The Labour Participation Response of Mothers to Changes in Early Childhood Education Costs

The labour participation response of mothers to changes in early childhood education costs

Selective review of the evidence

July, 2012

Contents

1.Main themes

1.1. Purpose and scope of this review

1.2. Effective tools for raising maternal employment

1.3. Effects of child care prices on labour participation

1.4. Cost-benefit analyses of ECE subsidisation

1.5. New Zealand context

2.Evidence from international comparisons

2.1. Jaumotte (2003) Female labour force participation: past trends and main determinants in OECD countries

3.Evidence from Australia

3.1. Gong, Breunig and King (2010) How responsive is female labour supply to child care costs – new Australian estimates

4.Evidence from the United States

4.1. Kimmel (2006) Child care, female employment, and economic growth

4.2. Herbst (2010) The labor supply effects of child care costs and wages in the presence of subsidies and the earned income tax credit

4.3. Fitzpatrick (2010) Preschoolers enrolled and mothers at work? The effects of universal prekindergarten

4.4. Tekin (2007) Childcare subsidies, wages. and employment of single mothers

5.Evidence from Canada

5.1. Lefebvre, Merrigan and Verstraete (2009) Dynamic labour supply effects of childcare subsidies: evidence from a Canadian natural experiment on low fee universal child care

5.2. Lefebvre, Merrigan and Roy-Desrosiers (2011) Québec’s childcare universal low fees policy 10 years after: effects, costs and benefits

6.Evidence from Norway

6.1. Black, Devereux, Løken and Salvanes (2012) The effect of child care subsidies on student performance

7.New Zealand context

7.1. Statistics New Zealand (2010) New Zealand childcare survey 2009

7.2. Treasury (2010) Labour market and broader outcomes. Introductory briefing to the ECE Taskforce

8.Other studies, in brief

9.Bibliography

10. Glossary of terms

1. Main themes

1.1. Purpose and scope of this review

This is a selective review of the evidence on the link between the cost and availability of early childhood education (ECE), and parents’, particularly mothers’, participation in work. This review is intended to inform policy advice on what impact changes in ECE costs would be likely to have on New Zealand mothers’ participation in work.

This review was selective, examining a number of studies, and selecting the most relevant, robust and recent, for discussion. A wider range of studies are briefly described, at the end of this report.

1.2. Effective tools for raising maternal employment

Child care subsidy is a policy that has been found, on a macroeconomic level, to be effective in stimulating female labour force participation. Other policies that have also been found to be effective include neutral tax treatment of second earners in a household, tax incentives to share market work between spouses, and paid parental leaves. Child benefits that are un-linked to child care participation are not effective in stimulating female labour force participation.

1.3. Effects ofchild care prices on labour participation

There is a general consensus that labour force participation among mothers of pre-school children is sensitive to child care costs, with elasticities usually in the range of -0.1 to -0.8. That is, for every 1% increase in the cost of child care, maternal employment rates drop by 0.1 to 0.8%.

At least one study (Lefebvre et al., 2009) shows that the increase in maternal employment stimulated by lower child care costs can be long lasting: mothers remained employed at higher rates than they would have been otherwise, even after their children had graduated from child care.

Decreases in the cost of child care (for example through provision of subsidies) consistently have a greater impact on child care use than on maternal employment, even when the subsidies include a requirement that the parent be employed. This may be in part due to substitution effects, where mothers who are already working are encouraged by the subsidy to move their children from unpaid informal care to paid formal care.

A number of contextual factors influence the extent to which maternal labour force participation responds to changes in child care prices. These factors include the following.

