Peter Baskerville, Eric W. Sager. Unwilling Idlers: The Urban Unemployed and Their Families in Late Victorian Canada. Toronto: University of Toronto Press, 1998. 294 pp. $24.95 CDN (paper), ISBN 978-0-8020-8144-5.
Reviewed by Lisa Y. Dillon (Institute of Canadian Studies, University of Ottawa)
Published on H-Urban (March, 1999)
Identifying the Unemployed
Unwilling Idlers is an innovative and comprehensive social and intellectual history of unemployed city-dwellers and their families in turn-of-the-century Canada. Authors Peter Baskerville and Eric Sager both work at the Department of History, University of Victoria, and are engaged in an ongoing collaboration, the Canadian Families Project, which uses the 1901 Canadian census. Based on sampled census data from six cities in 1891 and 1901 Canada, Victoria, Vancouver, Winnipeg, Hamilton, Montreal and Halifax, as well as a selection of surveys, contemporary interviews, reports, journal articles, policy statements and petitions, this book exemplifies how careful and imaginative use of qualitative and quantitative sources allows historians to compare rhetoric and reality. Unwilling Idlers is a successful study. Its main contribution to the historical literature, a precise explanation of the extent and reasons for unemployment at the turn of the century, is a result of judicious use of census microdata to establish the identity, living standards and opportunities of unemployed urban Canadians and their families.
Baskerville and Sager explore unemployment both as a concept and a condition, situating their work in a historiography which includes Alexander Keyssar's Out of Work: The First Century of Unemployment in Massachusetts (1986)[1], Daniel Hiebert's historical geographies of Toronto and Winnipeg[2], David and Rosemary Gagan's study of Victorian Ontario working-class living standards[3], Bettina Bradbury's Working Families: Age, Gender and Daily Survival in Industrializing Montreal (1993)[4], and Gordon Darroch and Lee Soltow's Property and Inequality in Victorian Ontario (1994).[5] The first substantive chapter sets the stage for the book by showing how late nineteenth-century craft unions and politicians only gradually began to understand unemployment as a structural problem resulting from market inequities rather than mere seasonal cycles. In Chapters Two through Four, the authors provide a detailed profile of the urban unemployed, discuss the impact of seasonality, technological change, and factory labour structures on unemployment, and explore the spatial distribution of jobless workers in city sub-districts. They agree with Keyssar that unemployment was a widespread phenomenon, affecting "women and men, native-born and immigrants, young and old," yet they disagree with his "lottery hypothesis" that unemployment was random, instead arguing that occupation, skill, social class and location influenced a person's chances of being unemployed (pp. 42, 51, 53, 68, 76, 78). They determined that more than one in three workers experienced joblessness, and that low average wages played a more important role than unemployment in reducing household income. Subsequent chapters detail the relative contributions of family members to household economies and the impact of joblessness on families, measure family living standards based on household income, and outline alternative family survival strategies. The authors assert that the capitalist labour market was unable to provide employment and incomes sufficient for one in seven urban working-class families to survive, and that these families had few alternatives to supplement their income beyond their own informal labour. In their final chapter, Baskerville and Sager return to the question of unemployment as a concept, exploring how trade unions and labour reformers identified unemployment as a structural condition and called for state intervention. They also survey more radical analyses by contemporary intellectuals which focused on deskilling and recommended the nationalization of production and distribution. Such arguments were opposed by persistent critiques that joblessness was rooted in moral deficiency and suggestions that Canada's large and undeveloped frontier offered unlimited opportunities to unemployed workers.
By combining census evidence with qualitative sources, Baskerville and Sager contextualize their findings in the light of turn-of-the-century rhetoric and reform surrounding unemployment. Though this is an urban history, in the first and final chapters, the authors deal with Canada's rural context: they dispute contemporary optimism that newly available western land could resolve urban unemployment by questioning whether unemployed workers could afford to move. It would have been difficult for the authors to draw more direct links between the rhetoric and the reality of unemployment. For the most part, Baskerville and Sager use their census-based investigations to establish how far workers' experiences with unemployment departed from government and business perceptions. The book is perhaps at its best when the authors can demonstrate more direct links between the two, for example when they show that certain workers' chances of being unemployed actually rose with higher wages, thus suggesting that workers in skilled occupations with frequent lay-offs commanded a higher wage through bargaining with employers (p. 147). In their introduction, Baskerville and Sager criticize the unsystematic character of some cultural studies of unemployment and take pains throughout the book to explain their choice of quantitative sources and methodology, as well as their choice of cities. Similar effort should have been applied to their qualitative research, explaining why they chiefly drew upon newspapers, labour journals and government reports which originated in Ontario, Quebec or in Ottawa.
