Social norms: the downside
"Surely under well-defined circumstances, ye CAN push yer grannie off a bus?"
If self-interest is the engine of society and the master key to social science, then social norms must come a close second. Humans are rule-followers who care deeply about what others think of us. Internalized rules are probably a key mechanism enabling us to pursue our goals and interests. The study of these rules, in particular the moral rules that society promulgates, could help to explain a lot of the social world. Almost every big organization works not just by paying its employees or ordering them about, but by developing rules of behaviour to guide them. Modern corporations put huge, deliberate effort into building the right culture. In short, norms are powerful.
They’re useful, too. Without norms, humans would find it much harder to cooperate. Norms set up expectations and help us to coordinate with each other. They push us to be less selfish. New norms, like a concern about animal welfare, are a force for moral progress, which is just as important as material progress.
Social norms also have downsides. It’s especially important to understand this today, because norms are becoming increasingly powerful, as people’s actions become ever more visible to their peers and potentially to the whole planet. When the investment industry turns on a dime to embrace ESG, when politicians apologize for the crimes of past history, when the biggest crooks espouse the highest ideals, these are signs we are living in a normative world.
How norms work
Here’s a simple model of a social norm. There is a norm to do X — wash your armpits, die for freedom, refrain from pushing yer grannie off a bus, whatever — if (a) people believe that one should do X and (b) people believe that other people believe that one should do X. So, norms involve what economists call second order expectations: beliefs about other people’s beliefs. On this account, social norms affect behaviour because humans care about conforming to other people’s moral beliefs.
This model straight away gives us norms’ first problem.
Norms have no efficiency guarantees
Under certain circumstances — free exchange of goods in the marketplace with no externalities — human self-interest leads to efficient outcomes, i.e. outcomes where nobody can be made better off without making someone else worse off. This is economics’ famous First Welfare Theorem, and the Second Theorem is like, namely this, that you can achieve any efficient outcome via free exchange, with maybe a one-off redistribution first. The theorems’ conclusions hold only under restricted circumstances, and a lot of the interesting parts of economics explain what happens when they don’t hold. Nevertheless, they do hold when they hold, and this truth may even, in some very broad way, have to do with why humans are much richer than they were one, two or three centuries ago.
Social norms don’t have any such guarantee at all. Given the model above, a powerful enough norm can get people to do anything:
Wear uncomfortable clothes.
Use elaborate forms of speech.
Cut down a tree every summer and dance round it.
Cut down a tree every winter and put glass baubles on it.
Bury your dead relatives.
Burn your dead relatives.
Eat your dead relatives.
Kill cockroaches.
Kill people.
Call people cockroaches, then kill them.
There’s no weird or harmful behaviour that this model is incompatible with. And indeed we see all these forms of norm-driven behaviour in the real world. Some people think that norms are usually efficient, but at the broadest level there doesn’t seem much evidence for that.
Extremely inefficient norms, like those which support outgroup genocide, are the darkest of downsides. They are talked about a lot, including in the academic literature, so I won’t discuss them here.
The lack of guaranteed efficiency affects norms in a subtler way. It means that even good norms may not respond to costs. Again, contrast that with self-interest. If I want an apple, but I also want a movie ticket, then I will have to trade off these two desires, and humans are often reasonably good at making those trade-offs in a satisfactory way. When apples get more expensive, I buy less of them.
Norms don’t have this built in, because moral beliefs are free. When the weather gets hotter, wearing a suit and tie becomes more uncomfortable. But thinking that other people ought to wear a suit and tie doesn’t get more uncomfortable.
There are examples where judgments are unresponsive to numbers — where people will spend the same, say, to prevent a public health hazard that kills a hundred lives or a thousand lives.1 I don’t think this is a general fact about human cognition. If you multiply the price of Mars Bars by ten, consumers will absolutely adjust their behaviour. It is a fact about costless cognition. Like Sissy in Hard Times, we think it’s important to say and feel that a hundred/a thousand/ten thousand deaths are very wrong, and so the same level of disapproval works in each case:
‘And I find (Mr. M’Choakumchild said) that in a given time a hundred thousand persons went to sea on long voyages, and only five hundred of them were drowned or burnt to death. What is the percentage? And I said, Miss;’ here Sissy fairly sobbed as confessing with extreme contrition to her greatest error; ‘I said it was nothing.’
‘Nothing, Sissy?’
‘Nothing, Miss — to the relations and friends of the people who were killed. I shall never learn,’ said Sissy.
The Yimby (“Yes In My Back Yard”) movement in the US has sprung up to fight for less restrictive zoning and more housing, often coming up against Nimby (“Not In My Back Yard”) opponents. Broadly, this is a contest of ideals against self-interest, and broadly I think the Yimbies are right. But the self-interest of Nimbies also carries some moral weight — it’s genuinely a loss to people when transport systems get more crowded, it’s even genuinely a loss when a given neighbourhood gets less “exclusive” than it was! I don’t expect Yimbies to be good at calculating the perhaps rare cases when that loss outweighs the gain.
