Main article

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

  1. Introduction
  2. Literature review
  3. Data and methodology
  4. Results
  5. Conclusion

1   Introduction

The last 30 years have witnessed substantial changes in the characteristics of Canadian families. Common-law unions have become more prevalent, the number of Canadians experiencing a divorce in their lifetime has increased, and blended or mixed families are more numerous than they were in the past. The labour market characteristics of families have also changed, reflecting the widespread entry of women into the paid labour force and their increasing contribution to family incomes. Attitudes regarding the roles of women and men in the labour force and in the home have also changed (Crompton, Brockmann, and Wiggins 2003). These developments have led researchers in a number of disciplines to re-examine the economic behaviours of families—how they make decisions about employment and consumption, how they balance paid and unpaid work, how they organize their income, and so on.

In this paper, data from the 2007 General Social Survey (GSS) are used to examine the income management strategies of older Canadian couples, that is, couples in which at least one spouse or partner is aged 45 or older: 1  the extent to which couples use an allocative strategy of income management, in which one spouse or partner manages the couple's income and allocates a share to the other; a pooled strategy, in which both spouses or partners pool their incomes, with each taking out what he or she needs; or a separate strategy, in which spouses or partners keep their incomes partially or completely separate. The demographic and financial characteristics associated with these strategies are examined, and the extent to which observable characteristics account for the marked differences in the strategies used by married couples and common-law couples are estimated.

Results show that the income management strategies used by couples, specifically the extent to which incomes are pooled or kept separate, are correlated with relationship characteristics, such as common-law status, duration of relationship, and the presence of children. Furthermore, the likelihood of using a separate approach is positively correlated with levels of educational attainment and with the amount of income received by wives or female partners.

The organization of income within households has several implications for public policy. For example, the intra-household distribution of income may have potential implications for the measurement of financial well-being. Measures of low income assume that individuals within families share resources equally and have the same standard of living. Hence, if a family's income is below the low-income threshold, all of its members are identified as living in 'straitened circumstances'. Because evidence on the intra-family distribution of income is not available, it is not possible to calculate low-income rates differently. However, Phipps and Burton (1995) assessed the sensitivity of low-income rates to different assumptions of intra-family sharing. Using 1992 data, they found that fathers and mothers in two-parent families with children had a low-income rate of 8.9% under the assumption of equal sharing and that the rate declined to 2.7% for fathers and increased to 17.5% for mothers under an assumption of 'minimal sharing.' 2  Similarly, Woolley and Marshall (1994) examined the aggregate distribution of household income and reported that an assumption of unequal intra-household sharing 3  yields a Gini coefficient 4  27% higher than is yielded by an assumption of equal sharing. The same point applies to income replacement rates among seniors, as estimates using couples or families as the unit of analysis assume that individuals within those units share the same standard of living, even though the replacement rates achieved by individual family members may differ.

Income management strategies also have implications for the use of tax provisions. For example, couples who view their financial futures as inextricably linked may be more willing than others to contribute their own income into the retirement savings account of their spouse—either inside or outside of spousal RRSP contribution provisions.

As well, the strategies that couples use to organize and manage their incomes has relevance from a legal perspective. For example, Treas (1993) found that American couples in which husbands are employed in 'high liability medical and legal professions' are more likely than others to use separate bank accounts. Given the elevated risks of such individuals being sued, she argued, couples use separate accounts as a potential defence against litigants. Treas also observed that couples' use of separate bank accounts is lower in states with community property laws (1993, 731) and argued that this is consistent with expectations since money in separate accounts is not treated as personal property under such legislation. Looking at a small sample of couples in which one or both partners had been married previously, Burgoyne and Morison (1997) found that older couples with children from a previous marriage were more likely than others to separate their incomes and that this was "…especially marked in the way they wished their assets to be treated after their death" (p. 363). This suggests that, for older individuals in complex family formations, estate planning and the laws governing it are factors influencing the organization of income.

Phipps and Burton (1995; 1996) provide a broad discussion of the policy relevance of household approaches to income, underscoring the relevance of new developments in the economics of the family to public policy.

