The riddle of weak productivity

 by Michael Collins, Investment Commentator at Fidelity International

July 2016

Fifty-one years ago one of the tech gurus who helped found US chip-giant Intel wrote an article in which he forecast that computing capabilities would double every year or so while costs would roughly stay constant. Gordon Moore said that cramming more components onto integrated circuits, as his article of 1965 was titled, would lead to “such wonders as home computers … automatic controls for automobiles and personal portable communications equipment”.[1]

Over the four decades following the publication of Moore’s article, the improved economics of computerisation – dubbed “Moore’s Law” after the first person to articulate it widely – led to such leaps in productivity that it spawned an industrial revolution no less. While economists may not agree on much, one thing they are settled on is that productivity, which is defined as output per hour worked, holds sway over long-term living standards. Productivity gains, for instance, explain why one worker’s output in the UK in 2014 was five times that of 1914.[2] Thus the world is a much wealthier place nowadays than it would have been without advancements in microchips, integrated circuits, storage, silicon wafers, software and other achievements of the IT revolution.

The recent decade has witnessed the fulfilment of Moore’s prediction of 1965 that people would one day have computers at home and carry smart phones. The past 10 years have hosted more tech-based innovation and an almost crazy burst of communications tied to the internet, from Snapchat to Facebook. Yet, almost unbelievably, such dynamism has been accompanied by slower productivity growth. The IMF says potential output growth has dropped in advanced economies since the early 2000s and has tumbled in emerging countries since the western financial crisis of 2008.[3] National statistics show similar crunches. US productivity growth has hovered at an annual 1.5% pace since 2009 compared with more than 3.5% from 1995 to 2003,[4] a drop that Janet Yellen, the chair of the Federal Reserve, said in June is “a serious concern”.[5] The Bank of England estimates that UK living standards are 16% lower than they would have been if productivity growth had not slowed since 2009.[6] Australian productivity growth has followed a similar trend of speeding up during the 1990s and slowing since the early 2000s.[7]

Discovering why productivity growth is sinking has short-term investment angles, beyond its long-term drag on living standards. Higher productivity allows an economy to grow at a faster speed without generating inflation. If low productivity persists, central banks, in theory, might need to raise interest rates sooner and by more than people expect to control inflation and governments face a harder struggle to reduce their debt ratios. There are numerous reasons given for the recent decline in output-per-hour growth. One of the most credible explanations might surprise.

Productivity depends on more than just advances in technology, it should be noted. Efficiency gains are tied to the framework for doing business such as secure intellectual property rights, low taxes, incentives for research, unhindered trade, flexible labour markets and anti-monopoly policies. They rely on the institutions that run a country; stable government, a competent bureaucracy, a modern education system, an independent judiciary, sound infrastructure and adequate healthcare. Even outside of industrial revolutions, productivity rises over time because its drivers feed on themselves in a virtuous way – more people become better educated, worker skills get honed and snazzier equipment becomes more widely used. At the same time, and just as importantly for productivity’s endless long-term rise, well-run firms crush their rivals. It can be dangerous to analyse short-term productivity trends because the business cycle can swamp structural shifts. Perhaps today’s drop in productivity is tied to the financial crisis. But sluggish productivity in advanced (and many emerging) countries predates the global financial crisis and has persisted too long for anyone to blame the business cycle.

There is, however, a huge caveat on analysing productivity growth. The methodology used to compile GDP figures might show stagnating productivity when no such problem exists. GDP numbers might undervalue output because an approach that was designed to measure the value of goods underestimates the worth of services, especially those provided by governments and those offered over the internet at no cost. GDP fails to accurately measure the benefits of innovation to consumers, especially the explosion of choice it heralded and the customisation that too much choice spawned. “The extent to which it undermeasures them is extremely large,” writes Diane Coyle is her book, GDP. A brief but affectionate history.[8] But any underestimation of output using GDP methodology also under-estimates longer work days (including all the time spent on work email on public transport) and is probably not large enough to explain most of the stalling in productivity growth. Alan Blinder, a former Fed vice chair, calculates that apps, social media and free internet services would have to be worth US$2.5 trillion a year to nullify the 1.6 percentage-point decline in US productivity growth over the past 10 years. “That’s not believable,” he says.[9] For context, the annual output of the US is estimated at about US$17.5 trillion.

Recession effects

How can productivity growth stall during an IT revolution? Economists see that productivity growth is derived from three sources: the extent to which firms use their capital and labour, the amount of capital invested per worker, and the usefulness of innovation.

The first element – the extent to which firms exploit their resources – is essentially cyclical. Companies look to muster their resources when demand is solid and conserve them when conditions darken. During downturns, productivity growth is hampered if firms keep their workforce but reduce production. The most usual reason companies preserve staff numbers is that they expect the drop in demand to be fleeting. They, therefore, don’t want to suffer the expense and hassle of firing staff and, shortly after, rehiring and training others. Another way productivity might drop during a recession is if companies steer resources away from production into business development in the hope of a longer-term payoff.

