WOMEN, FAMILIES AND UNPAID WORK
Stuart Birks*
Massey University, Palmerston North
NEW ZEALAND
______________________________
* This is a modified version of a paper presented at the New Zealand Association of Economists Conference, 19-21 August 1994, in Palmerston North. I wish to thank Gary Buurman, Chris Turner, Michael Pickford, Rolf Cremer, Paul Callister and others for helpful comments and discussion. Mention should also be made of the participants in the internet newsgroups alt.child-support, alt.dads-rights, alt.feminism, soc.men and soc.women for stimulating discussion and ideas.
ABSTRACT
Any attempt to take an economic approach to the family must consider the contribution made by unpaid workers. There are those who would contend that a greater contribution by one partner through paid work is balanced by a greater unpaid contribution by the other. Increased female participation in the paid labour force would indicate that changes in unpaid work contribution may also be appropriate.
Various approaches to the valuation of unpaid work have been described in the literature. Little has been said on actual payments made, however, although these are identifiable.
This paper firstly looks at relative contributions to paid and unpaid work, using census and time diary data. Working wives and full-time housewives are considered, as are intertemporal issues. Secondly, the nature of and rules for payments are assessed. Thirdly the general issue of fairness is addressed.
1. INTRODUCTION
There are two things that I stress to my students when discussing economic issues. They are firstly that, as all theories are simplifications, a "healthy scepticism" is required, and secondly that there are many perspectives that can be taken, and any theoretical view involves taking just one perspective. Debates between competing theories have at times been little more than disputes as to which alternative perspective is "best", as if they are assumed to be mutually exclusive. We are being very restrictive when we make such an assumption, and frequently we are wrong. It is as a contribution in this context, of healthy scepticism and many perspectives, that the following should be viewed.
In comparison to data and studies on paid work, there has been little analysis on the amount of unpaid work undertaken in New Zealand. Given the gender mix of many households and the gender debate going on at present, male-female distinctions are of topical interest. They are also an important area for analysis if we are to understand the nature of intra-family economics and the regulatory structures affecting families.
Firstly I shall consider paid work. Data are provided in the 1991 census giving numbers of men and women in paid work, grouped into five hour bands of hours worked per week. If we assume that each person was at the midpoint of the hours band, we can calculate hours worked by men and by women. This gives averages of 45.8 hours per week for men, and 36.1 for women. Note that only people "gainfully employed in the labour force" are included. Women who undertake paid work do so for an average of 9.7 hours per week less than men in paid work.
Gender differences in participation rates can also be observed. While they vary somewhat from quarter to quarter, figures in late 1991-early 1992 were about 74 per cent for men and 54 per cent for women. At that time, unemployment was running at about 11.5 per cent for men and 9.5 per cent for women. This means that percentages employed were approximately 65.5 and 48.9 respectively. At this time approximately 13 per cent of the 15+ male population were aged 65 or over, compared to 17 per cent for women. Assuming that no one over 64 was "gainfully employed", percentages employed in the 15-64 male and female populations are 75 and 59 respectively. Assuming a constant pattern, this means that men are employed for 37.6 years on average, while the corresponding figure for women is 29.5 years. Eliminating rounding errors, this means that men would undertake paid work for on average 8.2 years more than women.
The 1993 Official Yearbook gives average ordinary time hourly earnings at May 1992 as $16.22 for males and $13.13 for females. While this is only a partial indicator of earnings per hour, male earnings per hour are approximately 23.5 per cent higher than female or, in the more standard presentation, female ordinary time average hourly earnings are on average 81 per cent of male earnings.
Clearly, despite over a quarter of a century of change, women are not equal partners with men in the paid labour force. Programmes are in place to speed change, and slogans such as "Girls can do anything" have been used to encourage the necessary attitude adjustment whereby long-established barriers can be broken down. Nevertheless, they still work for less time, for fewer years, at lower pay than men. If we were to be looking for a roughly accurate, easily remembered slogan, we could say that, "compared to men, on average women work for 80 per cent of the years, at 80 per cent of the hours, and at 80 per cent of the pay". Financially, before tax, they each contribute on average about half as much as a man over their working lives.
