Retirement gets planned like a balance sheet. Contribution rates, drawdown, the state pension age, whether the mortgage clears before the salary stops. Most people approaching the end of a working life have some version of that arithmetic, and very few have the other document: the one that says what a Wednesday in March is for.

The financial side has decades of careful research behind it and an industry built on top of it. The other side, the picture of the days, tends to be treated as something that will resolve itself once the diary empties.

We are writers, not clinicians. What follows is a reading of the research, not medical or psychological advice.

What interests us is not the familiar advice to find a purpose. It is a narrower question the evidence can actually speak to: does the scale of the thing you intend to do matter, or does something else account for whether people follow through?

What retirement planning research usually measures

There is a well-developed literature on what its authors call retirement goal clarity. Goal Clarity and Financial Planning Activities as Determinants of Retirement Savings Contributions, by Robert Stawski, Douglas Hershey and Joy Jacobs-Lawson in The International Journal of Aging and Human Development in 2007, is one of the more cited examples, and it supports a chain that runs from clarity of goals, to planning activity, to how much people actually put aside.

Worth reading closely, though, is what the clarity is clarity about. In this line of work the goals are largely financial, and the outcome being predicted is savings contributions. It is a study of whether knowing what you want makes you fund it, not a study of whether the thing you funded turns out to be worth having.

That distinction gets lost when the finding is repeated. A worker who can state a target retirement income precisely, and has no idea what they will do at ten o’clock on a Tuesday, scores well on goal clarity as the field defines it.

What the purpose research supports, and what it does not

The evidence that a sense of purpose tracks with living longer is real and it is more than one dataset. Patrick Hill and Nicholas Turiano’s analysis of the Midlife in the United States sample, published in Psychological Science in 2014, found that people who reported greater purpose in life were less likely to die over the following 14 years, controlling for other measures of wellbeing. The detail most worth keeping is in their abstract: the benefit did not appear to depend on the participants’ age, on how long they survived during follow-up, or on whether they had already retired.

Andrew Steptoe, Angus Deaton and Arthur Stone reported something in the same direction in their 2015 Lancet review, using the English Longitudinal Study of Ageing. Among people in the lowest quartile of what they call eudemonic wellbeing, the sense that life has meaning, 29.3 per cent died over an average follow-up of about eight and a half years, against 9.3 per cent in the highest quartile.

That gap looks enormous, and the raw figures are not an effect size. People in poor health report less purpose, and people in poor health die sooner, which means the arrow can run in either direction. The authors adjust for baseline physical and mental health and report the association surviving that adjustment, but adjustment is not the same as settling the question.

There is a second limitation that gets skipped more often than the first. These studies measured purpose in people who already had it. None of them tested what happens when someone sets out to construct one. The finding is that purposeful people lived longer in these cohorts, not that deciding on a purpose at 64 buys anyone time.

The cognition argument is quoted as settled and is not

The other study that turns up in almost every article about retiring without a plan is Susann Rohwedder and Robert Willis’s Mental Retirement, published in the Journal of Economic Perspectives in 2010. Using comparable surveys from the United States, England and eleven European countries, and treating variation in national pension, tax and disability rules as a source of exogenous variation in who has stopped working, they reported an effect of early retirement on the cognitive ability of people in their early sixties that was, in their words, both quantitatively important and causal. The measure was immediate and delayed word recall.

It is a serious paper and it has been argued about ever since. Norma Coe and Gema Zamarro, working in 2011 with the same cognitive measure and the same instruments, found the effect no longer held once further controls including country dummies were added. Paul Bingley and Alessandro Martinello argued in the European Economic Review in 2013 that pension eligibility ages are not valid instruments unless schooling is controlled for, and that this accounts for a large part of the measured effect. Different specifications move the size and the significance considerably.

In our reading, this literature is unresolved, and quoting Mental Retirement without the fifteen years of dispute attached to it overstates what anyone knows.

Where the evidence is strongest is also where it is least dramatic

The most robust finding relevant to any of this has nothing to do with retirement at all.

Peter Gollwitzer and Paschal Sheeran’s 2006 meta-analysis, published in Advances in Experimental Social Psychology, pooled 94 independent tests of what they call implementation intentions: plans in the form “if situation Y occurs, then I will do Z”. In the authors’ own summary of that work, forming an if-then plan produced a medium-to-large effect on goal attainment across those studies, d = .65, with d = .61 for the specific problem of failing to get started and d = .77 for being derailed part way through.

The mechanism they describe is unglamorous. Naming a specific cue makes that cue easier to notice, and pairing it with a specific response removes the need to decide anything in the moment.

Their own worked example is the useful part. A plan whose then-part says “exercise more” and whose if-part says “tomorrow” does close to nothing, because the person still has to work out what to do and when. “If it is 5pm on Monday, then I will jog home from work” is the version that shifts behaviour. The two plans have identical ambition and completely different specificity, and the specificity is what carries the effect.

None of this was tested on retirees. The studies in that meta-analysis were mostly ordinary health and task behaviours over short horizons, and a fortnight of vitamin-taking is not a decade of unstructured mornings. Anyone applying it to retirement is extrapolating, and should say so.

What the material actually supports

Put the three literatures beside each other and the practical implication is smaller than the topic usually invites, and firmer.

Purpose is associated with better outcomes, in several large cohorts, with the direction of causation open. Whether stopping work erodes cognition is contested and should be described that way. And the one thing that has been tested repeatedly, under controlled conditions, is that intentions with a time, a place and a named action attached get carried out more often than intentions without them.

Notice what is absent from that list. Nothing in the evidence rewards the scale of the plan. The size of the ambition is not the variable that has been shown to matter, which is a reasonable argument for lowering the bar rather than raising it. A Tuesday morning at a specific place doing a specific thing is a fully-formed plan by the standard that has actually been measured. Writing a book is not, until it has a chair, an hour and a day of the week attached to it.

What we find useful about this is that it removes the excuse. The version of retirement planning most people avoid is the one that demands a grand answer to what the rest of a life is for. That question is hard, and it is also not the one the evidence asks.

The evidence asks a much duller question, which is what you are doing on Thursday, and where.