## 1korea2019002 ## Source details **Canonical URL:** [1korea2019002](https://www.imf.org/-/media/files/publications/cr/2019/1korea2019002.pdf) ## Other formats - [Markdown version](/-/media/files/publications/cr/2019/1korea2019002.pdf.md) - [Structured JSON version](/-/media/files/publications/cr/2019/1korea2019002.pdf.json) --- ### Monetary policy and financial cycles: coordination challenges - Financial sector is inherently procyclical and amplifies the real cycle through price and quantity channels; bank lending becomes more leveraged during credit booms. - Key coordination issues for policymakers: - Real and financial cycles operate at different frequencies. - Supply-side developments have constrained monetary policy in an inflation targeting regime. - Monetary policy is a blunt instrument for managing asset price bubbles. - Real vs financial cycles in Korea: - Financial cycles display greater persistence and larger amplitude than traditional business cycles. - Weak relationship between the credit gap and inflation (1990 to 2018); price stability does not guarantee avoidance of financial excess. - Supply-side and neutral rate dynamics: - Inflation targeting since 1999, combined with positive supply-side developments, has depressed inflation and monetary policy rates, contributing to a decline in the neutral real interest rate. - The neutral rate is model dependent and not directly observable. - Monetary policy and asset-price bubbles: - Theoretical decomposition of stock price into fundamental and bubble components; tightening should reduce the bubble component under conventional assumptions. - Empirical results (1990Q1 to 2018Q2) show the bubble component does not fall in response to rising policy interest rates; effect is persistent and statistically significant, consistent with rational bubbles. - Monetary policy shocks and financial quantities (inflation-targeting period, 1999m1 – 2018m8): - Identification approaches: Christiano, Eichenbaum and Evans (1999) and Rubio-Ramirez, Waggoner and Zha (2010). - A surprise 100 basis point non-systemic monetary tightening: - Leads to a decline in GDP and the price level. - Causes expansion in commercial and shadow bank assets; commercial bank and NBFC assets would have grown between 3-to-6 percent faster over the last 18 months in a counterfactual with policy rates up to 150 basis points higher (2017m1 – 2018m7). - Simulation of a systemic tightening (series of surprises leaving policy rates up to 150 basis points higher over 2017m1 – 2018m7) would curb output growth but accelerate expansion of commercial and shadow bank balance sheets. - Conclusion: - Monetary policy alone has been less effective in taming financial and credit cycles in the inflation-targeting period; dedicated macroprudential tools are required. ### Evolution of macroprudential policies in Korea - Twofold evolution: - Expansion of the macroprudential toolkit. - More aggressive use of macroprudential instruments over the financial cycle. - Shift in instrument mix: - Move away from monetary macroprudential tools (e.g., reserve requirements) toward borrower-based prudential instruments. - Drivers: reduced importance of reserve requirements after adoption of interest rate policy and inflation targeting; recognition that financial cycles (housing credit and house prices) have become longer, larger, and less synchronized with real/inflation cycles; shift toward explicit macroprudential objectives post-Asian crisis. - Usage patterns: - Korea has used prudential policies more actively since the Asian crisis. - An aggregate macroprudential policy index shows changes generally moved in phase with credit growth: higher credit growth often associated with tighter prudential policies and vice versa. - Probability of tightening macroprudential policies in response to a widening credit gap or house price growth is significantly higher in the post-Asian crisis period. - Macroprudential policy response is non-linear, particularly for the housing market. - Targeting: - Macroprudential policy changes have targeted Monetary Tools, Credit and Banking Sector, and Real-Estate sectors (period comparison: 1988-2000 vs 2000-2017 shows higher counts in Credit and Banking Sector and Real-Estate). ### Effectiveness of prudential policies and policy implications - Empirical conclusions: - Macroprudential policies are better tailored to financial risks and can moderate real and financial cycles with fewer unintended consequences on broader policy objectives. - Dedicated macroprudential instruments are required to achieve financial stability objectives, given monetary policy limitations. - Policy recommendations: - Design and deploy well-tailored macroprudential policies to moderate financial excess and build financial resilience in a low-interest-rate environment. - Coordination among policymakers is critical due to differing cycle frequencies and constraints on monetary policy. ### Methodology and empirical setup (pru­dential analysis) - Estimation method: - Local projections based on an adjusted local projection method (Teulings and Zubanov (2013) based on Jordà (2005)), using a robust regression that controls for the business cycle, monetary policy with fixed effects linking financial conditions and various macroprudential policy tools (equations (5) and (6) referenced). - Frequency and sample: - Model estimated at a monthly frequency from 2002 to 2016 to isolate prudential policy changes from other macroeconomic effects. - Controls: - Economic activity, VIX, monetary policy, interest rates; monthly economic conditions captured via a mixed frequency coincident index constructed using quarterly GDP and monthly retail sales, industrial production and service sector growth. - Prudential instruments analyzed: - (i) loan-to-value (LTV) - (ii) debt service-to-income (DTSI) - (iii) real estate tax - (iv) rise in risk-weights and provisioning - Notation and referenced equations: - Response equation (5); state equation (6); bank-resilience regression (7); bank lending interest rate rule (8); link to NIM via equation (9). ### Prudential instruments: timing, magnitudes, and bank resilience - Timing and impacts: - Changes in loan-to-value limits and risk-weighting appear, with a lag, to have their largest impact on the credit cycle. - Real estate–specific measures (raising real estate–related taxes or tightening LTV) directly reduce real estate price inflation. - LTV and DSTI enhance the banking system’s resilience to house price and income shocks and dampen procyclicality of credit and asset price growth. - Lagged effects imply macroprudential policies need to be forward looking to preempt financial excesses. - Bank-based prudential measures (index components): capital conversion buffer, reserve requirements, loan loss provisioning, changes in capital requirements including risk-weights. - Raising capital requirements both preempts and builds resilience. - Such policies have tended to be tightened in periods just before an economic slowdown or during one. - Quantitative responses: - A 1-year cumulated response chart presents values on a scale from -4.5 to 0 for Financial Corporation Debt*, Bank Leverage, Non-core funding, Credit gap* (* in percent fo GDP). Bank leverage and non-core funding decline by a similar magnitude following tighter bank-based prudential policies; financial corporation debt and the credit gap also decline. - Net Interest Margin (NIM): - Macroprudential measures (e.g., increased capital or reserve requirements) tend to increase banks’ costs and can raise NIM, affecting bank profitability and monetary transmission. - The link between macroprudential policy and lending rates is formalized in equation (8): banks’ lending rate expressed as a function of the policy interest rate and the interest margin (ߜ௧), with the NIM influenced by regulation (ݖ௧). ### Interaction with monetary policy: complementarity and caveats - Complementarity: - Macroprudential and monetary policies can complement each other; macroprudential policy can give monetary policy more room and help preserve monetary transmission in stress. - The impact of LTV tightening on the credit gap (in percent of GDP) is larger when macroprudential and monetary policy push in the same direction (accumulated responses chart with vertical axis from -2.5 to 1 and months labeled 1 13 25 37 49 61 73 85). - Empirical patterns: - Since 2001, tightening and loosening cycles in macroprudential and monetary policy have tended to move in phase. - Pairwise correlation of the two policy cycles reported as positive and significant at around 0.6. - Table 1 numeric correlations: - Monetary Prudential Tools: 0.59 - Nonmonetary Prudential Tools: 0.49 - Caveats: - Macroprudential measures are less effective if they lean opposite to monetary policy. - The sample contains insufficient episodes where real and financial cycles are misaligned to robustly test cases where policies push in opposite directions. ### Policy implications and recommendations (prudential–monetary coordination) - Empirical conclusions for Korea: - Macroprudential policy has two active contributions: - Preempting aggregate weakness by limiting buildup of risk and reducing occurrence of crises (reducing procyclical feedback between asset prices and credit). - Reducing systemic vulnerability by increasing financial system resilience and building buffers that help maintain credit provision under stress. - Trade-offs and governance: - Macroprudential policy is not free of costs; higher capital and liquidity requirements can enhance stability but raise the price of credit. - Objectives of macroprudential policy need clear definition and a strong accountability framework. - Practical guidance: - Macroprudential measures should be forward looking to preempt financial excesses given lagged effects. - Consider combining macroprudential tightening with complementary monetary easing when appropriate