Determinants of the number of hours doing housework of the wives in tow-earner households in urban HCM City

The study based on the theories of household behaviour and time allocation to test the effects of women and households' characteristics on the number of housework hours of the wives.. It

THE UNIVERSITY OF ECONOMICS

HO CHI MINH CITY

VIETNAM

INSTITUTE OF SOCIAL STUDIES

THE HAGUE THE NETHERLANDS

THE VIETNAM-NETHERLANDS PROJECT FORMA PROGRAM

IN DEVELOPMENT ECONOMICS

DETERMINANTS OF THE NUMBER OF HOURS

DOING HOUSEWORK OF THE WIVES IN TWO-EARNER

HOUSEHOLDS IN URBAN HO CHI MINH CITY

BY LUONG THI CHUNG THUY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE

OF MASTER OF ARTS IN ECONOMICS OF DEVELOPMENT

HO CHI MINH CITY, 2003

ACKNOWLEDGEMENT

The thesis in done as the fulfillment of the MA course of Vietnamese - Dutch Project for Development Economics Thanks to the Netherlands for her Aid and Scholarship during the course

I would like to thank Dr Gabrielle Berman, Dr Youdi Schipper and Dr Karel Jansen for their worthy teaching, suggestion and comments on my very first step of the thesis

I would like to express my special thanks to my supervisor, Dr Ho Ngoc Phuong, for his supervising and guidance during my writing I am deeply indebted to MA Ta Thi Bich Thuy for her special support and all materials she gave Many thanks are also given to Dr Nguyen Van Ngai for his suggestion on my thesis econometric model, Dr Nguyen Thi Hoa, Dr Tran Kim Dung for their helps The thesis would not have been completed without their support and encouragement

Many thanks are also given to all project teachers and staffs, especially Mr Tran Vo Hong Son,

Ms Tran Thi Ben, Ms Do Thi Anh Nguyet and Ms Dang Kim Chi for their helping during the course

I would like to thank Ms Dang Le Hoa for her checking of my writing in grammar, vocabulary

as well as her comments I also thank Ms Nguyen Phuong Chi, Ms Ha Dung, Ms Pham Le Hoa,

Ms Doan Thi Thu Huong, Mr Luu Tien Thuan, Mr Truong Khoa (RAND), and Ms Nguyen Thanh Binh (DOLISA), who gave me many supports in doing the thesis My sincere gratitude goes to all of my friends for their helping and discussions

I am also grateful to all people who helped me conducting the survey, especially Ms Phan Hong

Ha, Mr Nguyen Thanh Tung, other staffs and interviewers at ACORN Company I also thank all of respondents who gave me valuable information

Finally, I am thankful to my parents, my aunt·and my brother for their special supports I cannot do anything without their supports and encouragement

11

ABSTRACT

The study tries to examine the determinants of the number of hours doing housework of the wives in two-earner households in urban Ho C~i Minh City The study based on the theories of household behaviour and time allocation to test the effects of women and households' characteristics on the number of housework hours of the wives The Ordinary Least Square regression and descriptive method are employed in the research

The findings are that the higher their earnings, the less hours of housework they do; the more young children (aged less than 15) they have, the more hours of housework they do Moreover, the number of housework hours of the wives has negative relation with the number of other adult female in households Furthermore, the involvement in housework of husbands proves to

be the significant determinant of the number of hours doing housework of the wives However, the wives' age, the educational level of both spouses, expenditure per capita, appliance of modern equipment in home production, and number of other adult male in households have no significant relationship to the number of hours doing housework of the wives

TABLE OF CONTENTS

CERTIFICATION : 1

ACKNOWLEDGEMENT ! 11

ABSTRACT 111

TABLE OF CONTENTS IV ABBREVIATIONS VII LIST OF TABLES VIII LIST OF FIGURES IX CHAPTER 1: INTRODUCTION 1

1.1 PROBLEM STATEMENT : l 1.2 SCOPE OF THE RESEARCH 3

1.3 OBJECTIVES OF THE STUDY 3

1.4 RESEARCH METHODOLOGY AND DATA COLLECTION 4

1.5 THESIS' STRUCl"URE , ,,,,,,,,,,,,., ,,,,,,,,,, 5

CHAPTER 2: LITERATURE REVIEW 6

2.1 DEFINITIONS 6

2.1.1 Household 6

2.1.2 Time allocation ' 7

2.2 THE THEORETICAL BACKGROUND 8

2.2.1 The economics of household behaviour 9

2.2.2 A work-leisure model of household 10

2.2.3 The theory of time allocation 12

2.2.4 Division of work within household ' 15

2.2.5 Human capital theory 16

2.3 THE EMPIRICAL REVIEWS 17

2.3.1 In developed nations 17

2.3.2 In developing countries 19

CHAPTER 3: MODEL SPECIFICATION 26

3.1 THE VARIABLE JUSTIFICATION 26

3.2 THE MODEL SPECIFICATION , 30

3.3 THE SAMPLE SIZE 30

3.4 SAMPLING AND DATA COLLECTION METHOD 31

CHAPTER 4: DETERMINANTS OF NUMBER OF HOUSEWORK HOURS OF MARRIED WORKING WOMEN IN TWO-EARNER HOUSEHOLDS IN HO CHI MINH CITY 33

4.1 A GLANCE AT HO CHI MINH CITY 33

v

4.1.1 General Socio Economic 33

4.1.2 Profile of women ofHo Chi Minh City 34

4.2 DETERMINANTS OF NUMBER OF HOUSEWORK HOURS 38

4.2.1 Descriptive analysis 39

4.2.1.1 Numerical summaries of variables 39

4.2.1.2 Pattern of time allocation 39

4.2.1.3 Bivariate analysis 41

4.2.2 Quantitative analysis using econometric model 47

4.2.2.1 Regression results: 47

4.2.2.2 Diagnostic testing for regression 49

4.2.2.3 Interpretation of coefficients and regression results 50

CHAPTER 5 CONCLUSION AND POLICY IMPLICATIONS 56

5.1 BRIEF SUMMARIES OF THE THESIS FINDINGS 56

5.2 POLICY IMPLICATIONS 57

APPENDICES 59

BIBLIOGRAPHY 71

vi

Department of Labour, Invalid, and Social Affair

Ho Chi Minh City General Statistical Office Gross Domestic Products Ministry of Labour, Invalid, and Social Affair Ordinary Least Squares

