Income supplement: analysis of options

5. Modelling of Policy Options

We have calibrated the options set out in Table 5 in line with the first objective of the income supplement to achieve a 3 percentage point reduction in relative child poverty after housing costs. We then compared the weekly payments and the level of spending required.

The second objective of the income supplement is to ensure that there is a tangible impact on the depth of poverty. Therefore, more detailed distributional impacts are also considered to examine the extent to which families on very low incomes benefit relative to families closer to or above the poverty line under each option.

We model each of the five options based on an automated payment (Section 5.1) and an application-based payment (Section 5.2). Since we are not modelling the delivery costs and impacts, the difference between these two sets of options is in what they assume about take-up, with the former being based on full (100%) take-up by definition.

In addition to the modelling, we assess all options against a set of criteria to ensure that a range of aspects are considered to provide a guide for discussion and decision on the income supplement policy. This is covered in Section 6.

Table 5: Summary of policy options

	Policy Option	Design
1	Child Benefit based entitlement	A payment for every child in a household receiving Child Benefit Households in receipt of a reduced amount of Child Benefit (because their annual individual income is between £50,000 and £60,000) receive the full payment Households with annual individual income in excess of £60,000 do not receive the payment There is no taper and all families in receipt of the payment receive the same amount per child.
2	Universal Credit based entitlement	A payment for every child in a household receiving Universal Credit (regardless of the number of children) UC taper rate is not applied
3	Universal Credit based entitlement – targeted groups	A payment for every child in a household receiving Universal Credit (regardless of the number of children) Additional payment to households having one or more of certain characteristics, informed by the 'priority families' of the TCPDP. These characteristics include: lone parent; child under age of 5; 3 or more children; disabled child; disabled adult; and mother aged under 25 UC taper rate is not applied
4	Entitlement based on a means test	A payment for every child in a household with net earnings below a set threshold Size and composition of the household are not taken into consideration for setting the income threshold
5	Council Tax Reduction based entitlement	A payment for every child in a household that is receiving some level of Council Tax Reduction CTR or UC taper rate is not applied

All policy options have been modelled using the DWP Policy Simulation Model (PSM) that utilises the 2016/17 Family Resources Survey (FRS) dataset for Scotland. Policy costings, poverty impacts and distributional impacts of each option are modelled by projecting forward to the year 2023/24, which coincides with the planned full rollout of UC and makes long-term comparisons more appropriate.

PSM is a static microsimulation model of the tax and benefit system. Annex III provides more detail on the methodology underlying the PSM and the associated caveats. While the PSM framework is owned by the DWP, the assumptions, methodology, and analysis presented here are the responsibility of the Scottish Government. The modelling approach (as outlined in Annex IV) and the results have been scrutinised by DWP analysts.

Box 9: Microsimulation Modelling

The FRS is the main dataset used for modelling social security policy and is commonly used as the basis for UK microsimulation models. Examples include the Institute for Fiscal Studies' tax and benefit simulation model TAXBEN,^[33] the Institute for Public Policy Research Scotland's microsimulation tax-benefit model^[34] also used by Resolution Foundation and Joseph Rowntree Foundation and the Institute for Social and Economic Research's tax-benefit microsimulation model for the European Union, EUROMOD.^[35]

Although the fundamentals of these models are similar, some have a better developed tax modelling capacity. They also vary in their assumptions, for example about how take-up of benefits is modelled, what they assume about the future and the approach to modelling policies that are in the process of being rolled out (most notably UC).

In order to estimate policy impacts in the long-term, the PSM relies on a wide range of assumptions – the most fundamental being that past FRS data, suitably reweighted,^[36] gives a good representation of the future – which each add a degree of uncertainty to the impacts estimated. In addition, the size of the FRS Scottish sample adds to the uncertainty of estimates. As such, the figures presented in this section are best interpreted as providing illustrative examples of how the income supplement policy could work, using comparisons between the different options under consideration and the baseline of no policy change.

5.1 Automatic Entitlement Options

This section analyses each option assuming an automatic payment as presented in table 5 above. We do not believe there is a practical way to deliver automated payments under Option 4. However, we have presented this option for comparison and it can be interpreted as an application-based approach with full (100%) take-up (i.e. everyone who is eligible at any point in time applies and receives the income supplement). It should be noted that the results for other options can also be interpreted as application-based options with very high take-up from the modelling perspective.

