Select Committee on Work and Pensions Minutes of Evidence

Memorandum submitted by the Institute for Fiscal Studies (PC 23)


  The pension credit will restructure means-tested support for pensioners to address perceived shortcomings in the current system. So to understand its intended purpose, it is necessary to grasp the existing structure of benefits for pensioners. Most pensioners receive the flat-rate basic state pension, which in April 2003 is expected to stand at £77 per week for single people (£124 for couples).[1] Since the early 1980s, it has been indexed only with prices (which rise slower than earnings) and so it provides a living standard that is increasingly meagre in relative terms. Therefore, when recent governments have had money to spend on pensioners, they have tended to target it on those who would otherwise have only the basic pension to survive on.

Figure 1

Illustrative Budget Constraint under the MIG, for a Single Pensioner


  2003 benefit rates assumed.

  Income disregards, taxation and other means-tested benefits ignored.

  To do this, benefits that provide pensioners with an income "safety net"—the minimum income guarantee (MIG) and its predecessor, income support—have been increased.[2] These benefits "top up" the poorest pensioners' incomes. Crucially, topping-up is to a fixed level, regardless of whether the claimant has private income or not. So, for the poorest pensioners, any private income possessed is worthless—because, pound for pound, it reduces the amount of "top-up" received. Figure 1 illustrates how the system works for a single 65-year-old, with benefit rates in April 2003, when the MIG is expected to be £100. If his/her only income is the basic pension, the figure shows he/she will receive a £23 top-up (£100 MIG minus £77 original income). If, instead, the pensioner has the basic pension plus £10 of private income (total pre-means-tested-benefit income of £87) then we can see he/she receives only £13 top-up (£100 MIG minus £87 original income). So, final income is the same (£100) in the two cases. Indeed, the figure shows that a pensioner's original income has to exceed the MIG before it affects final income.

  This situation could be deemed problematic on two counts. First, it distorts the incentive to save: in particular, making saving unappealing for low-wage workers who expect to retire on a low income. The specific concern is that people in this situation would decide not to save at all, which would reduce their private incomes and so increase the means-tested benefit bill. Secondly, on grounds of fairness, it might be felt that it is unacceptable that those low-income pensioners who have saved to achieve some private income should go entirely unrewarded for their thrift.

  The situation has become more pressing recently as the Labour government has increased the value of the MIG far more rapidly than that of the pension. In the first year of the current government, 1997-98, income support for a 65-year-old stood at £68.80 (in current cash terms) while the retirement pension was £62.45, a gap of just £6.35, compared with the £23 we have seen is expected for 2003-04. This means that, in real terms, the private income that such a 65-year-old must possess before it has any effect on his/her final income will have roughly tripled in six years.[3]

  In the absence of reform, these tendencies would have been exacerbated in the more distant future, because the government "aspires" to increase the MIG every year in line with earnings, whereas the expectation is that the basic pension will continue to rise roughly in line with prices.[4] This means that the gap between the MIG and the pension should rise faster than earnings. In turn, this suggests that a growing number of pensioners will find that their basic pension entitlement together with their private income is below the MIG, and so their private incomes will have no effect on their final income.[5] The concern was, therefore, that the government's strategy of targeting help for pensioners "where it is needed most" implied ever more pensioners failing to be rewarded for their thrift.

  All this pointed to reform, and proposals for the pension credit followed. Figure 2 shows how it should work for an example single pensioner. Before the reform, each £1 of private income immediately above the basic pension left final income unchanged; afterwards, each £1 increases final income by 60p; ie abstracting—for now—from taxes and other income-tested benefits, this means the effective marginal tax rate implied by the pension credit will be 40 per cent, compared with the 100 per cent rate that the MIG currently imposes.[6] This 40 per cent withdrawal continues until all pension credit entitlement is exhausted, at around £135 per week. The chart shows that the lower withdrawal rate means single pensioners with non-means-tested-benefit income of anywhere between £77 and £135 will gain.[7]

  All this has been known for over a year, since the government unveiled its plans for a pension credit, as part of its substantial package for pensioners in the 2000 Pre-Budget Report.[8] But important additional details were not settled until the 2001 Pre-Budget Report.[9]

Figure 2

Effect of the Pension Credit on a Single Pensioner


  All incomes in projected 2003 prices.

  Income disregards, taxation and other means-tested benefits ignored.


