Participation in Agricultural Markets among the Poor and Marginalized: Analysis of Factors Influencing Participation and Impacts on Income and Poverty in Kenya

Abstract

Poverty in Africa has been found to be predominantly a rural phenomenon. About 75 percent of the world’s poor are believed to work and live in rural areas, and it is estimated that, by the year 2020, 60 percent of the poor will still be rural. Among the worst hit in these rural communities are women and those in marginal agricultural production areas. Meeting the challenge of reducing poverty and improving rural incomes in Africa, specially for these marginalized groups, will require some form of transformation out of the semi-subsistence production systems that currently characterize much of rural Africa to a more commercialized agriculture. Increased market participation by the poor has been found to be important as a means of breaking from the traditional semi-subsistence farming. It has been argued that market-oriented production can achieve welfare gains through specialization and comparative advantage, economies of scale and regular interaction and exchange of ideas. Unfortunately, the most vulnerable who need this kind of welfare boost may be constrained by several factors in their quest to participate in the market for their goods and services. This study aimed at assessing the extent of market participation among smallholder farmers in Kenya with a view to identifying potential market opportunities and constraints for the vulnerable and marginalized groups. The study is within the broader project by Tegemeo Institute of Egerton University, Makerere University and the World Agroforestry Centre (ICRAF) looking at market participation among marginalized groups in Kenya and Uganda. In this study, we use a three-year panel data set collected in 2000, 2004 and 2007 and across various agro-ecological zones of Kenya under the Tegemeo Agricultural Policy Research and Analysis (TAPRA) project. The study assessed the potential output market opportunities for different marginalized rural populations in Kenya and identified factors that could enhance their participation in both output and input markets. The study was carried out mainly through desk top analysis of the existing data set with some complementary secondary data. Additional information was received through a few key informant and stakeholder consultations. The data analysis mainly focused on the characteristics of these marginalized groups and their participation in different input and output markets. Critical questions under the study related to the type of markets that these marginalized groups access, their important enterprises and degree of commercial orientation, and market opportunities that could help them increase their incomes and transition out of poverty, among others. The study also ii looked at how access to financial services, agricultural information and participation in farmer groups impact on market participation. In addition, the study explored factors that affect participation in different markets for the marginalized groups. The marginal/social groups of interest include women, the poor (both income and land poor) and those in marginal agricultural production areas. For women, the study considers female headed households, who form about a quarter of the sample. Income poor households are those below the poverty line as generated by the Kenya National Bureau of Statistics (KNBS). Land poor households are those found to own less than the average per capita land size in Kenya of 0.2 hectares (0.49 acres) as indicated in the World Development Report (WDR, 2008). Information on land was not available for 2000 and is thus reported for 2004 and 2007 only. Marginal agricultural production areas are defined based on a binary variable representing low and high agricultural potential areas. The low potential/marginal agricultural areas comprise of the lowlands and other borderline regions while the high potentialagricultural areas comprise of the highlands and the high potential maize zone.Descriptive results show that female headed households consisted of about 12% of thesample in 2000 but increased to about a quarter in 2007 (over 80% of these females are widows). On the other hand, the proportion of income poor households decreased from 42% to 37% between 2000 and 2007, a scenario that is well consistent with the general reduction in the national poverty figures reported across that period. The share of land poor households remained stable at about 37% between 2004 and 2007. The proportion of the sample living in the marginal agricultural areas was 32%. Relationships among these groups reveal that at least half (and a third) of female headed households are income poor (land poor) and that there is a positive correlation between land and income poverty. As expected, we observe higher proportions of the marginalized groups in the low potential areas implying that if households are marginalized in one aspect, they are also likely to be marginalized in other aspects. In terms of their socio-economic characteristics, female headed households are less educated with over 40% of the heads having no formal education and only less than 10% having attained a secondary education. A slightly better scenario emerges for income poor households, with the proportion of heads with no formal education declining to about 26% and those with secondary education increasing to about 15%. In general, land poor households and those in the low potential areas had a similar distribution in education to the iii general population or average household. However, the proportion of heads with no formal education in marginal areas remained higher despite a slight decline across the years. The largest deviation from the average household was mainly observed with female heads and thus revealing the clear disadvantage that women/female heads have with respect to education. Female household heads were found to be slightly older than the average head and had smaller family sizes, while income poor households had higher than average family sizes. As expected, land poor households were headed by slightly younger persons. In the low potential areas, both the age of head and family size were similar to those for the average household. Overall, the number of persons per household in the sample declined from seven to six between 2000 and 2007. Compared to the average household, female headed households have significantly lower household incomes, lower asset values, and less land. Equally disadvantaged are land poor households with significantly lower incomes than the average household. These land poor households have an average of 0.3 acres per capita land size compared to about 1.2 acres for the average household. Households in the low potential areas, have lower income levels than the average household but higher than for the other marginalized groups. In general, however, incomes, asset values and land size of all marginalized groups are below those of the average household, and are, therefore, disadvantaged in these respects in the general population. In terms of income distribution, there is a significantly higher proportion of female headed households in the bottom lowest