ED 053 329 AUTHOR TITLE INSTITUTION REPORT NO PUB DATE NOTE AVAILABLE FROM EDRS PRICE DESCRIPTORS IDENTIFIERS ABSTRACT DOCUMENT RESUME VT 013 691 Allingham, John D.; Spencer, Byron G. Women Who Work, Part 2: Married Women in the Labour Force: The Influence of Age, Education, Child-Bearing Status and Residence. Special Labour Force Studies Series B, No. 2. Dominion Bureau of Statistics, Ottawa (Ontario). Special Manpower Studies and Consultation Div. Cat-No-71-514-Occasional Dec 68 19p. Dbminion Bureau of Statistics, Publications Distribution, Ottawa, Ontario, Canada ($.50) EDRS Price MF-$0.65 HC-$3.29 Age, Education, *Employment Level Family Income, Family Status, Foreign Countries, *Labor Force, *Manpower Utilization, *Marital Status, *Working Women Canada To followup an earlier study of the relative importance of age, education, and marital status as variables influencing female participation in the labor force, this research attempts to measure the relative importance of similar factors in determining whether or not a, woman wor:cs or wishes to work. Particular emphasis was given to such determinants as age, child status, education of married women, residence, and education of husband. One important finding of the study was that the presence or absence of a young child remains, in the aggregate, the most important single attribute affecting a wife's participation in the labot force until the age of 44. It was concluded that the general imprOvement in the level of education is likely to induce more and more *ives to remain in the labor force after they are married, and to return to it when their major child-raising responsibilities in the home are passed. Still other factors that are expected to influence the 'participation of'women in labor force are the use of effective birth control methods and the general continuing rise in the husband's income. Part 1 of this study is available as ED 033 i99. (Author/JS)
Transcript
ED 053 329
AUTHORTITLE
INSTITUTION
REPORT NOPUB DATENOTEAVAILABLE FROM
EDRS PRICEDESCRIPTORS
IDENTIFIERS
ABSTRACT
DOCUMENT RESUME
VT 013 691
Allingham, John D.; Spencer, Byron G.Women Who Work, Part 2: Married Women in the LabourForce: The Influence of Age, Education,Child-Bearing Status and Residence. Special LabourForce Studies Series B, No. 2.Dominion Bureau of Statistics, Ottawa (Ontario).Special Manpower Studies and Consultation Div.Cat-No-71-514-OccasionalDec 6819p.Dbminion Bureau of Statistics, PublicationsDistribution, Ottawa, Ontario, Canada ($.50)
To followup an earlier study of the relativeimportance of age, education, and marital status as variablesinfluencing female participation in the labor force, this researchattempts to measure the relative importance of similar factors indetermining whether or not a, woman wor:cs or wishes to work.Particular emphasis was given to such determinants as age, childstatus, education of married women, residence, and education ofhusband. One important finding of the study was that the presence orabsence of a young child remains, in the aggregate, the mostimportant single attribute affecting a wife's participation in thelabot force until the age of 44. It was concluded that the generalimprOvement in the level of education is likely to induce more andmore *ives to remain in the labor force after they are married, andto return to it when their major child-raising responsibilities inthe home are passed. Still other factors that are expected toinfluence the 'participation of'women in labor force are the use ofeffective birth control methods and the general continuing rise inthe husband's income. Part 1 of this study is available as ED 033i99. (Author/JS)
CATALOGUE No.
71-514OCCASIONAL CANADA
s ecial la our orce s u ies
A
11
women who work: part 2married women in the labour force: the influence of age, education, child bearing status, and residence
DOMINION BURyl/ OF STATISTICS
DOMINION BUREAU OF STATISTICS
Special Manpower Studies and Consultation Division
U.S. DEPARTMENT OF HEALTH,EDUCATION & WELFAREOFFICE OF EDUCATION
THIS DOCUMENT HAS BEEN REPRO-DUCED EXACTLY AS RECEIVED FROMTHE PERSON OR ORGANIZATION ORIG-INATING IT. POINTS OF VIEW OR OPINIONS STATED DO NOT NECESSARILYREPRESENT OFFICIAL OFFICE OF EDUCATION POSITION OR POLICY.
SPECIAL LABOUR FORCE STUDIES
Series B, No. 2
WOMEN WHO WORK: PART 2Married Women in the Labour Force: The Influence of Age,
Education, Child-Bearing Status and Residence
by
John D. AllinghamThe Australian National University
andThe University of Western Ontario
andByron G. Spencer
McMaster University
Published by Authority of
The Minister of Trade and Commerce
December 19681500.504
Price: 50 cents
The contents of this document may be used freely but DBS should be credited when republishing all or any part of it.
SPECIAL LABOUR FORCE STUDIES
Cataloguenumber
71-505 No. 1 Educational Attainment of the Canadian Population and Labour Force/1960 - 65OccasionalPrice: 75 centsEstimates based on supplementary questions appended to the Labour Force Schedules of February 1960 and 1965,
including relationship between educational attainment and labour status and activity, and a comparison between native-born Canadians and post-war immigrants.
71-505F N° 1 Niveau d'instruction de la population canadienne et de la main-d'oeuvre: 1960- 1965 Irreguller Prix: 75 centsEstimations d'apres des questions supplementaires annexees aux questionnaires de la main-d'oeuvre de fevrier 1960
et 1965, y compris le rapport entre le degre d'instruction et le niveau et l'activite de la main-d'oeuvre, et une compa-raison entre les Canadiens de naissance et les immigrants d'apres-guerre.
71-506 No. 2 Annual Work Patterns of the Canadian Population /1964 Occasional -- Price: 75 cents
Annual work experience of the Canadian Population is compared with data from monthly surveys, and includes ananalysis of long-duration unemployment, and part-year and part-time work.
71-507 No. 3 The Job Content of the Canadian Economy/1941- 61 OccasionalPrice: 75 centsA review of the theory and measurement of Job Content together with an attempt to estimate the kinds of jobs, in the
Canadian economy, by function and levels. A comparison is made with the United States.
71-508 No. 4 Geographic Mobility in Canada/October 1964October 1965OccasionalPrice: 75 centsMigration of the Canadian population between municipalities is analysed by age, sex and region. For male migrants,
aged 17-64, labour force status and reasons for leaving are also included.
71-509 No. 5 Women Who Work: Part 1OccasionalPrice: 75 centsAn evaluation of the relative importance of age, marital status, and education as factors influencing the participation
of women in Canada's work force. The study is based on special tabulations of 1961 Census data.
71-510 No. 6 Labour Force Characteristics of Post-war Immigrants and Native-born Canadians: 1956- 67Occasional--Price:75 cents
Differences between the labour force participation rates of post-war immigrants and native-born Canadians arecompared taking into account differences in the age, sex, marital status, regional and educational attainment distributionsof the two population groups.
71-511 Series B No. 1 The Demographic Background to Change in the Number and Composition of Female Wage Earners inCanada, 1951 to 1960OccasionalPrice: 75 cents
An evaluation of demographic change over the 1951-1961 decade and its impact on the composition and number offemale wage earners in 1961.
71-512 No. 7 Educational Attainment in Canada: Some Regional and Social AspectsOccasionalPrice: 75 centsAn examination of regional and occupational differences in educational attainment in Canada and its relation to
migration is followed by a study of intergeneration changes in educational attainment.
