Bernie Sanders and Bill Gates don’t agree on many things, but they both want to see a “robot tax,” a special levy on companies that replace workers with machines.
Sanders and Gates may be well-intentioned with this call for a robot tax. But that doesn’t make their call on this the right one.
Sanders is asking for this tax on automating labor to prevent what he calls the “race-to-the-bottom profiteering” of multinational corporations. But is disincentivizing automation the best approach to accomplish that objective, or would we be better served by policies that get and keep the fruits of technological progress shared equitably between employers, workers, and customers?
Is taxing automation to cover the cost of lost jobs even logically feasible? Yes, a robot may replace workers, but what about the jobs involved in creating that robot? Should those jobs be taken into account? How do we know when a given case of automation has eliminated more jobs than it has created? And how would a robot tax address automation already in place? Would existing automation be grandfathered in? If so, how would we treat employers who implement automation just to catch up with their competitors?
And what automation would be taxable, any technology that replaces labor or only “robots”?
Using tax policy to disincentivize otherwise cost-effective automation would be a radically wrong-headed step. But taxing automation in a manner that fosters a level playing field between machines and workers would not be at all radical. Taxing automation in this way would simply be fair.
Unfortunately, our current tax law does not provide a level playing field. Our current tax law provides an ongoing unfair subsidy to automation, and this unfair subsidy can cause workers to lose jobs even to machinery that turns out to be less cost-effective.
Here’s how this unlevel playing field works: In determining their taxable income, businesses in the United States have generally been able to claim as a deduction whatever they reasonably spend to make their products. Say a worker’s compensation, health insurance, and other expenses cost an employer $100,000 per year. That employer can deduct $100,000 per year in determining its taxable income. For a corporation facing a 21 percent tax rate, that would translate into a tax benefit of $21,000 per year, for a total tax benefit of $105,000 over five years.
Now suppose a $415,000 robot — that costs an employer $100,000 per year for five years, after taking interest expenses into account — could replace that same worker. On a level playing field, the cost of the robot, including the interest expense, would be deductible at the same $100,000 per year as the cost of the employee. But under our current tax law, an employer gets to deduct both the cost of the robot and the robot interest expense on a “front-loaded” basis.
What would this front-loading entail? A business that purchased the robot and had it operational in 2023 would be able to deduct about $375,000 of the robot’s total cost in that first year, generating a first-year tax benefit of $78,750.
Ultimately, to be sure, the total deductions and tax benefits for the robot over the entire five years would equal those for the cost of the worker. But the front-loading of the tax benefits that our current law allows makes the robot far more valuable.
The bottom line: Our tax law is now tilting employer decision making in favor of machines, even when those machines turn out to be no more cost-effective – or even slightly less cost-effective – than workers.
“Bonus depreciation,” a tax rule first put in place just over 20 years ago, is driving our current uneven playing field. Before 2002, businesses would depreciate — deduct — the cost of new machines evenly over the period the machines could be expected to last. That made the tax treatment of machine costs roughly equivalent to the tax treatment of the cost of the workers the machines replaced.
But the new bonus depreciation rules let businesses deduct the lion’s share of a new machine’s cost in the new machine’s first year. In 2023, businesses will be able to deduct 84 percent of the cost of machines they purchase and place in service this year right away, with the remaining 16 percent of the cost deducted evenly over the remaining four years of the machine’s expected life.
The logic behind this extraordinarily generous — to corporations — treatment? This treatment has no logic behind it. This special treatment only serves to incentivize automation not otherwise cost-effective, an approach no less wrong-headed than using tax policy to disincentivize automation that otherwise would be cost-effective.
The good news here? Under current law, bonus depreciation will phase out by 2027, a move that will go a long way toward leveling the playing field between workers and robots. The bad news? Congress is threatening to move in the opposite direction. Late last year, a plan to maintain bonus depreciation — and allow for 100 percent of the cost of a qualifying machine to be deducted in the year of purchase — gained appreciable bipartisan support. This effort to extend bonus depreciation did come up short in 2022, but we can expect more attempts to extend bonus depreciation in the current Congress.
Ultimately, in the years ahead, automation is most definitely not going to disappear, and we certainly don’t want to impede tomorrow’s technological progress any more than we would have wanted to impede the development of the forklift or, to go back to the dawn of civilization, the wheel. But progress — true technological advancement — doesn’t require tax subsidies. Tax policies like bonus depreciation encourage investments that otherwise would not be economically justified. Such policies don’t promote progress. They impede progress and, in the process, hurt us all, workers included.