Child Maintenance and Social Security Interactions: the Poverty Reduction Effects in Model Lone Parent Families across Four Countries

Abstract

In most developed countries, children in lone parent families face a high risk of poverty. A partial solution commonly sought in English-speaking nations is to increase the amounts of private child maintenance paid by the other parent. However, where lone parent families are in receipt of social assistance benefits, some countries hold back a portion of the child maintenance to reduce public expenditures. This partial ‘pass-through’ treats child maintenance as a substitute for cash benefits which conceivably neutralises its poverty reduction potential. Such neutralising effects are not well understood and can be obscured further when more subtle interactions between child maintenance systems and social security systems operate. This research makes a unique contribution to knowledge by exposing the hidden interaction effects operating in similar child maintenance systems across four countries: the United Kingdom, United States (Wisconsin), Australia and New Zealand. We found that when child maintenance is counted as income in calculating benefit entitlements, it can reduce the value of cash benefits. Using model lone parent families with ten different employment and income scenarios, we show how the poverty reduction potential of child maintenance is affected by whether it is treated as a substitute for, or a complement to, cash benefits.

Introduction

Children living in lone parent families face a high risk of poverty across many countries (OECD, 2011). In most English-speaking nations, poor lone parent families are eligible to receive social assistance benefits and, in most instances, private transfers from the other parent, known as child maintenance or child support payments. How countries treat income derived from private child maintenance payments has important implications for the reduction of child poverty, particularly for the poorest lone parent families who receive social assistance benefits. But some countries view child maintenance as a substitute for social assistance as some, or all, of the maintenance payments are held back to reduce public expenditures. That is, when a separated parent pays private child maintenance, the lone parent family either does not receive the full amount or receives less in public social security benefits. Thus, the poverty reduction potential of child maintenance is reduced. Alternatively, some countries treat child maintenance as a complement to social assistance and pass-through the full value to eligible families without reducing social assistance payments, thereby increasing household income and enhancing child maintenance's poverty reduction potential. But, despite the importance of this issue, we currently know very little about the interactions between child maintenance and social security systems and therefore our knowledge of the relationship between child maintenance and child poverty is limited.

In this paper, we begin to explore the anti-poverty effectiveness of child maintenance by conducting new research comparing the policy approaches in the United Kingdom, United States (Wisconsin ¹ ), Australia and New Zealand. These four countries were chosen as they share a common policy ancestry having built on each other's child support policies. This provides a solid foundation for comparable analysis. However, these countries’ systems are both alike and unlike in key respects and this facilitates cross-national learning to offer important policy insights.

We use a vignette technique which creates a fictitious lone parent model family, whose circumstances we vary in order to investigate the amount of child maintenance and other cash benefits the family receives. Through the vignettes, we highlight the extent to which both explicit and implicit interactions limit the anti-poverty effectiveness of child maintenance. One explicit interaction is apparent in the ‘pass-through’ mechanism and relates to the social assistance programme whereby some or all of the maintenance is held back by the state (treating maintenance as a substitute). However, we found other, more subtle interactions, when child maintenance is counted as income for the purposes of calculating cash benefits. Here too, child maintenance could be treated as a substitute, reducing its anti-poverty effectiveness. Ultimately, our findings reveal why policy makers and comparative research analysts need to understand these interactions and their potential effects on poverty reduction strategies that rely on private child maintenance payments.

Background

Comparing countries

The four English-speaking nations used in this comparative research share a common heritage by being clustered into liberal ‘welfare regime types’ by Esping-Andersen in ‘Three Worlds of Welfare Capitalism’ (Deeming, 2016). Typically this meant they had similarly residual or social assistance models of welfare provision in which claimants were often stigmatised and entitlements to social security benefits were strictly controlled. The social protection function of the welfare state to provide adequate incomes through public means was therefore minimal. The extent to which this typology holds true has come under constant scrutiny since Gøsta Esping-Andersen published ‘The Three Worlds of Welfare Capitalism’ in 1990.

Deeming (2016) argues that such liberal welfare regime classifications have always been problematic for these four countries because, among other things, not all social programmes are alike (the National Health Service in the UK, for example, fits better with a social democratic regime type) and because there have been long periods of social democratic government producing conflicting policy prescriptions. Additionally, Deeming (2016) highlights how neoliberal conceptions have since transformed the welfare state into a ‘workfare state’. Commonly work, not welfare, is seen as the ‘best route out of poverty’. Social investment in workfare activation policies (that facilitate return to work or deliver welfare sanctions if people fail to comply) has now become the main business of social policy, rather than providing adequate levels of social protection for the vulnerable and unemployed. Certainly, Deeming and Smyth (2015) illustrate how the UK, US, Australia and New Zealand tend to have the lowest unemployment benefit replacement rates among OECD countries. In particular, for lone parent families with two children, all four countries had the lowest rates varying between about 40 per cent (UK) to 48 per cent (Australia). In comparison, Switzerland had an 88 per cent benefit replacement rate.