If the slowing in productivity growth is cyclical, then the issue should right itself when demand improves. This explanation doesn’t suffice, though, for the weakness in productivity in the advanced world has endured for too long for it to be cyclical. While the Bank of England concedes that productivity growth is “procyclical”, it says the drop in the UK’s output per hour since 2010 is well below its normal relationship to GDP.[10]

If it’s not cyclical, then perhaps it’s more structural changes that explain the drop in productivity. The first source of productivity to look at here is the amount of capital invested per worker. Has the recession from 2008 (which Australia escaped) changed business behaviour in a detrimental way? It seems to have done so by making businesses more wary of investing, a problem exacerbated in countries where credit was restricted. Business investment in advanced economies has only averaged 20.7% of GDP since 2010 (having sunk to 19.5% in 2009) compared with 23.6% of output during the 1990s.[11]

Another structural explanation for the decline in productivity could be that infrastructure in the advanced world has worn out to such an extent it is hampering the efficient allocation of resources. Traffic jams, for instance, slow rail or gridlocked waterfronts that hinder the flow of goods lower productivity. Another more interesting explanation for a structural decline in productivity, and one proffered by the Bank of England, is that low interest rates are trimming productivity growth because they enable inefficient companies to survive – to the benefit of employment, output and social welfare it should be pointed out. (Higher productivity is thus not an absolute good in all circumstances.) Data from the US supports the thinking that economies are more lethargic these days as business dynamism – the rate at which jobs are created and vanish as companies come and go – is at a 30-year low, as measured by the US Census Bureau’s business dynamics report.[12]

The unlikely problem

The issues surrounding the third source of productivity growth – where innovation sits – are perhaps even more interesting. Economists assess innovation’s influence on output by accrediting to it any increases in productivity that can’t be attributed to traditional capital and labour inputs. This measure of innovation, which is called total factor productivity, is broadly defined to capture inventions and other improvements such as gains from new infrastructure and better work practices.

Analysis by the Federal Reserve Board of San Francisco shows declines in total-factor-productivity gains show innovation is less useful these days when it comes to boosting output. Numerous studies, including those that allow for cyclical impacts on productivity growth, conclude that total factor productivity has stumbled since the early 1990s. Amazingly, the biggest slump in total factor productivity is in industries that produce IT or were the most wired to IT. The conclusion? “More recent gains from IT might have been more incremental than transformative,” it says.[13]

While not all studies show technology industries give poorer readings on total factor productivity, it’s credible to claim that the IT revolution is nearing the end of its usefulness in boosting living standards. The improvements in business practices through the 1980s and 1990s propelled by leaps in technology such as instant communications have been implemented and can’t be honed much more. It’s a bit like how Formula One racing cars don’t change much these days – for all the millions of dollars spent on research – for they have reached the boundaries of aerodynamics. In fact, the innovations over the past decade could even have hampered productivity. Firewalls to ward off viruses slow work systems. Perhaps people bludge when working remotely. Internet access on office computers may well be distracting, or as Blinder says it might turn “formerly productive work hours into disguised leisure”.[14] Some technology businesses such as and LinkedIn are even winding back on simple PowerPoint technology because they have found that talking, eye contact and simple memos are better ways to communicate complex messages.[15] It could be that businesses have succumbed to the counterproductive “neo-mania” that Nassim Taleb of The Black Swan fame about improbable events warns about in his later book Antifragile. This “love of the modern for its own sake” … “makes us build Black Swans”.[16]

What can policy makers do if the IT revolution has lost its sparkle for living standards? They will have to work on all the other aspects that drive productivity. They will need to raise education standards, fix transport, advance competition policies, encourage risk-taking and toughen adherence to intellectual property rights. Those on the right side of politics will press for more business-friendly labour laws, even though their opponents will contend that their motive is to tilt rewards towards capital at the expense of labour rather than to bolster efficiency. Lifting productivity growth will be a slow, hard and contentious slog in a world where Moore’s Law is fading.

Some of the explanations for the drop in productivity growth have come the Bank of England’s analyst “The UK productivity puzzle”, found in its Quarterly Bulletin 2014 Q4. Other financial information comes from Bloomberg unless stated otherwise.

Important information

References to specific securities should not be taken as recommendations.


[1] Gordon E. Moore. “Cramming more components onto integrated circuits.” Electronics, Volume 38. 19 April 1965.

[2] Bank of England. “The UK productivity puzzle: an international perspective.” Martin Weale. 8 December 2014.

[3] IMF. World Economic Outlook April 2015. Chapter 3. “Where are we headed? Perspectives on potential output.” Page 71.

[4] Federal Reserve Board of San Francisco. “The recent rise and fall of rapid productivity growth”. John Fernand and Bing Wang. FRBSF Economic Letter. 9 February 2015.

[5] Bloomberg News. “Yellen productivity concern shows slump’s theory-to-policy shift”. 22 June 2016.

[6] Bank of England. “The UK productivity puzzle.” Quarterly Bulletin 2014 Q2.

[7] The Australian Bureau of Statistics measures productivity in its National Accounts among its labour market indicators by its  GDP per hour worked index. Trend series A23004328L.

[8] Diane Coyle. “GDP. A brief but affectionate history.” Princeton University Press. 2014. Page 124.

[9] Alan Blinder. “The mystery of declining productivity growth.” The Wall Street Journal. 14 May 2015.

[10] Bank of England. “The UK productivity puzzle.” Quarterly Bulletin 2014 Q2.

[11] IMF. “World Economic Outlook: Uneven growth – short and long-term forecasts.” Chapter 4. Private investment: What’s the holdup?” Page 116. April 2015.

[12] US Census Bureau. Business Dynamics Statistics. 2012 release. September 2014. The reallocation rate of workers was 24.2 in 2012, the last year available, and 24.0 in 2011, the lowest numbers since the survey began in 1976.

[13] “The recent rise and fall of rapid productivity growth.” John Fernald and Bing Wang. FRBSF Economic Letter 201504. 9 February 2015. 

[14] Blinder. Op. cit.

[15] “PowerPoint should be banned. This PowerPoint presentation explains why.” Katrin Park. Washington Post. 26 May 2015.

[16] Nassim Nicholas Taleb. “Antfragile. How to live in a world we don’t understand.” Allen Lane. 2012. Pages 312 and 7.