Part of the reason for this state of affairs is the effect of increased life expectancy. In order to maintain life expectancy at retirement more or less constant, it is likely that retirement age will have to increase by over 1.3 years per decade for males and almost 1.2 years for females. The need for this is not well appreciated.

However, there have been many other contributory factors, including tax changes, the increasing burden of pension legislation, accounting standards, poor investment returns during the financial crisis, lower expectations of future returns, low interest rates and increasingly risk-averse employers.

There is no point in looking backwards. The next big project is auto-enrolment and we need to make sure it succeeds. It is not perfect, since the required contribution level from employers is very low at 3%, and a total contribution of 8% of earnings is only about half what it should be to build up a reasonable pension. Moreover, employees can opt out. Even though they will be opted back in from time to time, there will still be significant numbers who remain opted out.

There is a chance, however, that auto-enrolment of most employees, and inertia, will result in many more people contributing to a pension plan than in the past.

For most, auto-enrolment means DC, with the individual carrying most of the risk. Attention should be paid in particular to three main issues – the need for low charges; a default investment option that performs strongly; and structures to minimise interest rate risk at the time of conversion to pension. For there to be an adequate focus on members, trustees should have a fiduciary responsibility for members’ interests.

A full solution also needs to focus on the interests of employers, as there will be thousands of small to medium-sized employers who are being dragged unwillingly into making pension provision for their staff. Auto-enrolment will inevitably have a cost for them, but the key issue is whether it can be implemented painlessly, ensuring that the complex requirements of the legislation are met and that contributions can start flowing with minimum hassle.

Some employers will be keen to provide employees with investment choice, but the reality is that very few will want to exercise choice, so the most important issue is whether the default fund offers good returns without too much volatility. There also needs to be a well-organised strategy for effecting the transition from a suitable return-seeking portfolio during the main part of a member’s career in order to minimise annuitisation risk as retirement age approaches.

The UK has one of the most favourable demographic profiles in Europe. Strong inwards migration is alleviating the effect of the ageing population, and our social security benefits are lower than they are in most European countries.

However, non-means-tested benefits are low and for auto-enrolment to work well – without reducing entitlement to means-tested benefits for lower-paid workers – the basic pension needs to be set at a significantly higher level than envisaged.

With a higher basic pension and market competition forcing auto-enrolment providers to offer pensions with lower charges, there is a good chance of creating a new environment for pensions that will begin to turn things around.

Ultimately, we need stronger incentives

for much higher minimum contributions, lower, more competitive charges, more intelligent default fund options and stronger protection around funds as retirement age approaches, so that members will no longer be so exposed to market and interest rate risk when they exchange their retirement pot for

a pension. a

Chris Daykin is a trustee director of NOW: Pensions. He was government actuary from 1989 to 2007