The math fail of basic income memes
The “basic income” meme has been making recent rounds in the US. I see a complete failure to do the basic math involved.
Let’s take for example this recent article proposing a $12k per year income for each person in the USA. There are 323M in the USA right now, so the bill for that is $3.9T per year. The federal budget for 2015 was $3.69T. The proposal is already bigger than the current US federal budget, including $439B in deficit for 2015.
In discussions I’ve seen on social media, there’s the “well but you have to subtract social security and welfare out of that because we’ll substitute”. The current average social security recipient receives $16,020 per year, so you’ve just given every Social Security recipient a huge cut in benefits. I doubt such a proposal would pass Congress, let alone be humane or just. Every argument I’ve seen about “exclude this or that group” fails along the same lines, it’s either grossly unjust or we’ll go broke trying to be fair.
Let’s take the original idea, that robots will increase productivity so much that there’ll be nothing left for anyone to do, so we might as well pay them to do nothing so that they can at least live. (The author of the article seems to think sitting around and watching 8 hours of TV a day is more rewarding than being a truck driver, but maybe I watched too much BJ and the Bear to think that. But I digress).
Let’s take Social Security as the metric to base calculations on. We’ll give everyone $16k per year, that way nobody loses benefits. Furthermore, in order to not distort the economy more than Social Security, we’ll need to keep the cost to 5% of GDP, and GDP currently about $17.5T. So we have supply $16k*323M = $5.2T. At 5% of GDP we’d need a GDP of 20* 5.2T = $104T, or 6x bigger than the GDP today. Note I could have done this as GDP per head, the above assumes constant population, but the result should be the same given we’re trading off labor for idle people — 6x GDP increase per capita.
Will robotics lead to a GDP per head of 6x bigger than today? Let’s look at the age of massive industrialization and productivity gains for which we have data for, 1950–2015. In that period productivity per worker in constant dollar terms increased by 2.2% per year (3.9x increase over 65 years).. Since we’re essentially saying that less people have to work because of robots, then the productivity will effectively have to increase 6x per worker in order to fund a non-distorting, fair basic income for everyone. Let’s do a little more math. By the rule of 72, 2.2% means productivity is doubling every 33 years. There are log2(6x) = 2.6x doublings to get to productivity 6x that of today. 2.6*33 = 86 years. Note this is assuming a lot of things about cost and productivity of services versus goods, but is in constant dollar terms and who knows maybe services will be cheap because of robots too. So let’s just stick with constant dollars and hope some key aspect of existence like food also have productivity increases of 6x. (yes, that’s highly optimistic, food productivity has plateaued since the Green Revolution, but there’s always vat grown protein from the scifi books right? Or soylent green. Pick your favorite).
So this means in 86 years we may be able to consider a basic income for all, assuming all goes well with the robotic revolution, and assuming that the results aren’t distributed as well as we’d like, assuming we have invented energy production that’s 6x more efficient than the cheapest today (coal fired plants), oh assuming all sorts of optimism.
86 years is a very long time even for optimists. Will Ray Kurzweil still be alive in 2101?. Let’s not count the chickens before they hatch, shall we?
[EDIT: Did the math wrong for productivity increase 1950–2015. At that rate it’s 86, not 62 years, to get 6x productivity increase]