The two kinds of social incentives

And how they can distort directional consensus

Epistemic status: Talking about a subclass of a more general phenomenon, and the boundaries of the subclass are fuzzy.

Incentives

The core insight of economics is that people respond to incentives, but it’s important to notice there are actually two different mechanisms by which this happens:

• Pursuing rewards (or avoiding punishment): If you offer employees a pay raise when they do a good job, they’ll try harder to do a good job so that they can get more money¹. Similarly if you punish your children whenever you catch them stealing cookies, they’ll be more reluctant to steal cookies (or at least work harder to avoid getting caught).

• Selection effects: If you fire employees that do a bad job, you’ll (hopefully) soon be left with only employees that do a good job. On a larger scale, businesses that do a good job succeed and grow while businesses that do a poor job die out. This is how darwinian evolution works (as opposed to, say, giraffes deciding to work hard to grow longer necks to reach more leaves).

These both push in the same direction, but it’s important to notice that selection effects scale and rewards don’t. A restaurant in a small town where it’s the only place to eat out isn’t going to face much selection pressure to improve its food². A restaurant in Manhattan with bad food or service will last approximately fifty three seconds. But offering bonuses to good waiters will work just about equally well in both places.

This is a problem, because it’s a social effect that changes as society scales up and becomes larger and more global and interconnected. People become more fungible, and selection pressures become more intense. Institutions built for villages or small towns or, at best, countries where moving between towns is hard and expensive for most people, start working in unexpected ways. This is most visible in rat race dynamics - people feel more stressed and more precarious even as, on an objective scale, most people become safer and more prosperous, because the selection effects for getting in on things (or getting pushed out) become more extreme. Dating apps in big cities also famously have issues with this. At the extreme, this turns into Korea’s Hell, Joseon.

Effects

Well, that’s the more general phenomenon. I mainly want to talk about a specific manifestation of it that makes some social structures act weird. In general, it goes like

1. There’s some tradeoff guided by social norms, where making something harder or easier helps some people and hurts others.

2. The people it helps are, in general, high status while the people it hurts are lower status.

3. Even in a small scale society, reward effects incentivise the people it helps to to talk about it more while the people it hurts talk about it less (since people want to look high status).

4. But in a large society with selection effects (like, say, the entire internet), you also get selection incentives: High status people get a megaphone (people both quote them and want to flatter them), while low status people have less reach.

5. So while in a normal society you could manage these tradeoffs without too massive a skew, once you’ve scaled up society you’re suddenly skewing much harder to one side than you should.

Let’s go through three examples here³.

Approaching women

This one has somewhat surprisingly faded a bit from public discourse⁴, but a few years back talking about how terrible men approaching women too much was and how men who did it were creeps (this discourse came to a peak with the Scott Aaronson affair). Meanwhile in real life, many of my female friends often enjoyed being approached or told me they wished they were approached more since they weren’t sure how to go about dating, and when they did complain about being approached it was in the context of someone doing something obviously creepy or wrong.

There’s a clear tradeoff here between being too permissive with approaching women (which can make women feel uncomfortable or harassed if it happens too much, even if the approaches are each individually nice and friendly). On the other hand women do enjoy dating and often enjoy social interactions with nice men, and it can feel nice and flattering to be approached even if you turn it down.

But this isn’t symmetrical. Very attractive women get approached too much and would probably like the approaching threshold to be higher. Less attractive women would prefer it to be lower. But being more attractive is higher status, so women on the fence will be incentivised to say they get approached too much - and more importantly the women who do get approached too much all have a social megaphone and can dominate the conversation. So the threshold will get pushed pretty hard towards their preferences.

NIMBYism

Again, there’s a tradeoff in how hard we make it for people to just build things. Making it too easy can lead to environmental damage, pollution, or Bauhaus architecture. Making it too hard leads to homelessness, poverty, and stagnation.

The status incentives in this case aren’t quite so clear-cut, but they’re still pretty biased. Being concerned with growth and wanting things to change is lower-status, since it implies dissatisfaction with your current place. Being concerned with the environment, or fancy architecture, is high-status. Supporting having to spend more on amenities or fancy extras implies wealth and is high status. And in local politics, people who’ve lived in the same place a long time and like how it used to be also have much better connections in city hall or local media and thus more power to push their side, while people who would like to move there if there was enough housing for them (but can’t) have zero sway.