• Maternal education levels and earning potential:morehighly educated mothers are more likely to be employed, presumably due to their higher earning potential, and their labour participation is less sensitive to changes in child care costs. That is, they are less likely than low education mothers, to exit employment when child care costs increase.
• Non-earned and household income: there is some evidence that mothers from higher income households are more sensitive to changes in child care costs, presumably because income earned by a partner, or via non-wage sources such as investmentsprovides the freedom to exit the labour force when child care costs increase. However, this effect may be small.
• Age of youngest child:in investigating the effects of child care costs on maternal employment, a number of international studies have included after-school care costs for older children, as well as early childhood care costs for pre-schoolers. Studies consistently find that mothers of pre-school children are more sensitive to child care costs than mothers of school-age children.
• Sole versus partnered parents: evidence is mixed on the differential responses between sole and partnered parents. The labour participation of both groups appears to besensitive to child care costs, but few studies compare the two, and comparisons between studies are difficult due to methodological and contextualdifferences.
• Type of working arrangement: the evidence is mixed on whether employment rates for part-time or full-time workers are more responsive to changes in child care costs. Some studies find that part time employment rates are more sensitive to child care costs, while others find that full time employment rates are more sensitive. This phenomenon may be context-dependent. For example, a subsidythat promotes full-time child care may have more of an impact on full-time employment decisions than on part-time work decisions.
• Under-supply of child care: any labour force response to a decrease in child care costs will be suppressed if the supply of child care places does not increase to meet the extra demand. This issue has been encountered in a number of studies.
• Universal versus targeted child care subsidies: the evidence on whether universal child care policies increase maternal employment is mixed, with some studies showing universal subsidies have a very large effect, and others showing little or no effect. The effects of a universal child care policy may depend on other factors, such as the extent to which targeted subsidies are also available.
• Shrinking elasticities: some of the more recent studies have found smaller female labour supply responses to child care prices than they were expecting. The authors speculate that, with the overall increase in female workforce participation, the ability of child care prices to influence maternal work decisions at the margin has decreased; mothers are now more likely to be working, regardless of child care price. Other work has demonstrated recent reductions in female wage response elasticities. It should be noted that this phenomenon is currently only found in a small number of studies, andthere are other contextual and methodological issues that may account for the smaller elasticities.
• Macroeconomic context: strong economic growth is likely to facilitate a positive employment response to decreased child care prices, as work is more available to mothers who are induced by the lowered prices to seek work.

In addition to the differences in context between studies, comparison between studies is compromised by the different methods that they use, and the resulting differences in their estimates of effects. Because of this and the complicated relationship between context and labour responses to child care prices, comparisons across jurisdictions and across time periods are difficult. It would be inadvisable to attempt to predict New Zealand responseson the basis of the findings from any one comparison country.

1.4. Cost-benefit analyses of ECE subsidisation

There is little robust evidence available on the cost-benefit or cost-effectiveness of child care subsidies. Estimates are very dependent on context, and on the variables that are chosen for inclusion in the equation.

One study (Tekin, 2007) has found that child care subsidies are more cost-effective than wage subsidies, in terms of extra hours of work generated, per dollar of subsidy. While wage subsidies are associated with a larger increase in the hours worked per percentage wage increase, in this study, child care subsidies were more tightly targeted and provided to fewer women, so were of lower cost, per extra hour generated.

Another study (Lefebvre et al, 2011) estimates the cost-benefit of a universal low cost child care policy, and finds that the revenue returned to the government, in terms of tax gains and welfare savings, fell short of the costs of the programme. This estimate is very specific to the policy contextand to the assumptions and variables incorporated into the cost-benefit equation. The finding is therefore not generalisable, but it does demonstrate that a short term benefit to government revenue cannot be assumed.

1.5. New Zealand context

No empirical evidence on New Zealand labour supply responses to ECE costs has been found.

While there was a sharp drop in ECE costs in 2007, due to the introduction of the ‘20 hours ECE’ policy, the publicly available data shows no associated detectable increase in labour force participation. However, this does not mean that there was no effect, as a number of other factors influenced employment at this time. In particular, the economic downturn may have masked any effects of ’20 hours ECE’. More detailed econometric work would be needed to investigate whether the policy had any impact on labour force participation.

Data on labour force participation and child care use shows that New Zealand has low employment rates among sole mothers, as compared to other OECD countries. In addition, a significant minority of parents who worked or wanted to work, especially mothers, report work difficulties relating to child care use, including having to turn down, or stop searching for paid work. There may therefore be room for New Zealand to increase its maternal labour force participation by addressing the child care-related constraints on employment.

1. Evidence from international comparisons

2.1. Jaumotte (2003) Female labour force participation: past trends and main determinants in OECD countries

Jaumotte (2003) used multivariate analysis across 17 OECD countries to assess the role of various policies and labour market factors in determining the pattern of female labour participation rates among 25-54 year old women. The time period covered is 1985-1999, although most countries only had data for a sub-set of those years.

Jaumotte (2003)found that, on this macroeconomic level, policies that stimulate female labour force participation included:

• a more neutral tax treatment of second earners
• tax incentives to share market work between spouses
• child care subsidies
• paid parental leaves.

Jaumotte found that child care subsidies and paid parental leaves tended to stimulate full-time rather than part-time participation. Child benefits, that is, per-child subsidies that were not linked to child care, had a negative effect on part-time female participation.

The following other issues were also found to correlatepositively with female participation.

