Designing an IPM Research Strategy to Benefit Poor Producers and Consumers in Honduras

Agricultural research organizations face pressure to justify investments in agricultural research and need tools to allocate their resources such that the resulting research portfolio has the largest impact across multiple objectives. These tools should use available information so that the allocation exercise is flexible and low cost. This paper describes a simple mechanism for assessing ex-ante impacts of agricultural research on population subgroups. Economic surplus analysis is combinedwith household data to predict the impacts of agricultural research in integrated pestmanagement technologies for eggplants, onions, peppers, and tomatoes onoverall efficiency and poverty reduction in Honduras.

Despite positive economic growth in the recent past, Honduras has exhibited persistent poverty and has some of the worst social indicators in Latin America and the Caribbean (World Bank 2006).Extreme poverty decreased from 61.5% in 1991 to 47.5% in 2006, but much of the decline is attributable to lower urban poverty and not to improved conditions in rural areas (de Hoyos, Bussolo, and Nunez 2008).Changes in rural poverty have generally only come through migration of the rural poor to urban areas.Thus, even as the poverty rate has decreased for those living in urban areas, the rate of extreme poverty among Honduras' rural populations has remained high.Today, about half of the 7.1 million inhabitants live in rural areas yet they comprise 73.7% of the poor and 86.1% of the extreme poor (World Bank 2006).
Of the three main economic sectors, agriculture represents the smallest portion of GDP in Honduras at 12.5%, with services and industry dominating at 60.4 and 27.1%, respectively (Central Intelligence Agency [CIA] Honduras 2010).Nevertheless, poverty is concentrated in agriculture; 81% of the extreme poor, 60% of the moderate poor, and only 35% of the nonpoor are dependent on agriculture (World Bank 2006).Thus, growth in agriculture could be an effective engine of poverty reduction.
Indeed, numerous studies have found that the agricultural sector has the strongest poverty linkages to the economy, providing jobs and raising incomes for the poor (Bravo-Ortega and Lederman 2005; Loayza and Raddatz 2010;DeJanvry and Sadoulet 2000).Ligon and Sadoulet (2007) estimate the effect of economic growth originating in agriculture versus nonagriculture on the poorest decile's expenditures.They found that outside of Latin America growth in the agricultural sector has a much larger positive effect on the poorest decile's expenditures than growth in other sectors.
Unfortunately, studies have also found that conditions in Latin America pose unique obstacles to agriculture-led poverty reduction.Because the extreme poor in Asia and Africa have little land, they gain disproportionately from growth in agricultural productivity through employment creation and lower food prices (Thirtle, Lin, and Piesse 2003).Lower food prices also benefit the urban poor who spend a large percentage of income on food.However, in Latin America, extreme inequality in income and land holdings prevents the poor from benefiting in similar fashion.In Honduras, the introduction of commercial pesticides and nontraditional export crops such as cotton and melon accelerated the redistribution of land, resources, and market opportunities from smallholders to large-scale transnational exporters (Murray 1991).Agricultural growth thus has potential to benefit Latin America's poor, but unless it is focused in specific subsectors, poverty may not automatically fall with increased growth in the sector.
The main engine of sustainable growth in agriculture is technical innovation (Norton, Alwang, and Masters 2010).Technical innovation in agriculture involves new ways of doing things, use of new inputs, and lower costs of production per unit of output.Lower production costs per unit of output can result from increased yields or lower input costs, or both.Agricultural research is necessary for technical innovation, and research involves discovery of new techniques and processes and adaptation of existing techniques for local conditions (Alston, Norton, and Pardey 1995).
Research administrators face pressure to justify investments in agricultural research and consequently want to ensure that such investments are allocated to meet their objectives.Unfortunately, the presence of multiple objectives, such as maximizing total benefits and reducing poverty, requires assessing tradeoffs when making such allocations.Increasingly, research managers are being asked to direct publicly funded research toward the needs of small-scale farmers and the poor (Byerlee 2000, Alwang and Siegel 2003, Moyo et al. 2007).When funding originates from external donors, allocation decisions should reflect donor objectives.For example, the USAID-funded Collaborative Research Support Programs (CRSPs) and particularly the Integrated Pest Management CRSP (IPM CRSP) provide funds for agricultural research in Latin America and the Caribbean, and research funded under the CRSPs is expected to contribute to USAID objectives.These objectives include increased exports, poverty reduction, reduced environmental contamination, and gender equity.
