gies (e.g., protective equipment) and management practices, (e.g., hygienic measures after spraying, compliance with re­entry intervals, safe storage of equipment and pesticides) were often found unfeasible in tropical climates and under the conditions of poor countries (Cole et al., 2000).
         A good of example of public sector AKST investment to mitigate the negative impact of pesticide use is IPM; IPM technologies are site-specific in that they need to be devel­oped for specific agroecological, socioeconomic and policy conditions. As a result a wide range of examples exist in both developing and industrialized countries. Despite the large amount of investment in IPM, global impact studies are rare. A meta-analysis for the CGIAR in 1999 showed that, although no aggregate ROR could be established, the ROR for IPM was above 30%, and this does not include the significant environmental and health benefits. In the in­dustrialized countries several successful IPM programs have been implemented in selected crops (e.g., Norton, 2005), but successes on the aggregate level remain questionable due to a lack of enabling policy conditions (Waibel et al., 1999).
        Iron, zinc and iodine deficiencies are widespread nu­tritional imbalances (WHO, 2002; FAO, 2004; Hotz and Brown, 2004; UN-SCN, 2004). The adverse health outcomes of micronutrient deficiencies include child and maternal mortality, impaired physical and mental activity, diarrhea, pneumonia, stunting or blindness, among others (Stein et al., 2005). Biofortification research aims to reduce malnutri­tion by breeding essential micronutrients into staple crops. The CGIAR HarvestPlus Challenge Program concentrates on increasing iron, zinc and beta-carotene (provitamin A) content in six staple crops species (rice, wheat, maize, cas­sava, sweet potatoes and beans). In addition, the program supports exploratory research in ten additional crops (Qaim et al., 2006). Most biofortified crops are in R&D phase, except for beta-carotene rich orange fleshed sweet potatoes and Golden Rice (Low et al., 1997; Goto et al., 1999; Ye et al., 2000; Lucca et al., 2001; Murray-Kolb et al., 2002; Drakakai et al., 2005, Ducreux et al., 2005).
         Thus far, only ex-ante economic analyses exist for bio­fortified crops. An evaluation of the potential health benefits of Golden Rice in the Philippines showed that micronutrient deficiencies can lead to significant health costs, which could be  reduced  through   biofortification   (Zimmermann   and Qaim, 2004). In an ex-ante impact assessment using dis­ability adjusted life years (DALYS) approach (Qaim et al., 2006) the estimated Internal Rate of Return (IRR) was very high, ranging 31 to 66% (pessimistic scenario) and 70 to 168% (optimistic scenario). Ex-ante studies on the expected impact of biofortification research under HarvestPlus have been conducted for rice in the Philippines, beans in Brazil and Honduras, sweet potato in Uganda, maize in Kenya and cassava in Nigeria and Brazil; health-cost reductions range from 3 to 38% in the pessimistic scenario and from 11 to 64% in the optimistic scenario depending on crop and loca­tion (Meenakshi et al., 2006).
         To find out if biofortified crops will be adopted by growers on a large scale requires research including ex-post studies building on observable data to verify the preliminary results. Further research is also needed on the bioavailability and micronutrient interactions in the human body. The key conclusion emerging from the available ex-ante studies is

that biofortification could play an important role in achiev­ing nutrient security in particular situations. However, its benefits will depend on the necessary institutional frame­work that can facilitate the effective introduction of these technologies as well as an enabling policy framework. Other impacts of AKST on health, both positive and negative, can be shown with the development of industrial livestock. Livestock products contribute to improved nutri­tion globally and are linked to disease, such as cardiovascu­lar disease, diabetes and certain types of cancer (Walker et al., 2005).

8.2.7 Spillover effects
The wide applicability of research results over a range of ag­ricultural production conditions or environments often cut­ting across geographical and national boundaries are gener­ally referred to as spillover effects. Spillover effects are a combination of four effects: price effects from the increased production caused by reduced costs which are captured in the supply and demand framework (Hesse and McGregor, 2006). Spill-over technology from country "Y" which can be adopted without any research in country "X"; spillover of technology from country "Y" which requires adaptive research before it is applicable in country "X"; and spillover of scientific knowledge which ultimately enhances future re­search in many areas.
         Technological spillovers increase the returns to research and can be spill-ins or spill-outs. Spill-ins take place when a country is adapting a technology developed elsewhere. This reduces the national research costs and shortens the time required for developing and disseminating the finished prod­uct. The gains from spill-ins are important to all research organizations, but are higher in smaller systems. Spill-outs take place when research findings are used by other coun­tries. Spill-outs are important when one is interested in the total benefits occurring to the country where the technology was developed as well as the country where it was adopted. This aspect is critical when performing impact assessment of a regional network (Anandajayasekeram et al., 2007).
          It has been long recognized that AKST spillovers are both prevalent and important (Evenson, 1989; Griliches, 1992). A study that fails to account appropriately for spill-ins will overestimate the benefits from its own research investment.7 Similarly if state to state or nation to nation spillovers are important—as in the case of regional research networks—and the study measures its own benefit at the national level and ignores the "spill-outs", this will underes­timate the ROR. Only 12% of the 292 studies in the sample of one of the aforementioned meta analysis made any allow­ance for technology spillovers; even fewer allowed for inter­national spillovers (Alston et al., 2000a). They also noted that by far the majority of research impact studies that have allowed for international agricultural technology spillovers were commodity specific studies, rather than national aggre­gate studies, and mostly they were studies of crop varietal improvements.
7 Farmer to farmer spill in/outs are also important, not just locally but where they happen through travel, guest worker return etc, but not easy to capture.