While the four countries in this paper were not chosen to test debates on welfare state typologies, this research can nonetheless make a contribution to that literature. As Deeming (2016:13) posits, it is better to focus on specific social programmes and their ‘cultural and historical antecedents’ than try to cluster whole countries into a regime type. Through the lens of child maintenance policies, this paper takes a traditional social protection approach to analyse the contribution of child maintenance to lone parent poverty (as opposed to a social investment approach which would have compared activation policies and earnings as the key means to reduce poverty). The countries were therefore chosen on the basis of their common ancestry regarding child maintenance policy development providing a solid foundation for comparable analysis. We now set the context for the research by reviewing the evidence on the poverty reduction potential of child maintenance across countries.

Lone parents, child maintenance and poverty

The overall poverty reduction effect of child maintenance payments for low-income lone parent families depends, among other things, on the amounts paid but also on the rate of receipt among the whole lone parent population (Bradshaw, 2006). Three of our four countries have national studies on this. In the UK, Skinner and Main (2013: 56) conducted a secondary analysis of lone mother households using the Family and Children Study wave 2008–09. They found that 34.3 per cent of eligible lone mothers reported receiving child maintenance and, of those that did, 14 per cent were lifted out of poverty ² . Whereas, Bryson et al. (2013: 9) conducted a dedicated survey of lone parent families in receipt of social assistance benefits and they estimated that, while 36 per cent received child maintenance payments in 2012, payments had the effect of lifting 19 per cent above the poverty threshold. In the US, Meyer and Hu (1999) found that, among lone mother families who receive child maintenance, over 20 per cent were brought above the poverty threshold. Overall, however the anti-poverty effect of child maintenance is limited, only one-third of separated parents received any formal payments (Grall, 2016). In Australia, Cook et al. (2015) conducted a secondary analysis of the ‘Household Income and Labour Dynamics in Australia’ (HILDA) panel survey (wave 11, 2011). They found that over half (55 per cent) of lone parent families received child maintenance in the previous year, and lone parent poverty was reduced by 21 per cent when child maintenance was received. This rate of receipt is comparable to that found in the Australian Longitudinal Study of Separated Families (Qu et al., 2014). In New Zealand, there are no administrative data showing the proportion of all lone parent families in receipt of child maintenance but research by Chapple and Cronin (2007) using aggregate 2005 data, concluded that ‘even under the most favourable assumptions, the impact of [the] child support scheme on poverty is almost certainly comparatively small’. These ‘within country’ studies show the important effects of child maintenance payments on poverty reduction, but their methods are not similar, making comparisons difficult. Some systematic international comparisons have, however, been conducted using the Luxembourg Income Data set (LIS) (Hakovirta, 2011; OECD, 2011).

Hakovirta's (2011) analysis of LIS data used a 60 per cent median income threshold to define poverty and found large relative poverty reduction effects for those lone parent families who received maintenance in the UK, Denmark, Norway and Sweden. Here, the proportion of children living in poverty before receiving maintenance decreased by between 41 and 50 per cent after maintenance receipt. The US fared less well, with an 18 per cent reduction in the proportion of recipient lone parent families living in poverty. The OECD's latest analysis of LIS (2011) found a similar ranking of countries with the largest poverty reduction effects occurring among lone parent families in Nordic countries, Poland and Switzerland. In the USA, UK and Australia, child maintenance reduced the proportion of lone parents living in poverty by 7.2, 8.7 and 19.8 per cent, respectively. However, these findings are not directly comparable with Hakovirta (2011), as they include all lone parents with a child under 18, rather than only those in receipt of maintenance, and rely on a 50 per cent of median income poverty threshold. The key point from these analyses is that, while the study countries have similar child maintenance systems, the poverty reduction outcomes vary considerably.

Overall, it is difficult to gauge the total effect of child maintenance on poverty reduction for lone parent families. The LIS data facilitate comparisons but they are now dated and the methods are not ideal. For example, consider a country in which child maintenance counts as income and thus lowers the value of benefits from an income-support programme. Lone parents responding to a survey would presumably only report the (lowered) amount of income support benefits, so child maintenance effects (of lowering these other benefits) would be invisible. It is therefore not possible from self-report survey data to see the potential interaction effects between child maintenance and social security systems which might reduce the value of maintenance ³ . Uniquely, the new research reported here tries to expose those interaction effects by using a model family approach to gauge the value of child maintenance payments to lone parents’ household income after various interactions with social security systems are taken into account. Before describing the methods in detail, we set out the policy context in each country; providing a brief description of the child maintenance systems and the relationships between child maintenance and a range of social security benefits.

