Profiles and Transitions of Groups at Risk of Social Exclusion: Lone Parents - November 2002

9. Reliance on Social Assistance

Previous

Contents

Next

9.1 Introduction

This section uses similar types of analysis to those used in previous sections of the study to understand why some lone mothers spend long spells on social assistance (SA), while others stay on SA for only short periods or do not rely on SA at all. Because of the close correlation between low income and SA, we expected that many of the findings of the previous sections with respect to low income will also hold for SA. For this reason, the analysis of SA was designed to be more selective than that of low income.

One of the challenges in analyzing SA with survey data is incomplete reporting of SA benefits. However, this problem is less serious with SLID than previous surveys, because of the increased use (with respondent consent) of income tax information. A related issue is that the definition of SA is not always clear, especially with respect to certain provincial income tested programs that tend to have a fairly universal application (e.g. income tested child benefits). To minimize this problem:

We begin the section with an examination of the incidence and duration of SA, as well as its impact on low income gaps. Then we examine the characteristics of SA recipients and attempt to identify the characteristics that are most closely associated with a high incidence of SA. Finally, we examine the duration of SA spells and the characteristics of SA recipients that are most closely associated with longer SA spells.

9.2 Incidence of Social Assistance

In this sub-section we use the 1998 cross-sectional SLID data to examine the prevalence of the use of SA among all lone mothers and among low income lone mothers. Chart 9.1 (and in more detail Table 9.1) shows that lone mothers not only have the highest incidence of low income, but they also have the highest incidence of SA among all low income families. In particular:

• both among all families and among low income families, lone mothers have the highest incidence of SA (40% and 68% respectively);
• the incidence of SA among lone fathers is less than half of that of lone mothers, but considerably higher than that of other family types with a male main income recipient.

There are several possible explanation why the incidence of SA among low income lone mothers is higher than among other low income families. For example: (a) before government transfers, the average low income lone mother is deeper in low income than other low income families (as revealed by comparisons of low income gaps in Table 4.1); (b) lone mothers may be facing fewer legal barriers or more sympathetic bureaucrats when applying for benefits; and (c) lone mothers may be applying more readily for SA, possibly because of non-cash benefits (such as child care or insured health benefits), or because they anticipate a longer stay in low income.

Table 9.1 Incidence of social assistance among all major income recipients, 1998¹

	Lone parent, kids<18	Unattached individual	Couple without kids<18	Couple with kids<18	Other economic families	All economic families
Male
All major income earners	110,545	1,392,465	1,309,674	2,551,716	361,897	5,726,297
- Received SA	20,501	157,578	67,777	134,373	61,717	441,946
- Incidence of SA	18.5%	11.3%	5.2%	5.3%	17.1%	7.7%
Low income major income
earners	18,506	400,738	45,288	149,940	35,069	649,541
- Received SA	***	133,555	16,956	53,071	***	223,595
- Incidence of SA	***	33.3%	37.4%	35.4%	***	34.4%
Female
All major income earners	630,731	995,847	548,226	682,179	268,630	3,125,613
- Received SA	253,806	131,167	27,808	53,268	60,389	526,439
- Incidence of SA	40.2%	13.2%	5.1%	7.8%	22.5%	16.8%
Low income major income
earners	246,407	349,879	41,682	85,805	45,745	769,518
- Received SA	168,011	113,944	***	22,809	26,498	339,141
- Incidence of SA	68.2%	32.6%	***	26.6%	57.9%	44.1%
Both genders
All major income earners	741,276	2,388,312	1,857,900	3,233,895	630,528	8,851,911
- Received SA	274,308	288,745	95,585	187,641	122,107	968,385
- Incidence of SA	37.0%	12.1%	5.1%	5.8%	19.4%	10.9%
Low income major income
earners	264,913	750,617	86,970	235,745	80,813	1,419,059
- Received SA	176,654	247,499	24,834	75,881	37,868	562,736
- Incidence of SA	66.7%	33.0%	28.6%	32.2%	46.9%	39.7%
(1) Sample of major income earners, age 16-55, in 1998. ***Less than 30 observations.

