Unemployment Insurance

The best anti-poverty governmental program is a guaranteed minimum income. But it’s not sufficient as a welfare system, because cutting the income of an engineer who was laid off from $75,000 per year to $10,000 will make him likely to scramble for any available job, even one that’s vastly beneath his skills. It’s in society’s interest that he’ll find work in his field, where he’ll be more productive than in an unskilled position.

The number one abuse of unemployment insurance systems is staying unemployed voluntarily until benefits run out. Anecdotally, I know of two professionals in Israel who were laid off, both of whom did not look for work until six months later, the time limit on Israeli workers’ comp. The obvious solution to that is to require the beneficiary to show that he’s looking for work, subject to certain constraints like location and skill level.

Obviously, different countries do it in different ways. Germany’s Hartz IV reforms require the unemployed to take any available job. In a system with guaranteed minimum income, this is not necessary, since the primary purpose of unemployment insurance in this scheme is to prevent deskilling. As such, for professionals the rule should be either that they need to show evidence of looking for work in their fields, or that the state will try matching them to available positions, which they’ll then have to take.

Since professionals who make large amounts of money can be expected to have savings, it makes sense to cap the benefits at a fairly low level by their standards. For instance, benefits might be 75% of last weekly income, capped at something like $800 a week, which corresponds to $55,000/year. In addition, to make sure people actually save, it makes sense to incentivize saving by exempting the first 10% or $10,000/year saved, whichever is lower, from income taxes. Benefits should be time-limited to one year or so, after which they should wane over a few months to the level of guaranteed minimum income. If it takes that long for a person to find a job in his field, it’s probably a lost cause and he should retrain.

Note that while guaranteed minimum income is best distributed based on household size, unemployment insurance should be strictly individual. This is because it’s intended not as a way of reducing poverty, but as a way of propping some people’s lifestyle for a short while to help them find jobs that best suit their skills.

Assuming the test for searching for work is rigid enough for there to be no voluntary unemployment due to workers’ comp, the cost can be computed based on average incomes above $31,000/year, at which level 75% of income is higher than the guaranteed minimum income level.

To simplify matters, let’s overstate the cost by saying that 5% of the >$31,000 population is unemployed and that there are 2 income earners per household so that the effective cap is $80,000/year. Since in the US, 65% of the population makes at least $31,000/year, the cost of the entire scheme will be around $800 per capita, as the bulk of the middle class won’t get anything approaching $80,000. Note that this assumes my doubly tweaked minimum income scheme, which already provides a base of $23,000 to the average household.

Apparently, decent social security costs around 4.5% of GDP in the US, so now you know that having a functional welfare system requires spending 11% of GDP, up from around 6% now. The existing poverty reduction programs, like food stamps and TANF, should be subsumed under guaranteed minimum income, so their cost is replaced with the admittedly higher cost of minimum income.