MIT’s Task Force on the Work of the Future has been providing perspectives, responding to media inquiries, and participating in virtual events to share viewpoints on the economy, government response, worker voice, and technological advancements.
Employers often ask whether a machine or a person is better suited to do a job, based on factors such as speed, precision, and what’s technically possible. But one of machines’ biggest advantages over a worker may be how they are taxed.
While labor has been taxed at an average rate of about 25% for the past four decades, things like equipment, software, and buildings that are classified as capital are taxed at a lower rate — and this rate has fallen steadily in recent years. For every dollar a worker receives, the business that employs them has to pay an additional 25 cents in taxes, effectively making the cost of labor to employers 25% higher. In contrast, the average tax rate on software and equipment stood at about 15% in the 1990s and fell to about 5% after a series of tax reforms in the 2000s and 2010s.
In essence, the U.S. tax system encourages companies to buy machines while discouraging them from adding workers.
While lower taxes on capital is a long-standing feature of the U.S. tax code, the issue now is more acute as new automation technologies move into a larger variety of worker tasks performed throughout the economy, from warehouses and factories to insurance firms and stores. Automation, which involves the substitution of machines and algorithms for tasks previously performed by workers, could be an engine of growth. But when it is excessive — for example, driven by tax incentives rather than for efficiency gains — it harms labor and fails to improve productivity. In this context, it is notable that the biggest drop in tax rates in recent years has been for software, followed by equipment, exactly the types of capital goods used to build automated systems, including robots, numerically controlled machinery, specialized software, and, more recently, artificial intelligence.
Acemoglu and Restrepo (2019) have argued that there has been faster progress in automation technologies over the last 20 years and much slower progress in technologies complementary to humans. If this view is correct, then U.S. technology looks much more substitutable for human labor than it did in previous decades. In the past, it was more common for every new machine to require at least one or more human operators, and subsidies to capital indirectly helped labor as well. This may no longer be so with rapid advances in automation technologies, making the asymmetric taxation of capital and labor much worse news for labor.
In this brief, we first examine the different forms of automation and then turn to how the U.S. tax system has led to excessive reliance on machines in ways that can hurt workers and subdue wages. We subsequently look at how the tax system evolved to favor capital and what U.S. policymakers could do to reverse this negative impact and level the playing field between capital and labor.