United Nations Development Programme Vienam Living Standard Survey

vii

Table 2.1

Table 3.1

Table 4.1

Table 4.2

Table 4.3

Table4.4

Table 4.5

Table 4.6

Table 4.7

Table 4.8

Table 4.9

Table 4.10

Table 4.11

Table 4.12

LIST OF TABLES

Summary of selected empirical studies of time allocation of housework 24

Definitions ofvariables 29

Percentage of households getting amenities of home life 34

Female labour divided by age groups 3 5 The active female aged from 15 divided by educationallevel 36

Sumtnary of variables 39

Pattern of weekly time allocation of spouses 40

Weekly housework hours of the spouses 40

The number of hours of housework by wives' educationallevel 43

The time used for housework according to the wives' age 43

Pattern of housework homs by household size 44

The effect of the presence of children aged 0-15 in the housework hours 45

The housework activities with respect to the domestic help 46

The OLS regression results 49

Figure 2.1

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

LIST OF FIGURES

The choice between Consumption and Leisure 11

The unemployment rate in urban Ho Chi Minh City in 1996-2002 37

The relationship between housework hours and wives' earJlings 42

The relationship between housework hours and market hours of wives 42

The relationship between housework hours and wives' age 44 Figure 4.5 The relationship between number of housework hours of wives and number of housework hours of husbands 4 7

CHAPTER 1: INTRODUCTION

In the 1990s, the world, especially the developing countries has experienced great achievements in economic development It is not only the achievements in economic growth, but also the improvement in the living standard of human beings, poverty and inequality reduction (UNDP, 1997) However, women seem to be less benefited from such achievements

as claimed by the United Nation that "women constitute half of the world's population and performed nearly two-third of the work hours, receive one-tenth of the work's income and own less than one-hundredth of the world's property" (Yarr, 1994: 111)

Traditionally, within household women are assigned to work at home while men are assigned to work in the market to earn for the household (World Bank, 2001) Women become the dependents of the men and thus they have few opportunities of participating in development process and enjoying the development's achievements (World Bank, 2001) The work of women at home is not considered as 'work' just because it is unpaid and then its contribution to the total welfare of household is ignored

In the modern society, together with the industrial innovation and women liberalization waves more women enter market work They in fact accomplish great achievements in many fields and their human characteristics and status are also improved (Alvarez· and Miles, 2002) However, they still have to maintain all the unpaid housework so that they are more burdened

by dual work (World Bank, 2001) Their load of work will increase when they get married and have children (Biau and Ferber, 1992) Due to the time constraint, time available for activities such as leisure, schooling and housework must be reduced as a result of an increase in time spent for paid work and social work However, which one will be reduced? Does time of doing

' housework reduce?

Vietnam has achieved rapid economic growth since the Doimoi policy implemented in 1986 Despite its great achievements, the inequality seems to rise (Gallup, 2002) In recent years, the issU((S of gender division of family labour at home have been paid much attention to Some

studies of time allocation of household members found that Vietnamese women are overloaded with disproportionate burden of housework (World Bank, 1999a) Vietnam female labor force increased 12 percent from 25,228,859 up to 28,325,968 persons in 1996-2000 (MOLISA, 2001) The second Vietnam Living Standard Survey in 1997 showed that the annual average hours of income-generating activities outside home are almost the same for both men and women aged over 15 On average adult male spent 1464 hours per year and adult female spent

1420 hours per year for income-generating activities (Desai, 2000) Although the hours per week that women spent on unpaid household activities had a trend to reduce from 15.52 hours

in 1992-1993 down to 12.75 hours in 1997-1998, their time spent for housework is nearly double that of men (GSO, 1999:197, 199) Moreover, in the survey of six localities' (Ho Chi Minh, Qui Nhon cities, Nam Dinh, Thai Binh, Dong Thap and Yen Bai provinces), it wasfound that although there exists the increasing trend of sharing housework among couples, the large proportion of housework still relies on the wives (Nguyen, 2000)

It seems that the reduction in household work cannot compensate for the increase in market work and that women expand their work at the expense of their time for rest, training, community activities and self-development (Yarr, 1994: 115) As a result, the heavy workload may have the negative impact on women's health, and in the long run, may affect the quality of their children (World Bank, 1999a) In other words, the women's traditional responsibilities within households may limit their participation in the society (Long et al, 2000) It raises an interesting issue that how to reduce women's burden workload and how to solve the confliction between their full-time market work and traditional roles at home when they participate in the paid employment outside home The problem is whether the women really liberate from the domestic chores within household when they increase the participation in the labour force, especially for married women It is interesting to examine, in this study, among various factors such as economic factors, household characteristics and composition which factors determine housework time allocation within the household in order to help married women who engage in income-generating activities out of the burden of household work In so doing, they will have more time to participate in other activities ~nd contribute more to the social welfare and economic development

2

1.2 SCOPE OF THE RESEARCH

In Vietnam, the problem of gender inequality has been recent concerns and many studies have been done However, most studies focused on the inequality in market work of women such as the wage inequality in Vietnam in the 1990s of Gallup (2002) and the gender wage gap in the study of Liu (2002) There are some studies on the inequality of division of labour between men and women in the household in terms of time allocation over various activities including productive, reproductive and leisure (Tran, 1997; Le, 1997, 1999a, 1999b; Hoang, 1998; Nguyen, 1998; Tran, 1998; Long et al, 2000; Le, Nguyen and Do (eds), 2002) Nevertheless, these studies just stand at the phenomenon and descriptive analysis None of these examines factors determine the household chores using the quantitative approach Moreover, most of these studies focus on the rural areas and ignore the urban areas such as Ho Chi Minh City where there are more job opportunities for women but the female unemployment rate is quite high (about 6-7%) (MOLISA, 2001) Therefore, in this thesis the urban area will be chosen

