Showing posts with label Welfare. Show all posts
Showing posts with label Welfare. Show all posts

Thursday, November 3, 2011

Randomised Compassion

The problem with giving to the homeless is twofold.

First, how do you tell which people are truly in bad shape, as opposed to merely being professionally homeless? There are lots of people who are really out of luck. On the other hand, back when I was at university, the same 'bum' worked the same corner for years near school, and didn't look that homeless. I subsequently found out that he actually lived in the same large apartment building that I did.

Second, how do you give money without creating bad incentives? In other words, if you give money to people that beg, you create incentives for more people to beg, as opposed to say apply for a low-wage job. The last thing you want to do is set up a situation where someone makes more money begging than they would at a minimum wage job, or you'll end up with the hard-working being punished on net.

One solution that I like is to to give large-ish amounts of money to people who aren't asking for it.

In other words, find someone who is clearly homeless and away from significant pedestrian traffic, so it's unlikely that they're doing things for the benefit of an audience. For instance, I came across I guy today near a petrol station who was going through the bins looking for aluminium cans.

Now, clearly this guy meets criteria #1. Nobody digs through rubbish bins unless they're clearly down on their luck - it's not an obvious way to win sympathy from people, unlike holding up a sign saying you're a homeless vet.

But even better, giving the guy money isn't going to change his incentives. Because the generosity was essentially random from his point of view, he'll keep doing whatever he was doing before - in this case, working to eke out a small existence recycling cans. If you give money to a guy begging, your generosity is not random - it's a response to him asking you, and thus you're creating incentives for him to beg more.

Now, if the guy is truly homeless, this honestly isn't such a problem - more people get hassled, but he gets a little more money. I can happily call that a wash, or even a gain overall.

The bigger problem is the incentives you create for people who aren't actually homeless - slacker hippy tourists, for instance, who see beggars getting money. Unless you're able to clearly distinguish genuine need from grifters, giving to homeless beggars will create incentives for non-homeless beggars. And those are the people that definitely should be applying for jobs instead. Even worse, being asked for money by obvious moochers tends to make the average person reluctant to give money to anybody, even those actually in need.

In other words, I'd rather give twenty bucks to a guy sleeping on the street than twenty cents to a guy begging.

This is not a problem-free solution, of course. The guy who is truly in real trouble is likely to ask for money, out of desperation if nothing else. And he's the guy you'd really rather not turn down. The problem, as always, is the hippie grifters, mooching off people's sympathy.

In this sense, when I do give to beggars, the question of whether they truly are homeless is dealt with using the strong application of the Ronin principle  - if there is any doubt, there is no doubt.

It's not perfect, but compromises necessarily aren't. Don't blame me, blame the moochers.

Wednesday, September 14, 2011

From the Department of Inconvenient Correlations

The Uber Blog has this hilarious post on prostitution arrest data:

Apparently prostitution arrests in San Francisco go up on Wednesdays. Weird, huh?

Oakland hump day

Apparently, they go up significantly more on the second Wednesday of the month than the first. Even weirder, huh?

Prostitution Wednesdays

What happens on the second, third and fourth Wednesday of each month?
Social Security benefits are paid each month. ...Generally, your benefit will arrive on the second, third or fourth Wednesday of the month, depending on the birthday of the worker on whose records you receive benefits.
Birth date on 1st - 10th, Benefits paid on Second Wednesday
Birth date on 11th - 20th, Benefits paid on Third Wednesday
Birth date on 21st - 31st, Benefits paid on Fourth Wednesday
Oh.. Ooohhhh.

Hmm.

The guys at Uber Blog try to put the best gloss on this that they can.
Keep in mind we’re only talking about 4-5 prostitution crimes each Wednesday. This is pretty low considering the cities we’re talking about have populations in the hundreds of thousands to millions. So before you go running off screaming about how the welfare state is subsidizing sexy times for retirees, chill out and keep that in mind.
Firstly, 4-5 arrests != 4-5 crimes. Not by a long shot.

Secondly, erectile disfunction being what it is, my mind wasn't turning to retirees, but people (some fraudulently) claiming social security disability benefits, which get paid on the same time line.

But it's the obligation of all right-thinking people to test their hunches against data. Thankfully (for this specific task, if not in general), the Megan's Law database lets you easily look up all the sex offenders in San Francscio.

As it turns out, there are a lot more retiree-age sex offenders than I would have guessed. Since few of them seem to have the year of conviction listed, it's hard to tell whether they actually committed their crimes when retiree age (as the authors seem to claim), or it happened years ago. I would guess the latter, but it's hard to know. Maybe it is retirees going to the hookers.

Also, very few of them seemed to be for prostitution offenses specifically. It's amazing how many of them are listed for 'Lewd or Lascivious Acts with a Child Under 14 Years of Age'. Yeesh.

Either way, we'll file this under the category of 'awkward data points for fans of the welfare state'. Which, given the proclivities of the mainstream media, means it gets flushed down the memory hole.