Résumé en anglais

The crux of New Public Management (NPM) is to use both externalization and private management criteria to evaluate the efficiency of public action and investments (Rouillard, 2003; Hood and Peters, 2004). Thus the efficiency of NPM is presented as depending on the willingness and capacity of public authorities to both delegate to the private sector and to learn management competencies and techniques that are transferred from private business to public administration. In this paper, we argue that there is a contradiction between the two goals of NPM. In effect, if the private sector is so much more efficient and innovative than the public sector and if public authorities are convinced that it is better to contract out, why would the latter bother to learn from the former? And indeed, why would the private sector be willing to properly share its knowledge and experience with public authorities if that helps these authorities use this knowledge to evaluate the activity of the private sector and even help them in becoming future and efficient competitors? Based on empirical research, we rely on a combined neo-structural and neo-institutional framework to show, at least in the case of PPPs, that public authorities do not learn much from private business and that, when they learn, the knowledge transferred is more ideological than technical. We first review the literature on inter-organizational collective learning and on risk identification and allocation in PPPs. We then provide a description of our empirical setting, in particular the system of actors involved in the promotion of PPPs in France between 2008 and 2010, as well as a qualitative analysis, based on in depth interviews, of their respective role in the main steps of the contract negotiation process, including their definitions of risks associated with very long term contracting. These actors include mainly representatives of the Ministry of the Economy, the banking sector, the building industry and its lobbyists, corporate lawyers, and consultants. Third, we present a network study of advice relationships among the 88 key actors in this system of actors. Fourth, we use network analytical tools to test our hypothesis about ‗non learning‘ by public authorities. By looking at who learns from whom in this system, we show that there is a two-step flow in this learning process and that public authorities learn mainly from lobbyists working for the building industry but heavily dependent on banks. The existence of this two step flow confirms our hypothesis about the abovementioned contradiction and about non learning by the public sector, or at least of learning highly ideologically charged knowledge diffused by lobbyists. Finally, we come back to the issue of the consequences of this pattern of learning relationships for actors who do not share the same frame of reference when defining risks, calculating costs and managing debt in contracting between the public and the private sectors. We conclude by questioning the theories that generalize the virtues of NPM based on communication between private sector and public authorities.