This study examines the recent divergence between unemployment rates and labor market conditions in South Korea from the perspective of changes in matching efficiency.

Prior to the COVID-19 pandemic, the unemployment rate in South Korea remained in the mid-to-high 3 percent range, before rising above 4 percent during the pandemic. It then declined sharply and, as of 2025, has stayed in the mid-to-high 2 percent range, a level significantly lower than its pre-pandemic level. These movements in the unemployment rate cannot be explained simply by changes in labor demand. The unemployment rate continued to decline even during periods when growth in labor demand stagnated, and it remained low even after labor demand began to decline.

These movements in labor market indicators suggest an improvement in matching efficiency. This study examines this possibility using actual data and analyzes the determinants of changes in matching efficiency.

This study first estimates a labor market matching function that incorporates heterogeneity in industries and job seekers’ characteristics. Incorporating multiple dimensions of heterogeneity allows us to examine how changes in the composition of the job seeker pool and cross-industry differences in labor market tightness affect changes in matching efficiency. Based on the estimated matching function, the matching efficiency of the South Korean labor market is found to have increased by approximately 14 percent over the analysis period (2015-2025). A substantial portion of this improvement is trend growth driven by advances in matching technology and the expansion of job intermediation services. However, even after accounting for this trend growth, matching efficiency is estimated to have increased by about 5 percent over the same period, largely due to the effects of demographic changes.

To examine this more closely, this study decomposes changes in matching efficiency - after accounting for trend improvements driven by factors such as technological progress - into job seeker characteristic composition effects, job seeker distribution composition effects, and a variance effect. The results indicate that improvements in matching efficiency during the 2015-2020 period were driven primarily by the job seeker distribution composition effect. As population aging increased the share of older job seekers and reduced the share of younger job seekers, job-search activity became increasingly concentrated in relatively efficient industries such as construction, while declining in less efficient sectors such as retail, accommodation, and food services. This reallocation of job-search activity contributed to an improvement in aggregate matching efficiency in the labor market.

After 2020, a reduction in cross-industry disparities in labor market tightness contributed to improvements in aggregate matching efficiency(variance effect). As the number of younger job seekers declined, relatively slack sectors such as retail, accommodation, and food services became tighter. In contrast, relatively tight sectors such as health and social services became slacker as the number of older female job seekers increased.

Improvements in matching efficiency are estimated to have reduced the structural unemployment rate in South Korea by approximately 0.5 percentage points, and to account for about 60 percent of the decline in the unemployment rate since 2015. These findings suggest that recent low unemployment rate is more likely to reflect structural changes rather than a temporary cyclical phenomenon. Accordingly, unemployment is likely to remain at levels substantially below its historical average going forward. Therefore, when assessing labor market conditions based on unemployment indicators, it is important to distinguish between structural and cyclical factors.

Abstract
Preface
Summary (Korean)

Chapter 1. Introduction

Chapter 2. Estimation of Matching Efficiency
　Section 1. Matching Function
　Section 2. Construction of Key Variables
　Section 3. Estimation Results of the Matching Function
　Section 4. Trends in Matching Efficiency

Chapter 3. Estimation of Matching Efficiency Accounting for Heterogeneity
　Section 1. Matching Function Accounting for Heterogeneity
　Section 2. Construction of Key Variables
　Section 3. Estimation Results of the Matching Function
　Section 4. Trends in Matching Efficiency

Chapter 4. Factors Affecting Changes in Matching Efficiency
　Section 1. Decomposition of Matching Efficiency
　Section 2. Composition Effect of the Job Seeker Distribution
　Section 3. Composition Effect of Job Seeker Characteristics
　Section 4. Dispersion Effect
　Section 5. Counterfactual Analysis

Chapter 5. Effects of Improved Matching Efficiency on the Decline in the Unemployment Rate

Chapter 6. Conclusion

References

Appendix