Policy Context

All four countries have similar child maintenance histories: in the US new legislation was passed in 1974 and the other countries followed in the 1980s–90s. The UK's 1991 Child Support Act closely mirrored both Australia's child support scheme introduced in 1988 and the US state of Wisconsin's scheme (Millar and Whiteford, 1993). The New Zealand 1991 Child Support Act tended to follow Australia. Not surprisingly, these systems are similar, with the UK, Australia and New Zealand operating administrative, as opposed to judicial, systems using standardised formulae to calculate child maintenance obligations. Wisconsin operates a hybrid scheme whereby either the courts or the child maintenance agency can set orders (Skinner et al., 2007). All countries have provision for collection and enforcement by a dedicated child support agency, but none guarantee child maintenance through public funds.

Table 1 describes the major social security benefits available to lone parents in each country and how these interact with child maintenance. The benefits and conditions of eligibility are very complex, but we have grouped them simply as follows: the main social assistance means-tested benefit in each country (for example, Income Support or Jobseekers Allowance in the UK ⁴ ); other child/family related benefits; work-conditional tax credits; and housing benefits. The two types of interactions relevant for our analysis are also shown. First, the ‘pass-through’ mechanism which explains how much of the paid child maintenance the recipients of a main social assistance benefit is effectively permitted to keep. Second, the extent to which child maintenance is counted as the income of the receiving parent when calculating other benefit entitlements. Where it is counted as income, it means that the state is effectively ‘clawing-back’ some (or all) of the paid child maintenance that it has already passed through. We refer to this as the ‘clawback’ mechanism.

TABLE 1. Social security benefits available for lone parents and their interactions with child maintenance

Notes: In Australia, PP, FTB(A) and FTB(B) include an Energy Supplement to compensate low-income earners for electricity price rises incurred as a result of the Carbon Tax.

Initially in the UK, US and New Zealand, none of the paid maintenance was passed through to recipients of the main social assistance programmes (except in New Zealand for a small number of cases when the maintenance paid was greater than the benefit, in which case the excess was passed through). Now, since 2010, the UK passes 100 per cent through. The US has set variable ‘pass-through’ rates across states since 1996 for recipients of Temporary Assistance for Needy Families (TANF). In Australia, since inception, all of the collected child maintenance has been passed through without any reductions to the main social assistance benefits. The picture regarding the means-testing of other benefits and clawback of child maintenance is more complex. In the UK, child maintenance does not count as income for any other benefits (bar Housing Benefit in some Local Authorities). In the US, clawback can occur in some benefit programmes but is disregarded in others. In Australia, child maintenance may affect Family Tax Benefit (Part A) and Rent Assistance benefits. In New Zealand, child maintenance is counted as income in respect of all major social assistance programmes for which lone parents may be eligible. The extent of the clawback therefore depends on the receiving parent's income and family circumstances. Pass-through and clawback mechanisms expose the possible interactions between social security programmes and child maintenance schemes and form the main part of our analysis described in the next section.

Data and methods

General approach

A model family approach (sometimes using vignettes) is frequently applied in comparative policy research (e.g. Bradshaw et al., 1996; Meyer et al., 2011; Skinner et al., 2007). Vignettes are intended to highlight how policy works in particular hypothetical family situations and thereby facilitate cross-country comparisons by simplifying reality. However, limitations occur as they only approximate reality and cannot account for all the complexities and interactions that do take place. Even so, they are still useful in exposing operational mechanisms in institutional systems.

We have created a simple model family of two separated parents (Mary and Paul) who have one preschool child and we vary their income and employment circumstances in ten different scenarios. The child resides solely with Mary and neither parent has a new partner or other children. While other model families with more complex circumstances could have been used (with shared care for example, or parents with new partners and children), the strength of our approach is that it holds family circumstances steady allowing us to test a range of earnings and employment circumstances and isolate the complex interactions between child maintenance schemes and a range of cash benefits. This offers a deeper comparative understanding of how the clawback and pass-through mechanisms might operate. However, the breadth of the analysis is reduced in terms of exploring different family types. Other model family studies have examined a wider range of family circumstances (Meyer et al., 2011; Skinner et al., 2007) but they tended to keep income constant in order to reduce complexity. In this study we take the opposite approach, which is similar to Summerfield et al. (2010) who used a single model family that varied income across a range of scenarios. Also for simplicity, we exclude health service subsidies and ignore activation policies (child care subsidies, parental leave provisions). Since the analysis is not focused on ‘workfare’, it is reasonable to ignore activation policies that support parental employment and also to ignore leave payments which are likely to provide a relatively short-term income source. Also, as health and childcare subsidies tend to be given to pay for service use, they do not necessarily add a surplus to the income package and so excluding them should not diminish the analysis.

