The Productivity Gap between Europe and the US: Trends and Causes Marcel P. Timmer Groningen Growth and Development Centre The EU KLEMS 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". EU-15 Labour productivity convergence to US has ended 100 (US=100) GDP per hour worked 90 80 GDP per capita 70 60 06 20 04 20 02 20 00 20 98 19 96 19 94 19 92 19 90 19 88 19 86 19 84 19 82 19 80 19 78 19 76 19 74 19 72 19 70 19 68 19 66 19 64 19 62 19 19 60 50 Alternative Explanations • • • Proximate causes: slower emergence of knowledge economy • • • • Small ICT-producing sector Limited role of ICT-investment Lower levels of skilled labour Less innovation (product and process) Ultimate causes: institutions • • • Role of labour markets (Blanchard, 2004) Product market regulations (Nicoletti and Scarpetta, 2003) End of catch-up (Aghion and Howitt, 2006): More R&D investment, Higher education system reform (Statistical myth) Sector Contribution to Labour Productivity Growth based on Shift Share Analysis VAi ,t H i ,t H i ,t VAt H t Ht VA VA ln vi ln vi ln ln VAt 1 H t 1 VAi ,t 1 H i ,t 1 i H i , 1 H t 1 i Source: Stiroh (2002) Sector contribution to market economy labour productivity growth, 1980-2005 3.0 2.0 1.0 0.0 EU, 80-95 EU, 95-05 Market services US, 80-95 Goods production Source: Updated from van Ark, O’Mahony and Timmer, Journal of Economic Perspectives, Winter 2008 US, 95-05 ICT production Market services important source of growth differences across Europe & US 4.0 Sector contribution to market economy labour productivity growth, 1995-2004 3.0 2.0 1.0 0.0 FIN US GBR AUT FRA NLD BEL GER DNK ITA -1.0 Market services Goods production* Source: Updated from van Ark, O’Mahony and Timmer, Journal of Economic Perspectives, Winter 2008 ICT production ESP EU KLEMS Growth Accounts (1) Gross output PF: Y j f j (K j , L j , X j ,T ) where Y is output, K is an index of capital service flows, L is an index of labour service flows and X is an index of intermediate inputs, Assumptions: competitive factor markets, full input utilization, constant returns to scale and using the translog functional form Growth accounting equation: ln Y jt v ln X jt v ln K jt v ln L jt ln A X jt v Xjt PjtX X jt PjtY Y jt K jt ; v Ljt PjtL L jt PjtY Y jt L jt ; v Kjt v L v K v X 1 PjtK K jt PjtY Y jt Y jt EU KLEMS Growth Accounts (2) Labour services based on hours worked by 18 types (education, age, gender) Capital services based on 8 asset types (ICT and nonICT) ln Lt vl ,t ln H l ,t l ln Kt vk ,t ln S k ,t k Measurement of capital services S i ,T i ,t I i ,T t Capital stocks for each asset t 0 S i ,T (1 i ) t 1 I i ,T t S i ,T 1 (1 i ) I i ,T t 0 ln Kt vk ,t ln S k ,t Aggregate capital services k and with contribution of capital services based on user cost equation v k ,T p k ,T S k ,T p k ,T S k ,T i p K k ,t p r k p [ p p I k ,t 1 t I k ,t I k ,t I k ,t 1 ] What is new in EU KLEMS? • • • • • • Data available for at least 30 industries and over 20 countries Systematic data collection based on national accounts and complementary official sources (LFS and other surveys) Long time coverage 1970-2005, with greatest detail for post-1995, harmonized methodologies on industry classification, deflation and aggregations Decomposition of inputs: • Capital assets in 7 asset types • Labour input in 18 categories (3 x skill; 3 x age and gender) • Intermediate inputs: energy, materials and services input Broad coverage of EU countries: • Growth accounts coverage of 14 EU new member states • Limited coverage of 11 other EU countries • Also comparisons with U.S. Korea, Canada, Australia and Japan www.euklems.net. Updated each year. Explaining growth in value added based labour productivity VA L L ln w ln H H w ICT ln Labour composition K ICT H N ICT K N ICT w ln H ln AVA ICT-capital per hour Non-ICT-capital per hour Multi factor productivity ICT investment contributes to labour productivity growth in market services growth in all countries, 1995-2004 3.5 2.5 1.5 0.5 -0.5 -1.5 US GBR NLD FIN FRA