Ho Chi Minh City is one of the most dynamic cities with the growth rate of Gross Domestic Output of about 10.2 percent per year Its proportion of female labour force participation is quite high (53 percent) and there are more opportunities for them to engage into the society than other provinces So the confliction of women's labour force participation and their traditional household chores may be higher Therefore, Ho Chi Minh City is chosen for the research

With reference to the Vietnam Labour Code, the working age ofwomen is from 15 to 55 (Tran

j

ed., 1996: 9, 239); and the Vietnam Law ofMarriage and Family, the marriage age ofwomen is from 18 (Dinh, 2003:14) The research focuses on households located in Ho Chi Minh City in which the wives aged from 18 to 55 and both spouses engage in the income-generating activities

The thesis tries to examine the factors that determine the allocation of time for housework of married, working women in two-earner households in Ho Chi Minh City In so doing, some

implications in reducing the housework burden of the women and in turn, helping women fully · integrate into the society

In order to answer to the above research questions, the following hypothesis will be tested

1 The women's labour monthly earnings negatively affect her number of hours of doing unpaid housework per week

2 The women's time of doing market work per week negatively affects her time of doing housework per week

3 The number of other adults (male/female) over 15 years old in household negatively affect the time of married women who do paid work allocate for housework per week

4 The number of their children aged under 15 in household positively affect the time of married women who do paid work allocate for housework

5 The husband's hours of doing housework negatively affect the time of married women who

do paid work allocate for housework

6 The use of washing machine negatively affects the hours doing housework of wives

The analysis will be performed via the descriptive analysis and the regression ofthe self-survey data of time allocation of married working women in two-earner households in Ho Chi Minh City,

4

The sample will focus on the households that the working wives, who are the main person in charge of the domestic chores and the working spouses are the main decision makers in the households The wives are between eighteen and fifty-five years old and do not use their whole time available for study like students

The primary data will be collected using the questionnaire method and the Ordinary Least Squares will be employed to test the hypothesis of the effects of women's characteristics, the household characteristics on her number of hours of doing housework

The thesis will be organized into five chapters

Chapter 1 gives the introduction and the relevant as well as the limitation of the study Chapter

2 reviews the theoretical background and the empirical evidences about the time allocation of women Chapter 3 will be devoted to the specification of model and data sources used for the study Chapter 4 will present the results found and interpretation of such results Last chapter will summarize the study and suggest some implications that may help to reduce the burden of household chores of married working women in two-earner households in Ho Chi Minh City

5

CHAPTER 2: LITERATURE REVIEW

'

This chapter includes three parts The first part is devoted to some definitions of important concepts The second part is to review the theory of household economic behaviour in which married women are belongs to, labour supply theory and time allocation theory within household The empirical tests related to the time of doing housework of women will be displayed in the last part

The research is to examine the determinants of the number of hours the married women devote

to household work So it is better to agree on some concepts: household and time allocation

2.1.1 Household

Household is an important concept that economists have traditionally used as the basic unit of analysis, especially in living standard survey (World Bank, 1999a) According to Blau and Ferber (1992:7-8), "household consists of one or more persons living in one dwelling unit and sharing living expenses" while "family consists of two or more persons, related by blood, marriage, or adoption, living in the same household" With this definition, all families can be considered as households whereas one-person households or households consisting of unrelated members are not considered families These two terms are sometimes confused This definition

of household is also consistent with the official definition ofthe Vietnam government statistics The Vietnam Living Standard Survey showed that the typical Vietnamese household is nuclear family consisting of one male adult, his wife and two or three of their children, and this type accounted for seventy one percent of all households (UNDP, 2002:3) Thus, the concept of household instead of family will be employed in this study

6

is difficult to exactly define which activities are household chores or leisure (Kooreman and Wunderink, 1997) For example, childcare is often considered as one kind of household chores but some parents consider it as leisure Moreover, it is hard to determine the time devoted to each joint activity when more than one activity is nearly done at the same time (Kooreman and Wunderink, 1997) Nevertheless, this division of time use is still proved to be the most appropriate way for analyzing Thus, this three-way division of time use has popularly been used by researchers in examining the pattern oftime use of individuals in households (Gronau, 1977; Wales and Woodland, 1977; Khandker, 1988)

Therefore, in this study the time allocation will be divided into three categories: time spent for market work, for household chores and for leisure assuming there are no joint-activities It helps to examine the time that women use for non-market work

Market work is defined as the work that the labourer is paid In the study, the terms of market

work, paid work and market employment are used with the same meaning of generating income activities

Household chores (housework, unpaid work or domestic work) will be consisting of

"childrearing and the satisfaction ofbasis human needs through the provision of meals, clothing and shelter within the home" (Eatwell, 1996:678) In this study, household chores consist of daily childcare, cooking, washing, shopping and cleaning up activities

With reference to the definition' and classification of household activities of Singapore household economics survey, househO"ld chores are defined as follows (Quah,

1993 :220-223)

7

Childcare consists of "all activities related to the physical care of children" such as

' bathing, feeding and dressing Childcare also consists of parents' teaching and helping their children studying at home, taking them to school and back

Cooking consists of cooking, meal preparation and after meal clean up

Washing includes washing and activities associating with washing such as hanging

clothes on line for drying, ironing, and storing clothes Washing time excludes the time taken by machine (if any)