  The basic shape of the pension credit proposal, as outlined in Section A, did not change between the government's Autumn 2000 consultation paper and the firmer proposals outlined at the time of the 2001 Pre-Budget Report. But there was some new important information—both changes made in the light of the consultation process and provision of details that had not previously been given—in the later document.

  In particular, we found out:

    —  the expected cost;

    —  that the main ("savings credit") reform will be restricted to men and women aged 65 and over;

    —  that the capital rules in the MIG system would be maintained but modified;

    —  how the reform will affect housing benefit and council tax benefit; and

    —  how the new means test will work.

  We will discuss each of these in turn. Before we do so, it is useful to expand on the government's terminology, which was also clarified in November.

  From 2003, the MIG—the guaranteed minimum income—will become a feature of "pension credit". This basic element will be known as "pension credit guarantee", effectively a re-labelling of the MIG, and will be available to exactly the same individuals. The main reform—the extra benefit available as a result of the reduction in the benefit withdrawal rate from 100 per cent to 40 per cent—will be known as the "savings credit" element of the pension credit. It will be available to a more restricted group. (See footnote 7 for more detail on how the government describes the benefit calculation.)


  The reform had previously not been costed in the public finances. In February 2001, an IFS Briefing Note estimated that it would cost at least £1 billion a year (in 2000 prices), but it was stressed that this was extremely dependent on the generosity of the changes made to other means-tested benefits at the time the pension credit is introduced.[10] As we discuss in Section B4, the government has decided to make housing benefit and council tax benefit significantly more generous for pensioners when it introduces the reform, making the reform more expensive.

  Another announcement concerned timing. The reform will not be introduced in April 2003, but October 2003, reducing by roughly 50 per cent its cost in the first year. The government projects (in current prices) that the reform will cost:

    —  £975 million in fiscal 2003-04;

    —  £2.025 billion in fiscal 2004-05, when it will be effective over the whole year for the first time.

  These numbers are very similar to the estimated additional benefit expenditure that our own tax-benefit modelling suggests will be needed.[11] In 2001 prices, we estimate a total of £1.9 billion extra benefit expenditure annually from the reform, of which £1.5 billion is being paid in pension credit itself, while £200 million is housing benefit and £200 million is council tax benefit. By contrast, when we modelled the introduction of the pension credit initially, we assumed that council tax benefit and housing benefit would continue to operate as now. The result was that expenditure on these benefits would actually have declined, by some £400 million, helping to offset the new pension credit expenditure.

  If we assume inflation at 2.5 per cent, then we estimate that the total extra expenditure will be £2.1 billion in projected 2004-05 prices, slightly higher than the government's estimate. But our modelling assumes 100 per cent take-up of benefit entitlement. The government's own most recent estimates suggest that only between 74 per cent and 86 per cent (by expenditure) of pensioner entitlement to income support is actually claimed.[12] Allowing for some degree of non-take-up could reduce our cost estimate significantly. There might also be additional differences between our estimation methodology and that used by the government.[13]

  We have not yet attempted to estimate the likely cost of the reform in the longer term. In practice, this will depend hugely on future government policy regarding indexation. The government's own estimates suggest that, by 2040, the cost of the reform could vary between 0.1 per cent of GDP and 1.1 per cent of GDP, depending on whether the MIG is uprated with prices or earnings beyond 2006.[14] Further analysis of these issues, which will crucially determine the impact of the reform on the rest of the pensions system, would be extremely valuable.


  The MIG is available to all low-income individuals (and their partners) of 60 or over. The savings credit element of the pension credit, by contrast, will only be available to those aged 65 or over (and their partners). The same is true of the main changes to housing benefit and council tax benefit. When the new system comes in, however, those in the 60 to 64 age bracket will benefit from the changes to the capital rules.

  Although there are large numbers of individuals in the 60 to 64 age bracket, their exclusion from the reform does not affect the cost very dramatically. The main reason is that 60 to 64 year-olds tend to be better off than older people, and so a lower proportion of them stand to benefit from this means-tested benefit increase. As a result, the immediate effect of applying an age limit of 65 is to save just £200 million compared with the alternative of using an age limit of 60. But this ignores behavioural responses, which could increase the cost of setting the age limit younger. Given that the savings credit makes the benefit system more generous, it could encourage potential recipients to decide that they could retire comfortably, when they would not otherwise have done so. As many of those aged 60 to 64 (unlike their older counterparts) are still working, if such behavioural effects were important, the cost of including under-65s could have increased considerably.