In Honduras, the IPM CRSP funds agricultural research conducted by the Fundacio ´n Honduren ˜a de Investigacio ´n Agrícola (FHIA).In addition to the IPM CRSP, FHIA receives funding from an endowment established by United Brands and USAID, and other donors.FHIA develops new technologies and adapts existing ones to specific agro-climatic conditions through field trials and local feedback received through its extension service.This arrangement fulfills stated goals of FHIA and the IPM CRSP that include reducing pesticide use, increasing farmer income, improving livelihoods of the poor through technology development and dissemination, and empowering women for pest management decisions.
Economic analysis can inform decision makers about trade-offs between benefits of pest management practices and their costs, including health and environmental benefits and risks (Sexton et al. 2007).IPM can be preferable to preventative pesticide applications as it has direct health and environmental benefits that also reduce pest control costs.For example, in the Philippines, an onion IPM package reduced specific insecticide applications by 50%, herbicides by 65%, and fungicides by 25% while producing health and environmental benefits of $150,000 to five villages where used (Cuyno et al. 2001).IPM will be successful where pest management practices are directly linked to farmers' overall objectives of improving yields and income or addressing the sustainability of production systems (Orr 2003).In Southern Benin, biological control was the only environmentally sustainable and cost efficient method to manage invasive water hyacinth.Managing the hyacinth with native varieties of weevil instead of costly herbicides raised incomes $30.5 million through increased fishing and trading of food crops (De Groote et al. 2003).
Organizations such as FHIA need tools to allocate their research resources so that the resulting research has the largest impact across multiple objectives.Such tools should themselves efficiently use available information such as baseline data and data from field trials, so that the allocation exercise will be flexible and relatively low cost.This paper describes a simple tool for assessing ex-ante impacts of agricultural research on population subgroups.The method can be used to adjust research portfolios so that impacts on intended beneficiaries are greatest.The tool is easy to use, makes use of existing data, and can be adapted to a wide variety of circumstances.Economic surplus analysis is combined with household level data to predict the impacts of agricultural research in IPM technologies for eggplants, onions, peppers, and tomatoes on overall efficiency and poverty reduction.
We find that when distributional objectives are incorporated into the analysis, the research portfolio should be designed differently than it should when aggregate impacts are maximized.That is, the tool shows how to design an IPM research portfolio to meet objectives other than overall efficiency.

Methods
Technical innovation resulting from agricultural research can impact poverty through direct, indirect, and induced effects.The direct effects of successful research translate into increased incomes for adopting producers because of lower cost per unit of production.As production increases and the market supply of the good increases, indirect effects such as lower market prices result.Lower market prices raise real incomes for consumers but also lower incomes for nonadopting producers.Indirect effects also include effects on labor markets and increased wages.If technical change causes an expansion of agricultural production, farms may raise wages, hire more labor, and purchase additional intermediate inputs.The poor often supply farm labor and will benefit from job creation and higher wages.Increased purchases of intermediate inputs stimulate production in related industries, leading to higher wages and increased job opportunities in these industries as well.Induced effects occur over the long run as greater productivity and more efficient resource use stimulate further economic growth.As induced effects (greater economy-wide growth because of increased resource efficiency) occur over a longer time horizon than the direct and indirect effects, they can only be estimated using a more complex model.
Economic Surplus Analysis.Economic surplus methods form the basis of measuring the direct and indirect welfare effects and prioritizing agricultural research (Alston, Norton, and Pardey 1995).As research-generated technologies are adopted by farmers, costs of production per unit of output decline, and the supply curve for the commodity shifts rightward (see Fig. 1 for an example for a traded product for which the price does not change and Fig. 2 for a nontraded product (also called a closed economy) for which increased supply lowers the market price).The supply shift (from S 0 to S 1 ) induces changes in producer incomes (producer surplus) and benefits to consumers (consumer surplus).Several parameters determine the overall size of consumer and producer gains.
The proportionate size of the supply shift resulting from adoption of agricultural research (also known as K; see Fig. 1) determines the direct effect-the change in cost per unit of output resulting from technology adoption.The extent and timing of adoption also affects the overall size of the supply shift.The magnitude of the indirect effects depends on market conditions such as equilibrium prices and quantities (P 0 and Q 0 , respectively in Fig. 2), and the percentage change in quantity supplied or demanded given a one percent change in price.These changes are known as elasticities of supply and demand; elasticity differences are reflected as differences in the slopes of the demand curves in Figs. 1 and 2. Labor market linkages and linkages to other sectors of the economy must also be measured.In an ex-ante setting, these calculations are made based on available data for all research programs; the resulting estimates represent forecasted impacts of alternative research programs.
