Final Food Consumption Level Based on Expenditure and Household Production

Danielle E. Aldershoff, Institute for Consumer Research (SWOKA), The Netherlands
ABSTRACT - This article gives a description of the volume of expenditure on food and appliances and of household production relating to food and meal preparation in different households. It summarizes money and labor inputs into one unit and examines how this total figure, the final food consumption level, differs in various types of households. For this purpose available Dutch money budget and time budget survey data from 1980 have been analyzed via cross-section.
[ to cite ]:
Danielle E. Aldershoff (1985) ,"Final Food Consumption Level Based on Expenditure and Household Production", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 209-214.

Advances in Consumer Research Volume 12, 1985      Pages 209-214


Danielle E. Aldershoff, Institute for Consumer Research (SWOKA), The Netherlands

[The author would like to thank her two colleagues W. Baak resp. H. Kasper as well as L. Goldschmidt-Clermont and anonymous ACR-reviewers for their helpful comments.]


This article gives a description of the volume of expenditure on food and appliances and of household production relating to food and meal preparation in different households. It summarizes money and labor inputs into one unit and examines how this total figure, the final food consumption level, differs in various types of households. For this purpose available Dutch money budget and time budget survey data from 1980 have been analyzed via cross-section.


In order to provide for their needs of food consumers spend a certain amount of money (on food and appliances) and perform a certain amount of household production (shopping, food and meal preparation) There are not many consumers in industrialized countries who can indulge themselves in always eating out and there is no opportunity for all to grow one's own food. In a Dutch household on the average 2196 of all expenditure and 3596 of the total household production is spent on food consumption. In The Netherlands circa 2.5 million unpaid labor years are spent by all households on household production relating to food consumption. At an estimate the share of this unpaid labor input varies from 26% to 33% of the Gross National Product and from 29% to 40% of the National Income (calculated from figures as described by Bruyn-Hundt, 1983).

Mostly the food consumption level is measured by the expenditure on eating and drinking, overlooking the foods and meals prepared by the households themselves. The aim of this article is to indicate the differences in the food consumption level between different types of households, including consumption of self produced goods and services. This level will be referred to as "final food consumption level", and is defined as the sum of money and labor inputs.

"Money input" refers to the summarized expenditure on food and appliances. "Food" includes all products and services relating to eating and drinking at home and out of the house. "Appliances" include purchasing costs of crockery, cutlery, tableware, kitchen utensils, small and major household equipment used in meal preparation.

"Labor input" refers to the volume of unpaid household production relating to meal preparation, expressed in number of hours, or in guilders by means of multiplication by a price or wage component. "Meal preparation" includes all activities relating to food shopping (including food fetching from snackbar, Chinese take-out, transport and waiting, putting purchases away) and preparation of meals (sandwiches/hot meals, baking, including laying and cleaning the table, washing up).

In this article the following two questions will be examined:

1. what is the volume of expenditure and household production spent on food consumption in different types of households

2. what is the distribution of the final food consumption level between different types of households?


Expenditures on food and appliances

The Dutch situation cannot, of course, be directly compared with that in other countries; in particular where possibilities of eating-out and growing one's own food are concerned, differences in expenditure will exist. However, results from Dutch budget surveys (f.e. Smits 1982) about the influence of the size and life cycle of the household agree with what has been described elsewhere (f.e. U.S. Department of Labor 1983 and Uusitalo 1979).

From budget surveys the following is known about the expenditure on food and appliances:

- the total expenditure on food increases with the size, the life cycle and the income of the household;

- the share of food in the total expenditure pattern is smaller with higher income;

- with larger households, the expenditure on food per person is smaller (independent of income level);

- one-person households consume a relatively large amount of food out of the house (in the Netherlands, below the age of 65 more than a quarter of their total expenditure on food);

- households without children devote a somewhat lower share of expenditure to the sum of food and appliances than households with children;

- summarized expenditures on food and appliances vary with the life cycle of households; older households spend relatively more per person than younger households (except 65+ households);

- summarized expenditures on food and appliances per person are slightly lower in households with a low income than in households with a high income;

- the expenditure on appliances increases with the size of the household and the level of income (in Dutch households its share forms about 496 of the total expenditure on food i.e. as far as purchasing costs are concerned; fuel and depreciation costs are excluded).