In this paper, we are particularly interested in analysing the potential of child maintenance to reduce poverty in lone parent families who are either unemployed or on low earnings. Therefore, Mary's earnings vary from being unemployed to having typical female part-time employment, to having a full-time job that provides two-thirds of median female earnings ⁵ . These scenarios allow us to calculate Mary's eligibility for benefits at three different points. Paul also has three levels of earnings: unemployed, two-thirds of median male earnings, and median male earnings. The various combinations of these benefits and earnings allow us to examine child maintenance expectations in ten scenarios to explore the ways that maintenance affects Mary's income package.

We conduct the analysis in three main ways: first, we apply the child maintenance rules in each of the four countries to work out Paul's child maintenance obligation. Second, we explore what happens if that amount is actually paid to see how much of the maintenance Mary receives after taking account of the pass-through mechanism. Third, we examine whether the maintenance is counted as income in calculating amounts of other benefits and whether Mary's household income is above a poverty threshold. This approach illustrates two things: how child maintenance is treated within the main social assistance programme for lone parent claimants, and how benefits might change when child maintenance is paid. This reveals the interaction effects and the extent to which the child maintenance schemes can ameliorate poverty, which we calculate with more precision than would have been the case if we had simply added child maintenance amounts to income.

Table 2 shows the ten different scenarios we devised for the analysis. In scenarios A–D, Mary is unemployed, but Paul's employment and earnings’ status varies from being unemployed (scenarios A and B), to becoming employed with two-thirds median male earnings (scenario C), to becoming employed with median male earnings (scenario D). Paul's increasing earnings are repeated across scenarios E–J, but this time Mary moves into work, receiving median female part-time earnings (scenarios E–G), and then receiving two-thirds median female full-time earnings (scenarios H–J). This allows us to analyse how much child maintenance is expected for each of the ten scenarios. In addition, by including scenarios without child maintenance (A, E, H), we can easily see how child maintenance combined with earnings and other benefits could lift Mary out of poverty.

TABLE 2. Income in the ten scenarios in the vignette family of Mary and Paul

*Scenarios A, E, and H are used simply to show the base line level of Mary's total income without any child maintenance being paid by Paul.

Data and measures

The data for this study consist of the child maintenance amounts (calculated from formulae used across countries in 2015) and the levels of annual income that result from the various scenarios. Calculations have been done based on publicly available benefit information and tax forms. We mainly use two measures of annual income: gross, and ‘net income after housing costs’. We first calculate Mary's gross income, which includes earnings, child maintenance, and cash benefits. We calculate the total of all cash benefits for which she is eligible, both those paid directly and those paid through the tax system. We then make two subtractions from gross income to give a measure of ‘net income after housing costs’. First we subtract payroll and income taxes, and then we subtract the net housing costs Mary would have if she lived in private rental housing (assuming a two-bedroom apartment in an urban area) and received any housing subsidies for which she was eligible.

To determine whether Mary is above or below the poverty line, we compare her total ‘net income after housing costs’ to a poverty threshold using 60 per cent of equivalized median gross income minus taxes (an available threshold used in comparative studies that is the basis for the UK and EU poverty line (e.g. Belfield et al., 2014; Förster, 2005)). We use Mary's ‘net income after housing costs’ even though the poverty threshold does not consider housing because of the importance of housing costs and subsidies across countries ⁶ . (Analyses conducted using Mary's gross income less taxes and ignoring net housing costs led to similar conclusions.) All figures in the analysis are shown in Purchasing Power Parity (PPP) adjusted US dollars for 2014. Table 3 shows the figures we used for calculating household incomes. They are based on the latest available earnings and housing-cost data, and benefit and tax schedules in effect at 2014 or 2015 (depending on availability). For each country it gives: the median values of male and female full-time earnings and female part-time earnings; ‘typical’ rental costs for a two-bedroom apartment, and the value of the 60 per cent poverty threshold.

TABLE 3. The country values of median earnings, housing costs, and poverty thresholds used in calculating total incomes in the 10 scenarios in the vignette family

Notes: The first figure in each cell is in each country's own currency; the second figures (italicized) are in purchasing power parity-adjusted US dollars for 2014 (authors’ calculations).

Source for earnings values: UK Office for National Statistics, 2013; US Bureau of Labor Statistics, 2013; Australian Bureau of Statistics, 2013; Statistics New Zealand, 2014.

