At some point in the construction of the “loneliness epidemic”, the lonely person disappeared. They were present in the opening paragraphs—the vivid description of the individual experience, the quote from someone who had not spoken to another person in three days, the photograph of an older woman looking out a window. They established the moral weight.
After that, they were replaced by a figure who was more useful to the argument being made: the economically significant unit of social deficit, whose isolation was costing the healthcare system, the labour market, and the GDP a measurable and substantial amount.
The lonely person becomes the socially isolated individual. The socially isolated individual becomes a population at risk. The population at risk becomes a burden on the health system. The burden on the health system becomes a cost to the economy. The cost to the economy becomes a productivity problem. The productivity problem requires a solution. The solution requires funding. The funding requires evidence of economic return. The evidence of economic return is calculated by multiplying the number of socially isolated individuals by the per-unit cost of their isolation to the healthcare system and the labour market. The calculation produces a large number. The large number justifies the funding.
The lonely person is in there somewhere. They are the unit of calculation. They are not the subject of the argument. The subject of the argument is the economy.
The productivity framing is not neutral. It carries an implicit claim about why loneliness matters—which is not that the lonely person is experiencing something painful and human that deserves attention because they are a person, but that the lonely person is generating costs that are falling on a system that would prefer not to bear them. The problem with loneliness, in this framing, is not that it hurts. The problem is that it is expensive. The solution worth funding is not the one that addresses the hurt but the one that reduces the cost.
This produces a specific kind of absurdity that the report language does not acknowledge. The isolated person who is not generating costs—who manages their own health, who remains in employment or has retired without claiming significant healthcare resources, who is lonely in the felt sense without being economically costly in the measurable sense—does not appear in the calculation. Their loneliness does not matter to the productivity argument because it does not impact the productivity figure. The loneliness that matters is the loneliness that costs, and the costs that matter are the ones that fall on the institutions whose interests are represented in the calculation.
The isolated person who is costing the healthcare system and reducing workplace productivity has committed, in the implicit logic of the productivity framing, an economic transgression. They have failed to maintain the social connections that would have kept them out of the healthcare system and in the workforce. Their isolation is creating a productivity issue. The framing positions them as a problem to be solved rather than a person to be encountered, which is a different orientation from the one the befriending service’s brochure describes, and which produces different consequences for how the response is designed and evaluated.
The epidemiological claim deserves more scrutiny than it typically receives. The loneliness epidemic is presented as a contemporary phenomenon—something that is worse now than it has ever been, a product of specific modern conditions that have disrupted social bonds that previously held. The evidence for this claim is more complicated than the epidemic framing suggests.
The measurement of loneliness across time is genuinely difficult. The instruments used to measure it now did not exist in previous generations. The UCLA Loneliness Scale, the De Jong Gierveld scale, and the other validated instruments that contemporary researchers use to produce the numbers the epidemic framing relies on were developed in the late twentieth century. They cannot be applied retrospectively to populations who lived before their development. What can be compared across time are proxy measures. Rates of social participation, household composition, marriage rates, reported friendship numbers—which are related to loneliness but not equivalent to it.
The proxy measures do show changes. People live alone more than they did in 1950. Household sizes have decreased. Rates of religious participation have declined. Club and civic organisation membership has reduced. These are real changes in the structure of social life. Whether they have produced more loneliness in the felt sense—the subjective internal experience of an unwanted gap between available and desired connection—is a question that the proxy measures cannot directly answer, because the proxy measures are measuring structure rather than experience.
The possibility that previous generations experienced the same felt absence of genuine connection, without a vocabulary or a measurement instrument or an institutional framework to name it, is not considered by the epidemic framing, which assumes that the absence of a named and measured epidemic in a previous period reflects the absence of the condition rather than the absence of the measurement. Chekhov’s characters were lonely in 1890. They did not have a loneliness scale score. The absence of the score did not mean they were not lonely. It meant nobody was measuring them.
The evidence that the offered solutions are working is thinner than the reports presenting them suggest. The systematic reviews of loneliness interventions—the meta-analyses that pool the results of multiple studies to produce an overall estimate of intervention effectiveness—consistently find effects that are small, heterogeneous, and often not maintained at follow-up. The interventions that produce the largest effects are typically the most intensive and the least scalable—the programmes that provide sustained, individualised support over extended periods, that bear almost no resemblance to the group activity referrals and social prescriptions that the commissioning frameworks fund at scale.
The interventions funded at scale produce effects that are, in aggregate, modest. The loneliness scale score at twelve-week follow-up is somewhat lower than at baseline for some participants in programmes. Whether the reduction in the score represents a genuine change in the felt experience of isolation, or an effect of the research context on the participant’s self-report, or a regression to the mean, or a transient improvement that does not persist beyond the measurement period, is a question the twelve-week follow-up cannot answer. The six-month follow-up data, where it exists, is less encouraging than the twelve-week data. The twelve-month follow-up data, where it exists, is less encouraging still.
The system reports the twelve-week data. The twelve-week data supports continued funding. The continued funding produces more twelve-week data.
The industrialisation of loneliness is visible in the job categories it has created. The loneliness researcher. The social prescribing link worker. The befriending service coordinator. The community engagement officer. The social connection programme manager. The loneliness policy advisor. The loneliness awareness campaign communications specialist. Each of these roles has a salary. The salary is paid from the funding that the epidemic framing unlocks. The funding is justified by the epidemic’s scale, which is measured by the instruments the loneliness researchers have developed. The instruments produce the numbers. The numbers justify the funding. The funding pays the salaries. The salaries sustain the sector.
The sector is not indifferent to the lonely person. The people who work in it are, in many cases, genuinely motivated by concern for the people their work is supposed to reach. The structure of the sector does not depend on this motivation. It depends on the continued production of evidence that the epidemic is real and serious and amenable to the kinds of intervention the sector can provide. The continued production of this evidence is what the funding supports. Whether the interventions actually reach the lonely person in a way that changes their experience is the question that the sector’s structure does not reward asking too directly.
The reduction of loneliness to a group social issue—a matter of societal impact, of collective cost, of population-level intervention—performs a specific erasure. It removes from the frame the specific quality of what the lonely person is experiencing, which is not a productivity deficit but a human one. The absence of the person who knows them well enough to notice when something is wrong. The absence of the conversation that goes somewhere rather than stopping at the surface. The absence of the trusted other in whose presence the monitoring can stop and the performance can be suspended. The absence of the slow space in which the three-hour conversation becomes possible.
These are not measurable in GDP terms. They do not appear in the healthcare cost calculation. They are not captured by the loneliness scale score at twelve-week follow-up. They are what the lonely person is actually missing, and they are absent from the framing that has organised the epidemic response precisely because the framing was not designed to describe what the lonely person is missing. It was designed to describe what the economy is losing.
The two descriptions are not the same description. The economy loses a productive worker and a light user of healthcare resources. The person loses the texture of a life that is genuinely inhabited rather than managed.
The response is calibrated to what the economy has lost.
It does not reach what the person has lost.
The person is in the opening paragraph, looking out the window.
After that, they are a unit of calculation.
The calculation does not look back.
The Vorpel Foundation