Abstract
This paper analyzes the implications of global rebalancing in the post-crisis period for Korea and how high leverage in the household and SMEs sectors could affect this process. The first section of the paper discusses implications of global rebalancing for Korea using simulations from the Global Integrated Monetary and Fiscal Model (GIMF). The second section focuses on how rebalancing growth in Korea is different from the rest of the region, and discusses the challenges of highly leveraged households and SMEs for the rebalancing process. The last section concludes with policy recommendations.
III. Labor Market Dynamics in Korea—The Role of Institutions and Shocks1
A. Introduction
1. The recent global financial crisis or “Great Recession” led to the deepest downturn in the global economy since World War II. It particularly hit economies open to trade and integrated with the global financial system, including Korea. However, the global economy is now recovering and Korea is at the forefront of the rebound, having benefited from a strong policy response, the normalization of international trade, and the return of investor risk-appetite.
2. During the “Great Recession” and the ongoing recovery, labor market dynamics have differed widely across countries. Some countries, such as the United States and Spain, have seen significant job losses and steeply rising unemployment rates, while others, including Korea, Japan, and Germany have experienced more muted dynamics for employment and unemployment. What explains this divergence in labor market dynamics?
3. This paper analyzes the determinants of labor market dynamics across a diverse set of advanced economies, including Korea, during recessions and recoveries over the past 40 years. In particular, the paper focuses on the importance of institutional factors such as the degree of employment protection, the generosity of unemployment benefits, and the share of temporary workers. It also focuses on the role played by the “nature” of recessions in determining labor market dynamics. For example, what are the implications if a recession was caused by a financial crisis and/or the bursting of a housing-market bubble? Also, do pre-conditions such as soundness of corporate balance sheets matter? Further to this, the paper analyzes the role played by policies in Korea and other countries, focusing in particular on government employment programs during the “Great Recession.”
4. Based on this analysis, the outlook for employment during Korea’s ongoing recovery is assessed and some preliminary policy implications are drawn. On the latter, the paper discusses the exit strategy from the employment support programs implemented by the Korean government during the crisis. Furthermore, it briefly discusses policy options to support employment growth and enhance labor market flexibility over the medium term, including the scope to reduce employment protection legislation, adjust the focus of training and education with future needs, and other steps to reduce the duality in the Korean labor market.
B. Labor Market Dynamics During the “Great Recession”
5. Korean labor markets did not escape the adverse spillovers from the recent global economic downturn, but the impact was cushioned by policy measures (Figure 1).
Korea—Labor Market Dynamics During the Crisis
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Labor Market Dynamics During the Crisis
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Labor Market Dynamics During the Crisis
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Labor force. Between June 2008 and December 2009, the labor force participation rate declined from 61½ percent (seasonally adjusted) to 60½ percent (seasonally adjusted), but then increased as workers re-entered the labor market to take up jobs offered under government work programs. The initial drop in the participation rate was most pronounced for women and workers with lower levels of education. They typically take up a larger share of temporary jobs and, in the latter case, lower skilled jobs and, therefore, are more likely to face layoffs and lack of job opportunities when a downturn hits. Among age-cohorts, the younger workers also left the labor market at a faster pace during the initial stages of the crisis. However, many of these more “vulnerable” groups have since returned to the labor market in conjunction with the introduction of the government work programs. However, the overall participation rate remains below pre-crisis levels, currently standing at 61 percent (seasonally adjusted).
Employment. Employment losses from the downturn were particularly concentrated in financial services, manufacturing, and other cyclically sensitive sectors exposed to the adverse spillovers from the global crisis. However, the government’s job creation program and other measures introduced in response to the crisis saw employment in public administration, education, and social services rise by close to 500,000 between June 2008 and August 2009, almost matching the decline in private employment during the same period. These measures, together with the high level of employment protection, helped sustain employment growth for regular workers throughout the crisis. The job-creation programs also helped support the employment of temporary workers, which declined significantly during the early stages of the crisis. On the other hand, the employment of daily workers and self-employed declined throughout the crisis. However, with the economic recovery gaining pace, job gains in the private sector have picked up in 2010 and are becoming more broad based.
Unemployment. Notwithstanding the decline in labor force participation, measured unemployment also took a hit during the first year of the crisis, impacting in particular the younger cohorts during the early stages of the crisis. While the government job-creation programs helped cushion the fall in employment, a delayed implementation of the latter stage of the program led to a temporary spike in the unemployment rate to almost 5 percent in early 2010 because labor force participation rose as workers signed up for the program. However, the corresponding increase in public sector employment did not take place until a few months later, which then led the unemployment rate to fall back again and it has since continued to decline.