Rental Values: these were calculated separately for each country based on ‘typical’ mid-range private rental costs for a two bedroomed apartment in an urban area.

Poverty Threshold: based on 60 per cent of the equivalized net income for a family of two.

Results

Child maintenance amounts

In Table 2 annual child maintenance liabilities are calculated in seven of the ten scenarios, but not in scenarios (A, E, H), as these are used to set a baseline income for Mary to show the changes in her total income when child maintenance is paid. The results (in PPP-adjusted US$) are given in Figure 1. The first set of bars show that when both parents are unemployed (scenario B), child maintenance expectations are quite low in three countries: in Australia only $278/year is expected, in the UK it is $521 and $629 in New Zealand. In the US (Wisconsin), a relatively high annual amount of $1860 is expected, partly because some states, including Wisconsin, assume that any non-resident parent could earn at least the full-time minimum wage, and set orders accordingly ⁷ .

*There are a total of 10 scenarios, but in scenarios A, E and H, no child maintenance is paid.

Figure 1. Annual Child Maintenance Obligations in Seven* Paying Scenarios

Notes: See Tables 2 and 3 for earnings levels.

Mary's ‘low-moderate earnings’ are 2/3 median female full-time earnings (see Table 2).

Paul's ‘low-moderate earnings’ are 2/3 median male full-time earnings (see Table 2).

Paul's ‘median earnings'are median male full-time earnings (see Table 2).

Source: Authors’ calculations of each country's respective policies.

The next three sets of bars show child maintenance obligations when Paul has low moderate earnings – i.e. two-thirds median male full-time earnings – (scenarios C, F and I), where Mary is unemployed, working part-time, and working full-time, respectively. Obligations are lowest in Australia and New Zealand, somewhat higher in the UK, and, again, substantially higher in Wisconsin/US. In Australia and New Zealand, Mary's earnings are not high enough to result in different maintenance obligations across scenarios C and F but, as Mary's earnings rise in scenario I, there are very small reductions in expected child maintenance. Neither the UK nor Wisconsin considers the receiving parent's income in the assessment of child maintenance amounts, so Mary's earnings make no difference to the amount due.

The final three sets of bars show child maintenance amounts when Paul has median earnings (scenarios D, G and J), with Mary's earnings increasing across the scenarios as before. Again, obligation amounts are by far the largest in Wisconsin/US, followed by the UK but, in these scenarios, the UK is more similar to Australia and New Zealand. In both Australia and New Zealand, Mary's increasing earnings across the scenarios from D to J results in slightly lower child maintenance obligations, but the differences are not large.

Overall, Figure 1 shows the value of child maintenance in all countries. Surprisingly perhaps, even when a non-resident parent has no earnings, all countries expect some maintenance from Paul, and substantially more maintenance as his earnings increase. Figure 1 also highlights the lack of impact Mary's earnings had in calculating child maintenance obligations in our scenarios. This is the result of explicit policy: in the UK and Wisconsin only the non-resident parent's income is considered; whereas in Australia and New Zealand, both parents’ incomes are considered. Moving on to the rest of the analysis, from now on we label amounts of child maintenance within countries as ‘low’ (scenario B), ‘moderate’ (scenario C, F, and I) and ‘high’ (scenario D, G and J). We can do this because, within each country, maintenance amounts stay essentially the same despite various assumptions about Mary's earnings.

Mary's income

Figures 2a–2d provide information on Mary's total income across the 10 scenarios. In each case, the total height of the bars above the vertical axis show Mary's gross income, separately indicating her three potential income sources (government benefits, earnings, and child maintenance). Below the axis, we show subtractions to gross income. These include Mary's net housing costs (the amount of rent less any housing subsidies), payroll taxes and income taxes she is projected to owe. Subtracting these amounts from gross income provides net income after housing costs, shown by the ‘dot’ within each bar.

Mary's ‘low-moderate earnings’ are 2/3 median female full-time earnings (see Table 2).

Figure 2a–d. Mary's Gross Annual Income and Income after Housing Costs, in Ten Scenarios

For the UK (Figure 2a), in scenario A with no maintenance, Mary has a benefit income of approximately $11,600. Although Mary's rent in the UK is the highest of all countries (Table 3), she receives substantial rental assistance, so her net housing costs are approximately $2,800; bringing her total income after housing costs to $8,800. All of the child maintenance paid in scenarios B–J goes to Mary and none of her other benefits are affected. Comparing scenarios A and E, when Mary begins earning $12,500 from part-time work, her benefits decline somewhat. She also owes a small amount of taxes and receives less housing benefit, so her income after housing costs increases by $7,000 in scenario E. Similarly, when Mary is working a low-wage but full-time job in scenarios H–J, her benefits decline somewhat, her share of rent increases, and her taxes increase, but she is able to keep all of the child maintenance paid, increasing her total income.