to mitigate adverse effects on real activity, recognizing empirical evidence that complementary policies yield stronger effects on the credit cycle. --- ### Financial conditions in Korea: construction and tail-risk projections - Financial Conditions Index (FCI) construction: - Dynamic factor model with 16 variables, including bank and NBFC credit, bond spread, equity prices, broad and narrow money, BOK Loan Officer Survey measures, leverage measures (household and corporate debt). - Framework purges real-cycle effects by including CPI inflation and a mixed frequency coincident indicator (real GDP quarterly; industrial production, services index, retail sales monthly). - Model controls for monetary-policy-driven changes by including reserve money and assumes time-varying factor loadings following an AR specification. - Forecasting and tail-risk methodology: - Forecasts constructed using quantile projections; decomposition into principal components of price of risk (p), financial leverage (Agg), domestic real conditions (y), and global/foreign variables (f). - Risk findings (as of Q3 2018): - Median projection and high probability outcome: financial conditions in Korea over the next 4 quarters will remain quite neutral; median FCI estimate of around 0. - Extreme event: FCI would be expected to tighten from 0 to -2 in a tail-risk materialization, with significant real-sector repercussions. - Rolling estimates of the 5th quantile since 2010 show tail-risks have not substantially risen; the tail has not become fatter or more skewed. - Sensitivity to external tightening: a standard deviation shock in the VIX would tighten financial conditions in Korea and increase probability of tail-risk outcomes considerably; heightened by high household debt linked to floating rates. - Link to authorities: - Finding that financial risks remained contained in 2018 is consistent with the Bank of Korea’s financial stability index (December 2018 Financial Stability Report). - Policy implications: - Maintain heightened vigilance during benign macro conditions that may facilitate accumulation of financial vulnerabilities. - Changes in the domestic price of external risk are potent signals for swift deployment of monetary easing and macroprudential policy actions. --- ### Labor market slack and the output gap: measurement and implications - Motivation and approach: - Output gap estimates are key for macroeconomic policy; labor is the most important input, so labor-market capacity estimates inform output-gap calculations. - Formulate a multivariate filter (MVF) based on Blagrave et al. (2015) linking labor market slack and the output gap through a production function. - Broaden labor utilization measures by augmenting regular unemployment along the extensive margin. - Alternative slack measures: - Unemployment augmented with discouraged workers (percent of labor force). - Unemployment augmented with persons inactive for unspecified reasons. - Levels and relationship with inflation: - The two alternative measures hover around 5.5 and 10.5 percent, respectively, and show more business-cycle volatility than regular unemployment. - The augmented measure with discouraged workers yields a negative Phillips curve for inflation since 2001, suggesting better capture of labor under-utilization. - Output gap estimates (2007-2023) under three labor-slack inputs: - Using regular unemployment: - Output gap around +1 percent in 2007/08; turned negative in 2009. - Returned to up to slightly below 1 percent in 2010-11; negative in 2012; widened to around -0.8 percent in 2012-14 before narrowing. - From 2018 onwards output gap closely correlated with labor market gap as TFP projected to evolve according to potential. - Using unemployment + discouraged workers: - Produces a more volatile labor market and output gap; more consistency between labor market gap and output gap up to 2012. - Both gaps deteriorate more in 2009; 2012-17 unemployment gap remains positive while output gap is negative. - From 2017 onwards the two gaps projected to move in tandem. - Using unemployment + inactive unspecified: - Also yields volatile labor market gap; output gap broadly similar to that with regular unemployment. - Indicates largest labor market gap in 2017; consistent labor market and output gap from 2017 onwards. - Robust findings: - Up to 2017 the labor market and output gaps display conflicting signals at times due to TFP deviations from trend. - From 2017 onwards both output and labor market gap projected to be negative across all models (TFP projected on trend; labor market gap drives output gap). - Broader slack measures increase measured volatility and improve the inflation-unemployment relationship. - Data and estimation: - Annual data from 2001 for GDP, inflation, employment, TFP, unemployment and alternative slack measures; historical data up to 2017; forecasts 2018-23; Bayesian estimation. --- ### SMEs, ALMPs, public employment, and labor policies - SMEs, R&D incentives, and youth employment: - Problem: Youth reluctance to work for SMEs due to fewer training opportunities and lower wages; subsidies may not effectively boost SME employment. - Recommendation: Redirect subsidies toward R&D activities to make SMEs more dynamic; target support to young and innovative firms. - Evidence: R&D tax incentives are particularly effective for small firms; re-orient some fiscal support toward R&D. - SME program scale (selected statistics): - As of 2017, 288 central government programs and over 1000 local government programs to support SMEs. - Central government spending on such programs estimated at around 0.8 percent of GDP in 2017. - Credit guarantees to SMEs were 3.8 percent of GDP in 2016 (second largest in the OECD). - Reduction in combined corporate and personal rates on capital gains due to preferential SME corporate income tax treatment estimated at 13 percentage points in 2014. - Active Labor Market Policies (ALMPs): - Korea’s overall ALMP spending: 0.4 percent of GDP (close to the OECD average). - Composition concerns: - Direct job creation represents more than half of government expenditure on employment support in 2016 (OECD average about 13 percent). - About 21 percent of direct job creation spending was targeted at workers above retirement age of 65 years in 2018. - PES and training: - Korea’s PES spending: 0.04 percent of GDP (about 10 percent of employment-supporting ALMPs); OECD averages: PES 0.1 percent of GDP and nearly 28 percent of total spending. - Korea’s training spending lower than OECD average. - Employment incentives: - Korea devotes more than 13 percent of ALMPs expenditure to employment incentives, about 0.05 percent of GDP in 2016. - Alternative estimate: Park (2016) reports Korea’s employment incentives spending in 2016 about 0.2 percent of GDP (20 incentives under four ministries; budget 2.8 trillion won). - Policy recommendations: - Reconsider ALMP composition; increase spending on public employment services and training; review effectiveness and targeting of employment incentives; consider targeted transfers for over-65 income support instead of employment programs. - Public employment: - Share of employment in general government relative to total employment: less than 8 percent. - Government plans to expand public sector jobs by 810,000 by 2022 (MOEF, 2018); 2019 budget intends to increase social services jobs by 60,000 to 94,000 and civil service jobs by 21,000. - Recommendations: Link public job creation to services not adequately provided by private sector; assess impacts on productivity and private-sector crowding out; compare public-private wages in setting hiring conditions. - Empirical determinants of employment (cross-country panel, 1985-2014): - Key regression coefficients (columns: All, Women, Men): - ALMP(‐1): 0.05 0.08 0.05; P‐value: 0.00 0.00 0.00 - Labor tax wedge: ‐0.18 ‐0.16 ‐0.14; P‐value: 0.00 0.00 0.00 - Childcare spending: 3.38 6.11 1.68; P‐value: 0.00 0.00 0.00 - Job‐protected maternity leave: 0.02 0.05 0.01; P‐value: 0.00 0.00 0.33 - EPL: 0.53 ‐2.17 1.86; P‐value: 0.42 0.01 0.76 - Output gap (‐1): 0.20 0.19 0.22; P‐value: 0.00 0.02 0.00 - Trade openness (‐1): 0.04 0.10 0.05; P‐value: 0.00 0.00 0.00 - Relative service employment: ‐0.04 ‐0.03 ‐0.05; P‐value: 0.00 0.00 0.00 - Union Density: 0.00 ‐0.04 ‐0.04; P‐value: 0.94 0.30 0.27 - Exposure to technology: 0.62 ‐0.04 0.24; P‐value: 0.00 0.90 0.10 - Tertiary education: 0.10 0.15 0.11; P‐value: 0.00 0.00 0.00 - Model fit and sample: - R‐squared: 0.93 0.93 0.86 - Estimation period: 1985‐2014 - Number of observations: 558 558 558 - Policy implications: - Family-friendly policies (childcare spending, maternity leave) support female employment. - Korea’s public spending on early childhood education and care was almost 0.9 percent of GDP in 2013; increasing this can support female employment. - Relaxing employment protection for regular workers could support female employment, accompanied by increased ALMP spending. ### Working hours reform: implementation and international context - Korea reducing maximum weekly working hours from 68 to 52 (amendment on March 20, 2018). - Key features: - Exceeding limit allowed if average weekly hours do not exceed 52 over reference period (initially 3 months; likely extended to 6 months). - Phased implementation over three years, starting with large companies and public sector; management positions exempted. - For employees under 18 the new ceiling is 35 hours per week. - Five industries (including transport and health care) exempt from limits; twenty-six were exempt under previous rules. - International comparisons: - EU Directive sets 48 hours maximum average; modalities vary across members (individual opt-outs common). - Japan (2018 law): cap 100 hours a month overtime and 720 hours per year; white-collar workers with annual incomes > 10.75 million yen exempted; effective April 2019 for large firms, April 2020 for