## _wp04200

## Source details

**Canonical URL:** [_wp04200](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04200.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04200.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04200.pdf.json)

---

### Key findings
- The bulk of the labor productivity deceleration in the euro area in the second half of the 1990s can be explained by slower capital deepening (slower growth in the capital-labor ratio), as opposed to slower TFP growth.
- The apparent slowdown in TFP growth for the euro area using aggregate national accounts data largely disappears once better, industry-level data for Germany are considered.
- Structural wage-setting changes that made labor cheaper induced firms to slow capital accumulation and to hire more workers; econometric estimates for France, Germany, and the Netherlands indicate wage-setting shocks would have forced capital-labor ratios to decline in the second half of the 1990s, though in practice capital-labor ratios grew at a slower rate rather than declined because cheaper ICT equipment partly offset the wage shock.
- The productivity growth differential with respect to the United States since the mid-1990s is explained by:
  - Faster labor productivity deceleration in traditional industries in the euro area, and
  - A surge in productivity growth in intensive ICT-using sectors (mainly wholesale and retail trade and financial intermediation) in the United States.
- Productivity behavior in ICT-producing sectors (computers, semiconductors, communication services) was similar in the euro area and the United States.

### Aggregate GDP per capita and productivity trends
- Long-run trend GDP per capita growth in the euro area has declined since the 1950s and convergence to U.S. levels halted in the 1970s.
- In the United States, labor productivity growth oscillated around "1½ percent" a year for many years until it trended up in the second half of the 1990s, surpassing the euro-area figures.
- GDP per capita growth in the second half of the 1990s was "about 1 percentage point" a year higher in the United States than in the euro area; increasing employment rates in the United States widened this gap.
- Aggregate growth-accounting decomposition (Table 2, annual rates, in percent):
  - Euro area, 1995-2000: Labor Productivity 1.6; Capital deepening 0.4; TFP 1.2.
  - Euro area, 1995-2003: Labor Productivity 1.2; Capital deepening 0.4; TFP 0.8.
  - United States, 1995-2000: Labor Productivity 2.1; Capital deepening 0.4; TFP 1.7.
  - United States, 1995-2003: Labor Productivity 2.1; Capital deepening 0.6; TFP 1.4.

### Industry-level decomposition and measurement corrections
- Use of the GGDC industry-level database (56 industries, 15 EU countries and the United States) corrects for ICT quality changes using U.S. statistical agency information and uses value-added weights at the industry level.
- Industry data show labor productivity decelerated by "0.7 percentage point" in the euro area in the second half of the 1990s, versus "1 percentage point" indicated by aggregate national accounts—i.e., the deceleration is smaller once industry-level corrections are applied.
- When compared to the United States, the faster productivity deceleration in non-ICT industries in the euro area accounts for "30 percent" of the gap between the two regions since the mid-1990s; the surge in ICT-using industries in the United States accounts for the remaining difference.
- If GGDC-based corrections are applied, euro-area TFP growth would be "0.35 percentage point higher than shown in Table 2"—about the size of the deceleration in TFP reported in aggregate data.

### Sectoral productivity patterns and numeric highlights
- Euro area labor productivity growth: 1.5 percent in the second half (acceleration of -0.7 percentage point).
- United States labor productivity growth: increased from 1.2 percent to 2.3 percent over the same periods (acceleration of 1.1 percentage point).
- Relative swing: Table 4 shows a 1.7 percentage points relative swing in these growth rates in favor of the United States (1.1 percentage points minus -0.7 percentage point, adjusted for rounding).
- U.S. surge in labor productivity growth in ICT-using industries: "1.2 percentage points or 70 percent of the gap."
- Italy contribution to euro-area deceleration: "about 40 percent of the 0.7 percentage point deceleration in labor productivity growth in the euro area in the second half of the 1990s."
- Germany weight in euro-area aggregate: "about 30 percent of total value added in the area."
- TFP differential dynamics:
  - U.S. TFP growth rose by "¾ percentage point" in the second half of the 1990s.
  - Euro-3 (France, Germany, Netherlands) TFP increased by "1/3 percentage point" in the second half of the 1990s.
  - Differential in TFP acceleration in favor of the United States: U.S. goes "from 0.13 percent ... to 0.87 percent" while euro area goes "from 0.69 percent to 1 percent."
  - All of the differential acceleration in TFP in favor of the United States originates in ICT-using industries: U.S. goes "from 0 percent to 2.34 percent" while euro area goes "from 0.79 percent to 0.63 percent."

