This article was downloaded by: [University of Chicago Library]On: 04 October 2014, At: 15:54Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Applied EconomicsPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/raec20

Valuing family household production: a contingentevaluation approachEuston Quah aa Department of Economics and Statistics , National University of Singapore , Kent Ridge,0511, SingaporePublished online: 28 Jul 2006.

To cite this article: Euston Quah (1987) Valuing family household production: a contingent evaluation approach, AppliedEconomics, 19:7, 875-889, DOI: 10.1080/00036848700000035

To link to this article: http://dx.doi.org/10.1080/00036848700000035

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in thepublications on our platform. However, Taylor & Francis, our agents, and our licensors make no representationsor warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Anyopinions and views expressed in this publication are the opinions and views of the authors, and are not theviews of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should beindependently verified with primary sources of information. Taylor and Francis shall not be liable for any losses,actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoevercaused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Applied Economics, 1987, 19, 875-889

Valuing family household production: a contingent evaluation approach EUSTON QUAH

Department of Economics and Statistics, National University of Singapore, Kent Ridge, 051 1 Singapore

The time of family members is often regarded as an important input used in family household production. Unlike existing methods used in the valuation of the time spent in home production, namely the opportunity cost of time and the replacement cost method, a contingent evaluation approach is proposed here. Consistent with the way non-market goods and services are valued, the method essentially elicits people's willingness to pay for household production whilst taking into account opportunity costs. Particular attention is also paid to the objective of valuation - whether it is for GNP accounting, matrimonial property settlements or for welfare ahd compensation issues-and the appropriate measures to be used -marginal, total and net valuation. Using a model of home production, these different measures of economic value are distinguished both conceptually and in their uses. The results from a modest study of household production using the methods described here are presented. These results are then compared to those derived from other studies.

1. INTRODUCTION

The 'new home economics' literature is now an established part of economic theory. Market purchased goods aside, the time and efforts of household members are now recognized as important inputs used in this production process (Becker, 1976; Gronau, 1977, 1980). The amount and economic value of time used to provide the daily services of cooking, cleaning and the multitude of other regular home maintenance chores that are demanded by households is clearly not insignificant. They represent a very substantial portion of the total productive time available to members of a household and to a society, and an accounting of this time and its worth that is commensurate with wages earned from other employments is an issue of increasing practical importance.

By convention, economists and home economists tend to value the time spent in home production on the basis of either the opportunity cost or replacement wage method. General criticisms of both methods are well known in the literature and, suffice to say, they will not be repeated here1. An alternative method, called 'contingent evaluation' is instead suggested. The method essentially involves posing alternative contingencies to households in terms of whether

'See for example, Peterson (1978), Gronau (1980) and Quah, (1986a).

0003-6846187 S03.00+ . l2 0 1987 Chapman and Hall Ltd. 875

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

876 E. Quah

they would rather pay specific amounts to reduce the time given over to household tasks or to keep the money and do the work themselves. Specifically, the method uses an auction bidding procedure to elicit households' maximum willingness to pay for the different contingencies posed for which their responses to varying price offerings then trace their marginal benefit and opportunity cost for doing different quantities of housework. This would then indicate the economic value of the time and effort devoted to household production. A small exploratory survey is used to demonstrate the approach.

The paper is organized as follows:

(1) a model of home production is presented showing a way of conceptualizing the household's behaviour. Using the model, the different measures of economic value -marginal, total and net -are distinguished both theoretically and in their uses. This aspect of valuation is important since, depending on the objective of valuation- whether it is for GNP accounting, matrimonial property settlements or for welfare and compensation issues-the appropriate measures would involve marginal, total and net valuation. Researchers to date have either ignored or failed to emphasize this; (2) the household survey and the contingent evaluation method is described; (3) empirical results which focused on the values at issue, their magnitudes are presented together with hypothesis testing of the determinants of household production values and time; and finally (4) the results are compared to those obtained from other studies which use conventional methods.

