Determinants of Promotion and Scientific Productivity: A Study on Italian and French Academic Physicists

Atlanta Conference on Science and Innovation Policy 2009This presentation was part of the session : Science and Innovation WorkforceBoth the sociology and the economics of science have dedicated many research efforts to explaining the determinants of academic scientists' careers and productivity. Most of the available evidence, however, is based upon US data, while cross-country studies are virtually absent; this is especially true of quantitative studies (Long, 1978; Allison and Long, 1990; Levin and Stephan, 1991; Lee and Bozeman, 2005). Such a gap in the literature stands in contrast with the need to explore the many differences between the US academic system and those of other countries, as well as the differences to be found in large regions, such as Europe, whose countries have for long pursued separate national higher education and science policies. The most important differences are related to the degree of autonomy of universities, the relative balance (in size and prestige) of universities vs. other public research organizations, and the size and flexibility of the academic job market. In this paper we study the determinants of promotions and scientific productivity of physicists in French and Italian universities. These two countries are characterized by political and scientific competition between universities and top public research organizations (especially CNRS in France, and CNR in Italy) and by quite a rigid institutional system. Almost all French and Italian universities are public, with limited financial and managerial autonomy; their professors are State employees, whose wages and duties are set according to tight ministerial rules and not subject to local negotiations; promotions and mobility across universities are administered by disciplines, that is by elective bodies that represent all tenured staff in specific scientific fields and exercise a great degree of control over We follow the existing literature in exploring the impact of age, gender, and social capital on the quality and quantity of academic scientists' publications on careers, and of all such variables on the same scientists' career advancements. In addition, we consider some institutional specificities of the French and Italian systems. In particular we observe that in both countries the recruitment process has been characterized by frequent stops-and-goes imposed by political events, which may have introduced noticeable cohort effects, as discussed by Stephan (1996). Both in France and Italy, in fact, the early 1980s saw a sudden peak in recruitment (respectively, in 1984 and 1980), which was meant to re-establish a satisfactory staff-to-student ratio after years of growth in the student population, but which was followed by a long dry spell during which many less new scientists got a tenured position. At the present stage, the paper uses a newly created panel of 3633 academic physicists of the matter in France and Italy active in 2004-05. Productivity is measured through the number of publications on 363 high impact journals and quality is measured using the five-year impact factor of the journals where the scientist's articles are published. We run different regressions for the different academic ranks established by law in the two countries, namely: Ricercatore (RU), Professore Associato (PA) and Professore Ordinario (PO) for Italy and Maitre de conference (MCF) and Professeur (PR) for France. We address the issue of endogeneity of promotion estimating five different models, one for each rank, accounting for the scientific productivity conditional to promotion. Sample selection issues are solved using a traditional Heckman two steps estimation. In our framework the probit selection equation gives information on the determinants of the probability of promotion. We account for individual heterogeneity by considering the scientist's average yearly productivity before promotion (quantity/quality before promotion). Our results show age and gender effects in the different institutional environments, by rank. We characterize the work environment using co-authors' productivity and affiliations and we find evidence of a significant impact of co-authors on productivity. Finally we show that scientists recruited or promoted after the reforms of the academic systems in France and Italy in 1984 and 1980 are on average less productive then scientists belonging to other entry cohorts. However this occurs only for Professori Associati, Ricercatori and Maitre de conference for which the selection criteria were less strict in those two entry cohorts. In the probit selection equations we find that Italian or French scientist's chances of being promoted grow with age but decline partially after scientists are 61. Women, coeteris paribus, have a significantly lower probability of being promoted both in Italy and France. In Italy, promotion is influenced only by the quantity of a scientist's publications while in France both quality and quantity play a statistically significant role. In both academic systems we observe big recruitments waves that affect significantly the probability of promotion. In the productivity equations we show that age has a negative impact on the quantity and quality of articles published, but only in France. In Italy we observe a negative gender effect only for RUs. In Italy therefore women at the early stages of their career are penalized in their publication activity. However if Italian women manage to be promoted to higher ranks they publish as much as Italian men. Conversely in France we observe a negative gender effects across all ranks. We find evidence that the work environment is very important for individual scientists' productivity. In particular having a high quality co-authors and co-authors form the US (or from other research organizations like CNRS, CERN or INFN) has a positive effect on the quality and quantity of published articles. Finally we show that big recruitment waves of the 1980s had a negative effect on the subsequent quality and quantity of publications of professors in the related cohorts. This effect is significantly negative in the case for MCFs in France, and RUs and PAs in Italy. On the contrary, the effect seems positive POs and PRs. We explain this result by recalling that recruitment waves interested only the lower ranks of the academic ladder (RUs and PAs in Italy, MCFs in France). As fot higher ranks (POs and PRs), no special provisions were made, but only above-average scientists managed to be promoted. We are currently working on collecting similar data for CNRS and CNR researchers, so that by the time of the conference we should also be able to compare the productivity and career patterns of these scientists with those of their academic colleagues. This is of great interest for France and Italy, being the two institutions in competition with universities for attracting the nest talents, and the policy-makers' attention and support