Effect of Labour-Out Migration on Farm Input use in Banana and Legume Systems in Great Lakes Region of Central Africa 1 Ochieng Justus 1, 3, Beatrice Knerr 1*, George Owuor 2, Emily Ouma 3 1 Department of Development Economics, Migration and Agricultural Policy and ICDD, University of Kassel, Germany. 2 African Economic Research Consortium (AERC), CMAAE Program, Nairobi Kenya. 3 International Institute of Tropical Agriculture (IITA), Burundi. Corresponding Author: Ochieng.Justus1@gmail.com 1. Introduction The role of rural out-migration in the development of agriculture is highly debated: on the one hand, it implies a loss of labour force for the household and the region of origin; on the other hand, remittances may contribute to livelihood security and provide a source for investment. Hence, migration holds a great potential to increase agricultural intensification and productivity in low-income countries, depending on the circumstances and also on the policy framework. However, there are limited studies on the impact of migration in East Africa, especially in the Great Lakes Region. Studies in the region did not find indications of a negative impact of migration, but find moderate capital investments in agriculture among migrant households (Hyden et al. 1993). Several studies have focused on the spending pattern out of remittances (Wouterse, 2010), but due to the fungibility of the remitted money the results are not suitable to explain the effect of migration on agricultural production. This study focuses on the migration status of the farm households to analyse the role of labour out migration on input use. 2. Theoretical Framework The Lewis and Fei-Ranis Model (Lewis, 1954) explains the causes and consequences of rural-to-urban labour migration in a dualistic economy characterized by labour surplus in the agricultural sector and labour scarcity in the industrializing urban sector. The agricultural production systems are characterized by a low level of external inputs and reliance on family labour. According to the Lewis and Fei-Ranis model, the outmigration in the initial phase does not entail any production loss in the agricultural sector as the marginal product of labour is zero. With increasing out-migration, however, labour becomes more and more scarce which means that it implies production losses. The national (rural-urban migration) and international migrations represent a loss of labour and capital to the household members left behind. The New Economics of Labour Migration (NELM), initiated by the work of Stark (1991), considers the household as the decision unit about the migration of family members. It focuses on the joint decision about migration and remittances, whereby remittances could serve different purposes: livelihood security; reduction of income risk by diversifying the income portfolio across different regions; and compensation of insufficiently functioning capital markets (for credits and insurances). 1 Paper presented at XXVII IUSSP International Population Conference on 26 th -31 st August, 2013 in Busan, South Korea. 1
3. State of Research Based on a large-scale survey in Cameroon Schrieder and Knerr (2000) found out that a major motivation for remitting to the left-behind family is securing a basic livelihood of the nuclear family, and the preservation of the inheritance which implies that in particular over periods of draughts, the amount of remittances received from the potential heir increased with the size of the left behind family s cattle herd. In the Ecuador Andes, land use and agricultural production of migrant households are not significantly different from those of non-migrant households (Jokisch 2002), as remittances can be used for other purposes and are not necessarily invested in agricultural activities. Migration positively influenced the adoption of new seed varieties in Mexico (Quinn, 2009), while in Albania migrant households preferred to invest in livestock production and less in inputs for crop production, such as fertilizers, seeds and herbicides (Mulluka et al., 2007). Alan de Brauw, (2007) found out that in Vietnam farm households with migrants spent less labour in farming activities than non-migrant households. These results demonstrate that the impact of migration and remittances on farm households is not homogenous, but, in contrast, various according to the specific household conditions and the regional context. Wouterse (2010) emphasizes in particular the difference between the impacts of overseas and intracontinental migration on the sending areas and on the life of the migrants. 4. Hypotheses and methodology Based on theory and the review of the state of research, we start our empirical analysis based on the hypotheses that migration and remittances have a significant influence on fertilizer, new plant varieties and hired labour. The study presented in this paper was done in the Consortium for Improving Agriculture-based Livelihoods in Central Africa (CIALCA) project action sites 2 in Rwanda and the Eastern Democratic Republic of Congo (DRC). It is based on a survey among 480 farm households which had been selected by a multi-stage sampling technique. 