Expensive, ineffective and unequal

Why incentives aren't the answer to the pensions crisis

Executive summary

Opponents of a greater compulsory element in the UK's pension system have suggested that incentives could provide an alternative way of increasing levels of private saving.

This paper considers the effectiveness of incentives and who they would benefit. It concludes that incentives are not an effective enough response to reverse the decline in pension provision in the UK.

Greater incentives are likely to:

• Have less of an impact amongst those who are currently un-pensioned, such as the low-paid;
• Increase complexity and, by extension, costs;
• Be ineffective at overcoming behavioural obstacles to saving; and
• Fail to significantly affect the overall savings level, but rather alter its composition.
• Therefore a greater element of compulsion will be required to increase savings.

Introduction

One of the central questions in the UK's current debate on pensions reform is how to increase the level of funded retirement savings. The TUC's position is that a greater element of compulsion should be introduced, requiring all employers and their employees to contribute to a pension.

Others argue that compulsion is a blunt instrument and that a preferable option would be to provide greater incentives to save. This is often portrayed as a less painful option, which would still achieve higher savings levels.

This paper considers the impact of incentives on the savings market and whether they could offer a solution to the pensions crisis.

Who isn't saving?

It is useful to start an analysis of the potential impact of incentives by considering what types of individuals are typically not currently saving for retirement. Here the results are probably what one would expect. Existing pensions coverage is lower amongst lower-income groups and part-time workers, it also varies by size of employer and the individual's occupation [1] .

For example the table below gives a breakdown of occupational pension scheme membership by broad socio-economic group.

Occupational pension coverage by socio-economic classification

	Men full-time	Women full-time	Women part-time
Managerial and professional	67%	71%	57%
Intermediate	64%	60%	44%
Routine and manual	42%	41%	26%
Total	55%	60%	33%

Source: Living in Britain General Household Survey 2002, table 6.4

It is also noticeable that the proportion of employees who are not members of an occupational scheme, even where the employer offers one, also varies by socio-economic group [2] .

Importantly those who are currently not saving are also less likely to consider themselves able to make financial decisions. Research by the Financial Services Consumer Panel has found a clear link between incomes and financial literacy [3] . Put simply those on lower incomes are significantly less likely to identify themselves as being financially literate.

These findings have significant implications for those who think that incentives can address the shortfall in savings. Consider that those who are currently not saving - into a pension and elsewhere - are generally on lower incomes and are less financially literate. In contrast those who are saving are generally better off, more likely to already be in a pension scheme and also regard themselves as more financially literate and therefore more confident in taking financial decisions.

For incentives to be successful it would require a certain element of financial literacy and a willingness to voluntarily commit savings. Many lower paid workers who do not currently save may not feel they have enough spare money to voluntarily commit to save or feel sufficiently confident to make financial decisions. In contrast the better paid who are more financially aware and already committed to saving may simply transfer their capital to the most tax advantageous vehicle.

In addition incentives add an element of complexity to the system, reinforcing the need for advice, yet as we have seen uncertainty about where to go for cost-effective independent advice is one of the factor contributing to an unwillingness to save. Again those who are better off are more likely to already use a financial adviser so are more likely to respond.

Who gets the money?

The next question to be considered is who predominantly benefits from financial incentives to save. Looking at the current system, the primary existing financial incentive in the UK's pension system, tax relief on contributions, comes at a significant cost. The bulk of tax relief goes to employers and the cost of relief solely on employers' pension contributions is estimated to be £10.65bn a year.

Cost of tax relief on pension contributions

	2001-2002	2002-2003	2003-2004
Employee contributions to occupational pensions	£3.3bn	£3.4bn	£3.5bn
Employer contributions to occupational pensions	£6.5bn	£7.7bn	£9.8bn
Employee contributions to personal pensions	£0.9bn	£1.15bn	£1.1bn
Employer contributions to personal pensions	£0.71bn	£0.81bn	£0.85bn
Employee contributions to FSAVCs	£0.17bn	£0.15bn	£0.13bn
Tax raised from pensions in payment	£7.5bn	£7.7bn	£8bn

Source: Inland Revenue table 7.9

As the Government has made clear employers do well from the current system:

'Employer contributions to a pension are one of the most tax-privileged forms of saving and the administrative costs of employer provision are often less than those paid on individually arranged personal pension schemes.' [4]

Within the tax relief on employee contributions there are two rates that correspond with income tax bands. This means that, even at present, the incentives in the pension system are skewed towards the better-off - those who pay income tax at the higher rate. Put simply a £1 pension contribution effectively costs a higher-rate taxpayer 60p, whilst it costs a lower-rate taxpayer 78p. Overall 55% it is estimated of tax relief it goes to 2.5 million high rate taxpayers, whilst the remaining 45% goes to 13 million lower rate taxpayers [5] . This distribution is illustrated in the table below.