11. A M O D E L O F H O M E P R O D U C T I O N : D I S T I N G U I S H I N G E C O N O M I C VALUES

Household production is normally provided by members of the same household consuming the services with the time allocated to this purpose determined largely by other demands on their time and energy and the utilities (disutilities) of providing and consuming these services relative to those associated with alternative production and consumption activities. These peculiarities of the provision and consumption of household services in no way make the time put to such production any less economically valuable, as people clearly demonstrate such value by willingly giving up other goods and services in order to enjoy the benefits of their provision. The non- pecuniary nature of the values only means that such values are not conveniently registered in market prices.

The economic value of the time spent in household production therefore provides a good indication of the economic value of the benefits produced - household commodities- using that time. An indication of the expected gains and losses of varying levels of production effort to a household is illustrated in Fig. 1. This figure shows the marginal benefit (MB)- the incremental benefit from an additional hour devoted to home production-and marginal cost (MC)-the additional cost incurred from devoting an additional hour to home production-curves (assumed here for convenience, linear and continuous) of a given household. The principle of declining MB indicates that for most households, at least some level of household services are

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production

S I

Fig. 1.

important, but successive quantities are likely to be less important. At point N, nothing could be gained by devoting an additional unit of time to home production. The MC curve is shown to have a positive slope indicating higher additional costs as more time is employed in household production. For most households, the MC are likely to be relatively small for some limited allocation of time to home production, and may even be zero or negative for some very small amount of time (OJ) but would probably increase for larger commitments. This rise in MC corresponding to increased household time spent in home production is the result of putting the more market productive household members with respect to the use of their time to household production. A household facing MC and MB as indicated in Fig. 1, would devote OR hours to household production. For by devoting OR hours to home production, the welfare of the household is maximized.

It is possible to identify and measure the different economic values of time devoted to home production depending on the purpose of measurement. First, there is the notion of marginal value, which is the value households place on the last unit of time (say hour) to home production. In the case of Fig. 1, this marginal value is represented by OW for the ORth hour of home production. Second, there is the notion of total value, which is the sum of marginal values of all the hours performed in home production. In terms of Fig. 1, this is shown as the area OVPR for OR hours of home production. Finally, there is a residual value, more commonly called the net value from home production. By net value is meant the residual benefit from the total value over total costs of home production. Since the total cost of home production is given by the area under the MC curve i.e. JPR for OR hours and the total value is OVPR, the net value measurement is the 0 VPJ (0 VPR - JPR).

These different concepts of economic value are important for measuring the value of home production for different purposes. Thus, if the objective of measurement is for national income accounting, then to be consistent with marketed output measurements, market prices in terms of marginal values are appropriate. On the other hand, if the purpose of measurement is for compensation and welfare issues, such as in tort litigation, then it is the net value measurement that is appropriate; for the objective of compensation is to return the injured party to its original

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

878 E. Quah

welfare position prior to the accident. The calculation of net values thus requires a knowledge of total values and total costs.

111. C O N T I N G E N T EVALUATION METHODS

The contingent evaluation method has been used, with growing success, to value other non- market goods and seems an appropriate approach to value household production as well. The technique essentially involves the construction of a hypothetical market for the good in question and information is given as to its quantity, quality, location, time dimensions and other attributes. The rules of operation of this contingent market are then established and the respondent is asked to indicate a reaction to some contingency that is posed. In its more common application, in which values have been placed on various non-market environmental amenities, individuals have been asked how much money they would be willing to pay to attain some improvement in natural surroundings (Bohm, 1979; Brookshire and Crocker, 1976,1981).

Contingent evaluation methods invariably use survey questionnaires to acquire the relevant data for analysis. This is seen as a means to allow individuals to participate in the choice of the provision of non-marketed good or service, thereby indirectly expressing his level of enjoyment - welfare - from current provision. The individual's responses will reflect his prospective sacrifices and gains in utility as a result of the proposed contingencies.

One form of the contingent evaluation methods is used here, namely the auction bidding approach. This method follows an iterative questioning procedure to elicit responses which enable the researcher to construct a demand curve for the good in question. Individuals are asked to indicate a 'Yes' or 'No' answer to a given bid of a proposed contingency. If the respondent indicates a 'yes' answer, the bid is raised to $X + X where x is some dollar increment. The process is then repeated with equal X increments until the respondent switches to a 'no' answer, with that point being the highest sum he is willing to pay for continued use of resource Y. Per contra, if the respondent initially indicates a 'no' answer, the bid is then lowered to $X - X

and the iterative process is repeated until the respondent switches to a 'Yes' answer, with again that point being the maximum sum he is willing to pay for the use of resource Y. By developing a series of contingent markets, the individual is clearly allowed to react more positively (or negatively) to a proposed contingency with the data then revealing the individuals' perception of value.

