Food insecurity in fragile lands : Philippine cases through the livelihoods lens

Food insecurity results from a web of problems involving human and non-human processes within certain environments. This thesis is both a metho­dological and a policy-oriented study. It explores the linkages in order to understand the food security situation in less favored areas (LFAs) in the Philippines. In the Philippines, food insecurity can be argued as most seriously felt in the LFAs which constitute about 65 percent of total agricultural land and where about 70 percent of the rural poor live. About 40 percent remain poor and food insecure in the rural areas. The two research villages (i.e. Alegre and Plaridel) in Leyte are representative of different types of LFAs: the flood-prone lowlands and the high risk eroded uplands and mountainous region in varied socio-economic settings. Methodological issues: fit, relevance and applicability

Since food security is multi-faceted, the three dimensions of food availability, access and adequacy are addressed using a combination of formal surveys and qualitative tools (e.g. focussed group discussions, case studies, key informant interviews), which were so designed that they mutually enrich the investigation. The formal surveys (i.e. household production and socio-economic survey; food consumption, nutrition and health survey) are parallels to the existing national surveys regularly conducted (FIES, NNS), and were so selected to build on them in order to gain insights at im­proving their usefulness, particularly for the comprehensive assessment of food secu­rity. The various non-formal survey methods and tools were designed to give context­ual information on the results of the statistical analyses (multi-variate analyses), and to qualify and enrich the interpretation of the significant variables that were identified in the regressions.

This thesis attempts to treat the food insecurity issue among households in LFAs using the livelihood approach within the framework of the inter­relationships between their biophysical and socio-economic environments. The variables used in the analyses also consisted of biophysical, socio-demographic and economic vari­ables. The households in the LFAs are engaged in farming to a greater or lesser degree. Iit is therefore logical that the biophysical environment is the starting point of analysis. This was done in the agro-ecology and land-use ethno-histories, the case studies, the community surveys and in the elaboration of Hypothesis 1. The latter deals with productivity determinants of the major food and cash crops. It was hypothesized that the niggardliness of the biophysical resource, and other biophysical-related and socio-economic factors contributed to low farm productivity, and thus, pushed house­holds to seek for other sources of income.

Hypothesis 2 deals with the factors that most likely affect the households’ deci­sion in composing their livelihood portfolio. Livelihoods are grouped into four types: same percentage distribution of income sources (LIVETYP1); farm income>50% (LIVETYP2); off-farm income>50% (LIVETYP3); non-farm income>50% (LIVE­TYP4), based on the income percentage criteria of the total full income of households obtained from these various income sources. This was found to be a useful classifica­tion because of the divergent trends found in LFAs, despite the similarities in cultiva­ted crop (coconut, rice, rootcrops) and livelihood activities (farming, backyard pigge­ry, domestic services, seasonal off-farm work, carpentry/construction). With this classification, it was found that one village counts more of the farming type of house­holds (Alegre), and the other more the non-farm type of households (Plaridel). The regressions on Hypotheses 1 and 2 combined with the community surveys and case studies explained and qualified the differences in livelihood portfolios. Also, it was found that livelihood types do impact significantly on the choice and productivity of crop systems and the nutritional status of children. This thesis argues for the useful­ness of the livelihood approach in exploring the linkages in food security analysis.

In addition, the data and regression results built around hypotheses 1 and 2 pro­vided insights in food availability and food access. Finally, food adequacy is dealt with in Hypothesis 3 which investigated the determinants of nutrient intake and nu­tritional status of children, the latter as proxy for health. Nutritional status was re­gressed on the livelihood types together with economic, socio-demographic and health-related variables. Coming full circle, the factors affecting nutritional status of children as well as their health conditions were identified.