71-514 Series B No. 2 Women Who Work: Part 2 Occasio-ml Price: 50 centsAn evaluation of the influence of age, education of the wife, education of the husband, child status and residence on
the participation of married women in Canada's work force, based on tabulations of the 1961 Census data.
Note: Les bulletins Nos. 2, 3, 4 et 5 seront bientot disponibles en frangais.
Remittonces should be in the form of cheque or money order,mode payoble to the Receiver General of Canada and forwardedto the Publicotions Distribution, Dominion Bureau of Sfotistics,Ottowa, or to the Queen's Printer, Hull, P.Q.
Roam DUHAMEL, r.s.s.c., Queen's Printer and Controller of Stationery, Ottawa
3
FOREWORD
In 1966 the Special Labour Force Studies was initiated by the DominionBureau of Statistics (see list on page 2). The research studies were designed toreach a broad audience interested in the changing nature and composition of theCanadian labour market. Some aspects of manpower development, however, requirea somewhat more technical analytical approach. For this reason, the presentseries (B), of which this is the second study, will be published as a companionseries of technical papers.
These studies are prepared under the direction of Dr. Sylvia Ostry, Director,Special Manpower Studies and Consultation.
The responsibility for the analysis and interpretation of the data is that ofthe authors and not, of the Dominion Bureau of Statistics.
WALTER E. DUFFETT,Dominion Statistician.
4
AUTHORS' PREFACE
This study was begun in the summer of 1967 in the Special ManpowerStudies and Consultation Division of the Bureau, and completed during the winterof 1967-68. We wish to express our thanks to Dr. Sylvia Ostry and to Mr. NormanDavis for many valuable discussions, and to Norman Davis for over-seeing thecomputer output.
Most of the calculations presented here were produced by REGRO-41, one ofthe multiple regression programmes prepared in the Scientific ProgrammingSection of the Bureau.
TABLE OF CONTENTS
Page
Introduction 9
Background to the Study, and its Purpose 9
Some Major Determinants of the Labour Force Participation of Married Women 10
Simultaneous Consideration of all Factors sited: The Technique of Multiple Regression Analysis 12
Analysis of the Multiple Regression Results 13
Analysis of the Wife's Labour Force Status: The Focus on Age and Residence Differentials 17
Summary and Conclusions 20
List of TablesPage
1. Analysis of the Labour Force Participation Rates of Married.Women, Camu la, 1961 13
2. Ranking of Labour Force Participation Incentives 15
3. Analysis of the Labcar Force Participation Rates of Married Women, Canada, 1961 (Standardised byResidence Variable) 18
4. Ranking of Labour Force Participation Incentives (Standardised by Residence Variable) 20
6
INTRODUCTION
"The role of women in a community is a mostimportant one and no nation can afford to ignoreit. Harnessed and properly controlled, theypresent a great and powerful force which can beused for the benefit and progress of the nation.Left to run wild, however, or simply ignored,they will be as locusts in the nation's corn-fields."'
In Canada the harnessing of the nation's woman-power, in the sense of bringing increasing propor-tions of the female population into the labour force,has been going on for some time, and has recentlygained substantial dimensions. The followingtrends have emerged:(1) the proportion of women in the labour force con-
tinues to rise: in 1967, 30.7 per cent of the la-bour force was comprised of women in contrastwith only 23.9 per cent in 1957, and 21.9 percent in 1947;
(2) an increasing proportion of married women is inthe labour force: in 1961 the figure was 22.5 percent, and rose to 24.4 per cent in 1963, 27.7 percent in 1965, and 30.5 per cent in 1967.2
Such marked changes in the pattern of female labourforce behaviour lead one to ask why and what con-tinuation might be expected in the future.
This study is an extension of an earlier probeinto some of the determinants of the labour forceparticipation of women in Canada.' The first studywas concerned with the relative importance of threevariables age, education and marital status ininfluencing participation in the labour force.
The further growth in the proportion of womenin the work force is likely to occur only with a risein the labour force participation rate of marriedwomen.4 At the moment, married women, while more
likely than ever before to be gainfully employed,still constitute the largest pool of potential labourin the country, outside the school-age population.
The considerable rise in the participation ofmarried women in the work force has undoubtedlyhad some effect on family life and marital relation-ships.' The implications of a continued upward trendin the employment of married women are unclear.There is, however, considerable evidence that child-bearing patterns are being affected,' suggesting aconflict between the roles of "mother" and "work-er".
Thus far, the work force has attracted from theranks of married women the greatest increase amongwomen in their late thirties and forties.' Given thatthe average woman is finished with childbearing byher mid-thirties,' the impact on fertility of recruitingthese "older" women into the work force will beminimal. The greatest potential effect on fertilitywould be achieved through a delay in childbearingamong women who choose to continue working forsome time after marriage.
The economic implications of a massive return-to-work movement by married women not now in thelabour force must be substantial. While there will beconsiderable debate about the implications of therising work force participation rate of married wom-en, there can be no doubt that such a rise has beencontinuing for some time and shows no signs ofdecelerating. This study is an attempt to assess therelative importance of several factors thought toaffect their labour force participation. By rankingthe importance of various influences one it hopeful-ly in a better position to undertake the importanttask of projecting future change in participationrates.
BACKGROUND TO THIS STUDY
The participation rate of married women in thelabour force will be greatly affected by the complexinterrelation of the demand for labour with factorsaffecting the numbers forthcoming on the market, in
An excerpt from an address given to women by agovernment leader from a developing nation, reported inHottel, A.K. "Women Around the World: An IntroductoryComment", The Annals of the American Academy ofPolitical and Social Science. January, 1967, p. 1.
2 Calculations based on figures published inDominion Bureau of Statistics, The Labour Force, variousissues.
3 Dominion Bureau of Statisticspecial LabourWForce Studies', No. 5, Women Who Work: Part I, by
Allingham, J.D., Ottawa, 1967.Some demographic factors creating this situation
are explored in, Dominion Bureau of Statistics, SpecialLabour Force Studies, Series B, No. 1, The DemographicBackground to Change in the Number and Compositionof Female Wage Earners in Canada: 1951- 1961, byAllingham, J.D., Ottawa, 1967.
particular with the community's values concerningthe"proper" roles of married women. The complexityof issues raised in discussing the supply of anddemand for female labour has been outlined by other
A summary of research in this area is reportedin he Employed Mother in America, by Nye, I. andHof1in,ln, L., Chia go, 1963.
6 Blake, J. "Demographic Science and PopulationPolicy" in Sheps, M. and Ridlsy, J.C. (eds) PublicHealth and Population Change, Pittsburgh, 1968,pp. 62 -7.
7 Dominion Bureau of Statistics, The FemaleWorker, by Ostry, S., One of a Series of Labour ForceStudies in the 1961 Census Monograph Programme,Ottawa, 1968.
Glick , P. and Parke, R. "New Approaches inStudying the Life Cycle of the Family", Demography,Vol. 2, 1965, pp. 187-202. The rapid decline in age-specific fertility rates in Canada after age 35 suggeststhat Canadian women are behaving like American womenin this respect.
si/
10
researchers and will not be pursued here.9 Our aimis less ambitious. A series of cross tabulations wereprepared from the data collected in the 1961 Censusof Canada in order to explore variations in the parti-cipation rates of married women having differentattributes." While recognizing the complexity of anexhaustive theory of labour force participation, webelieved that data on a sufficient range of attributescould be derived from the census to illuminate majordeterminants of labour force participation.