Historians interested in urban-dwellers will find particularly interesting Baskerville and Sager's comparisons of results across cities, including their analyses of the distribution of workers and ethnic groups across city sub-districts. Their most important contribution to urban history is their finding that the risks of unemployment were quite similar from city to city, and that most city differences in unemployment levels were actually rooted in "the distribution of occupations in each city." (pp. 62-63, 137, 141) Baskerville and Sager do carefully delineate when important differences among cities did exist, such as the heightened association between religion and chances of being unemployed among Montreal and Winnipeg residents. The author's discovery that about one in seven families earned less than what was required to survive, not only in Montreal, but also in Vancouver, Victoria, Winnipeg, Hamilton, Montreal and Halifax, is a major contribution to women's history: Baskerville and Sager emphasize that the informal labour of housewives that Bettina Bradbury identified in Montreal must have played an important role in family survival across the country.
The preface to Unwilling Idlers notes the manuscript reviewers' reservations about the authors' choice of census microdata as their main primary source. Fortunately for the historical community, the authors and the University of Toronto Press prevailed: it is through their in-depth analysis of turn-of-the-century census microdata that Baskerville and Sager have marshalled convincing evidence that unemployment was a common yet not random experience, that unemployment patterns were similar from city to city, and that the impact of unemployment on individuals was closely related to family life cycles. Both the 1891 and 1901 Canadian censuses asked questions about work and unemployment, the 1901 census asking fourteen separate questions on labour-related issues. No other single source on turn-of-the-century Canadians offers similarly broad and detailed information on their work and unemployment patterns.
The specific construction of the microdata used in this monograph ensures the representativeness of Baskerville and Sager's findings. The authors use three different microdata files, two of which are 10% samples by dwelling place. The sampling density of 10% was undoubtedly chosen to yield the large numbers needed to sustain analysis of separate cities and urban sub-groups. Nevertheless, these samples are far denser than typical of census microdata samples, including the historical U.S. public use microdata samples, which are 1% samples, and the 1901 Canadian census database, which is a 5% sample. Moreover, the third microdata file used by Baskerville and Sager is a 100% sample of all unemployed persons and other members of their dwelling in the six cities in 1891. Between their 10% samples of all persons in six cities in 1891 and 1901 and this 100% sample of the unemployed and their dwelling members in 1891, Baskerville and Sager have constructed data which represents the common experiences of Canadian urban employed and unemployed persons far more than any other primary source available at the time of writing. Including in their analysis the experiences of underenumerated persons, namely transients, the very poor and women, would only have increased their estimates of the extent of unemployment, as they discuss in Chapter Three.
The quantitative analysis on which five of the seven substantive chapters are based is presented with a clarity which makes this work accessible to "quals" and "quants" alike. The authors make shrewd choices about the degree of statistical sophistication they apply to different questions. For the most part, they use simple frequencies, turning to logistic regression analysis when exploring the relative weight of variables such as age, city, sex, occupation, household type, marital status, relationship to head of household and birthplace on chances of being unemployed. By using logistic regression analysis, Baskerville and Sager find evidence which contradicts Keyssar's argument that unemployment was a random phenomenon. Baskerville and Sager view the census as a constructed document, the product of government agendas, enumerator interpretations, as well as responses formulated by enumerated persons themselves. They explore the intentions with which this source was created, and test responserates to various questions. The high response rates to questions about work from workers, including women who reported an occupation to begin with, help to justify this choice of source.
There are a few aspects of Baskerville and Sager's organization of statistical evidence which could have been made clearer. Their logistic regressions in Chapter Three include both class of worker and occupation categories as independent variables. In these instances, Baskerville and Sager have organized occupations into industrial sectors, but do not elaborate their industrial coding scheme to the reader, either in a footnote or in their Appendix A: The Census as a Historical Source. The authors included in this regression persons "without an occupation who declared themselves unemployed." (p. 57); it is unclear where they classified such persons in the occupation variable, which does not include a value for "No Occupation Reported." Baskerville and Sager note problems with the coefficients displayed for the variable marital status in the 1891 regression, which excluded single young persons who had not reported an occupation yet may very well have been seeking work. I can think of another potential problem with the inclusion of marital status in this regression: could there have been a risk of endogeneity between being unemployed and being married? Since many men traditionally did not marry until they had established themselves in work, perhaps some men were married because they were not unemployed, rather than the other way around. In the 1891 regression, the value for "wife" does not appear with the independent variable, "Relationship to head," presumably because Baskerville and Sager have identified housewives as not at risk of unemployment for they already labour within the home. In the 1901 regression, however, wives do appear: is this because the 1901 regression is based on all persons who reported themselves to be employees in 1901, a group which now included some wives?
