The purpose of this paper is to present a new strategy of portfolio selection.
After making a comparative survey of different strategies of portfolio selection adopted by portfolio managers in Tunisia, the paper proposes a new strategy, which it calls weighted overreaction strategy. This strategy consists in over‐weighting the stocks having bad performances in the past.
The new proposed strategy turned out to be more performing than size, PER, and overreaction strategies in the Tunisian stock market via a mean equality test. Those who adopt it should create a loser portfolio and should sell it at a later period (12 months) and generate average annual returns of 241.75 percent.
This result deserves generalization to other stock markets. As the Tunisian stock market is marked by its looseness and low capitalization, applying this strategy over similar or more developed market would open the way for research aiming to define other strategies and to select the best one for each market. Indeed, it should investigate investors' behavior, which is certainly not the same in each stock market and outline the specific strategy for each market.
The weighted overreaction strategy generated a considerable gain compared to other portfolios.
The new proposed strategy turned out to be more performing than the other ones.