IV. CONTINGENT EVALUATION I N H O U S E H O L D P R O D U C T I O N

An explanatory study on the value of household production using the contingent evaluation method, was conducted over a Cmonth period involving sixty households in the city of Victoria and the Greater Victoria region in British Columbia, Canada. The participants were randomly selected from the telephone directory using a computer generated table of random numbers. In the process, all single person households were excluded from the sample so as to be consistent with the objectives of the survey.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production 879

Each of the participant households received a small package containing a sslrvey question- naire from which responses regarding personal data and information relating to household production were elicited. The participants were also contacted at varying times after they were in receipt of the mailed questionnaire to inquire about problems, to urge close attention and to answer any questions. Following the return of all completed questionnaires, each of the respondents was again contacted and a contingent bidding game2 was played.

The respondents were fist informed of the number of hours that their household was currently spending in household production per week (as reported in their completed questionnaire). The respondents were then told of the possibility of reducing this total number of hours with the magnitude of reduction in hours to be performed by a hired substitute worker. The respondents were told that the hours of replacement services would be available on a continuing basis for an indefinite time which would allow them to undertake more permanent reallocations of their time. They were also to assume that whatever housework done would be of equal quality and would involve no particular loss or inconvenience to them. The respondents were then asked the following questions.

(i) Whether they would be willing to pay $2 per hour to reduce their current hours of household production by 2 hours a week. If their answer was 'Yes', the bid was then raised to $4 per hour and the question repeated. The amount was increased in $2 increments until a negative response was obtained. On the other hand, if their initial answer was 'no', the bid was lowered to $1 per hour and the question was asked again. A refusal to pay $1 per hour was then recorded as a $0 bid. (ii) Whether they would be willing to pay $2 per hour to reduce their current hours of household production by another 2 hours a week, bringing a total of reduction by 4 hours a week. The iterative bidding process of (i) was then repeated. Subsequently, the entire question was repeated using decrements of 2 hours until a marginal $0 bid was recorded. Throughout this part of the interview, the respondents were reminded that under no circumstances were they compelled to purchase these substitute services.

The theoretical presumption behind the above questions was such that one would not expect households to be willing to pay more than their own marginal opportunity cost of household production. At most, it might be expected that households to be willing to pay an amount equal to their own cost of production. Their responses of bids to each of the 2-hour reductions then trace out the entire individual household's marginal cost function in home production.

After all the responses were recorded in the first part of the interview, the respondents were then told to assume that for some reason or other, their household could only perform 2 hours less than their current hours in household production. The respondents were than asked the same question as in (i) and the iterative bidding process repeated. Restricting the hours that can be performed by household members and replacing those hours by the hired substitute worker, questions were again asked on the household's maximum willingness to pay for household services. The difference between the questions in this part of the interview and the first was that

'Pioneered by Davis (1963) and applied to the valuation of outdoor recreation.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

E. Quah

the respondents were now restricted to perform a given number of hours and the remaining hours (up to current levels) could only be performed by the substitute worker. Thus, households could either leave the remaining hours undone, use market services outside the household (e.g. eating out, live in a room and board type of accommodation) or hire the substitute worker to perform services for them.

While the contingent questions in the first part of the interview attempted to generate the respective household's marginal cost function, the questions contained in the second part were aimed at tracing out the household's marginal benefit function. Given the latter situation, households would not be expected to be willing to pay more than they expect to gain from havingadditional hours to home production. At most, households would be willing to pay up to their marginal benefit from having the additional hours. Their responses or bids to the incremental changes constitute in effect their marginal willingness to pay for the additional hours. While the small size of the sample-necessitated by resource constraints and the intentionally exploratory nature of the study -limits the possible generalization of the specific findings, it appeared large enough to explore and illustrate the techniques and to obtain some rough indications of magnitudes.