Hours and wages. During the course of the crisis, the average work week declined by around 3 hours to 44 hours. Monthly wages also fell as hours worked and hourly wages were cut, declining on an annual basis between December 2008 and September 2009 for all industries. However, the growth rate has returned to positive territory over the past few quarters, although this primarily reflects a lengthening of the work week as hourly wages are still declining. The initial drop in average monthly wages was most pronounced for cyclically sensitive sectors such as manufacturing, construction, and financial services. By employment status, nonregular workers saw the largest declines in monthly pay.
6. Compared to the Asian Crisis, labor markets suffered less this time around (Figure 2). During the Asian crisis in 1997-98, the employment rate bottomed out close to 10 percent below the cycle peak, while it during the current crisis only declined by 1½ percent relative to the peak level. Correspondingly, the unemployment rate rose significantly more during the Asian Crisis, up by almost 6 percentage points compared to around 1 percentage point this time around. The divergence between the dynamics during the Asian and the current crisis to a large extent reflects that the economic downturn was much deeper in the late 1990s, with GDP per capita dropping by close to 10 percent from peak to trough. During the Great Recession, on the other hand, GDP per capita only fell by around 5 percent from peak to trough, supported by the proactive policy response and the much sounder fundamentals of the Korean economy this time around (i.e., stronger balance sheet positions of the government, financial institutions, and large corporates), making it more resilient to the adverse global spillovers. Moreover, wages and hours worked also adjusted more during the current crisis, cushioning the impact on employment.
Korea—Comparing the Asian Crisis and the “Great Recession”
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Comparing the Asian Crisis and the “Great Recession”
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Comparing the Asian Crisis and the “Great Recession”
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
7. Korean labor market dynamics diverged from those seen in other advanced economies during the current crisis (Figure 3). For example, the employment rate fell by 4 percent less than in the United States and the unemployment rate rose by around than 3½ percentage points less. This partly reflects that Korea’s economy has rebounded faster than the U.S. economy, employment protection is higher in Korea, and hours, wages, and participation rates fell more, cushioning the impact on employment and (measured) unemployment. These were also broadly the reasons why Korean labor markets were less adverse impacted than in countries such as Ireland, Spain, and the United Kingdom. However, employment conditions weakened relatively more in Korea than in Germany. This likely owes much to Germany’s higher level of employment protection and the German government’s massive expansion of a short-term work program. At the same time, the decline in the employment rate and increase in unemployment since the business cycle peak was broadly in line with the trend seen in Japan.
Korea—Comparing Korea to Other Countries
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Comparing Korea to Other Countries
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Korea—Comparing Korea to Other Countries
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
C. What Explains Labor Market Dynamics Across Countries?
8. As the previous section highlighted, the labor market dynamics during the “Great Recession” differed across advanced economies. While this partly reflected differences in output losses, it would appear that institutional factors were also at play. Moreover, the nature of the shock differed across countries, with some countries hit by a multitude of shocks (financial crisis, the bursting of asset bubbles, and trade shocks) and others, including Korea, primarily hit through the trade channel. In addition, policy responses varied in terms of both magnitude and type of measures.
Unemployment Rate and Output During Crisis
(Change in percentage points and percent, respectively, peak to trough)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Sources: CEIC Data Company; and IMF staff calculations.Unemployment Rate and Output During Crisis
(Change in percentage points and percent, respectively, peak to trough)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Sources: CEIC Data Company; and IMF staff calculations.Unemployment Rate and Output During Crisis
(Change in percentage points and percent, respectively, peak to trough)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Sources: CEIC Data Company; and IMF staff calculations.9. To analyze the respective roles of institutions and the nature of shocks, the papers applies Okun’s law as an organizing framework.2 Okun’s law captures the relationship between unemployment and output and can be expressed as follows:
where δu is the change in the unemployment rate, α is the intercept coefficient, and β is the elasticity of the unemployment rate with respect to changes in output. Based on the estimation of this simple equation for a diverse set of advanced economies and using a panel regression setting, the paper will assess (i) which institutional factors could explain the difference in unemployment responsiveness to output changes and (ii) to what extent the “nature” of shocks has a bearing on labor market dynamics during recessions and recoveries (proxied by the forecast errors of the estimated Okun’s law). Finally, we will assess the impact of some of the employment programs implemented in a number of countries.
The role played by institutional factors
10. To determine the responsiveness of unemployment to output changes, a dynamic version of Okun’s law is estimated for 21 advanced economies, including Korea.3 4 For each country, the dynamic version is estimated for the 20 years prior to each recession that the country has gone through based on quarterly data. Given that all countries have experienced at least one recession over the past 50 years, the estimated equations gives us a set of dynamic βs across countries and over time. The results show that the responsiveness of unemployment to changes in output has been higher in the years preceding the current crisis than it was in the 1990s. Moreover, the estimated coefficients for the dynamic βs reveal large cross country differences, with responsiveness very high in countries like Canada and Spain, while it is low for countries like Norway and Japan. For Korea, the dynamic /? is estimated to be at the low end when controlling for crisis periods by using a crisis dummy in the Okun’s law regression.