Overall, families and households (as well as individuals) are important units of analysis for the collection, organization, and study of statistical data. Nonetheless, a growing body of research testifies to the complex interactions and varied interests within these units and to the ways in which they may, or may not, behave as unitary actors. This paper serves as a reminder of this point, documenting the range and correlates of income management strategies among older couples.

Many qualitative studies drawing nuanced accounts of income management strategies from in-depth, semi-structured interviews with small numbers of couples have been published in recent years. This study provides a useful complement, estimating the incidence and correlates of these strategies across the population in a way that qualitative studies cannot. Furthermore, this analysis uses a more complete set of socio-economic variables than is often available in quantitative studies. For example, Vogler, Brockmann, and Wiggins (2006) as well as Heimdal and Houseknecht (2003) document differences in the income management strategies of married and common-law couples, but are not able to determine the extent to which this reflects systematic differences in the duration of these relationships or in the presence or parentage of children. Other studies focus on either married (Treas 1993) or common-law couples (Winkler 1997; Elizabeth 2001). Consequently, the objective of this analysis is to examine the socio-economic covariates associated with income management strategies using a large, nationally representative survey containing more complete information on respondent characteristics than has been available to date. Issues of gender inequality and power relationships, which are the focus of many studies of income management, are beyond the scope of this analysis.

The data source being used does not contain information on respondents' attitudes about gender roles 5  or about the importance attached to values such as personal independence, autonomy, or ownership of assets—factors found to be correlated with income management strategies (e.g., Vogler, Brockmann, and Wiggins 2006; Yodanis and Lauer 2007a). The extent to which between-group differences are attributable to these factors cannot be estimated. The data source for this study is limited to respondents aged 45 or older, although their spouse or partner may be younger. Results from other studies suggest a significant, but weak, correlation between age and income management strategies (more on this below). We were unable to identify any income management studies that tested for interaction effects between age and other variables, such as common-law status. In spite of the age limitation faced, the analysis yields robust results regarding the socio-economic characteristics associated with income management strategies among 'older' Canadians.

The remainder of this paper is organized into four sections. In Section 2, the context for the analysis is set by highlighting key themes in the research literature. In Section 3, the data source and methods are discussed, while in Section 4, findings are presented, including results from two multivariate models and one Blinder–Oaxaca decomposition. In Section 5, conclusions and implications are discussed.

2   Literature review

Economists have been grappling with economic behaviour at the household level for many years now. The traditional economic approach—the unitary model—largely extend individual models of economic behaviour to households (Samuelson 1956 and Becker 1973, 1981). The unitary model assumes that a household acts as a single decision-making unit, maximizing a single household utility function, subject to a single budget constraint. The model also imposes important restrictions, such as income pooling, whereby only total exogenous family income (and not its distribution across household members) matters for labour supply and consumption decisions, and provides for symmetry in the cross-wage effects on the labour supply of each household member, whereby the change in a wife's labour supply following a change in her husband's wage will be equal to the change in the husband's labour supply following a change in his wife's wage. Both income pooling and symmetry restrictions have been strongly rejected in empirical studies (Phipps and Burton 1996; Fortin and Lacroix 1997).The model has also been criticized on the grounds that it leaves no room for determining the intra-household allocation of consumption and labour supply. As Phipps and Burton (1995) summed up, the unitary model "…ignores the rather obvious fact that the family is made up of individuals with unique tastes and preferences who may or may not always agree—who may or may not have equal powers—who may or may not be equally well off." (p. 179).

Partly for these reasons, new approaches based on individual preferences which use game theory have been developed (e.g., McElroy 1990, Kooreman and Kapteyn 1990, Chiappori 1988). However, it is the collective model of Chiappori (1992) that has opened the "black box" of family economic behaviour. In the collective model, family members are characterized as having their own preferences and interests, and bargaining is assumed to take place among household members. The bargaining process is influenced by a sharing rule, with the model assuming only that bargaining outcomes result in Pareto efficient allocations of household resources. The model does not impose restrictions such as income-pooling or symmetry of cross-wage effects. While the unitary model has been rejected in empirical studies, the collective model could not be rejected on the basis of either expenditure (Bourguigon et al. 1993) or labour supply (Fortin and Lacroix 1997) data.

