Zakaj nas čaka dolgo obdobje nizke rasti

Medtem ko je lanska napoved Larryja Summersa o obdobju sekularne stagnacije (dolgo obdobje nizke rasti) naletela na dokaj razumevajoč sprejem v akademski sceni, pa je raziskava Roberta Gordona (Northwestern University) “Is U. S. Economic Growth Over? Faltering Innovation and the Six Headwinds” leto pred njim (avgust 2012) naletela na, milo rečeno, zelo kontroverzen sprejem. Jasno, Summers je stopil na pripravljen teren, medtem ko je bil zelo kritičen sprejem Gordonove raziskave predvsem posledica dejstva, da je bil prvi, ki je objavil tako šokantno napoved. Kritike so se vrstile, Gordon pa z novo, pravkar izdano raziskavo “The Demise of U.S. Economic Growth: Restatement, Rebuttal, and Reflections” (februar 2014) še utrjuje tedanjo napoved.

Razlogi so po njegovem v padajočih donosih zaradi močnih “protivetrov”, med njimi pa predvsem v padajočih donosih polpreteklih, sedanjih in bodočih inovacij oziroma natančneje v nizki mejni koristnosti zadnjih inovacij. Gordon to plastično opiše s primerom inovacij pametnih telefonov in tablic: če morajo biti v ekonomiji mejne koristi izenačene z mejnimi stroški, bo tudi mejna koristnost pametnih telefonov zaradi njihove cenenosti zelo nizka. Podobno velja za inovacije na področju 3-D printerjev, “big data” itd. Obdobje nizke rasti se je po Gordonovem mnenju sicer začelo že po letu 1972, zelo občutno pa se kaže po zadnji krizi po letu 2007, kar ponazarja s spodnjo sliko:

Gordon_low growth aheadVir: Robert Gordon (2014)

Historično je ameriška rast BDP per capita v obdobju 1891-2007 znašala v povprečju 2% letno, prihodnja rast do leta 2077 pa naj bi se ob upoštevanju štirih “protivetrov” (pozitivnega učinka demografije, višje izobrazbe, manjše neenakosti, zmanjšanja dolga) upočasnila na vsega 0.8% letno. Če pa odštejemo še “protiveter” zmanjšane koristnosti inovacij, pa naj bi se letna stopnja prihodnje rasti za spodnjih 99% prebivalstva zmanjšala na vsega 0.2% (glej sliko spodaj).

Gordon_growth without headwindsVir: Robert Gordon (2014)

Preden se “techno geeki” skregate s tem skeptičnimi videnjem prihodnosti, predlagam, da preberete celoten članek Roberta Gordona. Predvsem pa historično ilustrirano diskusijo o tem ali je inovacije mogoče napovedati vnaprej ter kakšen je potencial zadnjih inovacij za rast produktivnosti. Spodaj navajam nekaj odstavkov uvodnega zapisa Gordona. Da vas vpeljem v nadaljevanje teksta.

A controversy about the future of U.S. economic growth was ignited by my paper released in late summer 2012. The debate began with my prediction that over some indefinite period of time into the future, perhaps 25 to 40 years, the growth of real per-capita disposable income of the bottom 99 percent of the U.S. income distribution would average 0.2 percent per year, compared to 2.0 percent per year in the century before 2007.2 This prediction set off a firestorm of controversy with commentary, blogs, and op-eds around the world.

That paper was explicitly about the U.S. and recognized that there is plenty of room for “catch-up growth” in the emerging markets of the world. The growth concept is “potential” real GDP per capita from 2007, not actual real GDP from 2013. It attributed the slowness of future growth mainly to such “headwinds” as demography, education, inequality, and government debt. The subtitle “Faltering Innovation and the Six Headwinds” seemed to place primary emphasis on a future world with no innovation, and the primary role of the headwinds in predicting slow future growth escaped notice in the initial round of controversy about innovation.

This sequel paper recasts the debate to focus on the headwinds, which are formidable for the U.S. economy and may be less important in some other nations. In this sequel, there is no need to forecast that innovation in the future will “falter,” because the slowdown in the rate of productivity growth over the past 120 years already occurred more than four decades ago. This sequel paper explains why the pace of innovation declined after 1972. The future forecast assumes that innovations in the next 40 years will be developed at the same pace as the last four decades, but reasons for skepticism are provided for that prediction.

The epochal set of inventions that occurred between 1870 and 1900, with continuing benefits to 1972, represent the fruits of the “Second Industrial Revolution” (IR #2). The growth rate of American productivity in the eight decades before 1972 was 2.36 percent per year, compared to 1.59 percent per year since 1972. That permanent decline of 0.8 percent per year represents a new way to measure the extent to which the single-dimension digital “Third Industrial Revolution” (IR #3) has fallen short of the multi-dimensional IR #2.

This sequel paper compares the actual condition of the current 2013 U.S. economy with both optimistic and pessimistic future growth paths. By any measure, the U. S. economy continues to operate below even the most pessimistic projected post-2007 growth path, and we shall see that real per-capita income (for everyone, not just the bottom 99%) in 2013 was only one percent higher than six years earlier.

In addition to providing a new interpretation of the predicted future post-2007 path of potential real GDP per capita, this sequel paper assesses several different criticisms directed at the earlier paper. All of these issues involve future innovation, and none of the critics have denied that the headwinds facing the U.S. economy are formidable. Rather, the headwinds have largely been ignored by the critics who concur that any skepticism about future innovation reflects a lack of imagination.

An initial round of criticism pointed to the impossibility of forecasting future innovations. This paper counters with five historical examples from 1863, 1875, 1900, 1939, and 1949 in which many of the most important innovations of the following 50 to 100 years were accurately forecast and sketched out in substantial detail. In the same way, every techno-optimist has a list of revolutionary inventions that will occur over the next decade or two, including medical advances, small robots, big data, driverless cars, and oil-gas fracking. It is legitimate to compare today’s forecasts of future inventions with those late 19th century inventions that were accurately forecast in 1863 or 1875. Which are more important?

The set of criticisms based on a faith in the future of technology can be simplified if we use the term “techno-optimists” to describe those who foresee an acceleration in the pace of innovation to a rate faster than has occurred in the past 40 years. A recent theme of the optimists has been that official measures of real GDP understate the consumer surplus provided by free entertainment and information available on the internet. This paper reminds readers that real GDP measures have always missed vast amounts of consumer surplus since the dawn of the first industrial revolution almost three centuries ago, and few examples are more compelling than the replacement of horses and horse manure in urban streets by the motor vehicle between 1900 and 1930.

This paper points to a contradiction between the actual macro data on productivity growth and the techno-optimists’ predictions of accelerating growth. They fail to relate the “exponential growth of Moore’s Law” to the reality that the economy’s productivity performance for almost 10 years since 2004 shows no evidence of any any stimulus provided by the explosion of big data or invention of smart phones and tablets. Why should this be surprising? Marginal cost equals marginal benefit, and as electronic data becomes costless, its marginal benefit quickly approaches zero.

In the techno-optimist view, technological change is accelerating. Taken literally, this means that innovation will raise the growth rate of future productivity during 2013-23 faster than the rate of advance during 2003-13, which in turn must have been faster than the rate of change during 1993-2003. In contrast, this paper argues that diminishing returns have set in. Four pieces of evidence are provided that the post-1972 pace of technological change peaked in 1996-2000 and has been slowing down since then.

Vir: Robert Gordon (2014)

%d bloggers like this: