## wpiea2022213-print-pdf — INTRODUCTION

## Source details

**Canonical URL:** [wpiea2022213-print-pdf — INTRODUCTION](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022213-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022213-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022213-print-pdf.pdf.json)

---

### Major themes and structure
- Purpose: assess productivity and product market regulations in Korea and their empirical linkages using a cross-country, industry-level dataset.
- Core sections summarized:
  - COMPARISONS OF PRODUCTIVITY LEVELS BY INDUSTRY
  - PRODUCT MARKET REGULATIONS IN KOREA (OECD PMR Indicators; industry-level PMR indicators)
  - LINKING REGULATIONS AND PRODUCTIVITY (theory and empirical literature)
  - EMPIRICAL APPROACH (System GMM, data features)
  - REGRESSION RESULTS (baseline, heterogeneity, channels, robustness)
  - CONCLUSIONS

### Data, measurement, and scope
- Productivity measures:
  - Industry-level labor productivity (LP) and total factor productivity (TFP) from Groningen Growth and Development Centre; LP has broader coverage (29 countries), TFP skewed to 19 countries.
  - Analysis focuses on manufacturing and “market services”; excludes real estate and manufacturing of coke and refined petroleum products; construction and utilities noted as measurement-challenged.
- Time and sample:
  - Generally since 1995; OECD PMR indicators available at five-year intervals from 1998 with a structural break between 2013 and 2018 vintages.
  - Temporal aggregation uses non-overlapping three-year averages yielding seven time periods; regressions estimated over up to six periods due to PMR lagging.
- PMR measurement:
  - OECD economy-wide PMR indicator on a zero to six scale (higher = more restrictive).
  - IMF product market regulation economy-wide indicator on a zero to 1 scale (higher = more restrictive).
  - Korea’s aggregate PMR: 2.6 in 1998; later vintages indicate improvement but remain somewhat more restrictive than the average OECD economy.
- Industry exposure construction:
  - Three time-invariant industry-specific measures scaled 0–1: layoff exposure (US benchmark), capital-output ratio (industry average), and skill requirement (industry average).
  - Broad exposure measure equals the simple average of the three measures; combined with time-varying country PMRs to produce country-industry-time PMR measures.

### Key stylized facts on Korea’s productivity
- Long-run trends:
  - Korea’s trend growth rate has been in long decline predating Covid-19.
  - Economy-wide productivity rose through the 1990s then plateaued at around two-thirds of the most productive economies.
  - Korea historically relied heavily on capital and labor accumulation; scope for further input-driven growth is smaller due to demographics and structural shifts toward services.
- Cross-country benchmarks (mid-1990s to 2017):
  - Mid-1990s: Korea’s overall LP was 40 percent below the advanced economy median; by 2017 gap closed to less than 15 percent.
  - Aggregate TFP: just over half the advanced economy median in 1995 and about two-thirds by 2017.
- Industry patterns (2015–17 benchmarks):
  - Manufacturing (aggregate):
    - LP and TFP gaps with respect to the frontier are both about 50 percent.
    - LP is 120 percent of the median; TFP is 80 percent of the median.
    - Industry gaps mostly range from about 30 percent to 80 percent.
    - LP closest to frontier: basic and fabricated metals.
    - TFP closest to frontier: basic and fabricated metals and transport equipment.
    - Computer and other IT equipment ranks relatively highly due to semiconductors; U.S. advantage explains remaining frontier gap.
  - Market services (aggregate):
    - Productivity gaps about 60 percent with respect to the frontier for both LP and TFP.
    - Market services LP is 30 percent below the median; TFP is 40 percent below the median.
    - LP and TFP closest to frontier in professional and business services (exceed the median).
    - Gaps of 75 percent or more with respect to the frontier (about 35 to 50 percent with respect to the median) in wholesale and retail trade, transportation and storage, information and communication, and personal services.

### Main empirical approach
- Theoretical framework: Neo-Schumpeterian, with productivity drivers depending on distance to technological frontier (DTF).
- Core regression features:
  - Panel System GMM to handle short T dimension and dynamic panel bias; collapsed instruments using second or third lags to limit proliferation.
  - Core regressors: lagged dependent variable, lagged frontier growth, DTF (second lag), vector of PMR variables (lagged one period), time, industry, and country dummies.
  - Diagnostics: AR(2) for residual second-order autocorrelation in differences; Hansen test for instrument exogeneity; joint F-tests for PMR terms.
- Temporal treatment: use of non-overlapping three-year averages and lag structure to match PMR intervals.

