The U.S. R&D tax credit corrects a serious market failure and spurs U.S. companies to invest in more R&D, creating more jobs, more competitive companies and a stronger U.S. economy.
The R&D credit is on the list of “extenders” that Congress should take up in the lame duck session before they adjourn in December. Earlier this year, Michael Rashkin wrote in EE Times that “The R&D credit doesn’t work” and advocated for its elimination. Let’s hope Congress ignores Rashkin’s advice. Here’s why.
First, Rashkin’s claim that the credit is ineffective is wrong. He writes the “credit has been in existence for over 30 years. During that time, not one company has announced an invention for a product that was produced as a result of the R&D credit.”
Of course they haven’t because the credit is not intended for specific lines of research. It’s not as if IBM gets the credit to fully support its research on quantum computing, but gets no credit for its advanced battery research. The credit is covers the full range of a company’s R&D. This differs from research support from federal agencies like DARPA and NSF, which Rashkin cites, that support specific research projects. Judging the credit on this basis makes no sense.
Second, Rashkin dismisses the body of scholarly, peer-reviewed academic work on the credit, claiming it is based on invalid “what-if-calculations.” In fact, these reviews use sophisticated econometric analysis and have consistently shown that the R&D credit leads companies to do more R&D than they would otherwise.
Rashkin then goes on to say that these studies show that for every dollar of tax expenditure through the credit only one dollar of R&D is generated. In fact, most studies show a much higher effect, with the range being between $1.30 and $2.90 of corporate R&D induced for every dollar of R&D credit tax expenditure.
Rashkin’s claim that companies will conduct R&D even without the credit, which he acknowledges is wrong when he says the studies show a dollar-for-dollar effect, is beside the point. Without incentives, these companies will not conduct enough R&D, thereby making America worse off.
The single most important reason most economists support the R&D credit is to correct for a market failure called “spillovers.” It’s one thing if Apple, Intel, GE and others received all the benefits from their R&D. But they don’t. On average, U.S. companies get less than half of the total benefits from their R&D and the remaining “spillover” benefits society.
For example, Tewksbury, Crandall, and Crane examined the rate of return from 20 prominent innovations and found a median private rate of return of 27 percent, but a median social rate of return of a whopping 99 percent, almost four times higher. So even if companies invest “rationally” in R&D to maximize profits, they will still be under-investing from a societal perspective. The R&D credit is a key way to address this structural market failure.
Rashkin dismisses this by claiming that “spillovers” do not square with “free market” economics, whatever that is. I would challenge Rashkin to find any legitimate microeconomics textbook that does not discuss the notion of market failures and positive externalities (including spillovers). Only the most hardcore, doctrinaire free-market ideologues believe that markets always get it right and there are no market failures.
Rashkin also seems to think that doing R&D in the U.S. harms the U.S. economy when he claims “to give these companies a research credit in addition is to reward, in many cases, behavior that is detrimental to the U.S. economy.” But the R&D credit does not reward moving production or R&D offshore. It applies only to R&D performed in the U.S., an activity that without doubt is beneficial to the domestic economy.