Means testing in general creates horrible incentives: it means that you get punished for improving your situation, or sometimes even for doing things that seem like they could improve your situation.
The best case is when the "punishment" is proportional to the improvement. Like, when getting each extra $1 of income will reduce the received help by $0.2. The worst case is when people are put into discrete categories, and getting the extra $1 will put you in a different category and thus reduce the help by $100.
Complex rules on the level of accounting are necessary, because in real life the income is mixed with the expenses, for example it may be necessary to spend $50 on travel to get to the job that provides extra $100 income. This imposes extra costs for the sake of legibility, for example if you spend $50 on train, you have train tickets as a proof of your expense, but if a friend offers to take you by his car for $30 instead, you probably have to decline, because you couldn't produce an acceptable proof of expense. The need to do the complex accounting is also a burden for the poor people.
On the other hand, rich people can hire lawyers, accountants, tax advisors, etc., so it can actually be easier for them to generate evidence of their poverty.
I generally agree with this analysis, but there are two areas in particular that keep tripping me up:
Boxfoot
Every means tested program also adds to the implicit marginal tax rate. If benefits phase out as you earn more, it is as if you are being taxed at a higher rate.
Here are some more details from 2009, I'm not sure about the situation now. And it's hard to analyze nationally, because it depends on state and local programs as well.
This is the position taken by left wing economics including Paul Krugman and "atrios" - provide a level of benefit to everyone (what to provide is, of course, a political and values laden judgement) and pay for it with a progressive tax (studies have found that earning incentives aren't damaged for high earners until the marginal rate is over 70%). This minimizes the administrative costs, both for the government, and for the beneficiaries.
Further details can be found in the intertubes.
There are some things that we want everyone to have regardless of ability to pay, but providing them to everyone would be expensive. So we means test. A fancy college may agree to meet 100% of demonstrated financial need, which means they examine your finances and charge you the sticker price or the most they think you can afford, whichever is less. This makes a lot of sense, at least initially, but then people start adjusting their lives in response to incentives.
Take Medicaid and long-term care. Medicaid will cover nursing homes, but only if you have exhausted your savings first. To discourage people from transferring their money to relatives to intentionally become destitute, Medicaid has complex asset transfer rules. Still, they only look back five years, which means that with planning this is still quite practical and many people do it.
As the amount of money involved starts to get large enough, it makes sense for people to plan farther in advance and do stranger things to qualify. Here are two that I think are quite uncommon now, but if current rules persist I expect to see more of:
Intentionally destitute grandparents planning to go on Medicaid adopting their college-ready grandchildren. A few years ago I wrote a silly article on getting married to avoid having parents considered in college financial aid, and people pointed out in the comments that for the schools I was discussing marriage is not sufficient: CSS Profile still cares about parental income and assets. Adoption, however, is different: once you are adopted by your grandparents they are now your only legal parents. If they die, you're an orphan. Looking into adoption rules, it seems to me that adoption by grandparents generally just requires everyone involved to be OK with it.
Affordable housing is typically structured with income limits: you are only eligible if you make below some percentage (30%, 60%, 90%, etc) of the area median income. Today there are not very many affordable units, but if programs like this were expanded as much as proponents advocate then it could make sense for people to intentionally keep incomes low until accepted. There are also some places that let you intentionally opt an existing unit into affordable status, which is especially open to misuse. (more)
If you start trying to figure out how to make a system that is robust to this kind of adaptation, what you end up with is something that looks a lot like the tax code. Which makes me think we should generally just approach this sort of problem with taxes: provide a basic version for free and raise taxes to pay for it. For example, Medicaid could remove means testing and pay a fixed amount towards the long-term care of anyone who needed it. In many cases (ex: food) distributing cash would likely work better. One advantage of structuring things this way is that you can set it up as an income tax instead of a means-tested price, avoiding a lot of the problems around asset transfers.
Comment via: facebook