Tax incentives for occupational pensions

Median annual earnings (£)	£ millions	%
9,000	314	1.2
15,000	2,174	8.1
25,000	4,800	18.0
33,000	1,810	6.7
38,000	4,361	16.3
50,000	13,257	49.6

Source: ABI, 2005

The experience of stakeholder pensions is also worth considering. Stakeholders were launched with the aim of offering an affordable pension to many workers currently not saving. There was a specific focus on those on low to moderate earnings, and those in sectors with low pensions coverage.

Because of the way that stakeholders were set up, allowing contributions to be made on behalf of another person, wealthier individuals in particular have set up stakeholders for children and spouses. It now appears that many wealthy families are taking advantage of stakeholder pensions.

Contributions to stakeholder pensions

	2001-2002	2002-2003
Employee £0 - £9,999	£510	£880
Employee £10,000 - £19,999	£520	£980
Employee £20,000 - £29,999	£680	£1,500
Employee £30,000 +	£2,370	£3,930
Child	£1,320	£1,430
Full-time education	£1,730	£1,760
Carer	£1,570	£1,940
Unemployed	£2,620	£2,430
Other	£1,590	£1,500

Source: Inland Revenue table 7.10

According to analysis of contributions carried out by the Inland Revenue annual payments into pensions on behalf of children, students and the unemployed are typically higher than those made towards employees' stakeholder schemes for those earning below average wages.

The number of individuals involved is about 60,000 people in total out of an overall stakeholder membership of 1.06 million, but the amount contributed is significant. For example, in the 2002-2003 period total contributions to this group were half those in the same period for all employees earning under £10,000, despite the fact they represented a much smaller group [6] .

We can see that the existing incentive structure of tax relief favours the better off. In addition the experience of stakeholder suggests that any further incentives to save are more likely to be utilised by the better off. In turn the impact on the savings habits of those with lower incomes may be limited.

Do incentives result in savings displacement?

A further question that should be raised about incentives is whether they would actually affect overall savings levels. It is a common argument against greater compulsion to save that such an intervention would simply displace existing savings from elsewhere (although the experience in Australia provides a rather more complex picture [7] ). However there is no reason why the introduction of greater incentives to save for retirement would not have the same impact.

Based on an assessment of various studies the Sandler review concluded that the introduction of tax incentives maybe more likely to have an impact on the composition than the overall level of savings. There was significant research supporting this argument.

'Targeted tax incentives, which only increase the rate of return to saving in particular forms, are likely to result simply in portfolio reallocation.' [8]

Although this is simply a comment on the allocation of capital between pension and non-pension savings by individuals, actually there is evidence of switching between types of retirement savings by employers. Interestingly one of the most pertinent examples of this comes from America.

Many advocates of incentives believe there are lessons to be learnt from the introduction of 401(k) plans in the US and their subsequent success. These plans include incentives for directors to maximise coverage.

It is certainly true that these plans have been popular. From their introduction in 1978 they now account for a reputed one in every three dollars invested in employer-sponsored pensions [9] . But their impact on overall coverage is negligible. In fact a range of studies have concluded that rather than boosting pensions coverage 401(k) plans have merely changed the composition of retirement savings.

'Given their popularity and growth, one would have thought that the introduction of 401(k) plans should have boosted pension plan coverage in the United States. But... overall pension coverage has remained virtually unchanged. This means that the enormous expansion of defined contribution plans, especially 401(k)-type plans, has produced a sharp decline in the percent of the workforce covered under traditional defined benefit plans.' [10]

So whilst there has certainly been a significant increase in saving through 401(k) plans, at the same time there has been a decrease in other forms of pension savings. So at the level of pension savings as a whole what has really occurred is a form of 'portfolio reallocation'.

A recent analysis of the Australian compulsory superannuation system also queried whether greater incentives could affect overall savings levels.