SMEs. - Recommendation: - Monitor implementation to ensure widespread decrease in hours; evaluate impact on firms and productivity. --- ### Industrial structure, linkages, and shock propagation - Sectoral shares and concentration: - Manufacturing share of GDP: "12.7 percent in 1970s" to "31.6 percent in 2010s". - Services share of GDP: stable around "60 percent". - Share of the three largest manufacturing industries: "41.9 percent in 1990s", "57.6 percent in 2010s". - Services top three industries share: "39.5 percent in 2010s". - Interconnectedness: - Overall domestic vertical linkages increased to "47.1 percent in 2009 from 44.3 percent in 2000", then remained stable. - Upstream/downstream linkages: above "50 percent" in manufacturing; below "40 percent" in services. - Major manufacturing industries and trade linkages: - Major industries by 2014 GDP share: electronics, transportation, chemicals. - Upstream linkages highest in chemicals (almost twice electronics and transportation); downstream linkages around "60 percent" across the three industries. - Export/import share in total output increased by about "5 percentage point each" after the GFC. - Electronics and transportation show about twice higher export linkages than chemicals; import linkages highest in electronics. - Export and import exposures to top three trading partners account for almost "50 percent" of total trade linkages in major manufacturing industries (except export linkages in transportation). - Macroeconomic development: - Korea’s economic growth: "4.3 percent in 2000s" and "2.8 percent in 2010s". - Electronics contributes "almost half of the manufacturing growth since 2000s". - Largest three industries’ growth contribution: "26.5 percent in 1980s", "almost 70 percent since 2000s". - Growth volatility declined after the 1990s; manufacturing volatility higher than services in 2000s. - Shock transmission and empirical results (WIOD 2000-2014; network approach): - Supply-side domestic industry shocks have larger downstream effects than upstream effects. - External country shocks propagate mainly through own direct export linkages. - Quantitative impacts: - Domestic industry shock (one standard deviation) in chemicals: 0.52 percentage point productivity impact (largest). - China’s growth shock: 0.26 percentage point productivity impact for total industry (largest among partner countries). - Overall impact of one standard deviation shocks: - Total industry: 1.4 percentage points. - Manufacturing: 3.2 percentage points. - Robustness: - Effects larger for manufacturing due to stronger vertical and trade linkages. - For China and Japan, downstream and direct export linkages are significant; for the U.S., domestic downstream effects significant, direct export effects negative but not significant for total industry. - Structural vulnerability and policy implications: - High concentration and interconnectedness around a few major, relatively volatile industries (dominance of a few large firms) increase susceptibility to domestic or external shocks. - Recommended actions: - Diversify industrial structure to mitigate macroeconomic instability. - Promote fair competition between large corporations and smaller firms. - Foster innovation especially in SMEs. - Reduce regulatory burden on firms, especially in services. --- ### Conclusion and consolidated policy implications - Financial stability and macroprudential policy: - Monetary policy alone has limited ability to tame financial and credit cycles under inflation targeting; well-designed macroprudential instruments are essential. - Macroprudential tools should be forward looking, targeted (LTV, DSTI, taxes, risk-weights), and coordinated with monetary policy where possible. - Financial conditions monitoring: - Median FCI around 0 over near term (as of Q3 2018); tail events can tighten FCI from 0 to -2 with significant real effects. - External risk pricing (e.g., VIX) is a key signal for financial stability actions. - Labor market and employment policies: - Broader measures of labor under-utilization (including discouraged workers) improve output-gap estimation and the Phillips curve. - Reorient SME support toward R&D and young/innovative firms; periodically review and consolidate SME programs. - Rebalance ALMPs toward PES, training, and better-targeted employment incentives; reassess direct job creation and income support for elderly. - Implement working-hours reform widely and monitor effects on hours worked and productivity. - Industrial policy and resilience: - Address concentration risks in manufacturing (electronics-dominated growth); diversify exports and production linkages. - Policies to promote SME innovation, competition, and regulatory easing in services can mitigate vulnerability to shocks. *Source: IMF staff compilation from the Selected Issues chapter on Evolution of Macroprudential Policies in Korea and associated sections (content unit: 1korea2019002).* ### References _________________________________________________________________________________ 21 ### 1korea2019002 - References _________________________________________________________________________________ 21 ### Monetary policy and financial cycles: coordination challenges - Financial sector is inherently procyclical and amplifies the real cycle through price and quantity channels; bank lending becomes more leveraged during credit booms. - Key coordination issues for policymakers: - Real and financial cycles operate at different frequencies. - Supply-side developments have constrained monetary policy in an inflation targeting regime. - Monetary policy is a blunt instrument for managing asset price bubbles. - Real vs financial cycles in Korea: - Financial cycles display greater persistence and larger amplitude than traditional business cycles. - Weak relationship between the credit gap and inflation (1990 to 2018); price stability does not guarantee avoidance of financial excess. - Supply-side and neutral rate dynamics: - Inflation targeting since 1999, combined with positive supply-side developments, has depressed inflation and monetary policy rates, contributing to a decline in the neutral real interest rate. - The neutral rate is model dependent and not directly observable. - Monetary policy and asset-price bubbles: - Theoretical decomposition of stock price into fundamental and bubble components; tightening should reduce the bubble component under conventional assumptions. - Empirical results (1990Q1 to 2018Q2) show the bubble component does not fall in response to rising policy interest rates; effect is persistent and statistically significant, consistent with rational bubbles. - Monetary policy shocks and financial quantities: - Two identification approaches used for the inflation-targeting period (1999m1 – 2018m8): Christiano, Eichenbaum and Evans (1999) and Rubio-Ramirez, Waggoner and Zha (2010). - A surprise 100 basis point non-systemic monetary tightening: - Leads to a decline in GDP and the price level. - Causes expansion in commercial and shadow bank assets; commercial bank and NBFC assets would have grown between 3-to-6 percent faster over the last 18 months in a counterfactual with policy rates up to 150 basis points higher (2017m1 – 2018m7). - Simulation of a systemic tightening (series of surprises leaving policy rates up to 150 basis points higher over 2017m1 – 2018m7) would curb output growth but accelerate expansion of commercial and shadow bank balance sheets. - Conclusion on monetary policy role: - Monetary policy alone has been less effective in taming financial and credit cycles in the inflation-targeting period; dedicated macroprudential tools are required. ### Evolution of macroprudential policies in Korea - Twofold evolution: - Expansion of the macroprudential toolkit. - More aggressive use of macroprudential instruments over the financial cycle. - Shift in instrument mix: - Move away from monetary macroprudential tools (e.g., reserve requirements) toward borrower-based prudential instruments. - Drivers: reduced importance of reserve requirements after adoption of interest rate policy and inflation targeting; recognition that financial cycles (housing credit and house prices) have become longer, larger, and less synchronized with real/inflation cycles; shift toward explicit macroprudential objectives post-Asian crisis. - Usage patterns: - Korea has used prudential policies more actively since the Asian crisis. - An aggregate macroprudential policy index shows changes generally moved in phase with credit growth: higher credit growth often associated with tighter prudential policies and vice versa. - Probability of tightening macroprudential policies in response to a widening credit gap or house price growth is significantly higher in the post-Asian crisis period. - Macroprudential policy response is non-linear, particularly for the housing market. - Targeting: - Macroprudential policy changes have targeted Monetary Tools, Credit and Banking Sector, and Real-Estate sectors (period comparison: 1988-2000 vs 2000-2017 shows higher counts in Credit and Banking Sector and Real-Estate). ### Effectiveness of prudential policies and policy implications - Evidence indicates: - Macroprudential policies are better tailored to financial risks and can moderate real and financial cycles with fewer unintended consequences on broader policy objectives. - Empirical findings support the need for dedicated macroprudential instruments to achieve financial stability objectives, given monetary policy limitations. - Policy implications emphasized: - Design and deploy well-tailored macroprudential policies to moderate financial excess and build financial resilience in a low-interest-rate environment. - Coordination among policymakers is critical due to differing cycle frequencies and constraints on monetary policy. ### Empirical methods and key samples used - Asset-price bubble decomposition and response analysis: KOPIX dividends and KOPIX stock index, short-term interest rate, real GDP, GDP deflator; sample 1990Q1 to 2018Q2. - Monetary policy shock identification for inflation-targeting period: 1999m1 – 2018m8, using Christiano, Eichenbaum and Evans (1999) and Rubio-Ramirez, Waggoner and Zha (2010) approaches. - Counterfactual systemic tightening simulation: series of monetary policy surprises keeping policy rates up to 150 basis points higher over 2017m1 – 2018m7. - Shock experiment: a surprise 100 basis point non-systemic tightening. ### Figures and tables referenced (by label) - Figures: "Impact of Contractionary Monetary Policy Shock on Financial Intermediaries", "Simulation of a Leaning Against the Wind Monetary Policy", "Accumulated Impact of a Tightening LTVs on Credit and Asset Prices in Korea", "Accumulated Response of Credit and House Price Cycle to a Tightening in DSTI Ratio", "Financial Conditions Index", "Financial Conditions-at-Risk", "Financial Conditions Tail Risk (5%_)", "Financial Conditions-at-Risk to VIX Shock", "Financial Conditions Index", "Korea Credit Gap and Inflation", "Korea Macro and Financial Cycles", "Korea Neutral Interest Rate", "Korea: Response of Bubble Component (Qb) to a Tightening in Monetary Policy", "Figure 1. Impact of Contractionary Monetary Policy Shock on Financial Intermediaries", "Figure 2. Simulation of a 'Leaning Against the Wind Monetary Policy'", "Korea Macroprudential Policy Cycle and Credit Growth", "Korea Macroprudential Policies and Targeted Sectors", "Probability of a Tightening in Macroprudential Policies to Credit and Housing Cycles". - Table: "Correlation of Monetary and Macroprudential Changes", and macroprudential policy index and sectoral counts. *Source: IMF staff compilation from the Selected Issues chapter on Evolution of Macroprudential Policies in Korea (content unit: 1korea2019002).* ### 21. The effect of macroprudential policies on financial indicators is estimated using local ### 1korea2019002 - 21. The effect of macroprudential policies on financial indicators is estimated using local ### Methodology and empirical setup - Estimation method: local projections based on an adjusted local projection method (Teulings and Zubanov (2013) based on Jordà (2005)), using a robust regression that controls for the business cycle, monetary policy with fixed effects linking financial conditions and various macroprudential policy tools (equations (5) and (6) referenced). - Frequency and sample: model estimated at a monthly frequency from 2002 to 2016 to isolate prudential policy changes from other macroeconomic effects. - Controls: economic activity, VIX, monetary policy, interest rates; monthly economic conditions captured via a mixed frequency coincident index constructed using quarterly GDP and monthly retail sales, industrial production and service sector growth. - Notation and equations referenced: response equation (5); state equation (6); bank-resilience regression (7); bank lending interest rate rule (8); link to NIM via equation (9). ### Prudential instruments analyzed - The focus is on 4 prudential instruments most used in Korea: - (i) loan-to-value (LTV) - (ii) debt service-to-income (DTSI) - (iii) real estate tax - (iv) rise in risk-weights and provisioning ### Key empirical findings on credit, house prices, and resilience - Timing and impacts: - Changes in loan-to-value limits and risk-weighting appear, with a lag, to have their largest impact on the credit cycle. - Real estate–specific measures (raising real estate–related taxes or tightening LTV) directly reduce real estate price inflation. - LTV and DSTI enhance the banking system’s resilience to house price and income shocks and dampen procyclicality of credit and asset price growth. - The lagged effect of some prudential measures implies macroprudential policies need to be forward looking to preempt financial excesses. - Bank-based prudential measures (index components): capital conversion buffer, reserve requirements, loan loss provisioning, changes in capital requirements including risk-weights. - Raising capital requirements both preempts and builds resilience. - Such policies have tended to be tightened in periods just before an economic slowdown or during one. - Quantitative responses (reported measures and charts): - A 1-year cumulated response chart presents values on a scale from -4.5 to 0 for Financial Corporation Debt*, Bank Leverage, Non-core funding, Credit gap* (chart axis shows -4.5, -4, -3.5, -3, -2.5, -2, -1.5, -1, -0.5, 0). Note: * in percent fo GDP. - Bank leverage and non-core funding decline by a similar magnitude following tighter bank-based prudential policies; financial corporation debt and the credit gap also decline. - Net Interest Margin (NIM): - Macroprudential measures (e.g., increased capital or reserve requirements) tend to increase banks’ costs and can raise NIM, affecting bank profitability and monetary transmission. - The link between macroprudential policy and lending rates is formalized in equation (8): banks’ lending rate expressed as a function of the policy interest rate and the interest margin (ߜ௧), with the NIM influenced by regulation (ݖ௧). ### Interaction with monetary policy - Complementarity and caveats: - Macroprudential and monetary policies can complement each other; macroprudential policy can give monetary policy more room and help preserve monetary transmission in stress. - However, macroprudential measures are less effective if they lean opposite to monetary policy. - Empirical patterns in Korea: - Since 2001, tightening and loosening cycles in macroprudential and monetary policy have tended to move in phase. - Pairwise correlation of the two policy cycles reported as positive and significant at around 0.6. - Table 1 numeric correlations: - Monetary Prudential Tools: 0.59 - Nonmonetary Prudential Tools: 0.49 - Synergy effects: - The impact of LTV tightening on the credit gap (in percent of GDP) is larger when macroprudential and monetary policy push in the same direction (chart shows accumulated responses with and without complimentary monetary policy; vertical axis scale from -2.5 to 1 and months after policy tightening labeled 1 13 25 37 49 61 73 85). - Caution: sample contains insufficient episodes where real and financial cycles are misaligned to robustly test cases where policies push in opposite directions. ### Policy implications and recommendations - Empirical conclusions for Korea (summary points): - Macroprudential policy has two active contributions: - Preempting aggregate weakness by limiting buildup of risk and reducing occurrence of crises (reducing procyclical feedback between asset prices and credit). - Reducing systemic vulnerability by increasing financial system resilience and building buffers that help maintain credit provision under stress. - Trade-offs and governance: - Macroprudential policy is not free of costs; higher capital and liquidity requirements can enhance stability but raise the price of credit. - Objectives of macroprudential policy need to be defined clearly and supported by a strong accountability framework to contribute to financial stability and social welfare. - Practical guidance: - Macroprudential measures should be forward looking to preempt financial excesses given lagged effects. - Consider combining macroprudential tightening with complementary monetary easing when appropriate to mitigate adverse effects on real activity, recognizing empirical evidence that complementary policies yield stronger effects on the credit cycle. *Source: https://www.imf.org/-/media/files/publications/cr/2019/1korea2019002.pdf* ### Introduction ### 1korea2019002 - Introduction ### Key purpose and scope - External financial risks remain due to vulnerabilities, including rising corporate debt, house price misalignments and sovereign-banking sector links in some EU countries; domestically rising financial risks arise from higher household and corporate leverage. - The chapter examines current domestic and global macro and financial conditions and links them to a distribution of possible financial condition outcomes in Korea. - The approach allows assessment of whether slower domestic growth/higher household leverage and/or a tightening in global financial conditions is net financial‑critical and puts financial stability at risk in Korea. ### Construction of Financial Conditions in Korea - A dynamic factor model is estimated to capture aggregate financial conditions. - The model contains 16 variables, including: - bank and NBFC credit, - bond spread, - equity prices, - broad and narrow money, - credit developments based on BOK Loan Officer Survey, - measures of leverage including corporate bond issuance and household and corporate debt. - The framework purges changes in the real cycle from financial conditions by including: - CPI inflation, - a mixed frequency coincident indicator containing real GDP (quarterly) and industrial production, a services index and retail sales (all monthly). - The model controls for financial conditions driven by changes in monetary policy by including a measure of reserve money. - The model assumes time-varying factor loadings (weights of each variable in the factor) that follow an AR specification to account for structural breaks and non-linearities. ### Estimation and tail-risk methodology - Forecasts of the distribution of financial conditions are constructed using quantile projections. - The financial conditions projection equation (as specified) decomposes drivers into principal components of: - p: price of risk (asset prices and risk spreads), - Agg: financial leverage (household and corporate debt-to-GDP ratio, corporate bond issuance, NPLs, bank risk-weighted assets), - y: domestic real economic conditions (real GDP growth, exports, IP), - f: global or foreign variables (global growth and financial risk sentiment). - This approach disentangles contribution of changes in