Assessing the marketing and investment efficiency of Taiwan’s life insurance firms under network structures

• Published in: The Quarterly review of economics and finance 2019-02, Vol.71, p.132-147
• Main Authors: Huang, Tai-Hsin, Lin, Chung-I, Wu, Ruei-Cian
• Format: Article
• Language: English
• Subjects: Copula methods; Disaggregate data; Network SFA; Scale economies; Simultaneous equations; Taiwan’s life insurers; Technical change; Technical efficiency score; The maximum likelihood
• ISSN: 1062-9769; 1878-4259
• DOI: 10.1016/j.qref.2018.07.002

•This paper proposes the network stochastic frontier approach (SFA).•The production process is split into marketing and investment stages.•The model consists of a production and a cost frontiers with two share equations.•A copula-based econometric model is used to identify structural parameters.•Domestic, FHC, and new insurers outperform foreign, non-FHC, and old insurers. This paper proposes the network stochastic frontier approach (SFA) to fill the gap in the efficiency measurement literature, splitting the entire production process of life insurers into two stages: marketing and investment. A salient feature of the method is that it can characterize technologies undertaken by a series of stages without requiring disaggregate data for individual sectors of insurers. In the context of copula methods, the simultaneous equations can be estimated by the maximum likelihood, and the parameter estimates are used to compute measures of the technical efficiency score, technical change, and scale economies in the two production stages. We find that twenty-six of Taiwan’s life insurers have a higher average technical efficiency score in the investment stage than that in the marketing stage. Scale economies and technical advancements prevail in the two production stages over the sample period 2000–2012. Findings also show that domestic, FHC (financial holding company), and new insurers outperform foreign, non-FHC, and old insurers, respectively. The traditional single production stage model neither accurately describes an insurer’s production technology nor correctly evaluates its performance.