Relative poverty For each option, we have estimated weekly payments and the associated policy expenditure required to achieve a 3 percentage point (pp) reduction in relative child poverty after housing costs (AHC) in 2023/24.

We note that each option is assumed to have no impact on the UK median income and therefore the poverty line against which poverty in Scotland is measured.^[37] In addition, behavioural responses have not been considered in the modelling. We also note that estimated expenditure figures include the cost of the amount paid to income supplement recipients and not the cost of delivering the payments.

All illustrative weekly payments are modelled as multiples of £5, annual policy cost figures are rounded to the nearest 10 million and poverty impacts are rounded to the nearest 10,000 and 1 pp respectively. Because each household in the FRS sample represents many households in the real world and a number of other assumptions are made (as detailed in Annex III and IV), microsimulation modelling of this nature, especially given Scotland sample size limitations, cannot be too specific about the value of payment required or the impact on poverty. We therefore round the results to account for this uncertainty.

According to the modelling:

• A Child Benefit based automatic entitlement would require paying an additional £10 per child per week to bring around 30,000 children out of relative poverty in 2023/24, achieving a 3 pp reduction. This option is estimated to reach 870,000 children at an annual policy cost of £460 million.
• A Universal Credit based automatic entitlement would require paying an additional £10 per child per week to lift around 30,000 children out of relative poverty and achieve a reduction of 3 pp in 2023/24. This option is estimated to reach 480,000 children at an annual policy cost of £250 million.
• A targeted Universal Credit based automatic entitlement would require a payment of £5 per child per week for all families receiving UC and an additional £5 for families with one or more priority characteristics to lift around 30,000 children out of relative poverty and achieve a reduction of 3 pp in 2023/24. This option is estimated to reach 480,000 children at an annual policy cost of £240 million.
• An automatic entitlement based on a means test that (for illustrative purposes only) results in full take-up would require a payment of £10 per child per week to families with annual net earnings below £25,000 to bring around 30,000 children out of relative poverty in 2023/24, a reduction of 3 pp. This option is estimated to reach 460,000 children at an annual policy cost of £240 million. It should be noted that within this option there are a range of policy variations, as a number of different thresholds and weekly payments could be chosen to meet the policy objective of reducing relative child poverty by 3 pp. However, for simplicity purposes we are presenting a single illustrative payment here.
• A Council Tax Reduction based automatic entitlement would require an additional £45 per child per week to lift around 30,000 children out of relative poverty and achieve a 3 pp reduction in 2023/24. This option is estimated to reach 140,000 children at an annual policy cost of £330 million.

The results are summarised in Table 6 below.

Table 6: Impacts on child poverty – automatic entitlement options

	Policy Option	Illustrative weekly payment	Children benefitting	Change in children in relative poverty	Percentage point change in children in relative poverty	Annual policy cost
1a	Child Benefit based entitlement	£10 per child	870,000	-30,000	-3 pp	£460 million
2a	Universal Credit based entitlement	£10 per child	480,000	-30,000	-3 pp	£250 million
3a	Universal Credit based entitlement – targeted groups	£5 per child plus £5 for one or more priority characteristics	480,000	-30,000	-3 pp	£240 million
4a	Entitlement based on a means test	£10 per child per week to families with net earnings below £25,000	460,000	-30,000	-3 pp	£240 million
5a	Council Tax Reduction based entitlement	£45 per child	140,000	-30,000	-3 pp	£330 million

Decile analysis The second objective for the income supplement requires us to consider the impact on the depth of poverty. One way of assessing this is to look at the distributional analysis of the options modelled as presented in the charts in Box 10 below. The decile analysis divides the Scottish families with children into ten equal groups. The first decile group represents the 10% of households with children with the lowest incomes and the tenth decile group the 10% of households with the highest incomes.^[38]

When comparing the impacts of each option, the distributional analysis is presented on the basis of both the number of children in each decile that would receive the payment as well as the total spend per decile.^[39]

• The charts highlight that on both measures the impact of option 1a is spread more widely across the income distribution compared to other options as almost all children would benefit from this option. Less than 40% of the impact on both measures is concentrated in the bottom three deciles with around 13% of children benefitting and estimated cost in each of the bottom three deciles. However, the number of children benefiting and the corresponding spend in the 10th decile are notably lower (1% of the total number of children benefitting and spend), because those families with individual income of £60,000 and higher would not be eligible.

