When Good Proxies Turn Bad

Posted February 25, 2020 in politics  |  No Comments so far

Proxies are important in our society.

By “proxy” I mean a situation where something small and quantifiable happens, and we observe it, and then infer things from it about how to handle a much larger and unpredictable process.

A classic example of this is a job interview. It’s a linear process with a definite end-point. The phases are defined in advance and rarely change along the way. Each candidate moves along their own track in isolation from the others, until they either hit a brick wall or get hired.

Most actual jobs are very different from job interviews – they’re open-ended, there’s a degree of dynamism or unpredictability, different organisational forces are interacting all the time – yet, in general, performance in a job interview is seen as a good proxy for performance in a job.

There are, of course, very valid criticisms of this view, but that’s not really the point of this post so I’ll leave that alone for now.

A driving test is another example of this kind of proxy. If someone can adhere to a checklist of rules and procedures while driving around under observation for 40 minutes, we let them hurtle around for decades in huge steel boxes, entrusting them not only with their own lives but those of everybody around them.

And again, of course, there’s a discussion to be had about the validity of this proxy, but this has to be left alone too.

At the societal rather than individual level, there’s a proxy relationship that inescapably affects us all: elections.

In a typical democracy like the UK, the central idea is that the party that does the better job of fighting an election campaign will also do a better job of what comes after the election, namely running the country (or local council, or city hall, or whatever the election is for). Electoral performance is a proxy for performance once in administration.

This proxy has always been a bit shonky too, like the job interview or the driving test, but recently I’ve been wondering if the gap between electoral and administrative aptitude is actually widening. If so, it’s not good news.

Electioneering involves skills or traits that will come in useful once in power. Coalition-building, making and winning arguments, putting forward plans that can be turned into reality, promises that can be kept; these sorts of things do still play a part in elections, but they feel increasingly ancillary to the main job of fighting an election, which is, of course, waging meme warfare on Facebook through an army of automated sock-puppet accounts.

Maybe we’ve yet to learn that the ability to run disinformation and psy-ops campaigns on Facebook is indeed a reasonable proxy for the ability to administer a council, city or country. But maybe it isn’t? Governments still have a lot of things to do which are not really amenable to deepfaked propaganda, Pepe shitposts and behavioural microtargeting. Hospitals need to be run, tax needs to be collected, roads need to be maintained and so on. These things seem deeply boring compared to the use of deep learning networks to convert placid baby boomers into red-faced white nationalists, but they are and will continue to be part of government for the foreseeable future, and beyond.

Maybe someone deserves a government job because they had a successful run as a far-right misogynist troll on Reddit a few years ago, but maybe they don’t? Maybe the kinds of traits that work in the world of far-right online propaganda won’t work in the field of actual government? My hunch is that they won’t, and the usefulness of election campaigns as a proxy for administrative competence is in a state of decline. At least I don’t have to wait long, or go to any other countries, to find out if my hunch is correct or not.


No comments so far.  Post a comment

RSS feed for comments on this post. TrackBack URL

Leave a comment