Predicting Bank Stock Performance with Fundamental Relative Analysis: Simultaneous Multi-Dimensional Benchmarking as an Investment Tool

Abstract

Fundamental analysis or financial ratio analysis fails to capture the benefits of a simultaneous multi-dimensional benchmarking relative to a company's peers. The paper's main objective is to predict bank stock performance one year ahead by a composite efficiency metric based on an approach we call fundamental relative analysis. For the first time, the paper's approach brings together financial ratios commonly used in banking, generalized data envelopment analysis and simulated annealing to rank 68 Japanese regional banks on stock performance predicted from relative efficiency scores. An application of this ranking in a profitable investment strategy by designating long and short portfolios underscores the potential commercial value of the methodology. The study also makes a number of other methodological contributions such as using the highly flexible range-adjusted measure which captures non-proportional projections more suited to the reality of the business world and super-efficiency scores that allow better discrimination of efficiency scores. When set up as a self-updating system, the approach would be useful to investors and industry watchers, who will be better informed about which ratios to monitor in forecasting stock performance. The methodology demonstrated in the context of Japanese banks is generalizable to any industry in any country.

Keywords: stock performance, fundamental analysis, DEA, Japanese banks

JEL Classification: G21, L25