9.3 Impact of Social Assistance on Low Income Gaps

This sub-section examines the impact of SA in terms of reducing the low income gap. In particular, we want to see whether SA takes a bigger "bite" out of low income among lone mothers than is the case for other low income families.

Chart 9.2 shows that among all female main income recipients on SA, the low income gap before any government transfers is around 90%, while after government transfers the gap for most female SARs drops to about 30% with the exception of unattached female SARs where the gap drops to about 40%.

SA has a somewhat smaller impact on the low income gap of female SARs with children. However, this result is more than offset by other government transfers (such as the Child Tax Benefit).

Table 9.2 Impact of social assistance on low income, 1998, among all poor social assistance recipients (SARs)¹

	Lone parent, kids<18	Unattached individual	Couple without kids<18	Couple with kids<18	Other economic families	All economic families
Male
All SARs in 1998	***	133,555	16,956	53,071	***	223,595
Low income gap:
- before any transfers	***	93.0%	83.5%	89.8%	***	90.8%
- after transfers, exc. SA	***	85.8%	70.2%	68.1%	***	78.9%
- after all transfers	***	39.4%	25.2%	27.4%	***	34.7%
Impact of SA on gap	***	46.4%	45.0%	40.7%	***	44.2%
Female
All SARs in 1998	168,011	113,944	***	22,809	26,498	339,141
Low income gap:
- before any transfers	91.8%	95.6%	***	93.0%	91.0%	93.0%
- after transfers, exc. SA	70.0%	88.5%	***	71.6%	80.5%	77.3%
- after all transfers	28.0%	39.4%	***	27.4%	31.1%	31.8%
Impact of SA on gap	42.0%	49.2%	***	44.2%	49.5%	45.5%
Both genders
All SARs in 1998	176,654	247,499	24,834	75,881	37,868	562,736
Low income gap:
- before any transfers	92.0%	94.2%	85.0%	90.7%	87.1%	92.1%
- after transfers, exc. SA	70.1%	87.1%	72.4%	69.2%	76.3%	78.0%
- after all transfers	28.3%	39.4%	23.7%	27.4%	29.5%	33.0%
Impact of SA on gap	41.8%	47.7%	48.7%	41.8%	46.8%	45.0%
(1) Sample of major income earners, age 16-55, who received SA in 1998. ***Less than 30 observations.

9.4 Characteristics of Social Assistance Recipients

Table 9.3 shows the distribution of lone mothers who received SA in 1998, by various characteristics. It also shows the incidence of SA among all lone mothers. As in the case of the incidence of low income, the incidence of SA was highest among those with no paid work during the year (83%). Non-earning SARs accounted for 54% of all lone mothers on SA.

The two other most common characteristics of lone mothers on SA were: (a) not being in a union when their first child was born; and (b) the presence of pre-school age child (44%).

The impact of various characteristics on the probability of receiving SA was explored using logit regression (Table 9.4).¹⁰ The regression results show that the following characteristics have the strongest correlation with the incidence of SA:

(a) recent immigrants/ aboriginal/ disabled: 35% of lone mothers on SA had one at least of these three characteristics; their probability of being on SA was 28% higher than that of the rest of lone mothers;

(b) student status: 25% of lone mothers on SA were students (full-time or part-time); their probability of being on SA was 27% higher than that of non-students with high school education; this group is of relatively less concern, however, since the expectation is that their earning capacity, at least on average, will be greater once they graduate;

(c) not in a union when their first child was born: the most common characteristics among all lone mothers on SA is that they were not in a union when their first child was born (73%); their probability of being on SA was 21% higher than that of the rest of lone mothers;

(d) under 20 years of age when their first child was born: 31% of lone mothers on SA were under 20 years of age when their first child was born; their probability of being on SA was 19% higher than for the rest of lone mothers;

(e) high school dropouts: high school drop-outs accounted for 30% of lone mothers on SA that were not students; their probability of being on SA was 18% higher than that for lone mothers with high school graduation; however, higher education is not a sure way to avoid dependence on SA; in particular, 34% of non-student SARs had a post-secondary certificate or degree.