  The government's original plans for the pension credit seemed to suggest that the capital element of the MIG means test would be completely abolished when it was replaced by pension credit. Instead, only income generated by capital held would be subject to a means test.[15] The government's November 2001 proposals, by contrast, envisage keeping some capital rules, although they will be made significantly more generous than the existing rules, which penalise possession of financial capital very heavily. In particular, under the new rules:

    —  the upper capital limit above which no MIG can be received (currently £12,000) will be abolished;

    —  the amount of capital deemed equivalent to £1 of weekly income will double from £250 to £500.

  The government puts the changes down to arguments it received from pensioners' organisations and others during consultation.[16] Certainly, the changes are consistent with our submission at that time. We argued that there were significant problems with moving completely away from a capital test.[17] It would have:

    —  involved a loss of precision in the targeting of resources on the basis of need;

    —  discouraged the annuitisation of wealth, which might be a concern, given that other elements of government policy, such as stakeholder pensions, are aimed at encouraging this;

    —  made it far harder to avoid creating losers when the reform was introduced;

    —  quite possibly increased, rather than reduced, the informational demands on claimants.

  The measures the government is now proposing instead have none of these disadvantages, and yet they still address the concern that motivated the proposal to scrap the capital rules in the first place—they will allow saving amongst low-income people to be better rewarded.

  The abolition of upper capital limits in the pension credit, above which all benefit entitlement is forgone, is especially welcome. It should avoid the "cliff edge" situation where an individual with (say) £12,001 can find him/herself significantly worse off than if he/she had got exactly £12,000, as the extra £1 left him/her totally ineligible for benefits. Given this progress, however, it seems surprising that the government has chosen to maintain the £16,000 upper capital limits in housing benefit and council tax benefit for those pensioners who are entitled only to the savings credit. Although these will affect relatively few individuals, the reform package would surely look more coherent if these, too, were abolished.


  The government's original plans said nothing specific about how the introduction of the pension credit would affect housing benefit and council tax benefit (hereafter referred to as "rebates"). Rather, the government acknowledged that issues were raised, and appealed for views on these.[18]

  Interactions with the rebates are an issue because of two consequences of the way the benefits are withdrawn. Rent and council tax rebates are withdrawn respectively at 65 per cent and 20 per cent for those pensioners with an income of just above the MIG; where both benefits are received, the combined withdrawal rate is the sum of these two—85p in the £1. This withdrawal means two things. First, gains to pensioners from policies such as the pension credit that lifted them just above the MIG would be much reduced, as up to 85 per cent of any gain would be lost in forgone housing and council tax benefit entitlement. Secondly, this withdrawal could dominate the increased reward for saving an extra pound of income that the pension credit reform introduces. The pension credit substantially reduces the MIG withdrawal rate from 100 per cent to 40 per cent; but if both rebates are simultaneously being withdrawn, then the combined benefit withdrawal rate of a pension credit recipient would again rise, to 91 per cent.[19]

  With these problems in mind, the government has decided that when pension credit is introduced, both housing benefit and council tax benefit will be made more generous. In particular, the income that pensioners are allowed to possess before the rebates are reduced (their "needs") will be increased above the level of the MIG. For pensioners eligible for the savings credit, needs will be increased by the maximum gain from reform. This completely succeeds in eliminating the first problem that we have considered—a typical pension credit recipient, even one who gains by the possible maximum amount, will receive the same rebate entitlement as before, and so gain from the reform in full. There are some special cases, however, where particular care will be needed if the full gain is really to be guaranteed.[20]

  So, the government's decision to make the rebate systems more generous should be able to prevent benefit "claw-back" consuming any of the pensioners' gains from the reform in 2003. Indeed, pensioners whose income puts them just above the MIG will actually see an increase in their incomes that is bigger if they are on housing benefit and/or council tax benefit than if they are not—as the increase in needs will mean that they become entitled to full support from these rebates for the first time.

