Ever since the economist Joseph Schumpeter immortalized the phrase “creative destruction,” most economists have argued that technology innovation creates both winners and losers — but that the economy as a whole benefits.
In practice, however, it’s been very hard to isolate the specific link between a company’s pace of innovation and its future growth, and even harder to confirm how much tech innovation is just disruption, shifting money from “losers” to “winners,” versus how much it increases total wealth.
Now a sweeping new study of millions of patents issued over the past century offers concrete answers to both questions.
Coauthored by Amit Seru, a professor of finance at Stanford Graduate School of Business, the study confirms that the pace of innovation is indeed a disruptive force between competitors — perhaps more than previously thought.
But it also finds that periods of rapid tech innovation, such as the 1920s, the 1960s, and the 1990s, lead to a measurable increase in overall productivity and economic growth.
“If innovation were only about McDonald’s getting ahead of Burger King, we wouldn’t really care,’’ Seru says. “But our study shows that something more important is happening. When firms innovate, they are expanding the pie, and we see increased aggregate growth.”
Seru and his colleagues — Leonid Kogan at Massachusetts Institute of Technology, Dimitris Papanikolaou at Northwestern University, and Noah Stoffman of Indiana University — reached that finding by coming up with a new way to measure the economic impact of millions of patents that companies received between 1926 and 2011.
Thus far, many of the studies of patents and innovation have been based on measuring the apparent scientific importance of patents, usually by counting the number of times a patent is cited as “prior art” in later patents.
As the researchers note, however, some patents have considerable scientific value but very little economic value. IBM patented a system, for example, to reserve bathrooms aboard airplanes. The patent went on to be cited in many other patents, but IBM let the patent lapse without ever doing anything with it.
Seru and his colleagues took a different approach, looking at how much the news of a patent’s issuance had on the stock price of the company that received it. The researchers looked at nearly 1.8 million patents that were granted and assigned to publicly traded companies from 1926 to 2011.
After using statistical techniques to filter out random fluctuations and market noise, they found that news of a patent being granted had a measurable impact on the stock price of a company on the day or two after the news was announced.
The researchers then looked at what happened to companies that were faster or slower at getting market-moving patents. Sure enough, they found a strong correlation between the pace of that activity and a company’s future growth and competitive edge.
Specifically, Seru says, companies in the top 10% on innovation experienced annual growth rates during the subsequent five years that were between 1 and 3 percentage points faster than those of companies with only a median pace of innovation. Given that 10% is a fast rate of growth for most companies, that amounts to a significant edge.
By contrast, companies that lagged behind their industry’s innovative pace saw their annual growth over the next five years slow by as much as 2.5 percentage points.
Those results are similar to those of earlier studies that looked only at the scientific value of patents, but the new results show more evidence of “creative destruction.” Earlier studies didn’t find as much damage to the less-innovative competitors.
The bigger question, of course, is whether innovation increases the overall size of the economy. Do the rewards of innovation go only to the winners, at the expense of rivals or of workers whose jobs become unnecessary?
To answer that, Seru and his colleagues used their data on market-moving patents to create an aggregate “innovation index” and then looked at how the index tracked with increases in overall economic growth and productivity.
Again, the results confirmed the basic theory of creative destruction. A significant jump in overall innovative output (in statistical jargon, an increase of “one standard deviation”) correlated to an increase in annual economic output of between 0.6 and 6.5 percentage points. That’s a major boost, given that the U.S. economy usually expands by less than 4% a year.
If the study’s estimated range of that economic impact seems wide, Seru says that’s because the researchers plugged their results into a wide range of models for how the economy works. Using the simplest and most popular economic models, he says, indicates that an innovative surge increases total economic growth by at least 4.5 percentage points. That’s a huge impact.
“What our work shows,” says Seru, “is that Schumpeter was right after all.”