In methodological terms, the study attempts to provide a workable protocol for an integrative analysis of food security which can be applicable in development work (e.g. RDE programs and projects) or for policy purposes at local or national levels. This protocol is relevant and fitting particularly to the Philippine setting because of the recent moves to improve the local database of minimum basic data needs for food security assessments. The procedure and results of this study can be instructive and be tested, and further improved, in the current food security assessment initiatives by local government units with the assistance of the SUCs and research institutes. Research results: In this section, I will summarize the results and relevant policy concerns. But first, I will present a brief on the food security situation in the LFAs as gathered from the results of the ethno-histories, FGDs, case studies, community surveys and informal interviews in the research villages.Nature and man both played a role in the deterioration of the villages’ land­scapes over time. Flood-prone Alegre is located along the typhoon path. The unregulated quarrying operations of the politics-backed businessmen and the deforestation of the bigger watershed of which the village is part, greatly contributed to the high risk of floods which can be triggered just by heavy rains. This means high risk of flooding for the rice farms, vegeta­ble and rootcrop gardens, which contribute significantly to food and cash. This results in the erosion of capital which is usually sourced from the often usurious credit. With res­pect to farming, mainly upland Plaridel is no better, except for the fact that it has an additional cash crop, abaca, which provides added opportunities for off-farm wage work. Plaridel is also along the typhoon path, its hilly and mountainous farms ex­posed to wind and rain that cause erosions. Illegal logging by the influential and wealthy local businessmen has also contributed to the deterioration, and, in addition, has caused damage to the fish habitat leading to the disappearance of local fishing as an important liveli­hood. In both villages, farmers do not have the incentive for conservation farming. Farming income in general is not enough to feed and attend to the other basic needs of households. Both villages differ in many respects from each other and this led to differences in livelihood strategies and the food security situation. Plaridel is a bigger village in terms of non-farm (native craft industry, construction, trade) and off-farm opportunities (coconut and abaca) and is part of a more economically vigorous local economy, whereas Alegre has fewer non-farm (domestic service, local transport) and off-farm (smaller coconut and rice) activities. About 47 percent of households in Plaridel are depending mostly on non-farm incomes, and about 42 percent in Alegre mostly on farm incomes. The per capita income of Plaridel is about 0.90 USD and of Alegre 0.64 USD, both still below the WB poverty threshold of a dollar a day. In terms of income, households in these villages are poor. This below poverty threshold income presents serious limitations in terms of food access. Enough nutrient sources of food are available in the nearby town and village markets yet the households reported being food deficit two to three months a year during the off-harvest season when cash is also very low. Food access is the more limiting factor. Food habits are part of culture. Partly because of cultural factors and partly because of the village terrain and proximity of fields, Alegre households tend to have vegetable gardens more than in Plaridel, though quite limited in diversity. The use of the nutritious taro and sweetpotato leaves and petioles, and the preparation of many dishes with fresh coconut milk are very much a part of the Waray-waray food culture. Despite the relatively lower per capita income than Plaridel, Alegre household mem­bers turned out to have better nutrient adequacy for all nutrients tested and for all age groups. This could also be because Alegre mothers had more time for home and child care (associated with farm and off-farm type households) than Plaridel mothers (non-farm type households). The statistical regressions showed the high significance and importance of off-farm type households as positively affecting the nutritional status of children via the associated intensity of time allocation for home and child care. Though still below national averages, Alegre households are better off in nutrient adequacy for all nutrients than the better-income households in Plaridel. This allows for the conclusion that a higher income level per se is not an assurance of nutrient adequacy. Further, the home garden argument implies that improving the availability of nutrient sources from the home gardens provides opportunity for food and nutrient adequacy improve­ments especially among the poor households whose income levels will really take time, if at all, to adequately improve.

The village government of Plaridel was able to adequately provide for quality drinking water from three spring water sources in the mountains, while potable drink­ing water in Alegre was accessed outside the village due to poor quality of its ground­water. Plaridel has better school and health facilities than Alegre, where households are served by one school only and have some access to health services. There were more sources of credit in Plaridel, including the Church-organized cooperative for home industry and consumption uses, while Alegre households had to contend with the usurious local moneylenders, except for one or two farmers who accessed credit from the Rural Bank in town. Plaridel has relatively more traders who are also financiers of farm capital, i.e. seed, fertilizers or wage payment. Comparative­ly, Plaridel has more in terms of economic opportunities and support services, while Alegre has more of economic constraints. The economic gains however have not been translated in terms of nutritional status.

All these basic data provide strongly for the argument that the significant and important factors that contribute to food security are not only economic in nature. Economic factors such as income and food budget are necessary but not suffi­cient conditions for food security.

