Nudging people into saving for retirement is one of the policy success stories of the 21st century: 88 per cent of eligible workers are now regularly putting at least some money aside in a pension scheme, according to government figures.

It isn’t enough, of course. The minimum 8 per cent contribution from employer and employee together won’t deliver a prosperous old age. But it has been a better-than-expected start to the auto-enrolment project, which some sceptics warned would end in revolt or a mis-selling disaster.

Pension saving is now normal and mainstream.

There is, however, one glaring exception: the self-employed — the 4.4 million people, one-eighth of the UK workforce, who don’t have an employer to chip in parallel contributions and therefore tend to be much more sceptical about the value of pension saving.

Only 18 per cent of self-employed people are actively saving in a pension scheme of any sort, though three-quarters of them say they would like to. Half of them are headed for poverty in old age by one measure, twice the proportion of employees.

For several years now policymakers and pensions bosses have wondered whether the “nudge economics” that successfully helped push the employed into saving could be used in the same way to usher the self-employed over the line.

Employees are automatically put into pension schemes unless they actively opt out. Doing precisely nothing means they are enrolled and have deductions made from their pay packets. Inertia works its magic. Few people want to make active decisions on something as complicated as pensions but most people do realise they need to save.

So could something similar work for the self-employed? There have been doubts. After all, people who work for themselves tend to be more independent-minded than us wage slaves. Would they be prepared to be opted in, or would they regard it as unwelcome interference.

New research for the pensions platform Nest and the Department for Work & Pensions this week knocked back that concern at least. The 1,500 self-employed respondents gave an overwhelmingly warm reception to the idea, especially if it were designed in the right way. Many hadn’t realised that they would still get the benefit of tax relief even if there is no employer contribution to sweeten the pill further.

The hurdles to success are much greater though than for the employed. Categorising the self-employed in a single bucket is dangerous. They are far from homogeneous. At the top end of the income scale are skilled IT specialists, consultants and celebrities. At the bottom end is an army of gig economy workers — beauticians, joiners, buskers and delivery riders — sometimes juggling precarious lives.

One particular problem is the need to adjust for the sheer lumpiness and unpredictability of income. It is much harder for the self-employed to commit to a regular monthly pension contribution because they can’t count on a smooth workload or indeed on clients paying them on time.

Another stumbling block is the absence of an obvious touchpoint organisation to automatically enrol the self-employed person into a pension scheme. In the employed world, responsible employers may moan, but they rather like their paternalistic role in doing right by their staff.

A third obvious difficulty is the lifetime Isa — an existing tax-favoured savings vehicle aimed at both would-be home buyers but also the self-employed below the age of 40. For some younger self-employed on basic-rate tax, the Lisa is a more generous option than saving in a conventional pension scheme.

These problems can be solved, though they do add to the complexity of any new scheme. Policymakers are looking at adding an accessible rainy-day feeder pot, which once filled would then overflow into an inaccessible pension pot. Lots of flexibility in pausing contributions would also be needed.

Banks are obvious potential providers as almost every self-employed person already has an account out of which contributions could be taken. Lloyds Banking Group played a big role in the latest research and, with its Scottish Widows investment offshoot and existing roster of self-employed customers, would be an obvious candidate. So would NatWest, which is already pushing into employee auto-enrolment with its acquisition of the Cushon platform. Providers of accounting and payments software to small firms could also take on the role. What the latest research established is that the job need not be done by a public body such as HM Revenue & Customs.

Auto-saving for the self-employed as a concept has passed its first test. The government is keen. The pensions minister Torsten Bell was at the research launch this week and gave it a warm welcome. Nest is now pushing ahead with plans for a live trial and looking for a partner organisation to test the idea on live guinea pigs.

We are still, however, at least five years away from a working system complete with necessary legislation, most pension folk reckon. A decision from the Financial Conduct Authority on reforming the advice/guidance boundary, due this month, may help a little bit, but won’t go nearly far enough.

The biggest threat, perhaps, is that it will be very hard to design a system that contains literally no cumbersome and offputting customer consent requirement. Yet even the tiniest obstacle in the customer journey — like a tickbox or financial health warning — could prevent that journey being completed.

The difference in take-up rates between a service requiring customers to opt in to receive it, and one requiring them to opt out to avoid receiving it, is gigantic. Companies are already seeing that in vast differences in participation rates for employee savings schemes, depending on whether they are designed and presented as opt-in or opt-out.

Partly it’s natural human inertia and prevarication. Partly it is wanting the reassurance of doing what everyone else is doing. Good regulation, trustworthy platforms and prolonged cross-party support (as with auto-enrolment) will of course be needed to make this work. So too will be a grasp of psychology.