The purpose of the present study is to focus onthese various attributes of married women(their edu-cation, number of children, and so on) and to try tomeasure the relative importance of each of thesefactors, both separately and together, in determiningwhether or not a woman works or wishes to work.
Undoubtedly when a married woman considersgoing to work, or leaving work, there are many fac-tors which will be unique to her. There are also,however, factors which she will have in common withother married women of her age. One of these, forexample, may be residence grouping. Furthermore,the "environmental" characteristics may well domi-nate the weightings which most women give to thenumerous factors influencing their decision. We as-sume that there is sufficient homogeneity in the be-haviour pattern of married women that the commonattributes which we note do, in fact, dominate theirlabour force behaviour, the novelty of each casebeing largely subsumed. Accordingly, we focus ouranalysis on several such major attributes. The at-tributes which we have selected we designate thedeterminants of the married woman's decision towork.
MAJOR DETERMINANTS
Cross tabulations were prepared of participationrates of married women having all the possible com-binations of the following five attributes: age, childstatus, wife's education, husband's education andresidence." The theoretical rationale behind theselection of each of these attributes for examinationis outlined below.
Age
For most citizens the law sets the lower agelimit both for marriage and for gainful employment,and social custom sets an upper limit. In keepingwith these constraints, we examined participationrates of married women between the ages of 15 and64 years.
Although the patterns cl labour force participa-tion within the age range 15 -64 is greatly differentfor males and females," age per se, i.e., the matu-ration of the body, is unlikely to be of major impor-tance on either the supply side or the demand sideof the labour market within the age span noted.However, social custom is such that many jobs arerestricted to persons within fairly narrow ageranges."
"Changes in the participation of women withinthe 15-64 age range may best be understood ifage is considered as a dimension along whichcertain role changes occur to a majority ofwomen. These role changes tend to either facil-
g "The Interaction of Demand and Supply and itsEffect on the Female Labour Force in the UnitedStates", by Oppenheimer, Valerie K., PopulationStudies, Vol. 21 (3), pp. 239-59, November, 1967.
10 The tabulations were prepared at the DominionBureau of Statistics during the winter of 1967.
11 Tabulations were prepared for all "normalfamilies" i.e. husband present.
12 Dominion Bureau of Statistics, The EconomicStatus of the Aging, by Ostry, S. and Podoluk, J., p. 20,Ottawa, 1966.
13 For example, a contemporary citizen might wellbe shocked to see a 55-year-old airline stewardess.
itate employment or to impede employmentalthough none are inherently significant toparticipation decisions. Rather, the significanceof certain roles for participation in the labourforce must be seen as culturally relative, andas such, changeable.'
Thus, in order to isolate the influence of age,as socially interpreted, from obscuring factors suchas childbearing we have taken several potentiallymasking factors into account in our analysis.
Child StatusWe define child status as having one of the
following as an attribute: no children, all childrenunder six years of age, all children aged 6 -14 years,or "other" (children in both age groups or over age14)."
Several studies have suggested that the bearingof children and their rearing constitutes a majorbarrier to the gainful employment of women."
A conflict between the roles of mother andworker is of course not inevitable but largely re-flects the institutional structure of our society.
"The less the cost in transferring childbearingtasks to others, or in incorporating them withinthe job, the greater the compatibility. Compati-bility can stem from the nature of the task orfrom the social organization of child care"."
14 Dominion Bureau of Statistics, Special LabourForce Studies No. 5, op. cit., P. 11.
Is Ideally the number of children in any categorywould be added, but restrictions on tabulation preventedthis. The residual child status category "other"appears to be primarily made up of women who have bothchildron ;Alder age 6 and in the 6-14 age group. Onceagain, estrictions in tabulations prevented a finerbreakdown.
16 The Female Worker, op. cit.""Female Working Roles and Fertility", by
Stycos, J.M. and Weller, R.H., Demography, Vol, 4(1), P. 215. 1967.
11
Aside from cottage industry economies, work isusually in a location other than the home, thus re-quiring absence from any pre-school children duringworking hours. Family structure has altered in sucha way that few couples have relatives living withthem. In order to work, most mothers of pre-schoolchildren must hire child care. Unlike a number ofEuropean countries, Canada does not have a systemof low cost day-care centres for small chidlren."
In Canada, the conflict between the roles ofmother and worker is likely to be most keenly feltwhen any children are of pre-school age. After theyoungest child enters school, the custodial portionof a mother's role decreases considerably, freeingher for other activities. The deterrent affect on la-bour force participation of pre-school children hasbeen noted in the literature." Our analysis suggeststhat the presence of pre-school age children consti-tutes the single most important deterrent to labourforce participation.
"The far greater porportion of a woman's workoccurs after she has raised her family and re-entered the labour force.Most35-7ear old womenhave completed their families. The work lifeexpectancy of a 35-year old woman who hascompleted her family and who no longer hasresponsibility for the care of young children is24 years, essentially the same as for a newlymarried 20-year old woman.""
Education of Married Women
"Women's interest in employmentand theirsuccess in obtaining itare strongly influencedby the extent of their education and training.'
The direct relationship between education ofwomen and their participation rates has been ob--served by many researchers?' Such findings are tobe expected for several reasons, among them thefollowing:
(1) the demand for labour is generally stronger forthose with higher education;
(2) remuneration for employment is likely to behigher for a more highly educated woman, th.;sdiminishing the proportion of her income thatwould have to go for child care services;
(3) the more highly educated the wife, the morelikely she is to feel "unfulfilled" in the house-wife role and therefore the more open she islikely to be to the blandishments of the marketplace.
" "Day Care Services for Children of WorkingMother", Bulletin of the Women's Bureau, Departmentof Labour, Ottawa, 1964.
19 The Female Worker, op. cit.10 Work Life Expectancy and Training Needs of
Women, Report No. 12, U.S. Department of Labor, Man-power Administration, Washington, 1967.
31 p. 135.33 For example, see The Female Worker, op. cit.
Observation of participation rate differences byeducation alone will however create a false impres-sion of the importance of education for participation.The reason is that women in the different educationalcategories are also differentiated by average age,education of their husbands, residence, and, evencontrolling for age and marriage duration, by numberof children. Thus, in comparing the participationrates of married women, some with university educa-tion and others with elementary schooling or evenless, one obtains various confounding effectsforexample, the residence factor: university trainedwomen are more highly urbanizedor the child-status factor: university-educated women tend tohave fewer children.
In this paper the influence of education onparticipation is assessed independently of otherfactors for which we can control.
ResidenceOne might expect a higher participation rate for
married women living in more highly urbanized areasbecause the concentration of white collar jobsvaries directly with urbanization. Moreover, it is inthe cities that social change is manifest. Attitudesfavourable to working wives are more likely to bewidespread in the cities.
"In eight years of [city?] marriage, I have neverironed a sheet, washed a shirt, darned a sockor cooked three consecutive meals"
In the cities, both labour market conditions andcommunity attitudes are perhaps most favourable tothe employment of married women.
Any comparison of participation rates of marriedwomen by residence alone will be misleading be-cause, among other differences, women living in bigci'les are more highly educated than their less urbansisters, and have fewer children. The factor of resi-dence is therefore assessed independently of othervariables, which we suspect obscure the relation-ship."