Among poor households, labour is the most important resource. Labour as a variable, is a special case, and I will treat it as quasi-economic and quasi-social be­cause it has both an economic and a social value. The latter simply means that the consideration for labour use decision is not only for monetary aims but also for non-monetary rewards or benefits. This is especially true with regard to a mother’s or a wife’s use of labour time. In households where women play a dual role as both producer and consumer, valuation of labour is subjective because there is no ‘real market’ for own use of time partly because of social norms. This has an important bearing particularly in labour sub­stitution between wage and domestic work.

The hard evidence: in reply to the queries

The main research question that this thesis deals with is: “How do the livelihoods of farming households impact on their food security in a context of ecologically fragile environments?” The explanations of the variables or factors from the interrelationship between hypotheses 1 and 2, and between hypotheses 2 and 3 clarify the answer to this question.

It is clear at the outset that farming households in the LFAs are on average poor and engaged in diverse activities to make ends meet. In this study, the households are classified based on the percentage contribution of income from three groups of activities (farm, off-farm, non-farm). The ‘livelihood typology’ is used as a proxy for resource-use strategies. Households face difficult choices in allocating their two most important resources, labour and time, with very meagre contribution of other inputs, because of constraints and limited opportunities.

The livelihoods of farming households in less favoured areas are affected by the interplay of biophysical and socio-economic factors. These factors have differential impact on the different livelihood types because of the availability and competing use of labour of adult working members, the household size, the size of cultivated farm, value of farm produce, and the existence of idle lands. From direct observation and interviews it bacame apparent that idle land is the catch-all variable for poor quality land and/or the lack of capital to make use of technology and other inputs.

It will be noted that the final indicator of food security is that of nutrition ade­quacy, the latter of which is indicated by nutrient adequacy of the individual house­hold members and nutritional status of children. In the final analysis, the nutritional status of children, especially those under five, is highly significantly and positively affected when households income is obtained more from off-farm activities (LIVE­TYP3), and positively when more income comes from non-farm activities (LIVE­TYP4), and negatively when income is more from farming activities (LIVETYP2). The evidence for the latter two, however, is not conclusive.

Hypothesis 1: low farm productivity and multi-livelihoods

How does the situation in ecologically less favoured areas affect the choice and generation of livelihoods?

The first hypothesis has addressed this query by investigating the effect of the biophysical and socio-economic environments on farm productivity. Households have stopped depending only on their farms for sustenance because of low farm produc­tivity and low farm incomes. They maintain a dual or multi-activity livelihood portfo­lio because farming alone cannot provide adequately, even for their most basic needs. Households usually combine farming with other sources of income: in farm wage work (coconut, rice, and abaca) and/or non-farm employment (native craft, peddling, domestic service). The biophysical and related factors found to have highly significant effects on crop productivity include the season, fertilizer use, irrigation water and soil type for rice, terrain and variety for sweetpotato, cropping pattern for abaca; and idle land, indirectly, via livelihood choice (LIVETYP4) for coconut and rice. The higher the incidence of idle lands, the greater the tendency for households to engage in non-farm employment, and the more non-productive the farms are due to labour shifts away from them. Poor land quality with less or no capital and technology inputs to start with, results in low farm productivity. The socio-economic variables that have significant impact on farm productivity are: non-farm income, share tenancy, remittances, and dependency ratio for abaca, wage rate, production cost, farming experience for rice and coconut, age of farmer for sweetpotato, and livelihood type for coconut, abaca and rice. It is important to note that the cash crops abaca, coconut and rice are more sensitive to livelihood types es­pecially the non-farm (LIVETYP4) and off-farm (LIVETYP3) livelihood types. The negative relation implies a labour substitution effect (coconut, abaca) and the positive relation implies the capital plough-back effect (rice). Higher wage rate will have a greater likelihood of reducing rice yield due to labour shifts into more lucrative non-farm work, but not with coconut and abaca because these are the main cash crops. Similarly, the effects of the off-farm income and non-farm income variables depend on relative labour shifts away from coconut farms, and the capital plough-back in abaca farms, respectively. The higher the dependency ratio the more likely farmers intensify the use of resources for abaca farming because it is the more cash-earning crop, with a stable market and better prices. The more remittances a household has the more negative the im­pact on abaca production due to distance-induced inertia because the abaca fields are located three kilometres or more to the mountains. Share tenancy is significant but not a negative issue, especially with abaca and rice, because what is more relevant is the use-right of the land without the burden of tax (which goes with ownership). Working on the land for generations has generated family loyalty sealed with mutual benefits. Share tenancy, however, tends to nega­tively impact on coconut productivity because of labour shifts, which are likely due to the more diffused labour use among coconut farmers who are engaged more in off-farm and non-farm activities.