income quintiles compared to their male counterparts. This proportion however declined between 2000 and 2007. Conversely, there is a significantly lower proportion of female headed households in the top income quintiles and this increases slightly over time. Again, this trend across years is consistent with the general poverty improvement in the country between 2000 and 2007. Female-headed households compared to their male ounterparts have relatively higher shares of crop income in total household income, which are generally stable across years. Low income compared to high income households also have relatively higher shares of income from crops in total household income which seem to decline over time. They also have relatively higher shares of income from informal labour and business activities. High income households have relatively higher income shares from salary and remittance which also generally increase over time. Despite having smaller pieces of land, the relatively land poor iv households have a sizable share of income from crops and relatively higher share of income from informal labour and business activities compared to the households with larger land sizes. The share of income from livestock enterprises is much larger on larger farms. Crops and livestock account for a higher share of household income in the high potential than in the low potential agricultural areas. The reverse is true for the share of off-farm activities (business and informal labour activities and salaries and remittance). This may be an indication of greater off-farm diversification by households in the low potential areas as a response to the attendant low agricultural performance. In general, however, whether out of choice or a survival strategy, these results indicate that farming is an important livelihood source for the marginalized households and interventions that promote these households’ commercial orientation to farming could result into huge payoffs. On information sources, results of this study indicate that a similar proportion of both male and female headed households accessed agricultural information from public extension, perhaps an indication of lack of any gender bias in the provision of public extension. A slightly lower proportion of low income households as compared to their high income counterparts receive agricultural information from public extension though the trend is not clearly conclusive. Also, public extension agents are the most important sources of information in both the low and high potential agricultural areas. Further analysis on this issue would be critical in informing the debate on privatization of public extension service and the implications of the demand-driven extension. Generally though, more female headed and income poor households and households in low agricultural potential areas receive agricultural information through private agents/NGO’s and neighbours and friends, the latter clearly indicating the important role of social networks among these marginalized groups. In terms of access to financial services, the study results reveal only very minor difference between the proportion of households that sought credit across gender of head, and hardly any difference in success rate among those who sought. This result remains largely inconclusive and thus begs for a deeper and more focused study given the low proportions of households that sought credit in this sample and also the general nature of the questions posed during the survey. A clearer picture is observed with respect to income groups and agricultural potential whereby we observe a significantly smaller proportion of low income households and households in the low potential areas seeking credit as compared to their high income counterparts and those in the high potential areas. Generally though, across gender, v income groups and agricultural potential, the results indicate that the proportion of credit seekers has been declining over time, a result that could emanate from both the supply or demand side issues. This could indicate a need to come up with favourable credit packages that are both affordable and suitable to small scale farmers. Unfortunately though, close to half of the credit amount received was used to meet household needs and another quarter for education purposes with insignificant amounts invested in farming. The role of collective action in assisting small scale farmers through economies of scale, bargaining power and information haring/learning cannot be overemphasized. These attributes can enhance market participation and increase income of participants. This study indicates very high participation rates in these groups by the sample households; over 70% of all households belong to groups and this proportion increases with income. The causal relationship between social capital, households’ income and other factors, however, remains a question for further inquiry in this case. On participation in input markets, fertilizer adoption in the entire sample is generally high and has been increasing over time. Intensity of use has however been declining from an average of 72kg/acre in 2000 to 63kg/acre in 2007. Although a similar trend is maintained across gender and income, there are significantly lower fertilizer adoption rates and intensity of use among female headed households and low income households compared to their respective male and high income counterparts. The reduction in intensity of fertilizer use over time amidst growing adoption rates could be an indication of increasing awareness in the importance of using fertilizer but within a very constrained budget. This scenario clearly limits households’ ability to adequately gain from productivity levels possible with expanded fertilizer use. As expected, the disparity in fertilizer use between low and high potential agricultural areas is perhaps the largest with about 94% of households in the high potential areas using fertilizer compared to less than a half (38%) in the low potential reas in 2007. Generally, adoption of improved maize variety is relatively high across all households (over 70% of all households). While a higher proportion of male than female headed households planted improved maize varieties, female headed households had a larger proportion of their land under maize with improved varieties than their male counterparts. Adoption of improved maize varieties generally increased with increase in income and land size, suggesting the influence of income and land on household’s adoption decision for improved seed varieties. As with fertilizer, the disparity in use of improved maize variety between low and high vi potential agricultural areas is again the largest with about twice as many households in the high potential areas using improved maize seed compared to the low potential areas in 2007. Similar trends as described above are observed when considering all other improved seed varieties used by these households. There is generally high level of crop diversification among all households with marginalized groups having similar crop portfolios as with their counterparts. The results, however, indicate that the poor and female headed households generally produce lower volumes for most enterprises. In terms