With such differences, there could be significant inequities, and there are strong pressures for social change to correct for these. Attention is frequently drawn to the area of unpaid labour, what it is worth and how it is rewarded. It is to this topic that I would like to turn. Given the above information, clearly on average men are making a larger monetary contribution than the women who are in relationships with them. Essentially, a financial transfer, or payment, is being made by men to their partners. Assuming non-altruistic behaviour, it seems reasonable to ask what men are getting in return. Specifically:
1. what non-monetary benefits are they getting that outweigh those they are giving?
2. how much are they paying for these?
3. are they getting value for money?
It is to the examination of these issues that we now turn.
2. THE FIRST QUESTION: WHAT NET NON-MONETARY BENEFITS
ARE MEN GETTING?
There is a measurement problem with unpaid work that has long been discussed. Some of the more recent literature appears to overlook this. Waring, in her classic and widely acclaimed book, Counting for Nothing: What Men Value and What Women are Worth, describes economists' attitudes to housework. She states that, to an economist, a housewife is "totally unproductive ... and economists record her as unoccupied" (Waring 1988, p13). Describing two women who are heavily involved in unpaid work, she states:
"You, too, may believe that these women work full days. But, according to the theory, science, profession, practice, and institutionalisation of economics, we are wrong." (p14)
I think that she is saying that economists value unpaid work at $0. This is true in terms of the national accounts, but the problem is clearly recognised by economists, and was even discussed at length in standard texts at least twenty years before Waring's book was published [footnote 1]. The following gives some mainstream economics viewpoints on this limitation of GNP:
" ... the measurement of GNP does not include the value of services that are not exchanged in formal markets. The cooking services of a wife are not measured as part of national product, yet the same services bought in a restaurant are." (Alchian and Allen, 1968, p.516)
"GNP as a measure of the market value of output fails to include [some] productive transactions. Standard examples include the productive services of the housewife ... causing GNP to be understated." (McConnell 1978, p.174)
"The economic activity measured in the national income accounts constitutes the observed sector of the economy. But clearly a vast amount of production, consumption, and investment is never officially measured or taken note of. This unobserved sector of the economy incorporates ... the multi-billion dollar value of cooking, cleaning, and child-care services performed in the home." (Dolan 1983, p.141)
Eisner (1978) estimated a value of US$268 billion for unpaid housework in the United States in 1969. Kendrick (1979) gives a value of US$318.4 billion for 1973, this being 24.4 per cent of GNP.
As we all know, data measure what they measure, no more, no less. We must be aware of the bounds and limitations of the information available to us. I have yet to see an economist state that unpaid work is of no value. However, it has long been recognised that standard measures of economic activity do not include allowance for output from unpaid labour in the home. There are two problems involved, that of measurement of work done, and that of evaluation.
The question of interest at this stage in the paper is that of measurement. This has been addressed in the literature on unpaid labour in the home, including time diary studies (see for example Fletcher 1978, Horsfield 1988, Waring 1988, Ministry of Women's Affairs 1989, Department of Statistics 1991, Callister 1993). There is a widely held view that women do more housework, while men do more paid work. The standard unit of measurement is time spent on certain specified tasks. Clearly there are problems with this as time is a measure of input and the relationship between time and output is likely to differ widely over tasks. I shall use it, however, as (i) it is the standard approach used by researchers, and (ii) as with national accounts "it is what we have".
Horsfield refers to Fletcher as a source for New Zealand data. The New Zealand Department of Statistics conducted a pilot time use survey in 1990, with a few results being published in 1991. Callister summarises a 1987 Australian pilot survey [footnote 2].
Unfortunately the studies do not all measure the same thing, and results are only partially documented. I shall discuss the studies in sequence.
(a) Fletcher's study
Table 1 summarises the weekly position found by Fletcher (1978) in his 1978 study of 50 couples, half of whom included a working wife. Note that "outside employment" includes travel time. According to this study, the lower paid work time contribution by women coincides with a greater time contribution for unpaid work. Working wives undertake, on average, 65 per cent of the housework and childcare hours, with the result that over all they put in on average 1.8 hours per week more than their husbands. Full-time housewives undertake 76 per cent of the housework and childcare hours, working on average 11.61 hours per week less than their husbands. If we take into account the lower number of years that women spend in paid work, a couple staying together throughout their working lives may well put in approximately equal hours of work overall.