Dynamic Betas for 20 Years Prior to Great Recesssion
(Responsiveness of unemployment w.r.t output)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Dynamic Betas for 20 Years Prior to Great Recesssion
(Responsiveness of unemployment w.r.t output)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Dynamic Betas for 20 Years Prior to Great Recesssion
(Responsiveness of unemployment w.r.t output)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
11. A panel regression is estimated to gauge the role played by institutional factors, over time and across countries,. In the panel regression, the dynamic βs are regressed on a set of institutional factors5 (see Table 1, equation 1 to 5):
Employment protection. As expected, the panel regression shows that stricter employment protection makes unemployment less elastic to changes in output. This is because the higher costs of firing and hiring makes the employer more reluctant to both lay off workers during downturns and hire them during upturns. In Korea’s case, employment protection is relatively high by OECD standards despite a decline over the past few decades, which helps explain the relatively low dynamic β for Korea.
Generosity of unemployment benefits. Theoretically the impact of more generous benefits is ambiguous. During downturns, generous benefits can limit downward wage flexibility and cause more job losses, while they may constrain employment growth during upturns by keeping reservation wages relatively high. The panel regression shows that the former would appear to dominate the latter. Given the relatively low income replacement rates in Korea, unemployment benefits would tend to mute unemployment dynamics compared to other countries.
Share of temporary workers. A priori, a larger share of temporary workers should be associated with larger swings in unemployment during economic cycles due to the lower degree of employment protection for workers with temporary contracts and less sunk investment in them by employers. Moreover, the increased prevalence of this over time could have augmented this for Korea, which has a very high share by OECD standards. However, while the coefficient has the expected sign, it is not significant in any of the panel regressions.
Determinants of Unemployment and Employment Responsiveness to Changes in Output
Determinants of Unemployment and Employment Responsiveness to Changes in Output
| Un Employment Equation | Employment Equation | |||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | ||
| Strictness of employment protection | -0.05 | -0.08 | -0.07 | -0.13 | ||
| [0.050] | [0.007] | [0.007] | [0.03] | |||
| Generosity of unemployment benefits | 0.14 | 0.37 | 0.40 | 0.44 | ||
| [0.193] | [0.001] | [0.000] | [0.003] | |||
| Share of temporary workers | 0.005 | 0.005 | 0.01 | |||
| [0.148] | [0.334] | [0.134] | ||||
| Constant | 0.41 | 0.36 | 0.23 | 0.64 | 0.59 | 0.72 |
| [0.000] | [0.000] | [0.000] | [0.000] | [0.000] | [0.00] | |
| Number of observations | 70 | 66 | 59 | 48 | 39 | 34 |
| R2 | 0.06 | 0.02 | 0.04 | 0.27 | 0.49 | 0.33 |
Determinants of Unemployment and Employment Responsiveness to Changes in Output
| Un Employment Equation | Employment Equation | |||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | ||
| Strictness of employment protection | -0.05 | -0.08 | -0.07 | -0.13 | ||
| [0.050] | [0.007] | [0.007] | [0.03] | |||
| Generosity of unemployment benefits | 0.14 | 0.37 | 0.40 | 0.44 | ||
| [0.193] | [0.001] | [0.000] | [0.003] | |||
| Share of temporary workers | 0.005 | 0.005 | 0.01 | |||
| [0.148] | [0.334] | [0.134] | ||||
| Constant | 0.41 | 0.36 | 0.23 | 0.64 | 0.59 | 0.72 |
| [0.000] | [0.000] | [0.000] | [0.000] | [0.000] | [0.00] | |
| Number of observations | 70 | 66 | 59 | 48 | 39 | 34 |
| R2 | 0.06 | 0.02 | 0.04 | 0.27 | 0.49 | 0.33 |
OECD Measures of Strictness of Employment Protection
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
OECD Measures of Strictness of Employment Protection
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
OECD Measures of Strictness of Employment Protection
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Net Replacement Rate of Unemployment Benefits
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Net Replacement Rate of Unemployment Benefits
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Net Replacement Rate of Unemployment Benefits
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Share of Temporary Workers
(In percent)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Share of Temporary Workers
(In percent)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Share of Temporary Workers
(In percent)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
A panel regression looking at the determinants of the responsiveness of employment produces broadly the same results (Table 1).
