Cars for po' folks

While they're not generally my type of folks, I have over the course of my many years become friends with some pretty right-wing people. Of these, a few (maybe 5-10) are people that I am truly happy to call my friends, the rest are folks who I ended up meeting or working with and occasionally keep in touch with for whatever reason.

So, a post from one of them (not one of the close friends) got me thinking - he was referencing this article (which in turn references this article) and he had this to say about it:

Disgusting. The state ought to have no part of this - I have no problem with private charities and enterprises taking up the cause of providing for those who are lacking the resources to get by, but tax dollars that are automatically deducted by payroll cannot and should not be used for this purpose. I'm having a hell of a time paying for the car I have already and saving for a minivan, I'm not about to go buy the poor people in my neighborhood cars!!!Which, sadly, is so often the conservative approach to things.

So, my response to his post (which, due to a fat-finger, I didn't send):

From TFA: "Kehoe defended the program, saying the state breaks even by cutting welfare payments to the family - about $6,000 a year."
So I guess it evens out, although since there's apparantly a 20% recidivism rate (for lack of a better term), it still ends up being about $1200/yr/car, and at the rate of 65 cars/year, that's still a net cost of $78k/year for the program.
Also from TFA: "To get the cars, they must be unable to reach work by public transportation and have a clean driving record. The program is only available to families on welfare with children."
You know, this always bothers me; people come up with ideas like this, and in general I think they're a good idea, but could we PLEASE try to come up with a control group?!? I mean, I'm guessing there are a lot more families that qualify for the program than actually receive a car, so can we try to get some stats on families that qualified but for whatever reason (didn't ask for a car, a car wasn't available, etc) didn't get a car. Take the differences in control group vs. families who got cars, and if the cars statistically cost less than the amount of welfare payments you're _not_ giving to the families who receive them, then it's a good deal. If not, then it's a expense that needs to be seriously examined. Why can't we ever get data on these things?!?

I continued on with some other stuff (before I accidentally erased it), but you get the gist. I often wonder about government programs; I mean, it's impossible to account for all the ramifications of these things when you're talking about tens or hundreds of thousands of people, but at least small program like this, you could try! 65 cars per year, the program's been running since 2006, so let's call it 200 cars so far to make the math easier. Spend 10 minutes following up on each family that received a car, you might come up with something like this:

Out of the 200 families, 40 ended up losing the job within the first year but keeping the car. Out of those, 2 ended up re-employed and are no longer receiving welfare assistance. Out of the remaining 38, we're pretty sure that 15 of them intentionally gamed the program just to get the car. The rest, we don't know.
Out of the remaining 160 families, another 10 lost the job after the 1-year period was up and are back to full-time welfare assistance.
Out of the remaining 150, 100 families have been completely without state assistance since receiving the car. Out of those, 5 have risen to a family income of more than $60,000/year (or whatever middle-class benchmark you want to use), etc. etc.

To get most of that data would be a simple matter of looking up the case files of the folks who received cars - for the rest, you would have to talk to 200 families or case workers to catch the really interesting details (like, for example, that one guy who started his own business using the car and is now a successful entrepreneur, or those 10 families who realize that they were actually better off before they made the deal to get the car, and would gladly go back in time and reverse it if they could). Call it an average of 10 minutes per family / case worker, and you're talking 2000 minutes, or about 30 man-hours to get a decent picture of every single family who participated in the program - throw in another 40 hours (and I'm being pretty generous here - it's probably a lot fewer hours) to run data analysis and write a report on it, and you can have a detailed picture of this government program, what works, what doesn't work, and how it can be tuned to improve. Or, if the data is not at all encouraging, then maybe the program should be scrapped.

I've kinda lost momentum with this post due to being forced to work (gasp!) many times during the writing of it, so I'll stop here. I guess if I had to make a conclusion, it would be this: I expect people to be reasonably good at what they do, particularly when there's theoretically some sort of hurdle or test (like an election) that they need to pass in order to get the job. I would expect that folks who come up with programs like these would have a process and resources in place to determine the program's efficacy, hopefully accounting for some of the factors that can't easily be expressed quantitatively. Maybe everything I've described is already being done, but I'm guessing not...