Shopping consists shopping for daily use such as groceries Shopping time also includes

time going to market/shopping place and back

Cleaning includes daily cleaning activities such as mopping the floor, sweeping the

house, washing doors and windows Cleaning also consists of small maintenance of furniture and other equipments at home

Leisure is the time left from the housework and market work The leisure time consists of pure

leisure, personal care and studying Since this study excludes women whose main work is studying such as student, it is no need to distinguish studying from leisure

Currently, there are two views on the gender effect on the time allocation among households in

' developing countries (Ilahi, 2001) One is the sociological view that social norms of the group people belong to drives the time use pattern while the other considers the economic factors affecting the way people allocate their time over different activities (Kooreman and Wunderink, 1997) In this study, the economic factors will be employed to test the hypothesis

Within the household, all members are interdependent because the decision of this member will affect the decision of other members and may alter the way others allocate their time over various activities (Kooreman and Wunderink, 1997) So it is important to explore the behaviour

of household members that is the start f6r time allocation theory

2.2.1 The economics of household behaviour

Household is considered as an economic unit that adopts two functions: one as consumer and one as producer A household is a producer when its members use the goods and services bought in the market as inputs for household activities in the so-called home production And it

is a consumer when its members use the goods and services for intermediate and final

Thus, the general model of household behaviour is the optimal solution of (i) the functions representing the household members' objectives and preferences called utility function; (ii) the constraints that household facing such as budget, time, education, wealth; and (iii) the functions reflect the interactions among household members It is used as the framework for examining household behaviour (Kooreman and Wunderink, 1997:2-5)

There are two main strands in the household economics The first is the unitary model and the second is the collective or bargaining model

The unitary model (Becker, 1965) treated household as a single consumer Although the household consists of many members who have different characteristics, income and consumption, household is assumed to be a unitary decision maker In this model, the decision making process will be dependent on the preference of the dictator such as household head or the person who controls the money or in other words, all members of the household have the common utility function (Wildmalm, 1998) The model becomes the maximization of a single-member household utility function subject to a individuals' time constraint, household joint budget constraint and home production technology given market prices and wage opportunities (Schultz, 1999:8) This model proves its usefulness in the process of determining household consumer demands and labour supplY: However, the unitary model does not discuss how families distribute consumption and welfare among members within the household if the members have different preferences (Schultz, 1999:9)

To deal with the shortage of unitary model, the collective household models have been introduced as a second strand (Chiappori, 1997) All collective models relax the assumption of common preference and allow each household member to have different preferences The collective model assumed the Pareto efficiency for the group with cooperative solutions of bargaining process to maximize the total weighted utility of all members with respect to the budget constraint and time constraint (Schultz, 1999) The collective approach of household behaviour has divided into two branches The first is the cooperative model in which household decision is the outcome of the bargaining process, and the bargaining power of each member depends on his fallback position, and the resources they contribute to the household (Basu, 1999) The second is the non-cooperative approach in which each member cannot agree with each other, and each member acts on the condition of other members' actions (Quisumbing and Maluccio, 1999)

Although each approach has its own limitations, both the unitary and collective models help to explain the intra-household resource allocation Despite the advantage of collective model, it is more complicated to apply in empirical studies and the model has not widespread been tested (Aronsson and Daundeldt, 2001 ) Most of Vietnamese households are nuclear households in which the husbands work outside and play important role in decision making while the wives are the subordinates to their husbands (Le, 2002) Moreover, Vietnam has experienced thousand years of feudalism so that the Confucianism heavily influences its people With the rooted oriental lifestyle and thoughts of which the male has more power in the household and the wife is considered as the dependent of her husband, the unitary model seems to be more appropriate for the case of Vietnam Therefore, in this study the unitary model will be employed as a basis for analyzing

2.2.2 A work-leisure model of household

In neo-classical economics, household was treated as a single-consuming unit Hence household labour supply can be examined using the individual labour supply The theory of labour supply is based on the consumer behaviour in which individual maximizes his utility with respect to the constraints (Bojas, 1996) In the neo-classical model, consumer has assumed

to be rational and has full information on prices and quality of goods and services, income and wage rates in his decision to choose working or leisure (Wundei-ink, 1997)

10

The model is to maximize the utility function U(xJ ,x0 , T1)

Subject to

Y=W.Tw+Yo

where X0 , Pn are quantity and price of good n

T~, T w are leisure time and labour time

T is total hours of non-sleeping hours available

Y, Yo are total income and non-wage income1

W is the consumer's wage rate

The constraints (2.1 ), (2.2) and (2.3) can be combined into

C + wT1 = wT + Yo where C = C(x,p) is consumption of goods and services

(2.1) (2.2) (2.3)

(2.4)

The left hand side of equation (2.4) is the total consumption of goods, services and leisure; while the right hand side is the full income of the individual The optimal consumption is the point where the budget line and the indifferent curve are tangent

Figure 2.1: The choice between Consumption and Leisure Source: Dijkstra and Plantenga, 1997: 19-; figure 2.1

1

Non-wage income is the income that is not earned from paid work such as interest on bonds, gift received

Point S gives the optimal choice between consumption of goods and hours of leisure that the

' person yields maximum utility The optimal choice depends on the iso-utility function, the wage rate and non-wage income When non-wage income increases, it has income effect that raises the demand for leisure time and reduces labour hours When wage increases, the full income2 increases and the budget restriction will rotate outward to the origin The income effect causes the increase in demand for consumption and leisure However, the substitution effect will decrease the demand for leisure due to the higher opportunity cost of leisure The total effect is thus dependent on which effect dominates Therefore, when wage increases, the labour supply may either increase or decrease At low wage levels, substitution effect is stronger than income effect while the reverse happens at higher wage levels so that the labour supply curve is bending backward (Kooreman and Wunderink, 1997)