the price of risk, evolving financial leverage, and external shocks when forecasting risks to overall financial conditions in Korea. ### Risk to financial conditions — findings and projections - Based on current domestic macro and financial conditions and the global risk environment as of the 3rd quarter of 2018: - The median projection and high probability outcome suggests that financial conditions in Korea over the next 4 quarters will remain quite neutral, implying no real tightening in the ability of firms and households to obtain new financing. - The distribution reports a median FCI estimate of around 0. - In an extreme event with tail-risks materializing, financial conditions in Korea would be expected to tighten considerably, with the FCI falling from its current level of 0 to -2; this would have significant repercussions for the real sector. - Tail-risk dynamics: - Rolling estimates of the tail of the distribution (the 5th quantile) of financial conditions since 2010 show that since 2010 financial stability tail-risks have not substantially risen; the tail of the distribution has not become fatter or more skewed. - The materialization of tail-risk outcomes is susceptible to external financial conditions: a tightening in external financial conditions (proxied by a standard deviation shock in the VIX) would tighten financial conditions in Korea and increase the probability of tail-risk outcomes considerably. - The sensitivity to external tightening is heightened by high household debt, much of which is linked to floating rates. ### Link to authorities’ views - The finding that financial risks remained contained in 2018 is consistent with the Bank of Korea’s own financial stability index published in the December 2018 Financial Stability Report. ### Policy implications and recommendations - Tail-risk financial condition outcomes in Korea from recently evolving domestic and external real and financial conditions have not grown over the past year; however, should tail-risks materialize financial conditions are expected to tighten considerably with significant real-sector repercussions. - Policymakers should maintain heightened vigilance regarding risks to financial conditions during periods of benign macro conditions that may facilitate accumulation of financial vulnerabilities. - Changes in the domestic price of external risk are a potent signal of imminent threats to financial conditions in Korea, and can be useful for the swift deployment of: - monetary easing, and - macroprudential policy actions. *Source: 1korea2019002 - Introduction (IMF staff calculations and analysis, as presented in the provided chapter).* ### Labor market slack and the output gap — overview (selected material from subsequent section) - Motivation: - Output gap estimates are key inputs into macroeconomic policy making; labor is the most important input into production, so estimates of available capacity on the labor market are informative for output gap estimates. - Traditional unemployment understates true under-utilized labor resources because it excludes workers outside the labor force who are ready to work and part-time workers willing to work longer hours. - Proposed methodological contributions: - Formulate a multivariate filter (MVF) based on Blagrave et al. (2015) with a strong link between labor market slack and the output gap through a production function. - Broaden the measure of labor utilization by augmenting traditional unemployment on the extensive margin to better capture idle labor resources; this augmented measure shows a clearer relationship with inflation (the Phillips Curve). - Key model structure and equations (as presented): - Start from a Cobb-Douglas production function with capital (K), employment (E), and total factor productivity (A). - Employment E is the product of potential labor force (PLF) and the non-employed fraction (u). - The output gap y is written as the sum of the TFP gap a and labor market slack u, including stochastic components to represent uncertainty in gap estimates. - Structural TFP and the structural level of labor market under-utilization are modeled as stochastic processes with long-term steady states, transitory shocks, and stochastic trends (hysteresis and supply-side effects). - Labor market gap is defined as deviation between actual and structural under-utilization; positive gap implies labor market operates above potential, negative gap implies below potential. - A New Keynesian Phillips curve links current, future, and past inflation with the output gap, capturing price-setting inertia and other shock terms. *Source: 1korea2019002 - Introduction (Selection Issues Paper: "Labor Market Slack and the Output Gap" summary as provided).* ### 14. A key empirical input into the output calculation is the degree of under-utilization of ### A key empirical input into the output calculation is the degree of under-utilization of labor market resources ### Measuring labor market under-utilization - Regular unemployment (survey-based in Korea) has varied around 3-4 percent of the labor force since 1980, with spikes in 2000 and 2009. - The international definition of unemployment requires (i) without work; (ii) available to start within two weeks; and (iii) actively searching (International Labor Organization, 2018). This definition omits: - an internal margin: persons without work but not able to start within two weeks or not actively searching (often classified as outside the labor force); - an external margin: employed workers who want to work more hours (classified as employed). - Two broader measures constructed in the analysis (extensions along the extensive margin): - Unemployment augmented with discouraged workers expressed in percent of the labor force. Discouraged workers: legal working age, want to work, have been looking for a job over the last year, but not currently engaged in job search. - Unemployment augmented with persons classified as inactive for unspecified reasons (reasons other than childcare, house-keeping, schooling, old age, or disabilities). - Levels of the alternative measures: - The two alternative measures hover around 5.5 and 10.5 percent, respectively, and show more business-cycle volatility than regular unemployment. - The measure with discouraged workers is considered most relevant because it is confined to workers relatively close to the labor market, while still broader than regular unemployment. - The alternative measures relate to the Labor Utilization Indicator 2 published by Statistics Korea, but the short time period of that series prevents its use in the filter. ### Relationship with inflation (Phillips curve) - The Phillips curve links inflation and regular unemployment; typically a tight negative relationship as lower cyclical unemployment puts upward pressure on wages and then inflation. - For Korea the Phillips curve based on regular unemployment is weak and not negative for all time periods (Bhattarai, 2016). One interpretation: regular unemployment does not fully capture unutilized labor market resources. - The broader measure (unemployment plus discouraged workers) yields a negative Phillips curve for inflation since 2001, suggesting it better captures labor market under-utilization. ### Output gap estimation: model, data, and scenarios - Model framework: multivariate filter (MVF) with output and labor market slack linked through a production function (modification of Blagrave et al. (2015)). - Data and sample: - Annual data from 2001 for GDP, inflation, employment, TFP, unemployment and alternative labor slack measures. - Historical data up to 2017; forecasts applied for 2018-23. - Parameters estimated using Bayesian estimation. - Output gap estimates (2007-2023) under three labor-slack inputs: - Using regular unemployment: - Output gap around +1 percent in 2007/08, turned negative in 2009. - Returned to up to slightly below 1 percent in 2010-11, then turned negative in 2012. - From 2012 to 2014 the gap widened to around -0.8 percent before starting to narrow. - From 2018 onwards the output gap displays a close correlation with the labor market gap as TFP is projected to evolve according to potential. - Exceptions where output and labor market gaps have opposite signs (e.g., 2010, 2012, 2013) can occur when the TFP gap pulls in the opposite direction. - Using unemployment augmented with discouraged workers: - Produces a more volatile labor market and output gap. - More consistency between labor market gap and output gap up to 2012. - Both gaps deteriorate more in 2009; during 2012-17 the unemployment gap remains positive while the output gap is negative. - From 2017 onwards the two gaps are projected to move in tandem. - Using unemployment augmented with inactive persons for unspecified reasons: - Also yields a volatile labor market gap. - Output gap broadly similar to the gap produced using regular unemployment. - Indicates the largest labor market gap in 2017, and produces a consistent labor market and output gap from 2017 onwards. - Robust findings across all three models: - Up to 2017 the labor market and output gaps display conflicting signals at times, because deviations in TFP from trend also contribute to the output gap. - From 2017 onwards both the output and labor market gap are projected to be negative across all models, since TFP is projected to grow according to trend and the only factor contributing to the output gap is the labor market gap. - Labor market gaps computed using broader measures of under-utilization display larger fluctuations on business-cycle frequency. - The slack measure with discouraged workers may be most relevant (see paragraph 17). ### Conclusions and implications for measurement - The paper modifies the MVF by tying output and labor market slack together through a production function to reflect that lower labor market utilization ceteris paribus implies an economy running below capacity. - Regular unemployment is well-defined but narrow; it omits workers outside the labor force who are able and willing to work but not actively searching. - Augmenting unemployment with workers outside the labor force who are potentially able to take an incoming job offer produces a more standard Phillips curve and is therefore a useful input for output-gap estimation. - Using broader measures of labor under-utilization as input to the MVF affects the estimated output gap materially, increasing measured volatility and improving the inflation-unemployment relationship. ### Supplemental labor market and policy context (selected statistics and programmatic facts) - Korea employment and participation: - Employment rate (age 15-64) nearly 67 percent in 2018, 5 percentage points higher than in 2000. - OECD average employment rate about 69 percent in 2018. - Female employment rate about 57 percent in 2018, 7 percentage points lower than the OECD average. - Male employment rate 76 percent in 2018. - Long-term unemployed (12 months or longer) share 0.9 percent in Korea versus OECD average 30.5 percent. - Shares of temporary and self-employed workers: 21. 