Select Committee on Work and Pensions Minutes of Evidence

Memorandum submitted by PricewaterhouseCoopers (PC 06)


  1.  The pension credit, like other state pensions policies, can usefully be assessed against the following criteria:

    —  Does it help to ensure an adequate minimum level of income for all pensioners?

    —  Will it be affordable in the long run in the context of other demands on the public purse over this period?

    —  Will it adequately incentivise people to save for their own pension provision to the extent they can afford to do so?

    —  Will it allow people to plan for their retirement with some confidence about levels of future state provision and options for private provision?

  2.  Inevitably, no pensions policy regime will score perfectly on all criteria. There are, for example, unavoidable trade-offs between affordability and the minimum level of income provided and between simplicity and means-tested targeting. In the case of the pension credit we conclude, as discussed in more detail below, that (assuming earnings indexation of the minimum income guarantee):

    —  the pension credit should ensure a minimum income level of just over 20 per cent of UK average full-time earnings, with benefits also extending to pensioners on moderate incomes;

    —  the proposed new savings credit could add around 1 per cent of GDP (around £10 billion in 2001-02 values) to UK state pension spending by 2050; this would not be unaffordable in itself but would require either lower growth in other spending areas or a higher ratio of tax to GDP;

    —  by 2050, our estimates suggest that up to 70 per cent of pensioners could be eligible for some pension credit payment, which could have adverse effects on incentives for low and moderate earners to save for their own retirement provision; and

    —  the proposed new regime from 2003-04 (with the pension credit, the State second pension and stakeholder pensions) will be highly complicated, making decisions on retirement provision difficult for many people.


  3.  Provided that the minimum income guarantee (MIG), which will be incorporated within the pension credit (PC) from 2003, continues to be indexed in line with average earnings rather than prices, this should ensure a minimum income level of just over 20 per cent of UK average full-time earnings (and more than this for those with second pensions or other income sources). By 2050, however, the basic state pension element in this will have fallen to only around 6-7 per cent of UK average earnings, meaning that the majority of this minimum income level will have to come from some combination of the new State second pension (S2P) and the guaranteed element of the pension credit (for pensioners with little or no other income).

  4.  The guaranteed minimum income level of just over 20 per cent of UK average earnings is similar to that provided by the basic state pension (BSP) in 1980, prior to its gradual erosion through price indexation. But to achieve this from the current starting point would require not just indexing the basic state pension to earnings, but also increasing its level to match the MIG. This would make the pension credit redundant, but would be relatively costly since the benefits would go to all pensioners, not just those on low and moderate incomes. As discussed below, this could be funded by, for example, increasing the state retirement age.


  5.  We would recommend that the Government should publish its own estimates of the long-term costs of the pension credit, including comparisons with other possible policy options, as soon as possible. In the absence of any such official projections beyond 2004-05 at present, however, we have produced our own projections, as detailed in an earlier PwC fiscal policy briefing paper.[46] As with any such long-term projections, these are subject to significant uncertainties, but we consider the analysis to be useful in identifying likely broad trends in total UK state pension spending with and without the pension credit. The latest version of these projections is summarised in the table below.

Table 1


Per cent of GDP
Basic Pension
MIG only
Total spending(without SC)
Cost of savings credit1 (SC)
Total spending (with SC)

  1Additional cost of savings credit element in Pension Credit, over and above cost of MIG.


  PwC estimates based on an updated version of the methodology set out in the accompanying earlier paper (see footnote). The precise cost estimates in this table are slightly different from those in the earlier paper as a result of further work we have carried out, which is to be published in full later this year, but the general conclusions from the analysis are unchanged.

  6.  We can see from the table that the cost of the BSP trends down as a share of GDP due to price indexation, which generally outweighs the impact of there being an increasing total number of pensioners over time. The exceptions to this rule are the period between 2000 and 2010, due to recent above-inflation rises in the BSP, and the period between 2020 and 2030, when the baby boomers retirement rate peaks.

  7.  Lower BSP costs over time are offset by the rising cost of the new State second pension, in line with GAD projections, and by increased MIG payments. The latter arises from our assumption that the MIG is linked to earnings, so that the gap between the BSP and the MIG rises steadily over time. By 2050, for example, the MIG is still projected to be around 21 per cent of average full-time earnings, while the BSP is projected to be only around 6-7 per cent of average full-time earnings in 2050.[47]

  8.  Without the pension credit, total state spending on pensions might therefore show a broadly stable profile over time at around 5 per cent of GDP up to 2040 and then resume a gradual downward trend after 2040 as the ratio of workers to pensioners stabilises and as price indexation of the BSP dominates other effects on total pension spending. The regime without the pension credit therefore appears affordable on any reasonable definition.

