We all know that technological innovations are crucial for prosperity, and they’re changing our lives unprecedentedly. But how can we tell if this pace is faster than during Thomas Edison’s time? Some economists claim that today’s information revolution has a much smaller impact on our daily lives than the inventions in the late 19th century.
It’s challenging to separate genuinely epic inventions from those that are trivial or incremental, making it impossible to compare historical innovation trends. This makes it difficult to determine the conditions or policies leading to more breakthroughs.
Now, a group of researchers led by Stanford Graduate School of Business Professor Amit Seru has developed a novel approach that uses big data computing on several million patent documents to rank and identify the historical spikes of epic inventions.
Silicon Valley enthusiasts will be pleased to know that electronics and communications have dominated the industry during the last two decades.
Combustion engines, sewing machines, and elevators
As measured by the new yardstick, the most extensive boom was in the early-mid-1800s, when inventions revolutionized transportation, manufacturing, and the character of large cities. Charles Goodyear invented vulcanized rubber, used in tires, and improved many industrial products. Elisha Otis invented the elevator, allowing skyscrapers to be built, and helped create modern cities. Elias Howe invented the sewing machine, which revolutionized the garment industry.
The second wave of innovation began in the late 1800s. In that period, the most significant breakthroughs were the telephone (1876), the internal combustion motor (1877), the incandescent bulb (1880), the mechanical calculator (1888), and the first electric engine to run on alternate currents. Orville Wright received his patent in 1906 for the plane, which was a significant breakthrough, according to this study.
Seru says that the actual value of this new paper is that it offers a new way to measure innovation trends, which opens up new avenues for understanding economic forces and policies that could encourage more innovation.
Seru says that although economists agree about the importance of technological progress, we do not have any suitable methods of measuring it. This is especially true over a longer time horizon. Measuring technological innovation is crucial to see how policy changes affect large-scale changes. Does the right amount of creation take place? Does it shift from big to small firms or vice versa? Are they coming from private or public companies? Are there any shifts towards or away from government agencies or universities? Our metric allows us to answer these questions in first order.
Searching for Patents with Recurring Terms
The idea behind this new approach is simple: a vital invention is a product that differs from previous products and greatly influences future ones. Researchers searched for rare terms and phrases in older patents that appeared frequently in later ones to find these inventions.
The challenge of computing was enormous. Researchers had to find essential words in 9 million patents containing thousands of words. They then had to determine how often these terms appeared in the patents of previous and following years. The large “correlation matrix used to calculate the results made this task very computationally demanding. The ratio between these two measures was used to gauge the importance of a patent.
Seru claims that the new approach to measuring innovation has many advantages over other strategies. Counting the number of patents cited in later patents as “prior arts” has proven to be one of the most popular strategies. This method is flawed because patents did not include these citations consistently until the 1940s. As a result, they are less useful for analyzing historical trends or answering “first-order questions.”
A newer approach that Seru developed is based upon the increase in stock prices of a company when it receives a patent. According to Seru’s research, this metric is a good measure of the actual value of technology. However, it only applies to publicly traded companies, as no stocks are available for individuals, privately held firms, universities, or government agencies.
Aligned with historians
The researchers discovered that their textual-based measures of patent quality correlated well with other measures, indicating that their phrase-based approach is reliable.
The researchers also discovered that the top patents they ranked matched well with historians’ assessments. Researchers found that when they compared their ranking of the importance of patents to a list of 110 of the most influential American inventions from the late 1800s through the early 1960s – 40% of them ranked within the top 10%.
The top 1% of American inventions included many of the most famous ones. Some of these were obvious, like Thomas Edison’s electric lights, Alexander Graham Bell’s telephone, and Wright’s airplane. They also included lesser-known breakthroughs such as Gail Borden’s condensed milk and the safety pin.
Robert Noyce, Edwin Armstrong, and Philo T. Farnsworth are all within the top 10 percent of the list.
Researchers also found quantitative evidence that technological innovation boosts economic productivity. It may seem obvious, but in previous studies, predicting productivity using different technological innovation metrics was difficult. Seru says that, unlike previous studies, textual-based metrics could allow researchers to get a more accurate measure of technological innovations, especially breakthroughs.
Seru says the big news is the new ranking system for essential inventions. It appears accurate and allows us to understand better the conditions under which true innovation can flourish.