9 percent and 25.5 percent, respectively, versus OECD averages of 11.2 and 15.8, respectively (2016). - SME support in Korea: - As of 2017, 288 central government programs and over 1000 local government programs to support SMEs. - Central government spending on such programs estimated at around 0.8 percent of GDP in 2017. - Credit guarantees to SMEs were 3.8 percent of GDP in 2016 (second largest in the OECD). - Reduction in combined corporate and personal rates on capital gains due to preferential SME corporate income tax treatment estimated at 13 percentage points in 2014. - Evidence on fiscal support to SMEs and employment: - Park and others (2017) find some positive and significant impact of fiscal support to SMEs on employment over 2011-2013 in firms from different sectors. - MOSF and KIPF (2018) find a positive, but small impact of tax credits for SMEs on employment. - Chang (2016) finds recipients of 2009 public SME finance programs experienced lower productivity growth than non-recipients. *Source: IMF staff analysis in "Labor Market Slack and the Output Gap" (Republic of Korea chapter).* ### 10. SMEs’ difficulty in creating jobs is also related to the fact that Koreans, especially ### 10. SMEs’ difficulty in creating jobs is also related to the fact that Koreans, especially ### SMEs, R&D incentives, and youth employment - Problem statement: - Koreans, especially young people, often are unwilling to work for SMEs because SMEs provide fewer training opportunities and wages are considered low. - Providing subsidies to SMEs might not be an effective way to boost employment in SMEs. - Policy orientation: - Help SMEs become more productive and dynamic to make them more attractive to youngsters. - Redirect subsidies toward R&D activities that would help make SMEs more dynamic, rather than protecting jobs in weak firms. - Evidence and comparative points: - Korea’s marginal tax subsidy rate on R&D by SMEs is high compared to some other OECD countries (Figure). - Empirical analysis on advanced economies indicates that R&D tax incentives have a larger effect for small firms than larger companies (IMF, 2016). - Evaluation studies on selected advanced economies suggest that R&D tax incentives for small firms are two to three times more effective in promoting R&D investments than for an average size firm (IMF, 2016). - There may be scope to re-orient some other forms of fiscal support currently extended to SMEs toward their R&D activity. - Recommendation bullets from the chapter: - Subsidies should be better focused on R&D activities rather than protecting jobs in weak firms. - Important to support young and innovative firms that are more likely to achieve productivity gains and long-lasting increases in employment. ### Policy implications for SME support (explicit recommendations) - Given the large number of fiscal programs in support of SMEs, regularly monitor and periodically review their impact; consider setting up a comprehensive and integrated database on the different programs. - Instead of extending subsidies only on the basis of firm size, identify firms more likely to attain increases in productivity and profitability and bring about long-lasting employment gains, such as young and innovative firms. - Design fiscal incentives to help SMEs become more productive and dynamic to make them more attractive to youth. - Consider re-orienting some other forms of fiscal support currently extended to SMEs toward their R&D activity, given R&D tax incentives for small firms are particularly effective in promoting investment. ### Active Labor Market Policies (ALMPs): current stance and issues - Definition (OECD): programs specifically designed to support employment include (1) public employment services (PES); (2) training; (3) employment incentives; (4) direct job creation; (5) start up incentives; (6) subsidies for employment of persons with permanently (or long-term) reduced capacity to work. Programs to help the unemployed include unemployment benefits and programs facilitating early retirement of older workers. - Korea’s overall ALMP spending: - Korea’s spending on measures to support employment: 0.4 percent of GDP (close to the OECD average). - Composition and concerns: - Korea ALMPs are largely focused on direct job creation, with this spending category representing more than half of government expenditure on employment support in 2016. - This compares with the OECD average share of about 13 percent for direct job creation. - About 21 percent of direct job creation spending was targeted at workers above the retirement age of 65 years in 2018. - ALMP spending for the working-age population is somewhat lower than data suggest. - Alternative fiscal instruments (e.g., targeted transfers) may be more efficient and effective to support the elderly. ### Public Employment Services (PES), training, and employment incentives - PES and training: - Korea’s spending on PES: 0.04 percent of GDP and about 10 percent of employment-supporting ALMPs. - OECD average PES spending: 0.1 percent of GDP and nearly 28 percent of total spending on employment-supporting ALMPs. - Korea’s spending on training is lower than the OECD average and constitutes a relatively small share of total ALMP expenditure. - Employment incentives: - Korea devotes more than 13 percent of its ALMPs expenditure to employment incentives, amounting to about 0.05 percent of GDP in 2016. - Employment incentives are the second-largest ALMP item for Korea after direct job creation. - Rationale: compensate firms that would hire or preserve jobs when a worker’s productivity is temporarily lower than the wage, expecting productivity improvement through learning-by-doing, training, or monitoring. - Concerns about targeting and monitoring: - Unclear the extent to which individual cases are reviewed and monitored periodically in Korea. - Risk that employment incentives could be kept in place over an extensive period to preserve employment regardless of workers’ productivity. - Indications incentives might not be well targeted (e.g., incentives supporting hiring of highly educated workers who firms may have hired without subsidies). - Alternative estimate / note: - According to Park (2016), Korea’s spending on employment incentives in 2016 was about 0.2 percent of GDP, with 20 employment incentives in operation under four ministries for a total budget of 2.8 trillion won. ### ALMPs: policy implications (explicit recommendations) - Reconsider the composition of ALMPs. - Increase spending on public employment services that support job-matching and on training. - Consider supporting income for the over 65 through targeted transfers rather than spending on employment support. - Periodically review the effectiveness of employment incentives and reassess whether they are properly targeted. ### Public sector employment: size, plans, and trade-offs - Korea’s public employment: - Share of employment in general government relative to total employment: less than 8 percent (one of the lowest among OECD countries). - Note: these data reflect employment in general government, excluding SOEs. - In 2015 about 80 percent of people employed in the central government were between 35 and 54 years of age. - Government plans: - The government is planning to expand public sector jobs by 810,000 by 2022 (MOEF, 2018). - In the 2019 budget the government indicated the intention to increase social services jobs by 60,000 to 94,000 and civil service jobs (e.g., policy officers and) by 21,000. - Economic trade-offs and evidence: - Expanding public sector jobs can reduce unemployment in the short term and stabilize in recessions or disadvantaged regions, but efficiency for long-term economy-wide employment is questionable. - Theory: if public-sector production substitutes private-sector production, expansion can crowd out private jobs. - Empirical example: Algan, Cahuc and Zylberberg (2002) find that in a group of OECD countries in 1960–2000, on average the creation of 100 public jobs may have eliminated about 150 private sector jobs. - Public jobs that pay attractive wages and/or offer benefits tend to attract more workers and could crowd out private jobs. - Recommendations regarding public employment: - Link public sector job creation to developing or expanding services that cannot be provided adequately by the private sector. - Assess and take into account the possible impact on public sector productivity when creating public jobs. - Consider conducting comparisons between public and private sector wages to inform wage setting and hiring conditions in the public sector. ### Empirical analysis: drivers of employment rates and fiscal policy roles - Objective and framework: - Assess determinants of employment rates (ages 15-64) across