The themes of intra-household negotiation and bargaining evident in collective models proposed by economists are also central themes in other disciplines. For example, some consumer researchers argue that women's entry into the paid labour force and their receipt of earnings have strengthened women's bargaining position within the home and changed the way that consumption decisions are made. Belch and Willis (2002) report that women within couples now play a greater role in the purchase of automobiles, family vacations, and insurance and financial services than they did in mid-1980s and that decisions in these product areas have shifted from a husband-dominated process to a joint decision-making process. Intra-household negotiation and differences in power, material conditions, and risk are central themes in feminist research as well. The extent to which the entry of women into the paid labour has resulted in a renegotiation of household issues, such as domestic labour, work–family balance, child care, and income management, is a central theme in the literature (Vogler 2005; Tichenor 2005).

The organization of income is one issue within this broader theme of household economic behaviour and intra-household negotiation. In the 1980s, Jan Pahl conceptualized (1986) and applied a four-part income management classification capturing gradations of control over income exercised by spouses or partners (Heimdal and Houseknecht 2003). Pahl's typology of income management includes: (1) the whole wage system; (2) the housekeeping allowance system, in which one spouse or partner maintains control over the household budget and allocates a share or set amount to the other; (3) the pooled or shared management system, in which income is put into a common pot and used as needed by both spouses or partners; and (4) the independent management system, in which spouses or partners maintain separate control over their incomes and how they are spent. A variant of Pahl's typology has been used in more recent surveys, including the 1994 and 2002 International Social Survey Programme (ISSP). Specifically, ISSP respondents were asked:

  1. How do you and your spouse/partner organize the income that one or both of you receive? Please choose the option that comes closest.
  1. I manage all the money and give my spouse/partner his/her share.
  2. My spouse/partner manages all the money and gives me my share.
  3. We pool all the money and each take out what we need.
  4. We pool some of the money and keep the rest separate.
  5. We each keep our own money separate.

Other qualitative and survey-based studies have also focused on the prevalence of allocative, pooled, and separate-income management strategies (Burgoyne and Morison 1997; Burgoyne et al. 2007; Vogler and Pahl 1993). It should be noted that these strategies do not tell us about how equitably income and expenses are shared between spouses or partners. For example, the relative size of a spouse or partner's share of income under the allocative approach is not specified, nor is there any guarantee that spouses or partners using a pooled strategy have equal access to funds or draw comparable amounts for personal expenditures. Similarly, spouses or partners who separate their income may or may not contribute equally (or equitably) to expenses. Ashby and Burgoyne (2008) underscore the complexity of the issue, noting that couples who partially or completely separate their incomes perceive and handle money in a variety of ways (also see Vogler, Brockmann, and Wiggins 2006 (p. 478)). In short, broad categories do not capture the complex and varied ways in which couples manage their finances.

These caveats noted, responses to the income management question are suggestive of the degree of independence that individuals in couples exercise over their income. Tracking 42 couples over the first year of marriage, Burgoyne et al. (2007, p. 214) found that "those choosing more separation in money matters did so in order to maintain their financial identity and autonomy." Similarly, Vogler (2005, p. 12) states that individuals who use separate-income strategies "…tend to operate as two autonomous individuals each with their own separate accounting systems who then exchange goods and services between them on the basis of market-like relationships." 6 

A range of characteristics have been associated with the income management strategies used by couples, and these can be broadly organized into three themes.

2.1  Demographic characteristics

Several studies document the correlation between demographic characteristics of couples and the likelihood of using pooled or separate approaches to income management. Marital status is among these demographic characteristics, as common-law couples are more likely than married couples to use a separate approach (Heimdal and Houseknecht 2003; Vogler 2005; Woolley 2003). A number of possible explanations have been advanced for this. Expectations regarding the permanence of the relationship may be lower among common-law than married couples, with incomes kept separate to protect self-interests in the event of break-up. The values espoused by individuals who self-select into common-law unions—such as individualism, personal autonomy, and equality of partners' contributions—may also contribute to the separation of income (Burgoyne et al. 2006; Brines and Joyner 1999). Laws regarding the treatment of joint property are another consideration (Heimdal and Houseknecht 2003).