12. Simulations show that institutional reform in Korea could have a significant impact on the responsiveness of employment and unemployment to economic activity. Reducing the strictness of employment protection legislation (equivalent to half the current level of Korea’s current EPL index) would increase the elasticity of unemployment and employment (with respect to output) from -0.21 to -0.29 and 0.41 to 0.56, respectively, an increase of close to 40 percent. Aligning Korea’s unemployment benefits and share of part-time workers with the OECD average would also raise elasticity noticeably. A combination of lowering employment protection (cut in half), while raising unemployment insurance (to OECD levels), would increase elasticities of unemployment and employment by more than 50 percent.
Policy Simulations: Impact on Dynamic Beta’s, Unemployment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Policy Simulations: Impact on Dynamic Beta’s, Unemployment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Policy Simulations: Impact on Dynamic Beta’s, Unemployment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Policy Simulations: Impact on Dynamic Beta’s, Employment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Policy Simulations: Impact on Dynamic Beta’s, Employment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Policy Simulations: Impact on Dynamic Beta’s, Employment
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
The role played by the nature of the recession
13. Each recession is different and has different implications for the depth of the labor market shock and the speed of recovery in employment. While this clearly depends on the size of the economic downturn, the nature of the shock hitting the economy can also have an impact on the labor market dynamics during the recession and recovery phases.
14. To analyze the role played by the nature of recessions, a two-stage approach is applied. First, we compare the actual change in unemployment during previous recessions and recoveries for each country to the changes predicted by the estimated Okun law relationships. Second, these forecast errors, controlled for output changes, are then regressed on different shock types associated with previous recessions using a panel data setting (Tables 2 and 3): 6
Financial shocks. Historical evidence points to the protracted nature of recessions and recoveries following financial shocks. Using a financial crisis dummy with a value of 1 during recessions and recoveries accompanied by financial crisis, shows that financial shocks have implications for labor market dynamics. According to the panel regression results, the unemployment rate would be around 0.7 percentage points higher during recessions associated with financial shocks and about 0.2 percentage points higher during recoveries. Other measures of financial shocks, including a financial stress index, also suggest that unemployment will be higher during down-cycles associated with high financial distress, especially if there are also balance sheet vulnerabilities such as high corporate leverage.7
Sectoral shocks. A sectoral shock could be the bursting of a housing market bubble, which would primarily have a direct impact on the construction sector. However, it is also likely to have broader implications for households through wealth effects and the financial sector through solvency effects. A panel regression shows that a dummy variable with the value 1 during crisis accompanied by the initial bursting of a housing bubble can help explain a higher level of unemployment during the recession phase but not during the recovery phase. Another measure of sectoral shocks could be the dispersion in stock market returns across economic sectors, with a high level of dispersion indicating prevalence of sector-specific shocks. Including this measure in the regression confirms that sectoral shocks do tend to amplify the unemployment shock during recessions, but during recoveries, the coefficient on dispersion is not significant.
Unemployment Dynamics During Recessions Not Explained by Changes in Output
Unemployment Dynamics During Recessions Not Explained by Changes in Output
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Financial crisis | 0.73 [0.000] | ||||||
| Financial stress | 0.21 | -0.59 | 0.26 | 0.20 | |||
| [0.045] | [0.019] | [0.019] | [0.085] | ||||
| Financial stress X corporate leverage | 0.03 | ||||||
| [0.003] | |||||||
| Bursting of housing bubble | 0.09 | 0.08 | 0.07 | ||||
| [0.000] | [0.000]*** | [0.002] | |||||
| Stock return dispersion | 0.52 | 1.09 | |||||
| [0.078] | [0.007] | ||||||
| Constant | 0.21 | 0.12 | 0.04 | 0.06 | 0.29 | -0.17 | -0.28 |
| [0.032] | [0.346] | [0.735] | [0.658] | [0.007] | [0.241] | [0.059] | |
| Number of observations | 352 | 259 | 156 | 314 | 334 | 235 | 234 |
| R2 | 0.04 | 0.02 | 0.06 | 0.05 | 0.01 | 0.09 | 0.11 |
Unemployment Dynamics During Recessions Not Explained by Changes in Output
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Financial crisis | 0.73 [0.000] | ||||||