  The proposal for the rebates will not, however, completely avoid the problem of high marginal rates that we have discussed. For some, the problem is solved. Consider a pensioner on £105 of pre-means-tested-benefit income a week: the increase in "needs" would mean that an extra pound of private income will no longer give rise to reduced rebate entitlement. Even though such a person would be exposed to the pension credit taper, at 40 per cent, this would be less than the combined rebate taper (85 per cent) that he/she had previously faced. But for others, things are different. For example, a 65-year-old with pre-means-tested-benefit income of £120 a week who received partial rent and council tax rebates currently faces a taper on these of 85 per cent; post-reform, he/she would continue to face a taper on these benefits but would also become entitled to pension credit, taking his/her total marginal deduction rate to 91 per cent. Others, whose incomes were previously just too high to qualify for partial rebates—and the high deduction rates that this entitlement produces—will become entitled to them for the first time as a result of the increase in needs.

  For some pensioners, then, marginal deduction rates will increase, while for others, they will decline. We will consider the relative quantitative significance of these different types of cases when we look at effective marginal tax rates in Section C.


  In its November 2000 consultation paper, the government stressed its desire to reduce the degree to which applying for benefits was intrusive and troublesome for pensioners. In particular, it suggested that it wanted to examine the ways in which the existing weekly means test for MIG could be made more like the annual process of tax assessment that better-off pensioners (and others) currently undergo.[21] The government now seems to have gone further and decided on a five-year period of assessment, with reassessment necessary only where one's income drops (in which case it is voluntary) or where one's circumstances change dramatically. It has also come up with some welcome practical proposals about how the Pension Service can try to improve take-up.[22]

  When we first looked at the pension credit proposals, we welcomed the aim of simplifying the means test as something that should improve take-up.[23] Indeed, the actual—as opposed to the notional—results of the reform could depend just as much on its success in improving take-up as they will on the precise parameters of the new system. While applauding the aim, however, we expressed a number of concerns. In particular, we were concerned, first, that there was nothing in the move to an annual system that, per se, would necessarily significantly simplify the experience of applicants. The new practical proposals for telephone application etc. do, however, look as if they will help in this regard.

  Our second concern was the potential abuse of a system that fixed payments for a year—some groups of workers might be able to report a low income, achieved by reducing their hours of work, and be awarded high pension credit entitlement, only to go on and boost their income once this was fixed. The move to a five-year period of fixed payments potentially worsens this risk. The fact that pension credit will now be restricted to those aged over 64 should help considerably: fewer in this older age bracket will be the type of semi-retired people able to manipulate their earnings strategically. Only 3 per cent of those in our data aged 65 or over report themselves as currently working as employees, and only a fraction of these will be able to choose their hours at will. Still, for this small group, and for those who are self-employed, the potential gain from a strategic reduction in hours for the purposes of a means test could be significant now that the government has decided on a five-year period of fixed awards. The government may want to think about how to minimise the potential for this abuse.


  When we first modelled the pension credit, we had no idea about a whole range of details concerning how the new system would work. In particular, we knew nothing about how the interaction with the rebates system would be handled. We assumed the least expensive possible option compatible with the maintenance of the MIG as the net income safety net (ie that "needs" in the rebates would be increased in line with the MIG but no further), while noting that other options were possible that would have quite different effects.[24] But we now know that the government is going to be considerably more generous than this, meaning new estimates are needed. Our new modelling must also incorporate the effect of the changes to the capital rules and the new age restriction, amongst other things.

  In what follows, we model the reform by assuming the parameters used illustratively in the government's proposals.[25] Included in our "reform" is the increase in the pensioner tax allowance that the government will introduce to ensure that no single pension credit recipient (without special benefit premiums) need pay tax. Excluded is the real-terms increase in the rate of the MIG scheduled for April 2003.


  Looked at across the whole income distribution, the pension credit is a progressive measure, although it is worth less to the bottom 10 per cent than it is to the second poorest income decile, as Figure 3 shows. This turns out to be because—given the greater generosity of the benefit system to pensioners, even before the reform—few pensioners are in the bottom 10 per cent. Those pensioners who are in the bottom 10 per cent of the income distribution actually gain rather more in proportional terms than those in the second decile. This suggests that it might be more interesting to look at the impact across the distribution of pensioner incomes. But it is also worth considering one implication of Figure 3—if the concern is alleviation of the most acute poverty in general, then it is not obvious that directing more money to pensioners should be the top priority. Rather, those groups whom the benefit system treats less generously and who therefore dominate the bottom decile should, perhaps, be the immediate target.