Several studies show that couples in which one or both partners have been previously divorced are more likely than others to use a separate approach to income management (Yodanis and Lauer 2007a; Heimdal and Houseknecht 2003; Woolley 2003; Treas 1993). Prior divorce may lower expectations of permanence in the current relationship and may also increase the complexity of familial and financial arrangements—such as the payment or receipt of child support—making the separation of income more practical on a day-to-day basis.

The duration of the relationship is another factor. The separate approach is less prevalent (and the pooled approach more prevalent) among couples in longer-term relationships (Winkler 1997; Treas 1993). This may reflect increased expectations of permanence, the acquisition of common goods over time (Burgoyne et al. 2007), or increased likelihood of children being born to the couple (Winkler 1997; Treas 1993). The presence of children born to both partners is also positively correlated with income pooling (Winkler 1997; Heimdal and Houseknecht 2003).

Finally, a significant, but weak, relationship is often found between age and income management strategies. For example, Treas (1993) found a negative, but weak, relationship between the wife's age and the likelihood of using separate bank accounts. Similarly, in their bivariate results, Bonke and Uldall-Poulsen (2007) showed that income pooling is least prevalent among couples in their twenties, but that it varies by about 1 to 4 percentage points across couples in their thirties, forties, and fifties. Vogler, Brockmann, and Wiggins (2006) find a positive correlation between age and independent income management among male respondents but not among female respondents or among all respondents. As well, Yodanis and Lauer (2007a) find a significant, but weak, correlation between age and income management strategies. Consequently, while younger individuals are often in the types of relationships in which separate-income strategies are used (i.e., common-law union, shorter duration, without children), the effect of age per se appears to be weak relative to the effect of other characteristics.

2.2  Income and education

Some studies have examined the correlation between income management strategies and the income characteristics of couples. Treas (1993) addresses this in terms of absolute income levels, suggesting that the correlation with income management strategies may run in either direction. On the one hand, higher-income couples may be more likely to use separate accounts in order to avoid the inconvenience of sharing. On the other hand, they may be more likely to pool their income if there is less concern about 'free riding' or 'spendthrifts' when resources are more plentiful. Treas finds that the likelihood of maintaining separate accounts is positively correlated with couples' incomes. Heimdal and Houseknecht (2003) observe the same in the Swedish portion (but not in the American portion) of their sample, while Bonke and Uldall-Poulsen (2007) do not find a consistent pattern across household income quintiles. Vogler, Brockmann, and Wiggins (2006) do not include household income in their model, but find that the likelihood of separating income is higher among managerial or professional than among 'working class' couples.

Another approach to resources is the relative contribution of each partner to the couple's income. Certainly the circumstances for maintaining separate finances do not exist when only one spouse receives income. Beyond that, Bonke and Uldall-Poulson (2007) suggest that an unequal distribution of income within couples may increase the likelihood of separating income when spouses are 'egoistically inclined' or the likelihood of pooling income when they are 'altruistically inclined'. Bonke and Uldall-Poulson find that income pooling is not strongly associated with spouses' income contributions. This is also the case with Heimdal and Houseknecht (2003). In contrast, Yodanis and Lauer (2007a) hypothesize that when the relative economic contributions of spouses approach parity couples will be more likely to use a pooled, rather than an allocative, approach to income management. Their results are consistent with this view.

Net of income, Treas (1993) finds that higher levels of educational attainment among women are positively associated with maintaining separate accounts. However, considering relative levels, Bonke and Uldall-Poulson (2007) do not find a significant correlation between spouses' relative levels of education and income management strategies.

2.3  Attitudinal characteristics

Finally, some researchers have examined the correlation between income management strategies and other aspects of relationships, such as normative attitudes and personal values. For example, normative attitudes regarding gender roles are often based on how strongly respondents agree or disagree with declarative statements. Vogler, Brockmann, and Wiggins (2006) find that more traditional attitudes to income earning are correlated with allocative income management strategies; Yodanis and Lauer (2007a) report a similar finding. However, the measure of 'traditional gender ideology' used by Heimdal and Houseknecht (2003) is not significantly correlated with income management strategies among either the Swedish or American portions of their sample. Considering other values, Burgoyne et al. (2007) report that "…perceived ownership of income and other assets" and a desire to maintain financial identity and autonomy are positively associated with the separation of income.

