## 1. Growth Accounting and Labor Productivity

## Source details

**Canonical URL:** [1. Growth Accounting and Labor Productivity](https://www.imf.org/-/media/files/publications/selected-issues-papers/2026/english/sipea2026083.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2026/english/sipea2026083.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2026/english/sipea2026083.pdf.json)

---

### Introduction
- Iceland combines very high GDP per capita with comparatively low labor productivity and a sharp slowdown in productivity growth over the last decade.
- Recent growth has relied increasingly on high labor utilization rather than productivity growth.
- Four main factors explaining why firm-level productivity gains have not translated into stronger aggregate productivity:
  - Capital and labor are not allocated toward the most productive firms.
  - Productivity convergence occurs toward a frontier that itself remains below the European one.
  - Low-productivity firms remain in the market for too long.
  - Productive firms face a penalty when they try to scale.
- Possible policy options highlighted:
  - Improving allocative efficiency.
  - Strengthening entry/exit conditions and competition.
  - Improving framework conditions to allow productive firms to expand, including further developing the nascent venture capital and startup ecosystem and increasing access to skilled labor.

### Aggregate Backdrop
- Growth accounting (national accounts) key patterns:
  - The contribution of total factor productivity to output growth fell sharply between the pre-crisis and post-crisis periods.
  - The contribution of labor rose while capital deepening slowed.
  - Chart annotations indicate "2.9% p.a." and "0.8% p.a." (labels present in the figure).
- How Iceland sustains strong GDP per capita despite lower labor productivity:
  - Labor productivity is about "10 percent" below the Nordic-4 average.
  - The employment-population ratio is around "9 percentage points" above the Nordic-4 average.
- GEAD (generalized exactly additive decomposition) results:
  - The pure within-sector productivity effect (PPE) dominates the decomposition.
  - PPE falls from "14.8 percentage points" over 2001–14 to "3.7" over 2015–25.
  - Both reallocation level effects (price and labor) remained positive, but the labor reallocation contribution declined meaningfully; the reallocation growth effect was a small drag over the last 10 years.
- Sectoral reallocation compounding the slowdown:
  - Sectors gaining employment share (tourism, construction, business services) have weaker within-sector productivity growth.
  - Sectors losing labor share (manufacturing, fishing, finance) have stronger within-sector performance.
  - The absorption of a rapidly growing labor force into lower-productivity, labor-intensive sectors has reinforced aggregate productivity decline.

### Data and Measurement
- Firm-level data source and sample:
  - IMF RES-ORBIS database (Díez et al, 2021), sample covers 2000–2023.
  - About "4,800 observations per year" in labor-productivity specifications (with a two-employee floor).
  - Roughly "10,000–11,000" in Olley–Pakes and misallocation specifications (full firm-size distribution).
- Labor productivity measure:
  - Real operating turnover per employee (turnover deflated by producer price indices) used due to low value-added coverage in ORBIS.
- NACE2 sector exclusions and definitions applied consistently.
- Data limitations and coverage shifts:
  - Turnover-based productivity includes intermediate inputs and is less suited for cross-sectional level comparisons but informative for dynamics over time.
  - ORBIS coverage in Iceland increases sharply in 2016 (approximately doubling the number of firms captured). Analyses focusing on firm dynamics emphasize the post-2016 period; sensitivity checks compare pre- and post-2016.

### Drivers of the Aggregate Productivity Slowdown: Firm-Level Evidence
- Firm-level decomposition (Decker et al. 2017 analog) findings:
  - Within-firm productivity improvements (continuing firms becoming more productive) generate the largest share of overall productivity growth over time.
  - The typical continuing firm keeps improving; aggregate within-sector slowdown does not primarily reflect firms themselves becoming less productive.
- Entry, exit, and reallocation effects:
  - Firm entry has subtracted since around 2010 as new firms enter below incumbent firm productivity levels.
  - The exit margin has been positive over the last 15 years but its contribution has weakened over time.
  - Reallocation among surviving/continuing firms is close to zero on average.
  - Conclusion: mechanisms that normally translate firm-level improvements into stronger aggregate productivity (resource reallocation, firm entry/exit) have become less effective.

### Rapid Convergence to a Low Ceiling
- Laggard-frontier catch-up:
  - Estimation of the Griffith et al. (2004) error-correction catch-up model indicates Iceland’s national catch-up coefficient is the highest among economies in the sample, implying rapid domestic diffusion.
  - Laggard firms converge rapidly toward the national frontier.
- Convergence to the European frontier:
  - Convergence to the European frontier is consistently weaker than to the national frontier.
  - Largest shortfalls are concentrated in sheltered and professional services.
  - After improving between the mid-2000s and mid-2010s, the absolute productivity gap of Icelandic frontier firms relative to the European frontier has stopped narrowing.
- Implication:
  - Within-firm gains persist because laggards catch up to the national frontier, but the national frontier itself remains below Europe’s best performers and advances slowly. Rapid catch-up to a relatively low and slowly moving frontier compresses the productivity distribution without raising aggregate productivity.