166 of them were migrant households. Data were collected by using a structured questionnaire to obtain information on farm household s agricultural production, marketing activities, and migration details. In addition to descriptive statistics, a Tobit model was used to analyse the effect of migration and remittances on input use because it prevents unbiased and inconsistent estimates that could arise from corner solutions (Gujarati, 2004) which arise because some households do not use the investigated inputs (fertilizer, high yielding varieties and hired labour). 5. Results and Discussion 3 The major reason for migration was wage employment, followed by education, personal conflicts and visiting relatives. Most families had sent their children to institutions of higher education in the urban centres and abroad to study; therefore no remittances are expected from them. Migration for education is among the only legal ways for youth 2 Action Sites correspond to different administrative units in each of the countries ( Secteurs in Rwanda and and Localités in DRC). 3 Note that these are results from preliminary analysis. Comprehensive analysis will be included in final paper. 2
from low-income countries to enter developed countries and it provides opportunities to turn migration for education into labour migration (World Bank, 2007). Table 1: Reasons for migration Reasons Frequency Percent (%) Wage employment 86 54.4 Armed conflicts 9 5.70 Further education 45 28.5 Visit relatives 18 11.4 Total 158 100 Source: Authors results More than 30% of the migrant families receive remittances; almost 20% goods in kind, like clothes (Fig. 1). Some migrant farm households are less vulnerable to food insecurity shocks because they receive remittances mainly during emergencies. Almost half of the families left behind do not derive significant benefits from the migration, yet. Figure1: Benefits of migration Source: Authors results Migrant families had slightly more household members than non-migrant families. There was no significant difference in land sizes. The migrant households head were married, older and appeared little less educated than the non-migrants household heads. The difference in the age of household head could be attributed to the migration of young people from rural areas. However, the migrant household had significantly larger family size which contradicts other findings reported by Miluka et al, (2007) but confirms the findings by Vasco (2011). Most of the household heads were relatively less educated because the migrants are normally positively selected with individuals with secondary or higher education levels which help them to get employment faster in their destination are likely to migrate than those with lower education. The migrant households also had higher access to credit and owned more mobile phones although all were not significantly different from non-migrant households. The mobile 3
phones facilitate constant communication between the migrant and the family left behind. Migrant households also owned bigger farms and lived closer to the markets which facilitate timely delivery of marketable surplus to the markets. The market orientation levels of migrants households were higher than the non-migrants households indicating that the former were likely to have higher sales. Migrant families were hiring more labour during the peak planting season than non-migrant families which can be attributed to remittances. The remittances are mainly received quarterly (51%), followed by occasionally in case of emergencies or at the planting seasons (30%), and monthly in cash (19%) which is common with families who depend on remittances to survive. Farm households with migrants on average had higher number of livestock holding than those without. The households with migrants were likely to switch from crop to livestock production because they can afford the liquidity to purchase the animals. This confirms the experience that in the African context cattle are often used as a saving box for remittances, as, for example, observed in Lesotho (Knerr 1998). Table 2 Characteristics of migrant and non-migrants households Variables Non-migrant Migrant t-value households households Household Characteristics Female household heads 21.20 17.68 0.91 Education of the head (years) 10.13 10.39-0.73 Age of the head (years) 45.49 52.43-5.63*** Married heads (%) 64.24 73.17-1.98** Migration network (%) 4.43 34.15-9.52*** Dependants (persons) 3.06 3.87-3.44*** Household size (persons) 5.61 7.24-5.49*** House material (% with iron sheets or 80.38 85.76-1.18 tiles roof) Transport equipment ownership (%) 32.59 26.83 1.30 Mobile phone ownership (%) 55.75 62.80-1.70* Farm and community characteristics Farm size (hectares) 1.37 1.43-0.10 Distance to the market (km) 3.73 3.66 0.18 Livestock holdings (TLU) 1.05 1.92-2.32** Access to credit (%) 25.32 29.26-0.93 Extension contacts (number) 7.00 6.35 0.71 Country (1 if Rwanda %) 49.37 27.82 4.86*** CIALCA market orientation strategies 59.81 62.80-0.64 (%) Log expenditure on fertilizers 3.22 2.23 3.10*** Log expenditure on high yielding 3.18 3.25-0.41 Log expenditure on hired labour 3.06 3.11-0.25 Sample (N) 316 164 Notes: *, **, *** significant at 10%, 5% and 1% respectively. Source: Authors results 4
