Radical
Statistics
The gendering of spending within households
Jan Pahl
There is now an enormous literature on the domestic division of labour, which shows that in general women spend more time shopping than men. So for example, in 1996 fathers spent on average two hours 50 minutes shopping each week, compared with mothers' average of five hours 50 minutes (Office of National Statistics, 1997, 215; see also Anderson, et al., 1994; Miller, 1998; Morris, 1990). However, this literature leaves many questions unanswered about the details of who spends on what. Do men and women buy the same sorts of items or is there a division of labour within the general category of 'shopping'? What difference does employment status or age make to shopping patterns? And how are new forms of money, such as credit cards, being incorporated into the activity of shopping?
This article has three main aims: (i) to consider detailed patterns of spending within households, and in particular the extent to which spending is gendered; (ii) to explore patterns in the use of credit cards, since this research was done in the course of a study of new forms of money, such as credit and debit cards, telephone and internet banking (Pahl, 1999); and (iii) to demonstrate the wealth of data available within the Family Expenditure Survey (FES) and the opportunities it offers to researchers working in the area of families and households.
Methods
The Family Expenditure Survey is a long-running continuous survey, carried out by the Office for National Statistics (ONS), using a random sample of households drawn from every country in the United Kingdom (Office of National Statistics, 1996). Interviews take place in about 7000 households every year, and involve all 'spenders' over the age of 16. The overall response rate is 70 per cent. The survey collects information about the income and out-goings of the household as a whole, and in addition each spender is asked to complete an expenditure diary over a two week period. Respondents are asked to note in the diary every single item which they have bought, and to record if a credit card was used to make the purchase.
Since the study was concerned with the use of new forms of money within couple households, the re-analysis began by selecting households containing a married couple. The FES does not identify stable cohabiting households, so these could not be included in the analysis; also excluded were households where the 'head of household' was under 20. This produced a total of 3691 couples, or 7382 individuals. However, there were a few cases in which crucial data was missing, so the total sample for the analysis numbered 3676 couples.
The data derived from the 1993/4 expenditure diaries were re-coded in order to make the analysis more manageable. There were some 315,738 entries in the diaries kept by the couples who were the focus of the analysis, and in the raw data set they had been classified by FES into several hundred expenditure categories. These were re-coded to produce a total of 37 major categories, ranging from food, motor vehicles and household goods, which were the largest items in the diaries in terms of amounts spent, to clothes, holidays, gambling and alcohol. The re-coding involved grouping linked items together. So the category 'motor vehicles' included, not simply the purchase of cars and motor bicycles, but also petrol, vehicle insurance, tax, repairs and other running costs; the category 'household goods' included purchases of bedding, furniture, carpets, kitchen appliances, lamps and clocks.
It is important to remember that many items of household expenditure were not recorded in the diaries because they were paid by direct debits or standing orders out of bank accounts; these included payments for mortgage or rent, for council tax and for utilities such as water and electricity. Some people paid these by cash or cheque, and noted them in their diaries, but since many did not these items were excluded from the analysis.
The analysis also made use of the wealth of socio-economic information contained in the FES, and especially the information about age, education, employment, household composition and social class.
The gendering of spending
The analysis began by examining the different expenditure responsibilities of men and women within households. Table 1 is based on the whole sample, that is all 3676 couples. The expenditure categories in the table were chosen because they represented significant elements in most household budgets, with the two largest categories being spending on food and motor vehicles, which represented on average 16 per cent and 11 per cent of the total amount recorded in the expenditure diaries. The items were also chosen because they were unlikely to have been paid by direct debit or standing order; in other words all spending was likely to have appeared in the expenditure diaries.
Table 1 presents women's expenditure on particular items as a percentage of total expenditure by the couple on that item, and also gives information about credit card spending.