These differential effects on the various crops show that crops respond differ­ently to various factors, and this factor-specificity has important policy implications for targeting intervention options for productivity improvement.

Hypothesis 2: the livelihood portfolio – choices and dimensions

From Hypothesis 1, it is concluded that farming households have a diversified liveli­hood portfolio because low productivity and low incomes from farming cannot sus­tain their lives. In summary, livelihood type has high significant correlations with the variables farm wage worker (+), age of household head (+), size of cultivated area (-), actual household size (-), value of farm produce (+), number of working children (+/-). The evidence on the life cycle variable is not conclusive. Generally, the behaviour of the relationships shows labour shifts from one type of activity to another mainly because of household size, labour supply within household, the area and productivity of the farm.

The inclusion of the dimensions on assets and resources, sex, age, and life course was based on their assumed importance in the process of decision-making for the livelihood portfolio. In Hypothesis 2, these are included as in­dependent variables in the multiva­riate regressions for the livelihood types, and sought to address the basic question of what determines the households’ livelihood portfolio. Livelihood type one (LIVE­TYP1), the control, refers to household livelihood strategies which result in more or less equal distribution of income earnings from farm, off-farm and non-farm sources; livelihood type two (LIVETYP2), where the share of farm income is greater than 50 percent; livelihood type three (LIVETYP3), when off-farm income is greater than 50 percent; and livelihood type 4 (LIVETYP4), when non-farm income is greater than 50 percent. The non-farm livelihood type household tends to be highly significantly and positively related with idle land. There is a significant and negative relation with ac­tual household size. This clearly shows the movement towards more non-farm em­ployment when land becomes less productive, or with the lack of inputs or incentive to till the soil. A smaller household size means that one or two spouses can have more labour time shifted from domestic to non-farm work. This variable is highly signifi­cant for off-farm livelihood type. Farm wage worker is highly significantly and posi­tively related with both farm and off-farm based livelihood types. But with more working members, household labour will likely be shifted to off-farm work. Although not con­clusive, life cycle tends to have a negative relationship with the off-farm livelihood type, since the earlier in the life cycle with no child, the more labour time available for off-farm work. The value of farm produce is highly significant and positively related with farm livelihood type, showing that households will greatly rely on farm income if its value is high. This is a variable that captures relatively good land, better yield, and better market. Also the bigger the farm land cultivated, the more likely the negative im­pact on off-farm livelihood type because this would mean greater labour requirement on farm, and thus, less labour available for off-farm work.

The role of gender is critical in terms of type of labour available and employ­ment opportunity. Male labour is usually employed in off-farm activity of cash crops such as coconut and abaca, and female labour mostly for non-farm employment like native crafts and domestic services. Household size affects labour supply, espe­cially of women since their labour use on domestic chores is high. Hence, a bigger household size tends to negatively impact on their availability for off-farm (in rice) and non-farm work.

Hypothesis 3: nutrient adequacy and nutritional status

Following through the series of relations and results, the final answer to the main question is provided in this section on the food adequacy situation. On the whole, nutrient adequacy and nutritional status of households and individuals in the LFAs, as represented by the research villages, are far below the national averages.