V. EMPIRICAL RESULTS

In Table 1, the mean willingness to pay for household production, households' opportunity costs and the average number of hours households currently spend on household production per week, are presented. Households' mean total willingness to pay for household production was computed by taking the average of the areas under each household's marginal benefit-marginal willingness to pay function. The plot of the marginal benefit function was obtained from the household's responses or final bids to the contingent bidding process described earlier. In order to simplify otherwise tedious manual calculations and because log linear functions yielded insignificant result differences, a straight linear regression was used and the integral of the function from zero to current hours in household production taken. The area under this marginal benefit or demand function indicates the total value which a given household attributes to household production. The average of all these areas representing household's total value is $117.31 per week. Given the weekly average number of hours of household work that these households reported, this is equivalent to an average value of about $4.91 per hour.

Table 1. Mean willingness to pay for household production, opportunity costs and average number of hours in household production

Per week (S) Per hour (S)

Average number of hours 23.88 Mean total willingness to pay (total value) 117.31 4.9 1 Mean net willingness to pay (net value) 101.10 4.23 Mean total opportunity cost 16.21 0.68 Mean opportunity cost at the margin (marginal value) 2.22

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production

A similar computation is made to obtain the households' mean total opportunity costs in household production. By taking the average of the areas under each household's marginal cost function, the mean total opportunity cost is $16.21 per week or $0.68 per hour. However, the mean opportunity cost of the last hour in household production is about $2.22.

Finally, subtracting households' mean total opportunity cost from their mean total willingness to pay for household production yields a mean net willingness to pay of $101.10 per week or $4.23 per hour. In other words, households on average place a net value equal to $101.10 per week on household production. This gives a measure of the amount of welfare enjoyed by households from having household production.

Hypothesis testing

Some further insights into household production evidenced by the respondents might be gained with the use of regression analysis that statistically tests possible links between differences in the reported number of hours devoted to household tasks, their willingness to pay and the various factors that may be associated with these variations. The decision about whether or not variables should be included in any regression model should still depend largely on a priori theoretical considerations. Since, in any given demand relationship, the quantity demanded of, say, good X is a function of its own price, the prices of other goods, consumers' income and tastes, so is the quantity demanded for hours of household production a function of its own price as measured by the opportunity cost incurred by the household at the margin of provision, the prices of substitutes in the form of market replacements, households' income and households' tastes in performing housework, plus all other demand shift variables which one might think would have an effect on hours worked (e.g. size of household, types of dwelling, number of pre-school children and number of labour saving appliances).

A more difficult hypothesis to be tested using the same data is one involving households' willingness to pay for household production as the dependent variable. The difficulty arises mainly because of a lack of specific a priori theoretical determination of independent variables. Apart from the fact that willingness to pay for a good or service depends on the potential utility to be gained from consuming the good or service which is said to be reflected in consumers' demand function, a less satisfying ad hoc approach may be required in specifying the variables to be included in the willingness to pay regression model. Nevertheless, in specifying the willingness to pay regression model, the hypothesis to be tested is that household's willingness to pay is a function of demand variables and hours of household production.

Generally, one would expect the number of hours worked at home to be negatively related to the household's opportunity cost at the margin so that a rise in marginal opportunity cost will have depressing effects on the number of hours worked. The effect of household income on household production is indeterminate since household produced goods and services can be normal or inferior. Thus, if household production is normal, then it might be expected that as income rises, the quantity demanded of household production would also rise. But, if household production is inferior, then the reverse occurs so that, as income rises, the quantity demanded of household production falls. Prices of other goods and services are expected to influence household production; for example, prices of substitute goods and services are expected to react directly to household production. But because this is a cross-sectional study such that there is no

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

E. Quah

variation across the sample-every household faces the same prices of these other goods and services-this variable was omitted from the study. With regard to size of household and the number of pre-school children present, household production would be expected to vary directly with these variables: 'more people more work', together with the fact that small children require more a t t en t i~n .~ Similarly, the type of dwelling (apartment or house) is expected to have a positive influence on household production in that a physically larger space generally requires more household production. Finally, the effect of the number of household labour saving appliances on household production is indeterminate. It may be that most people would presume that the introduction of labour saving appliances to a household would reduce the total time spent in household production, but this is only true if one assumes a fixed amount of housework to be done. The possibility of observing an increase in hours performed is the case if the introduction of labour saving appliances increases the quality of housework as well as increasing the amount that can be accomplished in a given time.

