When a State Appoints a Minister for Loneliness, and What the Costing Did to It

In January 2018 the British Prime Minister added loneliness to the responsibilities of a junior minister who already had sport and civil society. In February 2021 the Japanese Prime Minister instructed a minister who already had the falling birth rate to take on loneliness and isolation as well. Neither country created a department. Neither created a budget line first. Both did something that is easy to file as a curiosity, was filed that way by a great deal of the coverage at the time, and is not one.

What a state does when it names loneliness as a policy object is to assert three things at once. That the feeling is not a character defect. That its causes lie at least partly outside the person having it. And that public machinery therefore has standing to act on them. Each of those is contestable and all three together amount to a serious philosophical commitment made, as such commitments usually are, in a press notice rather than a treatise. This chamber thinks the move deserves to be argued with rather than smiled at.

The argument this essay makes is not that the move was wrong. It is about what the move cost to make. Loneliness became legible to government partly because it was costed — translated into health expenditure, lost working days, staff turnover and foregone productivity. That translation is what got it funded, and it is also what deformed it, because the version of loneliness that appears in a cost model is the version that generates spending, not the version that hurts most.

There is a second problem underneath, and the two are related. A state can fund contact. It cannot supply intimacy. Contact is countable: attendances, referrals, sessions, cafés opened, volunteers trained. Intimacy is not, and what cannot be counted does not survive contact with a spending review. So the measurement drifts toward the thing the money can produce, and the gap between the two quietly becomes the gap between what was funded and what was wanted.

The precise arrangement matters, because the popular version of it is wrong in a way that changes the argument. Britain did not create a Ministry of Loneliness. On 17 January 2018 the Prime Minister accepted a set of recommendations from the cross-party commission set up in the name of the murdered MP Jo Cox, and one of them was to appoint a ministerial lead. The lead was a serving Parliamentary Under Secretary of State at the Department for Digital, Culture, Media and Sport, and loneliness was added to a portfolio that already contained sport and civil society. The strategy published nine months later carries her name on its ministerial foreword under exactly that double title.

That strategy is the document the rest of the world quotes without reading. Published in October 2018, it is the first national government strategy of its kind, and it is more careful than its reputation. It adopts a definition of loneliness borrowed from the academic literature — a subjective and unwelcome feeling of lack or loss of companionship, arising from a mismatch between the relationships a person has and the relationships they want. It insists that loneliness and isolation are different things, on the ground that isolation is observable and loneliness is not. It extends to England only. And it states plainly that it sets no quantitative target for change, because there is less certainty about what drives loneliness and what works against it than about the fact that it matters.

What followed was a national measurement standard rather than a national programme. The statistical office recommended a package of four questions — one direct, three adapted from an American research scale — and government committed to it as its standard way of measuring loneliness. Four annual progress reports were published between 2020 and 2023. The responsibility itself has travelled. It has passed between junior ministers across two governments, and by the middle of this decade it was being carried inside a single junior portfolio alongside youth, civil society, digital inclusion and gambling. It has never once been a department, and it has never once had a target.

Japan moved three years later and in the opposite order. On 12 February 2021 the Prime Minister instructed the minister responsible for the declining birth rate to take on loneliness and isolation as an additional brief, and within the week a dedicated office was stood up inside the Cabinet Secretariat. The circumstances were not abstract. Suicides in Japan had risen in 2020 for the first time in eleven years, to 21,081, and the rise was concentrated among women, who accounted for 7,026 of the total, an increase of 935 on the previous year, while the figure for men fell. The number for schoolchildren was the highest in a series running back to 1980. A government that had spent two decades building suicide-prevention machinery had a category of harm arriving through a door that machinery did not cover.

Then Japan did the thing Britain never did. The Act on Promotion of Policy for Loneliness and Isolation was enacted on 31 May 2023, promulgated on 7 June 2023 as Act No. 45 of that year, and came into force on 1 April 2024. Its first article defines the state of loneliness and isolation as a condition in which a person suffers harm to mind or body either by feeling alone in daily or social life, or by being isolated from society — a definition that deliberately holds a subjective experience and an objective circumstance inside one legal phrase. It establishes a promotion headquarters chaired by the Prime Minister and composed of the whole cabinet. It requires a priority plan, first adopted on 11 June 2024 and revised since. And it asks local authorities to endeavour to set up local councils that bring support organisations together.