Under the economic surplus approach, the change in total economic surplus (or income change) is derived from the K shift for a specific commodity as ⌬TS ϭ ⌬PS ϩ ⌬CS ϭ PQK(1 ϩ 0.5Z) where ⌬ represents change, TS is total surplus, PS is producer surplus, CS is consumer surplus, P is the initial price, Q is the initial quantity supplied, Z is the price reduction after technology adoption measured as Z ϭ K/( ϩ ), is the absolute value of the elasticity of demand, and is the elasticity of supply (Alston, Norton, and Pardey, 1995, p. 210).The change in consumer surplus is measured as ⌬CS ϭ PQZ (1 ϩ 0.5Z) and the change in producer surplus is measured as ⌬PS ϭ PQ(K-Z)(1 ϩ 0.5Z).When a commodity is marketed in a closed economy, the change in producer surplus does not just include a gain in producer surplus accruing to adopters through lower production costs but also a loss in producer surplus for all producers resulting from the lower market price.The loss in producer surplus is measured as PS loss ϭ PQZ(1 Ϫ 0.5Z), which when added back to ⌬PS ϭ PQ(K-Z)(1 ϩ 0.5Z), gives ⌬PS new ϭ PS loss ϩ ⌬PS, the total surplus change resulting from the rightward shift in supply.
The degree to which research benefits accrue to producers and consumers depends on elasticities of supply and demand, which are related to the tradability of the good.If a product is highly tradable such as eggplants or peppers, the producer captures most of the benefits in the form of producer surplus with few benefits going to domestic consumers (because the absolute value of the elasticity is high under such cases and local market prices are not affected by the outward supply shift; see Fig. 1).However, if the commodity is mainly sold in domestic markets such as tomatoes or onions, a portion of the benefits will be transferred to consumers as consumer surplus by way of lower prices and increased availability with fewer benefits going to producers.Indeed, studies of the poverty-reducing impact of the Green Revolution highlight the importance for poor consumers of lower prices.Even in cases where poor producers were unable to adopt new technologies, poverty declined substantially because poor consumers benefited from lower prices (Binswanger and von Braun, 1991).
Allocation of Surplus to Specific Groups.Forecasted impacts of agricultural technical change at the market level are aggregate: the surplus approach provides estimates of benefits to producers, consumers, participants in labor markets, and others.If we are interested in benefits to subgroups, such as the poor or specific groups of producers or consumers, we need a mechanism to allocate the forecasted change in surplus among the subgroups.To do this, we need to know the proportion of direct and indirect effects that accrue to these different groups.Moyo et al. (2007) present a means of such an allocation: they exploit information for the subgroup of interest (in their case, the poor) on the likelihood of adoption of new technologies, the shares of crop and farm income in total household income, and the shares of consumption expenses devoted to the goods in question.Data requirements affect the ability to apply this method: its application requires a substantial and representative household survey with detailed information on crops, production costs and consumption expenditures on all foods (and quite possibly all goods).
Our method is able to capture the direct effects to producers and some of the indirect effects such as changes in food prices for consumers, yet it is also able to allocate the shares of benefits accruing to poor households or other groups.This 'rough' method has several advantages over previous ex-ante surplus methods by capturing indirect effects ignored by other studies that estimate equilibrium changes in price and quantity (Alwang and Siegel 2003).For those wishing to use this method to guide policy decisions, it enjoys the added benefits of being easily adaptable to target a number of policy objectives beyond poverty reduction and of being widely accessible because of its relative technical simplicity.Benefits to population subgroups may be estimated as ⌬B i ϭ ⌬PS*w PSi ϩ ⌬CS*w CSi , where B are benefits for subgroup i, PS (CS) is producer (consumer) surplus, and w are the shares of each type of outcome for the i th subgroup.Producer surplus accrues to adopting producers (the direct effect), and adopters together with nonadopters lose a portion of producer surplus as market prices decline (an indirect effect).The portion of these changes accruing to the i th group (e.g., the poorest quintile) depends on its likelihood of adoption of the new technology and the share of production for which the ith group is responsible (its share in total value of national production; w Psi ).Consumer surplus accrues to population subgroups according to their share of the value of national consumption.