Education of the HusbandAn inverse relationship between the income of
husbands and labour force participation of wives ha:,been observed in Canada and elsewhere.25 Sinceeducation is highly correlated with income for males,
33 "The New Woman", by Fraser, S., Star Weekly,January 6, 1968, p. 18.
34 Duration of residence in an urban area may alsobe quite important. The recent migrant to the city islikely to reflect the attitudes of the area from which hehas come. The problem of residence duration is discussedin "Another Look at the Indianapolis Fertility Data",by Goldberg, D., Milbank Memorial Fund Quartely, Vol.38 (1), January, 1960.
3g "Labour Force Participation of Married Women:A Study of Labor Supply", by Mincer, J., Aspects ofLabor Economics, National Bureau of Economic Research,pp. 63-97, Princeton, 1962. For Canada see, The FemaleWorker, op. cit.
12it is not surprising that our results show an inverserelationship between education of husbands andlabour force participation of their wives. As ex-plained below, for a variety of reasons, includinglimitations of data, in this study we use the educa-tion of the husband as a proxy for his income.
The inverse relationship observed with hus-bands' income suggests that the economic motive
for working is a strong one for women.26 Such amotive may well decline with an increase in hus-band's income beyond a certain point. While wivesof men with a higher level of education are likely tobe more educated themselves, thus probably creatinga most, liberal climate for the wife's roles, desiresfor extra-familial fulfillment and freedom to realisesuch desires are perhaps not enough to counteractthe affluence provided by the husband.
SIMULTANEOUS CONSIDERATION OF THE FACTORS CITES:THE TECHNIQUE OF MULTIPLE REGRESSION ANALYSIS
The above analysis indicates that each of thefactors mentioned exerts a distinct influence on thelabour force behaviour of the married woman. Whatwe have still to 'earn is the relative importance ofthe factors: an vanced education moves a womaninto the labour force; if her husband has a high in-come she is deterred. Which effect is stronger? Doesthe presence of young children exert a stronger pullkeeping the wife at home than the apparent pull intothe labour force exerted by residence in larger cen-tres? Do the relative strengths, of these factors varysystematically with the age of the wife'?'
Above we have focused attention on one or otherparticular cha7acteristic. It is, however, evident thatthere are a large number of factors which simultane-ously affect the rate at which women in each agegroup enter the labour force. We shall, therefore,depart from our reliance on consideration of eachfactor acting alone and use a second technique whichis better suited to analyse the simultaneous andseparate impact of each of the factors we have out-lined. The technique' used here' is multiple regressionanalysis in which we subdivide each of the variablesmentioned into two or more characteristics, assign-ing "one" values when that characteristic is presentand "zero" values when it is not. As we shall see,this enables us to discover the relative importanceof each of the factors discussed above; we are alsoable to consider the factors in combination to gain,for example, a better idea, of a women's "marketa-bility" (especially as indicated by both her resi-dence and education characteristics) and of herprobable desire to be in the labour force (notablyher education and her husband's income, both cur-rent and projected).
To use multiple regression analysis in thismanner we assume that each of the factors outlinedabove has an additive and separate, impact on therate at which married women participate in the labourforce in Canada.27 The form in which we have testedthis hypothesis is the following:
26 Financial reasons are most often given for work-ing. See: Dept. of Labour, Married Women Working for Pay,Ottawa: Queen's Printer, 1958, and Nye and Hoffman, op.cit., pp. 23-26.
21 The assumption that there is little interactionamongst the factors is supported by preliminary analysisof variance runs. We found some interaction betweenresidence and child status and have therefore constructedseparate regression equations for each residence grouping.
P=bo
+b.E +b.Eh
+bk C+b1 R+Uj
whereP is the labour force participation rate for
each group of married women;bo is the average rate at which married women
participate in the labour force;E is the educational attainment of the wife,
w divided into four (i = 4) categories. Thecategories are (1) Elementary education orless; (2) One to three years of secondaryschool; (3) Four to five years of secondaryschool; (4) University, complete or incom-plete;
Eh is the educational attainment of the husband,divided into four (j = 4) categories. Thecategories are the same as for the wife;
C is child status, divided into four (k = 4)categories: (1) No children; (2) All childrenaged less than, six; (3) All children aged 6to 14; (4) All other;
R is residence, divided into four (1 = 4) cate-gories: (1) Metropolitan; (2) Other urban;(3) Rural non-farm; (4) Rural farm;
U the unexplained portion.and the coefficients b.
. kbj b and b1
indicate the
influence that the variables Ew' Eh, C, and R res-pectively, have on the participation rate, P.
The data which have been used are from the1961 Census of ,Canada. A complete tabulation ofmarried women was taken, and the results then crossclassified into a number of cells each comprising aunique combination of the above-mentioned charac-teristics (e.g., a cell 'in which the wife has, one-to-three years of secondary schooling, the husband hasfour-to-five years, the wife of age 15 -24, there areno children', and the family lives in a metropolitanarea). For each of these cells the labour force parti-cipation rates are noted. The primary reasons forfocussing on these particular characteristics havebeen discUssed, above. We can now proceed to ana-lyse the information obtained from the multipleregression , runs, the major resultS of which arereported in Tables 1 and 3.2°
26 The regressions were produced at the. DominionBureau of Statistics during the summer, fall and winterof 1967, The authors wish to acknowledge the extensivecontributions to this part of our analysis by Mr. NormanDavis of Special Manpower Studies and ConsultationDivision.
10
- 13 -
ANALYSIS OF TABLE 1: THE LABOUR FORCE PARTICIPATION RATES OF MARRIED WOMEN
In order to gain an over-all view of the labourforce activity of married women we proceed first withthe analysis of the labour force participation ratesof married women, aged 15-64. In this section wepay particular attention to the method of analysisand to the interpretation of the tables in order tofacilitate subsequent discussion when women aregrouped first by age and then by residence.
Interpretation of Table 1Consider the first column in Table 1. From it
we can learn the probability that an "average" mar-ried woman in Canada in 1961 was in the labourforce-the probability is 24.96 per cent, the figurebeside common term in the table. Suppose we weremore interested in married women between the agesof, say, 15 and 24. The figure 6.71 beside the 15-24age group tells us that a wife falling into this agegroup is 6.71 per cent more likely to be in the labourforce than the average-that is, of every 100 suchwomen we would expect, on the average, that 31.67(or 24.96 plus 6,71) of them would be in the labourforce. Similarly if we were primarily interested inmarried women, aged 15-24, who had, say, elemen-tary education, we find the probability that such awoman would be in the labour force to be (24.96 +6.71 8.03, or) 23.64 per cent.
Before proceeding to a further discussion of theimportance of the various attributes on which we arefocussing, let us ensure an understanding of the tableby a second example. Consider the following ques-tion: what is the probability that a married woman,who has 4 to 5 years of secondary schooling, whosehusband has university education, whO has no chil-dren, and who lives in a rural non-farm area will bein the labour force? The answer is found by addingthe (positive or negative) contribution of each of thefactors mentioned to the common term. In this par-ticular case the detailed calculation is as follows:
Common termPlus contribution of specified attributes:
24.96
Wife's education 1.62Husband's education 5.52Child status 14.84Residence 1.54
9.40
Total 34.36
Thus we conclude that of 100 such women wewould expect, on the average, approximately thirty-four of them to be in the labour force.