I have a question regarding the denominators chosen for Baskerville and Sager's regressions. In general, Baskerville and Sager are very careful to define who the census viewed as unemployed and which unemployed persons they choose to enter in the denominator in their measures. In Chapter Three, they note that the 1891 and 1901 censuses had not identified as "unemployed" persons who did not list an occupation to begin with. They reasoned that many such persons were also in search of work, including women who were not housewives and men who did not identify with a particular occupation. They include this potential unemployed labour pool in a simple frequency estimation of the true extent of unemployment (p. 50). What Baskerville and Sager really do in this estimation, one of the most important figures in the whole book, is show the ratio between two unemployment phenomena: the inability to get into the labour force to begin with, and the inability to stay there. The first unemployment experience largely concerns women and young persons, while the second, as they demonstrate subsequently, largely concerns low-skilled workers in vulnerable occupations. One could make reasonable arguments to analyze these unemployment phenomena separately or together. In their subsequent regression analyses for 1891 and 1901, Baskerville and Sager choose to focus on the second group. Their first regression, which used 1891 data, included "only people between the ages of fifteen and sixty-five who stated an occupation, as well as those without an occupation who declared themselves unemployed." (p. 57) Presumably Baskerville and Sager excluded other potentially unemployed persons from the 1891 regression to increase the comparability between the 1891 and the 1901 regression. Since the dependent variable in the 1901 regression, "unemployed or not," is based on a census question posed only to employees, it would have made no sense to include non-employees in this analysis. Through these regressions, Baskerville and Sager are able to establish which factors played the greatest role in preventing some employees from staying in the labour force. Their choice of denominator was both intellectually appropriate and methodologically necessary.
The only downside of this choice is that it prevents us from determining the relative weight of those factors which prevented some individuals from entering the labour force to begin with and others from staying in it. In other words, we still do not know the relative impact of being a woman or a young, single person versus being a general labourer on not being able to work. I believe such analysis would have been possible in an additional regression using the 1891 data. The results might then have been related more closely to a parallel historical discourse the authors describe in Chapter Two, the patriarchal language workers chose to defend their interests as breadwinners.
One final point concerns Baskerville and Sager's treatment of the relationships among family cycles, unemployment and household income. In Chapter Six, "Family, Work, and Income in 1901," the authors present the average income per family member in each family cycle stage, before children appear, when all children are less than 10 years, when some children are aged 10-14, when some children are over 14 years, and when the wife is aged 45 years or more and no children are present. By counting children under 10 as half-persons, Baskerville and Sager usefully build into this measure the fact that older children required more funds to support. Per-person income level dips with the appearance of children and rises again with teenaged children's entrance into the labour force. On the one hand, Baskerville and Sager emphasize the importance to the family economy of teenaged children's wage contributions, yet they simultaneously note that youths' wages and employment time could never match those of adult men. As a result, they point out, per-person incomes in family with teenaged children were only $187, far short of the average per-person income of $310 evident before children appeared. U.S. authors Brian Gratton and Frances Rotondo,[6] and Canadian historian Edgar-Andre Montigny[7] contend that the main direction of support in turn-of-the-century families with grown co-resident children went from older parents to their children, and emphasize on this basis the higher status and authority of older people. Elsewhere in Unwilling Idlers, Baskerville and Sager demonstrate the limited earnings of older Canadians and thus indicate their reservations about the status of the elderly. In their examination of family member contributions to household income, however, Baskerville and Sager do not take a direct stand on the issue of direction of support between generations. Perhaps this is because of the contradictions inherent in stage 4 families: older children made important contributions to household income yet could not compensate for the "extra mouths" they represented.
