Automation and Top Income Inequality

For almost 40 years, inequality within the top percentile of the income distribution, measured as the ratio of income share of top 0:1% to the income share of top 1%, has been increasing in the US. The income of super-rich people increased more than the income of rich people. In this paper, we show that improvements in automation technology (the number of tasks for which capital can be used) is an important factor contributing to this inequality. We consider a model in which labor has a convex cost and capital has a linear cost. This leads to a decreasing returns to scale profit function for entrepreneurs. As capital replaces labor in more and more tasks, the severity of diseconomies of scale diminishes, hence the market share of top-skilled entrepreneurs increases. If entrepreneurial skill is distributed according to a Pareto distribution, then top income distribution can be approximated by a Pareto distribution. We show that the shape parameter of this distribution is inversely related to the level of automation. Finally, we rationalize convex cost of labor using the theory of efficiency wage.

That is from a new paper by Omer Faruk Koru, via the excellent Kevin Lewis.


Economies are zero scale.

If robots replace most workers, I capital is built by workers to great abjndance, the return to cappital must fall to converge on interest paid on savings, and logically savings must be in great abundance, and thus approaching zero.

Thus, production has very low capital costs and very little labor costs, so prices must decline toward zero. GDP must thus fall as well, and there is no labor income to pay to buy GDP.

If you assume prices remain high, then either wages soar to maintain worker incomes, and consumption, as hours fall. Or capitalist end up with so much income, far beyond their ability to consume GDP, except by saving and paying workers to build even more robots.

Or loan, or give, their money to workers or the unemployed to pay to buy GDP.

Its simply not possible to pay workers only $50 and then charge workers $100 to buy what workers produced. The $50 taken by the capitalist that is not paid to workers to build new capital, must somehow be used to pay for GDP.

Zero sum.

Savings/capital are merely stored, or time shifted labor, as is debt. Savings/capital are labor in the past. Debt is labor in the future, but normally secured by capital, which is labor in he past.

Ie, a car loan is a promise of labor in the future secured by the car which is labor in the past.

Tanstaafl. Zero sum.

Until the robots replacing humans become consumers. The Amazzon Prime Electric Dreams episode 2, "Autofac", is a story of robots replacing humans after humans try to shutdown the robots. Once the humans are gone, the autofac has no purpose, there are no humans to serve. So, the autofac creates robots that think they are humans to consume what the autofac produces.

"Economies are zero scale."

That might just be the dumbest thing I've seen you write.

'is an important factor contributing to this inequality'

One assumes that Marx is properly credited for insights like that.

If acknowledging Marx is deemed too heretical at least Piketty can be acknowledged for having presented the same argument

"inequality...measured as the ratio of income share of top 0:1% to the income share of top 1%"

This is not a real thing. The fact that this is the inequality that left-leaning, educated, affluent elites actually care about is the reason why left-leaning elites are so alienated from mainstream America.
Emotionally healthy people do not obsess over Ivy League inequality.

I thought that the researchers were looking at the .1% to the 1% as an analytic tool to distinguish between lawyers, CEOs, surgeons, and bankers versus people who own businesses. All in service of an attempt to understand the structure of the economy and how it works, and to be able to show that the mathematical model is consistent with a real world data set.

And there results arguably do show something important for the “99%”, which is that income is increasingly going to business owners because production is more capital intensive. Which to me sounds like the perfect situation for collective bargaining/ collective action. Amazon wouldn’t be much without its servers, so striking workers could figure out a way to shut them down as a point of leverage, just to give one possibility.

Nah it's just the stock market, and a fractal of the general increase in (within country) inequality.

the computer helps those with ability much more than those with lesser ability. So wealth will grow faster. Unhappiness with income distribution is a political economy questions. Since no one knows when the Ayn Rand effect lowers economy growth, creators of wealth get to keep lots!

Wealth is increasingly attributable to rising asset prices, which isn't surprising because there's more capital (the savings glut as it's sometimes called) facing diminishing returns to productive capital and owners of capital bidding up asset prices in the search for a higher return. When companies go public at prices far in excess of what their earnings support, enriching further the early stage (already wealthy) investors along with the bankers and others who service the owners of capital, are we witnessing capitalism at its creative best or just another bubble that will burst absent help from a Fed that fears the collapse of asset prices worse than it fears the bubble?