As will be seen, we divide the population into income quintiles and use the formula ⌬PS*w PSi ϭ A ti *w Qi *(PS/⌺ i [A ti *w Qi ]) to allocate the change in producer surplus (w PSi ) by quintile (i) for producers of the export crops.A ti represents the adoption rate by quintile, w Qi is each quintile's share of total quantity produced, and ⌬PS is the change in producer surplus.For producers of domestic crops, we must also take into account the decrease in total producer surplus for the nonadopters because of the lower market price using a different formula.For domestic producers, the change in producer surplus is calculated as ⌬PS*w PSi ϭ A ti *w Qi *(PS new /⌺ i [A ti *w Qi ]) Ϫ (PS loss *w Qi ).Similarly, we use the formula ⌬CS*w CSi ϭ w Qi *CS where w Qi measures each quintile's share of total quantity consumed to allocate the change in consumer surplus by quintile.The formula is simplified by assuming no change in consumption patterns following the release of the new technologies.Adding the two measures together (⌬PS*w PSi ϩ ⌬CS*w CSi ), we derive an estimate of each quintile's share of the increase in total surplus according to crop and technology.

Data and Empirical Methods
The FHIA Research and the IPM CRSP.Agriculture in Central America is characterized by the dual presence of large-scale export oriented latifundios and smallholders devoted to subsistence production (Hilje et al. 2003).In Honduras, smallholders predominate and large-scale production is mainly limited to banana and some vegetable production.The tradition of the latifundio dates back to Spanish colonial times with massive estates concentrated in the hands of the political elite and was continued in the last century with the coming of large multinational fruit companies awarded large holdings by the Honduran government.Today, nearly two-thirds of households in rural Honduras work farms less than seven hectares (Jansen et al. 2005).Likewise, an IPM CRSP 2007 base-line survey found an average farm size of 6 ha for farmers in the Comayagua Valley.While the Honduran government has carried out numerous land reform measures since the 1950s to effect a more equitable redistribution, the reforms have failed to address the latifundios, preserving the unequal land distribution and hindering the land market's development (Nelson 2003).
FHIA is a private, nonprofit, apolitical research institution charged with conducting agricultural research directed toward traditional and nontraditional crops for export and crop diversification.Although internationally FHIA is known for its research on breeding of dessert and cooking bananas (plantains and Bluggoe types), substantial resources have been allocated for research on smallholder crops as manifested in its horticultural, diversification, cocoa, and agro-forestry programs.
The IPM CRSP, a global collaborative research project, began collaborating with FHIA on vegetables in 2005.Work started with trials on the FHIA experiment station in Comayagua, the main vegetable growing area.A base-line survey was conducted in 2007 to better tailor the research to producer needs.In 2009, the second five-year phase of IPM CRSP in Honduras began, and both FHIA and project investigators desire more information to target the IPM research to benefit poor and disadvantaged producers.
In many ways, the IPM CRSP is designed to address the needs of smallholders; it works directly with local scientists, growers, and grower organizations.IPM itself is defined as "a systems approach to reducing damage caused by pests to an acceptable level without harming the environment.IPM includes pest-resistant crop varieties; biological and physical control methods; environmental modification; biopesticides; and when absolutely necessary, nonresidual, environmentally-friendly and low mammalian-toxic chemical pesticides" (Muniappan 2010).
Tomatoes and onions are produced primarily for domestic consumption while peppers and eggplants are produced for export.Following the Caribbean Basin Initiative in 1984, tomatoes became a major Honduran export until 1987 when the United States banned imports of ripe tomatoes because of infestations of the Mediterranean fruit fly (Imbruce 2006).Tomatoes still comprise an important part of the local diet with 91% of production consumed domestically while the remainder represents 6% of nontraditional agricultural exports (FAOSTAT 2010, Instituto Nacional de Estadística 2008).Traditional agricultural exports in Honduras are bananas, coffee, cacao, and other processed commodities.Nontraditional exports include vegetables and some exotic fruits.Onions represent Ͻ0.3% of nontraditional exports, and are important in the local diet with 97% of production consumed domestically (FAOSTAT 2010, Instituto Nacional de Estadística 2008).
Eggplant is purely for export and plays no role in the local cuisine.Currently, eggplant exports constitute about five percent of the total value of nontraditional exports (Instituto Nacional de Estadística 2008).Peppers, both hot and sweet, also play little role in the local cuisine and have been produced mainly for the export markets.They have shown much promise by growing steadily from 11% of nontraditional exports in 2005 to over 23% in 2008(Instituto Nacional de Estadística 2008).