In sum, the work-leisure choice model provides the ba.sis for analyzing the household labour supply However, this simple model has considered all other activities except work for wage as leisure This is a critical assumption because many activities in household cannot be considered

as leisure but as productive (Becker, 1965) The New Home Economic that takes into account household production and time of consumers in the production process has been emerged in the study of Becker in 1965 (Kooreman and Wunderink, 1997)

2.2.3 The theory of time allocation

The simple model of household labour supply, which is based on consumer choice between work and leisure, has been discussed in the previous section In this section, the model has been extended to take into account home production

To capture the limitation of simple neo classical model of labour supply, Becker (1965), in his theory of time allocation, considered household as both consumer and producer According to Becker (1965), goods and services can only bring utility if they are combined with consumer's time in his production and he called such goods and services Z-commodit)? For example, food

itself has no utility until it is prepared for meal and then consumed (Kooreman and Wunderink,

I 997) Becker (1965) divided time into time of consumption and time of work in his model

However, this division still has limitation because many household activities are in fact pure leisure such as watching television and investment activities such as studying (Wunderink,

I 997) Nevertheless, Mincer (quoted by Gronau, 1977) pointed out that these two are not the same in nature, at least in the case of women According to Gronau (1977) to some extent a person can substituted her market goods' time for time used for work at home (or time doing housework), but leisure is not Moreover, her housework is done to serve for the others and themselves but leisure is for her only Therefore, Gronau (1977) has revisited the theory of time allocation in which he distinguished three time categories: market work, housework and leisure The actual choice of household members will depend on time and money available to the household, and the ability to convert them into the Z-commodity from which the household derives utility (Himmelwelt et at, 1998:157) So, the model becomes maximization of utility function U subject to time, budget constraints and its production function

In Gronau's model (1977), the household member will yield utility via consumption of three categories: home-produced goods (N), market goods (M) and leisure activities (L)

The household utility will be maximization of the utility function:

where V is non-labour income ; W is wage rate from market work; Imax is full income

'

On the other ,hand, assume that the home-produced goods is a linear function of the market purchased goods used for home production and so is the time used for home production

N=N(X, TN) where X= y.N and TN= 8.N

(2.8) (2.9)

The shadow price of the time and labor used for home production is denoted by the function

The above model has been described in terms of single person household Gronau also extended the model to household consisting of more than one person (e.g the husband, the wife) assuming the household is a single unit of decision making In other words, although each household member has different utility function, they are in the same household and care about each other Hence, they will prefer to act with respect to an agreed household utility function to their own preferences (Himmelweit, 1998) Gronau's findings in this case are similar to the case of single person household

Although household members assumed to have one common utility, different members have different wages and productivity So if they specialize in certain activities that they have

4

Non-labour income is income that does not result from employment and is 110t affected by the amount of

individual work (Bryant, 1992:124)

14

comparative advantages, household will gain more benefit from such division of labour within

' the household (Wunderink, 1997)

2.2.4 Division of work within household

Based on the theory of comparative advantage, Becker (1981 quoted by Himmelweilt (1998) and Wunderink (1997)) modeled the division of work between the husband and the wife in the household Traditionally women were thought to have comparative advantage in housework and men in market work This may be explained that women are biologically the ones who bear children and they are traditionally believed to be better suited for housework than men (Blau, 1992) Besides, men and women are traditionally raised with different expectations and hence receive different education and training (Blau, 1992) So women may want to spend more time

in houseworU and men to spend more time in market work The more time she engages in household work, the higher productivity she reached Likewise the more the man works in the market work, the higher productivity he gains Moreover, the discrimination against women in employment Jowers their market earnings (Blau, 1992) As a result, in the household the wives have higher value oftime spent at home relative to market earning power as compared to their husbands Therefore, the household will maximize utility if the wife specializes in household work and the husband in market work Both biological differences and human capital concepts support this gender division of household labour (Himmelweilt, 1998)

However, such specialization may not always be optimal solution (Blau and Ferber, 1992) They listed some reasons as follow First, it is often assumed that women have comparative advantage in household work as a whole But household work consists of various activities such as cooking, washing, shopping, gardening that the women not necessarily have advantages

in all those activities Among those activities, some such as repairs and maintenance may better be suited for men Thus if the men involve in such work, household seems to be more efficient Second, according to the on-the-job learning, the more people work on the job, the higher productivity they gain (Bojas, 1996) However, to some extent, people may experience

"declining efficiency as they work more hours at a particular task" (Jacobsen, 1998:78) The third is due to the utility of each individual from market work, leisure and housework If individuals dislike the work, market or household work, their utility will decrease when they have to perform additional hour on that work Moreover, within the context of marriage, one

spouse may feel bored if they enjoy leisure or doing housework alone So ifthe wife engages in

market work and also maintains all the housework, her leisure time will be reduced as a result, and then the husband will enjoy leisure while his wife does the housework Both spouses will

be disutility Finally, does the wife and the husband substitute or complement in performing the housework (Schultz, 1999)? If it were substitute, specialization would be better But if their housework was complement, both spouses should perform housework All of the above reasons suggest the sharing of housework between the men and the women, especially the spouses

In sum, traditionally household could gain more from the specialization in which the husband specialized in the market work for earnings while the wife specialized in the household work However, the issue of sharing housework seems to be more desirable (Blau and Ferber, 1992)

2.2.5 Human capital theory

The theory of human capital developed by Mincer and Becker in 1974 (quoted by Bojas, 1996) argued that one important factor that determined wage rate of one person is his ability Each person has his own ability and skills that lead to the different productivity between him and other people, and in tum, this productivity determines the payment he gets The human capital theory claimed that education is one of the effective ways to increase person's abilities and skills, and then increase the productivity and finally increase his earnings Moreover, abilities and skills can be accumulated through the practice and repetitions of the work This is called 'on-the-job learning' In other words, experience is another factor affecting productivity For the interest of the thesis, the household chores of women will be explored This type of work can be considered as manual and repetitive work and women may have to perform housework since her childhood When women become older, she may have higher productivity in doing housework due to her experience and skills learned during her life cycles, and she may save her time doing the same activities We can use the age of women as one indicator of experience in doing household work Therefore, in this study the age and education of women may help to explain the time allocation of women within household

