GDP growth by about 0.7 percentage point, similar to their estimate reported in our Outlook last year. Of course, not all experts believe that there is a mismeasurement problem. Notable among them is Chad Syverson of the University of Chicago, who raises four points in making this case.*° First, he states that the productivity slowdown has been global in nature and unrelated to countries’ consumption or production intensities of information and communication technology. Second, he states that estimates of consumer surplus are too small relative to his estimates of lost GDP due to slower productivity growth. Third, he argues that if such mismeasurement existed, the growth rate in the information and communication technology sector would be a multiple of its stated growth rate.
We believe that productivity in
health care is underestimated. Finally, while he acknowledges that gross domestic Consider IBM’s Watson Health, income has been higher than GDP since 2004 and an artificial intelligence system the gap might reflect the higher wages of workers
who are producing non-market digital services, he does not believe that this difference is evidence of mismeasured GDP because the trend started earlier
that can read 200 million pages of text in 3 seconds. As of 2015,
Watson had amassed 315 billion than the slowdown in productivity growth. data points representing health A somewhat similar line of reasoning has been records, lab results, genomic presented by Byrne, John Fernald of the Federal
Reserve Bank of San Francisco and Marshall Reinsdorf of the IMF in a paper titled “Does the United States Have a Productivity Slowdown or a
tests and clinical studies. The system processes patients’ cases
against its ever-growing database Measurement Problem?”*’ While they agree that and recommends customized productivity growth has been mismeasured in the treatment options. With such past, they argue that the mismeasurement has been
negligible in the 2004-15 period partly because computer hardware, for which mismeasurement was once a factor, now makes up a smaller part of of productivity in health care. GDP. Therefore, the impact is less in the 2004-15 period than it was in the 1995-2004 period when productivity growth was much higher. They also state that free digital products not only are non- Ihe joke oductivity growth market and should not be counted in GDP, but also were better measured, are not sizable enough to account for the level of decline in productivity growth rates.
developments, we expect continued improvement in growth
par ticularly in health and Experts’ opinions on mismeasurement continue
other services, the growth to evolve. In fact, in a subsequent publication co- authored with Carol Corrado of the Conference
rate would look better than Board, Byrne found a significantly higher level
is currently reported.” of software price mismeasurement than assumed
in his prior paper.*® He has also co-authored a Brookings Institution, September 2016. tablets such as iPads, proving that quality-adjusted price indices for tablets have fallen much faster than the broader price indices for computers and peripheral equipment.*?
16 | Goldman Sachs | JANUARY 2017
HOUSE_OVERSIGHT_014549