Hacker News Reader

Comment by rcthompson | Hacker News Reader

This reminds me of the anecdote where some city tried to deal with their rat problem by offering a reward for every dead rat turned in. Instead, people started breeding rats so they could have more to turn in for reward money, and the rat population actually increased.

Or something like that. I might be mis-remembering some details.

View on HN

There is actually a term for this in economics. It's called the "Lucas critique", and effectively states that any metric will be gamed eventually. It's slightly more subtle and is actually a critique about macroeconomic policymaking -- that people's actions are not structural and that making a new policy will affect how people act as they try to (ab)use the new policy -- but that's the basic gist.

EDIT: To clarify -- I'm not an economist, it's just something I read about in my spare time. I'm sure that the Lucas critique was pre-dated by other ideas (it's far too obvious to be entirely novel), and that there's a much deeper analysis that I wasn't aware of.

Lucas critique is predated by Campbell's law and Goodhart's law.

IMO the deeper history in Lucas Critique is the Problem of Induction.

"XYZ has always worked in the past" doesn't work. Imagine doing machine learning on this problem that Bird has. It could easily suggest that you should up the fees to get more scooter finders, rather than that you need to change the incentives.

I believe this is related to another principle in Economics called Goodhart's Law, which essentially states that as soon as something becomes a metric for measuring performance, it ceases to be a valid measure of actual performance. Essentially using something as a performance metric creates the incentive to game the system, inflate their values, and generally divorce the metric from the phenomenon it's trying to measure.

This was described in "The Wire" as juking the stats: https://youtu.be/_ogxZxu6cjM

https://en.wikipedia.org/wiki/Cobra_effect

Yep, that's the example I was thinking of!

A nice, classic variation of this theme, featuring day-care centers: https://www.nytimes.com/2005/05/15/books/chapters/freakonomi...

Not sure if it's analogous. The day care scenario replaced the guilt of leaving your child at day care too long with a fine that removed that guilt.

Clearly, you can fix the daycare scenario by 1) communicating to the parents picking up their children late the pain that causes the workers; 2) Ratcheting up the fine so that it is clearly painful to parents; 3) Ban parents who still flaunt the system.

My dog's daycare has this system! $35 for a full day of daycare (7am - 9pm). $15 for late pickup (9pm - 11pm).

with a fine that removed that guilt

Rather tha guilt I think the factor in play can be described as cheap overtime daycare. It's actually more valuable to pay $3 and get more flexibility in your day rather save $3 and be on time.

Need to see the dentist after work but you'd have to first pick your kid and have him sit in the waiting room while you're in? No problem, pay $3, let him play at the daycare and go to see your dentist first, then pick him up.

Isn’t this just a mispriced externality? I.e. if the late pickup fee was actually charged at market rates it would be what, $500? (random guess obviously)

My child’s daycare prices late pickup at $20 for pickup after time and an additional $10 a minute for minutes 5-10.

I don’t believe they often use this - maybe only against chronically late parents - but it is in the contract.

Mhm, isn't that one of the recommendations for freelancing when you get an offer that you don't really want to do and you don't want to say 'no' to their face?

Give them a price quote so ludicrously high they either walk away balking or take you up on it, which means that while you're doing the job you didn't want you also get a ton of money and some data on how much your services are valued.

In this case, apply to the daycarer actually staying late.

There are several examples of cash rewards for rat hunting like in Jakarta [1], but the "classic" version of where this went wrong is the "The Great Hanoi Rat Massacre" of 1902 [2] where rat hunters would just chop off live rats' tails to score the reward instead of killing it, because it would diminish their future income.

[1] - https://www.theguardian.com/cities/2016/nov/21/jakartas-rat-...

[2] - https://www.atlasobscura.com/articles/hanoi-rat-massacre-190...

This happened in Seattle but with human corpses, back in the early days. It worked out exactly how you’d expect.

People hoarded corpses. Weird.

And started making new ones when the supply dried up.

https://en.wikipedia.org/wiki/Cobra_effect

Reply on news.ycombinator.com