Many researchers have applied the above theories in explaining the determinants of household resources allocation of household members, of labour supply, education, health and time allocation (Gronau, 1977; Khandker, 1988; Ilahi, 2000, 2001; Aronson, Daunfeldt and

intra-Wilstrom, 200 I; Daunfeldt, 2002; and Alvarez and Miles, 2002) The theories did not limit in

I developed countries but they were also used to explain the behaviour of households in developing nations (Bryant, 1992)

In conclusion, the unitary model was useful in an-alyzing the allocation among households rather than individuals in the households while the collective approach seems to be more powerful in analyzing the inequality in allocation of individuals within the household However, it is not easy to measure the process of choice at individual level As Nelson (1995) claimed that there are too many restrictions and assumptions to make the bargaining models work In the context of Vietnam where people are still impressed by the Confucian thought that the husbands are the superiors and the wives are-inferiors, the unitary model proves to be more appropriate

The above theories have been used in both developed and developing countries in explaining the determinants of women time allocation within the household This section will present some selected empirical evidences with their findings

2.3.1 In developed nations

America

Gronau ( 1977: 113) has questioned if "an increase in unearned income reduce the work at home

of the non-employed while leaving the work at home of the employed unaffected." He uses the cross-sectional data of sample of 1,281 married women in which 660 were employed and 621 unemployed from the Michigan Study of Income dynamics to test the hypothesis The dependent variables in his model are time used for work in the market, time spent for housework and time for leisure The explanatory variables are women's age, education, work experience, husband's education and wage, family's unearned income, number of children and number of room in house and expected wage of wife He runs the Ordinary Least Square (OLS) regression and finds that there is a negative effect of unearned income and the husband's wage

on the number of hours of work at home when the wife does no!participate in the market work However, when women are employed, their wage rate becomes the major determinant of their

time use pattern Their wage rates have negative effect on the wife's hours of work at home and

' leisure The children have negative effect on her time use of housework and leisure ·

In 1994, based on the theory of human capital and the bargaining model, Hersch and Stratton used panel data from Michigan Panel Study of Income Dynamics for white, married workers aged 20-64 in the period 1985-1989 to find the determinants of division of housework between the couples and the determinants of the number- of hours of housework of each spouse They used the Ordinary Least Squares method to estimate two equations of housework time for each spouse and one equation of the husband share of the housework

On the one hand, the theory of human capital has argued that the household's gain the highest utility if specialization happened among household members due to their comparative advantages in household work or market work (Hersch and Stratton, 1994) On the other hand, the bargaining model has suggested that the household decision will be made through the bargaining process among its members and depend on the bargaining power of each member Whereas, the bargaining position of each member is assumed to be affected by the resources (e.g income) they bring to the household (Basu, 1999) Therefore, those theories suggest that the housework time allocation will be the function of relative position of both spouses, individual's characteristics, household characteristics and some other factors The explanatory variables are the husband's share of labour income, the household total labour income, the market hours of each spouse, educational attainment of each spouse, the number of children, the cohort and the regional variables

To my interest, only the estimation results from the determinants of wife's housework will be discussed Different from Gronau, Hersch and Stratton found a significant negative effect of total combined labour income and the husband's share of labour income on the hours of housework of the wife They also found that husband's characteristics such as age, education and time of market work have strongly effects on the housework time of wife Moreover, the more income the husband contributed to the household, the more housework the wife would do This is consistent with the bargaining t~eory that the higher contribution of income will lead to the higher bargaining power and induce specialization Hersch and Stratton empirically confirmed the positive effect of the presence of children on wife's housework This model can

be used as reference for the model used in the thesis in examining the determinant~ of number ofhousework hours of the working wife

2.3.2 -In developing countries

Bangladesh

Based on the theory of time allocation and household economic behaviour, Khandker (1988:111-26) explores the determinants of women's time allocation in rural Bangladesh The time allocation is divided into market production that is both inside and outside home for cash income and home production The maximization of household utility function subjects to the budget and time constraints yield ·the following reduced-form equations for women's time allocation:

where nH, Ow are the total time available for the husband and the wife

Pi is the price of good Z, X

V is other income

Wi is (wage rate) earnings from market work ofhusband and wife

8, y denote productivity in home production

Khandker uses the data of 444 households in rural Bangladesh to examine the determinants of time allocation of women The paper focused on the married women between the ages of 15 and 49, in which 195 women participate in the market work and 249 women work only at home The dependent variables are the time spent at home production and time spent at market production The explanatory variables are the women's age and education, husband's education and pre-marriage assets, landholdings, ?istance to schooling, distance to town, predicted male and female wage

He firstly used the probit model to estimate the women's labour force participation for the whole sample Then he used the Ordinary Least Squares method for estimation of home time

allocation Due to the 'work' or 'not work' in the market decision, the Tobit metho" was used

'

in estimation of market time allocation

Khandker (1988) finds that, for those who participate in the market work, women's wage rate and distance from home to school have positive effects on their time use in market production and negative effects on their time use in home production He found the different behaviors betWeen two groups of women: non-participants and participants in market work An increase

in wage of women will induce non-participants use more time for home production, which means producing more home products to substitute for market goods, while it induces participants to work more in the market work Since distance to town is also a factor that makes the market goods more expensive than home-based goods, women may decide to allocate more time for home production

Since the purpose of this thesis is to examine the time spent in home production ofwomen who participate in the market work, there's no need to use the Tobit method in estimation The Ordinary Least Squares method proves to be efficient in estimation the home time allocation of working women in urban Ho Chi Minh City

as regions Due to the interest of finding the determinants of the time used at home of women, only significant results of adult female in housework hours estimation equation will be discussed Skoufias found that female wage rate, years of schooling, age, marital status, the

number of boys aged 5-14, adult males and elderly and value of home capital have positive