### Structural labor market changes, mechanism, and empirical estimates
- Mechanism (right-to-manage wage-setting curve): W/A = f(u, B, τ, m) where lower m or reduced unemployment income (B net of wedge τ) lowers wage demands for a given unemployment rate.
- A downward shift in the wage-setting curve (wage moderation) initially: unemployment down, capital-labor ratio K/A decelerates, labor productivity growth falls; longer run profit-induced investment can reverse the capital-labor trajectory.
- Potential short-run TFP effects: reduced if newly hired workers are lower-quality or if labor quality is mismeasured; long-run TFP may benefit from better labor allocation.
- Wage-setting shock empirical construction: ln(W/(CP*A)) = γ + 1.0*ln(u)*(-0.1) + ξ, adopting θ = 0.1.
- Estimated panel regression linking percent changes in the capital-labor ratio to wage-setting shocks (∆ξijt) and changes in the user cost of capital (∆ηijt) yields key elasticity results:
  - WS shock: 0.64 *
    - Standard error: (0.31)
    - * significant at the 5 percent level.
  - country dummies: yes
  - industry dummies: yes
  - time dummies: yes
  - industry*time dummies: yes
  - country*time dummies: yes
  - country*industry: yes
  - Adj. R2: 0.40
  - Number of observations: 1,690
  - Number of industries: 26
  - Sample period: 1980-2000
  - Sources: GGDC; AMECO database; and author's estimates.
- Counterfactual implication: based on the evolution of cumulative wage-setting shocks, "capital-labor ratios would have declined in the euro area in the absence of further shocks." The contribution of capital deepening to annual labor productivity growth would have been "about -0.3 percentage point" as opposed to the "0.4 percentage point" shown in aggregate Table 2.

### Other factors affecting capital deepening and productivity
- Drops in the user cost of capital because of declining interest rates and cheaper ICT equipment partly offset the negative effects of wage-setting shocks on capital accumulation.
- Contribution of ICT capital deepening increased in all examined countries; contribution of non-ICT capital deepening declined and became negative in France and zero in the Netherlands (per Table 9).
- Labor quality growth contributed less to productivity growth in the Netherlands and Germany, but not in France.

### Policy implications and recommendations
- Labor market reforms
  - Reforms necessary to continue reabsorbing people into jobs will probably ultimately increase TFP growth.
  - Reforms should aim at:
    - Increasing the incentives to work vis-à-vis receiving social benefits.
    - Correcting incentives for human capital formation, with labor income better reflecting individual abilities and efforts.
  - Expected outcomes: increase in human capital accumulation and better allocation of labor across uses, boosting TFP growth in the long term.
- Product market reforms
  - Evidence points to the need for reforms in specific sectors, particularly ICT-using services (wholesale trade, retail trade, financial intermediation).
  - Policy implications: lowering barriers to entry and easing regulatory burdens on enterprise creation in Europe seem necessary.
  - Note: European Commission (2003) simulations suggest that even relatively rapid deregulation toward U.S. levels would not produce sufficiently large productivity gains over the next seven years to close the efficiency gap with the United States.
- Complementary measures to foster technological progress
  - Product market reforms could positively affect risk-taking activities that drive technological progress.
  - Investments in education and R&D:
    - Long-run productivity gains from education and R&D would directly increase TFP growth.
    - For R&D, focus should be on creating conditions that promote an endogenous increase in research spending rather than directly boosting R&D spending.
    - Two channels to stimulate endogenous R&D growth: higher product market integration and an investment environment that develops a more active market for risk capital.
- Interpretation caveat: it remains an open question whether the euro area is lagging the United States primarily in ICT adoption in some service-sector industries (and will catch up) versus structurally falling behind in dynamic TFP generation; this is left for future research.

*Source: _wp04200 - References*

### References..............................................................................................................