3   Data and methodology

3.1  Data source

Data for this paper were drawn from Statistics Canada's 2007 General Social Survey (GSS). The target population for the 2007 GSS was all persons 45 years of age and over residing in Canada, excluding residents of Nunavut, the Yukon, and the Northwest Territories, and full-time residents of institutions. The 2007 GSS was completed by 23,404 respondents, of whom 12,854 were living with a spouse or common-law partner at the time of the survey. From this group, respondents for whom proxy responses were provided by another household member, 7  who did not provide useable responses to the question regarding income management strategies, 8  or who did not provide complete information on the independent variables in this analysis 9  were excluded. Households in which family members aside from the couple received income were also excluded from the analysis. 10  This was done to remove the effects of income strategies adopted for managing the income received by parents, children, and other members, and focus solely on the strategies for managing the income of husbands and wives. A final sample of 9,489 respondents is used for the analysis.

2007 GSS respondents were asked:

  1. Which statement best describes how your household income is organized?
  1. You manage all the money and give your spouse/partner their share.
  2. Your spouse/partner manages all the money and gives you your share.
  3. You pool all the money and each takes out what he/she needs.
  4. You pool some of the money and keep the rest.
  5. You keep your own money separate.
  6. Other

For this analysis, the above categories are aggregated into three groups. The first two categories are combined because they capture the same underlying strategy identified as allocative. The third category is treated as a distinct group that is identified as the pooled strategy. And the fourth and fifth categories are combined into a single group identified as the separate strategy. 11 

The distribution of respondents across the three income management categories (as well as across the five original response categories) is shown in Table 1. Of the respondents who use the separate approach to manage their income (22.9%), about two-thirds (i.e., 15.3/22.9) reported that they and their partner "…keep [their] own money separate," while about one-third (i.e., 7.6/22.9) stated that they and their partner "…pool some of the money and keep the rest." In short, complete separation of income is more prevalent than partial separation within this group. In this respect, the conceptual distance between the pooled and separate approaches to income management is larger than it would have been had most respondents partially separated their income.

It is important to note that the 2007 GSS surveyed individuals rather than couples, with information collected from only one spouse or partner. Whether the other spouse or partner agrees with the respondent's assessment of the couple's approach to income management cannot be determined. In the aggregate, the responses provided by men and women are very similar. About 8% of both male and female respondents stated that the husband or male partner manages the couple's money and gives his wife or female partner her share, and about 12% of male and female respondents stated that the wife or female partner manages the income and gives her husband or male partner his share (Table 1). The proportions of male and female respondents reporting that they use a pooled or separate approach to income management are within 2 to 3 percentage points.

3.2  Independent variables and descriptive results

This analysis includes a broad set of socio-economic characteristics drawing on the information that respondents provide about themselves and about their spouse or partner. Respondents' sex and age and the age of their spouse or partner are included. The immigration status of both the respondent and his or her spouse or partner are included on the grounds that income management strategies vary across world regions, potentially resulting in differences between individuals born in Canada and individuals born abroad. 12  The base model includes a dummy variable for each spouse or partner (Canadian-born = 0, immigrant = 1), and a second model includes four dummy variables capturing the immigration statuses of both spouses or partners in combination. 13 

Information regarding the relationship itself is included. Distinction is made between individuals in common-law and marital unions, with the expectation that the former are more likely to separate their incomes, for the reasons discussed above. The duration of the current relationship and a dummy variable identifying respondents who have been married before are also included. Consistent with the literature, it is expected that the separation of income will be correlated with shorter relationships and with prior marriages. The characteristics and history of the family in which respondents currently reside are included. By definition, all respondents currently live with a spouse or common-law partner. In addition, the following are identified: (i) individuals who have previously had children with their current spouse or partner, but who are no longer residing with those children (i.e., empty-nesters); (ii) individuals who have not previously had children with their current spouse or partner and have no children residing with them (i.e., couple no children); (iii) individuals living with children born to both members of the couple (i.e., couple with children); and (iv) individuals living with children born to one member of the couple (i.e., other couples). The presence of children (now or in the past) is expected to be positively correlated with the pooled approach to income management.