' effects on the time spent at home while the number of female member including girls aged 5-

14, females aged 15-59 and elderly have negative effects on the time used for home production

of adult female

Peru

Ilahi (200 1) analyzed the determinants- of intra-household time use in Peru using the 1994 and

1997 panel data The different regressions were applied separately for women and men in rural and urban areas To my interest, only significant variables in housework equation of urban female will be presented The dependent variables are individual's share of the housework as percent of total work time The random-effect regression was used to estimate the housework equation It was found that the sickness of other adults, female's headship, individual's age, female aged over 5 have significant negative effects while children aged under 4 has positive effect on the female's share ofhousework

Vietnam

Mai (1997) used the random data of 100 households in four urban districts in Hanoi (Hoan Kiem, Ba Dinh, Hai Ba Trung and Dong Da) to test the effects of modern home appliance devices such as washing machines, refrigerators on the hours of housework done by the household members She compared the participation in household work of the spouses and number of hours they devote to housework before and after having the modern home appliances devices She found that the application of high technology in housework has the negative effect on the time of housework For example, without washing machines the women spend at least ten minutes to wash one clothe while they use only one minute to do the same work with the assistance of washing machine (Mai, 1997:11) Besides, the increasing share of husbands and other members in households in doing housework also helps to reduce the women's burden of housework However, these modern devices have weak effect on time women spend for shopping for daily use

21

Conclusion

' This section has given the theoretical and empirical literature in explaining the allocation of time of women among various activities The models used for analyzing can be summarized into two strands: the unitary model and the collective model Both approaches showed that the allocation of a women's time relies heavily on the composition of the household, the individuals' charactetistics which create her position within the household and the household resources Although the unitary model has been more criticized both by the theory and empirical studies than the collective model, it is still considered as the powerful approach Moreover, Vietnamese people still impressed by the Confucian thought that the husbands have more power in households and the traditional gender division of household labour so that the husband may act as the dictator Therefore, in this thesis the unitary model will be applied for analyzing the determinants of the number of hours that married working women allocate for household chores

• ••u•c ; ,:,urn mary 01 scJccrcu cmp1ncaJ sruu1cs ot nouscwork llours 111 L>oth tlevelopetl antl tlcveloping countries

Sign

-women

-\Stratton, spouses,

1994

-Husband and Wife's education (less/more high school)

+

+ women

-

I

Sign

'

-2001 rural, ' 1420 effect in total work (urban)

-Source: Gronau (1977), Khandker (1988), Skoufias (1992), Herch & Stratton (1994), Mai (1997), Ilahi (2001),

Note: Only the significant results are presented

-,

CHAPTER 3: MODEL SPECIFICATION

Following the theoretical and empirical reviews in chapter 2, this chapter will present analytical framework in examining the determinants of time allocation of working women

in two-earner households in Ho Chi Minh City First section will be devoted to the variable justification The model specification will be presented in second section And finally the data and sampling method will be discussed

The theoretical background and empirical review in the previous chapters have proved that household member's and household's characteristics can be employed to explain the time allocation for household chores of women The following will present the justification for the explanatory variables chosen and their expected sign

The wife 's and the husband's labour income

As mentioned before, the bargaining power of one member is often assumed to depend on the relative income that member brings in the household (Basu, 1999) and the decision of member having stronger bargaining power will influence the resources allocation within household, including time Also the gains from specialization in human capital theory suggest that individual earnings will decide who will devote more time to activities in the market or at home (Hersch and Stratton, 1994): Moreover, individual's earnings can proxy for the opportunity costs of her time of housework and the higher earnings may induce him/her to increase the participation in the market (Khandker, 1988) Therefore, the increase of the wife's earnings and her husband's earnings may affect the decision of her time allocation for the housework It may be expected that an increase in the wife's earnings will decrease her time of housework while an increase in her husband's earnings will increase her time of housework

The household expenditure

For all approaches presented above, income effect and substitution effects may lead to an increase in leisure or in market work of household members so that the work at home will reduce Thus, it is reasonable to use the variable of total household income However, due

to the difficulty in collection of precise non-labour income, the household expenditure will

be used as an alternative Household expenditure was also traditionally thought of as an effective alternative for the income indicator (Friedman, 1957 quoted by Gallup, 2002) It

is the direct method to measure the material well being attained by the household (Gallup,

2002) Moreover, the variable is needed because this can explain that the household may decide to substitute the market-produced goods for the home-produced goods assuming that there is no different utility in consumption of such goods Thus the inverse relation between the number of doing housework of the wife and the expenditure is expected However, to control for the correlation between the total expenditure and the number of members residing in the household, the expenditure per capita will be employed instead of the total expenditure

The market hour of both spouses

Given time constraint, the time of housework of the wife may be affected by the time she devotes to the market work Furthermore, within household one member's decision to do something may affect the time allocation of other members (Kooreman and Wunderink, 1997) In other words, if the husband decides to increase his time in market work, the wife may devote more time for doing housework given the amount of housework and the household production productivity In this thesis, the husband and the wife's market time are taken as exogenous Therefore, the husband's market time may positively affect his wife's time use in household chores and her own market time may negatively affect her time ofhousehold work

The educational attainment

In order to proxy for the earning opportunities for both spouses, I employ the variables of educational attainment In the human capital theory, the higher education the members obtain, the higher earnings they may have (Hersch and Stratton, 1994; Wunderink, 1997) And in turn, the higher earning the stronger bargaining power within household Then it may induce the members to engage more in the market work Thus the time devoted for the housework may reduce Moreover, the education level may also be proxy for the productivity of the member in the household production (Khandker, 1988) Furthermore, the education level may explain for the gender attitude (Hersch and Stratton, 1994) Therefore, the inverse relationship between the number of hours of housework of the wife and her educational level is expected while the expected sigl) between her housework time and her husband's educational level is ambiguous