| Financial stress | 0.21 | -0.59 | 0.26 | 0.20 | |||
| [0.045] | [0.019] | [0.019] | [0.085] | ||||
| Financial stress X corporate leverage | 0.03 | ||||||
| [0.003] | |||||||
| Bursting of housing bubble | 0.09 | 0.08 | 0.07 | ||||
| [0.000] | [0.000]*** | [0.002] | |||||
| Stock return dispersion | 0.52 | 1.09 | |||||
| [0.078] | [0.007] | ||||||
| Constant | 0.21 | 0.12 | 0.04 | 0.06 | 0.29 | -0.17 | -0.28 |
| [0.032] | [0.346] | [0.735] | [0.658] | [0.007] | [0.241] | [0.059] | |
| Number of observations | 352 | 259 | 156 | 314 | 334 | 235 | 234 |
| R2 | 0.04 | 0.02 | 0.06 | 0.05 | 0.01 | 0.09 | 0.11 |
Unemployment Dynamics During Recessions Not Explained by Changes in Output
Unemployment Dynamics During Recessions Not Explained by Changes in Output
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Financial crisis | 0.20 | ||||||
| [0.097] | |||||||
| Financial stress | 0.11 | 0.29 | 0.11 | 0.11 | |||
| [0.047] | [0.102] | [0.056] | [0.097] | ||||
| Financial stress and corporate leverage | 0.0 | ||||||
| [0.188] | |||||||
| Bursting of housing bubble | -0.01 | -0.02 | -0.02 | ||||
| [0.384] | [0.144] | [0.173] | |||||
| Stock return dispersion | 0.03 | 0.04 | |||||
| [0.822] | [0.842] | ||||||
| Constant | -0.18 | -0.10 | -0.07 | -0.12 | -0.15 | -0.05 | -0.06 |
| [0.000] | [0.041] | [0.268] | [0.077] | [0.005] | [0.531] | [0.474] | |
| Number of observations | 520 | 389 | 279 | 462 | 467 | 373 | 365 |
| R2 | 0.01 | 0.01 | 0.01 | 0.00 | 0.00 | 0.01 | 0.01 |
Unemployment Dynamics During Recessions Not Explained by Changes in Output
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
| Financial crisis | 0.20 | ||||||
| [0.097] | |||||||
| Financial stress | 0.11 | 0.29 | 0.11 | 0.11 | |||
| [0.047] | [0.102] | [0.056] | [0.097] | ||||
| Financial stress and corporate leverage | 0.0 | ||||||
| [0.188] | |||||||
| Bursting of housing bubble | -0.01 | -0.02 | -0.02 | ||||
| [0.384] | [0.144] | [0.173] | |||||
| Stock return dispersion | 0.03 | 0.04 | |||||
| [0.822] | [0.842] | ||||||
| Constant | -0.18 | -0.10 | -0.07 | -0.12 | -0.15 | -0.05 | -0.06 |
| [0.000] | [0.041] | [0.268] | [0.077] | [0.005] | [0.531] | [0.474] | |
| Number of observations | 520 | 389 | 279 | 462 | 467 | 373 | 365 |
| R2 | 0.01 | 0.01 | 0.01 | 0.00 | 0.00 | 0.01 | 0.01 |
15. These results also help explain why Korea was less hit this time around than during the Asian crisis and why Korea fared relatively better compared with other countries. During the Asian crisis, unlike this time around, the economy was hit by a dual financial and housing market shock. Moreover, the corporate sector was highly leveraged then, while at least the large corporates entered the recent crisis with strong balance sheets. Factoring in the relatively higher level of financial stress during the Asian crisis an the bursting of the housing market bubble can alone explain close to 2 percentage points (or most) of the difference in the change in the unemployment rate during the recession phases of the two crisis. During the current crisis, Korea, unlike the United States, did not face a banking crisis and a housing market collapse. In addition to the relatively lower responsiveness of unemployment to changes in output in Korea’s case, this can explain around 0.75 percentage points of the difference between the change in the United States and Korean unemployment rates during the “Great Recession”.
Explaining Different Unemployment Outcomes
(Percentage point contribution to change in unemployment rate)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Explaining Different Unemployment Outcomes
(Percentage point contribution to change in unemployment rate)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Explaining Different Unemployment Outcomes
(Percentage point contribution to change in unemployment rate)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
The role played by government employment programs
16. During the recent downturn, a number of advanced economies introduced short-term work programs to cushion the impact from the economic slowdown on employment. Countries making use of these programs included France, Germany, Italy, Japan, Korea, and the United States. There was a significant expansion in these programs early on in the crisis in both Germany and Japan, which saw the intake soar to 3½ and close to 4 percent of the labor force, respectively. In the United States and Italy, the increase was less pronounced, which partly reflected design features making it less attractive for employers to use the programs and for workers to participate. In Korea’s case, the government in the 2009 original and supplementary budget introduced various measures to support employment, both through subsidies and temporary public work programs. Aggregating employment in public administration, education, health, and social services, the increase in public employment during the first three quarters of 2009 accounted for close to 2½ percent of the labor force.
