Everything Is Obvious 3/10

Parts 1, 2.

3/ The Wisdom (and Madness) of Crowds

  • Circular reasoning – X succeeded because X had the attributes of X – pervades common sense explanations for why some things succeed and others fail. The fallacy here is that we don’t know any other meaningful attributes and tend to explain the success by the ones we know.

  • If something (previously unthinkable) happens, there’s no shortage of people explaining the outcome as a result of a social trend, but the only way to know that this trend exists is this very outcome. Circular reasoning again.

  • The micro-macro problem: society as we know it requires huge groups of people to care about rules and principles, but it’s up to an individual to actually act in a certain way to make sure the society exists. There’s no clear dividing line between the society and the individual.

  • Knowing everything about each individual atom doesn’t mean knowing everything about the composition of atoms. Ecosystems are much more than their individual members. Social systems have all kinds of interactions between people, companies, markets, society and the government (pretty much the list of stakeholders for S/ESG).

  • Said interactions are not necessarily commercial: they can be influential, too (an Instagram influencer tells me what to wear, my company sets rules of conduct for our suppliers, government talking heads hint at new policies, government officials get lobbied by firms, etc.).

  • The “behaviour” of social actors is a shorthand for the aggregate behaviour of a large number of individuals. Such actors are proxies for all kinds of behaviours (think market segmentation). What’s worse is that for policy planning firms are reduced to a handful of “representative firms” that supposedly possess more or less the same major features and expose the same major behaviours of the firms the government / analyst is looking for.

  • In practice this is not the way to do planning and one will be laughed off should they try to reduce lots of firms with vastly different internal features to the single agent. At least here the “common sense” is recognised as nonsense.

  • The behaviour of people making a crowd is NOT independent (thus the crowd’s behaviour is NOT the sum of people’s individual choices); if anything – it’s interdependent and very heavy on mimicking what others are doing. Also, the larger the number of misbehaving people in the crowd – the higher the chance that others would join in (switching their personal preferences from peace to riot).

  • The trigger for the negative behaviour depends on one’s threshold towards a certain activity. Anything below the threshold – and the person stays civilised but nearing or passing the threshold can turn the person into an evil monster. Everyone has a threshold but it’s different for everyone (and often dependent on the financial and social status of the person, but also on the inherent craziness). Individual thresholds in the crowd can be displayed as a distribution (not necessarily normal due to the self-selection) and can be analysed (maybe it even has predictive power?). But what’s even more important is that these thresholds can depend on other people’s thresholds.

  • In small groups people have been observed to change their preferences in order to fit in and also have a shared cultural context. [MK: I know it’s a cliché, but a useful one]

  • Popular items get cumulative advantage in a way that popular things tend to become even more popular; minor differences accumulate and lead to the “butterfly effect”. Other people’s choices affect our choices.

  • When individuals are influenced by what other people are doing, similar groups of people can end up behaving in very different ways. (Granovetter’s riot model)

Part 4.