Click HERE for Table 1: Women's Expenditure as a Percentage of Total Expenditure and of Credit Card Expenditure
Table 1 suggests that patterns of spending were highly gendered. Women spent more than men on food, women's and children's clothes, child care and educational courses. Men spent more than women on alcohol, motor vehicles, repairs to the house, meals out, gambling and holidays. The balance between men and women in spending was more even when it came to household goods, medical and dental expenses, tobacco and recreation. Rather surprisingly, perhaps, women were responsible for two fifths of the amount spent on men's clothes
The table also shows that credit cards were used quite selectively. The expenditure categories for which credit cards were most often used were men's clothes (32 per cent bought with a credit card), holidays (29 per cent), women's clothes (28 per cent) and household goods (24 per cent). By contrast, credit cards were rarely used for paying course fees and bills for household repairs, for gambling or buying tobacco.
Table 1 also shows that women were, in general, less likely than men to use a credit card to make a purchase. For nearly every item the percentage of credit card spending made by women was less than their percentage of the total spending on that item. The only real exception to this was spending on alcohol, perhaps because women were more likely to buy alcohol from supermarkets and off licence shops, where credit cards were acceptable, while men bought more alcohol in pubs, where cards were less welcome.
However, Table 1 leaves many questions unanswered. For example, do women buy men's clothes because they are not in employment, and so are more likely to be free when the shops are open, or is it something that wives are expected to do, even when they have a job? Do men spend more on meals out because they tend to pay when the couple goes out, or is their larger expenditure a reflection of the fact that being in employment involves eating away from home?
Table 2 suggests some tentative answers to these questions, and also illustrates some of the complexities which lie behind Table 1. It shows selected items of expenditure, chosen because relatively large sums of money were involved and because there was a range in terms of whether men or women were the main spenders. Some items, such as food and motor vehicles, appeared in the expenditure diaries of most couples; other items, such as holidays and repairs to the house, were purchased by a minority of couples but were quite costly when they did appear.
Click HERE for Table 2: Impact of Women's Employment on Expenditure patterns within Households.
The figures for mean spending are based only on those individuals who recorded expenditure on the item in question during the two weeks when the diaries were being kept. The table presents the differences between 857 couples where both partners were in full time employment and 688 couples where the man was in full time employment, while the woman was defined as 'economically inactive' or 'unoccupied', in the disparaging words of the ONS.
Table 2 suggests that women's employment status made a significant difference to patterns of spending within households, but that many other different variables were relevant. Women spent significantly more on food when they were not in paid work. Men with wives in full time employment spent a little more on food than those whose wives were not in paid work, but the difference was not significant, suggesting that the employment status of their wives did not cause these men to alter their behaviour. Women's continuing responsibility for buying the food was underlined by the fact that even when both partners were in full time employment women were still responsible for three quarters of spending. Women's employment status did not affect spending on household goods and on motor vehicles for either men or women.
Spending on holidays underlined the effects of women's employment status on the expenditure patterns of both partners. Compared with couples where both were in full time work, men who were the sole earners spent significantly more on holidays, nearly doubling their expenditure; by contrast, women who were not in paid work spent half as much on holidays as women in full time employment. Men's spending on their own clothes did not reflect their wives' employment status. However, women in paid work spent significantly more on this item, suggesting that women were responsible for buying men's clothes, not because they were free when the shops were open, but because for some couples buying her husband's clothes was part of a wife's domestic work, and when she earned more she spent more.
Spending on repairs to the house reflected employment status in quite complicated ways, though here only the male difference was statistically significant. Couples where both were in full time employment spent much more on this item than couples with only one earner, suggesting that they were paying for repairs to be done rather than doing repairs themselves. The fact that the difference was statistically significant for men but not for women was puzzling, but it may be that women without paid work were taking responsibility for painting and decorating, while men whose wives were in full time employment recognised that they would have to pay a professional to do this work.