'[W]hen tax concessions are introduced, voluntary superannuation can provide higher returns than other forms of saving. While tax incentives can encourage households to save more in superannuation, it is less clear whether they increase total saving. Households that would otherwise consume all their income might decide to save in tax-advantaged superannuation to take advantage of the higher returns and increase lifetime income. However, households that already save voluntarily may merely substitute into superannuation. They may even save less overall since they no longer have to save as much to achieve the same level of lifetime income.' [11]

So it appears that incentives result in exactly the displacement of savings often predicted to result from greater compulsion. In addition the displacement effect likely to occur in response to incentives may be exacerbated by some of the factors explored earlier. A lack of financial literacy is likely to mean that non-savers fail to react as intended, particularly the lower-paid. In contrast the well-off and more financially literate may recognise that pension saving had become more attractive and divert capital from elsewhere (although this maybe tempered by the knowledge that the capital invested was inaccessible).

Do incentives make the system more complicated?

A further problem with the introduction of further incentives into the pension system is that it would make matters even more complicated. Most commentators argue that the UK system is already very complicated and hard for those unfamiliar with pensions to understand. Yet incentives may add to this.

Enhancing the appeal of pensions as a form of savings through more favourable treatment would need to be communicated clearly if it were to act properly as an incentive. In addition it would further the need for advice for those not confident in their own financial decision-making. This would appear to make decisions more complicated, more costly (since advice would need to be paid for by someone) and therefore less efficient.

Having considered precisely these arguments, the Sandler review concluded that further incentives would add to complexity, and by extension to costs, and that the Government should therefore avoid them.

'[T]here is little evidence to suggest that tax incentives have a significant impact on overall savings levels, especially amongst the lower-income groups for whom increasing saving should be a priority. What is more, it is evident that such incentives generally increase the complexity of the regime as a whole, and that this complexity leads to higher costs.' [12]

Behavioural evidence against incentives

We also need to examine the role of individual behaviour in savings decisions. Economics typically assumes that individuals make rational decisions based on all the information available to them in a way that has the most beneficial outcome to them. If one assumes that savers are rational then greater incentives may appear to be a useful tool to increase savings levels. By increasing the attractiveness of a particular option, for example by altering the way it is taxed, it is hoped that rational individuals should assimilate the new information and respond to it.

Unfortunately it does not appear that that individuals necessarily do make decisions in this way. For example, the overall level of membership of particular pension schemes is often very low. In just one recent study by Hewitt Bacon & Woodrow the typical membership of a defined contribution scheme was under half of those eligible [13] . This was despite the fact that by joining the scheme the individual members could often obtain a significant employer contribution.

'What do employers have to do to entice people to join? If you give someone a 7 per cent pay rise they will jump at it, but if you say you are going to add 7 per cent to their pension they are reluctant.' [14]

Across the working population as a whole approximately 13% of full-time employees and 16% of part-time employees do not join pension schemes even when they are eligible to do so [15] . This is despite the fact that many employers contribute to pension schemes, and that there is tax relief on contributions.

It is instructive to look at the reasons that people give for not saving. Research by the consumer campaign group Which? found that by far the most common reason given was that they could not afford to save more at present [16] . This was followed by confusion over the products available, and the expense of financial advice. Research carried out by the Financial Services Authority found very similar results [17] . Looking from the other perspective Which? also found that the factor cited as most likely to encourage saving was having more spare cash.

Reasons for not saving

	Agree strongly	Agree slightly	Total
Cannot afford to save more at the moment	52%	22%	74%
Available savings products too confusing	21%	38%	59%
Too expensive to pay for the advice about saving	21%	26%	47%
Do not know where to find unbiased advice about the products on offer	15%	27%	42%
Savings products do not give more value for money	9%	29%	38%
Do not know where to get advice about saving	9%	24%	32%
It never occurred to me to save	7%	14%	21%
Too young to think about saving	7%	13%	20%

Source: Which? Financial Advice Market Research report, May 2002

Of course, if one believes that individuals respond rationally to the choices they face, it could be argued that those who do not join pension schemes are making a rational decision. Like the non-savers highlighted above, they may not feel able to afford to save and they may have other priorities, for example the need to pay off debts, that lead them to choose not to save.

However this interpretation is difficult to maintain when the experience of auto-enrolment is considered. Where pension schemes operate auto-enrolment the individual will become a member of the scheme automatically unless they make a conscious choice to opt out. Numerous studies show that when auto-enrolment is applied pension scheme membership as a proportion of those eligible is much higher [18] .

