I assume that's referring to the Florida "drug testing welfare recipients" program. That 2% stat rang a bell with me... I remember here in NC they floated a similar idea back in 2011 and now it's come back up again in 2013. The local newspaper (N&O), ran a few stories on the subject, and in 2011 the stats being tossed around felt awfully low to me. So, I googled around, found some interesting results, and ended up in a few discussions about it on the N&O website (via their disqus system). Unfortunately, while my comments are still available in my disqus account's timeline, the specific article itself has been expired so the discussions are longer available. :(
So, to make a short story long, I've gone into my comment timeline and cut-and-paste here some of the relevant comments I made, so I dont have to re-write them all, and so they're here in a much more accessible spot. I've also included a link to the study I based my comments on. I remember finding lots of good info out there, but this one seemed to have the best methodology and transparency on how they derived their conclusions. I'm always suspect of non-objective/biased sources, and this may qualify, but I couldn't find much fault with their reasoning so... "grain of salt" :)
I havent investigated the topic much since then, and I remember the study saying "we'll have to see how some of this pans out", and I havent looked at any of the relevant court cases or decided if I agree/disagree with any of the rulings. I'm happy to look at updated info though.
It looks like the study is still there:
My comments and thoughts about that study from back then:
It indicates that the 2% number is highly misleading, and that actually the state of Florida will likely save a decent chunk of change from their program. It suggests that it very well could save that state several million dollars (the estimate I think was at 9 mil, but I tend to look at such projections pessimistically, so I'd say a more realistic figure is somewhat less than that).
That 2% number was taken only after the first month of the law being in effect, and so it only counts the people who go to the state, fill out all the paperwork, get approved for benefits, and then go pay 30 bucks to take a drug test... all while knowing they've used drugs and will ultimately fail... and then they (surprise) fail the test.
If you count all the people who go through that entire process but then never actually go show up and take the test... the number was significantly higher than "2%".
If you also then go another step further and check out the additional deterrent effect by, say, comparing the dramatic drop in applications year to year and month to month, the figure rises another bit further.
The savings of preventing a drug abuser cash benefits for the average 4 months they are on assistance covers the cost of quite a few people who get reimbursed for passing drug tests. You really apparently dont need that much of a percentage to save money.
The study I read admittedly was from a conservative group, and it's still extremely early to judge the effects of these policies in Florida and make any declarations... but I can't argue much with any of their details or logic. I personally found it pretty objective, but other's mileage may vary :) At least, it quantified many of the questions I had when I see that 2% number and look more deeply at the issues involved.
Also, to the related question of "there is no evidence that people who receive public assistance use drugs at a greater rate than the general population".
Comment: That would be important/relevant if generally catching drug users was your goal. But it's not the case here. The goal is to prevent taxpayer money (in the form of cash benefits), to go towards furthering someone's drug habit instead of helping people in need get back on their feet, Since drug use can often be a contributing factor to the need for cash benefits in the first place, not being a drug user is a reasonable-to-me request of someone getting cash assistance from the taxpayer.
