Recognizing the strategic relevance of higher education, various countries have established ambitious objectives for the percentage of graduates in their population by 2020. In some cases, without effective policies aimed at increasing domestic rates of participation in higher education, such goals will not be met. However, successful measures are difficult to design and to implement when the phenomena they target are complex and encompass many personal, cultural, social and economic variables. Such is the case of aggregate demand for higher education. Therefore, a precise knowledge of its more likely influences is indispensable to perform the forecasting exercises required to inform the process of policy design and, subsequently, to monitor its success. In this paper, we employ the partial least-squares regression, which allows modelling with many variables even with relatively few observations, to identify the most relevant determinants of aggregate demand for higher education in Portugal. Our choice of country is dictated by the availability of adequate data but our results, suggesting that the most important factors are policy dependent, and may thus be managed to support strategic educational objectives, may be useful in other national contexts.