Two details in that statute are worth holding. The first is that the local council provision is an endeavour duty, not an obligation, which is the standard Japanese legislative move when the centre wants a structure it cannot compel. The second is that the only criminal penalty anywhere in a law about loneliness attaches to a member of one of those councils who leaks what they learned in its work: up to a year of confinement or a fine of five hundred thousand yen. A statute about the pain of having nobody ends by protecting, on pain of imprisonment, the confidences of people who were finally told something.

And then, in October 2024, loneliness and isolation disappeared from the formal title of the minister who held it. The portfolio was reorganised, and the words in the minister’s designation became co-existence and mutual aid. The statutory machinery did not go anywhere: the headquarters, the plan, the survey and the office all remain, and the duties now sit with the Cabinet Office because the Act put them there. Whether the disappearance of the word from a job title is institutional maturity or a quiet demotion is a question the public record does not settle, and this essay does not pretend to settle it either.

Strip away the machinery and the move underneath is genuinely large. For most of the modern period loneliness was understood, when it was understood at all, as a private condition with a private cause: shyness, bad luck, a difficult personality, a life not managed well. To make it a policy object is to say that some meaningful share of it is produced by arrangements — by how housing is built, how work is organised, how care is distributed, how old people are moved, how young people are sorted. That is a claim about causation, and it shifts the location of the blame.

The World Health Organization eventually made a version of the same move, which is worth noting because it removes the suspicion that two national governments had simply had a strange idea. A World Health Assembly resolution on social connection was adopted in May 2025, and the commission the organisation convened published its report at the end of June that year, estimating that loneliness was associated with roughly 871,000 deaths a year across the period it studied and that around one in six people worldwide experience it. Whatever one makes of those estimates, the institutional point stands: a feeling has been admitted to the register of things that public health bodies count.

This chamber has a standing reason to be interested in the causal claim, and also a standing reason to distrust it in one direction. Its essay on Denmark found that loneliness among the young survived a settlement which had solved almost every structural problem a country can solve — childcare as a utility, leave reserved in a father’s own name, accurate sex education for half a century, no meaningful legal obstacle to any relationship a young woman might want. If loneliness were simply downstream of arrangements, the Danish figure should have collapsed, and it did not. So the appointment’s underlying claim is right in one direction and incomplete in the other: arrangements produce a share of it, and removing the arrangements does not remove the feeling.

That is not a reason to dismiss the policy. It is a reason to be precise about what the policy is for. A state that removes an obstacle has done something real even if the person on the other side of the obstacle is still lonely. The trouble starts when the state, needing to show that the money worked, begins to count the obstacle removed as though it were the feeling removed.

Here is the essay’s actual argument, and it runs against the admiring reading of the whole episode. Loneliness did not enter British government because a commission proved it was painful. Pain had been available as evidence for centuries. It entered because, in February 2017, a retailer and a think tank published a figure: loneliness was costing employers in the United Kingdom about £2.5 billion a year. That number was issued in conjunction with the Jo Cox commission, it was reproduced inside the 2018 strategy twice, and it is the sentence that made the subject sayable in a room where budgets are set.

The strategy’s own framing follows the logic of the number rather than the logic of the definition it had just adopted. Having said that loneliness is a subjective and unwelcome feeling, the document then argues for action on the grounds that lonely people are more likely to be readmitted to hospital, to have longer stays, to visit a general practitioner or an emergency department, and to enter local-authority-funded residential care; that loneliness is associated with poorer performance at work; and that social interaction has been linked to increased productivity. It says, in as many words, that it is in all our interests to reduce loneliness. The appeal is to the public purse and to the employer’s balance sheet, and it is made because that is the appeal that works.

This is not a suspicion; it is a finding in the academic literature on the policy itself. A 2024 analysis in Critical Policy Studies by Jentoft, Sandset and Haldar read the British loneliness documents using Bacchi’s method of asking what a policy represents the problem to be, and found that loneliness is constructed there as a public health threat which both arises from and causes ill health, and as a strain on health and social care systems and on the economy through reduced productivity. Their further finding is the one that matters for this essay: the social determinants most closely tied to loneliness go largely unaddressed, in favour of solutions that route the lonely person toward the voluntary sector and hand them responsibility for their own reconnection.

