자동화의 결정 요인과 고용 효과에 관한 연구 : 노사관계의 조절효과를 중심으로 = A Study on the Determinants and Employment Effects of Automation: Focusing on the Moderating Effects of Labor-Management Relation

다국어 초록 (Multilingual Abstract)

This study examines the determinants of automation and its employment effects, with a particular focus on the role of labor-management relations. As digital transformation and smart manufacturing have increasingly reshaped production processes in the manufacturing sector, automation has become an important strategy for improving productivity and competitiveness. However, its implications for employment remain theoretically and empirically contested. This study therefore analyzes whether labor unions and the labor-management relations climate affect the level of automation and whether these labor-management factors moderate the relationship between automation and employment.
Using the 2019, 2021, and 2023 waves of the Workplace Panel Survey （WPS） conducted by the Korea Labor Institute, this study constructs a three-wave panel dataset of manufacturing establishments. The analysis proceeds in two stages. First, the effects of union presence and labor-management relations climate on automation are examined using static panel models, including pooled OLS, fixed-effects, random-effects, and hybrid models. Second, the effects of automation on employment size are analyzed by employment type, including total employees, regular workers, non-regular workers, and indirectly employed workers. For this analysis, both static panel models and dynamic fixed-effects models including lagged employment are estimated.
The main findings are as follows. First, union presence has a statistically significant positive association with automation. Establishments with labor unions show higher levels of automation than non-union establishments, and this result remains consistent across several model specifications. This suggests that labor unions do not necessarily function as obstacles to technological change; rather, under certain conditions, they may serve as institutional actors that facilitate or accommodate automation.
Second, the labor-management relations climate is also positively related to automation. In particular, the hybrid model shows that the between-establishment effect is statistically significant, indicating that unionized establishments with a more cooperative labor-management relations climate tend to have higher levels of automation. However, the within-establishment effect is not statistically significant, suggesting that the effect of labor-management relations climate is more closely associated with accumulated differences in trust and consultation practices across establishments than with short-term changes within the same establishment.
Third, automation does not have a statistically significant direct effect on employment size. Across all employment types, including total employees, regular workers, non-regular workers, and indirectly employed workers, automation shows no significant effect in either static panel models or dynamic fixed-effects models. These findings indicate that, based on the three-wave panel data used in this study, automation cannot be said to consistently increase or decrease employment size, nor is there sufficient evidence that its direct effect differs significantly across employment types.
Fourth, the moderating effect of union presence is partially supported. In the dynamic fixed-effects models, the interaction term between automation and union presence is positive and statistically significant for regular workers. This suggests that the slope of the relationship between automation and regular employment is higher in unionized establishments than in non-union establishments. However, this moderating effect is not found for non-regular workers or indirectly employed workers, indicating that the role of unions may be uneven across employment types. Given that a statistically significant direct effect of automation on employment is not confirmed, this result should be interpreted as limited evidence of a moderating role rather than as a definitive buffering effect against job loss.
Fifth, the moderating effect of labor-management relations climate is not supported. Among unionized establishments, the interaction between automation and labor-management relations climate is not statistically significant for any employment type. This implies that although a cooperative labor-management relations climate is associated with higher levels of automation, it is not identified in this study as a significant factor that changes the relationship between automation and employment size.
Overall, the findings show that labor-management relations play different roles in the adoption of automation and in subsequent employment changes. While labor unions and cooperative labor-management relations are positively associated with automation, automation itself does not show a significant direct effect on employment size. Moreover, the moderating role of unions is found mainly in relation to regular employment, while no significant moderating effect is found for non-regular or indirectly employed workers. These results suggest that the factors associated with the adoption of automation may differ from those shaping employment changes after automation is introduced.

목차 (Table of Contents)

• 제1장 서론 1
• 제1절 연구배경 1
• 제2절 연구목적 및 연구문제 4
• 제3절 연구방법 및 범위 5
• 제4절 연구의의 7
• 제5절 논문구성 8
• 제2장 이론적 배경 11
• 제1절 자동화와 제조업 디지털 전환의 개념적 배경 12
• 1. 자동화와 로봇의 개념 12
• 2. 스마트공장의 개념과 자동화의 관계 14
• 3. 제4차 산업혁명과 제조업 자동화의 확산 17
• 제2절 자동화 결정요인과 노동조합의 역할 21
• 1. 유도된 혁신이론과 노동절약적 기술도입 21
• 2. 노동조합과 자동화 도입 25
• 3. 노사관계분위기와 기술도입의 조직적 조건 29
• 4. 자동화 결정요인 연구 종합 33
• 제3절 자동화의 고용효과와 고용형태별 차이 35
• 1. 기술진보의 고용대체효과와 보완효과 35
• 2. 자동화의 과업 대체와 고용구조 변화 39
• 3. 내부자외부자 구조와 고용형태별 차이 43
• 4. 자동화의 고용효과 연구 종합 46
• 제4절 노사관계의 조절 메커니즘 49
• 1. 자동화의 고용영향과 조절조건으로서 노사관계 49
• 2. 노동조합의 조절 역할: 고용조정 경로에 대한 개입 50
• 3. 노사관계분위기의 조절 역할: 이해대변과 협의의 조정 방식 52
• 4. 고용형태별 조절효과의 차이 53
• 5. 노사관계 조절효과 논의 종합 56
• 제5절 이론적 논의 종합 57
• 제3장 연구방법 60
• 제1절 연구모형 및 가설 60
• 1. 연구모형 60
• 2. 연구가설 63
• 제2절 분석자료와 연구방법 70
• 1. 분석자료 70
• 2. 주요 변수의 정의 72
• 3. 통제변수 77
• 제3절 분석모형 및 분석방법 79
• 1. 자동화 결정요인 분석모형 80
• 2. 자동화의 고용효과 분석모형 82
• 3. 자동화의 고용효과에 대한 노동조합의 조절효과 모형 83
• 4. 자동화의 고용효과에 대한 노사관계분위기의 조절효과 모형 84
• 5. 분석방법 85
• 제4장 실증분석 88
• 제1절 기초분석 88
• 1. 기초통계 88
• 2. 주요 변수 상관관계 91
• 3. 노동조합 유무에 따른 자동화 차이 94
• 제2절 자동화 결정요인 분석 97
• 1. 노동조합 유무가 자동화에 미치는 영향 97
• 2. 노사관계분위기가 자동화에 미치는 영향 101
• 제3절 자동화의 고용효과 분석 103
• 제4절 자동화의 고용효과에 대한 노사관계 조절효과 분석 108
• 1. 노동조합의 조절효과 108
• 2. 노사관계분위기의 조절효과 115
• 제5절 가설검증 결과 요약 및 정리 121
• 제5장 결론 125
• 제1절 연구요약 125
• 제2절 연구결과 시사점 127
• 1. 학술적 시사점 128
• 2. 정책적 시사점 129
• 3. 실무적 시사점 131
• 제3절 연구 한계 및 향후 연구 과제 134
• 참고문헌 137
• ABSTRACT 149