My Life as a TaskRabbit

That's their all-time record. The odds of real numbers resembling it are low.

First, an unemployment rate of 0% is bad. This is because not all unemployment is bad.

Frictional unemployment is actually a good thing (to a certain extent), because the alternative is having people stuck doing jobs they don't want when better jobs are available.

Structural unemployment is what happens when $old_industry dies because $disruptive_tech_company changes the market. It's definitely bad for the unemployed (who will never be re-employed in the same field), but it's usually the result of positive improvements to society at large. The 'solution' is that those people losing their jobs need to be retrained to work in another field, which is admittedly difficult, or retire.

Cyclical unemployment is undeniably bad. That's the result of boom-and-bust economies - people over-hiring during bubbles, and then having to lay off en masse when the market contracts. (The problem is largely that the people are being dumped into the jobs marketplace all at once, creating a glut, instead of being more evenly distributed).

The rest - someone who's employed but at a salary lower than what they'd like, or someone who retires because they're laid off... that's not unemployment. Someone may dislike it and they may think it's a 'problem' that needs to be 'fixed', but they need to realize that they're talking about something completely different from unemployment.

As for real unemployment, the ideal value is a point of debate amongst economists because it is tricky to measure, but somewhere in the ballpark of 5% is a commonly accepted value for the target 'natural rate' of unemployment.

That is, a sustained, unwavering value of 5% unemployment is actually the target for a healthy economy.

1. voluntary - a person is capable of doing a job, that job is available and fits into that person's life, but (s)he chooses to work at a less demanding job not utilizing his/her skills. I don't see this as a problem, personal choice is OK.

2. involuntary - this is a case where a person or workforce has capacity to produce more, get higher wages, or otherwise be "better" employed, but market inefficencies[1] prevent the best employment. Again, this may not be a problem for the individuals - they may not be bothered by this, but it is a valuable measure for opportunity. If a company can employ these people for $fewer than the going rate, but $more than they are making and gain profit, they should, because it helps overall market efficiency - it helps lower the cost of employees overall, and helps raise the wages of the people who provide value. Everyone wins. Just like finding new sources of any raw material in demand, or using previously "too expensive" sources because the market price has shifted.

For a real example: company A had a systems office in my town employing some pretty good graybeards. A decided that they didn't want/need to be in that business anymore, and closed the office. For the large part, the graybeards were happy in town, and chose to stay and not work at as well paying jobs, for reasons outside of salary. The thing is - systems graybeards are in demand and expensive (usually). For a while they were underemployed, not making the money they should. But company B heard of this, and realized that they could pay cost-of-living adjusted salaries that saved them a lot of money on a pool of systems people with a proven track record. They opened an office hiring most of the underemployed graybeards and everything was in balance again.

Of course that example is almost the text-book ideal case of how efficiencies, indicators, and such should play out. There are harder cases of this. Finding ways to take advantage of the underemployed metric is more difficult.

Anyway -- that is the long way around of saying: underemployment may not be a good or even valid measure at a personal level, but at an aggregate level it is a good measure of where efficiency can be sought. Disbelieving it speaks more of personal/political biases than of a market understanding - probably because it is term associated often with jobs programs and other things many dislike.

[1] Like algorithms, just because a perfect theoretical algorithm has an efficiency of $X, a given implementation my only be able to achieve $Y efficiency (where $X > $Y).