  Figure 4 recalculates the figures using deciles constructed on the basis of family income for adults aged 65 or older only.[26] There are two immediate differences from Figure 3. First, the scale of the redistribution that the reform represents is far larger—gains for the bottom two deciles now represent an increase in income of significantly more than 5 per cent. This is because here we are only considering the effect on those people whose age means that they are potentially affected. Secondly, the poorest 10 per cent now gains the most: 8.1 per cent of income. This is because many households that gain that are not in the bottom 10 per cent of the overall income distribution are amongst the poorest 10 per cent of those over 65. Even in cash terms, it turns out that the poorest 10 per cent of pensioners gain most—some £11 a week on average, compared with gains of between £5 and £8 for each of the other deciles apart from the top three, which gain less.

Figure 3

Gains from the Pension Credit across the whole income distribution


  Income deciles are derived by dividing the total population into 10 equally sized groups according to household income adjusted for family size. Decile 1 contains the poorest tenth of the population, decile 2 the next poorest tenth and so on, up to the richest tenth in decile 10.


  IFS tax and benefit model, TAXBEN, run using data from the Family Resources Survey, 1998-99.

Figure 4

Gains From the Pension Credit across the over-65s Income Distribution


  Income deciles are derived by dividing the population of adults living in families containing someone over 65 into 10 equally sized groups according to household income adjusted for family size. Decile 1 contains the poorest tenth of the population, decile 2 the next poorest tenth and so on, up to the richest tenth in decile 10.


  IFS tax and benefit model, TAXBEN, run using data from the Family Resources Survey, 1998-99.

  Given that the pension credit is a reform aimed at helping those with incomes just above the MIG, it may seem surprising that it is the "really poor" rather than the "nearly poor" pensioners who seem to be gaining most. One reason is that there are a small but significant number of high-capital, low-income pensioners at the very bottom of the income distribution who can gain very substantially from the reform of the capital rules, as they can become entitled to means-tested benefits for the first time. Indeed, according to our data, average family savings in the bottom pensioner decile are actually higher than those in the second poorest decile (£5,000 rather than £4,000).

  Looking at the difference between couple and single pensioner families, we find that couples gain slightly more on average: £6.60 as opposed to £5.26. But given that pensioner couples have higher incomes on average, the proportional effect on single pensioners will actually be greater.[27]

  Like the estimates of cost we discussed in Section B1, these figures assume full take-up of benefit entitlement. If, in reality, take-up by expenditure is similar to that for existing entitlement, the average effects will be reduced by about 20 per cent.[28]


  The government is aiming to reduce pensioner poverty. One way in which it measures this is by the number of pensioners living below various thresholds of median income. The pension credit is one policy that might help in this regard. Table 1 shows its effect on poverty rates using two particular thresholds—60 per cent and 70 per cent of median income.

Table 1

Proportion of Adults Aged 65 or Over Living Below Different Poverty Lines, Before and After the Reform

Poverty line
60 per cent median income
10.6 per cent
5.7 per cent
70 per cent median income
18.5 per cent
15.9 per cent


  The poverty line is measured relative to the whole population median income (where households are weighted by the number of people in them). Income is measured net of taxes and benefits, but before deducting housing costs. Percentages are fractions of the 8.54 million people aged 65 or over who live in private households in Great Britain.


  IFS tax and benefit model, TAXBEN, run using data from the Family Resources Survey, 1998-99.

  Our estimates of poverty rates before the reform are considerably lower than those published in Households Below Average Income (HBAI), even though both sets of estimates ultimately derive from the Family Resources Survey.[29] The divergence is principally because the numbers here are based on a tax and benefit model, whereas the HBAI is simply recorded data. The figures in Table 1, therefore, show the notional working of the 2003 tax-benefit system—they can capture the pre-announced increases in benefits for pensioners that the HBAI numbers cannot, and they assume that all benefit is taken up, which the HBAI does not.[30] But the effect of a particular policy on poverty in a tax-benefit model should at least relate reasonably closely to the eventual effect that the policy will have on the HBAI figures when compared with the option of not introducing the reform.

  We can see that the effect of the reform might be to reduce by almost five percentage points the number of pensioners living below the 60 per cent median income poverty line; for the 70 per cent median income poverty line, the effect is more modest—a decline of 2.6 percentage points. One very important reason why the actual effect on the poverty rate might be rather smaller than this is that take-up of the entitlement created will be less than 100 per cent.