Fletcher makes the following statement in the introduction to his paper:
"Research in the U.S.A. ( ... ) and in a range of other countries ( ... ) has demonstrated that working wives, although spending considerably less time on household work than full-time housewives, tend to retain their housewife role with the consequent drastic increases in their total working hours." (p33)
TABLE 1
Weekly Allocation of Hours of Paid and Unpaid Work
Mean hours
per week
Outside Housework Total
employment and
Childcare
Working wives (ww) 32.77 35.37 68.14
(7.82) (8.88)
Husbands of ww 47.70 18.64 66.34
(6.21) (10.61)
Full-time housewives - 54.76 54.76
(fthw)
(-) (19.22)
Husbands of fthw 48.67 17.70 66.37
(6.45) (9.09)
Although working wives "retain their housewife role", they do in fact greatly reduce the time spent in that role. As a point of semantics, and bearing in mind discourse analysis, "role" here is defined as a qualitative variable; you either fill it or you do not. However, the cross-gender comparisons are made in terms of hours put in, a quantitative variable. There seems to be scope for much variation in hours and sharing of tasks with no change in role classification. We should not let ourselves be misled by this.
It is not until the last sentence of Fletcher's paper that it is pointed out that working men also put in long hours. In fact, according to this study, the "drastic" increase for working wives does little more than make up for an equally dramatic shortfall in the work hours of full-time housewives, compared to their husbands.
Other distinctions are apparent when we look at the breakdown of household tasks as covered in Fletcher's table 4. We may not be entirely happy with the use of hours as a measure of contribution under these circumstances. There is much diversity in the nature of both paid and unpaid work. There are also likely to be large differences in the output from that work. An extensive subjective assessment could be made by studying the various categories given. This would inevitably be open to debate, so I shall simply mention two observations that stood out for me. Firstly, although husbands of full-time housewives put in 29 per cent of the child care time, they were responsible for 41 per cent of the disciplining of children (the highest husband percentage of all the child care activities for this group). Secondly, one "household task" was "telephoning", of which full-time housewives did 86 per cent. It is not clear whether this includes all phone calls, or only "unpaid work related" ones.
Fletcher concludes that:
"The plight of working women in terms of the long hours they work has been well documented in overseas research, and commented on extensively in the media. The results from this research suggest that, in New Zealand at least, husbands share that plight." (p.39)
Nevertheless, it would appear that, on average, wives spend more time on housework and childcare than husbands.
(b) The Department of Statistics (1991) study
This was a pilot time-use survey of 418 people 12 years or older. The sample distribution only loosely matched the age/sex distribution of the New Zealand population. The publication, Testing Time, gives a very brief summary of the findings. Significant data are described in graph form in figure 4.3 of their publication. Hours have been estimated from the graph for table 2.
Perhaps the most immediate and striking observation is that, for both men and women, there are more than twenty-four hours in a day. This is because some activities may be done simultaneously. For example, the report states that care of children and other dependants as a main activity takes up one per cent of total time recorded, but this figure rises to eleven per cent when included as a secondary activity plus time required for a person who could not be left alone [footnote 3]. This shows that time spent is a poor measure of contribution. An hour of child
TABLE 2
Hours Per Day Spent on all Activities by Participants
Males Females
Labour force activity 7.2 4.6
Domestic work 2.2 3.6
Caring for children and other HH 3.4 14.8
members
Unpaid work in community 1.4 1.6
Education 3.2 2.6
Sleeping 7.8 8.2
Other personal care 3.2 3.4
Passive leisure (incl. TV) 5.2 6.0
Total 33.6 44.8
care could also be an hour of reading, watching TV, or some other leisure activity, or alternatively ironing, cooking, cleaning, sewing, gardening, or some other "household" task (or even paid work from home, such as marking!). The male and female totals suggest that women spend more time on simultaneous activities than men. This holds even if we assume that work and sleep can never be combined with other activities - the remaining hours are still a higher multiple of remaining time for women than for men. This might explain why the surveyed women recorded an additional hour per day on "other personal care" and passive leisure. Interestingly, female labour force activity is 64 per cent of the male figure, which exactly matches the "80 per cent of the years at 80 per cent of the hours" suggestion early in this paper. It leads us to question the meaning of the hours data in Fletcher's study. Is an hour of unpaid work an hour solely devoted to unpaid work, and how many tasks are being done simultaneously? Looked at in another way. How many hours are there in a man's day, 24, or 33.6, or somewhere in between? And for a woman, where does the woman's day lie in the range from 24 to 44.8 hours? Is an hour for a man identical to an hour for a woman?