Two other characteristics that were found to have a statistically significant impact in increasing the probability of receiving SA were:

(f) SA benefit rates: lone mothers living in provinces with above average SA benefit rates had a 16% higher probability of receiving SA, an indication of the connection between generosity of SA benefits and SA caseload; and

(g) regional employment rates: lone mothers in regions with below average employment rates had a 11% higher probability of receiving SA, an indication of the impact of labour market conditions on the SA take up rate.¹¹

Table 9.3 Characteristics of lone mothers on SA (SARs), 1998¹

	Distribution of SARs	Incidence of SA
Age
16-29	32.7%	68.7%
30-55	67.3%	33.5%
Age when first child was born
Under 20	31.2%	68.8%
20 or more	68.8%	34.2%
Marital status when first child was born
Not in a union	73.3%	53.8%
Married	22.9%	22.3%
Common law	3.8%	40.9%
Age of youngest child
0-5	44.4%	54.8%
6-11	36.1%	40.4%
12-17	19.5%	24.9%
Student during the year
Yes	25.0%	57.3%
No	75.0%	36.6%
Level of education of non-students
Less than high school	29.9%	57.1%
High school diploma	16.3%	33.2%
Some post-secondary	19.9%	47.8%
Post-secondary degree	33.9%	25.6%
Hours of work during the year
No work	53.7%	82.9%
1-749 hours	22.1%	72.1%
750-1499 hours	13.0%	37.4%
1500+ hours	11.3%	9.8%
Immigrant, aboriginal, or disabled 2
Yes	35.3%	62.5%
No	64.7%	33.7%
EI region employment rate
At, below average	43.4%	42.6%
Above average	56.6%	38.6%
Provincial SA benefit rates
At, below average	43.4%	36.2%
Above average	56.6%	44.0%
Broad region
Atlantic	13.0%	59.5%
Quebec	19.9%	31.2%
Ontario	42.1%	44.9%
Prairie	13.0%	33.8%
B.C.	12.0%	39.2%
All lone parents	100.0%	40.2%
(1) Sample of lone mother major income earners, age 16-55, in 1998. (2) Immigrated in last 10 years; or aboriginal origin; or work limiting disability.

Table 9.4 Logit regression estimate of determinants of incidence of SA among all lone mothers, 1998

Variable	Explanation	b-coef	Std err	t-stat.	Odds ratio	Linearized coefficient
Dependent
SAR98	Received SA in 1998
Age
GAGE(1)	- 16-29	0.428	0.199	2.146	1.534	10.1%
GAGE(2)	- 30-55		(omitted)
Age when first child was born
CAGE(1)	- 16-19	0.772	0.188	4.113	2.164	18.7%
CAGE(2)	- 20-55		(omitted)
Marital status when first child was born
CSPOUSE(1)	- did not have a spouse	0.964	0.146	6.609	2.623	21.2%
CSPOUSE(2)	- had a spouse		(omitted)
Age of youngest child
YKID(1)	- 0-5	0.140	0.169	0.828	1.150	3.4%
YKID(2)	- 6-11		(omitted)
YKID(3)	- 12-17	-0.534	0.173	-3.087	0.586	-12.0%
Level of education
STEDUC(1)	- student	1.106	0.279	3.964	3.021	26.8%
STEDUC(2)	- non-student: less than high school	0.762	0.237	3.215	2.143	18.4%
STEDUC(3)	- non-student: high school diploma		(omitted)
STEDUC(4)	- non-student: some post-second.	0.559	0.246	2.272	1.750	13.3%
STEDUC(5)	- non-student: post-sec. degree	-0.175	0.211	-0.829	0.840	-3.8%
Recent immigrant, aboriginal, or disability?
HIGHRISK(1)	- yes	1.138	0.159	7.136	3.121	27.6%
HIGHRISK(2)	- no (omitted)		(omitted)
Regional employment rate
REGER(1)	- at/below average	0.452	0.163	2.763	1.571	11.1%
REGER(2)	- above average (omitted)		(omitted)
Provincial SA benefit rates
SABEN(1)	- at/below average	-0.693	0.163	-4.256	0.500	-15.8%
SABEN(2)	- above average (omitted)		(omitted)
Constant		-1.491	0.231	-6.455	0.225
Nagelkerke R 2 (similar concept to OLS adjusted R 2)				0.336
Number of cases				1,234

9.5 Duration of Social Assistance

In this sub-section we use the 1993-98 longitudinal SLID data to see how many lone mothers are exposed to SA over a longer period and, once they are on SA, how long they tend to stay on SA.