### Baseline econometric findings and magnitudes
- Dynamics and convergence:
  - Lagged LP growth, lagged frontier growth, and lagged DTF significantly predict LP growth; similar for TFP though with weaker frontier spillovers.
- PMR effects:
  - PMR linear term often insignificant; PMR × productivity gap (DTF interaction) frequently significant and negative — industries closer to frontier are more adversely affected.
  - Weighting PMR by industry exposure (capital-output ratio, labor skill levels, layoff exposure) improves identification; simple average exposure yields strongest relationships.
- Quantitative magnitudes (preserve precision from source):
  - One standard deviation less restrictive PMRs than the sample average is associated with LP growth higher by around 0.7 percent per year at the median productivity gap.
  - For industries with DTF at the 75th percentile, only a small impact of PMRs on productivity; at the 25th percentile effects are twice as strong as at the median.
  - For TFP growth, one standard deviation less restrictive PMRs than the sample average is around 1.2 percent per year at the median; 25th–75th percentile range estimated between 1-1.5 percent.
  - For Korea: aligning overall PMRs to the average OECD level translates to an impact on LP growth of 0.5 percent per year (median DTF) and 0.8 percent per year for TFP.
- Selected regression diagnostics (example specification — simple average exposure):
  - Observations: 2,158
  - Country-industry groups: 381
  - Instruments: 39
  - F-test p-values for PMR terms reported across specifications including 0.00 and 0.01
  - Hansen test p-values and AR(2) test p-values reported (examples: Hansen 0.47; AR(2) 0.76)

### Detailed results by PMR subcomponents
- State control:
  - Aggregate state control indicator strongly negative for LP and TFP.
  - Level 3 indicators—public ownership and involvement in business operations—strongly negative.
  - Level 4 specifics: scope of SOEs, direct control over enterprises, and price controls jointly significant on LP growth; scope of SOEs, direct control, and command-and-control regulation jointly significant on TFP growth.
  - Selected level-4 coefficients (significance retained as in source):
    - Scope of state-owned enterprises: LP direct effect -0.09***; interaction -0.16***; joint significance p-value 0.01.
- Barriers to entrepreneurship:
  - Aggregate and level 3 indicators (complexity of regulatory procedures, administrative burden on startups, regulatory protection of incumbents) associated with lower LP growth; stronger nearer the frontier.
  - Level 4: all except barriers in service sectors jointly significant on LP growth; most jointly significant on TFP growth; service sector barriers especially impactful.
  - Example coefficient: Administrative burdens for corporations: LP direct effect -0.06***; interaction -0.08***.
- Barriers to trade and investment:
  - Aggregate indicator and level 3 indicators jointly significant for LP growth but not consistently for TFP.
  - Level 4: differential treatment of foreign suppliers jointly significant for both LP and TFP.
- Industry-specific PMR indicators:
  - Restrictions in network industries have wide-ranging negative effects on productivity economy-wide.
  - Restrictions in professional services significant barriers to TFP growth.
  - Marginal economy-wide effects detected for restrictions in retail sales.
  - Example coefficient: Restrictions in network industries: LP direct effect -0.03**; interaction -0.05***.

### Industry heterogeneity and exposure
- Broad industry-group interactions (manufacturing vs market services) generally not individually significant.
- Joint tests indicate:
  - Significant PMR effects on LP and TFP for high LP industries.
  - Significant PMR effects on LP for ICT-related industries.
- Interpretation: PMRs tend to exert larger effects on high-productivity industries (e.g., high-tech manufacturing, information-intensive services), implying restrictive PMRs hinder innovation and frontier activity.

### Transmission channels and aggregate implications
- Principal channel: real value added — tighter PMRs significantly reduce real value added growth, operating principally through lower TFP.
- Factor accumulation:
  - PMRs associated with slower growth in hours worked and gross fixed capital formation, but these effects are less consistent and smaller than productivity effects.
- Prices and costs:
  - More restrictive PMRs associated with somewhat higher value added deflators and substantially higher unit labor costs (ULC).
  - ULC effects are highly significant and stronger for state control and barriers to entrepreneurship, implying costs rise more than prices and profits fall.
- Spillovers:
  - Cost increases can propagate to downstream industries via input-output linkages, affecting broader economy-wide productivity.
- Aggregate policy-relevant implication for Korea:
  - Korea’s relatively tight PMRs likely constrain productivity in services and high-tech manufacturing.
  - Reforms could support reallocation toward dynamic, high-productivity sectors such as information and communications, pharmaceuticals, and biotechnology, and improve efficiency in broader market services.

