Modeling A Three-Stage Network Data Envelopment Analysis Model - A Case Of Efficiency Analysis For Exchange Traded Funds

In order to explore the efficiency of Taiwan's Exchange Traded Funds, we apply a three-stage network data envelopment analysis model to assessing overall and stage-level performance. These empirical results reveal that good overall performance, generally, may not suggest good allied process performance or good portfolio management performance. In the ETF performance appraisal literature, the focus is predominantly on the fund portfolio with main considerations being risks, returns, and costs. Our performance appraisal approach is beneficial to ETF management because findings are based on models that accommodate multiple measures of performance and in a comprehensive network representation of the overall fund management process. Out of the 87 ETFs considered in our empirical analysis of Taiwan ETFs, we find the top 10 best performers ETFs in terms of overall efficiency, each of these ETFs is either a bond ETF or an ETF that shortens (or bets against) the financials space. The portfolio managers can use the source of ETF's inefficiency revealed by three-stage network DEA to identify which ETFs that they could emulate in order to achieve efficiency in the future.