Household production relating to meal preparation

A method frequently used to measure the volume of household production is the time budget survey. As in all survey methods this also has some serious limitations inherent in the commonly used techniques for collection of time use data (see As 1982; Walker 1980, p. 127-129; Aldershoff et al. 1983, p. 5). On the basis of results from time budget surveys the following conclusions can be drawn on the labor input in household meal preparation (Walker & Woods 1976, p. 69-75; Santti et al. 1982; Aldershoff et al. 1983, p. 18, 61):

- the total time spent increases with the size and the life cycle of the household;

- with larger households, the time spent per person is smaller (independent of income level);

- lower income groups spend a little more time than higher income groups (regardless of household type);

- household production of meals varies with the life cycle of the household; older households spend relatively more working hours than younger households;

- the number of working hours per person in households with a low income is slightly higher than in households with a high income;

- higher income groups have in general more household equipment at their disposal (regardless of household type);

- appliances can result in higher time productivity, leading to an increased output, which does not necessarily results in net time saving (Gershuny, 1982) (except for the dishwasher).


Economic evaluation of household production

In order to bring the money and labor inputs under one denominator, both quantities must be expressed in the same unit. Since expenditure is expressed in money and the labor input mostly in time, the core of the problem is to find a suitable price or wage component to express both inputs either in money or in time. There is no ready-made solution to the choice of a particular price or wage component. This is inherent in the nature of the problem. After all, if household production were to be performed for payment (thus took place via direct market transactions), then it would not be what it is today, the structure of households would be different and so would the economic and social balance (Chadeau & Fouquet, 1981: 43).

The unpaid labor input in the household (expressed in the number of working hours) can be assigned an imputed value via two different approaches:

1. As a "foregone wage" or "opportunity cost of time" for the members of a household; that is to say, what income would they have been able to earn on the paid labor market instead of performing the household production themselves? Thus this refers to an estimated nett income exclusive of tax and social charges.

2. As a "foregone expense" or "saved expense" for the members of a household; that is to say, what costs has one saved by performing the unpaid household production oneself instead of buying market services or instead of letting someone else do it for payment? Thus this refers to estimated gross expenditure or gross wages inclusive of social charge and holiday allowance.

For both approaches numerous monetary measures are conceivable and are also used in research, each with certain advantages and disadvantages. (It does not come within the scope of this article to go into this in depth; see further Goldschmidt-Clermont 1982, p. 13-18).

None of the methods emerges as the best evaluation method.

After a methodological review of about 75 publications Goldschmidt-Clermont (1982) concludes that output based evaluations of household production are not numerous. Most studies evaluate the input, whereby they make use of a monetary measure. (For future research, the value added approach is considered as a promising alternative, see Goldschmidt-Clermont 1983 and Chadeau 1984).

A few examples of calculating the money value of unpaid household meal output are known.

In these studies the imputed value of meal preparation was calculated by multiplying the total number of meals prepared and eaten at home by:

- an average weighted price of meals in restaurants or snacks in cafes in France (Chadeau & Fouquet 1981, p. 47-49; Chadeau 1984, p. 5-6).

- labor cost prices of meals served by the State Catering Centre in Finland (Suviranta & Mynttinen 1981, p. 23-24);

- average expenditure of the household concerned for meals if purchased away f rom home in the USA (Volker & Bivens 1983).

In the French as well as the American study the added value of meal preparation was then calculated by subtracting from this output the costs of gas, electricity and depreciation of household equipment used in the preparation of meals. In the Finnish project the added value of home baking and preserving has also been calculated, consisting of the difference between the consumer price of bought products and the costs of corresponding homemade products, including price of raw materials, energy and packaging (Santti & Vesikkala 1983, P. 137-138).

Aggregation of money and labor inputs

As far as is known there are only a few studies, in which the financial resources and household production (in general, not separately for meal preparation) have been aggregated at the micro-level and brought under one denominator.

The view of Sirageldin ( 1969) is that a good measure of the welfare level of a household should account for differences in household structure as well as for the cost of time devoted to productive (paid and unpaid) activities. He therefore defines welfare as a product of "full income" (per equivalent adult unit) and leisure time available (as a ratio of non-sleeping time). "Full income" is defined as the sum of disposable income and total non-market output. The imputed value of non-market output is estimated by means of the opportunity costs of time as well as market prices.

