In the last two decades, an intense debate on the influence of financing constraints on the investment decisions of firms has emerged. One point of view (e.g. Fazzari, Hubbard and Petersen, 1988) argues that investment spending of a firm is affected by its level of internal funds. Another point of view (e.g. Kaplan and Zingales, 1997) argues that the investment-cash flow sensitivity cannot be used as a measure of financing constraints. Although, these two positions seem contradictory, this paper proposes a new test aiming at reconciling these two opposite views. This new test emphasizes the crucial role that the classification criteria used plays on the empirical results obtained. Therefore, two main empirical predictions were tested. Firstly, if firms are classified according to the degree of asymmetric information problems that they face in financial markets, a higher sensitivity of investment to cash flow would be expected for firms that face these problems at a higher level. Secondly, if firms are classified according to their financial status, investment by firms in a better financial position would be expected to be more sensitive to cash flow variations. The empirical findings obtained in this paper, using a sample of Portuguese manufacturing firms, did not allow to confirm these empirical predictions. In fact, from the regression results obtained, it was possible to see that the cash flow variable has a major impact for firms facing higher asymmetric information problems in financial markets and for firms in poor financial status. Thus, one can be led to think that these results would favour the predictions of Fazzari, Hubbard and Petersen (1988) instead of the one's proposed by Kaplan and Zingales (1997).