(c) Callister on the Australian survey
Callister (1993) does not describe the sample on which the survey is based. The data he presents are included in three specific examples from the survey as described in the report for the Office of the Status of Women, Department of the Prime Minister and Cabinet, Australia. They "show the very strong divergence in behaviour between men and women". To quote:
"When men are working a forty hour week they will undertake, on average, just under 15 hours of unpaid work per week, but even when they do no paid work the unpaid contribution rises to only just over 20 hours. For women when they work a forty hour week they will be undertaking just under 20 hours of unpaid work, and this rises to 40 hours if they have no paid work.
as a woman increases her paid work from 0 to 60 hours per week she can expect, on average, for the husband to increase his unpaid work by an hour.
on average new fathers do not increase their unpaid work by a single minute, and only among those whose youngest child is ten to fourteen years does time spent in unpaid work rise at all. However men make substantial adjustments to their paid work time, with fathers of pre-schoolers increasing their paid work by 58% to close to 50 hours per week (travel and breaks included). This represents a peak of men's commitment to work." (p.2)
While the above appears to paint a somewhat unflattering picture of men, a rearrangement of the data indicates that the results are not very different from those of Fletcher. Men working a 40 hour week (paid work) put in a total of 55 hours. Women who are not in paid work put in only 40 hours in total, and women doing 40 hours of paid work do 60 hours in total. A woman increasing paid work from 0 to 60 hours has a long way to go to catch up with the total number of hours put in by a man. Note also that unpaid work may involve simultaneous activities, not all of which need be work.
The fathers of preschoolers figures of 58 per cent and 50 hours seem suspect, but taking the 50 hours as valid, they are doing a total of 65 hours, compared to perhaps 40 (no paid work), or 60 hours (if 40 hours in paid work) by their partners. It seems harsh in this situation to point out that these men do not increase their unpaid work by a single minute.
In summary, women appear to be spending more time on unpaid work, and this may, on average, roughly equal the extra time that men spend in paid work. There are differences depending on labour force participation and the presence of children, in particular preschoolers. It could perhaps be argued that sole-income-earner men with preschool children are most severely disadvantaged compared to their partners in terms of effective hours of full-time effective contribution to the household.
Clearly this is not "disadvantage" if the value of hours spent in the various tasks differs sufficiently to balance this out. This issue is addressed in my questions two and three.
3. THE SECOND QUESTION: HOW MUCH ARE MEN PAYING?
Margaret Shields said, "The invisibility of women's unpaid work is a problem." (Ministry of Women's Affairs, 1989, p9). We could equally well say that the invisibility of men's payments for women's "unpaid" work is a problem.
The standard approach that has been taken in studies of unpaid work is to try to value it indirectly. This overlooks the fact that this work is actually being paid for directly in the "average" household as described above (one partner earning more and the other doing more unpaid work). We can easily construct a formula to identify the relationship under specified assumptions. Assume that one person is the primary income earner and earns more than the other. The income is part of matrimonial property and so is split evenly [footnote 4]. There is therefore a payment of half the difference in (after tax) incomes to the lower earning partner. Assume that the benefits of housework and childcare are shared equally by the partners. The housework done by each partner could therefore be considered half for oneself and half for the partner. The payment compensates for the difference in "unpaid" contributions. Half the difference in contribution earns half the difference in income. Alternatively, the difference in "unpaid" contribution could be valued at the difference in income. We can then scale up to estimate the total value that the household places on "unpaid" work, or at least to estimate the implicit price charged, assuming a constant price per hour.