### Robustness and extensions
- Tests show PMR findings are robust to:
  - Inclusion of additional structural drivers (labor market regulation, governance, R&D share, Fraser indicators, IMF Structural Reform Database liberalization indicators) — these were not significant and did not alter PMR findings.
  - Alternative lag structures (first lag of DTF, contemporaneous frontier growth) — similar coefficients though some specifications failed Hansen tests.
  - Excluding lagged dependent variable — stronger coefficients but exogeneity concerns for TFP.
  - Expanded industry coverage and alternative dummy variable choices — main results hold.
- Diagnostic caution: instrument proliferation and Hansen test sensitivity addressed via collapsed instruments and lag selection.

### Policy priorities and conclusions
- Main empirical conclusions:
  - Tighter PMRs are associated with slower LP and TFP growth at the industry level across PMR areas and many detailed indicators.
  - Effects are larger for industries nearer the global productivity frontier and for high-productivity industries.
  - Channels operate mainly through lower TFP and higher production costs (ULC), with moderate effects on hours worked and small effects on capital accumulation.
- Korea-specific policy priorities:
  - Lowering barriers in service and network sectors.
  - Reducing restrictions applying to trade and investment.
  - Evaluating scope of government involvement in the economy.
- Quantified potential gains:
  - Reforms aligning Korea’s PMRs with the average OECD economy could raise aggregate productivity growth by half to three-quarters of a percentage point per year.
  - For Korea specifically: aligning to OECD average implies LP growth +0.5 percent per year and TFP growth +0.8 percent per year (at median DTF).

*Source: wpiea2022213-print-pdf — INTRODUCTION (IMF Working Paper).*

### INTRODUCTION ...........................................................................................................

### INTRODUCTION

### Major themes and structure
- The chapter is organized to assess productivity and product market regulations in Korea and their empirical linkages. Key sections include:
  - COMPARISONS OF PRODUCTIVITY LEVELS BY INDUSTRY (page 7)
    - Productivity measurement and scope of analysis
    - Korea’s productivity levels
  - PRODUCT MARKET REGULATIONS IN KOREA (page 12)
    - The OECD PMR Indicators
    - Korea’s standing on PMR indicators
    - Industry-level PMR indicators
  - LINKING REGULATIONS AND PRODUCTIVITY (page 16)
    - Regulation-productivity linkages
    - Findings of cross-country, industry-level empirical studies
  - EMPIRICAL APPROACH (page 20)
    - Theoretical and empirical framework
    - Data features and System GMM estimation
  - REGRESSION RESULTS (page 22)
    - Baseline specification
    - Differences across industries
    - Transmission channels
    - Model extensions and robustness checks
  - CONCLUSIONS (page 32)

### Figures and tables referenced (chapter-level inventory)
- Figures cited include:
  - Figure 1. Productivity Levels Across Advanced Economies
  - Figure 2. Aggregate Productivity Levels Over Time
  - Figure 3. Industry-Level Labor Productivity Distribution
  - Figure 4. Industry-Level Total Factor Productivity Distribution
  - Figure 5. OECD Product Market Regulations:
  - Figure 6. OECD Product Market Regulation Indicators by Area
  - Figure 7. IMF Product Market Regulation: Economy-Wide Indicator :
- Tables cited include:
  - Table 1. OECD PMR Indicators Structure (2013)
  - Table 2. Effects of Product Market Regulations on Labor Productivity: Industry Exposure
  - Table 3. Effects of Product Market Regulations on Total Factor Productivity: Industry Exposure
  - Table 4. Effects of Product Market Regulations on Labor Productivity by Broad Areas and Subcategories
  - Table 5. Effects of Product Market Regulations on Total Factor Productivity by Broad Areas and Subcategories
  - Table 6. Effects of Product Market Regulations on Productivity: Detailed Indicators
  - Table 7. Effects of Product Market Regulations Across Industries
  - Table 8. Channels for Effects of Product Market Regulation

### Analytical focus and methodology
- Empirical emphasis:
  - Industry-level comparisons of labor productivity and total factor productivity.
  - Use of OECD PMR indicators and industry-level PMR measures to characterize regulatory stance.
  - Theoretical and empirical framework assessed in EMPIRICAL APPROACH.
  - Estimation employs System GMM to address panel dynamics and data features.
- Regression analyses:
  - Baseline specification and multiple extensions are presented, with attention to heterogeneity across industries and transmission channels.
  - Robustness checks and model extensions are included.

