Having looked at demographics last week to kick off our mini-series of big picture thoughts, we wish to move onto Productivity this week. Simply put, in the long-term, population growth plus productivity growth equals Real GDP growth. Below, we will look at productivity growth in recent decades drawing upon both our own work and that of Andrew Smithers, and then consider a thought provoking paper by Robert Gordon and why productivity growth maybe set to fall further in the decades ahead.
It has been an eventful week in markets and we will be sending our usual market commentary tomorrow.
Productivity growth, or the increase in real output per person, has been declining since at least the 1970s. This is illustrated in chart 1 below showing Real GDP per person employed for major developed economies.
Chart 1 – Real GDP per person employed (courtesy of Andrew Smithers)
Long term improvements in productivity are driven by the amount of capital invested and the efficiency of that capital. Chart 2 below shows the amount of Real capital investment per person employed for the same countries. As can be seen, the amount of capital invested grew strongly until the 2000s, and has broadly stagnated since. What is obvious from thinking about charts 1 and 2 is that productivity growth weakened between the 1970s and 2000 even at a time when the amount of capital invested rose. What happened is that the efficiency of the capital invested deteriorated over that time. In the last 10 to 15 years, there has not been enough new capital investment, and with the efficiency of new capital deteriorating again (chart 3), productivity growth remains near the lowest in post WWII data.
Chart 2 – Real capital investment per person employed (courtesy of Andrew Smithers)
The way to measure the efficiency of new capital invested is to measure how much new capital is required to generate an increase in productivity. Economists know this as the Incremental Capital Output Ratio (ICOR), and this is illustrated in chart 3 below. Just to be clear on this point, a higher number here illustrates less efficiency, and as can be seen, despite a brief improvement in the 1980s (after the disastrous economic performance in the 1970s), and again during the internet revolution between 1996 and 2004, the trend since the 1970s has been that it takes increasingly more capital to generate a given increase in productivity.
Chart 3 – Incremental Capital Output Ratios (courtesy of Andrew Smithers)
To sum up at this stage, we think that it is clear from the above that unless the efficiency of new capital investment improves, and the amount of new capital investment increases, then productivity growth will remain at low levels. We would also point out that, at the moment, capital efficiency is still deteriorating and new capital investment per person has not grown in the last decade or so. As such, we do not see any structural reason to expect any reasonable increase in productivity as measured by real output per person employed.
Now, let’s move onto something that caused a bit of a stir back in 2012, which was the publication by Robert Gordon of a paper questioning whether US economic growth is over. Basically, modern economics believes that economic growth is a continuous process that will persist forever. That said, academics and historians have illustrated how there was barely any real economic growth in the centuries before 1750, and Robert Gordon’s paper speculates whether the rapid progress made over the past 250 years will turn out to be a unique episode in human history.
The main thrust of his paper looks at the huge productivity gains made because of the Industrial Revolution (IR). He labels three phases of industrial revolution. IR#1 (steam, railroads) from 1750 to 1830; IR#2 (electricity, internal combustion engine, running water, indoor toilets, communications, entertainment, chemicals, petroleum) from 1870 to 1900 and IR#3 (computers, the web, mobile phones) from 1960 to present.
The paper provides evidence that IR#2 was more important than the others and was largely responsible for 80 years of relatively rapid productivity growth between 1890 and 1972. Once the spin-off inventions from IR#2 (airplanes, air conditioning, interstate highways) had run their course, productivity growth during 1972-96 was much slower than before. IR#3 created only a short-lived growth revival between 1996 and 2004. Many of the original and spin-off inventions of IR#2 could only happen once – urbanisation, transportation speed, the freedom of females from the drudgery of carrying tons of water per year, and the role of central heating and air conditioning in achieving a year-round constant temperature.
