Are we spending enough on R&D?
|Joachim Klement||May 31, 2019|
As a German, I am obsessed with productivity. On a personal level, I optimise my work and personal environment to be as productive during the day as I possibly can be. And as an economist, I am fascinated by the concept of productivity and its contribution to economic growth. As the more economically interested amongst our readers may know, productivity has largely been stable in the UK as well as in most other developed markets since the end of the Second World War but has declined markedly since the end of the Global Financial Crisis.
The reasons for the decline in productivity over the last ten years have probably been manifold, but a paper by Nicholas Bloom and his colleagues at Stanford University has shown that the problem might be far more fundamental than is comfortable. Traditionally, one assumes that productivity increases are a result of technological progress. And technological progress should be linked to the amount we spend on research and development (R&D) both in the private and the public sector (i.e. universities).
If a company like Apple has a business model based on technological and design innovation, it better spent a lot on R&D in order to deliver these innovations. However, the company needs to spend probably far more on R&D each year than management and shareholders are willing to accept because the paper by Bloom and his colleagues shows that the amount of R&D investments needed to get the same increase in technological innovation is rising at alarming speeds. Or to put it the other way around: new innovations seem to be harder and harder to find, as the example of Apple shows all too clearly. After all, even though Apple this week launched a new streaming service, this is hardly an innovation. And Apple has not presented any major product innovation for about ten years.
But Apple is just an anecdote and the plural of anecdotes is not necessarily data. So, let’s focus on the data on productivity growth and technological progress for different industries. Our chart shows the development of technological progress for two core industries. In the semiconductor industry, Moore’s Law famously states that every two years the processing power of a computer chip roughly doubles. However, we observe that the growth in computing power per billion US Dollars invested in R&D has consistently declined over the last three decades. IT companies have to spend roughly 18% more on R&D per year to keep doubling computing power every two years. Instead, the R&D budgets of semiconductor companies worldwide have been growing at just 8% per year between 2010 and 2015. Given this underinvestment in R&D it should come as no surprise that productivity growth is declining in the IT industry.
The second example we show in our chart is the number of new patented drugs per billion US Dollars spent on R&D in the pharmaceutical industry. Again, the same trend is visible. In order to develop a new effective drug that can be patented, more and more money has to be spent on R&D. In order to keep the number of newly-patented drugs per year constant, pharma R&D budgets need to grow by about 9.8% per year. Instead, they have grown at an annual rate of 3% between 2010 and 2015.
That private companies invest insufficient amounts of money in R&D is not surprising giving the pressure from shareholders to return excess profits in the form of dividends and share buybacks, and the focus on shorter-term earnings growth in favour of long-term growth. This pressure from shareholders slowly erodes investments in R&D and other long-term drivers of productivity growth and as a result slowly erodes economic growth and prosperity in developed countries.
Productivity of research and development spending
Bloom et al. (2019), Fidante Capital.