### Coverage and appendices
- APPENDIX. METHODOLOGY AND DATA COVERAGE (page 38)
  - Industry-level growth decomposition and productivity comparisons
  - Country groups
  - Industry categorization

*Source: wpiea2022213-print-pdf - INTRODUCTION (IMF Working Paper chapter contents).*

### Introduction

### Introduction

### Overview
- The regulatory environment has been identified as a key factor in keeping Korea’s economy-wide productivity at a relatively low level for a country with its per capita income (McKinsey, 1998; OECD 2014, 2016; Jones, 2009; Jain-Chandra and Zhang, 2014; Jones and Lee, 2016).
- Korea’s trend growth rate has been in a long decline dating from well before the Covid-19 pandemic (Cho and Kwon, 2017; Kwon and others, 2019; Zoli and others, 2018).
- The paper constructs a cross-country, industry-level dataset of productivity levels across advanced economies and compares Korea’s labor productivity (LP) and total factor productivity (TFP) with those of other advanced economies.

### Productivity trends and proximate drivers
- Korea historically relied heavily on accumulation of factor inputs—capital and labor—with a high rate of capital formation (Collins and Bosworth, 1996; Lee and Hong, 2012).
- Rising labor inputs during the high-growth period were driven by rapid working-age population growth, high hours per worker, and large increases in worker educational attainment (Barro and Lee, 2013).
- Economy-wide productivity levels increased substantially through the 1990’s but have since plateaued at around two-thirds of the levels of the most productive economies (Lee, 2016; Swiston, 2021).
- Structural constraints:
  - Korea’s already-high dependence on capital accumulation and labor inputs, together with the demographic transition to a more elderly population, imply the scope for input-driven growth will be smaller going forward (Kwon, 2017; Swiston, 2021).
  - The economy is likely to become more service-oriented as income rises; the Covid-19 pandemic may have accelerated structural shifts such as digitalization increasing the proportion of services.

### Cross-country and industry-level productivity comparisons
- Aggregate convergence:
  - In the mid-1990’s, Korea’s overall LP was 40 percent below the advanced economy median, but by 2017 the gap had been closed to less than 15 percent.
  - Aggregate TFP was just over half the advanced economy median in 1995, and increased to about two-thirds by 2017.
- Manufacturing versus market services:
  - Manufacturing has been the main driver of Korea’s overall productivity gains; since the 1990’s Korea’s LP in manufacturing moved from the bottom third to the top third of the sample.
  - TFP increases in manufacturing have been more modest, indicating capital accumulation played a large role in manufacturing output gains.
  - Productivity growth has been slower in market services, with no noticeable convergence toward other advanced economies.
- Detailed industry-level findings (2015-17 benchmarks and distributions):
  - For manufacturing in aggregate:
    - LP and TFP gaps with respect to the frontier are both about 50 percent.
    - LP is 120 percent of the median, while TFP is 80 percent of the median.
    - Industry gaps mostly range from about 30 percent to 80 percent.
    - LP is closest to the frontier for basic and fabricated metals.
    - TFP is closest to the frontier in basic and fabricated metals and in transport equipment.
    - Computer and other IT equipment manufacturing ranks relatively highly in LP and TFP owing to semiconductors; the U.S. has a large advantage explaining the remaining frontier gap.
  - For market services in aggregate:
    - Productivity gaps are about 60 percent with respect to the frontier for both LP and TFP.
    - Market services LP is 30 percent below the median, while TFP is 40 percent below the median.
    - Both LP and TFP are closest to the frontier in professional and business services, where productivity exceeds the median.
    - Gaps of 75 percent or more with respect to the frontier—or about 35 to 50 percent with respect to the median—exist in wholesale and retail trade, transportation and storage, information and communication, and personal services.

### Product market regulations (PMRs): measurement and Korea’s standing
- The OECD PMR Indicators are used as the primary, comparable measure of product market regulation across countries and time; indicators are questionnaire-based and aggregated into subcategories, three broad areas, and an economy-wide PMR score.
- OECD PMR coverage and features:
  - Indicators begin in 1998 and are available at five-year intervals; a structural break exists between the 2013 and 2018 vintages.
  - Indicators are economy-wide (with a few more industry-relevant items), but are highly detailed at the low-level indicator dimension.
  - The OECD PMR Indicators cover three broad areas: State control; Barriers to entrepreneurship; Barriers to trade and investment, each with subcategories and low-level indicators.
- Korea’s PMR trajectory and comparative standing:
  - Korea’s aggregate PMR improved from 2.6 in 1998 (scores range from zero to six).
  - Korea’s regulatory environment remains somewhat more restrictive than the average OECD economy.