• The prevalence of part time work (used as a proxy indicator of the flexibility of the country’s labour market; although this measure suffers from endogeneity[1]).
• Higher female education levels.
• Low overall unemployment.

In addition to the multivariate analysis, Jaumotte (2003) provides a useful review of the factors that are thought to underlie the increase in female labour force participation that occurred from 1981 to 2001. This review includes a summary of the pre-2003 evidence for the effects of child care subsidies on female labour supply, making the following main points.

• While child care subsidies reduce the relative price of child care and, therefore, increase the relative return of market work, child benefits have only an income effect and may lead to a reduction in labour supply.
• A number of studies have found negative elasticities[2]of female employment to child care costs. That is, employment decreases as child care costs increase.
• Studies vary in their estimates of the size of the effects of child care support on female labour participation, with elasticities ranging from 0.06 to -1.24 for the United States, and with some argument that the more plausible values are within the range of -0.1 to -0.4.
• There is evidence that elasticities may be higher for low income/less skilled mothers. That is, low income mothers are more likely than high income mothers to decrease their use of child care as the cost increases.
• The evidence on whether elasticities vary by marital status is mixed, but elasticities are consistently higher for mothers of younger children, as compared to mothers of older children.
• The effectiveness of child care support in encouraging labour force participation is compromised by substitution effects. That is, it may encourage working mothers to move their children from unpaid informal care to paid formal care. The increases in labour supply that are seen in response to child care subsidies are generally smaller than the increases in child care use.
• There is some mixed evidence on the direction of causation. For some countries such as the United Kingdom and the United States, there is strong evidence of causality from child care support to female participation, while in other countries, such as Denmark, the sequencing of policy changes and labour market changes appears to suggest a reverse causality.

1. Evidence from Australia

3.1. Gong, Breunig and King (2010) How responsive is female labour supply to child care costs – new Australian estimates

The degree of responsiveness of Australian women’s labour supply to child care cost has been a matter of debate, with four studies, published prior to 2010, finding that the level of responsiveness is very low or negligible (Doiron and Kalb, 2005; Rammohan and Whelan, 2005, 2007; Kalb and Lee, 2008). In the most recent paper on this topic, Gong et al.(2010) review this literature, and find that the previous papers suffered from measurement error arising from the following issues.

• Limited data on actual childcare prices, and the use of estimates derived from average child care costs divided by the number of hours worked. This leads to endogeneity, because the estimated child care prices are dependent on the parent’s hours worked: the variable that they are supposed to explain.
• Small sample sizes.
• The use of data that presents child care costs, net of government subsidies. Again this leads to endogeneity, as subsidies are partly determined by parents’ labour supply.
• The use of aggregated and noisy measures of average child care prices at the state level, which do not capture the variation in local market prices, to which households react, when making child care and labour decisions. This adds a large amount of measurement error.

Gong et al.(2010)undertake a new analysis of Australian labour supply and child care demand elasticities with respect to child care price, using data from the 2005-2007 ‘Household, Income, and Labour Dynamics in Australia’ (HILDA) survey. This data provides more detail than has been available previously on child care use and costs. The authors use an econometric model with a finer treatment of child care prices, addressing measurement error problems and estimatinggross costs relative tothe actual child care usage, tackling the endogeneity problems. Their analysis is restricted to married mothers only, as the sample size for sole mothers was too small.

Contrary to the prior Australian work, Gong et al.(2010) find that the cost of child care does have a statistically significant negative correlation with the labour supply of married mothers with young children. For an average mother, the elasticity of employment with respect to the gross child care price was -0.3 (for every 1% increase in the price of child care, the employment rate of those women dropped by 0.3%), and the elasticity of hours worked was0.7 (for every 1% increase in the price of child care, the hours worked by those women dropped by 0.7%). These findings are closer to international estimates, than those of Doiron and Kalb, 2005; Rammohan and Whelan, 2005, 2007; and Kalb and Lee, 2008.

Interestingly, Gong et al.(2010) found an indication that the price elasticity of child care demand was greater for higher income households (mothers from higher income households were more likely to reduce their use of paid child care, when the price increased), while the price elasticity of labour supply was smaller for lower income households (mothers from lower income households were less likely to decrease their labour participation in response to increased child care prices). This suggests that an increase in the price of child care may have a greater negative impact on the labour force participation of women from higher income households, while women from lower income households are more likely to continue to work, but to substitute in lower cost, informal child care options. However,it should be noted that this variation in responsiveness was small, and the confidence intervals were wide.