  9.  As Table 1 shows, the additional cost of the proposed new savings credit, estimated at around 0.2 per cent of GDP in 2004-05 (ie around £2 billion) by the Treasury, rises gradually at first to 0.4 per cent of GDP in 2020 and then more rapidly to around 1 per cent of GDP in 2040 and 1.2 per cent of GDP in 2050. This reflects the cumulative impact of four effects:

    —  the rising total number of pensioners from just under 11 million in 2000 to around 16 million in 2050, with the rate of increase being greatest between 2020[48] and 2040;

    —  the falling level of employment between 2010 and 2040 as the working age population is assumed to decline while participation and unemployment rates remain constant; this reduces real GDP growth below the assumed 2 per cent rate of productivity growth (and thus below assumed real average earnings); the denominator of the pension spending ratio (GDP) therefore rises by less than earnings in the period 2010-40;

    —  the increasing proportion of pensioners eligible for the savings credit over time; and

    —  the increasing average savings credit payment over time, relative to earnings.

  10.  The latter two effects require some further explanation. The key point to note is that the savings credit ceiling[49] (SCC) will tend to rise faster than earnings over time, assuming no change in the 60 per cent savings credit rate (SCR). This follows from the formula:

    (1)  SCC = MIG + [SCR/(1-SCR)]*(MIG—BSP)

  11.  In 2003-04, using the illustrative values in the DWP paper for a single pensioner, this gives:

    (2)  SCC = £100 + [0.6/0.4]*(£100—£77) = £134.50 per week (or 134.5 per cent of MIG)

  12.  Over time, the first term in equation (1) above, namely the MIG, will rise with earnings, but the second term, which is linked to the difference between the earnings-indexed MIG and the price-indexed BSP, will rise significantly faster than earnings. After 50 years, for example, with 2 per cent real earnings growth, the SCC in 2003-04 prices will be:

    (3)  SCC = £269 + [0.6/0.4]*(£269—£77) = £557 per week (or 207 per cent of MIG)[50]

  13.  By 2050 then, assuming no change in the savings credit rate, the SCC will be just over twice the MIG or around 43 per cent of average full-time earnings, compared to around 28 per cent of average earnings in 2003-04. Our modelling work suggests that the median recipient in the fourth pensioner income quintile will be eligible for at least some payment under the savings credit regime by 2050, implying that around 70 per cent of pensioners could receive means-tested benefits at that time, compared to a government estimate of around 50 per cent when the savings credit is first introduced in 2003-04.

  14.  Given projected pensioner numbers, this implies that the total number of recipients of the means-tested pension/savings credit is likely to more than double from around 5.4 million initially to around 11.2 million by 2050, although this is only an approximate estimate given uncertainties about the precise pensioner income distribution at that time. Furthermore, since in the long run some amount of savings credit will be paid on an ever-widening slice of income (ie from the BSP[51] up to the SCC), the average amount will also tend inexorably to rise relative to average earnings. Over a 50-year period, our estimates suggest that the average payment could more than double relative to average earnings, given our assumptions on the growth of additional pensioner incomes over and above the BSP. Again, our precise estimates are subject to many uncertainties, but the direction of change, and the intuition that the magnitude of the increases is likely to be material when accumulated over a long enough period, seems relatively clear and robust given the gearing effect implicit in the pension/savings credit formula.

  15.  In summary, our cost estimates suggest that:

    —  without the new savings credit, UK state pension spending looks set to remain broadly stable at around 5 per cent of GDP over the next 40-50 years; and

    —  with the savings credit, the long-term cost would increase by something of the order of 1 per cent of GDP (around £10 billion at 2001-02 values), to around 6 per cent of GDP.

  16.  It would be misleading to describe the latter projection as unaffordable. Both the increase in costs up to 2050 and the overall level of UK state pension spending in 2050 would still be low relative to government spending on pensions in other large EU countries, which face much more severe financial pressures from an ageing population given their more generous state pension schemes. In the UK, there would be many possible options to fund this additional cost through tax increases or net cuts in spending (as a share of GDP) in other areas. Given that pensions payments are clearly current spending not investment, however, the option of passing this cost on to future generations through additional borrowing would appear to be ruled out by the principles of the Golden Rule (ie borrowing is only allowed for net investment).

  17.  Although the costs of the savings credit could be met, however, our results are of some significance in that they belie previous long-term projections[52] published by the UK government, which have generally suggested that UK state pension costs will be stable or even declining as a share of GDP in the long run. These earlier projections suggested that this would leave room for an increased share of national income to be spent on health, education and other priorities without requiring tax increases or spending cuts elsewhere. The projections in Table 1 above suggest that the price for more generous state pensions provision will inevitably be hard choices elsewhere, even if not for some considerable time (ie until after 2020).