Gone are the days when owners of capital faced two choices, investing in productive capital (plant and machinery) or (conspicuous) consumption, the former subject to the laws of economics including overzealous investors in productive capital driving up wages and producing too many goods and driving down prices and profits. Today, owners of capital have greater choices, with global financial markets and complex financial instruments, augmented by an abundance of capital seeking higher returns from and promoting rising asset prices (in real estate as well as financial assets). Even Chinese billionaires who made their fortunes from productive capital are investing less and less in productive capital due to falling returns (entire new industrial cities in China are virtual ghost towns) and investing instead in real estate and financial assets in search of higher returns from rising asset prices (or at least stable asset prices thanks in part to cooperative central banks).

Can't asset prices continue to rise with the growth of the new information economy? After all, Google and Facebook generate enormous profits. Google and Facebook generate most of their earnings from advertising, and here's a surprising data point: the percentage of the total economy attributable to advertising is actually low by historical standards, it has just shifted to digital media (i.e., Google and Facebook). What about investment in productive capital attributable to new technology, such as self-driving cars? New technology, artificial intelligence (AI), will replace much of the labor that historically has gone into the production of vehicles (and other goods) and, hence, the capital share will rise while the labor share will fall. What about the growth in the global economy, as formerly very poor places prosper from economic development and the indigenous populations become consumers of goods produced there and elsewhere? That's the promise of globalization and trade, but the reality is political (and, thus, economic) destabilization and opposition to globalization and trade.

Am I wrong to have concerns about reliance on rising asset prices for prosperity? In his book, Thomas Piketty stated that central banks and governments will always intervene to prevent asset prices from collapsing and to support rising asset prices. Sure, Piketty could be right, but if he is wrong, if the political tide turns against a rising level of inequality and intervention to maintain it, the consequences could be disastrous. On the other hand, investment in surveillance technology may the the biggest new thing, making billionaires out of the pioneers in surveillance, including Peter Thiel (even though he already is a billionaire). I wouldn't bet against Thiel. If Palantir does go public this year, one would be well-advised to follow the pied piper even if it does contribute to the surveillance state.

Aren’t the industrialization of underdeveloped countries and displacement of labor by AI and robots two things that are mutually exclusive?

Or, the lowering cost of automation could make it available to smaller businesses and entrepreneurs, helping to level the playing field.

Or, automation could kick off a new economy of bespoke items made by boutique manufacturers. Much like 3d printing is eliminating economies of scale and enabling small shops to compete with giant manufacturers.

The priors in a study like this reveal more about the biases of the authors than a realistic view of the future. We are now in an era where automation coupled with 3d printing can empower an individual to be able to build things that would have taken millions of dollars of manufacturing hardware and a team of specialists and workers just a few years ago.

The story is far more complex than labor vs capital, and in an era when you can build complex technical products in a basement with a printer, where do you draw that line? I can point you to hundreds of businesses already in existence that use 3d printers to make custom components for industry and the public, and are doing so with a fraction of the capital required to do this kind of thing in even the recent past. That's democratizing and liberating, but you aren't seeing a lot of papers studying that effect - perhaps because such papers would not support the desired conclusion that capitalism is bad and things are getting worse for the common person.

It is true that 3d printers are cheaper to do some things than traditional shops but it is still not cheap enough and would involve a large business loan for the middle class. Given how risk averse our banking is when it comes to lending outside of housing, there is still more work to do to properly democratize manufacturing. Overall, I agree that this is a good trend.

You can buy a PLA printer for $300. You can assemble one from a kit for $150. You can get Autodesk Fusion 360 for free, and do 3d modeling with it. You can set up a shop on Etsy for zero dollars. If you need it printed in steel or titanium or any number of materials you can ship your 3d file to a service and receive the part in a day or two for not much more than the cost of the material.

There is a huge industry out there made up of hobbyists who make and sell small quantity items to each other. 3d printers themselves come in kits made up of many 3d printed parts.

The progression of technology has enabled small, low budget businesses to create viable products and compete in the market. It has allowed small companies like SpaceX to out-compete the large established aerospace firms. SpaceX's SuperDraco rocket engines are 3D printed, for example.

Technology has enabled the Gig economy, and global auction sites like eBay have allowed ordinary people to find value in things that otherwise would have been hard to sell.

And yet, everyone seems to assume that technology is going to empower the rich and freeze out everyone else.

Comments for this post are closed