Over the period 1993-98, 58% of lone mothers received SA in at least one year, while 24% of all lone mothers received SA in all six years over the period 1993-98. This means 41% of all SA recipients in 1993-98 received SA in all six years the highest rate for any type of family (Table 9.5).

The average "in-progress" SA spell for lone mothers was 4.3 years (Chart 9.3). The average completed SA spell is likely to be longer. The reason is that many of the spells may have started before 1993. Also, since 41% of SARs were on SA all six years, a significant number may have continued on SA past 1998.

Table 9.5 Duration of social assistance, 1993-98. Among all social assistance recipients (SARs)¹

	Lone parent, kids<18	Unattached individual	Couple without kids<18	Couple with kids<18	Other economic families	All economic families
Male
All major income recipients	96,912	1,078,953	1,263,516	2,298,221	313,254	5,050,856
At least one year on SA	***	26.9%	10.0%	9.8%	21.9%	14.6%
All six years on SA	***	6.2%	***	***	***	2.4%
Average years on SA	***	3.4	2.8	2.6	2.8	3.0
Average in-progress spell2	***	3.8	3.0	2.7	3.3	3.4
Female
All major income recipients	555,302	910,502	455,028	676,560	220,596	2,817,988
At least one year on SA	58.1%	21.4%	12.8%	16.6%	33.4%	27.0%
All six years on SA	24.1%	5.5%	***	***	***	9.0%
Average years on SA	4.4	3.4	2.9	3.7	3.7	3.9
Average in-progress spell2	4.3	3.8	***	4.4	3.9	4.1
Both genders
All major income recipients	652,215	1,989,455	1,718,543	2,974,781	533,850	7,868,844
At least one year on SA	54.1%	24.4%	10.7%	11.3%	26.7%	19.1%
All six years on SA	22.2%	5.9%	***	1.9%	6.5%	4.8%
Average years on SA	4.3	3.4	2.8	2.9	3.3	3.4
Average in-progress spell2	4.3	3.8	3.0	3.4	3.7	3.8
(1) Sample of major income recipients, age 16-55 in 1993. (2) SA spell of those who received SA in 1993. *** Less than 30 observations.

9.6 Determinants of the Duration of Social Assistance

In this section we analyze both in-progress SA spells (using a simple tabulation and an OLS regression), as well as the expected duration of new spells (using a "hazard" logit regression). The techniques are the same as those discussed and applied in Section 8.

The results show that the average in-progress SA spell is 4.3 years (almost a year longer than the average in-progress low income spell, which is 3.4 years). On the other hand, the average expected duration of new SA spells is 2.1 years, the same as for new low income spells.

The results suggest that in most cases, when lone mothers enter low income or SA, they stay for a short period (2.1 years on average). However, a number of lone mothers stay in low income or on SA for many years.

Both the OLS analysis of in-progress SA spells and the "hazard" analysis of new SA spells led to similar conclusions with respect to the following factors:

1. Change in family status: Both approaches identified this as one of the most important factors; the average in-progress SA spell and the average spell of new starts were about 1.5 years shorter than for the rest of lone mothers; this result is not surprising since, as was pointed out in an earlier section, changes in family status are one of the most significant factors behind low income exits.
2. High risk group: Hazard analysis shows that lone mothers who were also recent immigrants, disabled, or aboriginal, had longer SA spells by 1.5 years; the OLS approach did not confirm this but, as was discussed earlier in the context of low income spells, the most likely reason for this difference is that the OLS model does not properly handle variables that tend to change over time, such as disability status.
3. Age of youngest child: the OLS model shows that both the presence of pre-school age children and high school age children contributed to longer SA spells (by 1 and 1.7 years respectively), but most likely for different reasons:
   1. younger kids are a barrier to the employment of their mothers;
   2. however, this is less likely to be the case with older kids, suggesting that other barriers may be present such as the "scarring" effect of prolonged reliance on SA.¹²
4. Region: According to the OLS results, in-progress SA spells tend to be longest in Quebec; however, the "hazard" model shows that new spells tend to be longer for Ontario and shorter for Quebec and B.C.; these results suggest that most SA spells in Quebec tend to be short, but there is a core of long spells that show up year after year.