Since households' willingness to pay for household production is derived from their demand functions, it can be expected that similar relationships between willingness to pay and these variables would also hold.

The appropriate functional form for estimating a willingness to pay equation is still an unsettled issue. Previous empirical efforts have, for the most part, used a semi-log model specification, since the model is consistent with the assumption that the effect of a change in the explanatory variables depends on the level of these variables. Thus, the semi-log form permits the marginal effects of the explanatory variables to change as their level changes, a result that appeals to common intuition. For example, if the explanatory variable considered is one of size of household, then the larger is the size of the household, the smaller is the marginal effect of changes in the size of the household. Moreover, because some of the explanatory variables considered here have zero values, the log linear form cannot be used. Although a linear function was also estimated, the semi-log form performed consistently better and was therefore chosen as the basis for this study's findings. The general specification of the semi-log model is:

where E is the error term that is assumed to be lognormally distributed. Three semi-log equations were estimated using the OLS method. The first two equations take

the form:

and the third equation estimated,

where In H P = log of household production, defined as the number of hours households spent on performing housework per week;

In TV = log of total willingness to pay for household production per week;

30nly 11 households in the sample reported having pre-school children.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production 883

PR = opportunity cost incurred in performing household production at the margin; ZNCl = 1, if household income is between $20000 to $40000, inclusive; or = 0, if other income; INC2 = 1, if household income is greater than $40000 or = 0, if other income;

SZ = size of household as measured by the number of household members; PC = number of pre-school children; SA = number of household labour saving appliances;

D W = 1, if type of dwelling is a house or townhouse; or = 0, if type of dwelling is an apartment;

El , E2, E3 = random error component assumed lognormally distributed.

The OLS estimates of the three equations were (with t-statistics in parentheses):

These regression results largely confirm a priori expectations, despite the insignificance of four of the coefficients. Consider Equation 1; the income variables are largely insignificant in explaining variations in hours of household production. The insignificance of the income variables were further tested as a group using an F-test, the result of which showed that there is no difference in different levels of income on quantity demanded of household production. This was confirmed at either the 1 or the 5 % levels. Since the null hypothesis that the income coefficients are not significant from zero cannot be rejected it can be concluded that the income elasticities must be close or if not, equal to zero.4 What this means is that percentage changes in household income do not affect the quantity demanded of household production. Although the income coefficients display negative signs, suggesting that household production varies indirectly with income, care must be taken in their interpretation since the income coefficients are not significant from zero. If it were the case that the income variables were significant, then the negative signs of the income coefficients would imply that household production is inferior. This is, however, not to say that household production does not contribute to the welfare of the household. The inferiority only arises because of a particular sort of substitution relationship between goods. It might be that households, on receiving higher incomes, are now able to afford

'Income elasticity is defined as: In HPIINC. 1%/1n G, but since In HP/INC is not significant from 0, then 0 m/ln HP = 0.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

884 E. Quah

some other goods previously unattainable. An example would be one in which a household, on receiving a higher income, is now able to afford dining in expensive restaurants, or in general, eating out more often.

Two variables were significant in explaining variations in quantity demanded of household production: the size of the household (SZ) and the number of labour saving appliances (SA). An additional member to a household on the average adds 0.11 % to the total hours of housework performed while an additional household appliance on the average increases the hours demanded of household production by about 0.15

The remainder of the variables: PR, D Wand PC display statistically insignificant coefficient values but have the anticipated signs. In other words, households' opportunity cost at the margin, the type of dwelling (whether it be a house, townhouse or apartment) and the number of pre-school children have no significant impacts on hours of household production. That the opportunity cost at the margin is insignificant in explaining variations in hours of household production can be explained if the household's marginal benefit function is steeply sloped over the last few hours in household production, so that changes in marginal opportunity cost over that range do not produce significant changes in quantity demanded of household prod~ction.~ Finally, the insignificance of the pre-school children variable may be a function of the data used.

