Someone to count on

The single best-tested fact in the science of happiness is also the least surprising: people do better when they are not alone. Every year Gallup asks adults in some 150 countries a deceptively plain question — if you were in trouble, do you have relatives or friends you can count on to help you whenever you need them? — and the answer divides the world more sharply than almost anything else it measures. Those who say yes rate their lives, on average, at on the 0–10 Cantril ladder. Those who say no rate them at . The gap, about points, dwarfs the move from poor to rich within most countries.

It shows up in feelings, too, not just judgments. We keep the two strictly apart, because they are different things: the ladder is an evaluative measure, a verdict on your life; daily negative affect — whether you felt worry, sadness, stress, anger or pain “yesterday” — is an experiential one, a readout of the day. On the experiential side the supported report negativity of against for the unsupported. To have someone is to rate your life far higher and to feel the day far lighter.

None of that proves support causes well-being — happy, settled people also tend to keep their networks, so the arrow runs both ways. Read everything here as association. But the size of the association is not in doubt, and it raises a sharp policy question. If the warmth of a private safety net matters this much, can a public one stand in for it? Where the state spends heavily on health and citizens are richer, does the penalty for having no one to call shrink — the welfare state quietly playing the role of the absent friend?

The size of the premiumone coefficient per country, then the distribution

To ask the question cleanly we do not compare a supported person in one country to an unsupported person in another — response styles and circumstances differ too much. Instead we fit a regression inside each country, weighted by Gallup’s within-country weight, holding age, sex, education, marital status and income rank constant, and read off one number: how much higher is the outcome for someone with support than without. Call it that country’s support premium. Do it once per nation and a distribution of premiums emerges — one per country — whose spread is the evidence. A country enters only with at least 300 complete cases across at least two survey waves; clear the bar.

The premium is almost universal. The life-rating premium is positive in of countries; weighting each country by its population, the representative person’s premium is about ladder points, and the median country’s is . The experiential buffer is just as consistent in the other direction: the supported feel less daily negativity in of countries, a population-weighted buffer of about on the negative-affect index. Wherever you look, someone to count on is among the most valuable things a person can have.

The substitution test, reversedthe slope of the slopes runs the wrong way

The substitution hypothesis makes a falsifiable prediction: regress each country’s support premium on how strong its public safety net is, and the slope should be negative — the premium small where the state is large. The data deliver the opposite sign. The life-rating premium rises by about ladder points for every additional point of GDP a country spends on health, with a 95% interval from to that sits entirely above zero; national health spending alone accounts for roughly of the spread in premiums across the world. Sort countries into thirds by health spending and the average premium climbs from in the lowest-spending third to in the middle and in the highest. Richer countries show the same pattern: the premium also grows with GNI per capita, by about ladder points per additional thousand dollars.

The experiential side tells the same story from its mirror. The daily-distress buffer — already negative everywhere — grows more negative as health spending rises, deepening by about per point of GDP (95% interval to ). So the welfare state does not rescue your rating while leaving your feelings exposed, nor the reverse. It substitutes for neither. On both constructs the value of a private safety net is largest exactly where the public one is strongest. Health spending and national income are themselves correlated (about across countries), so the two moderators partly tell one story; but the health-spending gradient holds even when GNI is controlled in the same model, with a slope of about .

So why does the gap widen?who is left without support when the state is strong

A reversal this clean demands an honest mechanism, and the most likely one is not that generous states fail the friendless. It is a question of who the friendless are. Where the state is strong, support is nearly universal, so the people who report having no one are a smaller, more selected, more marginalized group — and their disadvantage runs deeper. The share of adults without anyone to count on falls steadily as health spending rises: from about in the lowest-spending third of countries to roughly in the highest (a correlation of about across countries). When almost everyone has support, lacking it is a more unusual and more telling deprivation, and the ladder leaves more room to fall.

That selection story also fits the extremes. The largest life-rating premiums sit in high-spending, high-income places — , and all post premiums above a full ladder point — where to be without anyone is to be conspicuously outside a near-universal norm. The smallest premiums cluster where support is scarcer and more evenly shared, so having it sets you apart less.

Where the premium is largestthe support premium on the world map

Mapped, the evaluative premium is brightest across the wealthy, high-spending world and darkest where public provision is thinnest — the geography of the reversal in a single view. Seven small states are absent from the base atlas and are left unshaded; the pattern across the rest is unmistakable.

What this does and doesn’t showreading the reversal honestly

This is an association, repeated inside countries, not a causal claim. Support and well-being feed each other; the same temperament that keeps a person buoyant helps them keep friends, so we never say support “causes” a higher rating. The widening of the premium is best read as a selection effect — a shift in who lacks support — rather than proof that the welfare state cannot help; both readings are consistent with these numbers, and we have shown the selection evidence rather than asserted the mechanism. The country context is a static snapshot: health spending and GNI per capita are joined from a single roughly 2015–2018 vintage and held fixed across all survey years, so they describe a country’s standing, not its history. The two moderators are collinear, which is why we report each alone and then together. Cross-cultural differences in how people answer survey questions are real, and raw ladder gaps smaller than about a fifth of a point should not be over-read. The population-weighted “world” figures lean on total population as a proxy for the adult population Gallup actually samples. And the negative-affect index is a coarse five-item average. With all of that granted, the central result is robust to weighting and survives controls: the value of a friend grows, rather than fades, where the state is strong.

The welfare state does many things, and easing material hardship is plainly one of them. What it does not appear to do is make a personal safety net redundant. If anything, the data suggest the opposite reading is the humane one: a strong state does not retire the question of whether someone has anyone to call — it sharpens it, by leaving the unsupported fewer and more exposed. The friend, it turns out, has no public substitute.