Research Finds Outright Grants of Cash are an Effective Form of Aid to the Poor

You're not accounting for those that have given up looking for jobs. Additionally, the report mentions that children are also included in the 46.2 million people below the poverty level. Also, if you know anything about the workings of Temporary Assistance for Need Familiies (TANF), the main federal block grant for welfare as defined under the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA), then you know that 1) welfare recipients must work for 24 months to continue receiving welfare, and 2) that there is a 5-year lifetime limit on being in welfare [1]. So, though your report mentions that most people below the federal poverty line, there's plenty of evidence to assume 1) those who aren't working are also not on welfare, with the possibility that 2) those who aren't looking for work have simply given up, as is the case for many, many Americans post-2008 recession.

> Your source [2] also disagrees with you.

Clarification, if anything the source disagrees with itself since the entirety of my use of it was within quotation marks. That said, I don't think it does disagree with itself. 1) We have no idea what the welfare/person spending was, as you note, so it's "silly" to assume that there were obvious net savings when we simply don't have the data. If you're not willing to trust the source on that, then at least trust the numbers for Arizona and Utah. Secondly, you say that "the actual number in Florida" was close to 2000, that 2000 people withdrew their applications when reaching the drug testing stage. You're obviously implying that they withdrew because they were on drugs and wary of not passing the test. You fail to mention the financial aspect to the drug tests--the welfare applicants themselves had to pay for the drug test according to the source you link to [2]. It was highly plausible, as noted in the article, that the ones who didn't take the test couldn't net that amount of cash.

As to Florida's program itself, it was deemed unconstitutional and stopped since it "unconstitutionally mandated searches without suspicion."[2] This was for the same reason you note that the Arizona program was "less effective"--namely because it didn't assume that just because a person was poor, they were using drugs.

[1] http://www.gao.gov/new.items/d0258.pdf

[2] http://usatoday30.usatoday.com/news/opinion/editorials/story...

Actually, this is the biggest reason for the build out of the Aadhaar project or the Indian biometric ID project. It is less for reasons of trying to track the common citizen (although I dont doubt for a moment, it will come to be used like that), but to ensure credible, direct debit of cash to the poor.

I certainly am accounting for those who've given up. They are "not looking for work", i.e. not unemployed. Accounting for children does change the fact that the poor choose not to work https://news.ycombinator.com/item?id=2130441 .

The fraction of the poor who choose not to work was low before the recession as well. http://www.bls.gov/cps/cpswp2006.pdf

As for your assertions that paying $30 for a drug test prevented people from applying for welfare, that doesn't pass the smell test. Anyone who passed was reimbursed, and welfare pays well over $30 anyway. If you want to hang your hat on the fact that I haven't found a source proving that welfare costs more than $425/person, be my guest.

Would the cost of doing that exceed the savings? If so (and that seems rather likely; the vast majority of social welfare recipients do _not_ cheat), it seems rather impractical.

In practice, people cheat in all systems, and a certain amount of cheating is built into the assumptions for any well-designed system. For instance, tax authorities do not audit everybody, because it would be impractical; in not doing so they accept a certain amount of tax fraud. Shops don't watch everyone who comes in the door all the time, because it would be impractical; in not doing so they accept a certain amount of theft (most of what accountants call 'shrinkage' is employee and customer theft).

Yes, I do think it's an obvious logical implication of the above that if you want to donate to where it will have the largest impact on welfare, you should be donating to Africa instead of the US. As a result, I live in the US and do donate to effective aid organizations that aim to help people in extreme poverty, who are usually in Africa. I don't think there's anything very obviously wrong with this.

I wrote a bit more about my approach to charity here: http://blog.printf.net/articles/2012/11/27/celebrating-seven...