Each of these crops faces pest pressures.Annual average crop losses because of insects, diseases, and other pests over the preceding 5 years were estimated at 36% for onion, 39% for eggplant, 40% for pepper, and 47% for tomato.These estimates are based on field surveys and expert opinion.White flies are a vector for Begomoviruses and still represent a major pest to tomatoes and peppers as well as, to a lesser extent, eggplants and cucurbits.Together with aphid vectored potyviruses, they are responsible for between 50 and 90% of crop loss in these crops within Honduras (IPM CRSP 2003).Rootknot nematode is the major pest affecting eggplants, and Thrips tabaci (Lindeman) is the major pest affecting onion production.Other major pests for these crops include fungi, bacterial wilt, Trips palmi Karny, spider mites (Tetranychus spp.) and Spodoptera worms.
Scientists at FHIA, with funding from the IPM CRSP, are in the process of developing, testing, and/or adapting crop-specific technologies such as eggplant grafting, backpack sprayers for onions, and cover crops for tomatoes.This research, which is at different stages of completion, will continue in the near future.Other technologies addressing pest problems common to multiple crops such as solarization and biological control agents, are similarly being developed and/or tested and adapted from elsewhere.Solarization involves placing a black plastic cover over a bed before planting; the heat generated by the sun kills weed seeds and other soil-born pests.The effectiveness of solarization depends on the particular pests and the procedure must be evaluated and compared with alternatives on a case-by-case basis.
Since 2004, rootstock grafting in eggplant has been successful in field trials in reducing losses because of root-knot nematode (Meloidogyne spp.).By grafting the introduced eggplants onto the resistant rootstock of the native "Friegaplato" (Solanum torvum Sw.), eggplant lives were extended by 2 months and pesticide use on experimental fields was dramatically reduced.FHIA scientists developed a pressure-regulating valve on a backpack sprayer for delivering pesticides to onions that has made progress in controlling T. tabaci.The motorized backpack sprayer applies pesticides more evenly and reduces pesticide use in field trials.Scientists at FHIA have experimented with cowpea (Vigna unguiculata [L.] Walp.) as a cover crop and green manure for tomatoes and other Solanum species that has proven effective in field trials.Cowpea intercropped with tomatoes provides habitat for predatory insects, suppresses weed growth, controls root-knot nematode, and adds nitrogen to the soil.
Solarization of seedbeds has shown success in dealing with bacterial wilt, fungi, and nematodes in onions, peppers, and tomatoes.Different biological control techniques have been applied for each of the four crops.The minute pirate bug (orius sp.) has been used to control fruit borers, aphids, spodoptera, and thrips in eggplant and onion, while parasitic wasps (Eretmocerus eremicus Rose and Zolnerowich) have been used to control whitefly populations in tomatoes and peppers.Additionally, trichoderma and mycorrhizal fungi have been used to control soilborne diseases in tomatoes and peppers (IPM CRSP 2009, 2010).
Each of these IPM techniques continues to be investigated by IPM CRSP-funded research at FHIA and form the first components of IPM packages that are being developed for dissemination to growers.Basic data on costs and effectiveness are being collected as a part of this research and, because of funding constraints, FHIA decision makers need information on expected ultimate impacts of continued research on these IPM components, as they eventually will on others as well.
Conducting the Economic Surplus Analysis.Scientists and extension agents working with FHIA provided data for parameters used in calculating the ex-ante economic surplus during January 2010.Scientists responsible for developing and testing each research component gave estimates on the probability of successfully completing the research, expected yield changes associated with the new technology, changes in input requirements, and total research costs per year.Their estimates were based upon results from field trials conducted at FHIAs research station in Comayagua.In some cases such as eggplant grafting, solarization, green manure, and the backpack sprayer, results have been uniformly consistent over several growing seasons.In others such as certain biological controls, the estimates are based upon as little as a single growing season because of the extreme success of the controls that reduced pest infestations completely during the trial as well as subsequent growing seasons.In these cases, the scientists say they are hoping to conduct further field trials to validate the results but can only do so if the pests return.
