Take every Swedish man who sat for the mandatory military cognitive test between 1969 and 1997 — a test built to sort conscripts into roles, not to settle a two-thousand-year-old philosophical argument — add roughly 3,000 women who took a similar civilian version, cross-reference all 1.2 million-plus of them against tax records on charitable donations, 2009 European Parliament voting rolls, and national vehicle registries, and then, to make sure the result wasn’t just wealthy families passing down brains and virtue together, run the same comparison on 5,786 pairs of twins where one sibling scored higher than the other. That’s the design behind An Inquiry into the Relationship between Intelligence and Prosocial Behavior: Evidence from Swedish Population Registers, published in The Economic Journal by Mikael Elinder and Oscar Erixson of Uppsala University, and covered Oct. 21, 2026 by PsyPost. People who scored higher on the cognitive tests were dramatically more likely to give to charity, vote, and own an eco-friendly car — and the specific skills doing the predicting were logical and verbal reasoning, not raw aptitude lumped together. Marcus Tullius Cicero drew almost exactly that line in 44 BC, in De Officiis, where he argued that ratio — reason, specifically, not cleverness in general — is the one faculty that builds justice, fellowship, and generosity out of raw human nature.

The Quick Verdict

Question	What we found
What was studied?	Whether cognitive ability predicts charitable giving, voting, and eco-friendly consumer choices, using Swedish administrative data
Who ran it?	Mikael Elinder and Oscar Erixson, Uppsala University
Where and when?	The Economic Journal, DOI 10.1093/ej/ueae105; covered Oct. 21, 2026 by PsyPost
Sample	Over 1.2 million Swedish men plus roughly 3,000 women, scored on mandatory military cognitive tests taken 1969-1997; a 5,786-pair twin sub-analysis checked the result against shared genes and upbringing
Core finding	A one-standard-deviation rise in cognitive ability predicted a 40% relative increase in charitable giving, a 31% rise in voting, and a 14% rise in eco-friendly vehicle ownership
The twin-pair check	Within twin pairs, the same rise in cognitive ability still predicted a 26% increase in giving, a 13% increase in voting, and a 13% increase in eco-friendly vehicle ownership
What Cicero already argued	In De Officiis, Book I, reason (ratio) is the faculty specific to humans that builds justice, fellowship (societas), and generosity (beneficentia) out of raw human nature

The Quick Version

A 2025 study of over 1.2 million Swedes, newly recirculated by PsyPost in October 2026, found that people who scored higher on cognitive tests were substantially more likely to donate to charity, vote, and buy an eco-friendly car — and a twin analysis, controlling for shared genes and upbringing, found the pattern held, just at a smaller size. Logical and verbal reasoning did more of the predicting than raw intelligence taken as one lump variable. Cicero named the mechanism in 44 BC: reason isn’t just a tool for figuring things out. It’s the specific thing that turns a self-interested creature into one capable of justice and generosity.

What the Study Actually Found

Most research on who gives, who votes, and who makes the ethically inconvenient purchase treats those as separate questions, usually answered with a survey and a few hundred respondents who may not remember accurately what they actually did last year. Elinder and Erixson skipped the self-report problem almost entirely. Swedish charities report donations to the tax authority automatically, so the giving data here isn’t someone’s flattering memory of their own generosity — it’s what the government’s records say happened, 2012 through 2015. Voting came from the 2009 European Parliament election rolls. Eco-friendly vehicle ownership came from national registries. None of it required trusting anyone to tell the truth about themselves.

Against that data, cognitive scores from decades-old military conscription tests did real work. A one-standard-deviation increase in cognitive ability predicted a 40% relative increase in the likelihood of charitable giving, a 31% rise in voting likelihood, and a 14% rise in eco-friendly vehicle ownership. The gap at the extremes was not subtle: men who scored highest on the cognitive test were more than five times as likely to donate to charity as men who scored lowest. Run the same question inside 5,786 twin pairs — where genetics and childhood household are held effectively constant, and only the gap in cognitive score between siblings is doing the predicting — and the pattern survives, smaller but real: 26% for giving, 13% for voting, 13% for eco-friendly vehicles. Whatever is producing this effect, it isn’t only “richer, better-educated families raise both smart and generous kids together.”

The detail that matters most for this comparison isn’t the headline number. It’s which cognitive skill carried the weight. The researchers broke general cognitive ability down into its components, and logical and verbal reasoning came out as the strongest individual predictors of prosocial behavior — stronger than raw aptitude scores considered on their own. Not memory. Not spatial skill. Reasoning, specifically, is where the prosocial signal concentrated.

