Three Stanford scholars explore how we measure innovation, how innovation drives productivity, and how productivity affects inequality.
In 1500, China’s economy was the strongest in the world. But by the 19th century, the U.S., Western Europe, and Japan had leapfrogged over China by churning out goods and services in vast quantities while the former superpower stalled.
Why? Some economists argue that China’s lack of free markets and unencumbered innovation in the West led to the shift.
But what is the relationship between innovation and markets, productivity, and inequality?
The answer to that puzzle and others were explored during a recent forum on the relationship of innovation to economic growth at the Hoover Institution.
Three Stanford professors, all Hoover fellows — Stephen Haberopen in new window, Edward Lazear, and Amit Seru — spoke on a panel moderated by Jonathan Levin, dean of Stanford Graduate School of Business.
The panelists offered thoughts on how innovation is measured, the role of markets, and what types of firms are likely to innovate.
They examined how productivity affects wages, skills, and social inequality, and considered what kind of policies might ensure that the pace of innovation remains brisk.
How Do You Measure Innovation?
Like art, everyone knows innovation when they see it, but defining and measuring it, says Amit Seru, “is a holy grail” for researchers.
Studying patents might be key to answering that question.
Seru and his colleagues used big data techniques to analyze 9 million U.S. patents filed over two centuries.
Although the Silicon Valley ethos holds that startups are the wellspring of innovation, the researchers found that established firms were also very innovative, as measured by high-quality patenting activity.
They also concluded that both private and public firms contributed to innovation and that universities and some government entities were also quite innovative.
The first step in that analysis was to construct a measure of high-quality innovation. The researchers did so by comparing the texts of all the patents in the database and tabulating the occurrence of important words. If there was little overlap between the text of a patent and its predecessors, the patent was likely a novel innovation.
If words in subsequent patents were similar, the subject patent was likely an important innovation that other patents had built upon.
Patents meeting both criteria, i.e., novel and important, were considered “high quality,” says Seru, a Stanford GSB professor of finance.
As a check, the researchers compared their list of high-quality patents to those already deemed significant by economic historians.
The two lists were quite similar, they found. Using this measure of high-quality innovation, the researchers examined which entities contributed to breakthrough innovations over time and what patterns were consistently associated with these events.