### Empirical approach and main econometric findings
- Methodology:
  - Panel regressions use System GMM to estimate the impact of PMRs on productivity, addressing the short time dimension and dynamics inherent in productivity growth typical of OECD PMR data.
  - The empirical strategy controls for each industry’s inherent exposure to the impact of regulations.
- Main results:
  - PMRs are a significant factor in industry-level productivity: a tighter regulatory environment slows industry-level growth of both LP and TFP.
  - Negative effects are found across all main areas of PMRs—state control, barriers to entrepreneurship, and barriers to trade and investment—and across most detailed OECD PMR indicators.
  - Reforms aligning Korea’s regulatory flexibility with the average OECD economy could raise aggregate productivity growth by half to three-quarters of a percentage point per year.
  - PMRs constrain productivity growth more when a country-industry observation is closer to the global technological frontier (evidence of effects through innovation).
  - No significant differences were found in PMR effects across broad industry groups (manufacturing versus services), but PMRs have larger effects in industries where LP tends to be higher (for example high-tech manufacturing and information-intensive services).

### Channels: how PMRs affect productivity and industry composition
- PMRs are associated with slower growth in real value added, transmitted principally through:
  - Lower TFP.
  - Secondarily through reduced labor input, especially in industries near the technological frontier.
  - Little impact on capital accumulation was found.
- Transmission mechanisms:
  - PMRs are associated with higher product prices and higher unit labor costs in affected industries, indicating regulation raises production costs and reduces competitiveness.
  - Cost increases can spill over into other industries through input-output linkages.
- Aggregate implications:
  - Korea’s relatively tight PMRs are likely constraining productivity in services and high-tech manufacturing.
  - Reforms could support reallocation toward dynamic, high-productivity sectors such as information and communications, pharmaceuticals, and biotechnology, and improve efficiency in broader market services.

### Policy priorities identified
- Key priorities for reform in Korea include:
  - Lowering barriers in service and network sectors.
  - Reducing restrictions applying to trade and investment.
  - Evaluating the scope of government involvement in the economy.

### Measurement scope and data notes
- Data sources and scope:
  - The paper begins with data from the Groningen Growth and Development Centre, which measures productivity by industry using industry-level purchasing power parities and accounts for volumes and types of capital and labor.
  - Analysis focuses on manufacturing and “market services” because a greater proportion of activities operate on a commercial basis and because measurement of inputs and outputs is more reliable than in non-market services.
  - Real estate and manufacturing of coke and refined petroleum products are excluded from industry-level analysis due to measurement inconsistencies; construction and utilities have greater measurement challenges.
  - LP data are available for a broader cross-section and longer period; TFP availability is skewed toward relatively wealthier countries and is more uncertain but conceptually superior as it controls for capital per worker.
- Sample and time coverage:
  - The analysis focuses on advanced economies in recent decades—generally since 1995.
  - Lack of cross-country availability of PPP-adjusted data precludes analysis of developments since the onset of the Covid-19 pandemic.

*IMF Working Papers — Productivity and Product Markets in Korea (Introduction).*

### 1.7 in 2018.

### wpiea2022213-print-pdf - 1.7 in 2018.

### OECD PMR indicators and Korea's standing
- OECD Product Market Regulation (PMR) indicators show Korea with an increasingly flexible regulatory regime albeit remaining more restrictive than other advanced economies.
- Comparison of the 2013 and 2018 indicators suggests methodological changes did not greatly affect the overall assessment of Korea’s regulatory environment relative to peers.
- Main areas where Korea is relatively more restrictive:
  - Barriers in services and network sectors.
  - Public ownership and involvement in business operations.
  - Barriers to trade and investment.
- Areas where Korea is near or better than peers:
  - Administrative burden on startups — near the OECD average.
  - Regulatory complexity/simplification — 2013 vintage shows Korea near the OECD average, while the 2018 vintage ranks Korea among the top 5 in the OECD.
- Indicator scales preserved as in source:
  - OECD economy-wide PMR indicator: zero to six scale, with higher scores more restrictive.
  - IMF product market regulation economy-wide indicator: zero to 1 scale, with higher scores more restrictive.
- Notes from source:
  - "Note there could be uncertainty surrounding the point estimates, though this improvement seems to be a general trend."
  - "Also, given the regression sample, the quantitative analysis uses only the 1998-2013 indicators."
- Figures referenced (as described):
  - Figure 5: OECD Product Market Regulation: Economy-Wide Indicator (zero to six scale).
  - Figure 6: OECD Product Market Regulation Indicators by Area (zero to six scale).
  - Figure 7: IMF Product Market Regulation: Economy-Wide Indicator (zero to 1 scale).