Graphically, the phases of productivity growth of the last 122 years are shown in the chart below from a follow up paper written by Gordon in 2014. The chart shows productivity (green bars), output per capita (red bars) and hours worked per capita (grey bars) for the discrete periods 1891 - 1972, 1972 - 1996, 1996 - 2004 and 2004 – 2013. As we are only looking at productivity this week, let’s focus on the green bars.
Gordon shows how productivity growth fell from 2.36% per annum between 1891 and 1972 to only 1.38% between 1972 and 1996. This is fact as confirmed by Andrew Smithers work, however, Gordon proposes that productivity declined as the benefits of IR#2 dissipated. Productivity briefly, and quite impressively, improved between 1996 and 2004 because of the internet. But relapsed back to 1972 – 1996 levels after 2004 as the benefits of the internet (e-mail, e-commerce, the web) proved short lived.
Chart 4 – Annualised growth in productivity, output and hours worked per capita (courtesy of Robert Gordon)
So, having looked at productivity from two perspectives, we can see that there has undoubtedly been a big slowdown in productivity growth since the 1970s. Andrew Smithers sets out how both insufficient capital investment and inefficient capital itself is a problem. It is our view that neither will improve any time soon. Robert Gordon has looked at productivity as part of a structural thesis about the demise of US economic growth, and within that, how productivity benefitted massively from the inventions of the second industrial revolution and the resulting spin-off inventions. Post 1972, productivity growth has declined sharply, except for a brief period driven by the internet revolution between 1996 and 2004. Gordon sets out what we believe to be a credible argument for why productivity and broad economic growth will falter in the decades ahead.
To try and balance the negative angle presented here for a moment, we need to ask whether there is potentially a technology revolution out there in the next decade or so that will lift productivity growth back to the 2%+ path seen prior to the 1970s. For the moment, we just don’t see one. Yes, there are some very exciting advances in the field of medicine that will help us live longer. If politicians can then raise the retirement age as we continue to live longer, then that will boost growth (via the hours worked per capita more so than productivity per se), however, there will be huge pushback to later retirement from some groups. A lot of the technology innovation of the last decade or so is nice to have stuff, but can Apps, mobile devices, social media, streaming and online commerce increase the productivity of our workforce in the same way the inventions and spin-offs of IR#2 did? Not in our opinion.
So, with the demographic argument made last week, and the productivity argument made this week, we believe that potential US real economic growth in the decades ahead will be less than that seen in the post WWII era to date. Put another way, with most economists and Central Bankers struggling to understand why growth remains sub-standard, we note that demographics (which were a strong tailwind between 1980 and 2007) is increasingly becoming a structural headwind. Productivity growth (aside from the internet revolution boost between 1996 and 2004) is now also acting as a structural headwind and could deteriorate a bit further in the years ahead.
What can policymakers do to overcome the demographic and productivity headwinds? Well, central bankers really cannot do much at all. Zero interest rates and trillions of dollars of QE have not encouraged companies to go on a capex boom (although US companies have gone on a borrowing binge to finance share buybacks). These policies will not create the equivalent of an internet revolution, never mind an IR#2. Will politicians come to the rescue and enact policies that will reduce red tape so that innovation can breathe more freely, leading to future gains in productivity? Not likely in our opinion.
Frankly, without some sort of technology advancement of the scale of IR#2, we see a slow economic growth coupled with low inflation as the likely economic environment for a number of years. If we are correct, then we all need to tone down our expectations for returns from financial assets on any reason time horizon for buy and hold investors.
Next week, we will have a look at another structural headwind in the form of debt and begin to relate all of these structural issues with current monetary policy. As a quick heads up (and it’s been quite a week for policymakers with Draghi promising another huge round of stimulus and China cutting interest rates for the 6th time in the last 12 months and reserve requirements for the 4th time this year), it is increasingly clear that monetary policy is becoming less effective in bringing forward a small amount of future incremental demand. Furthermore, it is increasingly our view that zero/negative rates and QE are actually deflationary and the unintended consequences of increasingly aggressive policies will be extremely unpleasant when they become apparent.