For the sceptical, the following is based on the assumption of equal contribution, which arises from the Matrimonial Property Act, 1976. Results depend on this assumption, no matter how absurd it (or the results) may be. The interpretation of results is based on the assumption of voluntary participants and competitively negotiated prices which can be revised over time. There may be other interpretations of the results depending on the perspective taken.
Let V be the value of unpaid work, Y refer to income earned and H refer to hours of unpaid work, with subscript 1 referring to the primary income earner and 2 to the other partner. We then have:
If we take the simple case where Y2 = 0 and Y1 is normalised to equal 1, the formula reduces to:
where x is the share of total unpaid work done by the "homemaker".
We can plot this relationship to give the value of unpaid work measured in terms of multiples of the income earner's earned income. The result is as in figure 1.
If we are to measure the welfare of the household as the sum of values of paid and unpaid work, the optimal point [footnote 5] is where the main income earner does just under half the unpaid work. Assuming no change in earnings, the household tends to a point of being infinitely well off as share of unpaid work increases to this point. Once this point is passed, the household becomes infinitely badly off. If the household is constrained so that the main income earner must do more than the other partner, then the second best solution is that the main income earner does all the unpaid work also. As with firms, there may be a shut down point which is preferable to this. Paradoxically (but note footnote 5), the optimum point can be reached either by one person increasing, or by the other person reducing, the unpaid work contribution. If unpaid work contribution is assumed to be independent of paid work contribution, household welfare increases as the primary income earner's income increases and the secondary income earner's income falls, as long as the primary income earner does less than half the unpaid work.
FIGURE I
To relate the above analysis to the hours worked data, we can say the following about "average" households. Using Fletcher's data for households with full-time housewives, the housewife contributes 37.06 hours more of unpaid work, which is balanced by the husband's 48.67 hours of paid work. This translates into one hour of unpaid work being worth 1.31 hours of paid work by the husband. For working wives, we should allow for the 25 per cent extra pay per hour obtained by the husband [footnote 6]. This means that the wife's 32.77 hours of paid work is equivalent to 26.22 hours by her husband. His extra 21.48 hours of paid work is balanced by her extra 16.73 of unpaid work. One hour of unpaid work is therefore worth 1.28 hours of paid work by the husband.
Department of Statistics data are not directly comparable because they do not relate solely to couples, some of the activities are simultaneously done (this may also be the case for Fletcher, but that is not spelled out), and labour force activity includes commuting time, which is unpaid.
The Australian data from Callister do not give averages. If both partners work 40 hours, applying the New Zealand average ordinary time earnings ratio, 8 hours of paid work by the husband is balanced by 5 hours extra unpaid work by the wife, and an hour of unpaid work is equivalent to 1.6 hours of paid work by the husband. Fathers of preschoolers married to full time housewives balance 50 hours of paid work against 25 hours of unpaid work, giving a ratio of 2:1.
All these studies would indicate (using this methodology) that unpaid work is more valuable than paid work. Given that the Matrimonial Property Act applies to all married couples, the equal contribution assumption also applies to all married couples. The finding that unpaid work is worth more on average therefore applies irrespective of the amount paid for the paid work, it being based solely on relative hours worked and not rates of pay.
Studies such as the above look at contribution during the relationship. They make no allowance for contributions before the relationship started or after it ended, nor do they consider methods of valuation of property for distribution on termination of the relationship. These are significant omissions. To illustrate their importance, I shall focus on the specific areas of superannuation and child support.
Superannuation
The following is taken from a standard document on valuation of superannuation contributions for matrimonial property purposes, as provided by KPMG Peat Marwick in 1993. Other valuations may vary in their approach, but it is my understanding that this particular one is not unusual. This view is possibly supported by the reference to legal cases to justify assumptions made:
"The valuation of an interest in a defined benefit scheme must observe the principles established in the leading case of Haldane v Haldane ....
The following is a summary of the key elements in their preparation of a Haldane type valuation.
(a) The separation date salary is used to determine the benefits payable at the first date of retirement which does not require consent of the employer. The deemed retirement date in this case is therefore aged 60.
