사업체별 재해빈도에 대한 노동조합 효과: 제조업을 대상으로 한 종단분석

Many research studies conducted in the US, UK and Australia have reported that, unlike our anticipation, the relationship between occupational injuries and illnesses and the union presence in the workplace remains positive and statistically significant. This study examines if this is the case in the Korea’s manufacturing sector where collective voice through unions have been relatively stronger than that of any other sectors in the country despite the overall declining union power. Our first analytical attempt using the cross-sectional dataset (the nationally represented Workplace Panel Survey in 2011 and 2013) led us to obtain the same result of the positive union effect on occupational injuries as those found in the previous studies. However, our alternative negative binomial longitudinal model suggests a somewhat different result from the previous one. The longitudinal model enabled us to control the unobserved characteristics of each establishment and, as a result, the statistically meaningful union-nonunion difference in the frequencies of occupational injuries faded away. This finding informed us that the positive coefficient showing the union effect on occupational injuries could be an overestimated one if unobserved firm characteristics are not effectively considered in the analytical model. Therefore, we could argue that it is critical to control organizations’ unobserved heterogeneity when it comes to examining union influences on occupational safety in the workplace.