Con Keating, head of research at Brighton Rock, makes the case for the insured, unfunded DB scheme.

This week I presented a small part of my new paper on UK occupational pensions, 'Don't stop believing'. The public launch event, which is open to all, will take place at the Global Policy Institute in London on 28 September.

One of the less significant findings of the research reported in the paper is that the massive increases in DB contributions made over the last decade, the radical shift observed in fund asset allocation and the advent of 'sophisticated' management strategies, such as liability-driven investment, have all been ineffectual, though very costly. It is this aspect that has garnered widespread press attention even though the more important point is that this is a direct result of misguided regulation, which generates costs for schemes and their members, but no apparent increase in member security or benefits. The regulator-inspired culture of reckless 'prudence' on the part of trustees is no small part of this.

Perhaps it is not surprising that the fact that DB arrangements are massively more efficient than DC has not attracted attention - this finding is scarcely new. Nonetheless, the scale of the inefficiency due to the collective risk-pooling and risk sharing is worth remembering - a pension benefit under a DC arrangement will cost at least 50% more in contribution cost than that same benefit provided under DB. The move to DC and lower contributions is a clearly evident trend and can only mean these 'pensions' will be grossly inadequate.

Another finding from the report is that schemes have lost sight of the importance of investment income in both the short and long term. The yield of pension funds has now fallen below 2% a year. This decline is greater than can be explained by falling bond yields (of any type) or the dividend yields on stock markets.  It cannot even be explained by fund management fees.

But possible explanations are much less important than the possible consequences. In the short term, if the sale of investment assets is to be avoided, either this investment income must be increased or further special contributions will be required. In the long term, this investment income is the principal determinant of the total achieved portfolio return - recall that, of the 5.5% realised long-term return of UK equities, 87% arises from investment income; less than 4% is due to changes in the market valuation basis. It means that the expected returns from scheme assets are severely depressed - something that is not justified by long-term global investment prospects, even from today's troubled times.

The paper dispels many more misconceptions - some are listed below:

Pensions are unaffordable and unsustainable Employers should not be in the business of providing pensions The individual should provide for his own pension Increasing longevity has been the principal cause of rising pension costs Pensioners are now better protected by UK pension regulations The accounting standards are merely the messenger of the parlous state of DB pensions Asset and liability management techniques can resolve the problem

Not one of these ideas stands close scrutiny - the evidence is that these commonly held beliefs are simply wrong.

The conclusions of the paper are that schemes should be defined benefit arrangements, since this delivers member pensions most efficiently. Funding a scheme is a grossly inefficient device to protect scheme members against sponsor insolvency, the sole risk that they face in the UK. Insuring the scheme against sponsor insolvency is vastly more efficient - and can actually be rewarding. The experience in Sweden, where such pension indemnity assurance is available (and widely used) is just that - dividends have exceeded premiums for many years, including our troubled recent times. Once insured, funding is entirely redundant, not merely an incomplete and inefficient security device. This frees the contributions for investment (as book-entry) in the sponsor employer, which, among other things, should increase their productive capacity and competitiveness. It also aligns more closely the interests of employees and the sponsor employer and enhances industrial relations.

It is important to recognise institutions such as the Pension Protection Fund and the US Pension Benefit Guaranty Corporation are mutual compensation funds, not insurance companies. The difference is material, not just idle pedantry. Insurance is a provision against adverse developments in the insured's circumstances - it is an asset for them. Participation in a mutual compensation fund is a pure sunk cost - the levies are expenses arising from the failure of others. However, the optimal form of organisation of a pension indemnity assurer is as an industry-owned mutual - perhaps we should be better occupied trying to organise that rather than discussing the failings of management techniques drawn from the theory of financial speculation.

The model proposed is insured, unfunded DB. It can actually be offered under current UK legislation, though only at small scale.

The final question is one of cost. Obviously, this depends upon the profitability of the sponsor employer - the capital is invested there. The pension indemnity insurance is certainly not expensive - the premium rate in Sweden averages 0.3% a year of scheme liabilities. Even if we see continuing increases in lifespan, we should expect that, for most companies, a two-thirds final salary scheme should have a contribution cost of between 10% and 20% of wages.

The final, but perhaps most important, attraction of this form of organisation is that the benefits of an ageing population are also captured, not just the pension costs. To quote Jay Olshansky, a leading researcher on longevity: "Older people should not be thought of as a burden, although they are often seen that way. People who make it healthily out to old age are a huge economic and social resource. We must emphasise the extreme value of the older population, not just the cost."

Con Keating is head of research at Brighton Rock