A reasonable model of politics is that left-wing parties are norm-driven and right-wing parties are driven by self-interest. I speak as a conservative! It may really be true that, as Tony Blair once said, people on the left are “pretty straight guys”. The downside of being ideal-driven is that your activists’ ideals may not involve much cost-benefit analysis. Canny right-wing politicians can slice off chunks of their opponents’ support by embracing idealistic causes that have little or no cost, and perhaps little or no benefit. David Cameron was great at this. Gay marriage, foreign aid and global warming were all ways to rebrand the Tory party. Some of these are unimportant in the grand scheme of things. Others are important, but carried serious costs only in the post-David-Cameron future, which arrived sooner than he expected. But the left wing is also quite capable of doing this to itself. The Scottish National Party has just spent scarce parliamentary time making it easier for teenagers to choose their gender. This is a non-issue for almost every person in Scotland, but presumably it makes SNP activists feel good.
In extreme cases, cost-free norms can become really absurd. On a flight to Bombay in the 1990s, the inflight magazine had an article: Is India Racist? The investigation proceeded by sending a black guy into various high-class shops to see how he was treated. 1990s India was a country of almost a billion people. The number of Africans or African Americans, even including ethnic groups like the Siddi who have been in India for centuries, is essentially a rounding error. By contrast it contained hundreds of millions of low-caste people, facing discrimination, entrenched poverty, debt bondage, you name it; and about a hundred million Muslims, who then as now were quite often — never mind discriminated against! — persecuted and murdered. But in the very rich US, people care about anti-black racism, and it’s important to be like the very rich US. So: how did Bombay shop assistants treat a black guy? With startled amazement, I suppose.
Because self-interest is a more reliable driver of behaviour than norms, social movements often start as norm-driven but devolve into self-interest. In early modern England, religious foundations endowed grammar schools, which helped to create a highly literate population for the era. By the nineteenth century, many of these schools were no longer fulfilling their function, and sometimes they were simply run for the benefit of the staff. The Free Software movement started in a blaze of idealism, but over time, the community has recognized that sustainable open source projects need ways to reward their contributors, either directly or indirectly. Charitable organizations have explicit missions, but the implicit mission is to keep the organization going. Inevitably much of their effort is spent on appealing to donors. They rarely decide that their funds would serve their explicit goals better in the hands of another, similar organization.
Doing the right thing for the wrong reason
Other problems of norms become visible if we focus on the mechanisms leading from our expectations about other people’s moral beliefs, to our own actions. Bluntly, norms often involve people doing good things for bad motivations. Valuable norms may be followed out of cowardice and conformity, enforced by bullying and sanctimony, and spread by hypocrisy and humbug. The Victorian lady puts money in the church collection because that is what everyone does, and for fear of gossip otherwise.
There is a social science tradition, going back to the eighteenth century, which sees this as a price worth paying. “Hypocrisy is the tribute vice pays to virtue”, says Rochefoucauld. In the Theory of Moral Sentiments, Adam Smith describes how our desire for others’ approval drives the economy and even innovation. Fair enough. If we relied exclusively on people’s innate virtue to ensure good behaviour, it would often come up short.
But the bad motivations are not a perfect substitute for innate virtue. They can make cooperation fragile. When the eye of public attention moves on, suddenly good behaviour collapses. Or people pay lip service. We buy expensive, environmentally costly reusable shopping bags, load them up and walk home, virtuous message displayed on the outside. Next week, having forgotten the bag at home, we do the same thing again. Couldn’t we just reuse an ordinary shopping bag?
Social media has exacerbated lip service. It can vastly increase the visibility of anything people say, but it has much less effect on the visibility of what they do. Twitter is a lip service playground. The 2010s created a whole vocabulary of villainous behaviour — gaslighting, punching down and all that — almost entirely focused on talk.
Like chimpanzees pant-hooting at a fallen alpha male, norm enforcement is a group activity. It’s an important way to keep powerful individuals in line. Power is a fragile and subjective thing, and the distinction between collective self-defence and bullying is subtle and sometimes arbitrary. If a group can enforce a norm, it may also be able to extract resources (I co-wrote a paper about this). Normative entrepreneurs, who coordinate attacks on norm violators, may exploit this for their own benefit.
People don’t always follow norms out of fear or conformism. Sometimes they truly believe in what they are doing. Unfortunately, one difference between acting from ideals and acting from self-interest is that there is no feedback loop. If I sell clothes to make a profit, and the market doesn’t want my clothes, I adjust my line or go out of business. If I give away clothes to feed the poor, there is no such mechanism. The extreme version of this phenomenon is when rich women distribute fur coats to the homeless. A certain amount of charity can be seen as high-end luxury consumption. Gulnara Karimova, the daughter of Uzbekistan’s enemy-boiling dictator, played at philanthropy while Uzbek schoolchildren were sent out of school to bring in the cotton harvest. The libertarian movement in the US is sometimes accused of catering to the self-interest of billionaires. If it did, it might be more coherent and impactful. Actually, it caters to their ideals — or you might say their whims. On the other side of the ideological divide, Socialist elites regularly pushed traditionalist norms, whether in art (Socialist Realism) or in standards of behaviour. Xi Jinping has similar inclinations. That’s what happens when you put old men in charge.