### _wp04200 - References

### Key findings
- The bulk of the labor productivity deceleration in the euro area in the second half of the 1990s can be explained by slower capital deepening (slower growth in the capital-labor ratio), as opposed to slower TFP growth.
- The apparent slowdown in TFP growth for the euro area using aggregate national accounts data largely disappears once better, industry-level data for Germany are considered.
- Structural wage-setting changes that made labor cheaper induced firms to slow capital accumulation and to hire more workers; econometric estimates for France, Germany, and the Netherlands indicate wage-setting shocks would have forced capital-labor ratios to decline in the second half of the 1990s, though in practice capital-labor ratios grew at a slower rate rather than declined because cheaper ICT equipment partly offset the wage shock.
- The productivity growth differential with respect to the United States since the mid-1990s is explained by:
  - Faster labor productivity deceleration in traditional industries in the euro area, and
  - A surge in productivity growth in intensive ICT-using sectors (mainly wholesale and retail trade and financial intermediation) in the United States.
- Productivity behavior in ICT-producing sectors (computers, semiconductors, communication services) was similar in the euro area and the United States.

### GDP per capita and productivity trends (aggregate national accounts)
- Long-run trend GDP per capita growth in the euro area has declined since the 1950s and convergence to U.S. levels halted in the 1970s.
- In the United States, labor productivity growth oscillated around "1½ percent" a year for many years until it trended up in the second half of the 1990s, surpassing the euro-area figures.
- GDP per capita growth in the second half of the 1990s was "about 1 percentage point" a year higher in the United States than in the euro area; increasing employment rates in the United States widened this gap.
- Aggregate growth-accounting decomposition (Table 2, annual rates, in percent):
  - Euro area, 1995-2000: Labor Productivity 1.6; Capital deepening 0.4; TFP 1.2.
  - Euro area, 1995-2003: Labor Productivity 1.2; Capital deepening 0.4; TFP 0.8.
  - United States, 1995-2000: Labor Productivity 2.1; Capital deepening 0.4; TFP 1.7.
  - United States, 1995-2003: Labor Productivity 2.1; Capital deepening 0.6; TFP 1.4.

### Decomposition and proximate causes
- Aggregate national accounts suggest both a significant decline in capital deepening and a decline in TFP growth in the euro area in the second half of the 1990s; however, industry-level data correct some of the apparent TFP slowdown.
- Reduced wage demands and reinsertion of unemployed workers into employment are associated with slower capital deepening because lower real hourly compensation encouraged hiring rather than capital accumulation.
- Real hourly compensation in the euro area in the second half of the 1990s grew significantly more slowly than in the United States for the first time in the available series; euro-area unit labor costs show a negative trend consistent with wage moderation and labor market reforms starting in the 1980s.

### Industry-level analysis (GGDC industry database)
- Use of the GGDC industry-level database (56 industries, 15 EU countries and the United States) provides corrections for ICT quality changes using U.S. statistical agency information and uses value-added weights at the industry level.
- Industry data show labor productivity decelerated by "0.7 percentage point" in the euro area in the second half of the 1990s, versus "1 percentage point" indicated by aggregate national accounts—i.e., the deceleration is smaller once industry-level corrections are applied.
- When compared to the United States, the faster productivity deceleration in non-ICT industries in the euro area accounts for "30 percent" of the gap between the two regions since the mid-1990s; the surge in ICT-using industries in the United States accounts for the remaining difference.

### Policy implications and scenarios
- Labor market reforms that improve labor utilization (pursuing Lisbon targets) could temporarily damp labor productivity growth through slower capital deepening as employment rises, but this is expected to be a temporary phenomenon that fades when the economy reaches a new equilibrium unemployment rate.
- Continued labor market reforms that sustain low wage growth while reducing unemployment should improve economic efficiency.
- Product market deregulation, particularly in wholesale and retail trade, would promote efficiency gains and help close the productivity growth gap with the United States.
- Policies that better reward individual effort (raising risk-taking, R&D spending, and human capital accumulation) could raise TFP growth.

