New methods for ordering multivariate data: an application to the performance of investment funds (Q2711722)

Usually, the performance of investment funds is analysed through linear econometric models. Recently, however, it has been recognized that linear models may be inadequate. At present, the widespread belief is that the market efficiency hypothesis must be buried. Attention has therefore been paid to non-normality or non-linearity and a large number of techniques have been suggested to cope with long-tailed distributions. However, models which use conditional heteroskedasticity have a good explanatory performance in-sample but a disappointing one for forecasting. Moreover, on the part of the operating units of financial companies, there is a lack of interest in statistical properties and asymptotic results.NEWLINENEWLINENEWLINEIn this paper, a simple and appealing approach is proposed to the evaluation of the performance of investment funds which is based on the construction of the robust bivariate ~box plot for each pair of variables. The suggested approach is easy to handle and can be conveniently used by operating units. The method is applied first to the original variables and then to principal components and canonical variates. An analysis of longitudinal data is also considered.