This disadvantage cannot be disentangled from norms’ essential advantage, that they free us from the short-run feedback of self-interest and let us dream of different worlds. To make his billions, Bill Gates provided incremental improvements to software for the developed world. He then dedicated that wealth to big, bold projects of human improvement, focused on the global majority. The second half of his career may be much more important than the first, if he gets enough big calls right.
Norms front collective self-interest
So far, I’ve talked as if norms are essentially arbitrary, and the model allows that. If enough people hold moral belief X, they will probably be able to enforce it, whatever X is. There is still room to ask what moral beliefs tend to be popularly held. One plausible regularity is that norms embody the collective self-interest of a group. They can do so explicitly: patriotism is a norm that favours a given nation, and most organizations similarly expect their members to support the team, not the world as a whole. (This might be a good thing! Collective self-interest can solve problems that either individual self-interest or pure global altruism cannot.) They can also conceal collective self-interest under a veneer of universalism. Late medieval anticlericalism was driven by the suspicion that the Church had its own interests at heart more than the interests of all Christendom. Marxists took the view that all of bourgeois morality was a swindle in this way. The modern rallying cry of social justice might really just mean “more for my group”. In war, the combatants usually claim they are fighting for some universal principle. In the Great War, Germany fought for Kultur; contemporary Russia is apparently struggling against wokeness, while Ukraine is defending Western civilization.
These universal claims needn’t be completely misleading. It is probably better for a ruling class to have a sense of noblesse oblige than to be straightforwardly dedicated to maximizing its take. On the other hand, naked oppression has the advantage of honesty. At least you know where you are with it. A cynical economic model would be that the guardians of any normative order, what Coleridge called the “clerisy”, must be given their due: not so much that they provide no social benefit, but enough that they have an incentive to maintain socially beneficial norms. Anyway, whether it’s open or hidden, norms can be a mechanism for one group to dominate and exploit others.
Further reading
Most norms are good. They encourage us to be cooperative and unselfish. The world would be a better place if norms were followed more at the margin. When a norm is explicitly invoked, it is usually in response to some perceived wrong. This makes it easy to imagine that norms are always good, and/or that we should wish for a world where all decisions are driven by norms. The reasons above — the bad motivations that lead people to obey norms, and above all their lack of efficiency guarantees — make me think otherwise. A completely norm-driven world could be quite dystopian, and though evil norms are rare, they can lead to bigger evils than ordinary self-interest.
Norms are a big deal, now more than before because communications technology is a norms enabler. Normative claims and counter-claims circulate around society, in parallel to the goods and money that circulate around the economy, and today they do so on the same global scale. Experimental economists and others have developed a reasonable “microeconomic” model of norms, along the lines sketched above. There should be a lot of potential in integrating it into macro models of economics and politics. The downsides of norms offer some interesting research avenues. This isn’t a literature review, but here are a few academic papers that are worth reading if you want to go deeper:
Cristina Bicchieri, The Grammar of Society. Book with a simple definition of social norms along the lines above.
Krupka and Weber, Identifying social norms using coordination games: Why does dictator game sharing vary? Important paper on how to measure norms.
Kimbrough and Vostroknutov, Norms make preferences social. Making the case for how social norms affect other-regarding behaviour in lab experiments.
Jonathan Haidt, The emotional dog and its rational tail: a social intuitionist approach to moral judgment. Hugely influential statement on the emotional bases for social norms.
Shweder, Culture and moral development. Old but still really interesting cross-cultural study of how different groups come to accept different norms.
Feddersen and Sandroni, A theory of participation in elections. Politics with ethical voters. A starting point for integrating norms into political economy models.
If you enjoyed this, you might like my book Wyclif’s Dust: Western Cultures from the Printing Press to the Present. It’s available from Amazon as a paperback/hardback/ebook, and you can read more about it here.
You can also subscribe to this newsletter (it’s free):
And, if you liked it, please share this post with others.
I couldn’t find a reference for this, but I think it is somewhere in Thinking, Fast and Slow.
Linked by Rob Henderson. I sent it on to a few and bookmarked it.
> There are examples where judgments are unresponsive to numbers — where people will spend the same, say, to prevent a public health hazard that kills a hundred lives or a thousand lives. (I couldn’t find a reference for this, but I think it is somewhere in Thinking, Fast and Slow.)
The canonical name is "scope insensitivity." The original experiment was about saving wildlife (bird) lives not humans, but it went as you said, and replication results do appear to track for harms of whatever kind. "Economic Preferences or Attitude Expressions?: An Analysis of Dollar Responses to Public Issues", https://doi.org/10.1023/A:1007835629236