### Cross-dataset corroboration
- The OECD PMR conclusions for Korea are corroborated by other measures despite coverage and focus differences, including:
  - IMF Structural Reform Database (product markets component).
  - Fraser Institute’s Economic Freedom of the World report (relevant components).
  - World Economic Forum’s Global Competitiveness Index.
  - World Bank’s Worldwide Governance Indicators.
  - World Bank’s Doing Business Indicators (noted as discontinued in source).

### Industry-level PMR measurement and empirical strategy
- Challenge: PMR indicators are measured at the country level, not industry level, complicating empirical estimation with industry-level data.
- Theoretical premise: Impact of PMRs varies across industries according to the degree to which regulations constrain each industry’s activity — the "industry exposure".
- Approaches to generate industry-specific indicators:
  - "Downstream impact" / "knock-on effects": Adjust raw PMRs for industry exposure using indicators that capture how regulated key industries (energy, transport, communication, retail distribution, business services) are, applied to the proportion of those industries' output used as inputs in other industries (see Conway and Nicoletti, 2006, and Egert and Wanner, 2016).
  - Industry exposure to structural factors: Use metrics such as firm turnover rates, layoff rates, and degree of dependence on external finance to capture industry-inherent characteristics mediating regulatory effects (Andrews and Cingano, 2014; Bassanini and others, 2009).

### Industry-specific characteristics used in interactions with PMR indicators
- To detect regulatory effects, PMR indicators are interacted with multiple industry-specific characteristics. Specifically, three measures are used (source provides detail on the first two):
  1. Layoff exposure:
     - Interpreted as a proxy for volatility inherent in an industry, reflecting reallocation of labor in response to shocks.
     - Calculated using layoff rates for United States, as the benchmark economy with generally the most flexible labor market reallocation policies among the group under consideration (Andrews and Cingano, 2014; Bassanini and others, 2009).
     - Thus, layoff exposure by industry is not dependent upon the degree of labor market regulation in any economy outside the United States.
  2. Capital-output ratio:
     - Taken as a measure of the capital intensity typical within a given industry.
     - Given the irreversible nature of investment in fixed capital, this could detect the effects of the regulatory environment on investment.
     - Calculated as the industry average across all available countries and time periods.
  3. (Measure name listed in source but detailed description not included in the provided excerpt.)

*Source: wpiea2022213-print-pdf - 1.7 in 2018.*

### 3. Skill requirement: This is used as a measure of the skilled labor requirements typical for a given

### wpiea2022213-print-pdf - 3. Skill requirement: This is used as a measure of the skilled labor requirements typical for a given

### Skill requirement measure and exposure construction
- Skill requirement: calculated as the industry average across all available countries and time periods to capture industry inherent characteristics, not idiosyncrasies from labor market regulation.
- For empirical analysis each measure is scaled between zero (industry with the lowest exposure) and one (industry with the highest exposure).
- A broad exposure measure is constructed as the simple average of three measures (capital-output ratio, labor skill levels, layoff exposure).
- Combining time-invariant, industry-specific exposure measures with time-varying PMR indicators yields PMR indicators varying by country, industry, and time.

### Linking regulations and productivity: mechanisms highlighted
- Regulation affects incentives to innovate: nonlinear effects on innovation; greater competition spurs innovation near the technological frontier but not in laggard industries.
- Regulation slows knowledge diffusion, limiting discovery of new technologies.
- Regulation affects allocation/reallocation of resources across industries, firms, and factors by altering relative returns/costs of capital and labor.
- Regulation shapes mix of establishments within an industry (entry costs, firing costs), impeding exit/entry and labor reallocation.
- Empirical literature generally finds more restrictive PMRs weaken productivity growth, operating through innovation, resource reallocation, and production costs; effects vary across industries and frontier position.