*Source: _wp04200 - References*

### 1.5 percent in the second half (acceleration of -0.7 percentage point). In the United States,

### _wp04200 - 1.5 percent in the second half (acceleration of -0.7 percentage point). In the United States,

### Key findings
- Euro area labor productivity growth: 1.5 percent in the second half (acceleration of -0.7 percentage point).
- United States labor productivity growth: increased from 1.2 percent to 2.3 percent over the same periods (acceleration of 1.1 percentage point).
- Relative swing: Table 4 shows a 1.7 percentage points relative swing in these growth rates in favor of the United States (1.1 percentage points minus -0.7 percentage point, adjusted for rounding).
- Partial explanation reported: a larger deceleration in non-ICT industries in the euro area of 0.5 percentage point or 30 percent of the gap.
- Note: source text continues beyond the provided excerpt; no additional details are available in this unit.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04200.pdf*

### Appendix III provides a listing of industries by ICT classification according to work

### _wp04200 - Appendix III provides a listing of industries by ICT classification according to work

### Sectoral productivity patterns and key numeric findings
- U.S. surge in labor productivity growth in ICT-using industries: "1.2 percentage points or 70 percent of the gap."
- Italy contribution to euro-area deceleration: "about 40 percent of the 0.7 percentage point deceleration in labor productivity growth in the euro area in the second half of the 1990s."
- Germany weight in euro-area aggregate: "about 30 percent of total value added in the area."
- TFP differential dynamics:
  - U.S. TFP growth rose by "¾ percentage point" in the second half of the 1990s.
  - Euro-3 (France, Germany, Netherlands) TFP increased by "1/3 percentage point" in the second half of the 1990s.
  - Differential in TFP acceleration in favor of the United States: U.S. goes "from 0.13 percent ... to 0.87 percent" while euro area goes "from 0.69 percent to 1 percent."
  - All of the differential acceleration in TFP in favor of the United States originates in ICT-using industries: U.S. goes "from 0 percent to 2.34 percent" while euro area goes "from 0.79 percent to 0.63 percent."

### Decomposition of labor productivity: capital deepening, labor quality, and TFP
- Main decomposition insight: Slower capital deepening (particularly non-ICT capital deepening) and slower labor quality improvements in euro-3 account for much of the euro-area lag versus the United States in the second half of the 1990s.
- Cross-country/cross-sector stylized facts:
  - Contribution of ICT capital deepening increased in all examined countries.
  - Contribution of non-ICT capital deepening declined and became negative in France and zero in the Netherlands (per Table 9).
  - Labor quality growth contributed less to productivity growth in the Netherlands and Germany, but not in France.
- Sectoral roles:
  - U.S. outperformance driven by: much slower deceleration in non-ICT industries and faster acceleration in ICT-using sectors.
  - Euro-area faster labor productivity acceleration in ICT-producing industries than the United States, but small aggregate impact because of small GDP share of ICT-producing industries.

### Cross-country heterogeneity within the euro area
- Several countries saw increases in ICT-using sector productivity between halves of the 1990s (Ireland, Netherlands, Portugal, Spain). Only Ireland exceeded U.S. growth.
- The three largest euro-area countries exerted downward pressure on ICT-using productivity growth and were responsible for most deceleration in the large non-ICT sector.
- Table 7 country contributions to euro-area deceleration (excerpted narrative):
  - Italy: contributed about 40 percent of the 0.7 percentage point deceleration.

### Measurement and database issues affecting TFP interpretation
- Growth Accounting Database (GGDC) versus aggregate AMECO/OECD data:
  - Using detailed industry-based GGDC data, German TFP "accelerates continuously" whereas aggregate data show a sharp decline.
  - If GGDC-based corrections applied, euro-area TFP growth would be "0.35 percentage point higher than shown in Table 2"—about the size of the deceleration in TFP reported in aggregate data.
  - This correction would align labor productivity deceleration (~0.7 percentage point) with the industry-data measurement (Table 3).