  Changes to the tax and benefit system can have two types of effect on the incentive of working individuals to save for retirement. The first are "income effects": these occur when people's understanding that the reform has increased (decreased) their income in the future will decrease (increase) the urgency with which they need to save. Because the pension credit reform represents a means-tested benefit increase, it increases low-income individuals' expected future income and thereby blunts their incentive to save through the income effect. The second channel through which the tax-benefit system affects saving incentives is the "substitution effect", ie it affects the rate at which an extra pound of saving now cashes in in terms of increasing the ability to spend in retirement. In what follows, we will focus on substitution effects as the reform is largely focused on these. But it is important to appreciate that these are not the only consideration that will determine the effect of the reform on saving behaviour.[31]

  The existing MIG system exerts negative substitution effects by clawing back the savings income of low-income pensioners in forgone benefit entitlement. As we saw in Section A, benefit withdrawal can mean that saving to provide for an extra pound of income ends up having no effect on the incomes of the poorest pensioners. Another way of expressing this is to say that the "effective marginal tax rate" (EMTR) is 100 per cent.

  We can use EMTRs more generally to look at the extent to which benefit withdrawal and taxes together reduce the incentive to save. Table 2 shows how the reform will alter these rates for adults of 65 or over. In general, the reform has reduced them. It has very substantially reduced the number of pensioners who face a marginal withdrawal rate of 100 per cent or more, from 32.1 per cent to 5.6 per cent. This is principally because most pensioners who are currently entitled to the MIG (and so face pound-for-pound withdrawal) will in the future face lower withdrawal rates—indeed, their EMTR will be just 40 per cent if they are on no other means-tested benefits.[32] There is also a decline in the number of people in the 80 per cent to 90 per cent bracket, because the increase in the generosity of housing benefit and council tax benefit needs has decreased the number of families whose marginal tax rate is made up of a 20 per cent withdrawal rate from council tax benefit and a 65 per cent withdrawal rate from housing benefit. A further effect reducing the number of people on high marginal tax rates is the government's decision to increase the tax allowances of pensioners so as to ensure that no single pensioner (without special income support premiums) can both receive pension credit and pay income tax.

Table 2

Proportion of Adults Aged 65 or Over Facing Various Marginal Tax Rates, Before and After the Reform

Effective marginal tax rate
(per cent points)
0 per cent
18.2 per cent
13.5 per cent
<50 per cent
42.2 per cent
56.7 per cent
50 per cent-69.9 per cent
0.6 per cent
14.9 per cent
70 per cent-79.9 per cent
0.5 per cent
0.4 per cent
80 per cent-89.9 per cent
5.5 per cent
1.3 per cent
90 per cent-99.9 per cent
1.0 per cent
7.7 per cent
100 per cent +
32.1 per cent
5.6 per cent


  The percentages represent fractions of the total population of adults aged 65 or over in private households in Great Britain. Their total number is 8.54 million. An effective marginal tax rate is the proportion of a small increase in income consumed through forgone benefits and taxation combined. Tax rates are estimated by considering the effect of an increase in private pension income.


  IFS tax and benefit model, TAXBEN, run using data from the Family Resources Survey, 1998-99.

  There are some effects working in the other direction. The number of individuals with a zero EMTR has declined, because the increased generosity of the means-tested benefit system necessarily means that more individuals are eligible for help, and so the number of people facing the effects of the withdrawal of benefits has also increased. Also, there is a very large rise in the number of people facing EMTRs of between 90 per cent and 100 per cent. The most important single reason is that the reform produces a group of individuals who will see their council tax benefit, housing benefit and pension credit simultaneously withdrawn: their marginal tax rate will be 91 per cent. These are mostly individuals who, before the reform, were already having both rent and council tax rebates tapered (giving rise to an EMTR of 85 per cent), so the increase in their EMTR is modest.

  A more general sense of the reduction in EMTRs is achieved by looking at the change in the mean EMTR for people aged 65 or over. Before the reform, it was 47.8 per cent; post-reform, it declines to 39.2 per cent. A final measure is the number of people experiencing increases or decreases: 25.3 per cent of adults aged 65 or over will see their marginal rate increase; 30.9 per cent will see it decline.

  So, it seems that EMTRs will generally fall, and they will fall most where they were previously highest. Although, as we discussed in Section B4, the increase in needs in the rebate system means some pensioners see their EMTRs increase sharply because they will become entitled to partial housing benefit for the first time, in practice their number seems to be small. A possible explanation is that there are relatively few pensioners living in rented accommodation whose incomes are sufficient to exhaust housing benefit entitlement, even before the increase in "needs". The immediate practical effect on EMTRs therefore seems strongly positive.

