• No results found

As for the third obstacle, Syverson analyses products and respective industries, associated with 2004- technologies which could be associated with problems resulting from the migration of value added to consumer surplus (Syverson (2016, p. 13)). More precisely, the author extracts data for real value added (and for the second potential calculation method revenues97) of those industries, providing the

goods which are most likely of the character of underestimating consumer surplus.

Firstly, the relevant sectors at central concern for the mismeasurement hypothesis are defined. Secondly, the growth in real value added of the industries within the period 2004-2015 is calculated.

96A combination of the two sources had to be used as the OECD does not provide data further than 2010 in this

category, nor does the Statistisches Bundesamt provide data going back before 2008.

Real values for 2004 are expressed in 2015-$ in order to provide the same price base98. As a result,

Syverson (2016) receives $545 bn. in total for the change in real value added in these industries. If the problem of (missing) migration from output to consumer surplus is to account for the gap in productivity (and therefore in favour of the mismeasurement hypothesis), the incremental consumer surplus has to be equal the missing portion of $2.7 trn. or at least close to it. Total change in real value added of $545 bn., however, is just around 20% of the ’gap’ in the statistics and makes underestimated growth in the ’problem’-industries very unlikely as potential explanation.

Data for the German calculation is taken via Eurostat from the Structural Business Statistics database (SBS). The statistics describe the “structure, activity, competitiveness and performance of economic activities within the business economy down to the detailed level of several hundred sectors” (see Eurostat (2018b)). Applying the method from Syverson for this pattern forces to include the same sectors like in his study. As the US and Germany make use of different sectoral classification systems, this does not only provide the need for a comparison of the NAICS-classification system (used in the US) and the NACE-classification system (used in Germany). In addition, the results of this calculation have to be used with an extra amount of care. At first, I make a comparison for both systems in order to filter the relevant sectors, which shall be included in the calculation. In original, Syverson includes (Syverson (2016, p. 13)):

• computer and electronic products manufacturing (NAICS 334) • the entire information sector (NAICS 51) and

• computer systems design and related services (NAICS 5415)

The comparison and resulting sectors (and sub-sectors) for the German analysis can be found in appendix V. The relevant sectors used for the German calculation (according to NACE 2) are:

• for the manufacturing sector: reproduction of recorded media (18.20), manufacture of electronic components (26.11), manufacture of loaded electronic boards (26.12), manufacture of computers

98Syverson approximates values for 2015 and for the price base 2015, as at the date of publication data was restricted

and peripheral equipment (26.20), manufacture of communication equipment (26.30), manufac- ture of consumer electronics (26.40), manufacture of instruments and appliances for measuring, testing and navigation (26.51), manufacture of watches and clocks (26.52), manufacture of irra- diation, electromedical and electrotherapeutic equipment (26.60), manufacture of optical instru- ments and photographic equipment (26.70), manufacture of other electrical equipment (27.90), manufacture of railway locomotives and rolling stock (30.20), manufacture of air and spacecraft and related machinery (30.30).

• for the service sector: book publishing (58.11), publishing of directories and mailing lists (58.12), publishing of newspapers (58.13), publishing of journals and periodicals (58.14), other publish- ing activities (58.19), publishing of computer games (58.21), other software publishing (58.29), motion picture, video and television programme production activities (59.11), motion picture, video and television programme post-production activities (59.12), motion picture, video and television programme distribution activities (59.13), motion picture projection activities (59.14), sound recording and music publishing activities (59.29), radio broadcasting (60.10), television programming and broadcasting activities (60.20), wired telecommunications activities (61.10), wireless telecommunications activities (61.20), satellite telecommunications activities (61.30), other telecommunications activities (61.90), computer programming activities (62.01), computer consultancy activities (62.02), computer facilities management activities (62.03), other informa- tion technology and computer service activities (62.09), data processing, hosting and related activities (63.11), web portals (63.12), news agency activities (63.91), other information service activities n.e.c. (63.99), photographic activities (74.20),

• Sector 91.01 “library and archive activities” is not included in this calculation, as no data are provided99. Its weight for aggregate value added in the sectors is limited and can be neglected

without concern.

In a next step, data from the SBS is taken (Eurostat (2019a) and Eurostat (2019b)), in order to calculate value added for the sectors chosen. I aggregate value added for these sectors for the years

99Please see <‌<https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Structural_business_statistics_overview>‌>

2015 and the 2004100. This results in nominal value added for the selected sectors. For the year 2004

it is €74.386,2 mil. and for the year 2015 the value is €102.592,6 mil.. In order to receive real value added for the sectors, I make use of sectoral price-deflators (the alternative approach of the Syverson- calculation, see Syverson (2016, p. 14 footnotes)). For the manufacturing sector, the calculation was pretty straightforward. A sectoral deflator is available and real value added of the industries can be constructed (Statistisches Bundesamt (2017c), Genesis Code: 61241 - 0001). GDP deflators for the service sector are not provided back to 2004 by the Statistisches Bundesamt. In fact, data only for a certain amount of services are available (Statistisches Bundesamt (2017c), Genesis Code: 61311-0003). In order to calculate real values, I make use of an approximation by taking the unweighted average of all service subsectors, providing data for 2004. The respective calculations can be found in the appendix V. Expressed in 2015-€ this yields real value added for 2004 of €62.410,8 trn. and €102.592,6 trn. for 2015, implying a change of €40.181,81 trn. for the selected sectors. Compared to the missing portion (€246 bn.) this is around 16% and provides similar results and implications as for the US by the Syverson-study (around 20%) - a result of minor value for the validity of the mismeasurement hypothesis.