Or more specifically, consider inclusionary zoning (the practice of requiring new buildings to include rent-restricted units that can only be rent to low-income people). They make all the other units more expensive (partly because they effectively have to cover the extra rent for the subsidized units, partly because this leads to less housing being built in total). But complaining about the difference between, say, a $2900 rent and a $3200 rent on the same apartment is very low-status - it implies you’re just barely rich enough to make it into the “rent a nice apartment” class, so you’re either a poser or a class traitor. Meanwhile advocating for more IZ is high-status - it implies you’re (a) concerned for the poor and (b) rich enough not to mind the extra rent. And of course for the potential housing lottery winners it’s nice to have the units. So again, we see a status skew.

Privacy

Same drill: Too much concern with privacy settings is inconvenient, too little and people spy on you. But concern with privacy is high status, partly because it implies you have big important secrets to hide (the rewards incentive component) and partly because famous or important people are a lot more likely to have people digging into their affairs (especially politicians subject to opposition research), which causes selection effects. So once again, we see a disproportionate concern with privacy.

Gotta have at least one image per post, so here’s one showing an inconvenient amount of concern with privacy

Note that the actual socially-imposed privacy laws we have are mostly kind of terrible. Everyone hates GDPR, everyone hates having to click on a cookie banner every website they visit, and while I haven’t seen opinion polling I’m pretty sure everyone hates the HIPAA thing where it’s impossible to get your own medical information because the doctor isn’t allowed to just email you your test results, he has to send you an indirect link to a portal on a HIPAA-compliant patient-facing website that doesn’t actually work⁵.

Limits and disclaimers

There’s an obvious disclaimer I don’t think needs to be said about how these skews don’t mean you should never consider the other side. Sometimes women do get approached too much, some places have seen environmentally disastrous projects built that should have been reconsidered, and it’s nice that we can keep our credit card numbers private. A general bias to one side should still be overrulable by case specific considerations.

A more interesting disclaimer is that we can’t just assume the right threshold is whatever we had as a smaller society. The actual object-level considerations on these also scale with society - a larger means women are approached more (and differently), that there’s more people affected in unexpected ways by any construction project, that there’s more people out to steal and misuse your personal information. There’s legitimate new considerations that aren’t just artefacts of misaligned social incentives.

It’s also worth noting that this bias is not self-reinforcing, which does limit the scale of the problem. As we move thresholds away from where people prefer them, more people are inconvenienced and this causes a backlash. We saw this with the first one - people are a lot more reasonable about it and less hysterical about calling ordinary male behavior creepy these days. There’s a few other dynamics involved⁶, but one big thing is that it just became a genuine problem for both men and many women (who did want to be approached more even if it was embarrassing to admit it), which created enough pushback for it to fade away. The rise of the YIMBY movement as NIMBYism started doing more and more damage is another example⁷.

Finally, it’s worth answering the obvious question - If there’s such strong social incentives to not talk about all this, how come I’m doing it? I get away with these through a mix of being exempt (I’m not a woman who wants to come off as attractive and no one ever complains about straight men being hit on too much, and I decided to embrace an online brand of using my real name, which gets me around the privacy issue by using a different status strategy), and genuinely not being subject to status cycles (e.g. I make enough money that no one’s going to question whether I’m complaining about NIMBYism because I can’t afford a nice apartment⁸. It’s not that I’m extremely brave and virtuous, I justh appen to be safe from the dynamic in these examples.

1. Which can be exchanged for goods and services.

2. Unless it gets bad enough that people just give up and start cooking at home; you can usually imagine some way to screw up so badly that you stop existing.

3. Note that there’s a similar but distinct phenomenon of evaporative cooling, where a group (either a political advocacy group, or an apolitical group that becomes political) becomes more extreme as extremity becomes a mark of status and non-extreme people are pushed out or leave. There’s some similarities and it leads to a similar form of bad tradeoff management, but that’s mostly a different thing.

4. Possibly thanks to dating apps making cold approaches less of a thing in general.

5. This isn’t getting into more controversial stuff, like security cameras and banks monitoring accounts for illegal activity, that I think is obviously good but doesn’t have the same consensus. I think it’s because the benefits are more indirect - having less crime isn’t easily noticeable or directly attributable - so people tend to have more ideologically based opinions on it, which means the social incentives come into play more.

6. I mentioned dating apps earlier, but there’s also the general alienation of society - younger generations are dating and socializing less in general, which is mostly sad but does push this tradeoff in the other direction. There’s also the fact that a lot of the original women leading the movement aged out of their peak attractive years and are no longer getting hit on too much.

7. I’m less sure there’s a privacy backlash, although people do seem to be complaining more about GDPR.

8. This is probably why the YIMBY movement mostly started with techies, instead of people getting narrowly priced out.