Most vegetable extension in Honduras is conducted by FINTRAC, a private U.S. contractor that has maintained a presence in Honduras since 1999.FHIA has a strong working relationship with FINTRAC.FINTRAC-supported extension agents provided estimates of the number of farmers expected to adopt the technology, the number of years it would take to reach maximum adoption, the number of years the technology would be used until it begins to be dis-adopted, the costs of inputs used in production, market prices, and yield data (Table 1).During interviews, the agents noted that they did not consider differences in sizes of farms when estimating the number of adopters; the estimated adoption rates can be considered to be proportions of total lands using the new technologies.In cases where cost savings are negligible or yield increases are small, FHIA scientists stressed that the economic surplus method misses the long-term benefits of environmental and sustainable practices that affect yields, costs, and productivity beyond the timeframe of this study.
The degree of tradability affects the choice of prices, quantities, and elasticities used for each crop in the analysis.The base prices and quantities for tomatoes and onions are four year averages for 2006 -2009, recorded by FHIA for its growers within the domestic market.Four years averages out unusual years, but is not long enough to have to worry about a trend effect.For eggplants and peppers, the base prices and quantities are 4 year averages for 2005-2008, taken from international trade data recorded by the Honduran Instituto Nacional de Estadística.The elasticities of supply and demand are estimated based on theory provided in Alston, Norton, and Pardey (1995).For each analysis, the elasticity of supply is chosen to be one (unit elastic).The elasticity of one is justified because with vegetable crops there is no long-term investment required until harvesting can begin such as with coffee and other tree crops.However, farmers may have to wait a year before adjusting production choices based on current prices.The elasticity of demand is estimated to be perfectly elastic for eggplants and peppers given that they are export crops traded on the world market in which Honduras has no influence on the price.The elasticity of demand for tomatoes is estimated to be Ϫ0.7, reflecting their importance in the Honduran diet, and is estimated to be Ϫ0.5 for onions, showing them to be a slightly more important component of the Honduran diet.
Using the above information, the estimated change in input costs is subtracted from the yield change and combined with the probability of research success (p) and the adoption rate (A t ) to arrive at the proportionate cost change, K t , using the formula K ))]pA t where t is years (Alston, Norton, and Pardey 1995, p. 380).Our analysis calculates net present values over a 15 year interval, using a 5% discount rate, beginning with the release of each technology.The discount rate reflects the real rate of return on alternative public investments.Thus, we calculate total economic surplus for each technology under investigation and compare them to determine which maximizes total benefits.
Additional analysis is needed to determine the distributional effects for the surplus to evaluate which technology has the greatest impact on the poor.We disaggregate impacts on the basis of which crop they produce or consume by income quintile.
Allocation of Surplus.Three main parameters are needed to allocate the changes in economic surplus among the subgroups of interest: the probability of adoption (Ai) and the shares of producer and consumer surplus (wi).Based on expert opinion from research scientists, extension agents, and select growers, we estimate the proportion of growers which will adopt the new technology according to their knowledge of modern production practices (basic, intermediate, or advanced) with basic producers comprising the first and second income quintiles, intermediate comprising the third and fourth income quintiles, and advanced comprising the fifth income quintile.We combine producer surplus with national household production data from the September 2007 Encuesta Permanente de Hogares de Propo ´sitos Mu ´ltiples (EPHPM) and the 2007 Honduras' IPM CRSP baseline survey to derive each quintile's production share.The EPHPM is Honduras' nationwide multi-purpose household survey consisting of 21,490 households including data on income, labor, and agricultural production across 16 of the country's 18 departments.The IPM CRSP Honduras baseline survey includes a stratified sample of 151 households from within the eggplant-growing region of the Comayagua Valley.The IPM CRSP survey was combined with EPHPM as the latter does not include information on eggplant producers whereas the former does.We further identify the proportion of income received from the target commodity, and identify what proportion of each subgroup is below the poverty line.This gives an estimate of the size of each technology's potential impact on the poor and allows for comparison across research programs.
Using national household consumption data from the 2004 Encuesta para el Mejoramiento de las Condiciones de Vida de los Hogares (ENCOVI), we are able disaggregate consumers of each commodity by the quantity consumed by each income quintile.ENCOVI is a nationwide survey of 8,064 households which quantifies the socioeconomic conditions of Hondurans in all eighteen departments including data on income and consumption.