(b) A real tax paid interest rate is used to determine present values of the entitlement. Currently a rate of 3% is appropriate taking into account interest yields in the market place. The rate was adopted in a recent High Court superannuation cases of Black v Black and Walkinshaw v Walkinshaw.
(c) Pre marriage contributions (if any) are deemed to be matrimonial property (Kilgour v Kilgour).
(d) The NZ Life Tables 1985-7 are used to determine life expectancy.
(e) The matrimonial proportion consists of the number of completed months of contribution up to the date of separation divided by the total anticipated months of contribution up to normal retirement.
(f) [A contingency deduction is necessary considering that a "willing purchaser" would wish to make allowance for change in tax regime, or interest rate, death at other date, loss of liquidity, etc.. For two such valuations for people in their early 40s, a global deduction of 30% was used, and legal precedents cited].
(g) The spousal benefit is to be regarded as part of the total package of benefits notwithstanding the possible loss of benefit through the subsequent dissolution of the marriage (Wilkinshaw v Wilkinshaw)."
The following points should be noted:
(i) From (a), retirement is assumed to be at 60. It is more likely that someone under 45 would retire at 65.
(ii) From (b), 3% seems a very low rate for discounting to present values. Given that the payment in cash is likely to result in increased mortgage for the payer and reduced mortgage for the recipient, a case could be made for a rate as high as 8% or more.
(iii) From (c), the inclusion of premarriage contributions means that work effort from before the marriage is also a contribution to the marriage. It is therefore inappropriate to consider only work done during the marriage, as occurs with time diary comparisons. In particular, work done during one marriage may also contribute to the matrimonial property of a later marriage. A person may not just be making superannuation contributions for the benefit of the current spouse, but also for future spouses.
(iv) From (f), the 30 per cent contingency deduction does not even cover the effects of later retirement (65 in place of 60) for a man, given a life expectancy of 74 years. Later retirement reduces the present value to 62 per cent of the original sum. The added effect of an 8 per cent discount rate in place of the 3 per cent used, means that the present value for a 42 year old, as in the two examples made available, is only 17 per cent of the calculated present value before contingency deduction. In this case the "half share" awarded to the spouse is in fact a double share, without any allowance for the loss of liquidity, loss of spousal benefits, and other risks born by the payer. Even discounting at only 4 per cent reduces the present value to 79 per cent with retirement at 60, or 47 per cent with retirement at 65.
(v) From (g), while I do not have figures for the size of the spousal benefit, it is clear that the entire burden of any loss in value resulting from separation is born by the policy holder.
The purpose of the above is to illustrate that some assets accumulated before marriage can and are considered part of matrimonial property for "equal" division on separation. Given the longer time in work and greater earnings of men, this means that, on average, there is a net payment from men to women out of earnings from before marriage. It should be noted also that only three years of marriage is required for a spouse to obtain this entitlement to 50 per cent of the estimated present value of a superannuation policy.
Also, in the specific case of superannuation, supposedly equal division of assets may be far from equal, at present at least. Women are generally likely to benefit from the inequality.
In this context, it is noteworthy that calls are being made for additional radical changes in favour of women. Bridge, a law lecturer writing on family law policy in New Zealand, (Bridge, 1992), calls for a "forward looking" perspective, where "Equality must move from its focus on asset division to the achievement of an equal financial result between the spouses" (p.252). She is saying that future earnings differences should be taken into account so that financial situations should be equal. This is really saying that, as far as finances are concerned, the "equal division" consideration within a marriage should be applied for life. She makes no such requirement in relation to unpaid work. Her suggestion is effectively saying that there should be no financial cost to women arising from separation, save the effects of no longer pooling assets and income. While some reference is made to children to "justify" her suggestion, her arguments at times ignore the effects of child support payments, and apply even if the woman does not have custody of children. Her suggested redistribution could occur solely on the basis of earnings differentials. Nevertheless, the use of children to support her position further links custody and finance in a way that can be very harmful to fathers and to children [footnote 7]. Proposals are currently under consideration that would extend the Matrimonial Property Act to apply to de facto relationships also.