The study found no significant differences in the nutrient adequacy levels be­tween age and sex groups, except a better-off picture of adequacy for all nutrients among children below 5 years of age. Overall, the energy adequacy level is only about 50 percent, which has serious implications for the growth of children and health of adults. Children under 5 and adults tend to have better nutrient adequacy, but chil­dren under 5 are critically deficient in Vitamin A, C, riboflavin and calories. The males, the children and adults tend to be more adequate in protein than the teenagers and the females. Also, there were no significant differences in the nutrient fair-share ratio by sex or age which indicates a fairly equitable distribution of nutrients within the household. However, there was a slight positive bias for all nutrients in favour of children below 5 years old, and women tend to be slightly favoured with calcium, vitamin C and thia­mine. This reflects priority given to the very young and mother’s nutrient needs. Despite the lower average per capita income in Alegre, it showed better ade­quacy levels than Plaridel for all nutrients. Obviously, not because of the income fac­tor but for a mix of reasons: there were more home gardens in Alegre, resulting in a more nutritious diet in the area, and more mother home care time as gleaned from their livelihood patterns. Nutrient intake is the major determinant for nutritional status. This implies that nutritional status is significantly dependent on variables affecting nutrient intake, such as the food budget, time allocated for home care, mother’s or caregiver’s educa­tion, the availability of non-bought food (own-production, given), household size, age and sex of the individual. There are differential impacts of each of these variables on specific nutrients based on the sensitivity of the nutrient to each of the above vari­ables. Thus, a greater probability that households and individuals in less favoured areas will have higher nutritional status when the food budget is improved, when mo­thers are better educated to take care of the home and its food needs, when household size is planned to a reasonable minimum, and when food is available from non-bought sources. As the food budget includes both bought and non-bought food, the food budget can be improved not only by improving income but also by intensifying the home production of nutrient-rich food crops. The separate regressions for nutritional status, using the weight-for-age of chil­dren as dependent variable, found as highly significant such variables as drinking water quality, education of mother, total farm full income, livelihood type, and short-term morbidity. Hygiene was significant particularly for children less than 5 years. Nutrient knowledge per se was not found significant in the individual nutrient intake regressions, but was found highly significant and positive with the nutritional status of school age children. Livelihood types matter in nutritional status via labour and time allocation and food priority effects. The off-farm livelihood type (LIVETYP3) was found highly significant and positively affecting the pre-schoolers’ health, most probably because of time allocation effects when the mother has more home care time with the child when the father is away for off-farm work. Though not conclusive, the farming livelihood type (LIVETYP2) variable is negatively related with health of school­children, and the non-farming livelihood type (LIVETYP4) variable is positively rela­ted with the pre-schoolers’ health via time allocation effects and priority to the very young with additional food from improved food budget, respectively. Total farm full income includes all produce of the farm that are sold and those consumed in the home. This variable is highly significantly and positively related with health of children under 5, yet significantly and negatively with school children’s health. This again highlights the contribution of non-traded or home-grown food and nutrient sources since the income variable per se has no evidence of significance. More significant is the trade-off between food, care and goods in favour of the pre-schoolers. These further confirm how pressing the food and income constraints are among the poor households that this trade-off is biased against the school children. With pre-school children, short-term morbidity is highly significant and impor­tant as they are more sensitive and prone to contracting illness of whatever kind. Hy­giene practices are a significant factor in the health of pre-schoolers, but not with pri­mary schoolchildren. This implies the relative sensitivity of children under 5 to hygie­nic practices and conditions. There is no conclusive evidence for garbage disposal. Firming up policy options

Table 11 presents the summary of policy actions necessary to address the significant and important variables identified in this study. The initiation level indicates where policy can originate. National refers to the national agency or relevant research, development and extension institute (RDE), and local refers to provincial, municipal or village governments. The indication of both national and local implies strong coor­dination with the relevant national agency or research institute and local government unit. Effective implementation presupposes good governance, cost-sharing for cost effectiveness, good human and non-human resources management, and collaboration and partnership with non-government organizations and community-based organiza­tions present in the respective areas.

