The secular stagnation discussion has a bad title and sounds wonkish but is extremely important – it is about nothing short of understanding what the economy is and should be about. If you want a lot of detail you can read through Marc Andreessen’s epic secular stagnation tweet storm and all the great source material that he links to.

At the heart of the secular stagnation discussion is the idea that somehow the economy is growing more slowly than it should. See for instance the following quote from a Larry Summers speech:

[…] through this recovery, we have made no progress in restoring GDP to its potential.

The rest of the speech goes on exploring the reasons for why this might be so with a particular focus on interest rates.

There is, however, another line of inquiry that is much more important and that is to question the very premise here by asking whether GDP makes sense as a measure for the economy going forward.

Let me give an analogy from the business world. Suppose you are newspaper and the KPI you are using to manage your business is your circulation. That works really well for a great many years. Then all of a sudden you are in board meetings that show a lack of circulation growth and possibly even a decline. You keep arguing about the reasons for why your management has “made no progress in restoring circulation to its potential.” The debate rages on all the while you completely ignore that your newspaper also has a website on which traffic has been growing steadily.

That is what is happening with the economy at large. We are de-materializing it and moving from a world of easily measured and priced atoms to a world of bits which are creating massive consumer surplus. So for instance: we stop printing and selling the Encyclopedia Brittanica because everyone is using Wikipedia instead – that shows up as a decrease in GDP even though many more people now have access and get the benefit of knowledge. If we take dozens of different devices that were all sold separately (film camera, video camera, maps, books, photo albums, pencils, tape recorders, etc) and combine them into a single device (the smartphone) that uses primarily software to accomplish these functions that too can easily result in a decrease in GDP. In a talk I am about to give at DLD in Munich I will give even more examples of this but you get the idea.

By the way, GDP was always a suspect KPI for the economy because of the challenge of negative externalities: if we make products that harm people (eg selling cigarettes that cause lung cancer) and then sell other products to correct the harm (eg lung cancer treatment drugs) both of these increase measured GDP. Again, there are tons of other examples like that.

To be clear, there are lots of intelligent discussions to be had about phenomena such as reductions in risk premia, the role of China and NAFTA, the influence of household deleveraging, etc. and their impact on the economy. But until we are willing to ditch GDP as the correct KPI for the economy we will be like the newspaper management team trying desperate measures to fix their circulation problem.

You can see me talk about all of this in a presentation recorded last May at DLD New York. In the coming days I will write more about this topic including the impact of technology driving down prices. Chris Dixon tweeted a great chart and you should check out my posts on education and healthcare. Imagine for a moment a computer than can correctly diagnose patients within seconds and then think about the implications for GDP.

I understand that this is a partial equilibrium argument and will expand on how it can be true for the economy as a whole. As part of that I will also suggest that our obsession with full employment – which is also part of the secular stagnation discussion – is also problematic as a KPI. Stay tuned.

Addendum 1: Turns out that Erik Brynjolfsson has given a great talk about the importance of free goods at Techonomy and has a working paper aimed at quantifying these effects (a couple % of measured GDP).

Addendum 2: One other important criticism of GDP as a KPI is that it is silent on distribution. We have seen major changes in that (a) away from labor toward capital and (b) within labor from normal distributions towards power laws (eg CEO comp relative to employees).

The secular stagnation discussion has a bad title and sounds wonkish but is extremely important – it is about nothing short of understanding what the economy is and should be about. If you want a lot of detail you can read through Marc Andreessen’s epic secular stagnation tweet storm and all the great source material that he links to.

At the heart of the secular stagnation discussion is the idea that somehow the economy is growing more slowly than it should. See for instance the following quote from a Larry Summers speech:

[…] through this recovery, we have made no progress in restoring GDP to its potential.

The rest of the speech goes on exploring the reasons for why this might be so with a particular focus on interest rates.

There is, however, another line of inquiry that is much more important and that is to question the very premise here by asking whether GDP makes sense as a measure for the economy going forward.

Let me give an analogy from the business world. Suppose you are newspaper and the KPI you are using to manage your business is your circulation. That works really well for a great many years. Then all of a sudden you are in board meetings that show a lack of circulation growth and possibly even a decline. You keep arguing about the reasons for why your management has “made no progress in restoring circulation to its potential.” The debate rages on all the while you completely ignore that your newspaper also has a website on which traffic has been growing steadily.

That is what is happening with the economy at large. We are de-materializing it and moving from a world of easily measured and priced atoms to a world of bits which are creating massive consumer surplus. So for instance: we stop printing and selling the Encyclopedia Brittanica because everyone is using Wikipedia instead – that shows up as a decrease in GDP even though many more people now have access and get the benefit of knowledge. If we take dozens of different devices that were all sold separately (film camera, video camera, maps, books, photo albums, pencils, tape recorders, etc) and combine them into a single device (the smartphone) that uses primarily software to accomplish these functions that too can easily result in a decrease in GDP. In a talk I am about to give at DLD in Munich I will give even more examples of this but you get the idea.

By the way, GDP was always a suspect KPI for the economy because of the challenge of negative externalities: if we make products that harm people (eg selling cigarettes that cause lung cancer) and then sell other products to correct the harm (eg lung cancer treatment drugs) both of these increase measured GDP. Again, there are tons of other examples like that.

To be clear, there are lots of intelligent discussions to be had about phenomena such as reductions in risk premia, the role of China and NAFTA, the influence of household deleveraging, etc. and their impact on the economy. But until we are willing to ditch GDP as the correct KPI for the economy we will be like the newspaper management team trying desperate measures to fix their circulation problem.

You can see me talk about all of this in a presentation recorded last May at DLD New York. In the coming days I will write more about this topic including the impact of technology driving down prices. Chris Dixon tweeted a great chart and you should check out my posts on education and healthcare. Imagine for a moment a computer than can correctly diagnose patients within seconds and then think about the implications for GDP.

I understand that this is a partial equilibrium argument and will expand on how it can be true for the economy as a whole. As part of that I will also suggest that our obsession with full employment – which is also part of the secular stagnation discussion – is also problematic as a KPI. Stay tuned.

Addendum 1: Turns out that Erik Brynjolfsson has given a great talk about the importance of free goods at Techonomy and has a working paper aimed at quantifying these effects (a couple % of measured GDP).

Addendum 2: One other important criticism of GDP as a KPI is that it is silent on distribution. We have seen major changes in that (a) away from labor toward capital and (b) within labor from normal distributions towards power laws (eg CEO comp relative to employees).