### Structural labor market changes and their modeled effects
- Mechanism summarized:
  - Right-to-manage wage-setting curve: W/A = f(u, B, τ, m) where lower m or reduced unemployment income (B net of wedge τ) lowers wage demands for a given unemployment rate.
  - A downward shift in the wage-setting curve (wage moderation) initially: unemployment down, capital-labor ratio K/A decelerates, labor productivity growth falls.
  - Longer run: higher profit rates induce investment, shifting short-run labor demand outward, capital deepening accelerates, and labor productivity growth surges before reverting to steady state.
- Potential additional effects:
  - Short-run TFP may be reduced if newly hired workers are lower-quality or if labor quality is mismeasured.
  - Long-run TFP may benefit from better labor allocation following reforms.

### Empirical estimation of wage-setting shocks and impact on capital deepening
- Wage-setting shock construction (industry-country-time): empirical specification mimics ln(W/(CP*A)) = γ + 1.0*ln(u)*(-0.1) + ξ, adopting θ = 0.1 as in the literature.
- Estimated panel regression linking percent changes in capital-labor ratio to wage-setting shocks (∆ξijt) and changes in the user cost of capital (∆ηijt) with controls F(.) (industry/country/year dummies and interactions).
- Key empirical result: wage-setting shocks estimated to affect capital deepening with an elasticity of "0.64" in the panel of France, Germany, and the Netherlands (used as representative of the euro area).
- Counterfactual implication: based on the evolution of cumulative wage-setting shocks, "capital-labor ratios would have declined in the euro area in the absence of further shocks." The contribution of capital deepening to annual labor productivity growth would have been "about -0.3 percentage point" as opposed to the "0.4 percentage point" shown in (aggregate) Table 2.

### Summary conclusions (sectoral and macro drivers)
- The U.S. productivity lead in the late 1990s owes mainly to:
  - Much slower deceleration in non-ICT industries in the United States.
  - Faster acceleration in ICT-using sectors in the United States.
- For the euro area, the decline in aggregate labor productivity growth in the second half of the 1990s is largely explained by declines in non-ICT capital deepening and by faster acceleration in hours worked rather than by markedly slower technological progress once measurement and coverage issues are corrected (per euro-3 analysis).
- Slower capital deepening emerges as the most important proximate cause of the euro-area labor productivity slowdown in the second half of the 1990s; structural labor market changes that reduced wage-setting pressures contributed materially to this decline in capital deepening.

*Source: Appendix III and supporting sections of the IMF working paper (industry classifications, tables, and analysis provided in the supplied content).*

### 2. Other factors, such as drops in the user cost of capital because of declining interest rates

### 2. Other factors, such as drops in the user cost of capital because of declining interest rates

### Key empirical results and statistics
- Table 11 (Elasticity of Capital Deepening to Wage-Setting Shocks) — dependent variable: ∆ln(Kijt/Lijt)
  - WS shock: 0.64 *
    - Standard error: (0.31)
    - * stands for significant at the 5 percent level.
  - country dummies: yes
  - industry dummies: yes
  - time dummies: yes
  - industry*time dummies: yes
  - country*time dummies: yes
  - country*industry: yes
  - Adj. R2: 0.40
  - Number of observations: 1,690
  - Number of industries: 26
  - Sample period: 1980-2000
  - Sources: GGDC; AMECO database; and author's estimates.
- Notes:
  - 1 Estimation uses industry-level data for France, Germany and the Netherlands. Standard errors are shown in parentheses and are corrected for AR(1) residuals.
  - 2 Wage-setting shocks measured as shown in equation (7). Consumer prices are measured by the implicit deflator for private consumption expenditures.

### Main findings on productivity and capital deepening
- Slower capital deepening—resulting from structural labor market changes—is identified as the main culprit behind the labor productivity slowdown in the euro area since the mid-1990s.
- The commitment of euro-area countries to increasing employment rates (Lisbon Summit targets, 2000) implies labor productivity growth might be dampened for many more years because higher labor utilization can reduce capital deepening.
- A large part of the labor productivity growth gap with respect to the United States can be explained by the surge in TFP growth in ICT-using sectors in the United States.
- Even if slower capital deepening in the euro area appears to be a by-product of “good” changes (increased labor utilization), flat TFP growth rates in sectors that are particularly dynamic in the United States are a concern.