• Over 60% of the impact of Options 2a and 3a, on both measures, is concentrated in the bottom three deciles with 20% of children benefiting in the bottom decile, 21% in the second decile and 20% in the third decile.^[40]

• For Option 4a, 65% of the impact on both measures is concentrated in the bottom three deciles, with 23% of children benefiting and estimated spend in the bottom two deciles respectively and 19% in the third decile.

• Over 70% of the impact of option 5a on both measures is concentrated in the bottom three deciles, with 43% of children benefiting and estimated spend in the first decile, 20% in the second decile and 8% in the third decile. This suggests that, under this option, children in the bottom income decile benefit the most.

Box 10: Distributional impacts of automatic entitlement options

Estimated coverage versus targeting In addition to the decile analysis, table 7 sets out for each option an estimate of children in poverty who would receive the payment both as a percentage of all children in poverty (what we refer to as 'coverage') and as a percentage of all children receiving the payment (what we refer to as 'targeting'). All figures are rounded to the nearest 5%.

Table 7: Estimated Coverage versus Targeting

	Policy Option	Illustrative weekly payment	Coverage: children in poverty receiving the payment as % of all children in poverty	Targeting: children in poverty receiving the payment as % of all children receiving the payment
1a	Child Benefit based entitlement	£10 per child	100%	25%
2a	Universal Credit based entitlement	£10 per child	90%	45%
3a	Universal Credit based entitlement – targeted groups	£5 per child plus £5 for one or more priority characteristics	90%	45%
4a	Entitlement based on a means test	£10 per child per week to families with net earnings below £25,000	95%	50%
5a	Council Tax Reduction based entitlement	£45 per child	40%	65%

The modelling suggests that Option 1a is the only option that ensures that almost all children in relative poverty would receive a payment. This is because of both the eligibility criteria (which captures most households) and high take-up of Child Benefit. However, this also means that only 25% of children in families in receipt of this payment are likely to live in relative poverty, suggesting that around 75% of children who could benefit from the policy would not be living in relative poverty. This is consistent with fact that the cost of this policy is higher than the other options modelled.

Under Options 2a and 3a, 90% of children living in relative poverty would be expected to receive the income supplement and 45% of all children receiving the income supplement are estimated to live in relative poverty, suggesting better poverty targeting than Option 1a.

At 50%, poverty targeting appears to be marginally better under option 4a. Coverage is also higher at 95%. It should be noted again that this option is for comparison purposes only as automated payments not linked to an existing benefit are not considered viable.

Under Option 5a, 65% of the children in families receiving the income supplement are estimated to live in relative poverty. Conversely the policy coverage decreases and only 40% of children living in poverty would receive the income supplement.

This analysis suggests that there is a clear trade-off between coverage and targeting. By introducing a more targeted income supplement, the risk that some children in poverty may not be included increases and the coverage of the policy falls. Conversely, the more universal the income supplement, the higher the likelihood of paying it to families whose incomes are substantially above the poverty line.

Severe poverty Another way to assess the impact on the depth of poverty is to look at the impact on severe child poverty (those in households with incomes below 50% of the median). Our modelling shows that the impacts on severe poverty for Options 1a to 4a are modest and we cannot report them with confidence. The modelling suggests that Option 5a has the potential to deliver a 5 pp reduction in severe poverty. This is because Option 5a is targeted at families with children further away from the poverty line, and therefore more likely to experience severe poverty.

5.2 Application Process Options

This section analyses the options set out in Table 5 but on the basis of an application process.

In terms of policy modelling, the key difference from the automatic payment options presented in the previous section is the assumed take-up rate. This allows us to make a more realistic assumption for options requiring an application as opposed to automatic payments. For Option 1 we have used the most recent take-up rate for Child Benefit, estimated at 93% as we expect take-up of a nearly universal income supplement to be similar to Child Benefit. As Options 2, 3 ,4 and 5 are directed at low-income families we have used Child Tax Credit take-up rate which is the closest comparator. The overall rate is estimated at 83% and we have modelled take-up rate by income bands. Because under Option 5 most of the families receiving Council Tax Reduction are in the lower income bands, the overall take-up rate for this option is 89%. Further detail on modelling take-up is set out in Annex IV.