The level of per-person incomes in the last stage of the family life cycle, however, suggests that the authors have the evidence to confront this issue more directly. When older parents no longer lived with their children, their per-person income rises to $227, the highest level since their children first appeared (p. 118). This income level is largely a result of the diminished number of household members. Baskerville and Sager pay no attention to this finding in their discussion of stage 5 families, instead emphasizing the instability wrought by the absence of co-resident children. The authors may have chosen not to discuss this statistic because their data furnished only 224 families in the fifth family cycle stage. However, the higher per-person income level characteristic of stage 5 raises an important question which goes beyond the number of families who actually experienced this family stage: if older parents enjoyed higher per-person income levels with children gone from their home, why didn't more older parents choose this family pattern? Did parents prefer to exist on the lessened income of their teenaged children? Although teenaged children who contributed earnings to their parents' home made sacrifices, as the authors point out on page 145, were they still better off at home than on their own? Much of the reason for the small number of families in stage 5 is probably demographic: the high fertility rates of older families in 1901 reduced the number of older parents who saw their youngest child reach an age at which they could feasibly leave home. Nevertheless, both the higher per-person income levels and the small number of families characteristic of stage 5 raises interesting questions about the direction of support between co-resident generations. It may have been more illuminating for Baskerville and Sager to compare the implications of per-person income levels in stage 4 families to stage 5 families, than to the first family stage. Although per-person income levels in the first familystage were, at $310, the highest ever experienced, this stage was a very brief period subject to even less demographic choice than stage 5, and perhaps should not have been used as the benchmark for per-person family earnings. Addressing the issue of inter-generational exchange more directly in this way might have allowed the authors to hypothesize how rural Canadian family dynamics, in which fathers played an important role in adult children's support and life course decisions, may have translated to the urban context.
The suggestions made above that the authors might have devoted additional attention to workers prohibited from entering the labour force as well as the direction of generational support in older families do not at all detract from the strong central arguments of this book. Unwilling Idlers is a carefully written and rewarding social and intellectual history, the quality of the authors' findings hinging on the quality and sophistication of their quantitative analysis. As an exploration of unemployment patterns in six cities across Canada, this monograph contravenes recent criticism of social history for being too particularist. Arguments presented in the book regarding the causes of joblessness contribute insights to the present day dialogue on unemployment. This book will prove useful for historians and other scholars interested not only in unemployment and work patterns, but also for those interested in families, inter-generational relationships, social welfare, politics, and intellectual history.
Notes:
[1]. Alexander Keyssar, Out of Work: The First Century of Unemployment in Massachusetts Cambridge and New York: Cambridge University Press, 1986.
[2]. Daniel Hiebert, "Class, Ethnicity and Residential Structure: The Social Geography of Winnipeg, 1901-1921," Journal of Historical Geography, 17 (1991); and Daniel Hiebert, "The Social Geography of Toronto in 1931: A Study of Residential Differentiation and Social Structure," Journal of Historical Geography, 21 (1995).
[3]. David Gagan and Rosemary Gagan, "Working-Class Standards of Living inLate-Victorian Urban Ontario: A Review of the Miscellaneous Evidence on the Quality of Material Life," Journal of the Canadian Historical Association (Ottawa, 1990).
[4]. Bettina Bradbury, Working Families: Age, Gender, and Daily Survival in Industrializing Montreal. Toronto : McClelland & Stewart, 1993.
[5]. Gordon Darroch and Lee Soltow, Property and Inequality in Victorian Ontario: Structural Patterns and Cultural Communities in the 1871 Census. Toronto: University of Toronto Press, 1994.
[6]. Brian Gratton, "The Poverty of Impoverishment Theory: The Economic Well-Being of the Elderly, 1890-1950," The Journal of Economic History,Vol. 56 No. 1 (March 1996): 39-61; and Brian Gratton & Frances M. Rotondo, "Industrialization, the Family Economy and the Economic Status of the American Elderly," Social Science History, Vol. 15, No. 3 (Fall 1991): 337-362.
[7]. Edgar-Andre Montigny, Foisted Upon the Government?: State Responsibilities, Family Obligations and the Care of the Dependent Aged in Late Nineteenth-Century Ontario. Montreal: McGill-Queen's University Press, 1997.
Copyright (c) 1999 by H-Net, all rights reserved. This work may be copied for non-profit educational use if proper credit is given to the author and the list. For other permission, please contact H-Net@H-Net.Msu.Edu.
If there is additional discussion of this review, you may access it through the network, at: https://networks.h-net.org/h-urban.
Citation:
Lisa Y. Dillon. Review of Baskerville, Peter; Sager, Eric W., Unwilling Idlers: The Urban Unemployed and Their Families in Late Victorian Canada.
H-Urban, H-Net Reviews.
March, 1999.
URL: http://www.h-net.org/reviews/showrev.php?id=2835
Copyright © 1999 by H-Net, all rights reserved. H-Net permits the redistribution and reprinting of this work for nonprofit, educational purposes, with full and accurate attribution to the author, web location, date of publication, originating list, and H-Net: Humanities & Social Sciences Online. For any other proposed use, contact the Reviews editorial staff at hbooks@mail.h-net.org.