### Empirical approach and data
- Neo-Schumpeterian framework: productivity drivers differ by industry’s distance to technological frontier (DTF).
- Core estimating equation includes: lagged frontier growth, lagged dependent variable, DTF term, vector of regulatory/structural variables (PMRs), time, industry, and country dummies.
- Policies and structural factors are lagged one period; frontier growth lagged one period; DTF uses second lag as explanatory variable.
- Data: productivity data since the mid-1990s for LP available for 29 countries and for TFP for 19 countries, typically across 26 industries; analysis focuses on industries operating commercially with reliable productivity measurement.
- Temporal aggregation: non-overlapping three-year averages yielding seven time periods; regressions estimated over a maximum of six periods due to lagging OECD PMR indicators.
- Estimation: System GMM with collapsed instruments using only the second or third lag as instruments to limit instrument proliferation.
- Diagnostic tests used: AR(2) for residual second-order autocorrelation in differences and Hansen test for instrument exogeneity; joint significance of PMR terms tested with F-tests.

### Regression results — baseline and magnitude
- Baseline dynamics: lagged LP growth, lagged frontier growth, and lagged DTF are significant determinants of LP growth; similar for TFP with weaker frontier spillovers and slower convergence.
- PMRs show significant effects on productivity only when interacted with position relative to the frontier (productivity gap): PMR linear term often insignificant; PMR × productivity gap frequently significant and negative, implying industries closer to the frontier affected more.
- Industry exposure matters: weighting PMR by any exposure measure (capital-output ratio, labor skill levels, layoff exposure) improves identification of PMR effects; the simple average of the three exposure measures gives the strongest relationships.
- Economic magnitudes:
  - One standard deviation less restrictive PMRs than the sample average is associated with LP growth higher by around 0.7 percent per year at the median productivity gap.
  - For industries with DTF at the 75th percentile, only a small impact of PMRs on productivity; at the 25th percentile effects are twice as strong as at the median.
  - For TFP growth, one standard deviation less restrictive PMRs than the sample average is around 1.2 percent per year at the median; 25th–75th percentile range estimated between 1-1.5 percent.
  - For Korea: aligning overall PMRs to the average OECD level translates to an impact on LP growth of 0.5 percent per year (median DTF) and 0.8 percent per year for TFP.

- Selected regression diagnostics reported (example from Table 2, column 7 — simple average exposure):
  - Observations: 2,158 (varied across specifications)
  - Country-industry groups: 381
  - Instruments: 39
  - F-test p-value for PMR terms (joint significance): values reported across specifications including 0.00 and 0.01
  - Hansen test p-values and AR(2) test p-values reported (examples: Hansen 0.47; AR(2) 0.76)

### Detailed results on structural reform subcomponents
- State control:
  - Aggregate state control indicator strongly negative for both LP and TFP.
  - Level 3 indicators—public ownership and involvement in business operations—also strongly negative.
  - Level 4: scope of SOEs, direct control over enterprises, and price controls jointly significant on LP growth; scope of SOEs, direct control, and command-and-control regulation jointly significant on TFP growth.
- Barriers to entrepreneurship:
  - Aggregate and level 3 indicators (complexity of regulatory procedures, administrative burden on startups, regulatory protection of incumbents) associated with lower LP growth; effects stronger nearer the frontier.
  - Level 4 indicators: all except barriers in service sectors jointly significant on LP growth; most jointly significant on TFP growth; service sector barriers impact especially large.
- Barriers to trade and investment:
  - Aggregate indicator and level 3 indicators jointly significant for LP growth but not consistently for TFP.
  - Level 4: differential treatment of foreign suppliers jointly significant for both LP and TFP.
- Industry-specific PMR indicators (network industries, retail, professional services):
  - Restrictions in network industries have wide-ranging negative effects on productivity economy-wide.
  - Restrictions in professional services significant barriers to TFP growth.
  - Marginal economy-wide effects detected for restrictions in retail sales.

- Selected level-4 detailed coefficients (from Table 6; significance preserved as reported):
  - Scope of state-owned enterprises: LP direct effect -0.09***; interaction -0.16***; joint significance p-value 0.01.
  - Administrative burdens for corporations: LP direct effect -0.06***; interaction -0.08***.
  - Legal barriers to entry: LP direct effect -0.05***; interaction -0.09***.
  - Restrictions in network industries: LP direct effect -0.03**; interaction -0.05***.

### Differences across industries
- Industry groupings analyzed: (1) manufacturing, (2) market services, (3) network sectors, (4) ICT-using industries, (5) ICT-related industries, (6) high LP industries.
- PMR × industry group dummy interactions not individually significant in most cases, but:
  - Joint significance tests indicate significant PMR effects on LP and TFP for high LP industries and on LP for ICT-related industries.
  - Interpretation: PMRs tend to have larger impacts on high-productivity industries; modest effects detected for ICT-related and high-LP groups.
- Policy implication for Korea: potential productivity improvements from structural reforms in information and communication, and finance and insurance.