TABLE 1. Analysis of the Labour Force Participation Rates of Married Women, Canada, 1961
Variableor
attribute
Age group
15 64 15 24 25 34 35 44 45 -.54 55 - 64
Age of wife:15 - 24 6.7125 34 3.2735 - 44 2.3945- 54 - 0.2355 - 64 -12.14
Common term 24.96 31. 50 28.22 32.41 29.79 18.67R2 . 72 .85 .87 . 83 .83 .79
Category excluded because too few observations available.
11
- 14
Discussion of Table 1: The Case of All MarriedWomen
The first column of Table 1 tells us much aboutthe impact of the determinants of labour force statusas they affect the entire group of married women inCanada.
Looking first at the age variable, it is evidentthat the youngest group of wives is much more likelyto be in the labour force than is the oldest group. Itappears that the influence of age alone is steadilyto reduce the rate at which wives participate in thelabour force.
The wife's education also has a marked effecton her labour force participation: the better educateda woman, the more likely it is that she will be eitherworking or looking for work. The probability that awoman in the most highly educated group (university)will be in the labour force is 19.3 percentage pointsgreater than for a woman having only an elementaryeducation. No doubt this observation reflects partlythe tendency of a more highly educated wife toachieve greater personal "fulfillment" by playingan active role in the labour force, partly her relative-ly greater "marketability" resulting in large measurefrom her advanced education, and partly her "oppor-tunity cost" the amount of income foregone if sheis not in the labour force.
The husband's education is, on the average,highly correlated with his income position: the morehighly educated the husband, the better his incomeand income prospects are likely to be. In this study,therefore, the education of the husband variable hasbeen used as a proxy for his income.29 For somepurposes it is more informative than his actual cur-rent income since, in the aggregate, and when com-bined with knowledge of his age," it also tells ussomething of his income prospects. A young, well-educated man should expect his income to rise at amuch faster rate than a less well educated man ofthe same age."
Using the husband's education as a proxy forhis income, Table 1 tells us that a wife is lesslikely to be in the labour force, the higher her hus-
° As explained above, education is used as aproxy variable for income also because of restrictionson tabulations. It would be desirable to include boththe husband's current income and his education in theanalysis.
3° The husband's age is typically within a fewyears of that of his wife. For the 1961 age distributionsee 1961 Census of Canada, Bulletin 2.1-11, Table 95.
31 Thus a man with a high level of education will,in general, earn much less early in his career than hewill later. On the other hand, a man with a lower levelof education will reach his peak salary early in hiscareer. This is not to say that he will not expect hisincome to rise over time with' the rise in the generaloutput of the nation. It means simply that he could notexpect his income to be much greater at any point intime than that of similarly educated people just enteringthe labour force. See Earnings and Education, by J.R.Podoluk, Advance Release from Census Monograph,"Incomes of Canadians", Dominion Bureau of Statistics,Ottawa, December, 1965.
band's income. A woman whose husband is in a rela-tively low income position is, on the average, 9,32per cent more likely to be in the labour force thanone whose husband has a relatively high income.
Looking now at the child status variable, forall married women the absence of children is a verystrong incentive to be in the labour force whereasthe presence of young children (those less than sixyears) provides a strong deterrent to entry. This, ofcourse, is as one would suspect. But perhaps thestrength of the two forces is surprising: for all wivesaged 15-64 the probability that a wife will work is30 per cent higher if there is no child than if thereis a child under six years. The presence of one ormore children aged 6-14 appears not to keep a wifefrom the labour force. Children in this age group aretypically in school during most of the working day,hence requiring less attention from the mother, andfreeing her to enter the labour force. The category"other" is less than ideal from a research standpointcomprising older children as well as some combina-tions of children in the less than six and 6-14 agegroups. One suspects that the "other" grouping isdominated by families having children in both theless than 6 and the 6-14 age groups.
The residence variable also has an impact, butit is the weakest of the factors considered here. Thelarger the centre in which a wife is living the morelikely is she to be in the labour force. But a metro-politan wife is only about 6.5 per cent more likely tobe in the labour force than her rural farm conterpart.
For all women the, measure R2 reported at thebottom of the table indicates the proportion of thetotal variation in the labour force behaviour of mar-ried women accounted for by the factors we haveconsidered. The .72 tells us that the factors wehave considered account for 72 per cent of the totalvariation. There are, as we have mentioned before,innumerable factors which are not widely shared, andwhich cannot therefore be taken into account, butwhich influence the labour force behaviour of indivi-dual married women. For example, some marriedwomen who would otherwise be in the labour forceare tending to aged or otherwise incapacitated rela-tions, retarded children, or large gardens. Perhapsit is surprising that as little as 28 per cent of thevariation in the labour force behaviour of marriedwomen remains to be explained.
Let us now proceed to rather finer breakdownsof this heterogeneous group, first standardising forage (in the rest of Table 1) and then for age andresidence simultaneously (in Table 3).
Analysis of ,,he Labour Force Participation ofMarried Women by Age Groups
First, let us look at the remaining columns inTable 1 where the impact of each of the variables isestimated for each age group separately. By focus-sing on one age group at a time, and then comparingthe five groups one with another, we can learn muchof the effects of the various attributes considered,particularly of the extent to which they have a dif-ferential impact from one age group to another.
12
15
In this process it is essential to have somemeans of comparing the impacts of the variousfactorssome basis for the statement that one factoris more important than another. The principle whichwe adopt here is to calculate, for each of the factors,the discrepancy between the classification provid-ing the strongest incentive to enter the labour forceand the one providing the strongest deterrent. Forexample, consider the impact of the education-of-wife variable on the labour force participation of theyoungest group of wives. A wife with an elementary
level of education has a "deterrent effect" of 10.86per cent to stay out of the labour force, whereas awife with university education has an "incentiveeffect" of 11.93 per cent. The difference (10.86 +11.93) is 22.79 per cent. For each of the otherfactors a similar calculation is made and the factorsranked in order of their relative impact on the labourforce participation rates of married women; the cal-culations are reported in Table 2.32 In the discus-sion in the text we round to the nearest decimalpoint.
TABLE 2. Ranking of Labour Force Participation Incentives
Variableor
attribute
Age group
15-24 25-84 35-44 45-54 55-64
Education of wifeEducation of husbandChild statusResidence
Education of wifeEducation of husbandChild statusResidence
Percentage gap between strongest incentive and disincentive
22.794.85
32.9214.83
16.4610.0637.238.58
19.2111.8924.266.87
22,9910.537.612.45
16.096.641.402.06
Rank ordering of the above factors
241
3
2314
231
1
234
1
243
Age Group 15 -24
Looking first at the youngest group, we see fromTable 1 that their average participation rate is rela-tively high (at 31.50 per cent). From Table 2 theimpact of the child status variable appears greatest,with the discrepancy between no children and thepresence of one or more children younger than sixyears at almost 33 per cent.
The variable having the, second greatest impactis that of the wife's education. A wife with only"elementary" schOol is 22.8 per cent less likely tobe in the labour force than a woman with "universi-ty" education. As suggested before, this reflectsboth the pull of the greater earning power of thewell-educated women (or, alternatively, the incomeforegone by staying at home) and the alleged desirefor a greater sense of fulfillment which she mightachieve by going to work.
Her residence is the third most important factorin determining whether a wife in this age group willbe in the labour force. A young married woman livingin a larger urban area is almost 15 per cent morelikely to be in the labour force than her rural farmcounterpart.
if3
Apparently the least important factor in deter-mining her labour force status is her husband's edu-cation. The data support the statement that thebetter educated the husband, the less likely his wifeis to work. Presumably the reason is the relativelyhigh income he earns. However, for this age group awoman whose husband has elementary education isonly 4.9 per cent more likely to be in the labourforce than a woman whose husband has universityeducation.