  18.  On the other hand, our model estimates suggest that the savings credit would be a cheaper option than earnings indexation of the BSP, in which case total UK state pension spending might rise to around 7.5 per cent of GDP by 2050. This could, however, be offset by scaling down the State second pension and/or increasing the state retirement age.


  19.  The new savings credit might increase incentives to save for those with additional incomes up to the MIG, where the "claw-back rate" would now be 40 per cent (although this is still a relatively high effective "benefit tax" rate) rather than 100 per cent. But it could reduce incentives to save for those with incomes between the MIG and the savings credit ceiling, since for these people the claw-back rate will rise from zero to 40 per cent. This is the familiar problem in welfare reform that, by lowering taper rates, you weaken "poverty trap" effects of means tests but extend them higher up the income distribution (perhaps as high as 70 per cent of all pensioners by 2050 on our admittedly rough estimates, compared to around 50 per cent immediately after the introduction of the savings credit in 2003).

  20.  The net effect on savings incentives in the population as a whole is therefore unclear, particularly when potential additional effects from higher expected future incomes due to the savings credit (and indeed other new initiatives of recent years such as S2P and the increase in the MIG) are taken into account.

  21.  Other policy options, involving a higher basic state pension and/or a more generous state second pension might score better on this criteria by reducing or eliminating means-testing, but at an additional cost that might need to be offset by, for example, raising the state retirement age.


  22.  This is inevitably a somewhat subjective criterion, at least compared to adequacy and affordability, but it is nonetheless important. Our judgement is that the pension/savings credit regime does not score particularly highly on this criterion because of the complexity of the formula, the relatively high level of means-testing implied and the fact that, together with the S2P, it greatly complicates decisions on whether people on low to moderate incomes would benefit from taking out additional private provision. In particular, it is not clear that there is now a good case for this income group to take out a stakeholder pension given the greater potential generosity of state provision (including the S2P).

  23.  The other problem with a complex regime of this kind is that it would be relatively easy for a future government to change the details of the regime in significant ways that most people would be unlikely to understand fully, as indeed has occurred with SERPS since the early 1980s. A simpler regime would be more transparent and, arguably, more robust against political change, although clearly this can never be guaranteed.

  24.  These considerations, together with the other arguments presented above, suggest that there might be a case for considering a simpler state pensions strategy based:

    —  either on a more generous state second pension, with the pension credit being phased out gradually over time as this comes into effect;[53]

    —  or on a significantly more generous, earnings-indexed basic state pension, with the additional costs of this being offset by raising the state retirement age and/or making the state second pension less generous (or even closing this altogether, while protecting existing entitlements under SERPS).

John Hawksworth

Head of Macroeconomics Unit

10 January 2002

46   "State pensions policy and the new pension credit", Fiscal Policy Briefing Paper No. 4, PricewaterhouseCoopers, December 2000. A copy of this report is being sent to the Select Committee for reference to accompany this submission (not published). This includes details of the methodology and key assumptions used, together with a sensitivity analysis. The future evolution of pensioner income distributions is a particularly important source of uncertainty with any such projections. Back

47   This may be a slight underestimate given the new 2.5 per cent minimum indexation guarantee for the BSP, depending on the average level and volatility of inflation. However, any such addition BSP spending will be clawed back in full for those receiving the pension credit and may also be taxed for higher income pensioners. Back

48   Between 2010 and 2020, the numbers over 65 rise significantly, but this is offset by the increase in the female state retirement age. Back

49   This is the level of non-means-tested pensioner income, including the BSP, above which no savings credit is payable under the pension credit regime. The SCC/SCR terminology is our own. Back

50   Note that, in the very long run, the BSP will fall to a negligible amount relative to average earnings and the PCC will asymptotically approach 250 per cent of the MIG, or around 53 per cent of average full-time earnings. Back

51   This assumes that the lower income threshold for payment of the savings credit remains at the BSP. If this threshold were to be raised over time towards the MIG, then the costs would decrease over time relative to our estimates. The Government has not, as far as we are aware, made its intentions clear on this point. Back

52   For example, the Treasury projections included in the March 2001 Red Book (pp 129-131), which took no account of the pension credit (and possibly did not include MIG spending either, although this was not clearly stated). Back

53   It is possible the Government has this first strategy in mind for the long run. If so, it would be helpful for long-term pension planning purposes if it made these intentions clear now. 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