With regard to education and income, the level of education of the respondent and that of his or her spouse or partner are included, broadly defined as the presence of a post-secondary credential. The base model includes a dummy variable for each spouse or partner (no post-secondary credential = 0, post-secondary credential = 1). The second model also includes the wife's level of educational attainment relative to that of her husband. 14  Treas (1993) suggests that additional years of education may enhance individuals' money management skills, thereby increasing their capacity to maintain separate accounts, or may strengthen their bargaining position within the household and increase their capacity to maintain separate incomes, if so preferred. Income level of each partner is included in the base model, while the second model also includes the wife's level of income relative to the husband's. 15 

Finally, a variable combining the respondent's place of residence (defined as residing in Quebec or residing elsewhere in Canada) and the main language which the respondent uses at home (defined as English, French, or other) is included. This yields six categories: English outside of Quebec (reference group); French outside of Quebec; others outside of Quebec; English in Quebec; French in Quebec; and others in Quebec.

The distribution of the sample across these characteristics and cross-tabulations with income management strategies are shown in Table 2. The vast majority of individuals in the sample (89%) are married, while 11% are in common-law unions. There is a 34-percentage-point difference in the share of married and common-law individuals using a separate-income strategy (at 19% and 53%). Most individuals in the sample have been in their relationship for more than 20 years (71%), while relatively few have been so for less than five years (6%). There is a strong bivariate relationship between relationship duration and use of separate-income strategies. About half of the sample (48%) are 'empty-nesters', about one-quarter (27%) live with a spouse or partner with whom they have never had children, and about one-fifth (22%) live with their spouse or partner and children born to both. Only a small portion of the sample resides in blended families (3%); it is interesting to note that the use of the separate-income strategy is quite prevalent among this group (at 40%). About one-fifth of the sample (22%) has been previously married; consistent with the literature, separate-income strategies are prevalent among this group. In terms of income, there is not a noticeable bivariate correlation between the husband or male partner's income and income management strategies (with the exception of the few who reported no income), while the use of separate-income strategies are positively correlated with the income of the wife or female partner. Finally, there are significant differences in the use of separate-income strategies among respondents residing in Quebec and respondents residing elsewhere in Canada, and among respondents who speak English, French, or other languages in the home.

4   Results

Our multivariate analysis proceeds in three steps. First, results from a multinomial logit model in which the dependent variable is comprised of the three income management strategies are presented. A second model is run on the same dependent variable, but both absolute and relative levels of education and income are included. This makes it possible to assess whether income management strategies are correlated with the absolute or relative characteristics of individuals—an issue that is central to the intra-household bargaining perspective on economic behaviour. Finally, differences in the use of separate-income strategies among common-law and married couples are examined by means of a Blinder–Oaxaca decomposition. All models are calculated using bootstrap weights to correct variance estimates for survey design. Results are shown as 'marginal effects', along with the standard errors of the estimates and levels of statistical significance.

4.1  Multivariate regressions

Most coefficients in the analysis yield results consistent with expectations. Sex and age are two exceptions. The predicted probability of using a separate approach to income management is 2 percentage points higher among women than among men (Table 3). Given the data available from the GSS, one cannot determine whether this is attributable to how husbands and wives within the same couple perceive their finances or whether this is attributable to other factors. A significant correlation between the age of the wife or female partner and the likelihood of using either a pooled or separate approach to income management was found as well. Again, no ready explanation can be found for this. Different categorizations of the age variables were used in earlier analyses but yielded similar results. 16  Immigration status is correlated with income management strategies; the probability of using a separate approach is almost 3 percentage points lower among male immigrants than among their Canadian-born counterparts. A similar result is found using the alternative specification of immigration status.

In terms of the characteristics of the relationship, a strong correlation is found between the duration of the relationship and income management strategies. Compared with individuals who have been in their relationship less than five years, those in relationships of 10 to 19 years are about 14 percentage points less likely to separate their income, while those in relationships of more than 20 years are almost 21 percentage points less likely. Moreover, individuals in common-law unions are far more likely than those in legal marriages to separate their income—a difference of 15 percentage points—even after the duration of the relationship is taken into account.