Age Group 25 -34
Turning now to the second age group, we againfind child status as the most important single factorin determining the labour force status of the married
32 We are aware that this criterion is influenced byour choice of classification for each factor. Thus itmight be possible to make the education of the wifeappear to be the most important factor simply by extendingthe classification to include wives with say only 4 yearsof schooling on the one hand, and those with post-graduate training on the other. However, with theexception of age, each of the factors were divided intofour broad groups, and since no attempt was made toidentify the most "extreme ", groups the criterion whichwe have adopted probably gives a good guide to therelative importance of the factors.
16
woman. Married women in this age group n.re verylikely to have young children at home, and for themthe presence of younger children again proves to bea very strong deterrent ;o labour force participation,so strong, in fact, that a woman with no children is37.7 per cent more likely to be in the labour forcethan a similar woman who has one or more youngchildren.
For this second age group the presence of anolder group of children (those aged 6-14) also be-comes a consideration. But for women whose onlychildren are in the 6-14 age group, and require sub-stantially less continuing attention than youngerchildren, there is iess of a deterrent to go to work.
The wife's education is the second most impor-tant factor:a wife with elementary schooling is 16.5per cent less likely to be in the labour force than awoman who went to university. The same argumentsmentioned above favouring the labour force participa-tion of the relatively well-educated wife apply tothis age group and to the older age groups.
The husband's education plays a relatively moreimportant role for wives in this age group than foryounger wives. The 25-34 year old. wife whosehusband has at least some university education is10.1 per cent more likely to be in the labour forcethan a woman whose husband has only elementaryeducation, this compared to 4.9 per cent for theyounger group.
The husbands of women in any age group are,on the average, about the same age as their wives.In many cases, and especially amongst the bettereducated, the young husband's current income willtypically fall far short of his expected income overthe next few decades. His wife, therefore, wouldhave an additional incentive to go to work in orderto even out the family's stream of, income receipts,allowing them to attain a current level of consump-tion more closely approximating their longer run ex-pectations than would otherwise be possible."
Residence ranks as the least important of thefactors considered in determining the labour forcestatus of wives in this age group, but not far behindher husband's education. A metropolitan wife is 8.6per cent more likely to be in the labour force thanher rural farm sister.
Age Group 35 -44
The overall participation rate of married womenis highest in this age group, at 32.4 per cent. Childstatus remains the most important single factor af-fecting a wife's labour force status, in that a wifewith no children is 24.3 per cent more likely to be inthe labour force than a wife who has children in the
33 The explicit suggestion that the "permanent"level of such income variables has an important impacton the labour force participation of the wife, distinguis-able from the effect of "transitory" levels, is made byJacob Mincer, "Labour Force Participation of MarriedWomen: A Study of Labor Supply", op. cit., p. 69.
"other" category (which is dominated by the pre-sence of children in both the less than six and the6-14 age brackets). The importance of child statushas, however, fallen appreciably in its influencevis-a-vis the effect it had for the two previouslyconsidered groups of younger women. As one wouldsuspect, the relative and absolute importance of thechild status variable continues to decline as weconsider still older groups of wives.
Education of the wife remains the second mostimportant determinant, but its role has increased forthe 35 -44 age group vis-a-vis the 25-34 group. Awife with university education is now 19.2 per centmore likely to be in the labour force than a womanwith only elementary education, this compared to16.5 per cent for wives aged 25-34. The increasedrole attributable to her own education reflects thetypical reduction in her child care duties, allowingher the freedom to enter the labour force.
The husband's education has also taken on an. nhanced role: a wife whose husband has elementaryschooling is both more likely to be in the labourforce than her younger sisters, and less likely if herhusband has university education. The, increasedrole attributable to the husband's education (andhence to his current income and income prospects)reflects primarily the diminished role of the childstatus variable. Once the wife is largely freed fromduties of raising young children she can give greaterweight to other factors, of which her own potentialearnings foregone (as indicated by her own level ofeducation) is one, and her husband's income (as ih-dicated by his level of education) is another.
The residence variable continues its decline inabsolute importance; a metropolitan wife in this agegroup is only 6.9 per cent more likely to be in thelabour force than a rural farm wife. For the 25-34group the figure is 8.6 per cent, and for the 15-24group, 14.8 per cent. The reduced importance ofthese residence and child status factors reflect thechange in the family's position in its life cycle, andalso the increased weight of consideration given tothe education (income) variables.
Age Group 35 -44
For the first time the child status variable nolonger dominates the labour force participation of amarried woman. The most important single, factor forthis age group is the wife's education: a universityeducated wife between the ages of 45 and 54 isalmost 23 per cent more likely to be in the labourforce than a wife with only elementary schooling.Such a woman is also relatively less likely to be inthe labour force than her younger sisters if she hashad little schooling, and more likely if she has atleast some university education.
The education of the husband increases in itsrelative importance to second place. A wife whosehusband has elementary schooling is 10.5 per centmore likely to be in the labour force than one whosehusband has university education, whereas for her35-44 year old sister the probability is 11.9 percent.
14
..41",,,RTFON
17
Child status is the third-ranking variable forthe 45 -54 age group. Clearly, for all but a few, thepresence of very young children is no longer a con-sideration. Hence the distinction between "no chil-dren" and "some" children has become more nomi-nal than real.
The residence variable again ranks last, havingdiminished substantially in its absolute importance:a metropolitan wife in this age group is only 2.5 percent more likely to be in the labour force than herrural farm counterpart.
Age Group 55 -64
In this, the oldest group of married women con-sidered in the present study, the labour force parti-cipation rate at 18.67 is markedly lower than in anyother age group. By the ti'lne a wife has reached thisage her family's needs have normally been reduced.Little further expenditure is necessary to educatechildren, and generally her husband's income willhave passed its phase of rapid growth. His incomemay rise in the few years remaining before retire-ment, but probably not by much. The wife, therefore,has lost the incentive of her younger counterparts tofill in the gap between her husband's current andexpected income. That motive may now be, of course,replaced by tl)e incentive to build up an adequateretirement fund, although a woman in this age rangeis less likely, than is a younger woman, to find sowide a range of attractive job opportunities open toher.
Both residence and child status have virtuallydisappeared as determinants of the wife's labourforce behaviour. By far the most important factorremaining is her own level of education (and hencepotential income from labour force participation): a
otmAn.rm
woman with university education is 16.1 per centmore likely to be in the labour force than a womanwith only elementary schooling. Thus even the im-portance of this variable has declined; the compara-ble figure for the 45-54 year bracket is almost 23per cent.
Her husband's education remains an importantexplanation of a wife's labour force behaviour, butits role has diminished. The wife of a husband whohas elementary schooling is 6.6 per cent more likelyto be in the labour force than one whose husbandhas university education. The comparable figure forthe 45-54 group is 10.5 per cent. It is evident,therefore, that n marked reduction in the importanceof this factor has occurred.
Summary
Recapitulating, Table 1 provides us with thebasis for an analysis of the labour force participa-tion rates for married women between the ages of 15and 64, first of all with age as a variable and thencontrolling for each of five age groups separately.From the table we have been able to considervarious "typical" wives and to suggest the causesof the divergent labour force participation behavioursfor women in the various age groups.