### Resource Misallocation
- Measurement approach:
  - Misallocation proxied by dispersion in marginal revenue products of capital and labor (Hsieh–Klenow 2009 framework); greater dispersion implies larger gains from reallocating resources.
- Key patterns:
  - Capital and labor misallocation increased into the mid-2010s and remains elevated, with no sign of compression that would indicate improved allocative efficiency.
  - Dispersion in returns is highest in professional services and tourism; dispersion is lowest in manufacturing.
  - Interpretation: manufacturing faces stronger international product-market discipline; sheltered sectors exhibit the largest distortions, suggesting weak competition and limited market discipline are important drivers of misallocation.

### Within-sector allocative efficiency and size-productivity link
- Olley–Pakes decomposition insight:
  - Aggregate productivity = unweighted average of firm productivity + covariance between firm size and productivity.
- Empirical pattern:
  - Within-sector covariance rises through 2009–11 and then declines toward much lower levels, indicating output is less concentrated in the most productive firms than a decade ago.
  - Sheltered sectors and certain services industries: output is not concentrated in the most productive firms.
- Within-sector allocative-efficiency measures (MRPK/MRPL, standard deviation) by period: "0.89, 0.95, 1.01, 1.02, 1.00, 0.73, 0.76, 0.82, 0.81, 0.82" (as presented).
- Sectoral dispersion (standard deviation, 2016-2023 average) highlights high marginal revenue product dispersion across sectors such as Manufacturing, Wholesale & retail, Fishing & aquaculture, Transportation & storage, Admin & support svcs, Construction, ICT, Tourism, Prof. & scientific svcs.

### Cross-sector allocation: labor versus capital
- Cross-sector Olley–Pakes covariance:
  - Capital: covariance between sector size and marginal product is close to zero (suggesting roughly efficient capital allocation across sectors).
  - Labor: covariance is large and rising (suggesting increasing labor misallocation across sectors).
- Implications:
  - Larger, higher-value-added sectors appear systematically understaffed.
  - Labor has flowed into labor-intensive, lower-productivity activities—consistent with a rapidly expanding, largely immigrant labor supply being absorbed by lower-productivity activities while capital- and skill-intensive high-productivity sectors drew proportionately less.
- Cross-sector Olley–Pakes covariance between sector size and productivity has declined steadily since 2010: output is not concentrated in the most productive sectors.

### Exit failure and zombie firms
- Zombie prevalence and productivity gap:
  - Zombie share has roughly doubled since the mid-2000s to around "11 percent" in 2023.
  - Productivity gap between healthy firms and zombie firms has widened to around "45 percent".
- Zombie firm definition used:
  - Firms older than "10 years" with two consecutive years of negative net income.
- Implication:
  - Increasing share of capital and labor tied up in older, less productive firms that are not exiting efficiently.

### Transition dynamics and persistence
- Productivity-quintile transition (bottom quintile firms over four years):
  - Among firms starting in the bottom quintile: roughly half exit and about one quarter remain in the bottom quintile after four years (similar patterns for entrants and incumbents).
  - Cohort evidence: little systematic improvement in productivity with age; survival alone does not imply convergence toward best practice.
- Exit trends by cohort:
  - Exit rate of bottom-quintile incumbents falls from about two-thirds before the global financial crisis to under one-half after 2016.
  - Share remaining in the bottom quintile stays relatively stable.
  - Firms that move up mostly move only to quintile 2 (around two thirds), and movement up is associated largely with cutting employment.
- Charted category values (productivity-quintile transition, bottom quintile, percent, 4-year change): "64.4, 59.3, 49.3, 43.7, 15.6, 25.2, 26.9, 27.0, 20.0, 15.4, 23.9, 29.3" (as presented).

### Scaling penalty and entrant dynamics
- Entrant survival and growth dynamics (age 1 to age 5, 2017/18 cohorts):
  - Firms starting in bottom quintile: around "70 percent" show positive productivity growth.
  - Firms starting in top quintile: more often lose ground over five years.
  - Result: compression of the productivity distribution from both ends.
- Scaling penalty regression estimates:
  - Negative coefficients in every sector and for both incumbents and entrants.
  - Economy-wide average coefficients: about "-0.30" for incumbents and "-0.36" for entrants (i.e., for each percentage point increase in employment, productivity declines by around 0.3 percent).
  - Penalty largest in tourism and administrative & support services, but present across the economy.
- Behavioral pattern:
  - Top-quintile entrants expand employment aggressively and incur large scaling penalties.
  - Bottom-quintile firms that improve often do so without employment expansion and frequently by shedding labor.