In Wisconsin/US (Figure 2b) in scenario A, Mary's benefits are comparable to the UK, but approximately 12 per cent lower. Because Mary remains in receipt of the main benefit programme (TANF) in scenarios B, C and D, only 75 per cent of the maintenance is passed through to Mary and 25 per cent is retained by the state to offset welfare costs. In most of the other benefit programmes, all of the child maintenance Mary receives counts as income so, as her child maintenance increases, the level of these benefits decrease. Similarly, Mary's contribution to housing also increases as the amount of child maintenance rises. When Mary works part-time (scenarios E, F and G), she no longer receives the main social assistance programme, but she does receive other benefits. In these scenarios, Mary's other benefits decrease due to child maintenance and her share of housing costs increase. This means that her income after housing costs increases by only a portion of the maintenance paid. This is similar when she works full-time in scenarios I and J.

In Australia (Figure 2c), in scenario A, Mary has benefits of approximately $19,800, but housing subsidies cover less than one-third of the housing costs, so there is a significant difference between gross income (about $20,000) and the dot that shows income after housing costs (approximately $13,000). Mary keeps the full amount of maintenance paid in scenario B, but in scenarios C and D when maintenance rises, some benefits decline slightly and her net housing costs increase. The remaining scenarios (E–J) show Mary with earnings. In these scenarios, where child maintenance is paid, her other benefits again decline slightly and housing costs increase somewhat, leading to an increase in income after housing costs that is less than the total amount of maintenance received.

For New Zealand (Figure 2d), benefits in scenario A are approximately $14,400, and net housing costs bring income after housing costs to $8,750. The next three bars show Mary's income when maintenance is paid. However, even though $629 of maintenance is due in scenario B, $2,263 in scenario C, and $4,320 in scenario D, if this maintenance were to be paid, none of it would be passed through to Mary since she is receiving the main social assistance benefit. Therefore, her income after housing costs remains unchanged across scenarios A to D. In scenarios E, F and G Mary is working part-time and earning approximately $12,700. She no longer receives social assistance, only broader benefits. Therefore all the maintenance is passed through to Mary. But, maintenance does not increase Mary's income because her after tax earnings plus child maintenance is low enough that she receives the Minimum Family Tax Credit which is reduced dollar-for-dollar with increases in her income. As such all her maintenance is clawed back by the government. When Mary has a full-time lower-wage job (scenarios H–J), she is above the threshold for the Minimum Family Tax Credit and so the higher child maintenance payments do increase her income.

Comparing across countries, Figures 2a–d show that when Mary is not working and not receiving maintenance (scenario A), her benefits (and therefore her gross income and her income after housing costs) are lower in the UK and the US than in Australia and New Zealand. Scenarios C–D show the effect when maintenance is paid, with the UK passing it all through to Mary and New Zealand passing none of it through. By scenario D, with high maintenance payments, Mary's income after housing in the UK has not only surpassed that of New Zealand, it is 59 per cent higher. Moreover, even though benefits are relatively low in the UK when Mary is unemployed, they do not decline much when she becomes employed (part-time or full-time work on relatively low wages). In contrast, in New Zealand, and especially in the US, benefits decline more sharply as Mary earns more. Consequently, when Mary is working full-time on relatively low wages (scenarios H–J), benefits in the UK are substantially higher than in the US and somewhat higher than in New Zealand. In Australia, total gross income and income after housing costs are highest in every scenario because of the relatively higher earnings and generous benefits that together outweigh the lower child maintenance amounts paid in Australia.

Child Maintenance, interaction with benefits, and poverty

Finally, we focus on the relationship between child maintenance and poverty (Table 4). The first column describes all ten scenarios grouped according to whether Mary has: (1) no earnings; (2) part-time earnings; or (3) full-time earnings. Under each grouping we show: first, whether child maintenance lifts Mary out of poverty; and second, how much the poverty gap is reduced for each dollar of maintenance paid. The poverty gap is the difference between Mary's income after housing costs and the poverty threshold.

TABLE 4. Child maintenance and poverty (based on income after housing costs) by scenario for each of the four countries

Notes: CM = child maintenance in Purchasing Power Parity (PPP) adjusted US dollars for 2014. See text/tables for definitions of scenarios and poverty.

Source: Authors’ calculations using given earnings levels and each country's policies.