Changes in employment for each subgroup can be estimated using production and labor market data.Staple crops such as maize and beans are much less labor intensive than are eggplants, tomatoes, onions, and peppers.Expanding production for the latter crops would create employment.For example, a single maize crop requires Ϸ60 person days per hectare that would cost 6,000 Lempiras ($317.40USD, one Lempira ϭ 0.0529 USD).In contrast, eggplant requires almost 250 person days, tomatoes and peppers require roughly 300 person days, and onion requires close to 350 person days per hectare with respective labor payments ranging from 26,000 -37,000 Lempiras ($1,375.40 to $1,957.30USD).A medium-run estimate (5-10 year out) of each technology's indirect effect on labor could be estimated through a sensitivity analysis based on the area under production changing as a consequence of greater profitability.This approach ignores labor effects including the immediate short-run changes in labor use resulting from adoption of the new technologies and the long-run induced effects occurring if the area under production changes as a result of greater efficiency.More detailed survey data on farm level decisions regarding own versus hired labor and individual labor market decisions are needed for the short to medium-run whereas an ex-post analysis is best suited for the long run.

Results
Research targeting these four crops would have a strong direct impact on poverty reduction based on the importance of these crops to household income.On average, onion and tomato producers receive Ϸ60% of their income from agriculture whereas pepper producers typically receive 53% (Table 2).In contrast, the average Honduran farmer only receives 31% of total income from agricultural production.Of their agricultural income, onion producers rely on onion production for Ϸ63% of total agricultural income, pepper producers rely on peppers for 55%, and tomato producers rely on tomatoes for 49%.Corresponding data for eggplant producers were not available in the 2007 Honduras' IPM CRSP survey.However, it did show all of the respondents to rely on agriculture for their livelihoods with close to half of eggplant producers relying solely on eggplant production for their agricultural income.Furthermore, roughly 89% of eggplant, 67% of pepper and onion, and 60% of tomato producers are below the poverty line (the poverty line is defined by the EPHPM survey as living on a daily per capita income of Ͻ34.62 Lempiras in rural areas and Ͻ67.58 Lempiras in urban areas) as opposed to 78% of farmers overall.This distinction effectively places all households within the bottom two income quintiles below the poverty line and close to two-thirds of households in the middle quintile below the poverty line.Therefore, any research program that distributes benefits to the bottom three quintiles will have the greatest impact on poverty reduction.
Economic surplus analysis reveals that aggregate benefits outweigh costs for each technology.The gross benefits accrue to producers in the case of eggplants and peppers and both producers and consumers in the case of tomatoes and onions while IPM CRSP pays the research costs.The net present value is $2.99 million for eggplant grafting, $1.44 million for the pressure regulating valve in onions, and $0.47 million for cowpea green manure in tomatoes.The net present value for biological controllers varies on the low end from $0.02 and $0.17 million in onions and eggplants, respectively, to $2.65 and $4.01 million in peppers and tomatoes, respectively, on the high end.The net present value for soil solarization ranges from $0.51 million in onions up to $2.12 and $3.18 million in peppers and tomatoes, respectively.
Overall net benefits can be evaluated on a crop or a technology basis (Table 3).IPM research on tomatoes results in the largest total economic surplus gain at $7.66 million.It is followed by $4.77 million for research on pepper, $3.16 million for eggplant, and $1.97 million for onions.Per technology, IPM research on biological control agents produces the largest economic surplus gain at $6.85 million followed by $5.81 million for soil solarization, $2.99 million for grafting, $1.44 million for the pressure regulating valve, and $0.47 million for cowpea green manure.
Survey data are required to disaggregate producer and consumer surplus by subgroup.Approximately 74% of all onion, 66% of all pepper, 67% of all tomato, and 94% of all eggplant producers are in the bottom three income quintiles (Table 2).However, producers from these bottom three income quintiles produce 55% of all onions, 5% of peppers, 20% of tomatoes, and 99% of eggplants.The most extreme difference is that the top income quintile produces 94% of all pepper yet represents only 11% of producers.These producers are large agro-exporting firms growing peppers in highly controlled environments using advanced production technologies and are atypical of the average Honduran farmer.
From the consumer side, there is an even a more pronounced difference between number of consumers by quintile and amount consumed (Table 2).Roughly 65% of onion consumers and 62% of tomato consumers are from the bottom three quintiles yet they each consume Ͻ0.2% of total consumption for either crop.This is in contrast to the staple crops of corn and beans where the bottom three quintiles comprise 86% and 66%, respectively, of total consumers yet consume 85% and 48%, respectively, of total production.Data indicate that there is zero eggplant consumption and very little pepper consumption given that these are export oriented crops.In the eco-nomic surplus analysis, this is reflected as zero consumer surplus change for these crops.
