The Missing Productivity Conundrum

Why has productivity been falling everywhere?

The latest issue of the Journal of Economic Perspectives has two papers on this problem.[1]The first is by Martin Feldstein, who focuses on the mismeasurement of real GDP, hence the mismeasurement hypothesis.[2] He says, “After studying the methods used by the US government statistical agencies as well as the extensive previous academic literature on this subject, I have concluded that, despite the various improvements to statistical methods that have been made through the years, the official data understate the changes of real output and productivity.” He emphasizes the inability to measure the “improvements in the quality of goods and services,” and especially for new goods and services.  

The second paper disputes the mismeasurement hypothesis.[3] Chad Syverson focuses on ICT. He shows evidence that the slowdown in productivity has occurred in many countries and not only in the U.S. and perhaps long before 2004. (See (Haldane) for the same in the U.K.) He also shows that the surplus created by the new goods and services (i.e., ICT) is far less than the missing output, which caused the slowdown in productivity. He also makes other points worth examining.

My argument is that mismeasurement explains a lot of the missing productivity in OECD countries.

First, labor productivity, measured by the ratio output/input, is pertinent to profit-maximizing firms. It makes no sense to talk about productivity, and to invest in non-profit maximizing firms and sectors.  

Second, GDP data include both profit-maximizing firms (and sectors) output and non-profit maximizing public sectors and government agencies’ output. Inability to measure the quality of goods and services notwithstanding, services output in particular is very difficult to measure, especially in non-profit maximizing sectors. Imagine the difficulty of measuring output in the public sector and government administration and services. It might be easier to measure the output of a private school or a private hospital services by inferring output from the profit function. The data, however, do not allow for such classification. Therefore, I argue that aggregate output (total real GDP) is not the right measure to use for calculating labor productivity because it includes a lot of “hard to measure outputs.”  

Third, the higher the share of services in GDP the lower measured labor productivity.

To illustrate, for New Zealand, I measure labor productivity using the sectoral data ANZSIC06 published by Stat NZ and employment by sectors (data for hours by industry are unavailable). The data are available from 2004 to 2017. Then I modify this labor productivity by subtracting the following “services” sectors data from total industry output and employment: Professional, Scientific, Technical, Administrative and Support Services, Public Administration and Safety, Education and Training, Health Care and Social Assistance, Arts, Recreation and Other Services, rental and real estate activity, and a column labeled “Not Specified.” I did not remove: Financial and Insurance Services because I think that this sector’s output could be measured or inferred from the firm’s profit function.[4]

I convert the two productivity measures to indexes (2004=100). Here is the plot.

It is clear that labor productivity (less services) is significantly higher than the conventional measure which includes public sector services.

Caveat: there must be a number of firms and agencies within the services sectors, which I removed, that are profit-maximizing. For example, in the public service sectors there are private schools, private hospitals, private arts and recreation etc. I cannot separate these from the data because I do not have enough information.

The average share of these services in GDP over the sample is 52 percent. It has been relatively stable.

This graph plots the growth rate of NZ labor productivity.

New Zealand missing growth is nearly 1 percentage point on average, which is quite a significant mismeasurement.

Labor productivity will continue to be understated until measurements of services are addressed by statistical agencies.  

razzakw@gmail.com

---

[1] Journal of Economic Perspectives Vol. 31, Issue 2, Spring 2017.

[2] Underestimating the Real Growth of GDP, Personal Income, and Productivity.

[3] Challenges to Mismeasurement Explanations for the US Productivity Slowdown.

[4] The Conference Board data KLEMS report figures for they call Market Activity whereby these sectors are subtracted from the total industry data.