So the counter-thesis is not that the costing was dishonest. The people who produced it were trying to get something funded that badly needed funding, and they succeeded. The counter-thesis is that a translation was performed, that the translation is what unlocked the money, and that nobody translated back.

The £2.5 billion figure is a serious piece of work and reading it closely is the fastest way to see the problem. The model is built from four pathways, all of them employer-borne. It counts staff turnover at about £1.62 billion, reduced productivity at about £665 million, working days lost by employees caring for someone with a condition associated with loneliness at about £220 million, and working days lost by lonely employees themselves at about £20 million. Roughly nine tenths of the total comes from wellbeing pathways and one tenth from health, and the authors explain why: many of the health costs are borne by the state rather than by employers, and so fall outside the frame. They also say, without evasion, that few of the underlying studies establish causality and that their findings otherwise reflect non-causal associations.

Now set that model beside what the national measurement actually finds. The Community Life Survey for 2024 to 2025 reports that about 7 per cent of adults in England felt lonely often or always. Disabled adults reported it at 14 per cent against 3 per cent for non-disabled adults. Adults in the never-worked and long-term unemployed classification reported it at 13 per cent, the highest of any occupational group, against 5 per cent among higher managerial and professional workers. Adults in the most deprived decile of areas reported it at 11 per cent against 4 per cent in the least deprived. Adults aged 16 to 24 reported it at 9 per cent, and adults aged 65 to 74 at 4 per cent.

Read those two paragraphs together and the deformation is visible without any rhetoric. An employer-cost model counts people in work. Its own summary estimates just over a million lonely workers. The measurement finds that the group with the highest reported loneliness in England is the group that has never worked or has been out of work a long time — people the model cannot price, because they generate no absence, no turnover and no productivity loss. Disabled adults, who report it at more than four times the non-disabled figure, enter the model chiefly as the people that employed carers take days off for. Deprivation, which shows a clean gradient across ten deciles, enters not at all.

This is what the essay means by saying the costing distorts. Not that the number is fabricated, but that a cost model is an instrument with a field of view, and the field of view is spending. Loneliness among the employed is a cost. Loneliness among the never employed is not a cost, it is a life, and a policy that arrived through the door marked expenditure will keep finding it easier to fund the first. The version of the problem that got it funded is systematically not the version that concentrates the harm.

Both countries measure loneliness the same way, and understanding the instrument is a precondition for reading either. The British standard, recommended by the statistical office and adopted across government surveys, is four questions: one direct — how often do you feel lonely — and three indirect, adapted from a research scale, asking how often a person feels they lack companionship, feels left out, and feels isolated from others. The guidance is explicit about why both kinds are needed: people vary in how they understand the word lonely, and some are reluctant to admit to it. The guidance names older men as a group that under-reports on the direct question, and still concludes that the direct question is the best available measure of self-perceived loneliness, which is an unusually honest thing for a methods note to say.

The Japanese survey uses the identical architecture, which makes the comparison possible and makes one particular result legible. It is a general statistical survey under the Statistics Act, drawing a random sample of twenty thousand people aged sixteen and over, delivered by post with the option to answer online. The round conducted on 1 December 2025 and published on 14 April 2026 obtained 11,873 valid responses, a response rate of 59.4 per cent. It asks the direct question and the three-item Japanese short form of the same American scale.

The two instruments then disagree inside a single questionnaire, and the disagreement is the most instructive number in this essay. On the direct question, 4.5 per cent of respondents said they feel lonely often or always. On the indirect scale, 6.5 per cent scored in the top band and a further 38.2 per cent scored in the band below it. Adding the softer direct categories — sometimes and occasionally — brings the direct total to roughly four in ten. So depending on which question you print, Japan is a country where one person in twenty-two is badly lonely, or a country where two in five carry something that a validated scale registers as loneliness. Both figures are from the same people on the same day.