The same variables which are significant in explaining variations in hours of household production also appeared significant in the total willingness to pay equation (Equation 2). All the variables in the In T V equation also have the same signs as those in the In H P equation. The results indicate that household income is not significant in explaining households' willingness to pay for household production. Even when both income dummy coefficients were tested as a group, using an F-test at either the 1 or the 5 % level failed to make it significant in explaining willingness to pay. The size of households continued to be significant withan additional member to present levels on the average raising households' total willingness to pay by about 0.16 %. To avoid repetitions, the coefficients of the remaining variables are similarly interpreted.

Finally, because of the high multicollinearity between hours of household production (HP) and the other independent variables in Equation 2, a separate regression (Equation 3) was run between households' total willingness to pay and hours of household production. The results indicate that hours of household production are statistically significant in explaining variations in willingness to pay. Thus, an additional hour of household production on the average, adds 0.03 % to households' total willingness to pay.

The adjusted multiple correlation coefficient of determination or R2, which indicates the percentage of variation explained by the regression equation, is quite high for all three

'The model specified can be written as:

Differentiating totally,

since dHP/HP = a,. In percentage terms, this is a, %. 60ver two-thirds of the household surveyed have steep marginal benefit functions over the last few hours worked.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production 885

equations, considering that this is a cross-sectional study. By means of an overall F-test, the null hypothesis that the multiple correlation coefficient of determination is not significantly different from zero, can be rejected,at the 5 % level of significance for all three equations.

V. SOME IMPLICATIONS AND LIMITATIONS

In the present study, the contingent evaluation approach was used to estimate the value of household production as recorded by the participant households themselves. The results indicate that on average, households' total willingness to pay-total value-for household production is $117.31 per week. The value measures how much the household would be prepared to give up (to pay) if necessary to obtain a weekly supply of household production. However, since households only pay an amount equal to their total opportunity cost in current levels of production, they are said to gain from having household production. This gain - total willingness to pay less total opportunity cost -measures the amount of welfare the household receives from household production. On the basis of the present survey, this measure of economic welfare amounts to $101.10 per week or an average net value of $4.23 per hour.

The average values placed on household production may or may not be indicative of the values that would result from a more extensive and more representative sample, but aside from the small size of the sample, there seems little to indicate that a larger sample would result in much different values. Although no specific tests were made for biases-efforts that would require very sophisticated statistical techniques and not the primary focus of this paper - arising from the use of contingent bidding methods, considerable care was taken in the design and conduct of the survey so as to ensure that the results obtained were as reliable as possible. The efforts included a number of trial tests in which problems relating to ambiguities in some questions and misinterpretations were detected and corrected before the final survey was launched. Interviewer bias was kept to a minimum with interviewers following a strict common questioning format. Regular cross-checks between interviewers were made to determine problems and to allow for common solutions. Furthermore, great efforts were taken to make sure that the hypothetical situations posed were as realistic as possible to the respondents; these efforts include careful explanations of the contingencies posed to respondents and allowing the interviewers to take their time in answering the questions asked. To the extent that these and other biases were minimized, the contingent evaluation technique applied to value household production would appear to have yielded the relevant estimates.

However, the rather small number of participants in the survey does not allow wide generalizations of the specific findings, but it would appear large enough for exploratory purposes in illustrating the technique involved.

VI. VALUE AND T I M E COMPARISON WITH OTHER STUDIES

Attempts to measure the amount and economic value of time devoted to household production are not new and many have originated from the USA. Table 2 shows some of both these and Canadian studies.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

886 E. Quah

Table 2. Summary table of past and recent studies on horcsehold productionl

Average Average hourly value annual value

Author@) Method used Hourslweek (S) (S)

Hawrylyshyn (1971) Replacement wst 47-63 5.06-6.78 16602 Walker and Gauger Replacement cost 69.3= 6.64 23940.31 (1971)b 24Sd Chase-Manhattan (1972P Replacement wst 99.6 7.57 39200.80 Kaye (1982)b Replacement wst 206' 4.19 448 12.41 Minton (1978lb Replacement cost 257.4' 7.17 96996.92 Ontario Status of Replacement wst 100 3.53 18356 Women Council (1977) Proulx (1978) Replacement wst 49 6.04 15394.70 Quah and Knetsch Indirect method of (1982) willingness to pay 27.25 1.51 2137.71 Present study Contingent bidding (1983) method of willing-

ness to pay 23.88 4.23 5252.64

aIn 1983 Canadian dollars. studies.

