In the near future, AI will be able to automate any existing method and process of the economy. Since AI can only mimic behaviors that already exist, humans are still needed to create new knowledge, new methods, and new processes. However, not everyone can become researchers, inventors, innovators, and entrepreneurs. Our economic system needs to change so that everyone can benefit from AI and automation. What is this new economic system?
Circulation of Money
Let’s start with a thought experiment on how automation displaces jobs and creates problems for capitalism. Imagine on a remote island there are 10 people. On this island there is one corporation, which is a farm. All 10 people are workers at the farm. Each day, they receive $1 for their work, the farm produces 10 potatoes, and each potato is sold on the market for $1 each. After work, each worker becomes consumer and spend their $1 salary to buy the potato, and the farm receives $10 from the consumers as revenue.
This economy is completely balanced and can function sustainably and perpetually. Money flows from the corporation to workers as wages, then workers turn to consumers, and spend the money to purchase goods and services from the corporation. Now let’s see how technology disrupts this perfectly balanced circular flow of money, and creates imbalance. Let’s follow the money.
Imagine someone ships a combine harvester to the island. Each worker’s productivity is greatly increased, and they can produce twice as many potatoes. However, since the land on the island is finite, 10 potatoes is the maximum that the island can produce every day. Therefore, at the end of the day, the farm lays off 5 workers. The 5 workers spent their last $5 to buy potatoes. After that, they no longer have any income, and they cease to exist within the economy. This is the first consequence of AI and automation to an economy, rising unemployment rate, or lower labor participation rate.
The farm still produces 10 potatoes each day, but they only need to pay $5 in wages. Therefore the farm is making $5 profit, this profit is basically the “salary” of the harvester and is kept by the farm owner. However, now the island only has 5 consumers, and each person can only consume 1 potato each day, so there is only demand for 5 potatoes. Each day, there are 5 potatoes that the farm cannot sell, and these surplus potatoes accumulate day by day. This is the second consequence of AI and automation, excessive capacity and overproduction.
The $5 profit made by the corporation can be used for investment. However, there is no investment opportunity as the island’s land is fully utilized. The $5 just sits in the corporation and does nothing. This is the third consequence of AI and automation, accumulation of money in the capital market.
So far we have assumed there is only one farm on this island. Now let’s imagine there are actually two and they are competing against each other. With the efficiency gain from automation and the reduction of wages, they each lower their price to $0.5 for each potato to compete. Everything else is kept equal. In this scenario, the 5 workers being laid off can survive for 1 extra day as their savings can now buy two potatoes. The corporation now has no surplus but the remaining workers each have $0.5 of surplus the first 2 days as they receive $1 wage but only spend $0.5 to purchase a potato. So the $5 from the lay off workers are now in the hands of the remaining workers. This is the fourth consequence of AI and automation: deflation. However, this only happens when the market is competitive.
On the third day, the corporation only receives $2.5 dollars from the remaining workers, but pays out $5 to the remaining workers. The farm then cuts down on capacity to match demand and only produces 5 potatoes each day. As the production is cut by 50%, the farm cuts workers by 50% again, leading to a vicious cycle of even weaker demand, more surplus cash, further production cuts, and so on and so forth. Until the economy vanishes.
In conclusion, there are 4 consequences of AI and automation:
- Lower labor participation rate.
- Excessive capacity and overproduction
- Accumulation of money in the capital market when the market is not competitive
- Deflation when the market is competitive
Many people may argue that there will always be new jobs for workers and new investment opportunities for corporations. However, this thought experiment aims to illustrate the broader impact of AI and automation on the entire economy across multiple sectors. When the economy reaches an ever-higher level of automation, a key issue arises: money becomes increasingly concentrated, leading to reduced circulation in the real economy.
There is a limit to how much more a wealthy person can consume relative to their wealth. For example, a person with $100 billion consumes far less than 100 million people with $1,000 each. Consequently, most of the wealth of the rich circulates within the capital markets rather than the real economy. The economy only grows when money circulates in the real economy. With reduced circulation, the economy contracts.
With the advent of deep learning and large language models, we now possess the technology to automate virtually any cognitive task. Therefore, the profound consequences of AI and automation on the economy are inevitable.
Examining historical data
In fact, over the past 60 years, those 4 consequences have been experienced in part or fully in many economies around the world. Let’s take the US auto industry as a case study. The number of vehicles produced increased 35.98% from 1960 to 2021, but the number of auto workers decreased 14.14% in the same period.
When workers are laid off due to automation, it first appears as a rise in the unemployment rate. However, unemployment only counts those who are actively seeking work. Initially, workers can still find other jobs with new sectors emerging. However, the trend of AI and automation means the rate of job displacement outpaces the rate of new jobs being created. Over time, those workers cannot find work and drop out of the labor force. The US labor participation rate peaked around 2000, and has since steadily decline.