What follows about what a self-report can establish is narrow and worth stating flatly. Within one instrument held constant, it can establish distribution across groups, and direction over time. It cannot establish incidence separately from willingness to disclose; a rise may be a rise in loneliness or a rise in the permission to name it, and after eight years of a government telling people that loneliness is nothing to be ashamed of, the permission has certainly moved. It cannot be compared across countries, because wording, translation and the cultural weight of the word do not survive the crossing — which is why this essay puts no British and Japanese figure in the same sentence as a ranking. And it cannot establish cause. The Japanese survey finds that people who almost never eat with anyone report the highest loneliness and those who eat with someone nearly every day report the lowest. It does not, and cannot, tell you which way that runs.

The instruments are the second problem; the first is what the money can buy. Britain’s strategy funds social prescribing, community spaces, transport, digital inclusion and an employer pledge. Japan’s priority plan funds local public-private platforms, local councils, consultation lines and the training of ordinary residents as connection supporters who notice a struggling neighbour and route them to a service. Every item on both lists is a form of contact. Not one of them is a form of intimacy, and nobody involved thinks otherwise; the documents are more honest about this than their critics allow. The Japanese plan explicitly notes that action on the subjective side touches the interior of a person and must be approached with that in mind.

The difficulty is that the countable item becomes the reported item, and then the achieved item. The Japanese survey contains a finding that ought to be read carefully by anyone who believes provision is the answer. Among respondents who said they have worries or anxieties in daily life, 87.0 per cent said they had received no support from government bodies or non-profit organisations. And the most common reason given by those people, at 63.1 per cent, was not that support was unavailable, or hard to reach, or shameful to accept. It was that they did not need it.

That single pair of figures does more work than any argument. It is not evidence that the services are bad. It is evidence that the thing the survey is measuring and the thing the services are offering are not the same object for most of the people who have it. A person who scores in the loneliest band of a validated scale and also declines support because they do not need support is not being inconsistent. They are telling you that what they lack is a particular person, and that a service is not a person.

This is the point at which this chamber’s earlier finding becomes load-bearing. The Library has argued that loneliness is a discrepancy rather than a headcount — the distance between the relationships a person has and the ones they want — which is, as it happens, precisely the definition the British strategy adopted in 2018. A discrepancy can be closed from either end, and public provision can only ever work on one of them. It can raise the supply of contact. It cannot lower the specificity of the want, and it should not try; a policy that succeeded in making people want less would not be a loneliness policy. The Library has also argued that being good at solitude is not the same as needing no one, which is the other half of the same caution: the person who looks resourced may be the one the survey cannot reach.

Both countries know the number has not moved. Britain’s strategy declined a quantitative target at the outset and the English series now sits above its own mid-2010s baseline. Japan has run five rounds since December 2021 and the most recent round reports no large difference from the one before it. Eight years of ministerial attention in one country and five in the other have produced institutions, a measurement standard, a statute and a great deal of funded contact, and have not produced a fall. The most rigorous meta-analytic work on loneliness interventions is consistent with that: pooled effects across many trials are small to moderate, and the confidence the reviewers place in their own estimates is graded low or very low.

The comparison is not a league table and the chamber refuses to draw one. Britain wrote the strategy and never wrote the law. Japan appointed the minister, wrote the law, built a headquarters chaired by the Prime Minister, and then took the word out of the minister’s title. Each country did the half the other did not, and a reader here who has been told that Japan is behind on this should notice that on the specific question of legal instruments, it is not. Japan has a statute; the United Kingdom has a strategy that extends to England only.

There is also a real difference in framing, and it runs the opposite way to the stereotype. The British case was opened with a productivity figure. The Japanese priority plan opens its account of causes with the transformation of the employment system — the erosion of lifetime employment, seniority wages and simultaneous graduate recruitment, the growth of part-time, fixed-term and dispatched work, widening income disparity — and describes the resulting condition as a poverty of relationship, and a matter of life. It reaches for a vocabulary of dignity where the British document reached for a vocabulary of cost. That is not because one bureaucracy is more humane than another. It is because Japan’s loneliness policy was born out of a suicide statistic and Britain’s was born out of an employers’ report, and each carried its birth certificate into its prose.