### Three reinforcing failures driving the productivity slowdown
- First: Weak reallocation both within and across sectors:
  - Size-productivity link deteriorated; labor shifted toward lower-productivity activities; marginal-product dispersion remains high.
- Second: Low and slow-moving domestic frontier:
  - Firms converge quickly to the national frontier, but the national frontier remains below European top performers.
- Third: Weak selection and exit:
  - Higher zombie prevalence, lower-tail of distribution does not clear efficiently, and productive firms face a pervasive scaling penalty.

### Policy implications and recommended reforms
- Strengthen business dynamism, competition, and resource reallocation:
  - Lower barriers to entry and exit to improve allocative efficiency.
  - Streamline licensing and permitting, reduce administrative burdens, strengthen competition in sheltered sectors.
  - Simplify insolvency and restructuring procedures to reduce capital sunk in zombie firms and enable reallocation.
  - Greater openness to foreign investment and entry to enhance competition and diffusion of technologies and practices.
- Support expansion of productive firms:
  - Improve access to finance for innovative and high-growth firms, including deeper venture-capital ecosystem.
  - Maintain wage settlements aligned with productivity developments; strengthen coordination in wage bargaining to preserve competitiveness and support productive-firm growth.
- Raise innovation, skills, and the productivity frontier:
  - Strengthen evaluation of R&D support schemes, improve monitoring of innovation outcomes, streamline research-support mechanisms, deepen collaboration between research institutions and private sector.
  - Develop venture-capital financing and international research linkages to support commercialization and frontier innovation.
  - Implement education reforms, strengthen STEM education, expand vocational training and reskilling, and improve labor-market integration of immigrants to address skills shortages and mismatches.
  - Strengthen digital skills and support adoption of new technologies.
- Address infrastructure bottlenecks:
  - Expand energy generation and transmission capacity and improve transport connectivity to ease constraints on investment and firm expansion.
  - Timely implementation of planned infrastructure projects with strong infrastructure governance to strengthen resilience and growth conditions.

### Conclusion: allocative mechanism is central
- Core finding: the problem is allocative, not primarily technological.
  - Weak exit ties resources to low-productivity firms.
  - A stagnant domestic frontier limits catch-up gains.
  - Scaling penalty prevents productive firms from delivering aggregate gains.
- Policy focus:
  - Strengthen competition, improve restructuring and exit, and create conditions for productive firms to scale; complement these with investments in innovation, skills, and infrastructure to raise the productivity frontier and support long-run growth.

### Technical annex: methodological summary (selected points)
- GEAD decomposition (Tang and Wang, 2004; Montebello and Darmanin, 2021):
  - Aggregate labor-productivity growth 퐺푡 decomposed into PPE, RLE, and RGE (equations I.1 and I.2 reproduced in the source).
  - PPE = within-sector productivity growth term; RLE = reallocation level effect (relative-price and relative-labor components); RGE = reallocation growth effect.
- Firm-level decomposition (dynamic Olley–Pakes; Melitz and Polanec, 2015; Decker et al., 2017):
  - Aggregate change in productivity 훥푃푡 decomposed into within-continuing firms change, covariance term, entry, and exit (equations I.3–I.4).
  - Within-sector means employment-weighted; across-sector aggregation uses turnover weights.
- Capital and labor misallocation (Hsieh–Klenow, 2009 gross-output implementation):
  - Log marginal revenue products defined in equation II.5 using gross output (PPI-deflated operating turnover), capital (PPI-deflated total assets), and wage bill.
  - Dispersion computed within each NACE2-year observation and value-added-weighted up to nine-sector grouping.
- Frontier catch-up (Griffith et al., 2004 error-correction model):
  - Regression specification in II.7; frontier defined as top 10 percent within-Iceland for national frontier and top 10 percent cross-country for European frontier; catch-up coefficient 훼 measures speed of convergence.
- Scaling penalty regression:
  - Regression of change in firm productivity on change in employment over horizon h (equation II.8); scaling penalty 훾 < 0 indicates productivity falls as employment rises.
- Estimation details:
  - Incumbent regressions pool baseline windows with firm-clustered (Liang–Zeger) standard errors.
  - Entrant regressions pool 2017 and 2018 cohorts at age 1→5.

- Key exact figures and sample details reiterated:
  - PPE: "14.8 percentage points" over 2001–14; "3.7" over 2015–25.
  - Labor productivity gap: about "10 percent" below the Nordic-4 average.
  - Employment-population ratio: around "9 percentage points" above the Nordic-4 average.
  - Growth accounting figure labels: "2.9% p.a."; "0.8% p.a.".
  - Firm-level sample sizes: about "4,800 observations per year" (labor-productivity specs); roughly "10,000–11,000" (Olley–Pakes and misallocation specs).
  - ORBIS coverage change: increases sharply in 2016 (approximately doubling the number of firms captured).

