the funds available for recurring and operating costs (Prem-chand, 1993; Eicher, 2001). The rewards of the agricultural staff tend to be misaligned leading to difficulties in keeping the best talent on the one hand, while the indexed salaries of employees without much concern for market wages tend to balloon the overall budget for this purpose. There is also the widely discussed problem of wage erosion, meaning the loss of salary purchasing power, which impacts negatively on commitment and morale of research staff.
          Sometimes allocations of public resources can lead to spending being spread too thinly across commodities, re­gions and research themes. There can also be other inef­ficiencies within public organizations leading to wastage of resources, corruption and poor planning in public-funded research. Public sector scientists can continue with research on commodities (crops, livestock, and natural resources) and technologies even when farmers move out of these areas due to economic reasons. Some studies show that returns to public sector agricultural extension became low due to the multitude of "non-extension" duties, and that extension agents were not the main sources of technical information to farmers (Isinika and Mdoe, 2001).
          During the 1980s, public research models were reformed to become more participatory. This was to make public re­search organizations more responsive to the requirements of farmers, especially those that are poor and live in resource-deprived areas (Kaimowitz, 1993). Only limited successes were achieved through such participatory research models. This could be due to the fact that the structure of public research organizations was not reformed. The channels of priority setting do not correspond to the funding channels, in other words, funding is provided from other sources than those setting research priorities (Hartwich and von Oppen, 2000). Another reason could be that the incentives for indi­vidual researchers were not always adequately oriented to participatory research. These incentives include not merely additional money but also additional facilities to carry out participatory research, but also intangible ones.
            Public research organizations have also responded to the criticisms on their inefficiency by adopting impact as­sessment of their efforts, priority-setting exercises, and also the introduction of operation and management reforms through measures such as decentralization, accountability, transparency and cost recovery among others (Hall et al., 2000). Moreover there have been efforts to give more au­tonomy to research organizations, remove them from civil service regulations and to provide greater flexibility to man­age their physical, financial and human resources (World Bank, 2000). One can see such examples from the industrial­ized world. There has also been decentralization of research and extension systems in developing countries including Uganda, Tanzania, Kenya, Zambia and Ethiopia (Ananda-jayasekeram and Rukuni, 1999). Similar examples of more pluralism in AKST systems were documented for various other African and Latin American countries (Shao, 1996; Byerlee, 1998; Echeverria, 1998; Heemskerk and Wennink, 2005). But the experiences in different countries are mixed. Research practices and administrative and financial proce­dures of national research systems have not witnessed any major changes in a number of countries. On the other hand, reforming compensation in the national agricultural system

of Chile has made the public research sector more attractive to talented agricultural researchers (Venezian and Muchnik, 1994).
         The  resource   allocation  for  public  sector  research, though ideally driven by considerations of social welfare, is determined in reality by the political economy9 , i.e., the struggle between the interests of different societal sections (social groups, regions, growers of specific crops, gender), and also those who dominate decision-making. Evidence from different parts of the world indicates the influence of such political-economy factors in resource allocation for agricultural research. Research and extension spending is linked to the political effectiveness of farm interests (Rose-Ackerman and Evenson, 1985). A study of 37 countries show that structural changes in the economy have impor­tant effects on the political incentives to invest in public ag­ricultural research (Swinnen et al., 2000).
          Thus even when agricultural research provides higher returns or has the potential to reduce poverty; it does not get enough investments in the public allocation of resources. Sometimes ideological considerations lead to high priority being placed on certain crops, thus making investments economically inappropriate. For example, concerns about food security in certain states of India have led to excessive research investments on some crops, and farmer adoption of commercial crops unsuited for the region (Santhakumar and Rajagopalan, 1995; Santhakumar et al., 1995). Gender is an area where political economy influences research invest­ments and outcomes. This manifests itself in certain situa­tions through inadequate investment in research on crops cultivated by women or technologies which would reduce the drudgery of female agricultural workers. In certain other situations, new technologies produced through research lead to the displacement of women workers. An example is access women have to ICT, which may be limited because of their reduced physical access to resources and infrastructure, social and cultural norms, education and skills, and poverty and financial constraints (Hambly Odame et al., 2002).
         Does democracy help in achieving the socially desir­able objectives through AKST investments? There is no straightforward answer evident from the literature (e.g., Diamon and Plattner, 1995) on whether democracy vs. the responsiveness of governments has a higher degree of such achievement. Even in democratic countries the political pro­cess can be captured by narrow interest groups, whose goals do not necessarily aim at overall social welfare. Even if the role of such groups are controlled, democracy is likely to be driven by the preference of the median voter, and there are situations in which the interest of such a voter need not be in tune with the maximization of the overall welfare of the society. Thus, though democracy is valuable by itself, and provides greater opportunity for wider participation in political decision-making, there is no assurance that it will lead to decisions that enhance the welfare of the society as a whole. The lesson for AKST is that democratic governance is not sufficient to ensure effective and efficient investments aimed at achieving larger development goals.
9 The term political economy is used within the framework of "new political economy" (e.g., Bardhan, 1997).