The Tobit model used to econometrically estimate the effect of migration on input use indicate that migration negatively influence use of fertilizers (p<0.01), high yielding varieties (p<0.05) while it positively influence use of hired labour though not insignificant. This means that the hypothesis that migration would increase intensification in banana and legume production does not hold. Similar negative findings have been reported among smallholders by Miluka et al. (2007) in Albania, Rozelle et al. (1999) and Luhua et al. (2013) in maize and wheat production in China respectively and Maharjan et al. (2013) in Nepal among others. However, the negative effect of migration conflicts the findings by Lihua et al. (2013) for fertilizer use in orchard production in China, de Haas, (2006) for fertilizer use in crop production in Morocco and Vasco, (2011) in rural Ecuador. The migrant household were able to hire more labour in their farms mainly from the occasional remittances which are normally received during crop land preparation and planting period. The factors such as farm size, family size, livestock holding, access to agricultural extension and education level of household head also influenced input use among smallholder farmers. 6. Conclusion Migration in Great Lakes region has the potential to play a major role in enhancing the intensification of farm production for families left behind. However, in Central Africa, it negatively affects agricultural intensification which still supports the NELM and Lewis- Fei- Ranis theory arguments. Migrant households are also slightly larger than nonmigrants households. Agriculture remains the main focus to achieve food security in the long run, with migration having potential to help to increase the purchase of farm inputs. This is possible if the remittances are invested into inputs required to improve agricultural productivity. The major question to investigate in this context is how far the remittances received are adequate to purchase food and agricultural inputs in the region. Probably the remittances received by farm households are invested in other non-farm activities or farm activities other than banana and legumes since they are involved in multiple cropping systems. Therefore, creation of enabling environment for both migrants and family left behind to invest in agriculture through improving basic infrastructure and efficiently channel the right production and marketing extension messages to smallholders are possible options. The region s governments could also subsidize the remittances sent for agricultural production activities. References De Brauw, Alan (2007): Seasonal Migration and Agriculture in Vietnam," Working Papers 07-04, Agricultural and Development Economics Division of the Food and Agriculture Organization of the United Nations (FAO - ESA) De Hass, H. (2006). Migration, Remittances and rural development in Southern Morocco, Geoforum, 37 (4): 565-580 Gujarati, D. N. (2004), Basic Econometrics, Fourth Edition, McGraw-Hill Publishers, New York, USA. Hyden, G., Kates,R.W.,Turner,B.L.I., (1993), Beyond intensification. In:Turner, B.L.I., Hyden, G., Kates, R.W. (Eds.), Population Growth and Agricultural Change in Africa. University of Florida Press, Gainesville, FL. 5
Jokisch, B. (2002): Migration and Agricultural Change: the case of smallholder agriculture in the highlands of South-Central Ecuador, Human Ecology 30: 523-550. Knerr, B (1998): Labour Migration from Developing Countries. Macro-economic Impacts and Policy Interventions. Kassel Lewis, W.A. (1954), Economic Development with Unlimited Supplies of Labor, the Manchester School, 22: 139-191 Lihua, L., Wang, C., Segarra, E., and Nan, Z. (2013): Migration, remittances and agricultural productivity in small farming systems in Northwest China, China agricultural Review, 5(1):5-23 Maharjan, A., Bauer, S., and Knerr, B. (2012): International Migration, Remittances and Subsistence Farming: Evidence from Nepal, International Migration, 51(1):248-263 Miluka, J. Carletto, G. Dans, B. and Zezza.A, (2007) The Vanishing Farms? The Impact of International Migration on Albanian Family Farming.Washington DC. World Bank Policy Research Working Paper 4367 Quinn, M.A. (2009): Estimating the impact of migration and remittances on agricultural Technology. The Journal of Development Areas, 43:199-216 Rozelle, S. Taylor, J.E, debrauw, A. (1999): Migration, Remittances, and Agricultural Productivity in China. The American Economic Review, 89(2):287-291 Schrieder, G. und B. Knerr (2000): Labour migration as a reinsurance in rural households in sub-saharan Africa: The case of Cameroon. Oxford Development Studies. Volume 28 (2). Stark, O. (1991): The Migration of Labour. Basil Blackwell, Cambridge, United Kingdom World Bank (2007): World Development Report: Development and the Next Generation. Washington DC, USA Wouterse, F, (2010): Migration and Technical efficiency in Cereal Production: Evidence from Burkina Faso. Agricultural Economics, 41: 385 395 6