### Policy analysis and recommendations
- Labor market reforms
  - Reforms necessary to continue reabsorbing people into jobs will probably ultimately increase TFP growth.
  - Reforms should aim at:
    - Increasing the incentives to work vis-à-vis receiving social benefits.
    - Correcting incentives for human capital formation, with labor income better reflecting individual abilities and efforts.
  - Expected outcomes: increase in human capital accumulation and better allocation of labor across uses, boosting TFP growth in the long term.
- Product market reforms
  - Analysis points to the need for reforms in specific sectors, particularly ICT-using services (wholesale trade, retail trade, financial intermediation).
  - Evidence is mixed; sectoral examples:
    - Retail food sector: labor productivity was 7 percent higher in France than in the United States in 2000; degree of IT use was about the same in France, Germany and the United States in 1999.
    - Barriers to firm entry and exit in the retail sector in Europe could underlie some productivity differentials; replacement of less productive by more productive establishments contributed to U.S. retail productivity growth in the 1990s.
  - Policy implications: lowering barriers to entry and easing regulatory burdens on enterprise creation in Europe seem necessary.
- Limitations of deregulation alone
  - European Commission (2003) simulations: even relatively rapid deregulation toward U.S. levels would not produce sufficiently large productivity gains over the next seven years to close the efficiency gap with the United States.
  - Gains from deregulation and privatizations may be static efficiency gains, not the dynamic efficiency gains needed to expand the technological frontier.
- Complementary measures to foster technological progress
  - Product market reforms could positively affect risk-taking activities that drive technological progress.
  - Evidence linking product market reforms to future labor market reforms suggests reforms can be mutually reinforcing and lead to increased human capital accumulation.
  - Investments in education and R&D:
    - Long-run productivity gains from education and R&D would directly increase TFP growth.
    - For R&D, focus should be on creating conditions that promote an endogenous increase in research spending rather than directly boosting R&D spending.
    - Two main channels to stimulate endogenous R&D growth:
      - Higher product market integration (e.g., completion of the single market program).
      - An investment environment that ensures development of a more active market for risk capital.

### Interpretation and open questions
- It is possible to argue that the euro area is lagging the United States primarily in the adoption of ICT technologies in some service-sector industries rather than in frontier technology creation.
- Product market reforms and other structural changes would likely speed the diffusion of technology in the euro area, but diffusion may nonetheless occur over time without dramatic policy moves.
- Evaluating whether the euro area is merely lagging in ICT adoption (and will catch up) versus structurally falling behind in dynamic TFP generation is outside the scope of the paper and left for future research.

### Data and methodological notes (databases and aggregation)
- Industry Labor Productivity Database (GGDC)
  - Contains information on value added, employment, and hours worked for 56 industries between 1979 and 2001 for the 15 EU member states and the United States.
  - Data series include: value added in current and constant prices (at basic prices), numbers of persons engaged, number of employees, total labor compensation, and working hours.
  - Homogenized treatment of quality changes in computer and semiconductor prices across all countries by using harmonized U.S. deflators for six ICT producing industries; U.S. value-added deflators corrected for differences in overall inflation between each country and the United States.
  - Törnqvist method of aggregation used to approximate an ideal Fisher price index for industry aggregates.
- Growth Accounting Database (GGDC)
  - Provides information for France, Germany, the Netherlands, the United Kingdom (not used here), and the United States for 1980–2000 with 26 industries.
  - Aggregations by ICT taxonomy based on a mapping between Appendix III listings and the 26 industries.
  - Labor share and ICT capital income share in total capital income provided; constant returns to scale assumed so the share of each capital type on value added can be recovered.
  - Labor quality changes measured by dividing total hours by skill level, weighting growth by wage share, and subtracting total hours.
  - Capital input measured with a Törnqvist capital service index comprising three ICT assets (software, computers, communications equipment) and three non-ICT assets (non-ICT equipment, structures, vehicles); shares based on user cost.
- Productivity growth accounting equation used (as presented):
  - ∆y = αl ∆l + αq ∆q + αict ∆kict + αnict ∆knict + ∆tfp
  - Rearranged decomposition for labor productivity growth (p):
    - ∆p = αict ∆(kict/l) + αnict ∆(knict/l) + ∆q + ∆tfp

*Source: _wp04200 - 2. Other factors, such as drops in the user cost of capital because of declining interest rates*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04200.pdf_