Kende (1978, p. 225) aims to produce a measure for the "total real consumption", which is the "total of socially valued resources available to a consumption unit during a certain period". With the aid of hourly wages of substitute household or market workers he calculates the value of the time spent on household production. He then examines the share of this unpaid housework in the total real consumption, which consists of the aggregation of value of: the dwelling, consumer durables, public services used, purchase of non-durables and services and imputed value of household production. In order to be used as an indicator for the level of living, the total real consumption must be divided by the total number of working hours (paid and unpaid). This new measure thus gives a higher degree of welfare to the household which acquires the same volume of goods and services with less working hours.

Finally, an estimation has been made of the consumption level in an average household including the volume of self-produced goods and services (Aldershoff et al. 1983, p. 16-17). In order to determine a minimal value of household production, the time input is multiplied by the net minimum wage (exclusive of children's allowance) for an adult married employee in industry. This value, aggregated with the average expenditure on food and drink, home and garden, clothing, health, insurance, and transport, gives a broad indication of the consumption level.

Final food consumption level

The ratio of money to labor input for food consumption is not necessarily a fixed one, on the one hand on account of the possibility of partial substitution of money and labor, and on the other hand because there are various interrelations (f.e. between income level, time available and acquisition of goods and services).

A low income can result both from a low hourly wage (in the case of full employment) or from unemployment or disability (no time spent on paid work). The volume of household production in its turn is also connected with the available time per household (i.e. time not tied to paid employment). The higher the income the more possibilities there are for households to acquire goods and services which can replace or back up their own labor input (for instance, convenience food, restaurant, household equipment). These goods and services can in themselves however call for or invite a certain labor input (for example, cleaning of appliances, more complicated and more extensive cooking activities with certain equipment).



Up to now no research has been carried out in The Netherlands, in which both the expenditure and the time spent by a household have been measured simultaneously. In 1980 there was, however, a national budget survey and also a national time budget survey (separate random samples).

The time budget survey data cover a survey period of one week in October 1980; circa 3000 respondents of 12 years and over recorded daily in a diary what they did with their time per quarter of an hour (Aldershoff et al. 1983, p. 11-14).

In the budget survey circa 3000 households kept account in a notebook of all their expenditure during one month in 1980, and of less frequent expenditure and consumer durables during the whole year (CBS 1983).


From relevant literature (see above), it was concluded that households should be differentiated according to composition (size, life cycle) and income level.

In the analysis the budget and time budget data for groups of households are made comparable on these characteristics (sort of group distribution matching, see Black & Champion 1976, p. 96-100).

On account of differences in the questions and coding it was unfortunately impossible to classify the households in a comparable manner by education, occupation, number of income earners, or urbanization.

Average time spent on meal preparation as well as paid work per type of household by all members aged 12 and over has been estimated by means of the time spending of individuals (Aldershoff et al. 1983. p. 13).

Economic evaluation method

As mentioned earlier, there is no one monetary evaluation method of household production which can irrefutably be called the best. It needs no explaining however, that different methods lead to different results. To illustrate this, the final food consumption level, i.e. the summarized money and labor inputs, of an average household has been calculated by means of four methods, using two for the earlier described "foregone wage" respectively "foregone expense" approach.

It seemed obvious in choosing a price or wage component to follow the calculations recently made by Bruyn-Hundt (1983) of the monetary value of household production in The Netherlands at the macro-level.

The following market wages were used (all for the year 1983):

- foregone wage approach:

I     net statutory minimum wage          for a married adult

II   net average earned hourly wage   employee in industry

                                                        (without children's allowance)

- foregone expense approach:

III   gross wage of a home help          inclusive employer's

IV   gross wage of a cook                 social security contributions

                                                        and holiday allowance

III concerns the wage of a professional home help employed in a home help service, who is expected to be able to stand in independently for the housekeeper. IV relates to market wages for a cook working independently, in the middle of regarding scale of wages.

The money value of the labor input naturally varies with the use of the different evaluation methods. The consequences for the final food consumption level and for the relative share of the money respectively labor input herein, are shown in table 1.

In the remainder of this article method III i.e. the gross hourly wage of home help, will be used. Where the unpaid meal preparation is concerned, the foregone expense approach is more realistic than the foregone wage approach, since the time spent on the required labor input is both fragmented and tied to certain periods of time. Paid work on the labor market cannot be considered as a very realistic alternative for this unpaid work input.