CNon-employed wife households. Employed wife households.

'Studies cited in Penny Kome's article, 'What Price Dusting?', The Financial Post, April, 1982.

In terms of value comparisons, the average of about $4.23 per hour of this study is far below those estimates that have been generated by methods of replacement cost. Indeed, in the Chase Manhattan Bank studies, the average per hour value of household production was reported to be $7.57, a difference as large as $3.34. The disparity in value measurements is even more marked when one considers the annual value figures. The Chase Manhattan study reported an average annual value of about $39 200.80, while in the study by Minton (1978), the average was asserted to be $96 996.92 per year, the difference with current results being nearly $34 000 and $91 000 respectively!

A major part of the reason for such high numbers in the table, particularly for those estimates just mentioned in the text, can be found by examining the range of wages used as replacement costs. Instead of using the average wage rates as values for the various estimation of each household function, particularly high wages and those usually earned by specialists were used. For example, in activities relating to cooking and dietary functions performed by the housewife, a wage rate of nearly $17.00 per hour was used as the equivalent market replacement costs!

The 1982 study by Quah and Knetsch was the first attempt to obtain direct estimates on the household's total willingness to pay for household production. For the first time in household production research on valuation, the traditional methods of replacement cost and opportunity

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing family household production 887

cost were discarded and a contingent evaluation of the indirect type was introduced. In that study, the respondents were presented with various prices charged by a substitute worker and they were asked the number of hours their household would contract for at each of those prices. Unfortunately, there was a methodological error in using such a contingent question-the question in effect only allowed households to compare the hiring cost to their own opportunity costs in household production with no measurement on benefits at all. Moreover, the contingent question postulated also measured the value of some additional hours that were currently not performed by households-the marginal benefit of those hours is less than the marginal cost- but would have been performed if the price of the substitute worker fell to such comparable levels as the households' marginal benefit of those hours. The low estimate of $1.51 per hour is thus a measurement of the value of market domestic servants and not the households' value of their own household production.

Attempts to assess the amount of household services performed by households have largely relied on respondents' recall of time devoted to such purposes. A typical survey poses questions to a member of the household asking for an estimate of the average number of hours that the person thinks are spent either on individual tasks such as meal preparation, cleaning, laundry and other household chores, or on the total time devoted to all such chores on a daily or weekly basis. The average of about 23.88 hours spent on household production by households per week reported in the present study followed such a method. Admittedly, a major weakness of the method is that people may not have any accurate awareness of the amount of time allocated to household production activities. The estimation is made even more difficult by the fact that while a disproportionate amount of the household chores may be performed by one person, some such work is usually done by nearly all household members. There are also other problems such as the definition given to housework and the joint production problem.

The present study has given a precise definition of household production in that household production is defined as those household activities that could be done to prescribed specifications and to the benefit of the household by someone outside of the household. Thus, making a cup of coffee and cleaning the floor are clearly included since these activities can be performed by hired help and to the benefit of the household, while watching television and sleeping are not as the latter activities will not benefit the household but rather only the hired help. The value of household production is seen as having it done and not by having it done by any particular individual or member of the family. Except for possible quality differences, and, of course, abstracting from any benefits of consortium - the companionship and conjugal relation aspects of non-market time - the value of the output and, therefore, of the benefits is not attributable to the input of any particular individual. Further, the joint production problem is resolved by adopting time estimates only for the major task performed; the other simul- taneously performed activities which were given only intermittent attention were thus ignored. As such, much of the methodological ambiguities usually associated with household production measurement were hopefully reduced.'

The estimate of about 23.88 hours is, however, supported by the time budget study -estimates based on an actual allocation of time-of Quah and Knetsch (1982) where the authors reported

'For a detailed treatment of methodological problems in household production measurement and valuation, see Quah (1986a).