The consequence of excessive capacity and overproduction can be seen in the decreasing trend of US capacity utilization.
The consequence of the accumulation of money in the capital market can be seen with the ever-increasing price-to-book value of US stocks.
The US did not experience deflation, partly due to excessive money supply from quantitative easing, but mainly due to a lack of competition with technology monopolies. Japan, on the other hand, has experienced deflation for the last 25 years. China is now also starting to experience deflation and is also accused of overproduction.
The simple fix
So what is the solution? How can the disruption of the circulation of money by automation be fixed? As mentioned in my last post, a Universal Basic Income payment (UBI) to the people will restore demand and provide a safety net to people affected by job displacement. But how can we fund UBI? Major governments around the world are already running huge budget deficits. Raising new taxes is not going to be easy; otherwise, it would have been done already. Where does the money come from?
Where does money come from?
Money is printed out of thin air.
The solution is surprisingly simple. Just print money and give it to the people in the form of UBI.
Wait, surely this will cause massive inflation? As mentioned before, with AI and automation in a competitive market, the consequence is deflationary. This will counteract the inflationary pressure from printing money to fund UBI. The deflationary pressure is driven by efficiency gains from automation, but also from the reduction of the workforce. The wages that would have otherwise been paid out to workers are now injected back to the consumers in the form of new money.
This is like rearranging terms in an equation from the left-hand side to the right-hand side. The net effect is zero. It is a short circuit to keep the circulation of money from corporations to consumers, bypassing workers, as workers are automated away. This restores the circulation of money, balances supply and demand, and solving the problem of overcapacity.
This is the new economic model in the age of full automation. A new form of monetary policy that controls money supply by adjusting the level of UBI payment, to directly drive consumption and achieve the target inflation rate of 2%. The level of UBI will be entirely determined by the central bank, instead of the government, to avoid populist abuse. This is another lever for the central bank to control money supply in addition to interest rates. The central bank now has two levers to achieve the target inflation rate and to promote GDP growth. Unemployment is no longer a target as automation will displace jobs.
As the economy achieves ever higher levels of automation, there will be more monetary headroom for the central bank to increase the UBI level. Governments should keep the conventional unemployment benefits in addition to UBI to ensure that people who are most directly hit by job losses due to automation are taken care of, and this should still be within the domain of fiscal policy and funded through tax revenues. It is important for UBI to be controlled by the central bank and to keep it the same for everyone. This keeps incentive structure for people to work.
Technology is the cause of inequality
When humans were hunters and gatherers, income was more or less equal. Everyone needed to hunt, and each person received a similar amount of food. This was the same in agricultural society, where each farmer owned their own tools and farmed their plot of land in family units.
It is only when technology developed to a point where tools became too large and complex for a single person or family to fully own that the situation changed. Tools then needed to be shared, and in capitalism, they are owned by capitalists, while everyone else became workers. The contradiction between the tool owner and the people who use the tools then emerged. The main problem is the disruption in the circulation of money. In our thought experiment of the remote island, in the end, no one has the money to buy the potatoes while there is a huge surplus of potatoes left to rot.
Karl Marx articulated all these problems of capitalism in his work “Das Kapital.” However, his solution was to have the state own the means of production and completely eliminate the role of the market. This is entirely wrong as communism destroys the incentive to work and eliminates competition. Competition is what drives progress and evolution, in both nature and human society.
It is likely that he didn’t see printing money to fund UBI as a solution, because monetary policy did not exist during his time. In the 19th century, currencies were based on gold standard, not the fiat currency we have today. There was no concept of creating money out of thin air, let alone controlling the money supply.
The core problem is the disruption of the circulation of money, and the fix is to inject new money into consumers to restore the circulation. The market mechanism and competition are still the best ways to drive progress and foster innovation.
However, before such a policy can be implemented, the economy needs to be sufficiently competitive to ensure automation results in deflation. Unfortunately, the US currently does not meet the condition to implement this policy as the economy is inflationary. The level of competition within the market first needs to be fixed, then as the level of automation increases within the service sector and service sector jobs are replaced, sustained deflation will also arrive. By which point the US can implement this policy.
Japan and China are possibly the only two economies ready to implement such a policy as they both experienced sustained deflations. China also has the unique policy tool of digital RMB to directly inject M0 money into consumers. I suggest starting with a small amount, like 1 USD a month, and observing the reaction of the economy.
The concept of printing money and distributing it freely to people is indeed radical and hard for many to accept. However, we must acknowledge that as automation increasingly takes over human labor, it becomes necessary to provide goods and services without requiring work. If we fail to adapt, the economy may falter, and market mechanisms could collapse. Despite this shift, the market remains crucial in directing energy, natural resources, land, and time to maximize our well-being.
In my next post, I will dive deeper into the cause of technology monopolies and the effective fix for this problem. In the meantime, tell me what you think about my proposal for a new economic model in the age of full automation.