*Some or all of the child maintenance is not decreasing the poverty gap because Mary moves out of poverty

When Mary has no earnings (scenarios A–D), her income after housing costs is below the poverty threshold in all four countries when no child maintenance is paid, but also even when it is paid. Maintenance does not lift Mary out of poverty, but it does reduce the poverty gap in three of the four countries. In scenarios B to D, the decline in the poverty gap for every dollar of child maintenance shows there are substantial variations across countries. In the UK, 100 per cent of every dollar of maintenance paid reduces the poverty gap, whereas the opposite is true for New Zealand. This demonstrates the effect of treating child maintenance as a complement (UK) or a substitute (New Zealand) to the main social assistance benefits. In Wisconsin, only 30 cents of every dollar paid decreases the poverty gap, partly because only 75 per cent is passed through when Mary receives the main social assistance programme, and partly because there are other benefit reductions when maintenance is paid. In Australia, the low child maintenance amount in scenario B fully decreases the poverty gap (when it is treated as a complement). However, in scenarios C and D, there is a decline in other benefits for lower-income families, so only 73–59 cents of maintenance decreases the poverty gap.

When Mary works part-time (scenarios E–G), if she receives no child maintenance (scenario E) her part-time earnings combined with benefits do not lift her income (after housing costs) above the poverty line. When maintenance is paid, Mary is still not brought out of poverty in any country, apart from the UK, and then only for the ‘high’ amount. As before, the UK treats child maintenance as a complement, and thus every dollar of maintenance increases Mary's income and decreases her poverty gap. In Wisconsin, Mary's earnings are high enough that she does not receive social assistance, so the full amount of maintenance is passed-through. However, every dollar of maintenance is counted as income in other programmes and in determining her housing contribution, so that only 43 or 50 cents in every dollar of maintenance paid is decreasing her poverty gap. In Australia, the poverty gap is reduced, but only by 73–59 cents for every dollar of maintenance, because maintenance increases Mary's housing contribution and decreases her other benefits when she is working part-time. As before in New Zealand, Mary's poverty gap is unaffected as none of the maintenance is passed through.

Where Mary works full-time for relatively low wages (scenarios H–J), if she receives no child maintenance (scenario H) her earnings combined with benefits do not lift her income above the poverty line ⁸ . The UK, as in all previous scenarios, treats child maintenance as a complement, increasing her total income from earnings and in-work benefits to lift her out of poverty. Similarly in Australia, for the first time, Mary is lifted out of poverty by maintenance payments in combination with low full-time earnings and other benefits. In Wisconsin/US and in New Zealand, approximately three-quarters of the maintenance paid decreases Mary's poverty gap, but this is not enough to bring her out of poverty.

Comparing across scenarios and countries, the UK stands out as having a consistent policy of complementarity in which child maintenance always increases Mary's income and, where her income is below the poverty threshold, always helps reduce her poverty gap. The other countries show a combination of approaches. In the US/Wisconsin, as Mary's earnings increase, she can keep more of the paid maintenance and thus, in combination with earnings, maintenance becomes more effective at reducing the poverty gap. In Australia, in contrast, the effect on Mary's poverty gap declines as the amount of maintenance increases (down from dollar-for-dollar at ‘low’ maintenance levels to 73 cents in the dollar at ‘moderate’ levels to 59 cents at ‘high’ levels). This contribution to reducing the poverty gap stays the same regardless of Mary's earnings (scenarios C, D, F, G, I, J). New Zealand – in all but two scenarios – is the only country in which Mary's poverty gap is unaffected by child maintenance. For example, when Mary has no earnings and receives social assistance (scenarios A–D) none of the maintenance is passed through. When she has part-time earnings (scenarios E–G), she no longer receives social assistance, so gets all the child maintenance. However, the loss of other benefits means maintenance does not increase her income. The interaction effect means any maintenance is effectively clawed back by the state and does not reduce her poverty. Only when Mary has low-wage full-time earnings (scenarios I, J) does maintenance reduce the poverty gap.

Discussion and Conclusion

This research makes a unique contribution to the poverty literature as previous studies have not explored the subtle and often hidden interactions between child maintenance schemes and social security programmes. Our analysis across four English-speaking countries with similar child maintenance schemes has revealed how two different policy mechanisms operate. First is the explicit ‘pass-through’. This is where the state decides how much maintenance that is paid is passed through to the lone parent when they are also claiming social assistance benefits. In New Zealand, none of the payments are passed through to recipients of the main benefit programme (except for any maintenance amounts in excess of the benefit) and, in the US, only a proportion is passed-through. In contrast, Australia has always passed-through 100 per cent of the child maintenance and, since 2010, so has the UK. We have shown how this mechanism variously reduces the anti-poverty effectiveness of child maintenance. But pass-through is only half of the story; our analysis also exposed more subtle interactions, what we call ‘clawback mechanisms’. These tend to be implicit and operate within social security programmes. They occur when benefit eligibility calculations treat child maintenance payments as income and the value of the benefits are reduced accordingly. We found this mechanism to also erode the poverty effectiveness of maintenance payments and that it operates in all four countries, although in the UK it is limited exclusively to Housing Benefits and only operates in some Local Authorities.