Age of wife

With the same explanation, the age of the wife can be proxy for the productivity and the experience gained from the on-the-job training in the production process It means the older

' the wife, the less time she uses to perform the same amount of housework However, the age is also proxy for the attitude of the gender division of household work and the changing

in social norms In other words, the older the spouses are the stronger attitude of traditional gender division of labour So the older the spouses are, the more time the wife devotes to household work Therefore, it is expected that the impact ofthe wife's age on her own time

of housework is not determined, depending on which impact dominates

The number of children

The empirical evidences prove that the presence of children, especially the pre-school children, is an important determinant of time allocation pattern of women (Gronau, 1977; Hersch and Stratton, I 994; Alvarez and Miles, 2002; and Daunfedlt, 2002) It is assumed that women have comparative advantage in child rearing and bearing due to the biological structure So the childcare and responsibility for bearing will r~ly on the wife However, the presence of children of different ages will affect the pattern of time allocation differently (Bryant, 1992) The burden of childcare and bearing will reduce over time when children grow up and the addition of children to the family may induce the mother to increase her housework (Bryant, 1992) Therefore, the women tend to spend more time for caring children when their children are very young so that in this model, three variables of the number of children at different age ranges (0-5, 6-10 and I 1-15 ages) are employed to capture the change of the time use in housework over time The positive relationships between the numbers of children of different age ranges are expected with respect to the housework

The domestic help

The domestic help will be employed in this model because the other people can substitute for the wife's time in household chores This help can be the labour hired from the market

or the labour provided by the relatives (Alvarez and Miles, 2002) To my interest, I divide this domestic help into two variables One for the help of husband in housework in terms of the number of hours of housework and the other is the help of other people in terms of number of adult female and male aged over 15

Finally, the variable of durables available in the household will also be employed The durables can proxy for the technology and productivity in home production However,

different home appliances have different effects in terms of time saving time for housewives For example, the washing machine may save more time than the electronic cook In this thesis, the dummy variable for using washing machine (technology) is employed Therefore, it is expect to have the inverse relationship between techn.ology and the number of hours doing housework of the wife

The variables used in the model are presented in Table 3 I

MKTHUSB Weekly hours of market work ofthe husband Positive

EDUWIFE Dummy for High school of the wife (1 =yes) Negative EDUHUSB Dummy for High school of the husband (I =yes) +I-

CHILD6_10 Number of children aged from 6 to 1 0 Positive CHILDll 15 Number of children aged from II to I5 Positive

FEMALE Number ofother adult female aged over I5 Negative TECHNOLOGY Dummy, equals I if washing machine is used Negative

3.2 THE MODEL SPECIFICATION

In this thesis, the determinants of the wife's time allocation for household chores will be

' examined The scope of the thesis is restricted to the household in which both spouses engage in the market work and the household resides in Ho Chi Minh City The Ordinary Least Squares method will be employed The dependent variables will be the number of hours doing household chores per week

Mainly based on the model ofKhandker (1988) and with reference to the models ofHersch and Stratton, and the theoretical literature review in previous parts, the estimate equation of the number of hours of housework will be the function of the wife's earnings, husband's earnings, husband and wife's characteristics, and the household's characteristics

(3.1)

where Ww, wh are the monthly earnings of the wife and the husband respectively

T mw T mh are the weekly hours of market work of the wife and the husband

T11h is the weekly hours of housework of the husband

Xi is the vector of the individuals and household characteristics

t: is the error terms

The objects of the study are women aged from 18-55 in the two-earner households In the

1999 Population Census, Ho Chi Minh City had 1,016,741 households residing in 940,963 houses (GSO, 2000a) However, the exact number of households where both spouses engage in paid work is unknown Thus it is better to choose the minimum required sample size According to Mason et al (1999:290-1 ), the correct sample size is determined through the level of confidence, the acceptable error and the standard deviation of the population as below

(3.2)

where z: standard normal value from the confidence interval

cr: estimated standard deviation of population

E: maximum error

n: the sample size

The level of confidence and the acceptable error will be decided by the researchers while the standard deviation can be calculated from the pilot survey (Mason et al, 1999) After conducting a pilot survey of fifty women, the standard deviation found is 6.609 The maximum error will be 1.1955, seven percent of the mean (19.925) The level of confidence is ninety five percent, so the standard normal value is 1.96 Substituting all these factors into the formula, the minimum required sample size is 87 responaents

After deciding the sample size, the next step is to specify the sampling and the data collection method There are four popular methods of probability sampling (Mason et at, 1999) They are simple, systematic, stratified and cluster random sampling Due to the time and ability constraints, the systematic and cluster random will be combined to collect the information The interest of the study is the urban women so that ten out of seventeen urban districts of Ho Chi Minh City will be randomly chosen Ten respondents in each district will be interviewed In each district, a random starting point is selected, and then every next ten houses will be assessed.· The total estimated sample size would be 100 This combination method helps to reduce cost of sampling a population scattered over a large geographic area and may be the representative of the population (Mason et al, 1999)

According to Ilahi (2000:3), there are two common types of time use survey One is traditional method in which the respondent is asked how much time he/she spent in activity

j in the reference period (last twenty-four hours or last seven days) This method is less cost and we can have more observations in the household survey However, respondents can miss activities, over reporting, or they report the supposed hours used rather than the actual hours The alternative method is time diary method based on a 24-hour recall Although this method gives more details, captures more non-routine activities, and reports actual hours, it is too costly, requires more time for recalling and also happens the double counting of activities

The data for the regression is the time use of the spouses However, time use survey has often the problem of measurement bias The dairy instruments are believed to be the most valid method of time use survey despite its high costs (Juster and Strafford, 1991) The reasons of the bias are that people cannot remember all activities and its durations and