

The income information gathered in the 2001 Census is sourced from Questions 25 and 26 below:

Question 25

Mark as many spaces as you need to show all the ways you got income in the 12 months ending today:

·        Other sources of income, counting support payments from people who do not live in my household.

Question 26

From all the sources of income you marked in question 25, what will the total
 income be

·        that you yourself got

·        before tax or anything was taken out of it

·        in the 12 months that will end on 31 March 2001?

Question 25 includes child support as a source of income for the custodial parent.  Question 26 then asks for the total income received from all of these sources, child support is therefore included in the custodial parent’s gross income.  No acknowledgement is made that child support payments received by the custodial parent are tax-free.  Thus the gross income of the custodial parent receiving child support will be understated by the tax on that amount of after-tax-income.  As argued in Birks (2001, pp.57-8) in comparison with someone who is not receiving child support there is an understatement of 64 percent of the child support received, given the top marginal tax rate of 39 percent.  The 2000-1 maximum child support is around $17,000, which results in a maximum understatement of $11,000, although in most cases it is much less than this.

Income of the liable parent is stated before taxes and child support payments are deducted, without recognition of the child support they paid taxes on.  Therefore, the gross income of the liable parent will be overstated, as the gross income includes both the child support to be paid and also associated tax.  Birks (2001, pp.57-8) argues that the income can be overstated by up to 164 percent of the child support paid, or a maximum of $28,000.  Again in most cases it is much less than this.

Further the current method of calculating gross income includes child support in both the gross incomes of the liable parent and the recipient parent, so there is an issue of double counting the child support.  As a result overall gross income is overstated.

For example, assume that there is a two-parent family with four children.  The man earns $60,000 and the woman earns $38,000, their gross income is $98,000.  If the couple then separates and the children live with the mother, the father has to pay $14,566 in child support.  Total gross income still remains $98,000.  According to the current definitions used in the Census, however, the man still earns $60,000 the woman is now assumed to earn $52,566 and total gross income is now $112,466.  As the child support to the woman is tax-free, her income is being understated.  If the child support were not tax-free, she would need to earn $59,849 to have the same after-tax income.  Finally, note that the man’s net income is equivalent that of someone who is not eligible to pay child support and earning $39,000.

Table 2.2.1

These income distortions have been identified and forwarded as written questions in the House:

Question:  Further to her reply to written question No 20558 (2000), do census income figures double count by including child support in the income of the recipient, while not deducting it from the income of the liable parent; if so, how is this consistent with the answer to question No 20558?

Reply:  The income questions in the 2001 Census of Population and Dwellings refer to gross income received by individuals.  Recipients of child support payments should record this figure as part of their income.  For liable parents, the payment will be part of their expenditure, which is not part of what gross income is measuring.  In this way the census is not double counting child support income.

The collection of detailed information on income is a complex topic best administered by a face-to-face interview.  For example the Household Economic survey provides detailed information on income and expenditure.  Such detail cannot be obtained from a census.

(Question for written answer no. 2132, 14 March 2001)

Question:  In relation to the Census, why is no adjustment made to allow for the tax-free status of child support, when it is included in the gross income of recipients, and does the Census data therefore overstate the incomes of liable parent households and understate those of child support households?

Reply:  Income before tax is asked in the Census of Population and Dwellings because it is the simplest way to obtain income in a self-administered questionnaire.

Gross income is used since it is most common income measure in use and it is not affected by changes in tax policy.  Also people are more likely to know their gross salary and wage figures.  It is more difficult for people to state their after tax income because the amount received may be net of a number of deductions such as rent or insurance payments.

The collection of detailed information on income is a complex topic best administered by a face-to-face interview.  For example the Household Economic survey provides detailed information on income and expenditure.  Such detail cannot be obtained from a census.

(Question for written answer no. 2133, 14 March 2001)

The first answer says that child support is income for the recipient, and is expenditure made by the liable parent, so should be counted in gross incomes of both.  The second answer is basically saying that although gross income is not the best measure, it will have to do because it is the easiest measure to obtain.  There is no denial that this method may overstate gross incomes for receiving parents and understate for paying parents. 

Both the household composition and household income distortions reinforce each other.  For example, sole parent households can appear worse off and single person households can appear to be better off than they actually are.  Therefore studies into income distribution that use such data are biased.

