Services productivity revisited

Abstract

This paper revisits the recent debate and evidence on productivity in services. Firstly, it discusses the main works analysing developments and impacts of services productivity in growth and overall productivity. Secondly, new data is presented concerning Baumol's ‘cost disease’ and the recent evolution of services productivity growth by main service branches. Decomposition of trends and cycles is also addressed to capture structural trends in isolation. Results confirm the continuing validity of such disease at aggregate level to a certain extent, but different subsectors show uneven patterns. Nevertheless, other complementary approaches and further research are needed to capture in a more comprehensive way the new developments and impacts of services productivity.

Keywords:

• Baumol's disease
• productivity
• services

Notes

1. A previous version of this section has been published in Rubalcaba (2007), which also includes complementary aspects to the ones presented in this paper.

2. Baumol himself (1989) configures a new classification of services, according to productivity, where sectors with a low productivity growth coexist with services registering productive growth that is equal to or higher than that of manufactured products. Along the same lines, more recent studies show that only one-third of the services sector can be identified as low productivity growth activities, while the rest includes sectors registering similar growth rates (transport and storage) or even higher (telecommunications) than the manufacturing sector. More recently, Baumol draws conclusions that highlight the importance of services and their innovation to economic growth (Baumol, 2000).

3. These results force us to reconsider previous literature on this subject. While the manufactured goods sector has increased – primarily due to improvements in technical and technological efficiency (productivity gains that will be translated into higher wage increments, and consequently, into increases of costs related to less progressive sectors of the economy) – the rapid growth registered in the services sector can be explained by the increase of capital and labour factors.

4. Regarding some specific services branches, such as business services, Baumol's limits of original interpretation are even more important than in other services sectors, as shown in the works by Oulton (2001) and Wölfl (2003) and Kox (2002, 2004). Baumol's model focuses on consumer services, while business services are intermediate inputs for other industries. Also, the possibility exists that the business services industry indirectly increases the productivity of other industries. This may be a consequence of the knowledge generated, and the possibility that underestimation problems are overlooked.

5.

Real estate services are not included in this figure due to their huge productivity level (601.2 thousand euros per employee), which is mainly caused by the peculiar estimate of their production, which is conventionally carried out through the inclusion of a cost allocation when buildings and land are owned. For this reason, and despite the fact that their levels are considerably higher than the rest of services, the growth rates during the period analysed have been negative (–0.59 per cent per year).

Total economy (TOT); Agriculture (AGR); Manufacturing (MANU); Energy (ENE); Construction (CONS); Services (SS); Distributive trade (TRA); Hotels and restaurants (HOT); Transport (TRA); Communications (COMM); Financial services and insurances (FIN); Business services (BUSS) – Equipment renting (REN); IT services (IT); R&D (R&D); Legal, technical and advertising services (LTA); and other business services (OTHER); Services provided by Public administrations (PUB); and Social and personal services (SOC). Real estate activities (REAL) are not included in the figure since their productivity level is much higher than the other areas.

6. From the lambda value suggested by the authors in the four-month series (1600), the Ravn and Uhlig (2001) formula is used, which leads to a value of 6.25 for the annual series. In the work performed by O'Mahony and van Ark (2003), the differences are shown with other values, such as the 6400 used by Gordon (2003), which only affects 4% of the growth observed.

7. Hence, it is not surprising that business services volatility is consistently higher from the 1980s onwards. From then, the processes of market flexibility and inter-industrial integration have taken place, which have had a very strong influence on business services.