It can be more efficient to increase farmers' countervailing power than to increase an agency's intervention power through investing in more vehicles, agent training or budget support. Farmer field schools (FFS) (Box 7-4) is an option that warrants further empirical research to determine the conditions under which this may be so and the kinds of policy environment that best enable empowerment strategies to be effective in meeting development and sustainability goals. (Van den Berg and Jiggins, 2007, for a review and assessment of IPM FFS literature).

Multistakeholder processes. A special participatory approach is the facilitation of multistakeholder processes (Leeuwis and Pyburn, 2002; Wals, 2007). Especially in resource dilemmas, where different categories of interdependent stakeholders make competing claims on common pool resources, sustainable solutions cannot come from regulation, technology or market interventions only. The way forward is a facilitated process of negotiation, shared (social) learning, and agreement on concerted action, based on trust, fairness and reciprocity. There is increasing evidence that humans are capable of agreeing on sustainable solutions and of creating institutional conditions that support the implementation of such solutions if drawn into appropriate knowledge processes (e.g., Ostrom et al., 1992; Blackmore et al., 2007). Multistakeholder processes increasingly are important with respect to climate change adaptation, when agreements have to be reached to avoid crisis or when loss of ecosystem services becomes a key cause of poverty.

The Chain-Linked Model. Commercial innovation studies give a central place to the entrepreneur who sees a possibility to capture an opportunity by mobilizing resources, including knowledge (Kline and Rosenberg, 1986). The driver of innovation in these situations typically is the entrepreneur spotting or creating market-related or social organizational opportunity. Policy support to innovation in these cases is provided by helping entrepreneurs to access specialized sources of knowledge, services and skills (Coehoorn, 1994; Crul, 2003). International experience of supporting innovation in small and medium enterprises in nonfarming sectors can be useful in guiding pro-poor agricultural enterprise development.

Strengthening farmer organizations. Investing in people's organizations is a policy option (Toulmin, 2005) with a long history. The experience of the USA and Europe shows that strong farmers' organizations can be a necessary condition for commercially efficient agricultural development (Bigg and Satterthwaite, 2006). An African example is provided by ROPPA in West Africa (Koning and Jongeneel, 2006; ROPPA, 2006). Organizations such as AGRITERRA in The Netherlands attempt to strengthen farmers' organizations in developing countries through training, delegating research funds to farmers' organizations, and building farmers' capacities as effective partners in the negotiation of contracts as well as in research-priority setting. Since farmers' organizations need allies in other sectors or at other levels if they are to become strong and act effectively in collaborative AKST partnerships (Wennink and Heemskerk, 2006) it is a useful policy option to invest in "platforms" (or organized

social arenas) where farmers and researchers can meet on a level playing field. The inclusion of small farmers' representatives on such platforms (as in the PRODUCE foundations in Mexico) may require special effort but may still end up favoring those with sufficient assets to seize commercial opportunity.

      One of the persistent experiences in agricultural development is that, while it can be relatively easy to promote pro-poor endogenous development, collaborative AKST partnerships and the mobilization of indigenous knowledge in pilot projects, the prevailing governance conditions make it difficult to scale up and embed successful pilot experiences in routine institutional behaviors The difficulty in part lies in social realities that position power and opportunity as highly contested zero-sum contests. In 1986, when Java's rice fields were devastated by resurgent waves of brown plant hoppers (BPH) resistant to pesticides destroyed the natural enemies or predators of the BHP, it took considerable time for the government to respond. The problem was a principle called