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

888 E. Quah

an average of about 27.25 hours of household production a week. The rather large disparity in time estimates between the present study and those of the other studies shown in Table 2, is largely the result of the joint production problem in that the other studies made no distinction between major and minor tasks for simultaneous activities and the accounting of time was simply doubled for two activities performed at the same time. Thus, it is quite possible for households to report more than 168 hours spent on household production per week as is the case in the studies of Kaye (1982) and Minton (1978).

Because of the rather involved nature and difficulty in obtaining respondents' cooperation, only very occasionally will time budget studies be done. Only two studies in Table 2 were reports on time use based on time budget studies- Walker and Gauger (1973) and Quah and Knetsch (1982). The rather large disparity in time estimates of the two time budget studies can be explained by the fact that the estimates of the former came from non-employed wife households, while the latter's sample consisted mainly of employed wife households. If employed wife households were taken in comparison, the disparity almost disappears since Walker and Gauger reported that only 24.50 hours were performed in an average week.

VII . C O N C L U S I O N S

Given the increasing importance being attached to estimates of the quantity and economic value of family household production in policy formation and in litigation, there would appear to be considerable point to appraisals that would yield more accurate and appropriate valuations. The present modest study has attempted to demonstrate methods that might lead to some improvement in these measures. If the values obtained by this application of the contingent evaluation method are even roughly correct, to an order of magnitudes, they are suggestive of the extensive bias inherent in other procedures.

A C K N O W L E D G E M E N T S

I am grateful for the financial support given by the Department of Economics at the University of Victoria, BC, Canada and from the Knetsch-Ferguson Research Funds where this paper was written up. I would like to thank Malcolm Rutherford, Lorne Rosenblood and an anonymous referee of the journal for very helpful comments. The usual caveat applies.

REFERENCES

Becker, G. S. (1976) The Economic Approach to Human Behaviour, University of Chicago Press, Chicago. Bohm, P. (1979) Estimating willingness to pay: why and how?, Scandinavian Journal of Economics, 81,

142-53. Brookshire, D. S. and Crocker, T. D. (1981) The advantages of contingent valuation methods for

benefit-cost analysis, Public Choice, 36, 235-346. Brookshire, D. S. and Crocker, T. D. (1976) The valuation of aesthetic preferences, Journal of

Environmental Economics and Management, 3, 325-46.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4

Valuing fmily household production

Chase-Manhattan Bank (1972) What is a wife worth? Chase Manhattan, New York. Davis (1963) Value of Outdoor Recreation, unpublished PhD Thesis, Haward University. Gronau, R. (1977) Leisure, home production, and work- the theory of the allocation of time revisited,

Journal of Political Economy, 85, 1099-123. Gronau, R. (1980) Home production -a forgotten industry, Review of Economics and Statistics, 62,408-16. Hawrylyshyn, 0. (1978) Estimating the Value of Household Work in Canada, 1971, Ottawa, Ministry of

Industry, Trade and Commerce. Mena, F. C. and Mullen, J. K. (1981) Expected encounters and willingness to pay for outdoor recreation,

Land Economics, 57,33-40. Ontario Status of Women Council (1977) Towards a positive image of housewives. Peterson, R. D. (1978) Problems in estimating the value of household services, American Journal of

Economics and Sociology, 37, 145-8. Pindyck, R. S. and Rubinfield, D. L. (1981) Econometric Models and Economic Forecasts, Second Edition,

McGraw Hill, New York. Proulx, M. (1978) A study of the Canadian housewife. Advisory Council on the Status of Women, Ottowa,

Canada. Quah, E. (1986a) Persistent problems in measuring household production, American Journal of Economics

and Sociology, 45, 235-45. Quah, E. (1986b) A diagimmatic exposition of household production, market substitutes and the terms of

trade, Singapore Economic Review, 31, 33-39. Quah, E. and Knetsch, J. L. (1982) Measuring the quantity and economic value of household production:

an exploratory study, mimeo, Simon Fraser University. Walker, K. and Gauger, W. (1973) The dollar value of household work, Information Bulletin No. 60,

College of Human Ecology, Cornell University.

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 1

5:54

04

Oct

ober

201

4