OECD economies since 1985 using cross-country panel regressions. - Policy variables included: labor tax wedge; wedge between tax rates of second earners and single individuals; public expenditure on ALMPs; stringency of employment protection regulations for permanent workers (EPL); public spending on early childhood education and care (percent of GDP); number of weeks of job-protected parental leave available to mothers. - Controls: output gap (one-year lag for potentially endogenous variables), ratio of employment in service vs industrial sector, share of urban population, trade openness, exposure to technological progress, share of population with tertiary education, union density, country and time fixed effects. - Key estimation results: - Public expenditure on ALMPs and childcare spending: coefficients positive and statistically significant for men and women, but larger for women. - Labor tax wedge: a high wedge (ratio between average tax paid by a single-earner family and corresponding total labor cost for the employer) is negatively and significantly related to employment. - Wedge between tax rates of second earners and single individuals: estimated not to have a significant impact on employment. - Longer leave available to mothers: positive and significant impact on female employment. - Stringency of EPL for permanent workers: more rigid regulation associated with lower female employment rate. - Among controls: employment depends on the business cycle; structural transformations shifting demand for certain worker types have significant effects; higher education positively associated with employment rates for both men and women; unionization not found to have a significant impact on employment. - Korea-specific fiscal implications drawn from results: - Family-friendly policies are effective in promoting labor market participation, especially for women. - Korea’s public spending on early childhood education and care was almost 0.9 percent of GDP in 2013 (higher than the OECD average of 0.7 percent), but 0.5 percentage points lower than the average of OECD countries that spend the most on this item. - Given low female employment, boosting spending on early childhood education and care would be critical. - Reducing the labor tax wedge is less critical in Korea because Korea’s tax wedge is relatively low. - Relaxation of employment protection for regular workers could support female employment; Korea’s restrictions on dismissal of regular workers have been stricter than the OECD average in recent years. - Fiscal measures can facilitate reforms to make the job market more flexible: increase spending on ALMPs to accompany lowering of employment protection, since looser firing restrictions can initially raise unemployment—more job training and matching services can reduce risk of extended unemployment spells. - Historical precedence: the Netherlands (2009), Finland (1992), and Portugal (1990) lowered employment protection and increased ALMP spending simultaneously (IMF, 2014, 2016b). *Source: Chapter excerpt from 1korea2019002 (IMF).* ### Conclusion and Policy Implications ### Conclusion and Policy Implications ### Boosting employment: priority and existing measures - Boosting employment is a priority for Korea both in the near and longer term. - The Korean government is implementing policies to support job creation, including: - transfers to SMEs to create or preserve jobs, - ALMPs (active labor market programs), - an expansion of public employment. - Review of measures suggests scope to enhance their design and implementation. ### Fiscal support to SMEs and recommendations - Korea has set up a large number of fiscal programs to support SMEs with the aim of creating or preserving jobs. - Recommendations to maximize effectiveness: - Periodically review the impact of SME support measures. - Consider establishing a comprehensive and integrated database on the different programs. - Focus fiscal incentives on making SMEs more productive and dynamic to attract younger workers. - Given that R&D tax incentives for small firms are particularly effective in promoting their investment, consider re-orienting some other forms of fiscal support currently extended to SMEs toward their R&D activity to help SMEs become more profitable and dynamic. - Rather than providing subsidies based only on firm size, focus support on firms more likely to experience improvements in productivity and long-term profitability (e.g., young and innovative firms). ### Active labor market policies (ALMPs) - ALMPs could be enhanced and used more effectively to boost employment in Korea. - Public spending on training and employment services that support job-matching is small in international comparison and could be boosted. - Periodic reviews of the effectiveness of employment incentives are important to reassess targeting. ### Public employment: scope and cautions - There is scope to increase public employment, given Korea’s share of public sector employment relative to total employment is low. - Job creation in the public sector should be driven by the need to develop or expand services that cannot be provided adequately by the private sector. - In expanding public employment, assess possible impacts on public sector productivity. - To minimize crowding out of private jobs, wage setting and hiring conditions in the public sector should take into account the compensation differential vis-à-vis the private sector. ### Empirical indications and fiscal policy levers - Empirical analysis on determinants of employment in OECD economies indicates fiscal policies that can support employment in Korea. - There is scope to increase public spending on early childhood education and care, given its effectiveness and Korea’s low female employment rate. - Korea’s employment, especially among women, can be supported by easing regulation on regular contracts, which has been comparatively strict in recent years. - Fiscal policy can play an important role in facilitating reforms to enhance the flexibility of the job market. ### Drivers of Employment: Regression Results (selected coefficients and statistics) - Regression results (columns: All, Women, Men): - ALMP(‐1): 0.05 0.08 0.05 - P‐value: 0.00 0.00 0.00 - Labor tax wedge: ‐0.18 ‐0.16 ‐0.14 - P‐value: 0.00 0.00 0.00 - Childcare spending: 3.38 6.11 1.68 - P‐value: 0.00 0.00 0.00 - Job‐protected maternity leave: 0.02 0.05 0.01 - P‐value: 0.00 0.00 0.33 - EPL: 0.53 ‐2.17 1.86 - P‐value: 0.42 0.01 0.76 - Output gap (‐1): 0.20 0.19 0.22 - P‐value: 0.00 0.02 0.00 - Trade openess (‐1): 0.04 0.10 0.05 - P‐value: 0.00 0.00 0.00 - Relative service employment: ‐0.04 ‐0.03 ‐0.05 - P‐value: 0.00 0.00 0.00 - Union Density: 0.00 ‐0.04 ‐0.04 - P‐value: 0.94 0.30 0.27 - Exposure to technology: 0.62 ‐0.04 0.24 - P‐value: 0.00 0.90 0.10 - Tertiary education: 0.10 0.15 0.11 - P‐value: 0.00 0.00 0.00 - Model fit and sample: - R‐squared: 0.93 0.93 0.86 - Estimation period: 1985‐2014 1985‐2014 1985‐2014 - Number of observations: 558 558 558 - Note: Bolded coefficients in the original are significant at at least 5 percent. ### Implementation of maximum working hours: international practices and implications for Korea - Korea is reducing the maximum weekly working hours from 68 to 52; implementation should be monitored to ensure an effective and widespread decrease in hours worked across sectors. - Key contextual facts and international comparisons: - Korea’s average hours worked per employee are the second highest in the OECD and about 19 percent above the OECD average. - On March 20, 2018, Korea amended the 1997 Labor Standards Act to include weekends in the definition of the weekly working period, reducing the maximum allowed working hours to 52 per week. - Exceeding the limit is allowed as long as average weekly working hours do not exceed 52 hours over a reference period. The reference period was initially set at three months but will likely be extended to six-months. - Implementation will be phased in over a three-year period, starting with large companies and the public sector. Management positions are exempted from working hours regulations. - For employees under 18 the new ceiling is 35 hours per week. Five industries, including transport and health care, are exempt from the limits, while twenty-six were exempt under the previous rules. - International experience highlights: - 40-hour or less is the most prevalent limit for regular weekly hours in advanced economies. - EU Directive 93/104/EC (later 2003/88/EC) established 48 hours as the maximum average working time for each seven-day period, calculable over a reference period of four months, with derogations for managers/professional staff and an individual opt-out option under conditions. - In practice, modalities vary across EU members; on average, 16 percent of workers continue to work 48 hours or more weekly in the EU. - U.K.: individual opt-out widely used; in 2011, 32.4 percent of workplaces had at least one worker who had signed an opt-out agreement. - France: stricter limits (weekly average of 44 hours over three months, 48 in a given week); company agreements allowed up to a weekly average of 46 over three months. - Japan (2018 law): cap of 100 hours a month on overtime and limit of 720 hours of overtime a year (average 15-16 hours a week); white-collar workers with annual incomes > 10.75 million yen are exempted, with an option to give up exemption; effective April 2019 for large firms and April 2020 for SMEs. - General international finding: caps on working hours can be an effective way to reduce hours worked; flexibility in implementation (reference periods, derogations, opt-outs) is common and reductions in hours have been observed across jurisdictions. _International Monetary Fund — Conclusion and Policy Implications (Republic of Korea)._ ### 11. In Korea, it will be important to ensure a widespread implementation of working ### 11. In Korea, it