As discussed in section 6.2, it is very difficult to predict the direction of take-up as there are several factors that can affect it. Therefore, the assumed take-up rates in this section are subject to a high level of uncertainty.

Relative poverty Based on the policy modelling:

• A Child Benefit based entitlement with an application process would require paying an additional £10 per child per week to lift around 30,000 children out of relative poverty in 2023/24, achieving a 3 pp reduction. This option is estimated to reach 810,000 children at an annual policy cost of £420 million.
• A Universal Credit based entitlement with an application process would require paying an additional £10 per child per week to lift around 30,000 children out of relative poverty in 2023/24, achieving a 3 pp reduction. This option is estimated to reach 400,000 children at an annual policy cost of £210 million.
• A targeted Universal Credit based entitlement, with an application process would require paying £5 per child per week to all families receiving UC and an additional £5 for one or more of the priority characteristics to lift around 30,000 children out of relative poverty and achieve a reduction of 3 pp in 2023/24. This option is estimated to reach 400,000 children at an annual policy cost of £200 million.
• An entitlement based on a means test would require a payment of £10 per child per week to families with annual net earnings below £25,000 to lift around 30,000 children out of relative poverty in 2023/24, a reduction of 3 pp. This option is estimated to reach 400,000 children at an annual policy cost of £210 million.
• A Council Tax Reduction based entitlement with an application process would require paying an additional £45 per child per week to lift around 30,000 children out of relative poverty in 2023/24, achieving a 3 pp reduction. This option is estimated to reach 120,000 children at an annual policy cost of £290 million.

These results are summarised below in table 8.

Table 8: Impacts on child poverty – application process options

	Policy Option	Illustrative weekly payment	Children benefitting	Change in children in relative poverty	Percentage point change in children in relative poverty	Annual policy cost	Assumed take-up rate
1b	Child Benefit based entitlement – application	£10 per child	810,000	-30,000	-3 pp	£420 million	93%
2b	Universal Credit based entitlement – application	£10 per child	400,000	-30,000	-3 pp	£210 million	83%
3b	Universal Credit based entitlement – targeted with application	£5 per child plus £5 for one or more priority	400,000	-30,000	-3 pp	£200 million	83%
4b	Entitlement based on a means test and an application	£10 per child	400,000	-30,000	-3 pp	£210 million	83%
5b	Council Tax Reduction based entitlement – application	£45 per child	120,000	-30,000	-3 pp	£290 million	89%

Decile analysis

• The charts below highlight that the impact of Option 1b is spread across the income distribution as with Option 1a. As a result less than 40% of the impact on both measures is concentrated in the bottom three deciles, with around 13% of children benefitting and estimated cost in each of the bottom three deciles.
• Options 2b and 3b could benefit families with children in the bottom three income deciles and result in nearly 65% of the impact on both measures, being concentrated in the bottom three deciles, with 20% of children benefiting in the bottom decile, 23% in the second decile and 21% in the third decile.^[41]
• Over 65% of the impact of Option 4b, on both measures, is concentrated in the bottom three deciles, with 26% of children benefiting and estimated spend in the bottom decile, 24% in the second decile and 16% in the third decile.
• Nearly 70% of the impact of option 5a on both measures is concentrated in the bottom three deciles, with 43% of children benefiting and estimated spend in the first decile, 17% in the second decile and 9% in the third decile. This suggests that, as with Option 5a, children in the bottom income decile benefit the most.

Box 11: Distributional impacts of application process options

Estimated Coverage versus Targeting Since the application process would be expected to have an impact on take-up, coverage of these options will be lower than for the automated process. This means that all application‑based options would be expected to reach a lower share of children than the automated options. The pattern of take-up, e.g. the extent to which those on lowest incomes are more likely to take up the income supplement will influence both coverage and targeting. Take-up is discussed in more detail in section 6.2.

• Looking at table 9 where each option is compared in terms of estimated coverage and targeting, under Option 1b the income supplement would reach around 90% of children in relative poverty; 25% of children benefitting from the income supplement would be living in poverty. As with Option 1, most of children who could benefit from the policy (75%) would not be living in relative poverty.
• Options 2b and 3b are estimated to reach around 80% of children who live in relative poverty and around 45% of children benefiting from these options would be living in poverty.
• Under Option 4b, targeting – estimated at 50% – is better as higher payments are made to families with lower earnings, suggesting that more children benefitting from this option were living in poverty.
• Option 5b, as with Option 5a, would reach a relatively smaller share of children living in poverty, estimated at 35%. However 60% of children receiving the payment would be living in poverty.