The presence of children—either currently or previously residing with the couple—is correlated with income management practices. Compared with couples who have never had children, empty-nesters are less likely to separate their income (a difference of almost 7 percentage points) and more likely to pool it. Similarly, the likelihood of using a separate approach is lower among couples who have children currently residing with them, whether the children were born to both spouses or partners (a difference of 10 percentage points) or to only one of the spouses or partners (a difference of 7 percentage points). Finally, while other studies have reported that the separation of income is positively correlated with one or both spouses or partners having a prior marriage, this is not the case in the multivariate results. 17 

Education is correlated with the type of income management strategy used. The predicted probability of using a separate approach to income is higher among women and men who have post-secondary credentials than among their counterparts who do not (differences of about 4 percentage points and 2 percentage points, respectively). Conversely, men and women with post-secondary credentials are less likely than others to use an allocative system. When the education of wives relative to that of their husbands is included in the model, no significant correlations are observed (Table 4); this suggests that it is the absolute levels of education, rather than the relative levels between spouses, that matter.

Income management strategies are also correlated with income, particularly with that of wives. When one of the spouses or partners is reported to have no income, the likelihood that the couple uses an allocative approach is substantially higher—almost 18 percentage points among men (relative to men with incomes of $20,000 to $39,999) and almost 14 percentage points among women (relative to women with incomes of $1 to $19,999). 18  Aside from this correlation, income management strategies are not associated with the amount of income received by the husband or male partner. However, the likelihood of using a separate approach to income management is correlated with the income of the wife or female partner. The likelihood of using a separate approach to income is about 5 percentage points higher among wives or female partners with incomes of $20,000 to $39,999 and about 8 to 12 percentage points higher among wives or female partners with incomes over $40,000 than among wives or female partners with incomes of $1 to $19,999. When the incomes of wives relative to their husbands' are included in the model, the variable is not significant (Table 4). As with educational attainment, it appears that it is the absolute level of income rather than relative level between spouses that is correlated with income management strategies.

The final variable in the analysis captures regional and linguistic characteristics of individuals. The reference group for this variable is Anglophones residing outside Quebec. 19  Compared with this group, Francophones residing outside Quebec are significantly more likely to use a separate approach to income management (a difference of 7 percentage points) as are Francophones residing in Quebec (a difference of 11 percentage points). However, while this suggests that there is a difference between official-language groups, it is also important to note that Anglophones in Quebec are also significantly more likely than Anglophones outside Quebec to use a separate approach to income—a different of 9 percentage points. Hence, the results indicate that income management strategies are correlated with both geographic and linguistic characteristics.

4.2  Decomposition analysis

The greater prevalence of separate-income strategies among common-law than married couples certainly stands out. To gain a clearer perspective of the factors accounting for this difference, a Blinder–Oaxaca decomposition technique is used to estimate the extent to which the between-group difference is attributable to socio-economic characteristics. Results from this decomposition are shown in Table 5.

There is about a 33-percentage-point difference in the use of separate-income strategies between married and common-law couples. Differences in the socio-economic characteristics of the two groups account for 14 percentage points—or about 42%—of this overall difference (Table 5). In other words, if the common-law individuals in the sample had the same socio-economic profile as the married individuals, the difference in the shares using a separate approach to income would decline from 33 percentage points to about 19 percentage points. Duration of relationship plays the largest role, accounting for about 8 percentage points, while family composition accounts for almost -3 percentage points. 20  The fact that a disproportionate share of common-law individuals are Francophones (who are more likely to use a separate approach to income management) accounts for 1.7 percentage points of the difference while the relatively large share of common-law couples residing in Quebec (where the separation approach is more prevalent) accounts for another 0.9 percentage points. Prior marriages and income characteristics account for another 2 percentage points, while demographic characteristics exert a modest countervailing influence. 21 

5   Conclusion

The characteristics of couples have changed dramatically over the last forty years. The increasing prevalence of common-law unions and changes in family structure are evidence of demographic change, while the widespread entry of women into paid employment has fundamentally altered the financial characteristics of couples. Cultural changes, such as attitudes regarding gender roles in the home and workplace, have also occurred. In this context, researchers from a variety of disciplines are re-examining how couples negotiate and manage a wide range of economic activities, paying particular attention to the resources, interests, and preferences unique to each spouse or partner. One theme in this work is the weight that spouses or partners attach to individual interests versus joint interests in their economic decisions.