*Source: IMF staff analysis in sipea2026083 (PDF chapter).*

### 1. Growth Accounting and Labor Productivity __________________________________________ 3

### 1. Growth Accounting and Labor Productivity

### Introduction
- Iceland combines very high GDP per capita with comparatively low labor productivity and a sharp slowdown in productivity growth over the last decade.
- Recent growth has relied increasingly on high labor utilization rather than productivity growth.
- Four main factors explain why firm-level productivity gains have not translated into stronger aggregate productivity:
  - Capital and labor are not allocated toward the most productive firms.
  - Productivity convergence occurs toward a frontier that itself remains below the European one.
  - Low-productivity firms remain in the market for too long.
  - Productive firms face a penalty when they try to scale.
- Possible policy options highlighted in the source:
  - Improving allocative efficiency.
  - Strengthening entry/exit conditions and competition.
  - Improving framework conditions to allow productive firms to expand, including further developing the nascent venture capital and startup ecosystem and increasing access to skilled labor.

### Aggregate Backdrop
- Growth accounting results (national accounts) show:
  - The contribution of total factor productivity to output growth fell sharply between the pre-crisis and post-crisis periods.
  - The contribution of labor rose while capital deepening slowed.
  - Chart annotations indicate 2.9% p.a. and 0.8% p.a. (labels present in the figure).
- Iceland maintains strong GDP per capita despite lower labor productivity because:
  - Labor productivity is about 10 percent below the Nordic-4 average.
  - The employment-population ratio is around 9 percentage points above the Nordic-4 average.
- GEAD (generalized exactly additive decomposition) results:
  - The pure within-sector productivity effect (PPE) dominates the decomposition.
  - PPE falls from 14.8 percentage points over 2001–14 to 3.7 over 2015–25.
  - Both reallocation level effects (price and labor) remained positive, but the labor reallocation contribution declined meaningfully; the reallocation growth effect was a small drag over the last 10 years.
- Sectoral reallocation has compounded the slowdown:
  - Sectors gaining employment share (tourism, construction, business services) have weaker within-sector productivity growth.
  - Sectors losing labor share (manufacturing, fishing, finance) have stronger within-sector performance.
  - The absorption of a rapidly growing labor force into lower-productivity, labor-intensive sectors has reinforced aggregate productivity decline.

### Data and Measurement
- Firm-level data source: IMF RES-ORBIS database (Díez et al, 2021), sample covers 2000–2023.
- Labor productivity measure: real operating turnover per employee (turnover deflated by producer price indices) due to low value-added coverage in ORBIS.
- NACE2 sector exclusions and definitions applied consistently (see source for exact list).
- Sample sizes vary by specification:
  - About 4,800 observations per year in labor-productivity specifications (with a two-employee floor).
  - Roughly 10,000–11,000 in Olley–Pakes and misallocation specifications (full firm-size distribution).
- Data limitations noted:
  - Turnover-based productivity includes intermediate inputs and is less suited for cross-sectional level comparisons but informative for dynamics over time.
  - ORBIS coverage in Iceland increases sharply in 2016 (approximately doubling the number of firms captured). Analyses focusing on firm dynamics emphasize the post-2016 period; sensitivity checks compare pre- and post-2016.

### Drivers of the Aggregate Productivity Slowdown: Firm-Level Evidence
- Firm-level decomposition (Decker et al. 2017 analog):
  - Within-firm productivity improvements (continuing firms becoming more productive) generate the largest share of overall productivity growth over time.
  - The typical continuing firm keeps improving, so the aggregate within-sector slowdown does not primarily reflect firms themselves becoming less productive.
- Entry, exit, and reallocation effects:
  - Firm entry has subtracted since around 2010 as new firms enter below incumbent firm productivity levels.
  - The exit margin has been positive over the last 15 years but its contribution has weakened over time.
  - Reallocation among surviving/continuing firms is close to zero on average.
  - Conclusion: mechanisms that normally translate firm-level improvements into stronger aggregate productivity (resource reallocation, firm entry/exit) have become less effective.

### Rapid Convergence to a Low Ceiling
- Laggard-frontier catch-up:
  - Estimation of the Griffith et al. (2004) error-correction catch-up model indicates Iceland’s national catch-up coefficient is the highest among economies in the sample, implying rapid domestic diffusion.
  - Laggard firms converge rapidly toward the national frontier.
- Convergence to the European frontier:
  - Convergence to the European frontier is consistently weaker than to the national frontier.
  - Largest shortfalls are concentrated in sheltered and professional services.
  - After improving between the mid-2000s and mid-2010s, the absolute productivity gap of Icelandic frontier firms relative to the European frontier has stopped narrowing (figure shows frontier-frontier log-point gap series by year).
- Implication: within-firm gains persist because laggards catch up to the national frontier, but the national frontier itself remains below Europe’s best performers and advances slowly. Rapid catch-up to a relatively low and slowly moving frontier compresses the productivity distribution without raising aggregate productivity.