### Transmission channels
- Dependent variables analyzed: real value added, hours worked, gross fixed capital formation, value added deflators, unit labor costs (ULC).
- Real value added: significant negative impact of more restrictive PMRs across all areas; sizable share of these effects operates via reductions in TFP growth.
- Factor accumulation:
  - PMRs generally associated with slower growth in hours worked and gross fixed capital formation, though these effects are less consistent and less frequently significant than productivity/value-added effects.
- Prices and costs:
  - More restrictive PMRs associated with somewhat higher value added deflators and substantially higher ULC.
  - Effects on ULC highly significant in most cases and stronger for regulations in state control and barriers to entrepreneurship areas.
  - Stronger effects on ULC than on final prices suggest net impact on profits is negative—costs rise more than any price increase from reduced competition.
- Overall transmission summary: tighter PMRs depress productivity by elevating business costs and product prices, moderately lowering hours worked, with minor impact on capital accumulation; higher costs can propagate through inter-industry input linkages affecting downstream productivity.

### Model extensions and robustness checks
- Additional structural drivers tested (not significant and did not alter PMR findings):
  - Labor market regulation (IMF Structural Reform Database; CBR-Labour Regulation Index; OECD employment protection legislation).
  - Governance (Worldwide Governance Indicators).
  - R&D expenditure as share of GDP.
  - Fraser Institute Economic Freedom indicators.
  - IMF Structural Reform Database indicators on liberalization of external trade, external finance, and domestic financial system.
- Alternative specifications:
  - Using first lag of DTF instead of second: coefficients similar but many specifications failed Hansen test.
  - Using contemporaneous frontier growth: similar to baseline with first lag.
  - Excluding lagged dependent variable: stronger significance for other terms including PMR, but exogeneity of instruments rejected for TFP and borderline for LP.
- Industry coverage robustness: main results hold when expanding sample to include public services, utilities and construction, agriculture and mining, and all industries.
- Choice of dummies: results hold when dropping industry dummies, replacing with country dummies, or including both country and industry dummies.

*Source: IMF staff calculations, from the IMF Working Paper "Productivity and Product Markets in Korea" (excerpt provided).*

### Conclusions

### Conclusions

### Main empirical findings on PMRs and productivity
- Tighter PMRs are associated with slower growth of LP and TFP at the industry level.
- Significant negative relationships are present across all the main areas of PMRs—state control, barriers to entrepreneurship, and barriers to trade and investment—as well as most of the detailed indicators.
- Barriers to competition in network sectors have negative effects on economy-wide productivity.
- The effects of PMRs tend to be larger for country-industry observations that are near the global productivity frontier.
- The effects do not vary substantially across broad industry groups such as manufacturing or market services.
- There is evidence that the highest-productivity industries globally (high-tech manufacturing and information-intensive services) are more affected by PMRs than other industries are, suggesting restrictive PMRs hinder innovation as well as imitation.
- The results described below are available from the authors upon request.
- Including country-by-industry dummies would result in an excessive number of instruments.

### Channels and transmission mechanisms
- Tighter PMRs slow real value added growth of affected industries.
- Negative effects occur principally through TFP and also through hours worked, with a minor impact on capital accumulation.
- Product prices and unit labor costs are higher when PMRs are more restrictive, indicating regulation raises production costs and reduces competitiveness.
- Competitiveness losses spill over into other industries through input-output linkages.

### Korea’s PMR stance and potential gains from reform
- Korea’s PMRs have become less restrictive since the 1990s, mirroring trends in advanced economy peers.
- PMRs in Korea remain on the restrictive side, especially:
  - relatively high barriers remaining in service and network sectors;
  - some binding restrictions still applying to trade and investment;
  - a relatively broad scope of government involvement in the economy.
- Reforms to bring the broad PMR environment into line with the average OECD economy could raise aggregate productivity growth by half to three quarters of a percentage point per year.
- Given large gaps in Korea’s productivity relative to other advanced economies, especially in market-oriented service industries, liberalization of these PMRs could help to rejuvenate aggregate growth.
- Regulatory reform is especially critical because Korea already relies heavily on relatively capital-intensive industries and is experiencing a demographic slowdown that constrains the growth of labor inputs.

*Source: IMF Working Paper — Conclusions (wpiea2022213-print-pdf).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022213-print-pdf.pdf_