Understanding these clawback mechanisms resulting from child maintenance and social security programme interactions is important for two main reasons. First, they show how poverty in lone parent families can potentially be decreased if maintenance is treated as a complement to cash benefits, rather than treated as a substitute for them. In that regard, the UK stands out. It mostly operates a policy of complementarity passing through all maintenance payments and disregarding it in calculating recipients’ entitlements to other benefits. Consequently, child maintenance (in combination with cash benefits and earnings) decreased the poverty gap for lone parent families in the UK more effectively compared to the other countries with similar child maintenance schemes ⁹ . Second, they reveal how child maintenance cannot simply be added to income when analysing survey data, as context-specific policy interactions can have vastly different effects on the incomes of a lone parent family. In our analysis, it was only in the UK that maintenance seemed to accord with a policy logic that conceived maintenance as a tool to reduce poverty. In the other three countries, only a proportion of the child maintenance paid increased the lone parent's income, and only in a sub-set of scenarios. Therefore, the policy logic was less clear, or even contradictory. As such, our findings indicate that caution must be taken when interpreting survey data reporting the poverty reduction effects of child maintenance payments, particularly for recipients of cash benefits and in comparative studies.

The comparative research conducted here is the first of its kind. While the methods have limitations, it nonetheless draws attention to hidden interaction effects that erode the poverty reduction potential of maintenance payments (at least in our scenarios). It also paves the way for policy analysts to evaluate the magnitude of these interaction effects by testing them in real life cases. This raises important policy implications. Our analysis suggests that it is the relative value of both maintenance and cash benefits (combined with earnings) that is important to poverty reduction, whereby complementarity was likely to be the most beneficial policy configuration. How the policy programmes that deliver these payments interact needs to be better understood to assess their poverty effectiveness. However, as policy currently stands in three of the four countries (bar the UK), we argue that child maintenance has only a latent potential to reduce child poverty given the interactions that mean it can be withheld and/or clawed back by the state. This is contrary to much of the policy rhetoric emphasising the value of child maintenance to poverty reduction.

This research also provides an important means to question neo-liberal rhetoric that extols the virtues of private responsibility and labour market participation as lone parent families’ best route out of poverty. The analysis shows the state retains an important role in providing social protection through cash benefits and setting the extent to which the full value of child maintenance and cash benefits are provided. As the UK case demonstrates, treating child maintenance as a complement can be effective in lifting lone parents out of poverty and in reducing the poverty gap, even at low levels of earnings. The policy implications are clear, all paid child maintenance should be passed through to recipient families and be unencumbered from subtle clawback mechanisms in interactions with social security programmes. To maximise the potential of child maintenance, policy should also focus on enforcement and ensure full payment compliance.

By examining these policy programmes and their interaction effects, our research also makes some contribution to the literature on welfare state analysis. This paper has highlighted the policy contradictions that can ensue across countries with apparently similar welfare state approaches and similar child maintenance schemes. While traditional welfare state analysis classified these four countries as similar liberal regime types based on social investment approaches, Deeming and Smyth (2015) argue for an alternative description of welfare states that also considers ‘social investment’ approaches. They describe welfare states based on this additional social investment criterion as either ‘light’ or ‘heavy’. ‘Heavy states’ are strong on both social investment and social protection approaches (as in some Nordic countries). In contrast, ‘light states’ do not invest heavily in either approach (including the UK, US and Australia). Arguably, if we accept Deeming and Smyth's (2015) categorisation, our research shows that even the ‘light’ approach is highly variable across similarly classified states, and can potentially produce very different outcomes within similar policy programmes. Our analysis scrutinises the likeness or otherwise of discrete social policy programmes across countries which Deeming (2016) argues is a better way to conduct welfare state analysis than regime classification. We make a contribution to that debate by highlighting the interaction effects and divergence in outcomes in programmes across four seemingly similar countries. The research in this paper takes a first step towards documenting a previously understudied policy research area; the interaction effects between social policy programmes dealing with child maintenance obligations. But we have ignored the fact that treating child maintenance as complementary would result in some financial losses for governments. Governments can try to recover costs in a variety of ways, as we have seen by the use of pass-through and clawback mechanisms, or by imposing fees. We have only scratched the surface here in understanding the combination of implicit mechanisms and their relative advantages and disadvantages from the taxpayers’ perspective. An important area for future research is to examine these issues from a governmental perspective and explore how mechanisms might result in lowered payments, increased system complexity, substantial programme savings, or have other effects.