### Resource Misallocation
- Misallocation measurement (Hsieh-Klenow 2009 framework):
  - Allocative efficiency is proxied by dispersion in marginal revenue products of capital and labor; greater dispersion implies larger gains from reallocating resources.
  - Capital and labor misallocation increased into the mid-2010s and remains elevated, with no sign of compression that would indicate improved allocative efficiency.
- Sectoral concentration of misallocation:
  - Dispersion in returns is highest in professional services and tourism.
  - Dispersion is lowest in manufacturing.
  - Interpretation: manufacturing faces stronger international product-market discipline; sheltered sectors exhibit the largest distortions, suggesting weak competition and limited market discipline are important drivers of misallocation.

### Key statistics and exact figures from the source
- PPE (pure within-sector productivity effect): 14.8 percentage points over 2001–14; 3.7 over 2015–25.
- Labor productivity gap: about 10 percent below the Nordic-4 average.
- Employment-population ratio: around 9 percentage points above the Nordic-4 average.
- Growth accounting figure labels: 2.9% p.a.; 0.8% p.a. (presented in the source figures).
- Firm-level sample sizes: about 4,800 observations per year (labor-productivity specs); roughly 10,000–11,000 (Olley–Pakes and misallocation specs).
- ORBIS coverage change: increases sharply in 2016 (approximately doubling the number of firms captured).

*Source: IMF staff chapter “1. Growth Accounting and Labor Productivity” (Iceland, July 10, 2026).*

### 16.      Productive firms are also no longer the largest firms, pointing to a deterioration in

### 16.      Productive firms are also no longer the largest firms, pointing to a deterioration in

### Within-sector allocative efficiency and size-productivity link
- Olley-Pakes (1996) decomposition: aggregate productivity = unweighted average of firm productivity + covariance between firm size and productivity.
- Empirical pattern:
  - Within-sector covariance rises through 2009–11 and then declines toward much lower levels, indicating output is less concentrated in the most productive firms than a decade ago.
  - Sheltered sectors and certain services industries: output is not concentrated in the most productive firms.
- Within-sector allocative-efficiency measures (MRPK/MRPL, standard deviation) by period:
  - 2000-2004: 0.89, 0.95, 1.01, 1.02, 1.00, 0.73, 0.76, 0.82, 0.81, 0.82 (as presented)
- Sectoral dispersion (standard deviation, 2016-2023 average) highlights high marginal revenue product dispersion across sectors such as Manufacturing, Wholesale & retail, Fishing & aquaculture, Transportation & storage, Admin & support svcs, Construction, ICT, Tourism, Prof. & scientific svcs.

### Cross-sector allocation: labor versus capital
- Cross-sector Olley-Pakes covariance:
  - Capital: covariance between sector size and marginal product is close to zero (suggesting roughly efficient capital allocation across sectors).
  - Labor: covariance is large and rising (suggesting increasing labor misallocation across sectors).
- Implications:
  - Larger, higher-value-added sectors appear systematically understaffed.
  - Labor has flowed into labor-intensive, lower-productivity activities—consistent with a rapidly expanding, largely immigrant labor supply being absorbed by lower-productivity activities while capital- and skill-intensive high-productivity sectors drew proportionately less.
- Cross-sector Olley-Pakes covariance between sector size and productivity has declined steadily since 2010: output is not concentrated in the most productive sectors.

### Exit failure and zombie firms
- Zombie firm prevalence and productivity gap:
  - Zombie share has roughly doubled since the mid-2000s to around 11 percent in 2023.
  - Productivity gap between healthy firms and zombie firms has widened to around 45 percent.
- Zombie firm definition used: firms older than 10 years with two consecutive years of negative net income.
- Implication: increasing share of capital and labor tied up in older, less productive firms that are not exiting efficiently.

### Transition dynamics and persistence
- Productivity-quintile transition (bottom quintile firms over four years):
  - Among firms starting in the bottom quintile: roughly half exit and about one quarter remain in the bottom quintile after four years (similar patterns for entrants and incumbents).
  - Cohort evidence: little systematic improvement in productivity with age; survival alone does not imply convergence toward best practice.
- Exit trends by cohort:
  - Exit rate of bottom-quintile incumbents falls from about two-thirds before the global financial crisis to under one-half after 2016.
  - Share remaining in the bottom quintile stays relatively stable.
  - Firms that move up mostly move only to quintile 2 (around two thirds), and movement up is associated largely with cutting employment.
- Charted category values (productivity-quintile transition, bottom quintile, percent, 4-year change): 64.4, 59.3, 49.3, 43.7, 15.6, 25.2, 26.9, 27.0, 20.0, 15.4, 23.9, 29.3 (as presented).

### Scaling penalty and entrant dynamics
- Entrant survival dynamics (age 1 to age 5, 2017/18 cohorts):
  - Firms starting in bottom quintile: around 70 percent show positive productivity growth.
  - Firms starting in top quintile: more often lose ground over five years.
  - Result: compression of the productivity distribution from both ends.
- Scaling penalty (regression estimates of productivity growth on employment growth):
  - Negative coefficients in every sector and for both incumbents and entrants.
  - Economy-wide average coefficients: about -0.30 for incumbents and -0.36 for entrants (i.e., for each percentage point increase in employment, productivity declines by around 0.3 percent).
  - Penalty largest in tourism and administrative & support services, but present across the economy.
- Behavioral pattern:
  - Top-quintile entrants expand employment aggressively and incur large scaling penalties.
  - Bottom-quintile firms that improve often do so without employment expansion and frequently by shedding labor.