Starting with the group of all women together,we found that we could explain about 72 per cent ofthe variation in their labour force behaviour. Inlooking at each of the five age groups separately wefound that we could explain between 79 per cent and87 per cent of the variation. By looking at morenarrowly defined subdivisions of the group of majorinterest we were able not only to increase thedegree of explanation achieved, but also to learnmuch more about the changing roles of the variousimportant attributes which we have discussed.
ANALYSIS OF THE WIFE'S LABOUR FORCE STATUS:THE FOCUS ON AGE AND RESIDENCE DIFFERENTIALS
In an attempt to carry the study one step further,it was decided to look at still finer subdivisions ofmarried women by analysing the determinants of thelabour force status of wives in each of the residencecategories by age groups. The result is presented inTable 3, in which each age group of married womenis subdivided by the residence classification withthe results of a multiple regression analysis reportedfor each of the subdivisions. The ranking of thefactors in order of their importance for each of theage and residence groups is reported in Table 4.The interpretation and analysis of these resultsproceeds on the same basis as for Tables 1 and 2.
Overall Measure of Goodness of FitLooking first at the row reporting R2, the coef-
ficient of multiple determination, one can see thatthe proportion of the total variation in the labourforce behaviour which can now be explained has
risen substantially. For the youngest group, forexample, while the measure on the basis of theentire group taken together was .85, the R2 measuresfor three of the four subdivisions are higher than .85,and the fourth is .80. A similar statement holds foreach of the other finer breakdowns.
Analysis by Age Groups
Age Group 15 -24
The breakdown by the residence variable of theyoungest group of wives in our study leads to a con-firmation of what one would probably have expected.
First, the mean participation rates (the commonterms) vary from a high of 40.07 for metropolitanwives to 25.24 for rural farm wives. The generallygreater availability of employment opportunities forwomen in larger centres would account for much of
r-II
- 18 -
the difference. And perhaps differences in socialattitudes would also have some impact cf it isgenerally better accepted in large centres than insmall that married women should work.
For each of the residence groups the childstatus variable is of great importance. But itsrelative importance is clearly greatest for the metro-politan wife. For the rural non-farm and rural farmwives, and even for other urban wives, the educationof the wife is of approximately equal importance indetermining a wife's labour force status. Why thechild status variable should be of diminishing im-portance as one moves from larger to smaller centresis open to conjecture. Perhaps it is easier andcheaper in smaller centres to arrange for child careand for other domestic help since the travelling timeis usually les:, and wages rates are generally lower.
The education of wife variable is an increasing-ly important factor the smaller the centre. Perhapsthis reflects a relative scarcity of well educated
women in smaller centres, such that those with arelatively advanced education, experience a muchstronger pull into the labour force than would acomparable metropolitan wife. It is not unusual fora well educated woman living in a smaller centre tobe actively sought out and offered employment, apractice less common in larger centres.
The education of husband variable, which isincluded as a proxy for his income, is consistentlyfound to be the least important determinant of theyoung wife's labour force status. However, it isagain interesting that this factor plays an increas-ingly important role as one moves from larger tosmaller centres. That is, a relatively high incomeearned by the husband is a stronger deterrent tolabour force activity on the part of his wife in asmaller centre than in a large one, an outcome re-flecting the differences in social attitudes in centresof varying sizes, and perhaps also the lower costsand more limited range of social activities avail-able.
TABLE 3. Analysis of the Labour Force Participation Rates of Married Women, Canada, 1961(Standardized by residence variable)
Variableor
attribute
Age group
15 - 24 25 -34 35 -44
Metro Otherurban
Rural Ruralnon-farm farm Metro Other
urbanRural
non-farmRuralfarm Metro Other
urbanRural
non-farmRuralfarm
Education of wife:Elementary1 -3 years secondary4 -5 years seconaaryUniversity
Education of husband:Elementary1-3 years secondary4 -5 years secondaryUniversity
Child status:No childrenChildren less than 6Children 6-14Other
Common term 30.82 29.84 30.53 28.17 17.82 17.80 20.58 18.50R2 .92 91 .38 .91 .88 .85 .83 .83
Category excluded because too few observations available.
19
Age Group 25 -34
For the second youngest group of married womenthe enhanced explanatory power achieved by workingwith each residence group separately is againevident from the increased value of R2.
Child status remains the most important singlefactor influencing the labour force status of thisgroup of women. Again its strength appears greatestin the larger centres, a conclusion surprising tothose who think that there are better child carefacilities available in larger centres. In fact, thescarcity of child care centres is a problem in mostplaces, large or small." Perhaps the presence ofyoung children in a smaller centre is less of alabour force deterrent, first, because the costs ofhiring a sitter are usually less (distances are short-er, and wage rates typically lower) and, second,because the family structure in small centres isfrequently such that young children can be left inthe care of relatives.
The wife's education remains the second mostimportant determinant of a wife's labour force status,and again is a more potent factor in small communi-ties than in larger ones (excepting the rural farmcategory). While this outcome may appear surprising,one could hypothesize that it reflects both the rela-tively high level of demand in smaller centres forwell educated women and the relatively low level ofdemand for the poorly educated.
The husband's education (income) still ranksthird in the order of importance, but is much closerto a second-place position than it was for the young-er wives. This circumstance may result more fromthe reduced importance of the role of her educationrather than an enhanced role attributable directly toher husband's position.
Age Group 35 -44
This group ranks in the centre position of thefive we have defined; it also has the highest parti-cipation rate of all, the second highest being for theyoungest group. Intervening is a period of high fer-tility rates for married women at which time manyleave the labour force either permanently or tempo-rarily.
Here again the subdivision by type of residenceis most informative, and the proportion of the varia-tion in labour force behaviour which can now beexplained has risen in each of the cases.
As we noted with the previous cases, theaverage participation rate trends to be lower thesmaller the centre.
For this middle group the child status variablecontinues to be an important one. And again itsrelatively greater importance in urban centres vis-a-vis smaller centres is striking.
34 "Day Care Services for Children of Working-Mothers", op. cit.
1. 1
'7+
While the child status variable continues todominate for metropolitan and (to a lesser extent)other urban wives, it is of no greater importance thanthe education of the wife in rural farm communitiesand is actually displaced by the education of thewife variable in rural non-farm centres. The educa-tion of husbi.nd variable has gained in strength, ascompared to its importance for younger women.
In other words there has been a substrdittalevening out of the relative importance of the factorswhich a wife apparently weighs regarding her labourforce status. The child status variable has lost theclear dominance which it had for younger wives, andboth the education of the husband and of the wifehave risen in importance.
Age Group 45 -54
The overall participation rate for married womenin this age group remains high for all residence cat-egories. There is, in general, a tendency for womenin the 45-54 age bracket to be somewhat less activelabour force participants than the 35-44 year olds,and the difference is most marked amongst metro-politan wives.
For women in all residence groups the educationof wife variable has become clearly dominant, withits relative weight being greatest in the smallercentres. Education of the husband has become thevariable ranking second in importance in determiningthe wife's labour force status. The enhanced roleattributable to this variable reflects primarily thegreatly diminished role of the child status variablefor women in this group. The education (and henceincome) of the husband has therefore become a muchmore important matter for consideration: there isless to keep the wife out of the labour force, whilethe forces tending to place her in the labour forcemay have gained in strength.