The data and analysis described here imply the following limitations:

- Quantification of the final food consumption level based on money and labor inputs gives no information on the quality of the output, nor on the evaluation by the household of its attained consumption level.

- The figures presented hereafter all refer to averages. Thus account must be taken of dispersion in the guilders and hours reflected.

- Budget data and time budget data are in itself of a descriptive nature. They reflect how much money or time is spent in a particular period, without clarifying the reasons or considerations.

- No attention is paid to the input of foodstuffs (i.e. ingredients, type of food and meals, use of semi-prepared foods etc.).




Table 2 shows the volume of the separate money and labor inputs in different types of households, which information has been used to determine the final food consumption level. Apart from that, the data in table 2 correspond with the earlier described patterns of expenditure on food and appliances resp. household production relating to meal preparation.

Table 3 shows to what extent the addition of the labor input for food consumption alters the differences between households based on the picture of their expenditure inequality.

It appears that the expenditure inequality between households with children ( varying by number of children and age of the eldest child) does not essentially alter, nor does the difference between one- and two-person households. Within the group of one- respectively two person households the addition of the labor input appears to increase the expenditure inequality.

Thus by neglecting the consumption of goods and services produced by the households themselves misleading or even wrong conclusions about differences in the (final) food consumption levels between households might be drawn. In short, the concept of final food consumption level appears to be a useful one. This is true throughout the life cycle of households. This corresponds to the results of Bivens & Volker (1983). Reasons behind different final food consumption levels are unknown from these data. Possible reasons might be due to differences between households in buying foodstuffs, that need home production added (Uusitalo 1979, p. 81-83).





In table 4 the final food consumption level is given per household as well as per person.

It appears that the final food consumption level per household is higher:

- for larger household size;

- (in households with children) the older the eldest child is;

- for people living alone aged 35 and over compared with younger singles;

- for couples without children where the wife is 45 or over compared with younger couples.

It is also shown from table 4 that the final food consumption level per person is higher:

- for smaller households;

- in older households with children (see above), with the exception of two-parent families with children under 5.

Table 4 presents in the third column the ratio of labor to money input in attaining the final food consumption level. As shown in table 1, the precise ratio greatly depends on the evaluation method used. Since a realistic foregone market wage cannot/may not be lower than the net minimum wage, this market wage has been used to impute a value of the labor input in order to calculate the ratio labor/money input. In short, this figure reflects a minimal ratio, and indicates the way in which households attain their consumption level.

In all types of households the labor input is higher than the money input, except the younger singles. The final food consumption level of older households without children (especially the 65 ) is also proportionally effected by a higher labor input than younger households (see Bivens & Volker 1983). This might be due to generation differences in attitudes towards household work, and/or differences in time spent on paid work on the labor market (see hereafter). In two-parent families, money and labor are spent in a rather constant ratio since separately these inputs show the same pattern (see table 2).



The fourth column in table 4 contains the total workload of the household consisting of the paid working hours, which are performed by all household members of 12 and over together per year, added up with the unpaid working hours needed for meal preparation.

The fifth column in table 4 shows the ratio of the labor input in meal preparation to the paid work input on the labor market. The 65+-ers perform, as is to be expected, no, or scarcely any paid work (retired people did this earlier in their life and are receiving payment afterwards).

What are the conclusions to be drawn from table 4?

1. A more or less the same final food consumption level can be reached in very different ways.

A final food consumption level of about f 23,000 - 24,000.C per year occurs, for example: with older couples without children, one-parent families, two-parent families with one child, and two-parent families where the eldest child is between 6 and 11 years old. However, the total workload in these households differs greatly, both as regards volume and share of labor input for meal preparation. There are also great differences in the labor/money input ratio between these households.

2. As has been mentioned, in two-parent families the final food consumption level increases both with the number of children and also the age of the eldest child; the total workload shows the same pattern. In the total workload the share of unpaid labor in meal preparation remains fairly constant; in other words, there is presumably little difference between different types of two-parent families as regards their final food consumption level.

3. Of the households without children the younger households have a relatively low final food consumption level, a relatively high total workload (consisting mainly (8096) of paid work on the labor market) and relatively more money is spent on food and appliances.

4. Taking into account the size of the household two-parent families have relatively a slightly lower final food consumption level than the one- and two-person households, except in case of older children.