If all non-joiners were making a rational and conscious choice there should be no significant difference between the membership rates of schemes with auto-enrolment and those without. This suggests that failure to join, or contribute to, a pension scheme is not simply a rational economic decision but the result of a combination of factors.

Similar evidence of sub-optimal financial decision-making emerges from when the asset allocation decided upon by participants in defined contribution schemes is analysed.

'[W]hen people are confronted with a large number of [investment fund] options, they typically adopt a strategy of dividing their contributions equally among the options - 50-50 if there are two; 25-25-25-25, if there are four; and so on. What this means is that whether employees are making wise decisions depends entirely on the options that are being provided for them by employers.' [19]

The result in practice is that schemes that offer proportionately more equity funds find that their members have proportionately more invested in equities [20] . Once again this undermines the idea that the choices made by individuals in relation to saving are inherently rational.

In fact it is increasingly accepted that financial decision-making is considerably less optimal than is sometimes assumed. The growth of the fields of behavioural finance and behavioural economics, which seek to use insights from psychology to understand how individuals act, represent a significant development.

The interest in behavioural finance within financial institutions is particularly telling since it is a recognition that even the high priests of finance operating within the capital markets can exhibit profoundly irrational behaviour [21] . Such views are, of course, not new as the irrationality prevalent in the stockmarket was a topic discussed at length by Keynes. It does however appear that questions of actual human behaviour are being taken more seriously in policy debates on savings.

For example, the insight provided by behavioural economics is discussed in the preliminary report by the Pensions Commission produced last year [22] . It highlighted 'procrastination' and 'inertia' as two factors that influenced the decision of employees to join a scheme or begin saving.

Significantly these factors can continue to have an impact even when non-joiners are specifically targeted with information. When presented with evidence of their under-saving individuals often say they will act, but research suggests that actually few follow through on their initial intention.

The idea that people do not necessarily make economically rational decisions is challenging for those who advocate greater incentives to save. Given considerable encouragement at present many employees choose not to save, and even when targeted with information they continue to choose not to do so. It is not clear therefore why greater incentives would overcome these factors.

Would further incentives make much of a difference?

The ultimate question to be answered by proponents of incentives is would they have a significant impact on savings levels.

As we have seen a range of factors make it unlikely that incentives would have a significant impact. Savers do not always make the best decisions possible and may not respond even when saving more would clearly benefit them. Those who currently do not save are less likely to be responsive to incentives, due to their lower incomes and lack of financial literacy. In contrast the better off are more likely to take advantage of favourable tax treatment. Finally, evidence suggests that overall incentives would probably have more of an impact on the composition of savings as opposed to overall levels.

Perhaps not surprisingly then the Sandler review concluded tax incentives would have little impact in terms of increasing overall savings levels. In fact it went further and recommended that the Government actively avoid the introduction of further tax incentives:

'The Review therefore recommends that, in future, governments should avoid introducing new tax-based savings incentives if their aim is to increase aggregate savings levels. The core objective of policy in this area should be simplification.' [23]

Research for Age Concern carried out by the Pensions Policy Institute reached similar conclusions.

'? There is no evidence that tax incentives increase the overall level of saving. They are complex, do not appeal to their target group, and do not solve the basic problem for most low income people; that they do not have the money to save.

? Tax incentives can encourage pension rather than other types of saving.

? But tax incentives appear not to have been effective in generating enough pension saving for future pensioners.' [24]

A final and more fundamental point should also be considered. Although incentives are sometimes presented as a 'painless' alternative to compulsion if they can achieve the same coverage as a compulsory system the cost to employers and employees must be broadly the same (since the same levels of contributions are being made). In fact there would undoubtedly be extra costs in terms of Government expenditure, and the continuance of inefficiencies in administration, and sales and marketing costs that could be reduced or eliminated under a compulsory system.

Compulsion versus incentives

To date there has been little direct comparison of specific proposals for incentives against those in favour of greater compulsion in the system. However in June 2005 the Association of British Insurers produced a report comparing at the impact of its proposed Pension Contribution Tax Credit for employers with other possible policies such as auto-enrolment and compulsion.