will be important to ensure a widespread implementation of working hours ceilings across sectors. ### Working hours reform: rationale and priorities - Policy recommendation: - Ensure a widespread implementation of working hours ceilings across sectors. - Rationale and expected effects: - Korea’s hours worked per employee are well above the OECD average. - The new maximum working hour limit is high relative to the EU limits. - Effective implementation of Korea’s new legislation will be essential to improve people well-being and support female labor force participation and fertility. - The impact on firms and their productivity will have to be regularly evaluated. ### Industrial structure: concentration and interconnectedness - Sectoral shares and evolution: - Manufacturing share of GDP: "12.7 percent in 1970s" to "31.6 percent in 2010s". - Services share of GDP: stable around "60 percent". - Other sectors (non-manufacturing and agriculture): declined over time in the share of GDP. - Concentration: - Manufacturing follows a U-shaped relationship between development and diversification. - Share of the three largest manufacturing industries: "41.9 percent in 1990s", "57.6 percent in 2010s". - Services top three industries share: "39.5 percent in 2010s". - Interconnectedness (input-output relationships): - Overall domestic vertical linkages (share of intermediate input in total output) increased to "47.1 percent in 2009 from 44.3 percent in 2000", then remained stable afterwards. - Upstream and downstream linkages: above "50 percent" in manufacturing; below "40 percent" in services. - In manufacturing, vertical linkages increased until the global financial crisis (GFC) and then declined; in services they remained stable. ### Major manufacturing industries and trade linkages - Major manufacturing industries defined (by 2014 GDP share): electronics, transportation, chemicals. - Upstream/downstream patterns: - Upstream linkages highest in chemicals at almost twice higher than electronics and transportation. - Downstream linkages around "60 percent" across electronics, transportation, and chemicals. - Upstream and downstream linkages in the three major manufacturing industries declined slightly after the GFC. - Trade linkages: - The share of export and import in total output increased by about "5 percentage point each" in the aftermath of GFC. - Manufacturing has substantially higher export and import linkages and showed a sharp increase during the 2000s. - In services, export and import linkages remained stable at around "5 percent" and "3 percent", respectively. - Electronics and transportation show about twice higher export linkages than chemicals; import linkages highest in electronics. - Export and import linkages to China have shown a drastic increase over the 2000s in major manufacturing industries; linkages to the U.S. and Japan have declined. - Export and import exposures to the top three trading partners account for almost "50 percent" of total trade linkages in major manufacturing industries (except for export linkages in transportation). ### Macroeconomic development: contributions to growth and volatility - Historical growth: - Korea’s economic growth: "4.3 percent in 2000s" and "2.8 percent in 2010s" (after averaging above 6 percent over several earlier decades). - Contributions to GDP growth: - Manufacturing and services remain key contributors but their contributions have declined over time. - Concentration of growth contribution in manufacturing: - Electronics contributes "almost half of the manufacturing growth since 2000s". - Largest three industries’ growth contribution: "26.5 percent in 1980s", "almost 70 percent since 2000s". - Service sector major-three contribution has remained around "30-40 percent" since the 1980s. - Volatility: - Growth volatility declined after the 1990s (Asian Financial Crisis period) and through the 2000s. - Level of growth volatility higher in manufacturing than in services and other sectors in 2000s. - Electronic and transportation industries have been relatively volatile compared to other manufacturing and service industries. - Risks: - High concentration and interconnectedness around a few major, relatively volatile industries, dominated by a few large firms, can increase susceptibility to domestic or external shocks. ### Shock transmission channels and empirical approach - Transmission channels: - Domestic shocks propagate via upstream/downstream linkages (input supply and input purchase/use). - External shocks propagate via direct export/import linkages and indirectly via domestic industries’ trade linkages. - Empirical strategy: - Uses industry-level international input-output data from WIOD (2000-2014). - Extends Acemoglu et al. (2016): focuses on the role of network linkages in transmitting (i) domestic industry supply shocks and (ii) external country GDP growth shocks from Korea’s three largest trading partners (China, the U.S., and Japan) to industry labor productivity growth. - Estimation examines how upstream/downstream/linkage measures and export/import exposures mediate the impact of shocks on industry labor productivity growth. - Model elements (as presented in source): - Variables include industry labor productivity (ܲܮ ௜௧), real value added (ܻ ௜௧), capital intensity (ܭ ௜௧), upstream and downstream linkage weights (ܱܫ), export and import linkage weights (ܺܧ, ܯܫ), and industry/sector-year fixed effects. - The model considers first-order direct trade and vertical linkages between Korea and other countries and between two industries. *REPUBLIC OF KOREA INTERNATIONAL MONETARY FUND* ### 18. Industry growth is estimated to be significantly affected by both downstream ### Industry growth is estimated to be significantly affected by both downstream domestic industry shocks and own export shocks ### Estimation results: propagation channels and directionality - Supply-side domestic industry shocks have larger downstream effects (due to supplier shocks) than upstream effects (due to customer shocks); industries are more likely to be affected by seller’s growth shocks rather than buyer’s growth shocks. - External country shocks are propagated to domestic industries mainly through own (direct) export linkages — industries highly involved in exports directly to foreign countries are significantly affected by the country’s growth shock. - Separate estimation for manufacturing industries: - Propagation of domestic industry shocks via both upstream and downstream linkages is significant. - External shock propagation via direct export linkages is effective for manufacturing. - Other controls: - Lower productivity industries have a higher growth (i.e., convergence or catch-up). - Industries with higher capital intensity growth are growing faster. ### Key quantitative impacts on productivity growth - Domestic industry shock impacts (total industry): - Growth shocks in chemicals (one standard deviation) are estimated to have the largest productivity impacts: 0.52 percentage point. - Growth shocks in electronics and transportation are also expected to have substantial productivity impacts. - External country shock impacts (total industry): - China’s growth shocks would have the largest productivity impacts for total industry: 0.26 percentage point. - Growth shocks in the U.S. and Japan would have smaller but substantial productivity impacts. - Manufacturing vs total industry: - Effects are expected to be larger than twice for manufacturing compared with total industry due to stronger vertical and trade linkages. - Overall impact of one standard deviation shocks: - Total industry: 1.4 percentage points. - Manufacturing: 3.2 percentage points. - Implication: economic shocks in the largest three industries and/or in the top three trading partners can lead to large swings in the overall economy by transmission through vertical and trade linkages. ### Robustness and country-specific findings - Propagation of external country shocks through downstream and direct external linkages is robust to individual country shocks. - For China and Japan: - Significant role of downstream linkages and direct export linkages in transmitting shocks, consistent with baseline specification. - For the U.S.: - For total industry, domestic downstream effects are significant while direct export effects are estimated to be negative but not statistically significant. - For manufacturing, domestic upstream effects are estimated to be significant for the U.S. and Japan. ### Structural features and vulnerability - Specialization and concentration: - The electrical and electronic equipment industry has contributed almost half of the growth in the manufacturing sector since early 2000s. - A small number of dominant industries are highly interconnected via upstream/downstream linkages and with foreign markets via export/import linkages, and these industries are dominated by a few large firms. - Increased vulnerability: - Tighter vertical and trade linkages have increased the vulnerability of the economy to domestic and external shocks. - High trade linkages to China and the U.S. may lead to vulnerability to unfavorable economic episodes in these countries (for example, recent trade tensions between China and the U.S. are expected to have significant impacts on growth and stability in Korea). ### Policy implications and recommended actions - Diversification of industrial structure would help mitigate macroeconomic instability from propagation of economic shocks. - Further efforts to mitigate concentrated economic structure could include: - Promoting fair competition between large corporations and smaller firms. - Fostering innovation especially in SMEs. - Reducing the regulatory burden on firms, especially in the service sector. *Source: IMF staff calculations and analysis in the Korea country study chapter.* --- _Source: https://www.imf.org/-/media/files/publications/cr/2019/1korea2019002.pdf_