Employment Work Program
(In percent of the labor force)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Employment Work Program
(In percent of the labor force)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Employment Work Program
(In percent of the labor force)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
17. These programs can, therefore, also help explain the “unpredicted” change in unemployment during the crisis. The economic slowdown, rising financial stress, and the housing markets busts (in some of the countries) can help explain a significant portion of the increase in actual unemployment from peak to trough during the recent downturn. However, in some countries, including Korea, Germany, and Italy, the predicted change in the unemployment rate by Okun’s law exceeds the actual change and the expansion in the job programs in the latter two countries can explain some of this. In Korea’s case, the public job expansion did not start until after the trough of the crisis in end-2008 and, therefore, only help explain the forecast errors during 2009.
Explaining Change in the Unemployment Rate
(Peak-to-trough and peak to end-2009 change in percentage point)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Explaining Change in the Unemployment Rate
(Peak-to-trough and peak to end-2009 change in percentage point)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Explaining Change in the Unemployment Rate
(Peak-to-trough and peak to end-2009 change in percentage point)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
D. What do the Empirical Results Suggest for the Employment Outlook?
18. Employment is likely to pick up with the rebound in the Korean and global economy, although the recovery may prove protracted. Korea’s economy has rebounded impressively since the recession in the second half of 2008 and is at the front line of the global recovery. This has already been reflected in an improvement in labor market conditions, including a pickup in employment across sectors and professions. However, the economic growth momentum is expected to slow in coming quarters, partly as a “technical payback” for the fast recovery and as macroeconomic stimulus is scaled back.
19. Even so, employment in Korea is predicted to grow faster than in most other advanced economies, but this also reflects faster growth. Scaling the employment growth predicted by the estimated Okun laws with expected GDP growth, shows that the employment gains are less buoyant in Korea. This is in line with the estimated lower dynamic beta in Korea’s employment equation. However, these predictions do not factor in that employers may be somewhat cautious about hiring due to lingering uncertainty about the economic outlook. Also, they do not factor in the potential impact from a gradual scale back in public employment programs in Korea and other advanced economies.
Model Predicted Employment Recoveries
(Employment q/q SAAR divided by GDP q/q SAAR)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Model Predicted Employment Recoveries
(Employment q/q SAAR divided by GDP q/q SAAR)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Model Predicted Employment Recoveries
(Employment q/q SAAR divided by GDP q/q SAAR)
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
20. Absent some of the labor market rigidities caused by institutional factors, employment could have recovered faster. If Korea’s employment protection level had been lower and unemployment benefits had been higher, employment would have declined more during the crisis, but would also have picked up much faster during the recovery. Comparing the through-cycle prediction of employment growth using the dynamic beta estimated for Korea with the employment growth predicted if employment protection had been half its current level and unemployment benefits in line with OECD levels (resulting in a higher beta), suggests that the overall employment level would end up much higher in the latter case. In addition, panel regressions show that annual employment growth in Korea (and other countries), controlling for output growth, could have been 0.15 percentage points higher in the past for each 1 point decline in OECD’s employment protection index and 0.36 percentage point higher for each percentage point increase in the income replacement rate of unemployment benefits. For Korea, this corresponds to additional employment of 350,000 and 850,000, respectively, over a 10 year period. However, the positive coefficient on the income replacement ratio should be interpreted with caution. It may simply pick up the fact that some countries with high income replacement ratios have institutional features supporting labor market flexibility, including a low level of employment protection, decentralized wage-setting, etc. Indeed, if both employment protection and unemployment benefits are included as explanatory variables for employment growth, they turn out to be insignificant.
Model-Predicted Employment
(Growth q/q SAAR (LHS) and cumulative change in persons (RHS))
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Model-Predicted Employment
(Growth q/q SAAR (LHS) and cumulative change in persons (RHS))
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
Model-Predicted Employment
(Growth q/q SAAR (LHS) and cumulative change in persons (RHS))
Citation: IMF Staff Country Reports 2010, 271; 10.5089/9781455204311.002.A003
E. Concluding Remarks
21. Korea’s economy and labor markets were hit by the “Great Recession,” but the job losses were lower than in the past and elsewhere. When the adverse spillovers from the global economic and financial crisis hit the economy during the second half of 2008, employment declined, especially in the more cyclically sensitive sectors and for non-regular workers. However, the job losses were smaller than during the Asian crisis, which partly reflected the smaller output loss but also a larger adjustment this time around in wages and hours worked. The latter also partly explains why job losses were less severe in Korea than in many other countries during the “Great Recession.”
