Forecasting and competition policy

Top B2B Technology Companies for AI in Sales ForecastingValues are observed in actions and choices, and rather less so in words. Competition policy has been applied with great relish to the labour market – at least at the bottom end. (Subject to our relatively generous basic and award wage arrangements). So restrictive practices of warfies and workers in manufacturing plants have been ended, and where they haven’t compromised worker safety or other valuable things, I expect this is a very good thing.

But back the top end of town, rigid demarcations remain, between doctors and nurses, baristas and solicitors (ok – barristers and solicitors, but you get my meaning). You might think this is because of the raw power of powerful professions in our society. And you’d be right. 1 But the thing is that economists’ and policy makers’ imagination seems to align with the power. The desirability of reforming professional licensing was always in the Chicago School catechism. 2

But in the case of Australian opinion leaders, not so much. I’ve been in numerous ‘How Good is Australia? Not good enough without more Reform. Why can’t we get back to the glory days of Reform?’ conferences of the great and the good when it’s not been mentioned. I brought it up at such a session run by the Victorian Competition and Efficiency Commission and managed to get it onto the list, but not with any great enthusiasm from anyone.

And as many Troppo-didly-dillians will know, a few years ago the Abbott, though it may have been the Turnbull or even Gillard Governments, convened the Harper Review into competition policy which at some stage got it into its head that we should open up social services. Admittedly they did so in a manner that led Gary Sturgess to warn them at their conference “stop it or you’ll go blind”. No-one on the panel knew much about the human services they were arguing should be opened up and it occurs to me that if the draft report had had one of those mascots that are now de rigueur at the Olympics, it would be called “Thread Bear”.  The final report purported to a little more knowledge, but not very convincingly. This then fed into a PC report which recommend five areas that might be opened up.

Be that as it may, it occurred to me that here’s another area ripe for opening up. Nick Kamper and I have previously argued for the opening up of government forecasting models – their release as open source models. But forecasting is exactly the kind of thing that should be opened up to all comers.

Currently our situation is not so unlike Philip Tetlock’s picture of forecasting in our society:

Accuracy is seldom determined after the fact and is almost never done with sufficient regularity and rigor that conclusions can be drawn. The reason? Mostly it’s a demand-side problem: The consumers of forecasting – governments, and the public – don’t demand evidence of accuracy. So there is no measurement. Which means no revision. And without revision, there can be no improvement. Imagine a world in which people love to run, but they have no idea how fast the average person runs, or how fast the best could run, because runners have never agreed to basic ground rules – stay on the track, begin the race when the gun is fired, end it after a specified distance – and there are no independent race officials and timekeepers measuring result, How likely is it that running times are improving in this world? 3

Indeed, in doing so government could move towards its classical role which is the ring master, designing forecasting tournaments – with forecasts in probabilistic form – which would actually generate retrospectively accountable information about the quality of forecasting and in so doing both rescue forecasting from suspected political bias as well as start identifying whose forecasts are best and so create the incentives for them to improve.

  1. To be as generous as possible, let’s say “largely right”. One can concede that restrictive practices in law and medicine do have some consumer protection and integrity functions, though these could be easily dealt with if we were serious about achieving anything here.
  2. Milton Friedman cut his academic teeth with extensive research into occupations and occupational licensing (under the tutelage of Simon Kuznets if I’m not mangling what I read a while back).
  3.  Tetlock, P.E. and Gardner, D., 2016. Superforecasting: The art and science of prediction. Random House, p. 14.
This entry was posted in Competitions, Cultural Critique, Economics and public policy, Innovation, Politics - national. Bookmark the permalink.

3 Responses to Forecasting and competition policy

  1. paul frijters says:

    forecasting quickly becomes a closed shop, true. opening up to people like your kaggle mates would certainly improve forecasts and reduce cost. But usually the market is not for great accuracy but for vaguely right plus a story. The thing about the top forecasting techniques is their lack of story, though I guess one could make those up anyway and separate those elements. A related problem is that much of the data gathering is set up for very specific stories: to find something else with them would be very discomforting for the whole chain.
    The big one though is that the market for accuracy is virtually absent in government. Only in some places accuracy would shine, like financial markets (which indeed use whatever works), niggly tech stuff (translation, indeed using whatever works), but not really the gov program areas or the econ areas. to make them a race would require one to take the measurement away from the bureaucracy. if you do that they lose interest.
    so you must consider the whole chain. what would a different chain look like and where could one trial a new one?

  2. Moz of Yarramulla says:

    We’ve also opened up the taxi market quite dramatically, so perhaps we could open up forecasting the same way… have a bunch of tech companies offer predictions sourced from random app users, on the proviso that they not allow you to select the provider of a specific prediction… but you can rate the provider once you get your prediction, albeit only on politeness and promptness, not accuracy (so no change there :).
    It would be interesting to see an article on prediction in financial markets, since Paul seems to think they have accurate predictions available (despite the ample evidence to the contrary, unless we’re talking the equivalent of same-day weather forecasts “it will keep raining (unless it stops)” and “the market will keep rising (until it doesn’t)”). I’m thinking of those graphs Jericho is is so fond of, showing treasury predictions of growth against real growth, or the wage growth guesses in this article.
    https://www.theguardian.com/australia-news/2019/apr/02/federal-budget-2019-the-seven-graphs-that-expose-the-coalitions-2019-budget-fairy-tale

  3. conrad says:

    You are partially wrong about doctors and nurses (I see it as in-between). The distinctions are less rigid than they used to be as many of the allied health professional jobs would have once been done by doctors (cardio, clinical neuroscience, respiratory, sleep, etc. . ). We also have nurse-practioners who can perform some of the functions of doctors, lower level people who are not nurses but do some of what they used to do, and clinical psychologists (who they shouldn’t never let prescribe drugs IMHO, but want to) cut into the market for psychiatrists.

Leave a Reply

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.