Problems also arise if equivalence scales were used in these studies.  Equivalence scales are a tool for adjusting incomes to allow for household size and composition.  An assumption is made that the households pool their incomes and spend it for the greater welfare of the household.  Households can differ in size and composition.  For example an income that provides one person with an adequate standard of living may be inadequate for a family of four.  Although larger households require a larger income to achieve the same standard of living, there are economics of scale in sharing of many expenses.  Household incomes are divided by factors from the scale to give an equivalent income.  Currently one of the most common scales used in New Zealand is the Jensen scale.  Although, theoretically equivalence scales could give valuable insight, some problems relating to the definitions in the data exist.  There would be problems with children that live in more than one household, as the cost of the children would be spread over both households not just the one as would be assumed.  This would result in an overstatement of the equivalent income of one household and an understatement of the other.

These definitions also do not allow for the existence of informal inter-household transfers such as payments from Whanau or parents who live in other households.  These may exist when a parent is paying extra child support privately or an extended family member is lending or giving money to a family member who lives in another household.  In a study on income sharing in New Zealand families, Fleming (1997) found that inter-household transfers between ‘family members’ were especially prevalent in Maori and Pacific Island families.  Fleming found that ‘Money given to Whanau sometimes took priority over the household budget’ (1997, p.11).  Fleming also found that many parents continue to help support or supplement their children’s incomes even when they live apart and have families of their own.  Although the existence of inter-household transfers is known, the current definitions assume that relations who are members of different households are essentially strangers and have no more financial or emotional contact with Whanau (or other relations) than they would with any other household.  Thus a whole cultural factor of New Zealand society is essentially ignored.  This also has implications for income distribution, as some households would actually be receiving a greater income than that recorded in the surveys.

3.      Studies into Family Structure

In 1998 Statistics New Zealand published a study outlining the current state of families (Statistics New Zealand, 1998).  This study used Census data to attempt to explain what the family is and how it has changed over time.  The study also conducts an in-depth analysis on the different family types that the Censuses collect data on.  The current definition of family only distinguishes between three main family types:

Family Type

Couple without children

Couple with child(ren)

One parent with child(ren)

This is sometime broken down further to include child dependency status; whether the child is a dependant, or an adult.  These further classifications are shown in appendix 1. 

3.1    Results and Findings

Some of the major studies into family structure have been conducted by Statistics
New Zealand (1998), Fleming (1995) and (1999), Maxwell (1989) and McPherson (2000).  Studies have found that the structure of families in New Zealand has changed over time.  The predominant family type for much of the last century was the married heterosexual, two-parent, childrearing family.  Now there are many different types and the family is constantly changing and diversifying.  Some of the more prevalent types of family referred to are: solo-parent families; extended families; same-sex couples; blended families; remarriage families; and couple-only, and the range is ever increasing.  Although these changes have occurred, the Census data is only able to identify and distinguish between the three main types listed above.  Recently same sex couples have been included in the data and are identified as subsets of couple only and two parent families.  Of the identified family types, one-parent and couple-only families are growing the fastest, increasing by 10.9 and 15.5 percent respectively.  Two parent families are growing by the slowest rate at 0.6 percent from 1991 to 1996 (Statistics New Zealand, 1998, pp.13-14).  In 1996 there were a total of 949,497 families, of which 354,585 were couple only and 426,567 were two parent families.  There were 2,571 same sex couple families with no children, or 0.72 percent of couple only families.  There were 684 two-parent same sex families with children, or 0.16 percent of two-parent families (Statistics New Zealand, 1998, p.13). 

3.2    Distortions

As previously discussed, the household focus results in distortions in the data when two parents separate and are in a shared parenting situation.  A child is taken as living in the house that they are in on Census night.  The other parent who cares for the child part of the time is assumed to be childfree. 

If the custodial parent had re-partnered, then according to the current definitions, this arrangement would be defined as a couple with child(ren).  The new partner is considered as a parent, there is no distinction made to indicate that they are not the natural parent.  According to the definitions, once a couple separate, the non-custodial parent losses their parent status and can be replaced by a new partner.  This assumption implies that there is no difference in both the emotional and financial support given to children between ‘natural’ and ‘new’ parents.  Fleming (1999, pp.75-6) suggests that the relationships between new parents and children and the children and their natural parents are markedly different.  In many cases the new partner does not have a parenting relationship with the child(ren) at all.  Therefore, it is inappropriate not to have a distinction between natural and step relationships.  Also, for many children, both their natural parents play a very real part in their lives even if they do not live in the same household.  Some acknowledgement has to be made that natural parents are very often a part of their child’s lives even when they do not live together.  This invisibility of the non-custodial parent in the statistics has significant implications for the effectiveness of policy in this area.  An example here is where a parent is on the domestic purposes benefit and may receive subsidies for child-care even though the non-custodial parent may be willing to care for the child.

