On the application of cointegration analysis in enhanced indexing

We investigate the application of cointegration techniques in designing trading portfolios that outperform a market benchmark. Of particular interest is the situation of enhanced indexation with incomplete portfolios, that is, by imposing a limit on the maximum number of assets included in the portfolio. We present a technique for solving cardinality-constrained portfolio selection problems using cointegration analysis. We investigate the empirical performance of cointegration-based trading strategies in the context of benchmarking portfolios relative to a common stock market index.