The End of innovation (Part 2)?: A new hope

At the end of my last column the prospects for the future looked bleak. If the standard theory of economic growth is correct, then: i) R&D spending produces ideas; and ii) ideas lead to TFP growth. In this case, the fact that R&D spending has been increasing while TFP growth has been falling necessarily implies that ideas are getting harder to find. And if ideas are getting harder to find, then innovation led economic growth is going to peter out — with all of the terrible consequences that this entails.

Fortunately, new research indicates that the standard view of how innovation happens is incomplete. So, before we conclude that innovation is over, we need a deeper understanding of how innovation happens.

Think, for example, about the wonder drug penicillin. Alexander Fleming famously discovered the idea of penicillin in his lab in 1928. But, this discovery remained a mere scientific curiosity for almost two decades because the method that Fleming used to isolate penicillin in his lab could not be scaled up to produce medicine. It took years of engineering by research labs and pharmaceutical firms to process the idea of penicillin into an actual pharmaceutical innovation.

The example of penicillin shows that it is processed ideas — not ideas alone — that create the innovations that drive TFP growth. The decline in TFP growth that we observe could then be due to either a shortage of ideas (the standard view) or to a shortage of the idea processing capability required to turn a plentiful supply of ideas into actual innovations.

Processing ideas takes long time horizons.

A fascinating study of innovation in science by Jay Bhattacharya and Mikko Packalen illustrate this point perfectly. They find that while scientists want to engage in innovative science, they also need to attract outside support to pursue their careers (for example, a grant and/or an academic position). This leaves scientists with a difficult choice. They can choose longer horizon projects that do produce innovative science but that don’t provide immediate feedback to universities or foundations. Or, they can choose quick-win projects that demonstrate their skill to those outside parties but which are not very innovative. Scientists have been shifting from innovative strategies to quick-win strategies over time (perhaps due to increased competition for funding and positions), and the progress of science has stagnated as a result.

Firms too must choose between longer horizon innovative projects and shorter horizon quick-win projects. And, just as scientists have to choose their projects with an eye towards what universities or foundations will reward, firms must choose their projects with on eye towards what financial markets will reward.

If financial markets work well — that is, if firm accounts credibly disclose key information about the firm, if corporate governance arrangements align the interests of management and shareholders, if markets are clean and a firm’s stock price accurately reflects its value, etc. — then firms can convince the market that a long horizon approach makes sense. But if markets work poorly and it is very hard for shareholders and other stakeholders to tell whether or not a firm is wasting money, it will be hard to get shareholder and/or stakeholder buy-in for a long horizon approach. In this case, shareholders and stakeholders will push firms to choose a quick-win “show me the money” approach instead.

So, financial market effectiveness will have a significant impact upon a firm’s choice between a long horizon and a quick-win approach. Since only firms with long horizon approaches develop the capability to process ideas into innovations, financial market effectiveness will play a key role in determining an economy’s idea processing capability.

Now, if ideas are in plentiful supply, then average TFP growth will be determined by the economy’s ability to process those ideas into innovations. Idea processing capability in turn is a function of financial market effectiveness. So, if this is the case, then the evolution of average TFP growth over time will be a function of financial market effectiveness (and not idea supply).

To test this proposition, Akshay Kotak, Dimitri Tsomocos, and I devise a way to track financial market effectiveness in the US over the 1899/2019 period. We find that financial market effectiveness is low in the essentially unregulated period of 1899/1930. The stock market crash of 1929 led to a major reform drive in the 1930s, and this effort produced the highly effective Post-WW2 markets of 1951/1970. Starting in the 1970s, market effectiveness began to decline (we think because regulation did not keep up with market developments). By 1980, market effectiveness fell back to its level in the unregulated 1899/2019 period. This period of low market effectiveness has lasted to the present day.

If market effectiveness determines average TFP growth, then average TFP growth will rise and fall with market effectiveness over time. As one can see in Table below, this is precisely what happens.

The idea processing theory of TFP growth the only theory that can explain the evolution of average TFP growth over time. So, it is at least plausibly the case that TFP growth has fallen in the US due to a decline in idea processing capability rather than because we are running out of ideas.

Of course, this result does not change the fact that TFP growth is still depressingly low. But it does open up new ways to attack the problem of low growth.

I will explore the implications of idea processing for growth policy in my next column.

{{{short}}}
{{#more}}
… Read More
{{/more}}

{{/totalcount}}
{{^totalcount}}

Start a Conversation