The rest of the characteristics such as student status, level of education, or level of provincial SA benefits did not have a statistically significant effect on in-progress SA spells.¹³

The most surprising result is that the level of education has no effect on the length of SA spells. It would appear that education has a positive effect in helping lone mothers stay out of SA. However, once on SA, the level of education makes no difference on how long they stay on SA.

Instead, the length of SA spells appear to be dominated by factors over which public policy has little effect such as change in family status; high risk characteristics; presence of young children; and region of residence. Although these characteristics cannot be influenced directly by public policy, they can be used as indicators for targeting programs for assisting lone mothers to exit SA.

Table 9.6 Length of completed SA spells by personal characteristics among lone mothers on SA, 1993-98

	Distribution of 1993 SARs	Length of average in-progress SA spell (years)
Age in 1993
16-29	38.5%	4.1
30-55	61.5%	4.5
Age of youngest child in 1993
0-5	55.4%	4.4
6-11	27.7%	3.8
12-17	16.9%	4.8
Student in 1993
Yes	22.0%	4.3
No	78.0%	4.3
Level of education of non-students in 1993
Less than high school	42.4%	4.5
High school diploma	19.8%	4.7
Some post-secondary	15.6%	4.2
Post-secondary degree	22.1%	3.8
Recent immigrant, aboriginal, or disability in 1993 1
Yes	31.2%	4.5
No	68.8%	4.3
Lone mother/major earner/child under 18 in all years
There was a change	50.9%	3.7
There was no change	49.1%	4.9
Moved to another region after 1993
Yes	13.3%	3.4
No	86.7%	4.5
EI regional employment rate in 1993
At/below average	38.0%	4.3
Above average	62.0%	4.3
Provincial SA benefit rates in 1993
At/below average	44.6%	4.1
Above average	55.4%	4.5
Atlantic	9.4%	4.4
Quebec	16.5%	5.0
Ontario	59.0%	4.5
Prairie	15.1%	3.5
B.C.	***	***
All	100.0%	4.3
(1) Immigrated in last 10 years; or aboriginal origin; or work limiting disability. *** Less than 30 observations.

Table 9.7 OLS Regression of determinants of completed SA spells, 1993-98 among lone mothers on social assistance in 1993

Variable	Explanation	B-coef	Std err	t-stat.
Dependent
SASPEL93	Length of completed SA spell (yrs)
Age in 1993
GAGE93(1)	- 16-29	-0.349	0.231	-1.508
GAGE93(2)	- 30-55 (omitted)
Age of youngest child in 1993
YKID93(1)	- 0-5	1.035	0.246	4.207
YKID93(2)	- 6-11 (omitted)
YKID93(3)	- 12-17	1.707	0.324	5.269
Level of education in 1993
STEDUC(1)	- student	-0.146	0.316	-0.463
STEDUC(2)	- non-student: less than high school	-0.046	0.300	-0.154
STEDUC(3)	- non-student: high school diploma (omitted)
STEDUC(4)	- non-student: some post-second.	0.057	0.372	0.152
STEDUC(5)	- non-student: post-sec. degree	-0.482	0.352	-1.369
Recent immigrant, aboriginal, or disability in 1993
HIGHRISK(1)	- yes	0.100	0.211	0.471
HIGHRISK(2)	- no (omitted)
Lone mother/major earner/child under 18 in all years
FAMILYF(1)	- there was a change	-1.508	0.213	-7.078
FAMILYF(2)	- there was no change (omitted)
Moved to another	region after 1993
FEIR(1)	- yes	-0.661	0.287	-2.305
FEIR(2)	- no (omitted)
Broad region
REGION(1)	- Atlantic	0.045	0.341	0.132
REGION(2)	- Quebec	0.522	0.285	1.832
REGION(3)	- Ontario (omitted)
REGION(4)	- Prairie	-0.651	0.290	-2.245
REGION(5)	- B.C.	-0.600	0.404	-1.485
Constant		4.591	0.315	14.575
Adjusted R-squared Number of cases			25.4% 294

