The Korean life insurance industry has undergone profound changes, such as the beginning of the variable insurance in July 2001 and the bancassurance enforcement in August 2003. However, little empirical research has analyzed data that includes the bancassurance of life insurance companies operating in Korea. In response to this lack of research, this paper applies DEA (data envelopment analysis) models to measure and decompose their efficiency. We discovered that life insurance companies operating in Korea are a little different in their composition ratio of inputs and outputs, due to the increased variety of distribution channels and new products. We provided efficiency scores, return to scale, and reference frequencies. We also decomposed CCR, BCC, and SBM efficiency into scale efficiency and MIX efficiency. So, we try to investigate whether the sources of inefficiency were caused by the inefficient operation of DMU, disadvantageous conditions, the difference of the composition ratio in inputs and outputs with reference sets, or any combination of the above. Most companies in the sample display had either constant or decreasing returns to scale. The efficiency rankings were less consistent among models and efficient DMUs. In response to this problem, we used the super-efficiency model to rank them and then compared the rankings of the DMUs among the various models. It was also concluded that the availability of panel data, rather than cross-sectional data, would greatly improve the validity of the efficiency estimates.