Child Support
Child support is a complex and controversial topic. Issues of payment and access, financial and non-financial responsibilities to children, plus benefits of contact versus the payment-but-little-contact situation of many non-custodial parents are all relevant. There are also problems in separating benefits for children from benefits for parents. Hence, in terms of desired outcomes, a claim that "children's living standards" should be protected (Bridge, 1992, p.252) is hardly distinguishable from a claim that "custodial parents' living standards" should be protected.
Given these complexities, the payment question that this paper is addressing is not as clear cut as in the case of superannuation. While I shall not go into detail on these issues, the following points would be relevant in some cases.
Two of the objectives of the Child Support Act 1991, as described in Child Support Review 1994: A consultative document, (1994, p.5) are:
To improve equity between custodial and non-custodial parents in respect of the costs of supporting children.
To put in place a more effective system of offsetting costs of income support to the State by ensuring that non-custodial parents make a fair contribution.
The same document (p.11) states that the maximum liability "reflects the maximum cost of keeping a child". There are clearly situations where the non-custodial parent is paying to support the custodial parent, and cases where the non-custodial parent is paying the full cost of supporting his/her children. This is in terms of statutory obligations only, and makes no allowance for voluntary outlays or costs in relation to NCP contact with children, which could include up to 40 per cent of nights over an extended period. According to the consultative document, 84 per cent of lone parents are women. There are therefore some cases where payments continue to be made over and above an equal share of the cost of raising the children from the relationship, although this may not be universal. Where this translates effectively into ex-spousal support, financial contributions continue after the end of a relationship. This would not be the case for non-custodial parents on low income whose ex-partners are not on a benefit, for example. The situation is further complicated by the continued question of financial versus non-financial contributions to the upbringing of the children, and the extent to which the obligations are as would have been chosen voluntarily. The presumption of a 50-50 split of custody, paralleling the matrimonial property presumption, might be appropriate, and the disregard for custodial parent income in the child support assessment formula conflicts with the equity objective stated above [footnote 8].
4. THE THIRD QUESTION: ARE THEY GETTING VALUE FOR MONEY?
Rather than address this question directly, I would rather consider the nature of the decision that a man makes to purchase the product of unpaid labour, and the associated price setting mechanisms. This has parallels with the approach of looking at market structure to determine whether prevailing market prices are appropriate. The existence of distortions can be taken to indicate that the resulting prices may be undesirable.
Several distortions are apparent:
(i) The financial price paid is set according to the circumstances, past, present and future, of the payer. This is the reverse of the case in normal market transactions, where the price is specified in terms of the good or service being purchased. If two men, A and B, wish to marry the same woman, but A earns twice as much as B (and they have no assets), the price for A would be twice the price for B. It is not clear that A would be getting any more than B in return. This is a case of perfect price discrimination [footnote 9]. It is also likely to be highly inefficient in terms of utility maximisation. This is because the share of income spent on a spouse is fixed, so it is not possible for A to pay B's price and have more money for other purchases, retirement, or children from an earlier marriage, for example.
(ii) As a man's income changes, spousal payments change proportionately. This is entirely independent of spousal contribution to the relationship. Disregarding any public good components to household expenditure, this means that he faces a very high effective marginal tax rate (perhaps we should talk about a "relationship tax" which is levied to provide transfer payments to his spouse). This will affect the labour-leisure decision, but in which direction is not clear.
(iii) The amount of unpaid work done by the spouse is independent of the payment made, ceteris paribus.
(iv) Once a marriage has been entered into, there is no competition for the supply of unpaid labour. Given the high entry and, perhaps more importantly, exit costs to the purchaser, it is doubtful that the market could even be called contestable. If we view the relationship as one of employer-employee, there is the extraordinary situation that redundancy or resignation carries a redundancy (or resignation) payment equal to half the value of the "company". If we view the relationship as a partnership, distribution of income and assets within the partnership is based on presumed equal equity holdings and (usually) takes no account of possible differences in labour input.
(v) A case could be made that legislation and discussion about legislation is strongly influenced by the presumption that marriage is for life, with the minor allowance for marriages of "short duration", i.e. up to three years. The reality for many couples is very different. The regulations set by law are therefore often likely to be inappropriate, particularly for second and subsequent marriages.