### Three reinforcing failures driving the productivity slowdown
- First: Weak reallocation both within and across sectors:
  - Size-productivity link deteriorated; labor shifted toward lower-productivity activities; marginal-product dispersion remains high.
- Second: Low and slow-moving domestic frontier:
  - Firms converge quickly to the national frontier, but the national frontier remains below European top performers.
- Third: Weak selection and exit:
  - Higher zombie prevalence, lower-tail of distribution does not clear efficiently, and productive firms face a pervasive scaling penalty.

### Policy implications and recommended reforms
- Strengthen business dynamism, competition, and resource reallocation:
  - Lower barriers to entry and exit to improve allocative efficiency.
  - Streamline licensing and permitting, reduce administrative burdens, strengthen competition in sheltered sectors.
  - Simplify insolvency and restructuring procedures to reduce capital sunk in zombie firms and enable reallocation.
  - Greater openness to foreign investment and entry to enhance competition and diffusion of technologies and practices.
- Support expansion of productive firms:
  - Improve access to finance for innovative and high-growth firms, including deeper venture-capital ecosystem.
  - Maintain wage settlements aligned with productivity developments; strengthen coordination in wage bargaining to preserve competitiveness and support productive-firm growth.
- Raise innovation, skills, and the productivity frontier:
  - Strengthen evaluation of R&D support schemes, improve monitoring of innovation outcomes, streamline research-support mechanisms, deepen collaboration between research institutions and private sector.
  - Develop venture-capital financing and international research linkages to support commercialization and frontier innovation.
  - Implement education reforms, strengthen STEM education, expand vocational training and reskilling, and improve labor-market integration of immigrants to address skills shortages and mismatches.
  - Strengthen digital skills and support adoption of new technologies.
- Address infrastructure bottlenecks:
  - Expand energy generation and transmission capacity and improve transport connectivity to ease constraints on investment and firm expansion.
  - Timely implementation of planned infrastructure projects with strong infrastructure governance to strengthen resilience and growth conditions.

### Conclusion: allocative mechanism is central
- Core finding: the problem is allocative, not primarily technological.
  - Weak exit ties resources to low-productivity firms.
  - A stagnant domestic frontier limits catch-up gains.
  - Scaling penalty prevents productive firms from delivering aggregate gains.
- Policy focus: strengthen competition, improve restructuring and exit, and create conditions for productive firms to scale; complement these with investments in innovation, skills, and infrastructure to raise the productivity frontier and support long-run growth.