22. Institutional factors also explain the small increase in Korean unemployment during the recent crisis compared to trends in other countries. Korea’s unemployment rate rose by a mere 0.1 percentage point between the pre-crisis cyclical peak in June 2008 and the trough of the crisis in December 2008. In fact, regression analysis reveals that the responsiveness of unemployment to output changes is relatively low in Korea compared to elsewhere, especially when controlling for crisis episodes. Based on a cross-country panel regression using advanced economies, it is found that the low responsiveness in Korea can partly be explained by the high level of employment protection, which is particularly high for regular workers. It also partly relates to the relatively low level of unemployment benefits in Korea. Further to these institutional factors, Korea’s labor markets fared relatively well because the economy entered the crisis with sound fundamentals and was not, unlike other countries, hit by both banking crisis and the bursting of housing market bubbles. Moreover, the Korea government took decisive steps to counter the adverse fallout from the crisis through macroeconomic stimulus and an expansion of employment programs.
23. To avoid a “job-light” recovery, the authorities should manage the unwinding of policy support measures carefully and take further steps to increase labor market flexibility. While the rebound in economic activity in Korea has been impressive so far, macroeconomic and labor market policies have played an important part. These will, therefore, have to be scaled back carefully to safeguard the recovery in the labor market, although it will be important to ensure that the work programs do not become permanent features and restrain private-led employment growth. In this context, it will also be important to step up efforts to reduce the high level of employment protection for regular workers. While it reduces the decline in employment during a crisis, panel regressions suggests that it also slows the subsequent recovery and, more generally, has a negative impact on employment growth. Furthermore, the high level of employment protection has served to amplify dualism in the labor market, which leaves the nonregular workers more exposed during downturns, raising precautionary savings given the limited access to unemployment benefits. It also gives employers less incentives to invest in on-the-job training for the large share of non-regular workers (more than one-third), with likely negative implications for potential growth. Reducing employment protection, especially for regular workers, should go hand in hand with an enhancement in social protection programs to help lessen the income losses during unemployment and smooth consumption. Moreover, to help give nonregular workers a foothold, further efforts are likely needed to strengthen their training and education.
APPENDIX
The description of data sources methodology in this appendix borrows from Chapter 3 of the IMF World Economic Outlook, April 2010.
Data sources
Data Sources
OECD = Organization for Economic Cooperation and Development.
Data Sources
| Descriptor | Source |
|---|---|
| Employment | OECD,1 Labor Force Statistics |
| Labor Force | OECD, Labor Force Statistics |
| Unemployment Rate | OECD, Labor Force Statistics; Haver Analytics |
| Read GDP | GDS (raw data from Haver Analytics) and CEIC for |
| Korea | |
| Employment Protection Legislation | OECD |
| Unemployment Benefits (average replacement ratio | IMF Structural Reform Database |
| for first two years) | |
| Share of Temporary Workers | Eurostat, OECD |
| Hours per Employee | Haver Analytics, National Sources |
| Sectoral Stock Market Returns | Datastream |
OECD = Organization for Economic Cooperation and Development.
Data Sources
| Descriptor | Source |
|---|---|
| Employment | OECD,1 Labor Force Statistics |
| Labor Force | OECD, Labor Force Statistics |
| Unemployment Rate | OECD, Labor Force Statistics; Haver Analytics |
| Read GDP | GDS (raw data from Haver Analytics) and CEIC for |
| Korea | |
| Employment Protection Legislation | OECD |
| Unemployment Benefits (average replacement ratio | IMF Structural Reform Database |
| for first two years) | |
| Share of Temporary Workers | Eurostat, OECD |
| Hours per Employee | Haver Analytics, National Sources |
| Sectoral Stock Market Returns | Datastream |
OECD = Organization for Economic Cooperation and Development.
Business cycles
This paper employs a “classical” approach to dating business cycles by focusing on turning points in the level of output rather than deviations from a trend. The procedure—based on Harding and Pagan (2002)—uses a set of statistical criteria to determine the window over which an observation is classified as a local peak or trough and to determine the minimum duration of a complete cycle and the minimum duration of a phase of a business cycle. In this chapter, the observation window is set at two quarters, the minimum duration at five quarters, and the minimum phase at two quarters. Although the criteria for the minimum duration of a cycle and a phase are occasionally binding, the procedure generally dates the start of a recession as the quarter during which output is higher than the two quarters preceding and following it. This implies that a period of two quarters of negative growth is a sufficient, but not necessary, condition for a recession. Likewise, the end of a recession is generally marked as the quarter during which output is lower than the two quarters before and after it. With these criteria in place, local peaks and troughs are identified, which define recessionary and expansionary phases of the business cycle.