Using the most optimistic assumptions the report concludes that the ABI's proposed tax credit would only increase savings by up to £1.2bn [25] , whilst compulsory contributions set at 3% from both employer and employee would increase savings by £5.3bn [26] . Even auto-enrolment, which as explained earlier is in essence a form of compulsion, is estimated to be able to potentially boost savings by £2.5bn. This is after taking into account displacement of other savings.

As the ABI report stresses, simple figures on overall savings levels need to be balanced against the cost to Government (assuming tax relief is retained) and consideration of whether the right groups are saving. However it is fair to draw the conclusion that using compulsion in either membership or contributions would result in a significantly greater levels of total savings than an incentives-based approach.

Auto-enrolment: compulsion by another name?

Some opponents of compulsion suggest auto-enrolment as an alternative option. Under this approach employees automatically join their employer's scheme unless they actively chose to opt out. In some cases it has further been suggested that individuals should only be allowed to opt out if they can demonstrate that they are making other arrangements. As we have seen auto-enrolment does have a significant impact on membership of pension funds.

However if introduced on its own auto-enrolment could potentially create further inequality in the pension system. Those sectors where employers generally offer a pension would see coverage increased further. In contrast in sectors where employer pension provision is patchy the impact would be much less. This would have the unintended consequence of increasing labour costs for some sectors whilst leaving others relatively unaffected.

It should be noted that support for compulsion amongst some employers is already rising, in part because they do not wish to be undercut on pension costs by employers who do not provide or contribute to a scheme [27] .

Therefore in order to combat the inequality in employer pension costs it is likely that it would be necessary to also auto-enrol employees who are offered access to a stakeholder pension scheme by their employer. Since this should apply in all businesses with 5 or more staff where there is no occupational scheme this would certainly boost coverage.

However there would be a further difficulty. If left at this basic level auto-enrolment would mean compelling many employees to join DC schemes without a contribution, since this is how most stakeholder schemes are structured. This would clearly be a questionable policy. Therefore if auto-enrolment were extended to include those employers at present only offering access to a stakeholder it would be sensible to also require that the employer contribute to the pension to make it worthwhile for the employee to join.

Such a proposal sounds reasonable. It is also almost the same in practical terms as a compulsory system. The TUC argues that all employees should be members of a pension scheme and that all employers should be required to contribute a minimum amount. Auto-enrolment with a minimum employer contribution is the same in practice.

Assessing the evidence

It should be clear that an incentives-based approach to increasing the amount of funded retirement saving will have a limited impact at best. Incentives do not have a strong enough pull to change the behaviour of significant numbers of savers, and those that do respond to them are likely to come from groups already saving. The result may simply be a change in the composition of savings rather than an increase. In addition further incentives could increase complexity and costs.

In contrast compulsory approaches are likely to significantly increase savings overall and amongst the right groups. Whilst there have been concerns this could lead to over-saving, or divert savings from elsewhere, these must surely be outweighed by the evidence that compulsion would overcome the significant behavioural barriers to save. The experience of auto-enrolment demonstrates that many non-savers are not making an active choice.

The TUC therefore welcomes the developing debate on the merits of auto-enrolment, but we believe this policy should be seen for what it is - a form of compulsion. In addition it is important to stress that unless auto-enrolment is applied widely it may exacerbate the differential nature of pensions coverage.

Throwing money at the problem

Put in stark terms the argument in favour of greater incentives is that the way to increase saving is by the Government using taxpayers' money to make the private savings system more attractive and hence boost savings. Regardless of the merits or otherwise of this as an approach it is clearly an inefficient way to tackle the problem. As former NAPF chairman Alan Pickering has argued, if that is the approach it might simply be better to spend the money on bigger state pensions.

'[C]areful thought needs to be given to the value for money that will flow from further positive incentives, particularly when we are calling for an increased use of taxpayers' money to boost the basic state pension. Bigger state pensions may, in fact, be a very cost-effective and non-distorting way of increasing the nation's savings.' [28]

In addition, as has been argued above, the impact of incentives to save is likely to be most on the wealthier. So the overall result may be that the Government spends taxpayers' money topping up the retirement savings of those higher up the income scale. This is unlikely to be politically acceptable if it becomes widely understood.

If we understand the type of people who are generally not saving enough, and the reasons why they are not currently saving, it becomes clearer that greater incentives are not the right answer.