Producer and consumer surplus is disaggregated by income quintile using estimated adoption rates and survey data.Extension agents provided estimates on adoption rates according to whether the producer used basic technology (subsistence farming), intermediate technology (some incorporation of advanced techniques such as IPM practices or hybrid seeds), or advanced technology (large-scale commercial farms).For example, biological control agents are more likely to be adopted by producers using advanced technologies because of the higher cost associated with upfront purchases and their greater understanding of and faith in advanced technologies (Bentley 1989).Solarization of seedbeds is more likely to be adopted by producers practicing basic levels of technology because of its higher labor inputs and because most advanced producers would be producing in controlled greenhouse type environments anyway.Some technologies such as eggplant grafting are adopted evenly regardless of producer level because of limited substitute technologies.For example, the 70% of total producers adopting solarization for onions include 72% of basic producers, 72% of intermediate producers, and 57% of the advanced producers.
Using the guidelines by which the first two quintiles belong to the basic level, the third and fourth quintiles to the intermediate level, and the fifth quintile to the advanced level, we use the formulas presented above to allocate the change in producer surplus (w PSi ) by quintile (i) for producers of the export crops.For producers of domestic crops, we also take into account the decrease in total producer surplus for the nonadopters because of the lower market price using a different formula (see above).We derive an estimate of each quintile's share of the increase in total surplus according to crop and technology (Table 4).Based on the total surplus generated by each research program, eggplant has the largest poverty reducing impact with almost $5.12 million going toward the first three income quintiles.Roughly $4.76 million of this comes from grafting alone.IPM research in tomato has the second largest impact on the bottom three quintiles at $0.97 million, followed by onion at $0.71 million, and finally pepper at $0.30 million.Behind grafting, solarization provides the most benefits to the first three quintiles with $1.02 million, followed by biological controllers with $0.52, pressure regulating valves with $0.49, and cowpea green manure with $0.08 million.
On a benefit-cost basis, the results are similar (Table 5).Grafting in eggplant provides the most benefits for the bottom three quintiles at $234.10 for every research dollar spent from a total of $236.50.In contrast, biological controllers in tomato provide the most benefits per research dollar at $720.50.However, only $8.90 reaches the bottom three quintiles.Behind eggplant, the ordering is the same as above for each research program with tomato, onion, and pepper, respectively, generating the most benefits for the poor out of every research dollar spent.Similarly, grafting provides the most overall benefits per research dollar followed by solarization, biological controllers, pressure regulating valves, and cowpea green manure.

Conclusions
Methods such as this one for guiding research decisions and for priority setting in agricultural development can assist research managers in helping alleviate poverty.Use of commonly available household survey data to disaggregate total economic surplus among individual producers and consumers makes this method particularly powerful beyond normal ex-ante studies in evaluating policy tradeoffs such as efficiency and equity.Furthermore, this approach can be easily tailored to assess impacts on any subgroup of interest, such as by geographic region, level of education, gender, or other identifying characteristics.This information broadens knowledge typically available from base-line surveys and expands the focus beyond a particular region.
We chose to assign arbitrary weights to compare different commodity oriented research projects in terms of the trade-off between overall efficiency and poverty reduction.On a per crop basis, the surplus analysis reveals that IPM research achieves the greatest efficiency with tomato followed by pepper, eggplant, and onion.However, when considering the poverty reducing impact of each crop, eggplant research has the greatest impact followed by tomato, onion, and pepper.Similarly, there are trade-offs between efficiency and equity by technology.IPM research on biological control produces the largest economic surplus followed by soil solarization, eggplant grafting, pressure regulating valves, and cowpea green manure.Eggplant grafting has the greatest poverty reducing effect followed by soil solarization, biological control, pressure regulating valves, and cowpea green manure.
An extension could incorporate more detailed labor market data.Even though owners and employees of large agro-exporting firms, such as those producing peppers, tend to be concentrated in the top income quintile, employment at these firms is substantial.This could lead to an underestimate of the benefits distributed to the lower income quintiles by our current model.As mentioned previously, benefits accruing to labor from eggplant, onion, pepper, and tomato far exceed those from traditional crops.

Fig. 1 .
Fig. 1.Research benefits in a small open economy.

Fig. 2 .
Fig. 2. Research benefits in a small closed economy.

Table 1 . Market and research parameters affecting economic surplus analysis
Scientists and extension agents working for FHIA and FINTRAC.Research costs reflect the operating costs for the experiments and represent the portion of total research costs paid by the USAID-funded IPM CRSP project.