Table 9.8 "Hazard" analysis of Social Assistance spells, 1993-98

In all cases, the regression variables reflect the status at the end of each spell (time dependent)	Logitcoeff.	Stand.error	t-stat		Exit rates by duration of SA (yrs)					Median spell
				1	2	3	4	5	6
Dependent variable: probability of exiting SA
Change in lone motherhood status
there was no change in status	-0.747	0.288	-2.594	24.4%	19.8%	15.8%	12.6%	9.9%	7.7%	3.2
status changed (omitted)				40.5%	34.2%	28.4%	23.3%	18.8%	15.1%	1.5
Was recent immigrant; disabled; aboriginal
yes	-0.785	0.307	-2.557	19.0%	15.2%	12.0%	9.5%	7.4%	5.8%	5.2
no (omitted)				34.0%	28.2%	23.1%	18.7%	14.9%	11.8%	1.9
Age of lone mother
16-29	-0.717	0.354	-2.025	19.1%	15.3%	12.1%	9.5%	7.5%	5.8%	5.1
30-55 (omitted)				32.6%	27.0%	22.0%	17.8%	14.2%	11.2%	2.0
Age of youngest child
0-5	-0.090	0.303	-0.297	27.1%	22.2%	17.9%	14.3%	11.3%	8.8%	2.7
6-17 (omitted)				29.0%	23.7%	19.2%	15.4%	12.2%	9.6%	2.4
Level of education
student	0.685	0.498	1.376	38.2%	32.0%	26.5%	21.6%	17.4%	13.9%	1.6
non-student: less than high school	-0.083	0.478	-0.174	22.3%	18.0%	14.3%	11.3%	8.9%	6.9%	3.8
non-student: high school diploma (omitted)				23.7%	19.2%	15.4%	12.2%	9.6%	7.5%	3.3
non-student: some post-second.	-0.179	0.531	-0.337	20.6%	16.6%	13.2%	10.4%	8.1%	6.3%	4.4
non-student: post-sec. degree	0.524	0.454	1.154	34.4%	28.6%	23.5%	19.0%	15.2%	12.0%	1.8
Region
Atlantic	0.392	0.443	0.885	26.2%	21.4%	17.2%	13.7%	10.8%	8.5%	2.8
Quebec	1.000	0.360	2.778	39.5%	33.3%	27.6%	22.6%	18.2%	14.5%	1.5
Ontario (omitted)				19.4%	15.5%	12.3%	9.7%	7.6%	5.9%	5.0
Prairie	0.603	0.405	1.489	30.5%	25.1%	20.4%	16.4%	13.0%	10.3%	2.2
B.C.	1.453	0.466	3.118	50.7%	44.0%	37.5%	31.4%	25.9%	21.1%	1.0
Spell duration in years (continuous independent)	-0.269	0.151	-1.781
Constant	-0.290	0.576	-0.503
Dependent variable: probability of exiting SA				32.2%	24.4%	18.0%	13.0%	9.3%	6.5%	2.1
Nagelkerke R2 (similar concept to the OLS adjusted R2):		16.7%		Number of cases:			369

• ¹⁰The results of an alternative specification of the logit regression (replacing EI regional employment rates and provincial SA benefit rates with dummies for the five major regions) are shown in Appendix D.
• ¹¹As explained earlier, the employment rate was calculated by dividing the weeks of work of each woman (0 to 52) by 52. The ratio was averaged within each of the 54 regions that are designated by the EI program.
• ¹²In the "hazard" model we tested only the presence of pre-school children. We found no significant effect. The most obvious reason is that the positive effect of the youngest child being 6 to 11 years of age, was offset by the negative effect of the younger child being 12 to 17 years of age.
• ¹³The results with respect to provincial SA benefit rates are based on a similar regression; the results of this regression are not shown here.