In my last blog post, I discussed the rise of automation. The basic problem is that, as artificial intelligence systems and robots become more advanced, the value of human labor falls. Unskilled labor–particularly manual labor–may eventually become almost worthless. It’s easy to say that the invisible hand will magically create jobs for the newly unemployed, but to me, this argument seems to be less based on reason, and more based on a religious faith in capitalism. Thus, we need to examine other solutions to the inequalities created by automation.
We’ve known for decades that there is a correlation between employment and IQ. For instance, one 1994 study found 14% of people with an IQ less than 74 had been unemployed for at least a month in the preceding year, versus only 4% of those with an IQ greater than 126. It seems likely to me, by eliminating unskilled jobs, automation should exacerbate this relationship because many people who lose their job to automation will be incapable of retraining to do more skilled jobs.
Even if you are skilled enough to keep your job, this trend is a big problem. The more income disparity increases, the greater the instability in society. If we end up with, say, a third of the population without any reasonable prospects for employment and the rest of the population with the highest standards of living of all time, then it’s likely to result in conflict.
It’s ironic that, while a significant portion of the population will be struggling to achieve subsistence standard of living, they will be doing so at a time when human productivity will be at an all-time high, when a small fraction of the population can produce enough to address the needs of all.
And that suggests one possible solution, the guaranteed minimum income (GMI).
Will there also be chicks for free?
The guaranteed minimum income is exactly what it sounds like–money for nothing. Every month, every adult, regardless of whether or not they are working, gets a cheque from the government. Those who make more than the monthly minimum income would have some of the amount taxed back.
The government would levy income or consumption-based sales taxes to pay for guaranteed income. In effect, in a world where people aren’t required to produce a multitude of goods, the government will be ensuring that the goods are more widely distributed than they otherwise would be.
Isn’t that communism?
The GMI model is significantly more free market-based than communism. Communism typically has a centrally-planned economy–the government deciding which goods get produced in what quantities. In this model, consumer demand will still determine what gets produced, since consumers will decide what they want to purchase with their income.
Second, communism, by definition, has communal ownership of wealth. In this model, there is still private ownership of wealth, but there would be fairly high taxes required to support the guaranteed minimum income, redistributing a significant portion of the wealth. So, GMI isn’t communism, but is certainly more socialist than what we have today.
One of the problems with communism is that, because individuals don’t own the product of their labor, there is little incentive to work hard. Why work if you don’t get anything as a result of your effort? And that is also a potential problem with the guaranteed minimum income. If people don’t have to work to survive, will they?
One possible solution to this problem is to be smart about the tax rate on incremental income above the guaranteed income. If the income tax rate ramps up slowly, then the incentive to work is still relatively high. For instance, perhaps there is no income tax on any income up to $10,000 above the guaranteed minimum income, and the tax rate on the next $10,000 is only 5%.
This strategy will result in a highly progressive tax system, where the top earners will pay a huge percentage of the taxes, but that’s already happening today anyway. There is often discussion about how this will cause the huge earners to not bother trying to become a huge earner. But I believe most people making stupidly large amounts of money are working mostly for reasons other than money (e.g. power, social validation, curiosity, fun, desire to make the world a better place, community). So, while high income earners will whine, it seems unlikely to me that most of them would slack off because of taxes.
Another option is to make the tax system to support the GMI mostly through taxes on consumption so that there are no income tax-based reasons to stop trying. However, this method is perhaps sub-optimal as it would be highly regressive, increasing inequality, the opposite of the goal of GMI.
Would these methods be sufficient to ensure that everyone wants to work? It’s doubtful, but in this highly-automated economy, you don’t need everyone to work. You just need enough to keep the economy chugging.
And now for a contradiction
Another interesting way to implement the GMI might be to make payouts revenue-based. In other words, the government will provide a guaranteed minimum income, but the minimum will vary on a monthly basis depending on the tax revenue and the number of people being supported. You could do this by putting all the GMI tax revenue into a bucket that is divided evenly among all eligible adults. This approach might address the incentive problem–when production falls, the payouts fall as well, and the incentive to work increases.
For instance, suppose that you have a population of 10 people. Eight of them are making $20,000 and have their income taxed at 50%, while two do nothing. Then the tax revenue bucket would be $80,000, divided among 10 people, or $8,000 per person. So the workers make $10,000 from their jobs, and $8,000 from the GMI, or $18,000. The non-workers make just $8,000, just under half of what workers make.
But then suppose three of the workers decide they’d rather not work and just live off GMI. Then, since we only have five workers, the amount of money in the bucket falls to $50,000. The GMI payments are still divided among 10 people, so fall to $5,000. That increases the incentive for the non-workers to start working, since their income is now much lower, $5000–only a third of workers’ $15,000 income. What’s more, with the reduced supply of labor (only five workers rather than eight), those five workers will be more in demand, and their salaries might increase as a result, further increasing the incentive to work.
I have no idea if this bucket approach is actually practical. But GMI is an interesting solution to problem of excess unemployment as a result of automation and should be seriously considered. It’s not something to rush into, since it seriously messes with incentives in the marketplace and therefore has the risk of unforeseen side-effects. But, with some more thought and exploration, we might be able to discover ways–such as the bucket approach–to reduce the disincentive impact of the GMI while still lowering inequalities. And that would be the best of both worlds.