The way that couples organize their income can be examined in these terms. While the broad categories used here likely do not adequately capture the complexity of the arrangements that couples use, they are suggestive of the degree of independence involved. The socio-economic characteristics correlated with income management strategies are also consistent with this interpretation. Given the focus of this paper on couples aged 45 or older—a group largely characterized by longer-term marital unions in which children are or have been present—one might reasonably expect income pooling to be most prevalent. This is indeed the case, as 57% of respondents pool all the money with each taking out what he or she needs and a further 20% pooling income, with one spouse or partner managing and allocating it. The data in this study do not allow us to examine how these strategies correlate with the actual distribution of income between spouses or partners. However, even among older couples, income management strategies are often characterized by a degree of independence, with 23% of the respondents in the sample either partially (8%) or completely (15%) separating their incomes.

The characteristics of relationships matter a great deal in this regard. As often noted in the research literature, separate-income strategies are more prevalent among common-law couples than among married couples—a raw difference of almost 34 percentage points. However, a considerable portion of this difference (42%, or 14 percentage points) is attributable to socio-economic characteristics that systematically differ between these groups, most notably the duration of the relationship. The extent to which these characteristics account for between-group differences has not received much attention in the literature. Still, the majority of the married or common-law difference remains unaccounted for in the decomposition. Between-group differences in values and attitudes, such as independence, autonomy, perceived ownership of assets, and expectations regarding the permanence of the relationship may account for some of this.

The likelihood of using a separate approach to income management is strongly and positively correlated with the wife's or female partner's absolute income, but is not with the wife's or female partner's income relative to that of her spouse. Resource theory suggests that the relative contributions of spouses are the "…key factor promoting more or less equal arrangements [within couples]", with the implication that, "…when couples earn equal amounts, they are more likely to manage their pooled income jointly" (Yodanis and Lauer 2007a, p. 1309 and p. 1320). The results from this study do not support this view, as no significant correlation is evident between relative income and income management strategies. Several possible explanations may be advanced for the positive correlation between wives' or female partners' absolute income and the use of separate-income strategies, such as greater importance attached to independence or autonomy, desire to maintain control over personal income, and/or perceived ownership of personal assets among higher-income women. Convenience may also be a consideration, following Treas's (1993) findings.

The growing number of Canadians in second marriages or blended families raises the question of whether complex families have complex finances. The descriptive results from this study are suggestive of such a relationship, as the share of individuals in blended families using a separate-income strategy is twice as large as the share of individuals in families with children born to both spouses or partners doing so (40% and 17%, respectively). However, the multivariate results indicate that it is the presence of children—regardless of parentage—that is positively associated with income pooling. Readers are reminded that the sample used in this study is restricted to individuals aged 45 or older. A sample of couples in their twenties and thirties may yield different results. The issue of complex families and complex finances is also reflected in the descriptive results from this study, which show that individuals with prior marriages are about twice as likely to use a separate-income strategy as individuals with no prior marriages (37% and 19%, respectively). This is consistent with findings in the literature. Again, however, this correlation is non-significant in the multivariate results, perhaps because a range of covariates broader than that employed in most other studies is used here. Interestingly, two of the characteristics most strongly and positively correlated with separate-income strategies—common-law status and women's income—have changed markedly in recent years. While this might appear suggestive of a trend towards this type of income management strategy, the analysis presented above is based on data from a single point in time and offers no evidence regarding trends.

Overall, this study provides an opportunity to look at how couples arrange their incomes and documents considerable variation in this regard. However, the data used do not explicitly tell us about how equitably incomes are shared between spouse or partners or whether there are differences in the living standards of household members. These issues are relevant to public policy and information remains scarce in this area.

Date modified: