Elected leaders around the globe often take office with promises of driving growth through innovation. But deciding which policies to champion in order to achieve this goal can prove difficult. Nicholas A. Bloom saw the challenge firsthand when he served as an advisor in the UK treasury department in 2001.

“Ministers get lobbied by endless firms, academics, and other politicians. It’s hard for them to conclude what works and what doesn’t, because each group only presents evidence favoring their own policy,” Bloom says. “I remember wishing there was some overview by an independent group based on actual research.”

Nearly two decades later, Bloom, now an economics professor at Stanford School of Humanities and Sciences and by courtesy at Stanford Graduate School of Business, did exactly that. Along with economists Heidi Williams of Stanford University and John Van Reenen of the Massachusetts Institute of Technology, he came up with an evidence-based toolkit for governments looking to spark innovation.

Research shows that innovation is the main driver of a country’s long-term economic growth. At the same time, companies in free market economies are likely to underinvest when it comes to research and development, Bloom says. That’s because an individual firm bears all the costs for developing a new solution, but competitors reap some of the benefits by either imitating the invention or waiting until a patent runs out. That’s where the government comes in.

The Top Five Policies for Boosting Innovation

Outlined in a recent paper in the Journal of Economic Perspectives, here are five policies that Bloom and his colleagues say can effectively drive innovation:

1. Offer Tax Incentives for R&D

The research is clear: Government tax subsidies and grants are the most effective way to increase innovation as well as productivity. Studies show that reducing the price of R&D by 10% increases investment in innovation by 10% in the long run. The U.S. has one of the least generous tax credit policies among developed countries, only cutting the cost of R&D spending by around 5%. Countries like France, Portugal, and Chile are the most beneficent, slashing the cost of R&D spending by more than 30%.

2. Promote Free Trade

Existing evidence suggests that opening trade can spark innovation by increasing competition, allowing new ideas to spread faster and dividing the cost of innovation over a bigger market. For example, researchers who summarized the findings of more than 40 papers in 2018 concluded that freer trade generally increased innovation in South America, Asia, and Europe (results from North America are more mixed). This policy can bear fruit in the medium run and doesn’t cost much to implement, but can increase inequality.

3. Support Skilled Migration

Even if there’s more funding for innovation, you won’t see it unless you have more scientists to do the research. The most direct way to increase the supply of researchers is to allow more high-skilled immigrants into the country. One paper showed that increasing the population of immigrant college graduates in the U.S. by one percentage point increased patents per capita by up to 18%.

4. Train Workers in STEM Fields

Another way to increase the supply of researchers in the long term is to invest in training them domestically. One option is to promote programs that boost the number of people studying science, technology, engineering, and math (STEM). Another is to expose more would-be inventors from disadvantaged backgrounds to role models and mentoring.

5. Provide Direct Grants for R&D

Compared to tax incentives, government grants — often to university researchers — can target projects that are likely to have the most long-term benefits. Research shows that grants to academics in turn results in more patents filed by private firms. However, it’s tough to track the impact of grants, since it’s possible that private R&D funding would have stepped up in their absence.

Other Strategies That May Boost Innovation

Bloom and his colleagues found that three policies may go some way toward increasing innovation, but aren’t bolstered by much evidence.

1. Providing Incentives for University Researchers

Some evidence exists that academics who retain full rights over their work are more likely to patent inventions and found startups. However, there’s not enough research to make this policy a priority.

2. Engaging in Intellectual Property Reform

Debates continue to rage about whether certain systemic reforms, such as curbing “trolls” who buy patents just to go after supposed infringers, would improve innovation. The researchers conclude there’s no strong evidence tying a particular IP policy change to greater innovation.

3. Embarking on Mission-Oriented Projects

“Moonshot” R&D projects focused on specific sectors have led to remarkable innovations, including NASA landing a man on the moon. The researchers point out that moonshots might be needed to innovate in response to pressing crises, like climate change. But proving their effectiveness is difficult, and funding is likely to be dictated by politics rather than the most beneficial technologies.

What Not to Do

There’s one policy the researchers squarely warn against: Establishing patent boxes.

Patent boxes involve taxing revenue based on patents at a lower tax rate than other corporate revenues. Originally introduced in Ireland in the 1970s, the policy was in use in 16 developed countries as of 2015. Bloom and his colleagues conclude that these tax breaks don’t actually encourage R&D spending, but rather incite firms to move their patents around to different jurisdictions and distort the tax system.

Falling Behind in Federal Funding for Innovation

In the 1960s, the U.S. spent more on innovation as a share of GDP than any other developed country. Today, it spends 2.7% of GDP, still above average among developed countries but below Germany and Japan, and federal funding for R&D has declined relative to funding from private companies, which Bloom says has slowed economic growth in the U.S. “Businesses tend to focus more on short-term but lower economy-wide payoffs, whereas federally funded R&D tends to focus on big picture stuff that takes longer but pays off big time and benefits entire industries.”

If policymakers in the U.S., or anywhere, want to double down on innovation policy, Bloom has one request: “I’d love it if they could just look at this toolkit first before meeting the lobbyists.”