It is important to recognise that the evidence from behavioural economics poses some challenges for advocates of compulsion too. Individuals do make irrational decisions. The factors of procrastination and inertia mean that many may choose to only pay the minimum contribution under a compulsory system even when they are exhorted to pay more. In addition the minimum contribution may also act as an 'anchor' for decisions. The range of fund options within any compulsory funded system will also need careful consideration in order that it does not inadvertently lead savers into an inappropriate asset allocation.

However this merely reinforces the need for a compulsory system to be properly structured with the minimum contributions set at a level that will deliver a decent income in retirement. In addition it is clear that whatever issues will need to be addressed under a compulsory funded system it will still be significantly more efficient at addressing the central issues - coverage and levels of saving - than a voluntary system employing incentives. If there is any role for incentives it will have to be on top of a compulsory system.

Conclusions

Incentives are not an effective enough response to reverse the decline in pension provision in the UK. Greater incentives are likely to:

• Have less of an impact amongst those who are currently un-pensioned, such as the low-paid;
• Increase complexity and, by extension, costs;
• Be ineffective at overcoming behavioural obstacles to saving; and
• Fail to significantly affect the overall savings level, but rather alter its composition.
• Therefore a greater element of compulsion will be required to increase savings.

---

[1] More detailed analysis is available in Uncovered: workers without pensions, TUC, July 2002

[2] Living in Britain General Household Survey 2002, table 6.5

[3] Financial Services Consumer Panel Annual Survey 2000

[4] Simplicity Security and Choice: Informed choices for working and saving, DWP (2004)

[5] Tax Relief and Incentives for Pension Saving, A Report by the Pensions Policy Institute for Age Concern England , page 19, PPI, October 2004

[6] Calculation based on information in Inland Revenue table 7.10

[7] ' We estimate that around 38 cents of each dollar in superannuation contributions are offset, or in other words, 62 cents in each dollar are saved additionally.' The Impact of Superannuation on Household Saving , page 25, Reserve Bank of Australia, 2004

[8] Household saving in the UK, Banks, J. and Tanner, S., 1999 (as referenced in Sandler review)

[9] Mutual Funds and the Retirement Market, Fundamentals, Vol. 9, No. 2, Investment Company Institute, May 2000

[10] How Important are Private Pensions ? Alicia H. Munnell, Annika Sundén and Elizabeth Lidstone . Center for Retirement Research at Boston College.

[11] The Impact of Superannuation on Household Saving, page 14, Reserve Bank of Australia, 2004

[12] Sandler Review: Medium and Long-Term Retail Savings in the UK, page 208, HM Treasury, July 2002

[13] 2005 Defined Contribution and Additional Voluntary Contribution (DC & AVC) scheme survey, Hewitt Bacon & Woodrow

[14] Chris Cairns of Hewitt Bacon & Woodrow. http://www.management-issues.com/display_page.asp?section=research&id=2180

[15] Living in Britain General Household Survey 2002, table 6.2, page 74

[16] Which? Financial Advice Market Research report, May 2002

[17] Better Informed Consumers, FSA, April 2000

[18] For example B&CE Benefits Schemes saw a 466% increase in membership of one plan after introducing auto-enrolment, 'Auto-enrolment green light is right signal, says B&CE', page 10, Pensions World, August 2005

[19] The Paradox of Choice, page 28, Barry Schwartz

[20] Pensions: Choices and Challenges, page 209, Pensions Commission

[21] For example see Excessive volatility or uncertain real economy? From: Boom and Bust: the equity market crisis - Lessons for asset managers and their clients, European Asset Management Association

[22] Pensions: Choices and Challenges, pages 208-209, Pensions Commission

[23] Sandler Review: Medium and Long-Term Retail Savings in the UK, page 208, HM Treasury, July 2002

[24] Tax Relief and Incentives for Pension Saving, A Report by the Pensions Policy Institute for Age Concern England , page 9, PPI, October 2004

[25] Bridging the Savings Gap: An evaluation of voluntary and compulsory approaches to pension reform, paragraph 7, page 34, ABI, June 2005

[26] Bridging the Savings Gap: An evaluation of voluntary and compulsory approaches to pension reform, Table 6, page 23, ABI, June 2005

[27] A survey in 2004 by the Engineering Employers Federation found two-thirds of employers supported compulsion, primarily to create a 'level playing field' on pension costs. http://www.eef.org.uk/UK/mediacentre/mediareleases/uk/2004/newsrelease10122004.htm

[28] How government can get us saving again, page 3-4, Adam Smith Institute, Spring 2004