What Did Cicero Mean By This?

In De Officiis, Book I, Cicero writes that it is reason — “by which he comprehends the chain of consequences” — that sets humans apart from animals. He then names what that faculty builds: “the first principle is that which is found in the connection subsisting between all the members of the human race; and that bond of connection is reason and speech.” Reason and speech, in his account, bind humans together “in a sort of natural fraternity” that no other creature shares.

Cicero wasn’t writing a treatise on cognition. De Officiis is a manual on duty, composed for his son, laying out what a good Roman owes other people and why. But his argument rests on a specific claim about where duty comes from, and it isn’t instinct, and it isn’t training alone. He opens Book I by naming four sources of moral goodness — “the full perception and intelligent development of the true,” then justice, “with the conservation of organized society, with rendering to every man his due,” then greatness of spirit, then propriety. The first item isn’t decoration. Everything downstream of it — justice, fellowship, eventually generosity — flows specifically from reason’s capacity to perceive truth and work out consequences, not from good instincts or inherited manners.

The move that matters here comes when Cicero connects that same faculty to kindness toward others. Having established that reason and speech are literally what binds the species into societas, he turns to beneficentia — kindness, generosity — and calls it something that “appeals more to the best in human nature” than almost anything else, folding it into the rational fellowship he just described rather than treating it as a soft virtue bolted on afterward. Justice and generosity aren’t, for Cicero, products of good feeling that reason happens to accompany. They’re things reason itself produces, once a creature capable of grasping consequences and shared standing with others actually does the grasping.

Why Does Reasoning Specifically Predict Prosocial Behavior?

1. Giving and voting both require seeing consequences you won’t personally experience. A donation doesn’t benefit the donor. A single vote has a vanishingly small chance of swinging an outcome. Logical reasoning works out why the action still makes sense anyway — the exact chain-of-consequences faculty Cicero credited reason with in Book I.
2. Verbal reasoning tracks the ability to follow an argument for why someone else’s claim on you is legitimate. Justice, in Cicero’s framework, means rendering to others what’s due — which requires first following the reasoning for why it’s due. That’s closer to a language and logic skill than to memory or speed.
3. Raw aptitude, measured as one lump score, blurs a distinction the data says is real. Logical and verbal reasoning outperforming overall cognitive ability as predictors is itself evidence that “smart” isn’t one undifferentiated thing doing the moral work — some specific component is, the same move Cicero made when he singled out ratio rather than crediting human capacity indiscriminately.
4. None of this requires assuming reasoning produces warmth. Cicero’s account doesn’t claim reason makes people feel more for others. It claims reason lets a creature recognize a shared fellowship and act on obligations that follow — a colder, more structural claim than “smart people are nicer,” and one the twin analysis supports better than a simple warmth story would.

Why the Study and the Philosophy Rhyme

The honest version of this isn’t “a Roman statesman predicted a 2025 Swedish regression.” It’s narrower, and it holds up better for being narrower: Cicero’s specific claim was that reason — not instinct, not upbringing, not raw mental horsepower in the abstract — is the faculty that generates justice, fellowship, and generosity in humans. Elinder and Erixson weren’t testing Cicero. They were testing whether cognitive ability predicts three behaviors in Swedish administrative data. But the detail where their finding gets most precise — reasoning specifically, not intelligence broadly, carrying the weight — lines up with the exact distinction Cicero drew when he credited ratio rather than cleverness in general.

It connects to the Heraclitus piece on West Point cadets, which found conscientiousness beating raw IQ as a predictor of leadership — a reminder from the opposite direction that “intelligence” and “good character” aren’t the same variable, even when a study finds them correlated, as this one does for reasoning specifically. It sits close to the James piece on disgust and voting: both are, at bottom, about which internal faculty actually does the work behind a political behavior that looks like a simple choice from outside. It rhymes with the Tocqueville piece on democratic backsliding — Tocqueville argued abstract commitment to democracy isn’t enough to produce democratic behavior under pressure, and this study’s voting finding raises an adjacent question: whether reasoning itself is part of what closes that gap. And it stands in useful tension with the La Rochefoucauld piece on strategic kindness, which found generosity driven by Machiavellian calculation rather than real fellow-feeling — a reminder that reasoning can serve self-interest as easily as it serves Cicero’s societas.

