Unpacking Patent Boxes

An intellectual property box is a special lower tax rate for profits associated with intellectual property (IP). There are over twenty around the world, and over the last few years bills have been introduced to the U.S. Congress for a U.S. IP box. IP boxes always include profits from patents but often include profits from some copyrights or even technical know-how.  For additional background and discussion, Agustin Redonda has previously written here on IP boxes.

Supporters give two key reasons for an IP box, and both are spurious: that an IP box promotes innovation, and that an IP box will win the tax competition with other countries.

Diane Feinstein, a senator from California, states on her web site (in part directly cut and pasted from an industry think tank) that “companies are experiencing a significant decline in innovation….  One strategy to reverse this trend is [a patent box].”

The assignee of a patent is the owner of the rights embodied in the patent, and is the person or entity to whom royalties are paid.  The typical case in large firms is that the inventor is contractually obligated to assign the patent to his or her employer or a holding company, which could have its official address anywhere, including locales commonly identified as tax havens such as the Cayman Islands or Luxembourg.

The core purpose of the IP box offer of a discount on royalties or other IP-related profits is to attract assignees, not inventors. The country gains inventors only indirectly, should the assignee choose to use its tax savings to hire more inventors in the country that offered the IP box.

There is a long literature on lowering tax rates to attract profitable firms, known either as tax competition or more descriptively, the race to the bottom.  If country A is offering a 20% tax rate, country B can win over some mobile firms and thus gain more tax revenue by setting a 15% tax rate — but then country A can undercut to 10%. After the game settles to an equilibrium, will the tax rate be above zero?

In a CEP working paper I consider two race-to-the-bottom games. In the first, country A has a unified tax rate for both profits from fixed capital such as manufacturers that need local resources or vendors selling to local residents, and for mobile capital like IP that can be assigned anywhere. Then the race to the bottom does not reach zero: every time A lowers its tax rate it improves its chance of winning the mobile capital, but at the cost of losing revenue from the fixed capital. The paper presents proofs that these two opposite forces balance at an equilibrium with strictly positive values.

In the second game, both countries have two separate rates, one regular rate for revenue from fixed capital and an IP box rate for mobile capital. Now, country A bears no loss from lowering its IP box rate, so there is nothing keeping the rate from eventually dropping to zero. The paper shows that, even with some frictions, the payoff from mobile IP capital must be lower with two distinct rates than with one unified rate.

In short, setting a distinct IP box rate to attract mobile capital accelerates the race to zero. The only way to win the IP box game is to not play.

Technical issues

The Feinstein web page features pictures of people producing silicon wafers and assembling electronics, but it is not just the electronic device on which you are reading this that involved patents: manufacturing the textiles on your back, the chair you are sitting on, and the building materials surrounding you almost certainly involved some patented tool or process, making them eligible for the proposed IP box rate.  In the U.S. there are even thousands of valid patents for financial transaction techniques, so purely financial firms may be able to join in as well. A broad enough IP box may simply be an across-the-board change in the tax rate, by another name. It is not unreasonable to claim that any given country needs to lower its tax rates, but rather than doing so with a tax incentive so broad that it covers almost every firm that can file the right paperwork, it may be simpler to achieve a lower effective corporate tax rate by lowering the corporate tax rate.

If the IP box does not cover everything, then some firms would have some revenue taxed at a higher rate and some at a lower rate. As an exercise for the reader: if you were running a firm with a two-tier tax structure, how could you lower your tax burden without doing any new research or development? If you need a hint, the working paper linked above lists seven methods of doing so.

Conversely, there is no lawful way to claim a research credit for direct R&D expenses without spending on R&D. If our goal is to attract inventors and not assignees, a credit for inventor salaries is far more direct, simple, and effective than a reduced tax on profits derived from products that use patents on results from research by inventors.

The one clear benefit of an IP box is that every politician and voter wants to support innovation. Directly lowering the corporate tax rate is a steep challenge for any politician; an incentive for innovative firms is not so daunting, even when the definition of innovation is so vacuous as to include the entire economy.  A tax discount to draw offshore profits back inside a country’s tax purview looks like a handout to wealthy investors; a lowered tax rate for patent-associated profits is support for innovation. But in truth, these outcomes bear a tenuous link to supporting invention, and bring inefficiencies and distortions relative to the alternatives. For example, if our honest goal is to expand research and development funding, we can do so by expanding research and development funding, via grants to public research institutions and tax credits for private R&D. When all is considered, the IP box is neither an efficient way to promote domestic R&D nor a beneficial strategy in international competition.