Given the results of this study, significant policy concerns can be drawn and need to be addressed in order to respond to the persistent food insecurity in LFAs. But there are major over-arching concerns that must be considered in drawing out policy options: the continued deterioration of the biophysical environment (watershed management, resource conservation), the movement away from the farm in livelihood pursuits and the increasing role of non-farm employment, a rationalized RDE in agri­culture, and agrarian reform. Land suitability should be an integral part of agriculture RDE, both at the national and local levels. These concerns will need a longer time perspective, a broader scope of policy action, and concerted efforts in order to effect­ively help poor households, especially in the targeted LFAs. Addressing these con­cerns will help provide both the necessary and sufficient conditions for the policy actions, herein summarized, to have impact on the significant variables (biophysical factors, land tenure, non-farm livelihood, remittances). There is need for a review and assessment of the planned and implemented programs related to these concerns. In the medium and long runs, the RDE programs for cash and export crops, like coconut and abaca, need special attention in order to improve their productivity and comparative advantage vis-à-vis the world market. These cash crops are main farm income sources in the LFAs and provide greater opportunities for off-farm work, which tend to impact significantly and positively on the nutritional status of children. These also may yet provide the only significant livelihood option for those who are left in the farms to earn a living since local non-farm opportunities are quite limited. It is important to highlight what could be addressed by policy in the short and medium-terms that will most likely improve food security especially of the poor farming households in the LFAs. In particular, I am referring to the need to improve the coordina­tion and implementation of the food-based program on own-food production through home gardens to be integrated with the non-formal education program of women and men. The latter includes components of food-nutrient diversity, home gardening, pri­mary health care, reproductive health, preparation of nutritious local food dishes, micro-enterprise livelihood, home management, and hygiene and sanitary practices. These comprise existing programs of the Department of Health, Department of Social Welfare and Development, Department of Agriculture, the BIDANI, and some local governments, but have not been fully implemented or well coordinated for various reasons. These programs address most of the factors, which were found to be highly significant and important, to improve nutritional status. Such programs can be initia­ted at the local levels and in collaboration with national initiatives where applicable.

In relation to the food-based nutrient diversity approach, there is also the critical need to develop and integrate an RDE program in existing food crop improvement agenda of relevant research institutes aiming at improving the knowledge base of the nutrient content of food crops. Of the 80,000 species or so of plant sources of nutrients, only a very limited number are grown in own fields and gardens. Promoting the diversity of plant foods that can be produced in farms and gardens through im­proved food-nutrient information and nutrition campaign program have increasingly received attention in recent years as vehicles to alleviate poverty and food insecurity (Gari, 2004; Frison et al., 2006; Toledo and Burlingame, 2006).

Further, the methodological issue that needs policy action is the coordination of the family income and expenditure survey (FIES) and the national nutrition survey (NNS) so that they can be more useful in a holistic food security assessment. The organizational structures to conduct the surveys are in place but may need modification in order to stratify for LFAs or otherwise, and ensure the consistency of the sampling design, the reference period, and the timing of surveys in order to apply the whole range of analysis in the production-consumption continuum. This can save a substantial amount of time, budget and effort in the development and implementa­tion of food security targeted programs or projects at various levels.

Table 11.    Summary table of significant variables and policy options

Policy variables	Policy option type	Initiation level
Biophysical environment and technology
Idle land	Technology, RDE infrastructure	national, local
Land, soil quality	Technology, land-use suitability assess­ment, RDE infrastructure	national, local
Cropping pattern, variety	Technology, land-use suitability assessment, RDE infrastructure	national, local
Fertilizer use	Technology, regulatory, RDE	local
Irrigation	Regulatory, RDE infrastructure	local
Socio-economic environment, demographic and health
Farm livelihoods  (LIVETYP2)	Cash crop RDE (coconut, aba­ca, rootcrops, rice, vegetables) Food-based program (home gardens, own-food produc­tion), RDE in nutrient content of own-food produced and wild foods. Support services (capital, market)	national, local   local   national, local
Off-farm livelihoods (LIVETYP3)	Cash crop RDE (coconut, abaca) Trade and industry, regulatory	national, local   national, local
Non-farm livelihoods (LIVETYP4)	Technology, design, RDE Support services (capital, mar­ket, quality control)	national, local national, local
Drinking water quality	Utility service/ infrastructure	local
Mother’s education	Women non-formal training program (food preparation, food-nutrient information, home gardens, home manage­ment)	local
Nutrition knowledge	Women non-formal training Food-nutrient RDE, education and information campaign (social marketing)	national, local
Remittances	Social service (adult education/ training, savings mobilization, capital investment)	national, local
Land tenure	Agrarian reform program review (with assessment of CARP)	national
Household size Number of working children Dependency ratio	Population management (pro-active local implementing in­frastructure) Non-formal education for women and men (reproductive health, home management and relations)