The final food consumption level has been indicated by means of the summarized money and labor inputs, using data from two separate samples. A survey in which both expenditure as well as time use is measured in one and the same household will probably be impracticable because of very high financial costs. it would therefore be desirable to pay attention to household production in a budget survey and to the pattern of expenditure .. a time budget survey. it is also recommended that this type of data should be comparable with surveys, in which the inputs and outputs of food in the household are measured in the physical sense.

One should keep in mind that on the one hand the choice of a particular price or wage component for imputing a value for household production is determinant in the results, on the other there is no one best evaluation method available. However, in order to impute a value to the unpaid labor input for food consumption, the foregone expense approach seems more suitable than the foregone wage approach. In the case of labor input in meal preparation, which is so fragmented and at the same time bound to certain periods of time, there cannot actually be any weighing against the performance of paid work on the labor market.

The input based approach of the final food consumption level in this article needs further refinements. A combined measure must be developed in which both expenditure, household production as well as other inputs such as fuel, number and type of durable goods and foodstuffs are incorporated.


Aldershoff, D.E., A.C.L. Zuidberg, W. Baak (1983), Household production in The Netherlands, SWOKA Interim Report no. 14, The Hague.

As, D. (1982), Measuring the use of time, OECD, Special Studies, no. 7, Paris

Bivens, G.E. & C.B. Volker (1983), Value added through meal preparation in U.S. households: The young, newly-old, and older-old, Proceedings of Family Economics - Home Management Preconference Workshop, AHES, Madison, Wisconsin.

Black, J.A. & D.J. Champion (1976), Methods and issues in social research, John Wiley & Sons, New York.

Bruyn-Hundt, M. (1983), De waarde van onzichtbare arbeid gemeten, Economische Statistische Berichten, 27 juli, 666-669.

CBS (1983), Technical notes to budget surveys 1978 and later, Heerlen, (memorandum, not-published).

Chadeau, A. (1984), Presentation of an output related evaluation of unpaid household work in France, Pre-congress Workshop, Second International Interdisciplinary Congress on Women, Groningen.

Chadeau, A. and A. Fouquet (1981), Peut-on mesurer le travail domestique? Futuribles, Decembre 1981, 33-35.

Gershuny, J.J. (1982), Social innovation, Change in the mode of provision of services, Futures, December, 496-516.

Goldschmidt-Clermont, L. (1982), Unpaid word in the household. A review of economic evaluation methods, International Labor Office, Geneva.

Goldschmidt-Clermont, L. (1983), Out-put related evaluations of unpaid household work: a challenge for time use studies, Home Economics Research Journal, December, 127-132.

Kende, P. (1978), Les biens et les services autoproduits dans la consommation des menages francais, in: Michel, A. (ed.). Les femmes dans la societe marchande, Presses Universitaires de France, 225-244.

Kilpio, E. (1981), The concept of unpaid housework and the determination of its value, and Description of the study and sample, Official Statistics of Finland, XXXII:17, Helsinki (Housework Study part I and II).

Santti, R. & H. Vesikkala (1983), Cooking work, Official Statistics of Finland, SVT XXXII: 94, Helsinki (Housework Study Part XII).

Sirageldin, I.A.H. (1969), Non-market components of national income, Institute for Social Research, University of Michigan, Ann Arbor.

Smits, W.F. (1982), De uitgaven aan voeding, Voeding, Vol. 44, no. 4, 108-113.

Suviranta, A. & A. Mynttinen, The value of unpaid cooking work, Official Statistics of Finland, SVT XXXII; 76, Helsinki (Housework Study Part VI).

SWOKA Research Report no. 21 (1984), Consumer and household production, The Hague, (forthcoming).

U.S. Department of Labor (1983), Consumer expenditure survey: diary survey, 1980-1981, Bureau of Labor Statistics, Bulletin 2173, Washington.

Uusitalo, L. (1979), Consumption style and way of life, The Helsinki School of Economics, Helsinki.

Volker, C.B. & G.E. Bivens (1983), Value added: Accounting for more than time use in preparation of food in households, in: K. Goebel (ed.), Proceedings of 29th ACCI Conference, 35-38.

Walker, K.E. & M.E. Woods (1976), Time use: a measure of household production of family goods and services. The Center for the Family, AHES, Washington.

Walker, K.E. (1980), Time measurement and the value of nonmarket household production, in: C. Hefferan (ed.). The household as producer, A look beyond the market, AHES, Washington, 119-138.