Differences in other areas of spending seemed to reflect the greater spending power and increased autonomy of women who were earning. Spending on meals out was greater for both men and women when both were in full time employment, but the difference was particularly marked for women. There are many possible explanations for this finding. It may be that the additional spending of women represents eating lunch at work or going out for drinks with work mates; it may be that couples in full time employment have less time and energy for cooking and more money for eating out; or it may be that they are more likely to share the bill when they go out. Women who were in full time work were significantly more likely to spend money on gambling, though they still spent much less than men.
Up to this point the focus has been on the ways in which spending is gendered, with men and women tending to take responsibility for different items of expenditure. The next step was to look for patterns in the amounts spent and in the use of credit cards.
Gender and the costs of things bought
Table 3 underlines the enormous differences between men and women in the number and costs of the items which they buy. Women made many more low cost purchases than men, and many fewer high cost purchases: so nearly three quarters of items under £10 were bought by women compared with under two fifths of items over £1000. The result was that, over the two weeks when they were keeping the expenditure diaries, the women involved bought 217,410 items, as opposed to the 98,328 bought by men.
Click HERE for Table 3: Expenditure on Items of Different Cost by Gender of Purchaser and Use of Credit Cards.
Previous research has suggested that people prefer to use different forms of payment for purchases of different cost, with a trend away from cash towards credit cards as cost increases, except for very expensive items which tend to be paid by cheque. There is also a trend over time away from cash and towards credit cards as a means of payment. Thus a survey of 2000 adults, carried out for Girobank, showed that cash was the preferred means of paying for items costing under £10. For purchases of £100, cash was preferred by 25 per cent of those surveyed, down from 28 per cent a year earlier, while cards were preferred by 43 per cent, compared with 39 per cent a year earlier. For purchases costing £1000 cheques remained the first choice for 46 per cent, compared with 33 per cent who said they would use a card and 10 per cent who would use cash (Graham, 1997).
The new analyses of the FES data produced a similar pattern, with credit cards being used less often for items costing under £10 or over £1000, as Table 3 shows. Credit cards were most likely to be used for goods costing between £100 and £1000, where 15 per cent of payments were made in this way. More detailed analyses, not shown in the table, revealed that the peak of credit card use came in the price range of £250 and £500, where 38 per cent of payments involved credit cards.
In general, women's share of credit card expenditure was similar to their share of total expenditure, though men were slightly more likely than women to use a credit card. The greatest gender difference came in the price range of £500 and £1000, where only 33 per cent of all the credit card purchases were made by women.
Treating the sample as a whole, however, obscures variations between different sub-groups: as we have seen, women's full time employment made a significant difference to their expenditure patterns. However, this does not take account of the divide between full time and part time employment for women; nor does it explore the effects of unemployment and retirement. The next stage in this examination of spending and of the use of credit cards was to sub-divide all the couples in the sample according to employment status.
Employment and spending patterns
Previous research had suggested that two aspects of employment were particularly relevant. These were, first, the employment status of women, with a distinction being made between full time, part time and no employment, and, secondly, household employment status, with distinctions being made between 'work rich' and 'work poor' households in particular. In order to examine the effects of these two aspects of employment, a variable was created which combined the employment situations of both the man and the woman. This involved re-coding the employment variables given by the FES to make six broad household employment categories. (For further details about the categories see Pahl, 2000).
Table 4 suggests that gender, employment and age all have an impact on the use of credit cards. The table gives the percentage of men and women, in each employment category, who used a credit card to make a purchase over the two week period during which they kept the diary for the FES. Significance was tested by chi squared test, N being the number of households with expenditure on the item in the two week period.
The results suggest that between a quarter and a third of all those who took part in the survey used a credit card over this period, with men being slightly more likely than women to use cards. Differences in employment status were associated with differences in the use of credit cards. When the man and the woman were both in full time employment they were equally likely to have used a credit card. However, women in part time employment were less likely, and women without employment very significantly less likely to have used a credit card, by comparison with their employed husbands.