Also the inclusion of the social element in the definition of parenting may be problematic in that it serves to further cut out or downplay the biological links to the natural parent no longer living with the child.

While statistics are collected on three types of families, there are others for which no data are available; these family types are included in one of the three defined types.  This means that the three types are not homogenous, although these differences are important, and therefore the data are misleading.  Such an invisible yet different type is the blended or remarriage family.   In 1996 13.3 percent of men and 14.6 percent of women aged 16 and over who had previously been married were either separated or divorced.  This figure is up from 10.5 or 11.2 respectively in 1986 (Statistics New Zealand, 1998, p.21).  In 1996 36.5 percent of marriages were remarriages for at least one partner.   This is compared with 30.5 percent in 1986 (Statistics New Zealand, 1998, p.23).  These statistics indicated an increase in the number of blended families.  However, they only give a partial view, because nowadays many couples never marry or delay marriage till later life.  As a result, de facto families can form and dissolve more frequently without any official record of their existence, thus indicating that there are more blended families than indicated by the data. 

Remarriage family households are structurally different, in that they are flexible, changing in size as children come and go according to access arrangements.  They are economically different, as parental responsibilities for the support of children may extend beyond household boundaries, and adults within the household may not be held financially responsible for all the children who live there.  They are relationally different, in that there will be adults and children in the household who do not relate as parents and children.  The idea that a family household is a unit is contested by the second marriage family.  Rather than acting as a unit, this type of family is often characterised by fragmentation and flexibility.  Its boundaries are not well defined and are likely to be redefined differently by each family member.

(Fleming, 1999, p.162)

These observations about remarriage families can be stretched further to include blended families i.e. the parents are re-partnered but may not be married.  Currently blended families are not identified as such in the official statistics.  As a result, they are essentially an invisible subset of New Zealand families.  As they are grouped together with other families, the observed characteristics may not accurately describe any specific type.

Some of these problems have been highlighted by the following answers to written questions in the House:

Question:  Given the Government’s wish to acknowledge the diversity of family types and to encourage the parenting involvement of both parents when they live apart, why does question 41 of the 2001 Individual Census Form refer to children according to membership of a household rather than membership of a family?

Reply:  The categories for the activity question (41) in the Census of Population and Dwellings have been designed to achieve consistency with the definitions used in the 1998/99 Time Use Survey.  In the Time Use Survey, reliable data was collected on the context of different activities, including who [sic] activities are done for.  The Time Use Survey distinguished between children and adults, between people who live in the household and people who live in other households, and whether children or adults are being cared for because they were ill or had a disability.  The Census activities question maintains these classification boundaries, and will therefore provide complementary data to the Time Use Survey results. 

(Question for written answer no. 20556, 17 November 2000)

Question:  Given the Government’s wish to acknowledge the diversity of family types, why does question 4 of the 2001 Dwelling Census Form fail to explicitly recognise ‘blended’ families?

Reply:  The census of population and Dwellings is not a suitable vehicle for collecting information on blended families and shared parenting arrangements when parents live apart.  The complexity of this information is such that it is very difficult to measure accurately with a self-completed questionnaire.  The basic principle of the census is to count every person once.  Family concepts that extend beyond the household would result in some people being included in more than one family.  This would lead to double counting and would make it impossible to measure the number and characteristics of families in a consistent and meaningful way.  A separate interviewer-administered survey would be needed in order to collect quality information on blended families and shared parenting arrangements. 