For all residence groups child status ranksthird amongst the factors considered. It has becomea quite unimportant consideration for wives in thesmaller centres, while it retains substantial, thoughgreatly diminished, significance for wives in largercentres. For all age groups we have found the childstatus variable to be a much more important consid-eration for wives in larger centres than in small.For this age group of women, for whom the presenceof children younger than six is uncommon, one couldnot argue that the relative importance of the childstatus variable in the larger centres results from thelesser availability of child care services. But it mayreflect the phenomenon that women working in thelarger centres usually find it necessary to be awayfrom the home a larger portion of the day (e.g., be-cause of greater travelling time), which would leavethem less time to be home caring for their (older)families.
Age Group 55 -64
At 18.67 per cent the mean participation rate forthe oldest group of married women in the currentstudy is also the lowest. The sharp reduction in
- 20 -
labour force participation occurs in all residenceareas, with the largest centres showing the greatestdecline.
For each residence group the rank ordering ofthe three factors is the same as for the 45-54 agegroup: education of wife, education of husband, andchild status. In line with the generally lower parti-cipation rates is the reduced importance of each ofthe factors separately. Child status has virtuallydisappeared as a consideration, and the educationof the wife continues to be the dominant factordetermining her labour force status.
For most families in which the wife is in thisage group, major family expenditures have declined
and many probably ceased. By and large, childrenare no longer financially dependent; most majorassets have been acquired and perhaps fully ...Idfor. In other words, many of the forces which causedthe wife to be in the labour force have now subsided,and she remains in the labour force more out ofpersonal preference than was the case earlier in hercareer. This is reflected in the increasingly impor-tant role which our study attributes to the educationof the wife. At all ages a better educated person isable to demand the more desirable forms of employ-ment; with the need for continued income reduced,the older wife is able to give increasing weight toher own judgement of how desirable she finds heremployment.
TABLE 4. Ranking of Labour Force Participation Incentives(Standardized by residence variable)
Variable orattribute
Age group
15-24 25-34 35-44
Metro Otherurban
Ruralnon-farm
Ruralfarm Metro Other
urbanRural
non-farmRuralfarm Metro Other
urbanRural
non-farmRuralfarm
Education of wifeEducation of husbandChild status
Education of wifeEducation of husbandChild status
Education of wifeEducation of husbandChild status
Education of wifeEducation of husbandChild status
Percentage gap between strongest incentive and disincentive
14.442.11
45.08
25.276.53
38.14
32.946.45
28.74
18.508.30
19.74
11.0910.4840.68
19.3312.2541.71
25.419.60
33.14
12.918.374.04
14.6714.1432.96
20.6815.6926.98
17.164.10
16.60
25.5013.6020.49
Rank o dering of the above factors
231
Metro
2 1
3 31
Otherurban
231
231
231
231
231
231
231
1
32
1
32
-54 55 -64
Ruralnon-farm
Ruralfarm Metro Other
urbanRural
non-farmRuralfarm
Percentage gap between strongest incentive and disincentive
17.7413.0312.34
24.0414.269.09
29.2111.084.62
23.405.764.67
12.687.823.30
16.357.952.08
22.166.301.34
15.624.461.54
Rank ordering of the above factors
1
23
1
23
1
23
1
23
1
23
1
23
1
23
1
23
SUMMARY AND CONCLUSIONS
In the preceding pages we have looked closelyat the determinants of the labour force participationrates of wives of various age and residence group-ings, attempting to distinguish the influence ofvarious important factors. Can we now draw someconclusions, and suggest the probable future courseof married women participating in the labour force?
First, by way of summary, we have obtained agood quantitative impression of how various impor-tant influences affect the labour force participationof the wife at various stages in her life cycle, andin her family cycle. We have been able to distinguishthese separate impacts for wives living in centresvarying in size, and to compare one with another.
21
We have seen that the presence or absence of ayoung child remains, in the aggregate, the most im-portant single attribute affecting a wife's participa-tion in the labour force until age 44. However, theinfluence of this factor is uniformly stronger in thelarge centres than it is in the small, and actuallydrops to second place for wives in the rural non-farmand rural farm areas who are in the 35-44 age group.We have suggested that the greater relative impor-tance of the child status variable in larger centresvis-a-vis the smaller ones reflects the greater avail-ability of inexpensive child care in the smallercentres resulting in part from the shorter distancesand greater ease in local travel. As a result thepresence of a young child provides a substantiallysmaller deterrent to the labour force participation ofwives in the smaller centres.
As we look at older groups of women the pres-ence of a young child becomes increasingly rare.For wives in both the 45-54 and the 55-64 agegroups the child status factor is the least importantfactor of those considered. However, as for youngerwomen, it remains more important for "urban" thanfor "rural" wives.
The education of the wife is a very importantvariable in explaining her labour force activity.Other things equal a better educated woman is muchmore likely to be in the labour force than one whois less well educated. For all women the educationfactor increases in relative importance with age,reflecting primarily the diminished importance of ti.echild status variable. Again there appears to besome distinction by the size of the centre in whichthe wife lives. By and large, for wives in the smallercentres, education plays a more important role thanit does for their city sisters.
The education of the husband variable, whichwe use as a proxy for both his current income andhis anticipated future income positions, is of rela-tively little importance for younger wives, but gainsin relative significance for the older wives. Foryounger wives the child status factor is of suchdominating importance as to obscure somewhat theeffects of the others. No doubt the husband's incomeis always an important consideration for a wife'slabour force participation, but it can play an impor-tant role only if the wife is free of household duties,especially the care of young children. It is only forsomewhat older wives (the 45-54 and 55-64 groups)
1
that her husband's education (and hence income)becomes an important enough consideration to dis-place the (declining) influence of child status.
Considering the influence of all these factorstogether what can one predict of the future course ofthe labour force participation of married women?Forecasting is notoriously difficult and inaccurate,and we shall not attempt more than to indicatetendencies.
1. The general improvement in the average level ofeducation is likely to induce more and more wivesto remain in the labour force after they are mar-ried, and to return to it when their major child-raising responsibilities in the home are passed.This factor alone would suggest, therefore, acontinuing increase in the labour force participa-tion rate of married women in most age groups.Even for women in the most fertile age group (25-34), an increased availability of day nurseriesand other substitutes for the mother's constantattention would permit increased participationrates in the future.
2. Any change in fertility patterns, brought about,for example, by the wider use of highly effectivemeans of birth control, could be expected to havea pronounced impact on the labour force partici-pation rates of married women. If couples decideto wait a few years after marriage before havingchildren the wife is very likely to be in the labourforce during that period. And there is evidencethat such family planning has become a wide-spread phenomenon."
3. The general continuing rise in the husband's in-come will continue to act as a deterrent to thewife's labour force participation. It may be,however, that it will not become much more im-portant than the evidence we have discussedabove suggests that it was in 1961. What it maymean is that women with little formal education,who cannot command very high income:1, willincreasingly tend not to be in the labour force.But a continued rise in the average level of edu-cation would counteract such a tendency, and verylikely outweigh it substantially.
35 Oral contraceptives were first available byprescription in 1961. In metropolitan Toronto in 1967,about half the women using any method of contraceptionwere using the pill. See "Oral Contraception and theFertility Decline in Canada, 1958-1968. A First Look ata Crucial Component in the Argument", by Kantner,J.F., Allingham, J.D., and Balakrishnan, T.R., a paperpresented to the Population Association of America,Boston, April 17, 1968.
