Let’s start from the standard objection to proposals to tax the rich more to redistribute to the poor. If we tax the rich heavily, they might work less, or emigrate, so we’ll not raise more money. And if we subsidize the poor, we’ll reduce their incentives to work their way out of poverty.
This triggers what is usually a sterile debate between “right” and “left”. The “right” says incentives matter enormously, the “left” downplays them. Yada yada yada.
This is where tagging comes in, as described in an ancient paper (pdf) by George Akerlof.
Let’s say we could identify a group of people who can’t work their way out of poverty - say, because they are disabled or otherwise disadvantaged. We could redistribute quite heavily to these, in the knowledge that we’re not affecting incentives because, arguendo, they can’t get out of poverty anyway.
Equally, if we could identify some people with high ability, we could tax them on their ability, not their income. This would allow us to raise money from high earners without affecting incentives - because high ability people would have to pay the same tax even if they work less.
So, if we could “tag” low-ability or high-ability people, we could tax and redistribute according to these tags, rather than (just) according to income. This would then mitigate the trade-off between equality and efficiency.
The question, then, is: are there such tags, which are well correlated with high or low ability? The important thing here is that the tags be easily identifiable and incentive-compatible. This rules out some possibilities. For example, high IQ is not a good tag. If we taxed folk according to this, smart people would pretend to be stupid, or even deliberately reduce their IQ. Nor is a lack of qualifications a good tag; if we gave a grant according to this, we’d give kids an even greater incentive to bunk off school. For the same reason, obesity is a bad tag, as Homer Simpson showed.
However, this still leaves loads of possibilities. We know that black men do badly in the labour market, so could use this as a tag, and pay men for being black. This is just - it tackles poverty. And it’s efficient. It doesn’t distort incentives, as it would give black men an incentive to earn more, as they’d still get the payment even if they earn decent money.
Equally, there are lots of correlates with ability - and earnings - that cannot be hidden, such as height, looks, gender or being privately educated. All of these are good candidates for tags that could attract higher taxes. After all, if you tax me on my income I might respond by reducing my pre-tax income. But if you tax me on my height, I‘ll not make myself shorter*.
In theory, then, taxing and redistributing according to tags is more efficient than doing so according to income - as it’s more income-compatible.
However, although economists have shown lots of interest in this theory, there is almost no interest in it among politicians. This is despite the fact that they have in recent years moved towards adopting other principles (pdf) of optimum taxation theory.
Why is this? Does it show that there’s a big gap between what’s just and efficient and what’s “politically acceptable” (taxing public schoolboys and subsidising black males would be a hard sell!)? Or are there other flaws with taxing tags?
* I’m ignoring the problem of emigration here, but there are tags that could deal with this, in theory.