Header Logo

Universal Basic Income: Paying For It Through Automation

How and why we could use the productivity wage gap to fund Universal Basic Income in the United States

November 26, 2017-4 min read

 — This is a follow up on why we should implement Universal Basic Income. The original post, Universal Basic Income: The Maslow Argument — 


Now how do we pay for it? There’s a range of ideas. It would take all of them to keep things budget neutral.

If we’re going to do it we should do it all the way. Let’s not nickel and dime. It should be $1200–1500 per month and it should be truly universal, everyone gets it, no questions asked.

One idea that’s worth floating into the mix is to place a tax on business productivity increases.

Hear me out.

It’s no secret that worker productivity has tripled while wage growth has only risen 30% since the 1970’s.¹ Clearly this is a trend that we can attribute to automation. Automation is not just robots putting cars together, it’s also Microsoft Office, email, algorithmic trading, Google maps, etc. It’s anything (but lately it’s software) that lets you do more than you were doing before in the same amount of time. When one person can do the job that three used to it’s because part of their job has been bundled into a new technology that lets them do it faster.

Lagging that increase in productivity is the stagnation of wages, particularly hourly wages.

Since the ubiquity of automation is increasing we should be looking at the gap between what businesses pay for an employee and how much they save from letting software eat the employee’s job.

Let’s say a shipping company with 10 employees buys a new suite of software that allows them to ship double the amount of cargo that they were shipping with the exact same number of employees. That company is now getting double the labor for very close to the same price they were paying, they just have to purchase the software. That’s awesome, let’s keep doing that. But let’s take a look at payroll taxes for a second.

That doubled amount of labor doesn’t require the business to pay the payroll taxes required of actual human workers. A human worker at a $50,000 salary costs every business roughly $3,500 a year in payroll taxes. By doubling productivity with the same number of people the effective payroll tax for the company has been halved. What would have cost $35,000 in payroll taxes for the 10 people they would have had to hire now costs exactly $0 because you don’t have to pay payroll taxes on software or machines. Effectively, the company is now paying $1,750 per unit of employee output.

Given productivity growth numbers, employers are getting a discount on payroll taxes at an effective rate of about 70%². I’m not saying that we should recoup that exact amount, that would disincentivize the drive towards higher productivity. What I am saying is that we could feed the UBI pot by requiring automation to pay a share. A share that will continue to grow as time goes on; a sustainable source of revenue into the pool.

Ultimately, keeping the “automation” tax rate lower than the current payroll tax rate would encourage businesses to automate. When choosing between paying the higher rate for a new employee verse the lower rate for automation the business is going to choose automation. Encouraging that choice would funnel more money into the UBI account. Little by little, as more and more becomes automated and productivity continues to increase, that account will reap the benefits of innovation. As a kicker we wouldn’t have to feel guilty about automating jobs away.

I’m pro-business, a capitalist that’s had multiple small businesses. This is certainly not a rant against business. My premise is simply that we will unlock an endless cornucopia of human potential if we allow people to focus on something other than how to pay rent.

The “why” we should do it is the easy part. The “how” is the hard part. It’s going to require a different way of approaching things, a true paradigm shift. Instead of striving for incremental changes we need to aim for what seems impossible.



¹You can spin the productivity, compensation numbers in a dizzying number of ways — Heritage Foundation, BLS, EPI

An excerpt from Heritage:

“Economic theory predicts that employers pay workers according to the value of their marginal product — the benefit to the firm of hiring them. Marginal productivity depends on the price at which a business sells its goods, not the price of other consumer goods. Economists would expect a U.S. company to pay more if higher demand raised the price of its goods or services. Economists would not expect the company to increase compensation because, for instance, oil imports became more expensive.”

On a micro scale that’s relevant. But the employee of that company does not live on a micro scale. While the price of oil imports might not matter to her employer it most definitely matters to her living in the same macro environment as all other humans.

That argument accounts for a 21% of how the gap is less than 70%.

The article rests most of its logic on measuring compensation as a function of Implicit Price Deflator (goods and services that businesses sell to other businesses or export to other countries, as well as those sold to consumers) as opposed to the Consumer Price Index. Since we are not measuring the state of business (IPD) profit verse productivity but instead worker (CPI) compensation verse productivity this seems to me to be an invalid comparison. One that I see as comparing apples vs. oranges — a statement the author uses to illustrate his point in the exact opposite direction that I do.

And that argument accounts for the other 26% of how the gap is less than 70%.

²productivity — total compensation




© 2019 Phil Andrews