BEL DNK GER ICT capital deepening contribution AUT ESP ITA PLUS Improvement in labour composition 3.5 2.5 1.5 0.5 -0.5 -1.5 US GBR NLD FIN FRA ICT capital deepening contribution BEL DNK GER AUT ESP ITA Labour composition change contribution PLUS Non-ICT deepening 3.5 2.5 1.5 0.5 -0.5 -1.5 US GBR NLD FIN FRA Non-ICT capital deepening contribution Labour composition change contribution BEL DNK GER AUT ESP ICT capital deepening contribution ITA … but MFP contribution makes the big difference between fast and slow growth 3.5 2.5 1.5 0.5 -0.5 -1.5 US GBR NLD FIN FRA Non-ICT capital deepening contribution Labour composition change contribution BEL DNK GER AUT ESP ICT capital deepening contribution MFP growth ITA What is MFP growth? • • Ratio of output growth over input growth Disembodied technological change, under assumptions: • • • • • • • Constant returns to scale Technical efficient Allocative efficient Competitive output and input markets Outputs and inputs are measured without error All outputs and inputs are included Measure of our ignorance (Abramovitz)? • Deviation from assumptions above What is MFP growth? • • • • • • • Constant returns to scale ? Perhaps at sector level justified Technical efficient? Many firms/sectors are not on the frontier. Allocative efficiency? ICT investments…. Competitive output and input markets? If not, prices do not reflect marginal revenues and costs • • Non-competitive product markets: Mark ups? Regulation in capital and labour markets Outputs and inputs are measured without error? • • • Capital (and labour) capacity utilization Quality adjusted quantities of inputs Quality adjusted quantities of outputs All outputs and inputs are included? • • • Missing inputs (R&D, intangibles, management) Measure of our ignorance (Abramovitz)? Deviation from assumptions above Good news: many of this are (or can be) tackled How to explain differences in MFP growth? Distance to the frontier-approach 1.2 1 MFP-level (US=1) in Market Services 0.8 0.6 Source: Inklaar and Timmer (2008) GGDC Productivity level database JA P HU N CZ E DN K NL D US A G ER IR L BE L FR A ES P SW E FI N UK AU S IT A AU T SV N 0.4 Distance to the frontier-approach Unconditional convergence: ln MFPi ln MFPF MFPi Conditional convergence: ln MFPi ln MFPF MFPi Z Z ln MFPF MFPi Imitation Innovation Interaction Potential conditions include: 1.Skills, 2. ICT-capital and 3. Product market regulation (Inklaar, Timmer and van Ark (2008), “Market Services Productivity”, Economic Policy) Regulation Some evidence at detailed industry-level Dep var: MFP growth Technology gap Barriers (average) All industries 5 6 0.019*** 0.009 (0.004) (0.007) 0.000 -0.004 (0.007) (0.007) Barriers (industry-level) -0.010 (0.007) Barriers*Technology gap Number of observations Selected set 7 8 0.009 0.010 (0.006) (0.012) 2376 0.015 (0.009) 2376 715 -0.009 (0.010) -0.002 (0.017) 715 Post and telecom 8 0.068*** (0.023) -0.060*** (0.021) 0.037 (0.029) 264 Source: Inklaar, Timmer and van Ark (2008), “Market Services Productivity”, Economic Policy, Table 10 Main findings • • • • • Importance of having detailed measures of productivity at industry level, both growth and level estimates End of convergence in productivity across Europe and the US, mainly driven by differences in MFP growth in market services New productivity gap has opened up Regulatory practice shows no clear impact on MFP, only for post and telecommunications Market regulation, competition and Scale and/or lagged adjustment to ICT in Europe? Some ideas for future research based on methods teached in this course • • • • • • • • • • Breakdown of TFP growth (DEA,SFA)? Allow for increasing returns to scale? Imperfect competition? Localised innovation? (DEA) Stochastic frontier estimation of MFP-gaps Efficient use of ICT? (.Dot boom and bust) Complementarity skills and ICT Variable lags and interactions in effects of skills, ICT and regulation Innovation, trade and R&D (manufacturing) Sectoral convergence in the world economy?