  This is a much more emphatic conclusion than we were able to reach when we last analysed the pension credit.[33] In part, this is because we were concerned with government pension policy more broadly last time (in particular, we considered the increase in the MIG as a reform, whereas here we take it as a given). But, in part, it is because the increase in the 'needs' in the rent and council tax rebate system that the government has announced has dramatically reduced the numbers who will be left facing very high marginal rates after the reform.

  In the long term, though, there are factors that will tend to increase the number of pensioners on high EMTRs. First, this is because the gap between the state pension and the MIG is set to increase very rapidly, which should increase the number of pensioners who will have pension credit entitlement and so will face the effects of its withdrawal (see Section A and footnote 5). Secondly, because of the rapid growth in typical MIG entitlement that this produces, there will likely be a growing range of incomes over which pensioners would be liable to have both rebates and pension credit withdrawn at once.[34] Thirdly, as the married couple's allowance for pensioners continues to be phased out, there will be an increasing number of pensioners who both pay tax and receive pension credit.[35]


  Since we last analysed the pension credit, the government has announced significant new details concerning the implementation of the reform. We have learnt about how council tax benefit and housing benefit will be made more generous for pensioners. This development has very significantly increased the cost of the reform, but it ensures that more pensioners will gain in full from the reform and that their effective marginal tax rates will typically be lower after the reform than they otherwise would have been. In addition, the policy towards capital has changed in an advantageous manner, and we have learnt a little more about how the new means test will work.

  The main outstanding questions concern longer-term issues. In particular, there are questions about how much the pension credit will cost in the future and about its effects on long-term incentives to save, given that it seems likely that it will cover a growing proportion of pensioners. Most particularly, it remains unclear how the credit fits in with other parts of the government's pensions strategy, notably stakeholder pensions for those on moderate incomes and the state second pension. For the growing number of middle-income pensioners who seem likely to be floated onto the benefit, given current government policy, the gains from having built up pension rights will be reduced.

Tom Clark
Senior Research Economist

February 2002

1   Precise figures will depend on the inflation rate. The couple rate applies where one partner is without a full National Insurance contribution record. If both have full records, they receive twice the single rate. Where one record is partial, the couple receives whichever is the greater of the couple rate and the sum of their individual entitlements. Back

2   This happened under both Conservative and Labour governments. For a 65-year-old pensioner, the real-terms increases in the maximum means-tested benefits were 7 per cent between the introduction of income support in 1988 and 1997 and 25 per cent between the election of the Labour government in 1997 and 2001. Back

3   Total inflation between April 1997 and April 2003 would be 16.3 per cent if the government hits its 2.5 per cent inflation target in 2001-02 and 2002-03. This would imply that the gap between the MIG for a 65-year-old and the pension would have to have been just £7.40 in 2003 to have maintained its 1997 value. Back

4   This is in spite of the Chancellor's announcement on pension uprating in the 2001 Pre-Budget Report, which will only produce higher pension increases when inflation falls below the target of 2.5 per cent. Back

5   This holds even if pensioners' private incomes rise in line with prices over time, as long as patterns of state pension entitlement and the distribution of private pensioner incomes have fixed shapes. Back

6   This is true of unearned income. The MIG disregard means that the first few pounds (standardly, £5 for a single person and £10 for a couple) of earnings are exempt from the 100 per cent withdrawal rate. Back

7   The government describes the calculation for those over 65 in two parts-a pound-for-pound income top-up (100 per cent withdrawal) followed by payment of a "savings credit" (at 60 per cent of private income possessed). But the process is more simply captured by being described as a single calculation with 40 per cent withdrawal (100 per cent minus 60 per cent). Back

8   The Pension Credit: A Consultation Paper, Cm. 4900, The Stationery Office, London, 2000. Back

9   Department for Work and Pensions, The Pension Credit: The Government's Proposals, London, 2001. Back

10   T Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no 17, IFS, London, 2001 ( Back

11   Source throughout this section: TAXBEN run on Family Resources Survey 1998-99. TAXBEN produces output in 2001 prices. Modelling assumptions are described at the beginning of Section C. Back

12   Department for Work and Pensions, Income-Related Benefits: Estimates of Take-Up in 1999-2000, 2001 (p 15) (available from DWP website: Back

13   Our modelling assumes constant demographics and that the distribution of private pensioner incomes has a constant shape. The government may invoke more sophisticated assumptions on these points, which could give rise to an additional estimated expenditure. Back

