Over the past half-century, a troubling trend has emerged. America’s research has drifted away from deep innovation toward incremental innovation. Deep innovation is about exploring uncharted territory— where the payoffs are not obvious and the time scales may not be short. Some deep innovation can be transformative and the stuff of economic revolutions. Who would have guessed when Russell Ohl studied how current flowed through a cracked crystal in 1939 that it would lead to the P-N junction, the basis for semiconductors, enabling a revolution in modern computers and telecommunications, which have markets now worth hundreds of billions of dollars? Who would have guessed, when Barnett Rosenberg applied electric fields to growing bacteria in 1965, that the platinum metal from the electrode would leach into the solution and hinder cell division, resulting in cisplatin, the penicillin of cancer drugs, now a billion-dollar industry?
Since World War II, deep innovation in America has largely come through federal funding of academic research. Government funding is essential for supporting national needs like defense and for helping our technology industries. Some research simply is not done by industry. It is considered too basic, too blue sky, too expensive or too diffusible, meaning that certain research efforts can leak out beyond a company’s walls and pay off for other companies. Take the computer mouse. It was researched and developed by Xerox Corp. But the mouse paid bigger dividends for computer companies than for copier companies like Xerox. These days, industrial research is more targeted; it is mostly “D,” not “R,” aimed at making better widgets for tomorrow’s store shelves. The government plays an irreplaceable role in supporting research that is deep, broad and diffusible— the kind of work that has the potential to transform our science and create new industries.
By definition, deep innovation is often unpredictable
Imagine having written years ago a grant proposal to invent the iPad. Reviewers would have rejected it. There were already larger versions (computers). There were already smaller versions (iPhones). And there was no apparent market: Existing tablet computers were flops. So on the day the iPad appeared, it was panned by the media, and Apple stock dropped 4 percent. The result? The iPad sold 15 million units the first year. Deep innovation is less about guessing the future than it is about supporting the people who can lead us to it. Seeking, finding and supporting deep innovation is very different than promoting incremental innovation. A reviewer of an incremental-innovation proposal decides whether the problem is important and the solution feasible, whether the preliminary results give sufficient proof of success, the possible ways it might fail, and how quickly the investigator can achieve his or her goals.
But judging deep innovation requires nearly the opposite set of determinations. Rather than asking whether the problem is important, one must ask whether it is interesting. Rather than hoping for an expected outcome, one must hope for surprises. Instead of minimizing the time to payoff, the goal is to maximize how long the consequences will reverberate. Instead of more scrutiny and longer proposals, deep innovation requires less scrutiny and shorter proposals. Controversy is not something to be avoided. Instead of looking for the next iPad, those interested in deep innovation look for the next Steve Jobs. Does the investigator have the passion to focus on a problem, the sensitivity to recognize a small signal in a large amount of noise, the ability to connect the dots, the tenacity to withstand the objections of critics and the perseverance to follow a road wherever it may lead? It’s not that we need better reviewers. Nobel Prize winner Linus Pauling objected that quasicrystals could not exist when Dan Schechtman, winner of this year’s Nobel Prize, discovered them in 1982. With deep innovation, even our best reviewers are usually wrong.
Some agencies have made progress supporting deep innovation. The National Institutes of Health, for example, has recently developed Pioneer Awards, the Eureka Program, Transformative RO1s and New Biomedical Frontiers at the Interface of the Life and Physical Sciences.
But we need much more
The two bottom curves on the figure (http://www.nsf.gov/statistics/) show that our federal R&D (divided by gross domestic product, to normalize for the size of the economy) has diminished since the 1960s while industrial R&D has grown. In the 1960s, the U.S. invested $2 in basic research for every $1 companies invested. Now, it’s the opposite: The U.S. invests $1 for every $2 invested by companies. If this trend continues, who’s going to generate tomorrow’s industrial revolutions?
Taxpayers tend to focus on immediate threats. The peak of R&D spending in the ’50s and ’60s arose from military threats. To end World War II and respond to Russia’s launch of the Sputnik satellite, the U.S. created new funding agencies. Taxpayers see more clearly the benefits of research that is mission-oriented (i.e., targeted: the U.S. Department of Energy’s mission is energy, the U.S. Department of Defense’s mission is defense, NASA’s mission is aerospace, the NIH’s mission is health), rather than discipline-oriented, such as that of the National Science Foundation.
Our research universities are looking for quicker payoffs too. Universities are becoming more entrepreneurial as their federal support shrinks. For example, as our pharmaceutical industry grows more risk averse, academics are stepping up. By one estimate, nearly 80 academic research units have sprung up in the last four years focused on pharmaceutical discovery, which is traditionally the business of industry. Of course, we all benefit from the start-up companies that universities spin off. But there’s also a potential downside. If our universities divert too many resources to short-term payoffs, we risk losing the basic-science wellspring of tomorrow’s science and technology.
The top panel of the figure shows the shifting balance between the two main funders of academic research. Until recently, NIH budgets roughly have kept pace with the economy. But NSF budgets have not. The NSF’s discipline-driven funding is a shrinking part of university funding.
And, look at the shifting balance in university departments. A typical university used to have one department each for biology, chemistry and physics. Now a university may have five to 10 different flavors of biology departments (biochemistry, systems biology, bioengineering, genetics, physiology and so on), while still having only one for each of the other basic sciences. Biology, of course, is crucially important. But the NSF provides the broad underpinnings of all sciences. We don’t know where to look for the next unexpected, transformational technology. We should not foreclose our options. Tomorrow’s science, engineering and technology may spring up from discoveries that are unimaginably unrelated to the disciplines in which they are born.
What I propose
Here’s what I would say to a national innovation czar, if we had one. First, to solve a country-size problem— like our current jobs crisis— by creating the next $100 billion-per-year technical industry, we need to increase the federal R&D budget to 1.7 percent of GDP. That was the level that worked in the 1960s when President Kennedy made good on his commitment to land a person on the moon. Increased funding would mean raising federal R&D to two-and-a-half times its current level. Today’s NSF budget should be around $17 billion, and the NIH’s should be around $78 billion. That seems like a lot of money, but, as former U.S. Rep. John Porter, R-Ill., once told me, the issue is not dollars; the issue is priorities. Even at those amounts, we would still be unable to fund many meritorious proposals. And today’s entire NSF budget (around $7 billion) is smaller than the single-company R&D budgets of Pfizer, Merck, Microsoft or Ford.
Second, I would suggest that our innovation czar develop new initiatives to protect deep innovation. After the 1960s and ’70s, our portfolio of deep innovation has become inadequate to power an economy as large and technology-based as ours. For the disruptive technologies that have driven economic revolutions— and that could create the jobs of the future— we need more academic research, and we need a protected portfolio of deep innovation.
Thanks to Alberto Perez and Jim Larimer for their assistance.
Ken Dill (firstname.lastname@example.org) is the director of the Laufer Center for Physical and Quantitative Biology and is professor of physics and chemistry, Stony Brook University.