Gordon and Lawrence Summers both make the same mistakeby Anatole Kaletsky / February 18, 2016 / Leave a comment
Published in March 2016 issue of Prospect Magazine
Read the two pieces from last month’s Prospect that Anatole Kaletsky is replying to:
Last month Prospect published two articles about economic growth by two of the world’s most distinguished economists, Robert Gordon and Lawrence Summers. These reminded me of why, 40 years ago, I decided not to become an academic economist and drop out of the Harvard PhD programme where, as it happens, Summers was my fellow student. Instead of shedding light on business and social realities, academic economists seem to prefer to debate among themselves over the intricacies of statistical methods and largely spurious mathematical models. As the joke most famously made by Ronald Reagan runs, an economist is someone who, when he or she sees something working in practice, then wonders if it works in theory too.
A good example of this intellectual distortion is the question whether technological progress has accelerated or slowed to a crawl. For anyone except economists, the answer is so obvious it is hardly worth debating—technological progress is shooting forward. But Robert Gordon assumes that weak economic statistics are proof that, however much new technology we see around us, progress has slowed down.
Gordon is the most influential proponent of the view that he describes as “technological pessimism,” developed in his new 700-page book, The Rise and Fall of American Growth. Summers, a former United States Treasury Secretary, is neither a technological pessimist nor optimist, but believes that the world faces “secular stagnation”: a broad slowdown caused by a combination of changing demographics, globalisation and economic policy mistakes.
These two theories of the world neither contradict nor support one another—they hardly intersect at all. But they have one thing in common. They both fail to address the crucial practical issue—why has the link between technological advances and economic growth apparently broken down?
Gordon claims to provide an answer, but his evidence mainly relates to a totally different question: what happened in the 1970s, after the post-war “Golden Age” of rapid economic growth. The answer is obviously “yes.” Between 1920 and 1970, US output per person grew by 2.41 per cent annually. From 1970 to 2014, the rate was 1.77 per cent.
Gordon explains fairly convincingly why this slowdown happened. Technological change accelerated from 1920 onwards, as the great scientific advances of previous decades were fully incorporated into production processes and new consumer goods. After 1970, the main benefits of electrification, telephony, car and aircraft transport had become almost universal, at least in the US. Hence the post-1970 slowdown.
But what does the slowdown from 1970 onwards—which was also connected with changes in government policies on inflation and unemployment—have to do with the apparent malaise in productivity today?
The “technological pessimism” that now underpins his policy conclusions refers to a slowdown in productivity that suddenly occurred in the past decade, not in 1970. It was only in the past 10 years or so that continuing technological advances such as broadband internet or automated barcode logistics diverged completely from stagnant economic activity.
The decade from 1994 to 2004, by contrast, saw strong productivity growth, almost matching the pre-1970 golden age. Gordon acknowledges that the breakthroughs of the 1990s amounted to a “Third Industrial Revolution” based on information and communications technologies (ICT). But he argues that most of this new technology was based on ideas developed in the 1970s and 1980s, which were fully deployed in businesses by the late 1990s—and the benefits of ICT were suddenly “exhausted” by 2004.
“There is a preference for theory over practice in academic economics”
In jumping to this conclusion Gordon mainly relies on anecdotes: smartphones and tablets have reached saturation in their potential markets; e-commerce accounts for only 6.4 per cent of retail sales; robots cannot match humans in folding laundry; employees waste working time on Facebook; mobile phone keypads have become too small for human fingers.
And strangely for a historian, Gordon loses historical perspective once he reaches the present day. He extrapolates the poor productivity performance that has lasted only 10 years since 2004 into the indefinite future. Yet much of his book’s first half is devoted to debunking the early 20th-century experts who made exactly the same mistake by assuming that the economic benefits of electricity, cars and so on had been largely exhausted by 1918.
What draws one of the world’s top economists into such an ahistorical contradiction? I believe it is the preference for theory over practice in academic economics.
There is, after all, a convincing explanation for the sudden divergence of technological progress and economic growth, which only began in 2008 and is separate from the gradual productivity slowdown. It starts with the mismanagement of macroeconomic policy after the global financial crisis, which prevented a rebound in global demand. This inadequate demand growth has been compounded by the failure of our statistical measures to capture new economic activity during periods of rapid technological change.
Cuts in public spending and tax hikes, motivated by irrational paranoia over public borrowing, have been so severe that it has not been possible to offset their effects through low interest rates. Inadequate demand, combined with labour deregulation and globalisation that would have been healthy if conditions had been normal, have squeezed wages downwards, reducing incentives for investment and aggravating inequality, which in turn has exacerbated the weakness of consumer demand.
To make matters worse, inflation is systematically exaggerated in official figures. If the true level of inflation since 2008 has been negative, as appears quite likely, then even zero interest rates were too high to stimulate rapid growth.
Neither Gordon nor Summers takes these issues seriously because they conflict with deeply embedded theoretical biases. Since the 1980s, academic economics has been dominated by a theoretical paradigm which insists that productivity is independent of macroeconomic conditions and that monetary and fiscal policy cannot affect long-run growth. Beyond that, academic economists have an obvious professional interest in believing that standard statistical methods offer the best available information about what is happening in the real world.
In trying to understand why advancing technology is failing to translate into visible economic growth, economists prefer to deny the practical evidence of progress rather than re-examine their theories. Thus Gordon keeps referring back to the weakness of measured GDP growth since 2008 as clinching evidence of slowing technological progress, instead of an empirical anomaly that probably falsifies his theory about how technology relates to growth. Summers is clearly uncomfortable with Gordon’s conclusions, but can only retort lamely: “It is hard to see how technology could be displacing huge numbers of workers without raising measured productivity… and if innovators are capturing huge [profits from technological advances], why do we not see more evidence of increased output? I do not have compelling answers.”