The Gap Cicero Didn’t See Coming

Here’s where the comparison is more interesting than a clean match.

Cicero was writing for an audience of one — his son, training to take a place among Rome’s governing class — and his argument assumes a fairly uniform picture of what reason is and does. He had no way to anticipate that “reason” might fracture into separable components under measurement, with logical and verbal reasoning outpredicting general cognitive ability on the exact behaviors he cared about. His ratio is a single, unified faculty. The 2025 data suggests the thing doing the moral work might be a specific slice of it, not the whole.

He also had no concept of a population-level correlation running at a fraction of a standard deviation across 1.2 million people he’d never meet. Cicero’s examples are individuals — specific Romans making specific decisions about specific duties. The study’s claim is statistical and aggregate: on average, higher reasoning scores track a higher likelihood of three particular behaviors. That’s a different kind of claim than “a reasoning person will act generously,” and his framework, built for advising one young man, was never built to distinguish the two.

What This Study Doesn’t Show

Be precise about the size of this. The framing invites more confidence than the data earns.

Overwhelmingly Male, and Built on a Military Test

The core sample is over 1.2 million men, because Sweden’s conscription test was mandatory for men across these decades and not for women. The roughly 3,000 women included took a similar but separate, non-mandatory test. A finding this heavily weighted toward one sex, measured through a military screening instrument, isn’t automatically a finding about cognitive ability and prosocial behavior in general.

One Country, One Welfare State

Sweden’s giving norms, voting infrastructure, and the cultural meaning of owning an eco-friendly car all sit inside a specific welfare state with specific donation tax incentives and green-vehicle policy. Whether the same relationships hold under different incentives or institutional trust is untested here.

Correlational, With a Genuine but Partial Causal Check

The twin sub-analysis is a real strength — it rules out the simplest confound, that wealthier, better-educated families produce both smarter and more generous children together. It doesn’t rule out every confound. Twins still share a household and a set of parental values even when their cognitive scores diverge. The study narrows the causal story. It doesn’t close it.

Three Specific Behaviors, Not Virtue in General

Charitable giving here means tax-deductible donations reported between 2012 and 2015. Voting means participation in one election, the 2009 European Parliament vote. Eco-friendly vehicle ownership means owning one type of car during a specific registry window. These are real, well-measured behaviors — but not a general measure of kindness or moral character, and the study doesn’t claim to be one.

What Actually Follows From This

1. Being “smart” in general isn’t the finding here — reasoning specifically is. If this result holds up, the useful takeaway isn’t “intelligent people are better people.” It’s that two particular cognitive skills, logical and verbal reasoning, track a willingness to act on obligations that don’t pay the actor back directly — which is a narrower and more interesting claim.
2. Cicero’s order of operations is worth taking seriously on its own terms, independent of the regression table. He didn’t argue that generosity comes first and reason gets attached to it afterward as a polite justification. He argued reason comes first, and justice and generosity are things it builds. Whether or not the 2025 data holds up under further scrutiny, that’s a claim worth sitting with regardless: fellow-feeling that can survive scrutiny is usually fellow-feeling that’s been reasoned through, not just felt.
3. Ask Cicero’s actual question about any claim of generosity you encounter, including your own: what did the reasoning behind it actually trace out? Reflexive kindness and reasoned kindness can look identical from outside. Cicero’s bet was that only the reasoned kind reliably survives contact with a situation where generosity is costly — which is close to the territory this study’s twin analysis, and its emphasis on logical reasoning over raw aptitude, is quietly gesturing toward.

Cicero never ran a regression on 1.2 million Swedes, and Elinder and Erixson weren’t trying to vindicate a Roman statesman’s account of duty. But a study built to find out whether cognitive ability predicts giving, voting, and green consumer choices — and which specific component of cognition does the predicting — landed close to the distinction Cicero insisted on in 44 BC: it isn’t cleverness in general that builds justice and generosity out of a self-interested creature. It’s reason, doing the specific work of tracing consequences and recognizing what’s owed to people who will never pay you back for noticing.

---

This post discusses a 2025 study published in The Economic Journal (Elinder and Erixson, DOI: 10.1093/ej/ueae105), newly covered by PsyPost in October 2026, and Cicero’s De Officiis, Book I. The study is correlational, drawn overwhelmingly from a Swedish male conscription sample, and its twin sub-analysis narrows but does not eliminate the possibility that shared upbringing, rather than cognitive ability itself, drives part of the result.