Click HERE for Table 4: Percentages of Individuals Using a Credit Card for Making a Purchase over a Two Week Period, by Household Employment Categories.
The table highlights the exclusion of unemployed people from the credit card economy, with only a very few individuals in this category using a credit card during the two weeks. This is consonant with research on access to credit more generally, which shows that low income households find it hard to obtain credit; if they do have to borrow they tend to be forced to contact more expensive money lenders than the typical credit card company (Ford, 1991; Kempson et al., 1994; Kempson and Whyley, 1999).
Retired people were the other group which appeared to be relatively excluded from the credit card economy. This may partly have been a result of low income and lack of credit worthiness. But it may also have been a consequence of a lack of financial confidence in new forms of money and of a general mistrust of getting into debt. There was a significant difference between men and women among retired couples, which may have reflected gender differences in income or a tradition of male dominance in financial matters among this age group.
Conclusions
Detailed analyses of spending patterns reveal that some very traditional gender roles persist. The fact that women are overwhelmingly responsible for paying for child care is very relevant to debates about provision for the children of employed parents, and about who pays for child care. Debates about the domestic division of labour have not so far acknowledged that 40 per cent men's clothing is actually bought by women, presumably because men cannot be bothered to buy their own. And gambling and drinking are clearly still male preoccupations. The evidence suggests that women continue to be responsible for the collective expenditure of households, whereas men are more preoccupied with individual as opposed to collective expenditure. (For a brilliant analysis of shopping for the family as an act of love, see Miller, 1998).
There are also gender differences in the average cost of the items bought. Typically women are responsible for buying the mass of small items which every household requires, while men buy the larger, more expensive, more rarely-purchased items. So the time which men and women spend 'shopping' tends to be differently spent and more time-consuming for women.
The use of credit cards also reflects gender, employment status and age. Overall men are slightly more likely than women to own and to use a credit card. However, women without incomes of their own, retired people and unemployed people are all significantly less likely to use a credit card to make a purchase: their access to credit cards and other new forms of money is constrained by their social and economic circumstances (Pahl, 1999)
Acknowledgements
The work reported in this article is part of a study on 'New Forms of Money and Family Financial Arrangements', which was funded by the Joseph Rowntree Foundation and the University of Kent at Canterbury. I am grateful to the funders, and also to the ESRC Data Archive, at the University of Essex, and to the Office for National Statistics, for making the Family Expenditure Survey available to researchers. Lou Opit worked on the analyses of the FES until four days before his death: this article is dedicated to him with my love.
References
Anderson, M., Bechhofer, F. and Gershuny, J. (1994) The Social and Political Economy of the Household, Oxford: Oxford University Press.
Ford, J. (1991) Consuming Credit: Debt and Poverty in the UK, London: Child Poverty Action Group.
Kempson, E., Rowlingson, C. and Whyley, C. (1994) Hard Times: How Poor Families Make Ends Meet, London: Policy Studies Institute.
Kempson, E. and Whyley, C. (1999) Kept out or Opted out: Understanding and Combating Financial Exclusion, Policy Press: Bristol University.
Miller, D. (1998) A Theory of Shopping, London: Polity Press.
Morris, L. (1990) The Workings of the Household, London: Polity Press.
Office of National Statistics (1996) Family Expenditure Survey, London: Stationery Office.
Office of National Statistics (1997) Social Trends 27, London: Stationery Office.
Pahl, J. (1999) Invisible Money: Family Finances in the Electronic Economy, Policy Press: Bristol.
Pahl, J. (2000) Social polarisation in the electronic economy, in Renewing Class Analysis, edited by R. Crompton, F. Devine, M. Savage and J. Scott, Oxford: Blackwell.
Jan Pahl
School of Social Policy, Sociology and Social Research
University of Kent at Canterbury
Canterbury, Kent
CT2 7NY
Tel: (01227) 764000
Email: J.M.Pahl@ukc.ac.uk