(vi) Law changes can occur during a marriage that alter the terms of the agreement voluntarily entered into. There is little scope for renegotiation at that point.
(vii) There are no penalty clauses if either of the parties fails to deliver, as with say a wife who refuses to undertake paid work, or who chooses a more pleasant, but lower paid job than was expected, or otherwise underperforms in relation to unpaid work. (The same could be said for men, but we are looking here at what the man receives) [footnote 10].
As a further point, there is no guarantee that the results of paid and unpaid labour during a relationship benefit both partners equally. The question of equality on the financial side has been studied (McKinlay and Easting, 1994), but I have not come across any studies on the extent to which unpaid labour is applied to meet individual partners' needs.
5. CONCLUDING REMARKS
To conclude, I shall return to the three questions raised in the introduction. On average, in return for the larger financial contribution, men do appear to receive more non-monetary benefits than they provide. The measures available for this are poor, and there are issues of definition of tasks, simultaneous performance of activities, and the use of an input measure (hours spent) as a proxy for output [footnote 11].
It is notable that, on average, unpaid work appears to be valued more highly, on an hourly basis, than paid work. The greater men's contributions and the smaller women's contributions become to unpaid work, ceteris paribus, the greater the difference in values becomes. As the results of paid work (i.e. earnings) are transferable over time and distance, it is possible for financial transfers to be made outside the time of a relationship. This would tend to penalise men as their financial contributions are generally greater.
In summary, there are major distortions in the domestic unpaid labour market. It is likely that some men lose and some obtain a net benefit. It is highly unlikely, however, that optimal outcomes are frequently achieved, and it would appear that, on average "unpaid" work is highly priced.
FOOTNOTES
REFERENCES
Alchian, A.A. and Allen, W.R. (1968), University Economics (2nd ed.), Belmont: Wadsworth.
Bridge, C. (1992), "Reallocation of Property After Marriage Breakdown: the Matrimonial Property Act 1976", in Henaghan, M. and Atkin, B. (eds.) (1992), Family Law Policy in New Zealand, Oxford: Oxford UP.
Callister, P. (1993), "Girls Can Do Anything So Why Can't Boys?", New Zealand Association of Economists Conference, Dunedin, September.
Child Support Act Working Party (1994), Child Support Review 1994: A consultative document, Wellington: Inland Revenue.
Cowan, C. and Kinder, M. (1986), Smart Women Foolish Choices, New York: Bantam.
Department of Statistics (1991), Testing Time: Report of the 1990 Pilot Time Use Survey, Wellington: Department of Statistics.
Dernburg, T.F. and McDougall, D.M. (1969), Economics (3rd ed.), New York: McGraw-Hill.
Dolan, E.G. (1983), Basic Economics (3rd ed.), New York: Dryden.
Eisner, R. (1978), "Total Incomes in the United States, 1959 and 1969", Review of Income and Wealth, 24(1), pp.41-70.
Fisher, R.L. (1977), The Matrimonial Property Act 1976, Wellington: Butterworths.
Fletcher, G.J.O. (1978), "Division of Labour in the New Zealand Nuclear Family", New Zealand Psychologist, 7(2), pp.33-40.
Horsfield, A. (1988), Women in the Economy, Wellington: Ministry of Women's Affairs.
Juster, F.T. and Stafford, F.P. (1991), "The Allocation of Time: Empirical Findings, Behavioural Models, and Problems of Measurement", Journal of Economic Literature, XXIX, pp.471-522.
Kendrick, J.W. (1979), "Expanding Imputed Values in the National Income and Product Accounts", Review of Income and Wealth, 25(4), pp.349-363.
McConnell, C.R. (1978), Economics: Principles, Problems, and Policies, (7th ed.), New York: McGraw-Hill.
McKinlay, R. and Easting, S. K. (1994), "Pakeha Couples' Access to Family Money", paper at the Public Health Association of New Zealand 1994 National Conference, Palmerston North.
Ministry of Women's Affairs (1989), Synopsis of Valuing Women's Unpaid Work Project 1989/90, Wellington: Ministry of Women's Affairs.
Waring, M. (1988), Counting for Nothing: What Men Value and What Women are Worth, Wellington: Allen and Unwin.