5 Discussion and conclusion

View the most recent version.

In both Canada and the United States, rising employment among older (40 and over) single mothers occurred mainly in the 1980s and was largely the result of changing composition. In contrast, changes among younger single mothers were mainly confined to employment levels, not earnings, took place in the 1990s and had little to do with rising education or other compositional changes. The gains in annual earnings among older employed single mothers mainly reflected their increased earnings power as a result of higher education levels rather than increased labour supply. In contrast the modest gains of younger lone mothers mainly reflected an increase in weeks worked.

These broad similarities, however, mask two large differences in the magnitude of these effects. While compositional changes dominated the account for older single mothers, the magnitude of these compositional effects was substantially larger in Canada than in the United States, despite the fact that compositional changes were somewhat larger in the United States. The share of older single mothers with some postsecondary schooling or a university degree rose by 32 percentage points in the United States, compared with 28 percentage points in Canada. Instead, differences in trends among older single mothers reflected the larger Canadian gains in mothers' employment rates more generally. Over the two decades, employment among Canadian married mothers rose from 58% to 80% (22 percentage points), compared with an increase from 59% to 73% (14 percentage points) in the United States.

The pattern was reversed among younger single mothers. Despite starting from a higher base, the employment gains among U.S. single mothers under 30 were two to three times higher than in Canada and mainly due to differences in 'behavioural' changes. While our data do not allow us to demonstrate the reasons for this difference, the much stronger behavioural effect in the United States is consistent with the standard set of stylized facts concerning differences in the extent to which social provisions in the two countries enable single mothers to remain at home.

Traditionally, both nations have relied on welfare provisions for single parents that provided strong disincentives to employment, namely means-tested programs that were reduced dollar for dollar as earnings rose. Nevertheless, significantly higher social transfers have arguably made non- employment a somewhat more viable alternative in Canada than in the United States. Blank and Hanratty's simulations (1993) with 1986 data show that single-parent poverty in the U.S. population would decline substantially under a Canadian-style transfer system. Moreover, over the 1980s, income transfers to single mothers were falling in the United States and rising in Canada (Hanratty and Blank 1992).

During the 1990s, both countries went through welfare reforms that were aimed at increasing work incentives for single mothers. In the United States, however, these reforms were national in scope, introduced as a result of the Personal Responsibility Act in 1996. In Canada, significant cuts in social assistance and changes in eligibility criteria were at the discretion of the provinces, and large reductions were mainly confined to Ontario and Alberta.⁸

Equally important, both countries have introduced 'carrots' as well as 'sticks' to encourage employment (Myles and Pierson 1997). In the United States, the Earned Income Tax Credit (EITC), introduced by Richard Nixon, has been expanded under every subsequent administration except for the most recent. The Canadian variant has been evolving since 1978, culminating with the adoption of the National Child Benefit (NCB) in 1997 but with an important difference. Since access to EITC is conditional on earnings, single mothers must be employed in order to benefit. The NCB, in contrast, also provides income-tested benefits to the non-working poor. Indeed, stay-at-home mothers with zero earnings receive the highest benefits (Mendelson 2003). While neither country's benefits are generous by international standards (Rainwater and Smeeding 2003), both the level and changes in incentives to work are undoubtedly much stronger for American single mothers.

The finding that the labour market behaviour of younger and older single mothers in the same social policy environment are very different is hardly novel. Dooley (1999) reports sharp differences between the labour force participation rates of single mothers under 35 compared with single mothers over 35. Analysing changes in welfare participation of Canadian single mothers in the 1973-to-1991 period, he finds that single mothers under 35 had an increasing reliance on social assistance, accompanied by stagnant wages and declining labour force participation. Single mothers over 35, in contrast, showed decreasing reliance on social assistance, accompanied by rising wages and earnings. Women who become single mothers later in life are more likely to have significant labour force experience and to be exiting from unions with considerably higher family earnings. Juby, Le Bourdais and Marcil-Gratton (2003: 20) show that single mothers' financial circumstances after separation are closely linked to the intact family income before separation. The most affluent single mothers, for instance, are those from the most affluent intact families, who were already fully involved in the labour force before they separated.

At the end of the 1990s, the increase in single mothers' employment was welcomed largely because of the associated decline in single mother poverty rates. Based on Luxembourg Income Study (LIS) poverty measures (families with incomes less than 50% of median income), single mother poverty stood at 52% in the United States and 49% in Canada at the end of the 1970s. During the 1980s, rates rose in the United States—to 61% in 1991—but fell back to 49% in 2000 (LIS key figures). In Canada, rates were more or less stable over the 1980s but then declined to 41% in 2000.

It would be naïve, however, to conclude that the declining poverty rates of the 1990s are indicative of a longer-term trend that is likely to continue into the future. At 82 %, the labour force participation rate of single mothers in the United States was reaching saturation levels by 1999 and it was significantly higher than the rate for married mothers with children, which stood at 73%. Moreover, the majority of employed U.S. single mothers were employed almost full year (40 weeks and more). There is little room for future gains from encouraging or compelling even stronger labour market attachment by additional benefit reductions.⁹ There is more room for additional gains among Canadian single mothers, where employment levels reached 73% in 2000, compared with 80% for married mothers. But the demographically driven gains of Canadian single mothers in the past quarter century are unlikely to be repeated in the future. The aging of the baby boom was a one- time event that will only be faintly 'echoed' as their children enter their child-bearing years.

 

^8. Separate analyses for these two provinces also show a somewhat different pattern from national trends. During the 1990s, employment levels rose by 8.8 and 7.5 percentage points among lone mothers in Alberta and Ontario, respectively. Quebec lone mothers, however, had the highest employment growth in the 1990s (9.4 percentage points) and, by 2000, their employment levels were 4 percentage points higher than in the rest of Canada. Compositional shifts accounted for only 28% of the growth in the 1990s, results that are consistent with conclusions concerning the effects of liberalized child-care provisions in that province (Baker, Gruber and Milligan 2005; Lefebvre and Merrigan 2005).

^9. From 1994 to 2003, maximum monthly Aid to Families with Dependent Children and Temporary Assistance for Needy Families benefits for single parents fell in all U.S. states except 5; in 25 states, maximum benefits fell by 18% (Green Book, Table 7-10). Since 1970, maximum benefits have fallen by 40% or more in 42 states. (Green Book, Table 7-13).