(Question for written answer no. 20558, 17 November 2000)

Birks (2001, p.56) points out that it is said that quality is best obtained by a face-to-face interview.  The Time Use Survey offers such an opportunity.  However, it seems that the Time Use Survey was structured to be compatible with the Census and then the Census was structured to be comparable with the Time Use Survey.  Perhaps this is merely an excuse.  One of the reasons stated by the Minister for not changing the current definitions is that it will have implications for the reliability of data for time series analysis.  This suggested reason is a weak one.  Changes to the current definitions have been made over recent years, such as the inclusion of same sex couples introduced in 1995 and also ‘persons in a parenting role’ included in the last review of the definition in 2000.  These changes to the definitions have implications for the effectiveness of time series analysis, as it would be misleading to directly compare data that stem from these slightly varied definitions.

Although it was said by the Minister that more detailed information is obtained by such face-to-face surveys as the Household Economic Survey, Statistics New Zealand states that ‘The New Zealand five yearly Census provides the most comprehensive source of information on families’ (1999, p.10).  It is the most comprehensive because it surveys the whole population, not because of the detail of the questioning.

Birks (2001, p.56) also argues that the Minister suggests that the current approach ‘measures the number and characteristics of families in a consistent and meaningful way’.  Although the Minister said this, 300,000 children have a parent liable to pay child support.   In other words the child is not living primarily with that parent.  Birks questions if it is meaningful to ignore this situation, as this will result in a large proportion of the census data being inaccurate.  The researchers would not know which data is inaccurate, as it cannot be separated out.  So there would be general problems with analysis.  Birks continues by saying that ‘it can result in inaccurate data on the household composition and costs, on time spent caring for children, on the nature of parenting relationships between households and on household income’
(2001, p.7).

An important feature of family life in New Zealand is the extended family.  Statistics on extended families are especially important for Maori and Pacific Islanders as a greater proportion tend to live in these types of family groups.  Statistics New Zealand acknowledged the need for information on extended families especially in the Maori and Pacific Island sections of the community.  Information on extended families was first gathered in the 1996 Census.  However, as with the other definitions, the definition of extended family is household based.  The definition is then limited to extended families that live in the same household.   Thus any interaction between related individuals who reside in different households is not accounted for.  Statistics New Zealand wrote ‘the family type classification does not include categories for extended families.  It is based on the concept of family - the family nucleus- that is inappropriate for Maori and Pacific Island groups’ (Statistics
New Zealand, 1995, p.41).  Statistics New Zealand expects to derive a wide range of information on extended families from the 2001 Census.  Some of the information is expected to include extended family type, number of people aged 15 years and over, number of children, sources of extended family income and also some other characteristics.  If Statistics New Zealand themselves acknowledge that a household based definition of the extended family is inappropriate for Maori and Pacific Islanders while stating that this is the group for whom the data is most relevant, why then do they produce statistics on this topic at all?

4.      The Family According to the New Zealand Immigration Service

Recently the New Zealand Immigration Service in conjunction with the Department of Labour undertook a review of family sponsored immigration called Immigration Research Programme: Family Structures (Elliott & Gray, 2000).  Central to the issue of this paper was identification of the appropriate definition of family.  The report was based on a ‘literature review and 15 interviews with people who have immigrated to New Zealand from such places as Asia, Africa, Eastern Europe and the Pacific’ (2000, p.1).  Some of the major conclusions of the research were:

·               The family is not a biological unit as used in arguments about property rights and such like, but it is a social unit.  All of the people surveyed perceived the family as a social unit and most cultures saw the family as more important than the individual.  This idea of family being a social rather than biological unit is very interesting as it is in line with Statistics New Zealand’s move to incorporate persons in a parenting role in the current NZSCHF definitions.  As stated earlier, however, the inclusion of the social element in the definition of parenting may be problematic in that it serves to further cut out or down play the importance of links to the natural parent who no longer lives with the child.  Although the report concluded this, it also found that:

Although the literature often focuses on family living arrangements, the study concludes that family membership includes obligations across and between generations, no matter where the family members are living.  ‘While co-residence over some periods of life can reinforce obligations, evidence indicates that co-residence is not a key characteristic of highly interdependent relationships.  Families attempt to maintain obligations with family either within or between countries and in spite of living apart.  … Family membership is not an individual matter, nor restricted to a nuclear family definition or even co-residence.  … At a minimum, definitions of families in the cultures studied include parents, grandparents, adopted members and siblings, and depending on circumstances, can extend beyond that to clan, tribal or village associates’
(Elliot& Gray, 2000, pp.49-50).

·               The research concludes very strongly that co-residence is not an essential characteristic of families.  If this research found differently, there would be no point in family sponsored immigration.  This conclusion is inconsistent with the assumption of the NZSCHF definitions that families are household based.  The findings state that families can still act and interact as families while living in different countries, while the Statistics New Zealand’s definition implies that people cannot function as a family if they do not reside together, even if they live in the same street.

·               The research findings were the same as a 1992 UNESCO report (pp.6-7), which also found that it was important to distinguish between a household and family unit.  In this report five distinct family types were identified; a nuclear family, stem family, lineal family, an extended or joint family and a compound family (definitions are in
appendix B).  Of the five only three would be acknowledged by the current Statistics New Zealand’s definitions.

5.      The Family in International Conventions

In 1993 New Zealand signed and ratified the United Nations Convention on the Rights of the Child (UNCROC, 1990).  By doing so New Zealand committed to implementing UNCROC’s principles into our own law.  Two of the articles state:

9.3  States Parties shall respect the right of the child who is separated from one or both parents to maintain personal relations and direct contact with both parents on a regular basis, except if it is contrary to the child's best interests...

18.1  States Parties shall use their best efforts to ensure recognition of the principle that both parents have common responsibilities for the upbringing and development of the child.  Parents or, as the case may be, legal guardians, have the primary responsibility for the upbringing and development of the child.  The best interests of the child will be their basic concern.

(UNCROC, 1990).

As can be seen this international convention does not match the view of the family being household based.  It states that parents have common responsibilities for the child without mentioning the need to reside together.  Also article 9.3 implies that if parents do not live with the child they should be allowed regular direct contact with the child.  This is opposite to the implication of the statistical definition that if a parent does not live with a child they will have no more contact with that child than with any other child in any other house, that is, they are essentially strangers.  Also the definition of parent is biological although some wider scope is given through the mention of legal guardians.

6.      Recent Policy and Policy Debate

In New Zealand a variety of definitions of family are used in policy.  An example is the Child, Young Persons, and Their Families Act (1989) (CYP&F Act), this definition acknowledges both legal and functional relationships and also tries to incorporate cultural differences into the definition.  This definition is:

A family group including an extended family, in which there is at least one adult member with whom a child or another adult member has a biological or legal relationship; or to whom the child or other adult member has a significant psychological attachment; or that is the child’s or other adult member’s Whanau or other culturally recognised group.

(Child, Young Persons, and Their Families Act 1989, s2)

This definition is very different from the statistical definition that is household based.  It allows for both vertical and horizontal relationships within and between households.  A cultural element is also included as Whanau and other culturally recognised groups are classified as family.  This definition leaves open the question of how far a culturally recognised group might extend, what criteria for determining family relationships are acceptable or who has authority to speak or make decisions on their own or on a family member’s behalf (Elliott & Gray, 2000p p.6).  As can be seen this is much broader definition than the statistical one.  Many of the definitions of family used in every day policy are broader and less restrictive than the statistical definition.  The recent push toward ‘biculturalism’ indicates that the inclusion of cultural elements, especially Whanau, is only going to increase, so as to represent a greater proportion of society.

Even though many different definitions and meanings are used in various types of policy, most rely to some extent on official statistics in order to identify problems and to then formulate and target appropriate responses.  If the data that are used are gathered using definitions differing from the policy makers understanding, then this could have serious implications on the effectiveness of targeting and appropriateness of policy, as well as putting into question many policy documents.  The following section will briefly take a few recent policies and official reports and discuss possible distortions and problems.

6.1    The Ministry of Social Policy

The Ministry of Social Policy, now known as the Ministry of Social Development, is responsible for social welfare, dealing with benefit expenditure, and now a greater focus on social development.  As a result, the Ministry deals with a lot of issues regarding the family.  In the post-election briefing paper one of the key areas identified for Government focus in the next three to five years is: ‘identifying policies that will improve the circumstances of Maori and Pacific people by strengthening their families and Whanau’.  It goes on further to say, ‘An important means to achieving these objectives will be better information on the effects of current policies, on which programmes work and which don’t work… … There is a need for a fundamental look at family incomes policy’ (Ministry of Social Policy, 1999a, p.30).  Currently most of the statistics used by the Ministry of Social Policy are from Statistics
New Zealand.  This is a problem as there are no statistics on Whanau.  The closest are on extended families but these, as with the others, are household based.  The Ministry identified a need for better information but did not detail how this would be achieved.