*Source: IMF staff analysis in sipea2026083 (PDF chapter).*

### References

### sipea2026083 - References

### References cited
- Abdulla, K., 2025, “Quantifying the Productivity Effects of Education–Job Mismatch: A Cross-Country Analysis,” Education Economics 33 (4): 595–612.
- Adalet McGowan, M., and D. Andrews, 2018, “Design of Insolvency Regimes Across Countries,” OECD Economics Department Working Paper No. 1504, OECD Publishing, Paris.
- Adalet McGowan, M., D. Andrews, and V. Millot, 2018, “The Walking Dead? Zombie Firms and Productivity Performance in OECD Countries,” Economic Policy, Vol. 33(96), pages 685-736.
- Adilbish, O., D. Cerdeiro, R. Duval, G. Hee Hong, L. Mazzone, L. Rotunno, H. Toprak, and M. Vaziri, 2025, “Europe's Productivity Weakness: Firm-Level Roots and Remedies,” IMF Working Paper WP/25/40, International Monetary Fund, Washington, DC.
- Andrews, D., and F. Petroulakis, 2019, “Breaking the Shackles: Zombie Firms, Weak Banks and Depressed Restructuring in Europe.” ECB Working Paper No. 2240, European Central Bank, Frankfurt.
- Bank of England, 2025, “Unlocking Growth: What Can the Literature Tell Us About What's Holding Back High-Growth Firms?” Bank Overground, October 2.
- Cho, W. H., C. Criscuolo, I. Desnoyers-James, M. Reinhard, and R. Verlhac, 2024, “Diagnosis and policy action for sustainable and inclusive productivity growth," OECD Science, Technology and Industry Working Papers 2024/7, OECD Publishing.
- Decker, R., J. Haltiwanger, R. Jarmin, and J. Miranda, 2017, “Declining Dynamism, Allocative Efficiency, and the Productivity Slowdown,” American Economic Review Papers & Proceedings, 107(5), 322–326.
- Díez, F., J. Fan, and C. Villegas-Sanchez, 2021, “Global Declining Competition,” Journal of International Economics 132 (2021): 103492.
- Griffith, R., S. Redding, and J. Van Reenen, 2004, “Mapping the Two Faces of R&D: Productivity Growth in a Panel of OECD Industries,” Review of Economics and Statistics, 86(4), 883–895.
- Hsieh, C.-T., and P. Klenow, 2009, “Misallocation and Manufacturing TFP in China and India,” Quarterly Journal of Economics, 124(4), 1403–1448.
- Hu, B. and Ugur, C., 2025, “Allocative Efficiency, Firm Dynamics, and Productivity in Latvia,” IMF Selected Issues Paper SIP/2025/135.
- IMF, 2025, “Iceland—2025 Article IV Consultation Staff Report,” Country Report No. 25/141.
- Kim, H., 2026, “Transportation Infrastructure and Total Factor Productivity: Development Heterogeneity and Resilience under Adverse Shocks,” World Bank Policy Research Working Paper No. 11410, World Bank, Washington, DC.
- Melitz, M., and S. Polanec, 2015, “Dynamic Olley-Pakes Productivity Decomposition with Entry and Exit,” RAND Journal of Economics, 46(2), 362–375.
- Montebello, R., and J. Darmanin, 2021, "Labour productivity growth in Malta: a sectoral decomposition analysis," CBM Working Papers WP/04/2021, Central Bank of Malta.
- Mosiashvili, N., and Y. Jin, 2025, “Firm-Level Productivity Dynamics in Iceland,” OECD Economics Department Working Papers, No. 1847.
- OECD, 2013, OECD Reviews of Innovation Policy: Sweden 2012, Paris: OECD Publishing.
- OECD, 2024, “Productivity and Business Dynamism,” OECD Topics page.
- Olley, G. S., and A. Pakes, 1996, “The Dynamics of Productivity in the Telecommunications Equipment Industry,” Econometrica, 64(6), 1263–1297.
- Tang, J., and W. Wang, 2004, “Sources of Aggregate Labour Productivity Growth in Canada and the United States,” Canadian Journal of Economics, 37(2), 421–444.

### Annex I. Technical Annex — Methodological summary
- The Technical Annex presents methods used for decomposition and firm-level analysis, reproducing specifications from Tang and Wang (2004), Montebello and Darmanin (2021), Melitz and Polanec (2015), Decker et al. (2017), Hsieh and Klenow (2009), Griffith et al. (2004), and related applications.

H3: A. Generalized Exactly Additive Decomposition (GEAD)
- Aggregate labor-productivity growth 퐺푡 is decomposed following Tang and Wang (2004) with the price/labor extension of Montebello and Darmanin (2021). The three-term form:
  - 퐺푡 = ∑ 푌푖,푡−1/푌푡−1 퐺푖,푡 푖 + ∑ 푍푖,푡−1/푧푡−1 훥(푝푖,푡/푙푖,푡) 푖 + ∑ 푍푖,푡−1/푧푡−1 훥(푝푖,푡/푙푖,푡)퐺푖,푡 푖  (I.1)
- Definitions:
  - 퐺푖,푡: within-sector productivity growth.
  - 푌푖,푡−1/푌푡−1: nominal value-added share.
  - 푝푖,푡: relative price level.
  - 푙푖,푡: employment share.
  - 푍푖,푡−1/푧푡−1: sector's relative productivity level.
- Interpretation:
  - First term = pure productivity effect (PPE).
  - Second term = reallocation level effect (RLE), decomposable into relative-price and relative-labor components.
  - Third term = reallocation growth effect (RGE).
- Expanded algebraically equivalent form reproduced (I.2) with explicit RLPE, RLLE, and interaction terms:
  - 퐺푡 = ∑ 푌푖,푡−1/푌푡−1 퐺푖,푡 + 푌푖,푡−1/푌푡−1 (푝푖,푡/푝푖,푡−1 − 1) + 푌푖,푡−1/푌푡−1 (푙푖,푡/푙푖,푡−1 − 1) + 푌푖,푡−1/푌푡−1 (푝푖,푡/푝푖,푡−1 − 1)(푙푖,푡/푙푖,푡−1 − 1) + … + interaction terms × 퐺푖,푡  (I.2)
- Note: The summation runs over all sectors i.