Table 9: Estimated Coverage versus Targeting

	Policy Option	Illustrative weekly payment	Coverage: children in poverty receiving the payment as % of all children in poverty	Targeting: children in poverty receiving the payment as % of all children receiving the payment
1b	Child Benefit based entitlement	£10 per child	90%	25%
2b	Universal Credit based entitlement	£10 per child	80%	45%
3b	Universal Credit based entitlement – targeted groups	£5 per child plus £5 for one or more priority characteristics	80%	45%
4b	Entitlement based on a means test	£10 per child per week to families with net earnings below £25,000	90%	50%
5b	Council Tax Reduction based entitlement	£45 per child	35%	60%

As with the automatic entitlement options, this analysis shows the clear trade-off between coverage and targeting. The more universal the income supplement the higher the coverage of the policy and at the same time the higher the likelihood of paying it to families whose incomes are substantially above the poverty line. On the other hand, by introducing a more targeted payment the risk of not including some children in poverty increases.

Severe poverty As with the core options, the impact each application-based option may have on severe poverty has also been considered. All options with the exception of Option 5b are found to have a small impact on severe poverty that we cannot report with confidence. Option 5b is estimated to reduce severe poverty by 4 pp.

5.3 Comparing the Options

The previous sections have analysed five policy options for the income supplement. All options were assessed on the basis of both automation and application.

1. Child Benefit based entitlement

2. Universal Credit based entitlement

3. Universal Credit based entitlement with targeted groups

4. Entitlement based on a means test

5. Council Tax Reduction based entitlement

Below we bring together the results of the modelling set out above.

The key assessment is the effectiveness of each option in reducing child poverty. Table 10 summarises how each option performs against the primary objective of the income supplement to achieve a 3 pp reduction in relative child poverty.

Within each option the delivery process – automation or application – would achieve the same reduction in child poverty at different costs. This is because fewer children would be captured if an application process is in place.

Table 10: Effectiveness in achieving 3 pp reduction in child poverty

	Policy Option	Automation vs Application	Illustrative weekly payment	Children benefitting	Change in children in relative poverty	Percentage point change in children in relative poverty	Annual policy cost
1a	Child Benefit based entitlement	Automation	£10 per child	870,000	-30,000	-3 pp	£460 million
1b		Application		810,000			£420 million
2a	Universal Credit based entitlement	Automation	£10 per child	480,000	-30,000	-3 pp	£250 million
2b		Application		400,000			£210 million
3a	Universal Credit based entitlement – targeted groups	Automation	£5 per child plus £5 for one or more priority	480,000	-30,000	-3 pp	£240 million
3b		Application		400,000			£200 million
4a	Entitlement based on a means test	Automation (full take‑up)[42]	£10 per child per week to families with net earnings below £25,000	460,000	-30,000	-3 pp	£240 million
4b		Application		400,000			£210 million
5a	Council Tax Reduction based entitlement	Automation	£45 per child	140,000	-30,000	-3 pp	£330 million
5b		Application		120,000			£290 million

As presented in figure 3, Options 2, 3 and 4 are estimated to have similar annual policy costs ranging between £200 million and £250 million. Option 1 would require the highest policy expenditure, estimated at £460 million with automation and £420 million with application, followed by Option 5, estimated to require £330 million with automation and £290 million with application. It would also require the highest payment per child at £45 per week as opposed to £10 per week required for all other options.

Figure 3: Comparing estimated weekly payment and annual policy expenditure

The difference in estimated policy costs is explained by the trade-off between coverage (how many children in poverty the payment can reach) and targeting (how many of all children receiving the payment are in poverty) as summarised in figure 4.

Option 1 would reach a large number of children who are not in relative poverty and therefore would be more expensive. Option 5 would also require a relatively higher policy expenditure, because it is paid to a smaller share of children in relative poverty and therefore larger weekly payments would be required to achieve the 3 pp reduction.

Options 2, 3 and 4, whilst requiring the same payment as Option1, are found to be more targeted at children in poverty.

Figure 4: Comparing estimated coverage and targeting