What the Japanese record then shows is that the vocabulary does not settle the outcome. Two years after the Act came into force, the local councils it authorises existed in seventy-two local bodies — three prefectures and sixty-nine municipalities — and the public-private platforms in a hundred and seventy-four. The commonest reason local authorities gave for not having established one was not cost and not disagreement. It was that they did not know how to set one up or how to go about the work. A statute can create a structure and cannot create the competence to run it, which is the administrative version of the essay’s whole argument.

The mirror for a reader here, then, is not that Japan should copy Britain or that Britain should copy Japan. It is narrower and more useful. If a country with high social trust and a functioning welfare state still cannot move the figure, and a country that legislated still cannot move it, then the residue is not a national defect. This chamber reached the same conclusion from Denmark by a different route, and from the United States by a third: a society can be free, or equal, or legally equipped, and still contain people who go home and find that the person they wanted to tell about their day does not exist. You are not lonely because of where you were born. That explanation flatters the diagnosis and is unkind to the person.

This house sells paid companionship, which makes it the most interested party imaginable in an essay about loneliness policy, and the concession has to come before anything else. Everything above — that a state can fund contact and not intimacy, that provision does not reach the specific want, that the number has not moved after eight years of public effort — is, read commercially, an argument that the market should be left to do it. The house has an obvious motive to write exactly that essay, and the reader should weigh the argument in that knowledge rather than in spite of it.

So the house states its own position against itself. What it sells is contact too. Better-designed contact, chosen rather than prescribed, unhurried, with attention as the substance rather than as a side effect of some other service — but contact, purchased, for a defined evening, from someone whose presence is arranged. It is not exempt from its own critique. If a public café and a public referral cannot supply the particular person somebody is missing, a paid evening cannot supply that person either, and any suggestion that it can would be the same conflation dressed better.

What the house cannot settle is the question that matters most, and it will not pretend otherwise. There is no evidence available to it that anything it sells moves the thing the two national surveys are trying to count. It has no trial, no follow-up, no measure, and it would be improper to construct one from customer sentiment, since satisfaction with an evening and the discrepancy at the centre of a person’s life are different objects. Nothing here is a health intervention, nothing here treats loneliness, and nothing here is offered as a substitute for a relationship, a friendship, a clinician, or a public service.

What it can honestly claim is smaller and is worth stating precisely because it is smaller. An evening in which someone is attentive, unhurried and not owed anything is a real experience of being attended to, and a person who has not had one recently may find that having one tells them something about what they have been missing. Whether that knowledge is useful, and whether it survives the walk home, is not in the house’s gift.

This essay does not claim that appointing a minister for loneliness was a mistake. Its argument is the reverse: the move rests on a defensible and serious claim about causes, and the chamber would rather argue with it than condescend to it. What the essay claims is that the costing which made the move possible also narrowed it, and that the narrowing is visible in the documents themselves rather than inferred from suspicion.

It does not claim that either country has failed. Eight years and five years are short in the life of a social policy; the absence of movement in a self-reported series is not proof that nothing worked, since an unmeasured counterfactual may have been worse, and the essay has no way to rule that in or out. It states the flatness and declines to convert it into a verdict.

It does not compare British and Japanese levels of loneliness, and the refusal is technical rather than diplomatic. The measures are not safely comparable across languages and survey traditions, the word itself does not carry the same weight in both, and any ranking built from these two series would be an artefact of translation. Where figures appear, they describe one country against itself.

It does not claim that the cost estimate is dishonest or that the people who built it acted in bad faith. They stated their own limits, including that few of the underlying studies establish causality, and this essay has used those stated limits rather than alleging hidden ones. Nor does it claim that the academic critique of the policy framing is settled; it is one reading, published and peer reviewed, and it is treated here as an argument rather than a finding.

It does not name the current holder of the British brief, because ministerial responsibilities are reassigned frequently and a name would date the piece without strengthening it. It does not characterise the internal reasoning behind the Japanese portfolio change of October 2024, because the public record shows the change of title and does not show the reason.

It makes no diagnosis of any reader, offers no generalisation about the character of either nation, and endorses no organisation, campaign or named individual mentioned in it. And it does not claim that this house has an answer to the problem it has spent nine sections describing. It has a product, it has stated what the product is, and it has stated what the product cannot reach.