H3: B. Firm-Level Decomposition (Decker, dynamic Olley–Pakes)
- Uses dynamic Olley–Pakes form (Melitz and Polanec, 2015; Decker et al., 2017) and applied in Hu and Ugur (2025).
- Aggregate change in productivity:
  - 훥푃푡 = ∑ 푤푠 ( 훥푝푠퐶 + 훥푐표푣푠퐶 + 푒푛푡푠 + 푒푥푡푠 )푠  (I.3)
- Component definitions:
  - 푐표푣푠퐶 = 푃푠퐶 − 푝푠퐶
  - 푒푛푡푠 = 휃푠퐸(푃푠퐸 − 푃푠퐶)
  - 푒푥푡푠 = 휃푠,푡−1푋(푃푠,푡−1퐶 − 푃푠,푡−1푋)  (I.4)
- Variables:
  - 푤푠: turnover share used to aggregate across sectors.
  - 푝푠퐶: unweighted mean labor productivity for continuing firms.
  - 푃푠퐶: employment-weighted productivity mean for continuing firms.
  - 푐표푣푠퐶: Olley–Pakes covariance between employment share and productivity.
  - 푃푠퐸, 푃푠푋: employment-weighted mean productivity for entering and exiting firms.
  - 휃푠퐸, 휃푠푋: employment shares for entrants and exits.
- Notes:
  - Within-sector means and shares are employment-weighted; across sectors aggregation uses turnover weights because employment coverage is 70–90 percent whereas turnover coverage is near-complete.
  - Results are broadly robust to weighting scheme.

H3: C. Capital and Labor Misallocation (Hsieh–Klenow)
- Implements the gross-output version of Hsieh and Klenow (2009).
- Firm i in sector s production with Cobb–Douglas; gross-output capital elasticity 훼푠, labor elasticity 훽푠 = 1 − 훼푠 − 푚푠, intermediate-input share 푚푠.
- Log marginal revenue products:
  - 푙푛 푀푅푃퐾푖푠 = 푙푛 훼푠 + 푙푛 푄푖푠 − 푙푛 퐾푖푠
  - 푙푛 푀푅푃퐿푖푠 = 푙푛 훽푠 + 푙푛 푄푖푠 − 푙푛 푊푖푠  (II.5)
- Variable definitions and data sources:
  - 푄푖푠: gross output (PPI-deflated operating turnover).
  - 퐾푖푠: capital (PPI-deflated total assets).
  - 푊푖푠 = 푤푖푠 퐿푖푠: wage bill (푤푖푠 wage per worker, 퐿푖푠 employment).
  - Data sourced from ORBIS; elasticities calibrated using national accounts from Statistics Iceland.
- Dispersion computed within each Nace2-year observation and value-added-weighted up to nine-sector grouping; results robust to weighting.

H3: D. Olley–Pakes Decomposition
- Aggregate productivity equals unweighted firm mean plus covariance between size and productivity (Olley and Pakes, 1996):
  - Φ푡 = 휙̅푡 + ∑ (푠푖푡 − 푠̄푡)(휑푖푡 − 휑̄푡)푖 = 휑̄푡 + 푐표푣(푠푖푡, 휑푖푡)  (II.6)
- Interpretations:
  - Φ푡: aggregate (turnover-weighted) labor productivity.
  - 휙̅푡: weighted mean of firm-level productivity.
  - 푠푖푡: firm i's share of sector turnover; 푠̄푡: cross-firm mean.
  - A larger covariance indicates larger firms are more productive; a declining covariance signals deteriorating allocation.
- Results robust to weighting scheme.

H3: E. Frontier Catch-Up (Griffith et al., 2004)
- Error-correction regression of firm productivity growth on lagged gap to frontier:
  - 훥ℎ 푙푛 퐴푖푠푡 = 훽 훥ℎ 푙푛 퐴푠,푡퐹 + 훼 ( 푙푛 퐴푠,푡−ℎ퐹 − 푙푛 퐴푖,푡−ℎ ) + 훿′푋푖𝑡 + 휂𝑠 + 휏𝑡 + 휀𝑖𝑡  (II.7)
- Elements:
  - 퐴푖푠푡: firm labor productivity in sector s.
  - 퐴푠퐹: sector-s frontier productivity.
  - 푋푖𝑡: firm controls (including age and size).
  - 훥ℎ operator uses h = 3 in baseline.
  - Frontier defined as top 10 percent of within-Iceland productivity distribution for national frontier or top 10 percent of cross-country distribution for European frontier.
  - 훼: catch-up coefficient (error-correction speed; 훼 > 0 and larger 훼 = faster convergence).
  - 훽: response to contemporaneous frontier growth.

H3: F. Scaling Penalty Regression
- Regresses change in firm productivity on change in employment over horizon h:
  - 훥ℎ 퐴푖 = 훾 훥ℎ 퐿푖 + 훽′푋푖 + 훿𝑠 + 휀𝑖  (II.8)
- Variables and interpretation:
  - 훥ℎ 푙푛 퐿𝑃푖: h-period change in firm i's log labor productivity.
  - 훥ℎ 퐿𝑖: h-period change in log employment.
  - 푋𝑖: firm controls (including initial log labor productivity).
  - 훿𝑠: sector fixed effects.
  - Scaling penalty 훾 < 0 indicates productivity falls as employment rises.
- Estimation details:
  - Incumbents pool two baseline windows with firm-clustered (Liang–Zeger) standard errors.
  - Entrants pool 2017 and 2018 cohorts at age 1→5.
  - Regression estimated separately by nine-sector grouping and pooled across all sectors.

*sipea2026083 - References*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2026/english/sipea2026083.pdf_