Do Tax Policies Spur Innovation?

There is broad support in the United States for faster economic growth—a goal that ultimately depends on boosting labor productivity. In the long run, that requires more innovation, a big component of which comes from private research and development (R&D). But what, if anything, can countries do to increase domestic innovation?

Over the last few decades, a growing number of countries have adopted tax policies to spur innovation. The two most important of these policies are research tax credits and patent boxes. Tax credits lower the after-tax cost of R&D by allowing companies to deduct part of their expenses from their tax liability. Patent boxes apply a lower tax rate to profits derived from intellectual property, such as patents. The question is whether these policies really result in more innovation. A recent literature review by a noted innovation economist summarizes some of the available evidence and finds that tax credits produce more private R&D along with strong social benefits, and patent boxes can help countries retain the intellectual property they produce.

Economist Bronwyn Hall has written many of the leading papers on innovation policy. In a recent paper for the National Bureau of Economic Research, she reviews the literature on whether the economic benefits from tax credits and patent boxes exceed the lost tax revenue. She starts by reviewing the basic rationale for subsidizing research—namely, that private companies are unable to capture all the benefits from their R&D. A large percentage of these benefits spills over to benefit society in general. But private companies aren’t charities; they don’t consider these benefits when deciding how much R&D to conduct, so they do too little of it. Policies that lower the after-tax cost of R&D can encourage companies to do more.

Hall points out a number of complications in analyzing the impact of R&D tax credits. One is that a firm’s innovation strategy has many components other than the amount of R&D it conducts, including employee training, organizational change, and capital investment. Tax credits may have little effect on these. In fact, because the most important share of innovation spending is on new equipment, computer hardware, and software, accelerated depreciation can have a bigger impact on total private R&D than policies specifically aimed at research. Moreover, the people actually making decisions about research budgets may not be aware of the relevant tax policies. If so, these policies will not affect company actions.

Despite these complications, Hall cites evidence going back to the 1980s showing R&D tax credits “are generally effective at increasing business R&D, with a price elasticity of one or higher,” meaning that a 1 percent increase in lost tax revenue increases private R&D by at least 1 percent. Other studies show the increase in R&D spending balances or even exceeds the value of the lost tax revenue. Moreover, significant spillovers continue to exist. Hall cites one study arguing that socially optimal R&D investment in the United States would be at least four times the current level. She concludes that R&D tax incentives should be even larger than they are.

The evaluation of patent boxes depends on whether their primary purpose is to encourage more domestic innovation or to deter companies from moving the profits associated with new or existing IP to low-tax jurisdictions. Because patent boxes are still relatively new and fewer countries have implemented them, research results are imprecise and sometimes conflicting. The link between policies and outcomes is also complicated by a recent OECD agreement that requires companies to develop the IP within a country in order to qualify for a lower rate. This reduces the incentive for companies to move existing IP in order to qualify for the benefits of a patent box. It should strengthen the link between tax benefits and domestic innovation, but unfortunately this link is also weakened by an EU requirement that forbids member countries from discriminating against research done in other EU countries. The United Kingdom, for example, must treat any innovation done in France as being done in the United Kingdom for purposes of qualifying for the credit.

Hall concludes patent boxes are mainly defensive: “[Although] the encouragement of innovative activity and IP creation may be motives for lowering taxes on IP income, countries are effectively forced to do this by the presence of many low tax jurisdictions around the world into which such income could migrate.” In response to this pressure, some countries have pushed for a minimum tax on worldwide income at the headquarters level. This would reduce the benefit of moving activity to low-tax jurisdictions. The United States recently took two steps in this direction by enacting the Global Intangible Low Taxed Income (GILTI) and Base Erosion and Anti-Abuse Tax (BEAT) provisions in its recent tax reform.

Patent boxes can be effective in keeping IP income from moving abroad. One study shows that a 10 percent decrease in the tax rate on patent income tax leads to a four or five-fold increase in transfers of existing patents over the next three years. However, this impact disappears if companies are required to develop the patents within the country. There is little evidence so far that they increase the amount of local innovation. But this is partly because many countries have not yet linked their patent boxes to domestic activity. This should change going forward now that countries are required to link the tax benefits to innovation done within the country.