IT in Europe and the US

1
InnovaTion, InvesTment and ImiTation How
Information and Communication Technology Affected
European Productivity Performance Bart Los and
Marcel Timmer, University of Groningen (Faculty
of Economics, Groningen Growth and Development
Centre) This project is funded by the European
Commission, Research Directorate General as part
of the 6th Framework Programme, Priority 8,
"Policy Support and Anticipating Scientific and
Technological Needs".
2
End of European catch-up process
3
Slowdown in most countries
4
ICT as a GPT

• ICT as GPT (see David, AER 1990 Brynjolfsson
  Hitt, JEPersp 2000 Hall Trajtenberg 2004 NBER)
• Europe has exhausted imitation in old
  technologies, and is lagging in application of
  new ICT-based innovations (Aghion and Howitt
  2006, JEEA).
• Divergence is possible. Degree to which imitation
  can lead to productivity gains depends on
  technology operated (appropriate technology,
  Basu Weil, QJE, 1998)
• Traditional growth accounting findings Total
  Factor Productivity growth in market services in
  US, but not in Europe (see Jorgenson, Ho and
  Stiroh 2005 Triplett and Bosworth 2004 Inklaar,
  OMahony and Timmer, RIW, 2003 Timmer Van Ark,
  OxEP 2005.)

5
This paper

• TFP should be divided into pure technological
  change (innovation) and efficiency changes
  (imitation) through estimation of global
  production frontier. (see Los Timmer, JDevE
  2005, Timmer and Los, JPA 2005)
• ICT capital is a critical input
• Questions
• How many years are European countries lagging
  behind in ICT?
• How efficient are European countries in using
  old-vintage ICT?
• Does this differ for various sectors?
• How much of European growth is due to innovation
  and how much due to imitation?

6
Falling behind, catching up or leapfrogging?
7
Frontier Estimation

• Data Envelopment Analysis on 1 output (GDP) and 3
  inputs (labour, IT and non-IT), assuming CRS
  (Färe et all, 1994, AER)
• Non-parametric approach with very few
  restrictions on production technology
• Advantage is flexible functional form, which
  allows for localized technological change
• DEA with intertemporal dataset, to avoid
  technological regress.
• Frontier for year y based on all observations
  from 1980 up to y
• First frontier for 1990.

8
Frontier estimation (1)
9
Frontier estimation (2)
10
Frontier estimation (3)
11
Frontier estimation (4)
12
Data

• Fourteen countries EU-15 (minus Luxembourg and
  Ireland), and U.S.
• GDP (at PPP), total hours worked and capital
  stocks for 1980-2004
• Harmonised Capital stock estimates for six
  assets, using Perpetual Inventory Method,
  aggregated into two groups Non-IT and IT capital
• Non-IT machinery, transport equipment and
  non-residential buildings
• IT office and computing equipment,
  communication equipment and software.
• All data from The Conference Board and Groningen
  Growth and Development Centre, Total Economy
  Database, May 2006, (www.ggdc.net) (Updated from
  Timmer Van Ark, OxEP, 2005)

13
Lagging ICT stocks in Europe
14
Frontier Results 1990

• Frontier points include UK (80,97,88), US (89,
  90), DK (80, 85, 86), France (80, 90)

15
Frontier Results 1995

• New Frontier points include UK (95), US (94,
  95), DK (95)

16
Frontier Results 2000

• New Frontier points include US (98,99,00), UK
  (97, 00), DK (96,00)

17
Frontier Results 2004

• New Frontier points include Fr (04), US (04),
  DK (02,04)

18
European countries in 2004
19
ICT Lags ...
20
But some have succesfully imitated, or even
innovated
21
Efficiency scores
22
European manufacturing is innovating
23
Main findings and road ahead

• Global production frontier is driven by
  investment in ICT capital goods
• European countries lag US in application of ICT
  technology (4 to 16 years)
• Some countries are succesful imitators (FR, UK,
  DK), but others face divergence (IT, PT, ES)
• Different pattern at industry level innovation
  in manufacturing in some countries

24
Main findings and road ahead

• Industry-level analysis, in particular services
  EUKLEMS data project
• Explanation of divergence in terms of Imitation
  (inefficiency model) including regulation and
  skilled labour supply as determinants