14   Department for Work and Pensions, The Pension Credit: Long-Term Projections, London, 2002 (available at Back

15   The Pension Credit: A Consultation Paper, Cm 4900, The Stationery Office, London, 2000 (pp 22-3). Back

16   Department for Work and Pensions, The Pension Credit: The Government's Proposals, London, 2001 (p 5). Back

17   For more details, see T Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no 17, IFS, London, 2001 (Section B(4)) ( Back

18   The Pension Credit: A Consultation Paper, Cm 4900, The Stationery Office, London, 2000 (p 20 and p 24). Back

19   Withdrawal rate of 91 per cent calculated as follows: [pension credit taper] plus [(rebates taper) times (income kept after pension credit taper)], ie 0.40 + 0.85 (1-0.40) = 0.4 + (0.85 × 0.6) = 0.4 + 0.51 = 0.91. Back

20   The savings credit is potentially worth more to people with especially high MIG entitlement, such as those with severe disability premiums. The reform's maximum worth is the reduction in the taper rate (0.60) multiplied by the MIG entitlement affected. In the typical case, the benefit affected is £23 (the difference between a basic pension of £77 and MIG of £100), making the maximum value of the reform £13.80. Needs for the rebates will rise by just this amount. But where the relevant MIG is higher, then the affected MIG entitlement is also higher, as is the maximum gain. To ensure that none of the gain is consumed in forgone rebates in such cases, needs would therefore have to be increased especially sharply. Back

21   The Pension Credit: A Consultation Paper, Cm 4900, The Stationery Office, London, 2000 (p 23). Back

22   Department for Work and Pensions, The Pension Credit: The Government's Proposals, London, 2001 (p 10). Back

23   T Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no 17, IFS, London, 2001 (Section B(5)) ( Back

24   T Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no 17, IFS, London, 2001 ( Back

25   However, these numbers are expressed in 2003 prices, and we do not want to have to make too many guesses about what the rest of the 2003-04 tax-benefit system will look like. Instead, we deflate the illustrative numbers by assumed inflation of 2.5 per cent between 2003-04 and 2002-03, and then model the reform as if it were being introduced in 2002-03. Back

26   Each adult aged 65 or over is given equal weight. Couples where both partners are over 65 are therefore given twice as much weight as single pensioners. Back

27   These averages are across all "pensioner families", ie single people over retirement age or couples where the man is 65 or over. This implies differences from the 65+ definition we have used in constructing Figure 4 (see footnote 26). If the same basis as for Figure 4 is used, then the average gains are £5.73 for single people and £6.81 for couples. Back

28   This figure is approximately in the middle of the range of estimates for the non-take-up of entitlement that we cited in Section B1. Back

29   See Department of Social Security, Households Below Average Income, 1998-99, London, 2000 (p 127), which is also based on the 1998-99 Family Resources Survey. It suggests that 21.5 per cent of pensioners lived with below 60 per cent of contemporary median income and that 35 per cent lived with below 70 per cent of contemporary median income. Back

30   There are other possible reasons for the divergence-for example: the income definition is slightly different from the before-housing-costs measure in the HBAI; the table is only for those aged 65 or over, whereas the HBAI includes women of any age if they are married to men aged 65 or over. Back

31   For a fuller discussion, see T Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no 17, IFS, London, 2001 (Section B(6)) ( Back

32   The main group of pensioners left with an EMTR of 100 per cent are those with pre-means-tested-benefit incomes of less than the basic state pension. Back

33   See T. Clark, Recent Pensions Policy and the Pension Credit, Briefing Note no. 17, IFS, London, 2001 (Section B(3)) ( Back

34   This will happen, for example, if rents and local taxes increase more slowly than the maximum income required to exhaust pension credit, which will rise more rapidly than earnings. In these circumstances, over time, some of those initially entitled to rebates but not pension credit would become entitled to pension credit and so face an increased EMTR. In addition, if rebate "needs" increase more slowly than the income that exhausts pension credit entitlement, then the proportion of the band of incomes over which the savings credit is received before the tapering of rebates begins will decline. Back

35   This will apply in couples where neither partner was born before 1935 and where the joint income is sufficiently low to attract pension credit, but where income is concentrated on one partner, which can mean that his/her individual income exceeds his/her personal tax allowance. Back

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2002
Prepared 12 April 2002