Stock market dispersion
Measure of Stock Market Dispersion The measure of dispersion in stock market returns follows Loungani, Rush, and Tave (1990). Stock market returns at the sectoral level for each country are obtained from Datastream. The data generally begin in the early to mid-1970s. For each country i, the time series of the stock market dispersion measure (SDt) is computed as follows:
where ωnt is the share of total market capitalization of sector n in quarter t, Rnt is the quarterly return on the sector n index, and
Okun’s law and dynamic beta
For each recession episode in a particular country, a dynamic version of Okun’s law is estimated using quarterly data for the 20-year period leading up to the peak in output just before the start of the recession. The general form of the equation that is estimated is as follows:
where Ait and Ay refer, respectively, to the change in the unemployment rate and the level of output growth. DR is a dummy variable that takes on a value of 1 if the economy is in a state of recession. The use of the dummy variable allows the coefficients related to the responsiveness of changes in the unemployment rate to output growth to take on different magnitudes depending on the state of the business cycle. To allow for different dynamics across countries, the lag lengths (p1, p2and q in the specification above) are chosen using a Bayesian information criterion for each country and each episode. For most countries and episodes, the criterion suggests the use of fewer than two lags. The procedure used to estimate the Okun’s law equation for changes in employment is carried out in a similar manner, with the change in log employment as the dependent variable.
To demonstrate how the dynamic beta (β) is derived, we use the example of one lag on output and unemployment. This gives the following expression for Okun’s law:
The dynamic beta (DB) measures the long-term impact of a one-unit change in output growth on the change in unemployment. Based on the specification above, the dynamic beta can be written as follows:
When there is a one-unit change to output growth during period t and no change during other times, the equation reduces to:
The summation in the last term can be written as
Assuming that the there is initially no change in unemployment (i.e. Δut_1= 0), we get:
Rearranging then gives the expression for the dynamic beta:
References
Bernanke, Ben, and Mark Gertler, 1989, “Agency Costs, Net Worth and Business Fluctuations,” American Economic Review, Vol. 79 (March), pp. 14–31.
Boeri, Tito, and Jan van Ours, 2008, The Economics of Imperfect Labor Markets (Princeton, New Jersey: Princeton University Press).
Cardarelli, Roberto, Selim Elekdag, and Subir Lall, 2010, “Financial Stress and Economic Contractions,” Journal of Financial Stability (forthcoming).
Harding, Don, and Adrian Pagan, 2002, “Dissecting the Cycle: A Methodological Investigation,” Journal of Monetary Economics, Vol. 49, No. 2, pp. 365–81.
International Monetary Fund, 2009, World Economic Outlook, April 2009, (Washington D.C.).
International Monetary Fund, 2010, World Economic Outlook, April 2010, (Washington D.C.).
Knotek, Edward S., 2007, “How Useful Is Okun’s Law?” Federal Reserve Bank of Kansas City Economic Review (Fourth Quarter), pp. 73–103.
Loungani, Prakash, Mark Rush, and William Tave, 1990, “Stock Market Dispersion and Unemployment,” Journal of Monetary Economics, Vol. 25, No. 3, pp. 367–88.
Moosa, Imad A., 1997, “A Cross-Country Comparison of Okun’s Coefficient,” Journal of Comparative Economics, Vol. 24, No. 3, pp. 335–56.
Okun, Arthur M., 1962, “Potential GNP: Its Measurement and Significance,” American Statistical Association, Proceedings of the Business and Economics Statistics Section (Alexandria, Virginia: American Statistical Association).
Reinhart, Carmen, and Kenneth Rogoff, 2008, “Is the 2007 U.S. Sub-Prime Crisis So Different? An International Historical Comparison,” NBER Working Paper No. 13761 (Cambridge, Massachusetts: National Bureau of Economic Research).
Sharpe, Steven A., 1994, “Financial Market Imperfections, Firm Leverage, and the Cyclicality of Employment,” American Economic Review, Vol. 84, No. 4, pp. 1060–74.
Prepared by Leif Lybecker Eskesen. This paper is expected to be published in the Korea and World Economy Journal.
The paper follows the methodology used in Chapter 3 of the April 2010 IMF World Economic Outlook, which was authored by Ravi Balakrishnan, Mitali Das, and Prakash Kannan.
The dynamic (J captures the long-term impact of changes in output on changes in the unemployment rate, which essentially corresponds to the impact of lagged changes in unemployment and output on current changes in unemployment.
The countries included are: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Korea, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, and the United States.
As a measure of employment protection, OECD’s employment protection legislation (EPL) index is used and the generosity of unemployment benefits measures the income replacement rates.
This again follows the approach used in Chapter 3 of the April 2010 IMF World Economic Outlook.
The financial stress index used here was developed by Cardarelli, Elekdag, and Lall. See for example, Journal of Financial Stability, 2010 (forthcoming).