of the crops that are most important for commercialization, traditional cash crops e.g. tea coffee and sugarcane lead, though vegetables and fruits are also substantially marketed. Thus, market orientation for these higher value crops is greater than for cereals, roots and tubers among these households. Although both male and female households sell various crops, a significantly higher proportion of male than female headed households sold maize, vegetables, roots/tubers and cash crops other than tea, coffee and sugar cane. Female headed households always sell smaller amounts, with other cereals (sorghum, millet etc) and bananas being more important as measured by their share in total value of production. On livestock, female headed households not only generally lag behind their male counterparts in keeping improved cattle, but they also keep small herd sizes. As with female headed households, the poor are generally less market oriented. Maize is very important among the poor, accounting for over 42% of total value of crop production in 2007, closely followed by vegetables. Yet, even though maize is important, most is subsistent as the marketed volume is barely 10%. Income poor households lead in keeping indigenous cattle and lag behind in keeping improved cattle. Generally, market participation among the poor is low for all livestock, but relatively greater participation is observed for sheep, goats and chicken. Milk production and sales is also generally lower for the poor. Market orientation is less in the low than in the high potential areas. With the exception of fruits, volumes sold for most of the crop enterprises are higher in the high than in the low potential areas. In the low potential areas, market orientation was highest for the cash crops, bananas, fruits and vegetables while in the high potential areas cash crops, bananas, fruits, vegetables, maize and roots and tubers had the highest market orientation. On livestock, market participation among the households in low potential areas is low for all livestock, but vii relatively greater participation is observed for cattle. Participation in marketing small ruminants and chicken is low but increased between 2000 and 2007. Milk production and sales is lower for households in the low than in the high potential areas. Overall, market participation is dominated by 20% of the households, who sell well over two thirds of the marketed volume for various crops. There seems to be a strong link between income and the share of revenues from sales of tea, coffee and sugar cane, indicating that these crops have helped high income households out of poverty. Fruits and maize appear to be more proincome poor while vegetables, bananas, fodder and maize show higher levels of being proland poor. With regards to livestock marketing, results have shown an increasing importance of small ruminants (sheep and goats) and chicken among women and the poor. Dairy production (in terms of the percentage of households) is high among the households, but the income poor and land poor households hardly produce or sell. Results from the econometric analysis of determinants of input market participation show that all the marginal groups are less likely to use fertilizer than their respective counterparts. We find that the degree of commercialization of crop enterprises by female headed households is significantly and positively affected by land size, membership in farmer groups and access to credit. The degree of commercialization by the income poor households is significantly and positively influenced by membership in farmer groups, access to credit, ownership of transport and communication equipment, while market participation by the households that exited poverty is positively influenced by group membership and ownership of communication equipment. Access to credit and group membership also affects market participation among the land poor households and those in the low agricultural potential areas. Therefore, credit and collective action appear to be attractive investment areas for efforts aimed at increasing market participation by all these marginalized groups. On market participation and transitioning out of poverty, the results of this study show that households that exited poverty registered significant increases in both the proportion of households selling and the proportion of marketed crop and milk production. The increase in the proportion of marketed production was particularly large for fruits, bananas, other cash crops (such as pyrethrum, tobacco, cashew nuts, coconuts, etc), roots and tubers and vegetables. Increased market participation by poverty exiters is an indication of the importance of linking these smallholder farmers to markets for income growth and poverty viii reduction. These results show a strong correlation between market participation and exiting poverty. In conclusion, the study finds that the identified social groups are clearly disadvantaged in many aspects than the average household. Female headed households are locked in some kind of marginalization with over half of them being income poor, and over 90% having at most a primary education. Income poor households are also equally marginalized with over 50% of them also being land poor and about 80% having at most a primary education. Both female headed households and low income households are characterized by high crop income shares yet with low incomes and thus implying some locking up in low productivity–low return farming activities with limited alternatives outside of agriculture given the low income shares from business of salaried wage activities. Further, these marginalized groups lag behind their respective counterparts in use of productivity enhancing inputs and thus are limited in their ability to produce sufficient marketable surplus. Although crop diversification is also high for these marginalized groups, they are, however, less market oriented and occupy a very small portion of the market compared to their respective counterparts. Given the results that positively link increased market participation with transitioning out of poverty, efforts to link these marginalized groups to markets should be seen within the broader goal of poverty reduction.Regarding the enterprises in which opportunities may exist for the marginalized groups, vegetables, fruits and bananas have substantial market orientation and would hold promise in integrating the poor in agricultural markets. Cereals such as sorghum and millet, pulses such as beans cowpeas, green gram and groundnuts and bananas are important to female headed households and could hold key to unlocking their potential to participate adequately in markets. On livestock, small ruminants (goats and sheep) and chicken are gaining increased importance in commercialization among both the female headed and income poor households. Targeting these livestock could also be important in efforts aimed at increasing market participation among women and the poor. In addition, promoting collective action and facilitating access to credit are also key intervention areas that can significantly enhance access to both input and output markets and thus integrate women and the poor to market for income growth and poverty reduction.