当前位置：文档库 › Assessing the impact of migration on food and nutrition security

Assessing the impact of migration on food and nutrition security

Alberto Zezza a ,?,Calogero Carletto b ,Benjamin Davis c ,Paul Winters d

The World Bank,Via Labicana 110,00184,Rome,Italy

The World Bank,1818H Street NW,Washington,DC 20433,USA c

UNICEF,Regional Of?ce for Eastern and Southern Africa,P.O.Box 44145,Nairobi 00100,Kenya d

Department of Economics,Roper 203,American University,Washington,DC 20016-8029,USA

a r t i c l e i n f o Article history:

Available online 3December 2010Keywords:Migration Nutrition

Food consumption

a b s t r a c t

Migration has become a key component in the livelihood strategies of an increasing number of house-holds across the developing world and remittances have expanded dramatically in the last decade.This has come at a time when an increased emphasis has been placed on reducing malnutrition to achieve Millennium Development Goal (MDG)targets.While this is the case,there has been little attention on the interface between migration and nutrition even though migration can in?uence nutrition through a number of channels.The objective of this special issue is to present state-of-the-art analyses of the link between migration and nutrition in developing countries.In this paper,an overview of the conceptual and empirical issues in identifying the link between migration and nutrition are considered.Further,the results from seven country case studies are synthesized and policy implications are drawn.

Introduction

Despite countless global initiatives and individual country and regional efforts,malnutrition among children under-?ve and for the general population remains ubiquitous in many developing countries.An unacceptably high number of countries are unlikely to meet Millennium Development Goal (MDG)targets of halving the prevalence of underweight children under-?ve and the propor-tion of the population below minimal level dietary energy con-sumption (MDG Target 1.C).Malnutrition and limited food access remain stubbornly high in much of Sub-Saharan Africa and South Asia,but also in some Latin American countries and in other parts of Asia.The latest estimates from the Food and Agriculture Organi-zation (FAO)put the global number of undernourished at 925mil-lion,below the 1.02billion found for 2009during the peak of the global recession,but remaining very high and greater than the 2005–2007estimates (FAO,2010).

Against this backdrop,migration and remittances have become a key component in the livelihood strategies of an increasing num-ber of households across the developing world,with large numbers of people seeking better earning opportunities in richer countries or in more developed areas within their own country.Remittances have expanded dramatically in the last decade,driven by an up-surge in international migration,and are increasingly viewed as an important factor in promoting economic development and poverty reduction.Worldwide,the number of people living outside their country of birth has more than doubled to almost 200million since 1975,representing about 3%of the world’s population.Global remittances have grown steadily over the past several years,with formal remittance ?ows in 2003—which grossly underestimate to-tal remittances—being second only to foreign direct investment as a capital ?ow into developing countries (Ratha et al.,2010;Ratha,2003).In addition,as countries develop and rural areas become less attractive to new generations,an increasing number of rural individuals are seeking better employment opportunities in urban centers.Migrant and remittance ?ows,however,have also been impacted by the 2009recession.After years of steady growth,remittances to developing countries declined by 6%in 2009although they are expected to recover in 2010and to continue to grow in 2011(Ratha et al.,2010).

In the development economics literature,there has been little attention to the interface between these two major phenomena,migration and nutrition.Migration,whether domestic or interna-tional,can in?uence nutrition through a number of channels.Remittances from migration may have a positive and direct income effect on food consumption and the use of nutrition services.They can also have a positive indirect income effect through the poten-tial relaxation of binding liquidity and insurance constraints and subsequent impact on production and investment decisions.How-ever,migration may also impact the allocation of household labor to both productive and reproductive activities,including child care time and home production.These latter effects may be substantial,especially with widespread female migration and in the absence of adequate surrogate child caretaker services within the household

?Corresponding author.Address:Development Research Group,The World Bank,

Via Labicana 110,00184Rome,Italy.Tel.:+390677710203;fax:+39067096046.

E-mail address:azezza@https://www.wendangku.net/doc/fa6129443.html, (A.Zezza).

or the community.Furthermore,migration is likely to affect nutri-tional habits back home through exposure to different diets and health practices in destination countries,which may have both positive and negative effects on the quantity and quality of food consumption as well as other nutritionally-relevant behavior.The overall effect of migration on nutrition is thus dif?cult to sign a pri-ori and depends on idiosyncratic conditions at the individual, household and community level.

The objective of this special issue is to present state-of-the-art analyses of the link between migration and nutrition of migrant-sending households in developing countries.This is particularly relevant given recent worsening economic condition in destination countries which is likely to affect both the propensity to move—whether to move at all and if to move across destination countries or back to origin countries—as well as migrants’propensity to re-mit.The propensity to remit may in fact be countercyclical although it depends on the economic situation in both the destina-tion and the point of origin.The?ndings of the seven papers in the special issue,both individually and as a set,are relevant for inform-ing food and nutrition policies within developing countries as well as policies designed to assist the?ow of remittances to migrant-sending households.The individual research papers analyze,both theoretically and empirically,the different channels through which different types of migration affect nutrition-linked outcomes in origin households.

The in?uence of migration on food consumption and nutrition For migrant-sending households,the main changes that occur as a result of a migrant’s departure are that the household(i)has the potential to receive remittances from the migrant which can have direct and indirect effects on consumption,(ii)has the possi-bility of obtaining information from the migrant,and(iii)has fewer members leading to lower consumption requirements but also less family labor available.The overall in?uence of these changes on the food consumption and nutrition of the household can be posi-tive or negative and here the channels through which these im-pacts occur are considered.

Through remittances,migration can generate a positive,direct income effect on the sending household,raising the household’s ability to access important nutritional inputs like food,sanitation facilities and health services.The income elasticities of these nutri-tion-related expenditures can vary considerably and the empirical evidence on the subject,although rather rich,is still quite contro-versial.1Furthermore,given the fungibility of remittances,it is not clear whether spending behavior on nutrition-related inputs follow the same pattern of income from other sources and may depend on who controls this income within the household.Of course,the process of migration of household members,particularly the migra-tion of adult males,may alter intrahousehold dynamics changing the manner in which income is managed.The overall direct impact of remittances on nutrition-related expenditures may depend on who has access to the remittance income.

In addition to this direct income effect on consumption,remit-tances may indirectly affect household nutritional outcomes by relaxing insurance and credit constraints for productive and hu-man capital investment.The role of migration and remittances in relaxing credit constraints and providing insurance mechanisms has been widely recognized in the New Economics of Labor Migra-tion(NELM)literature.From this perspective,the decision to mi-grate may be considered a joint household decision if the household uses migration as one mechanism for diversifying risk and gaining access to capital in the presence of market imperfec-tions in the credit and insurance markets(Stark and Bloom, 1985;Stark and Levhari,1982).The migration decision is thereby viewed as an active attempt by households to overcome market imperfections in order to enhance the ability of the household to purchase consumption items or productive inputs and to invest in agriculture and non-farm businesses or,in human https://www.wendangku.net/doc/fa6129443.html,r-mation can also be provided by migrants on productive activities, thus potentially altering the choice of these same activities,or improving skills,thereby increasing the indirect income effects of migration on nutrition.

Migration,through increased exposure to improved practices, may enhance the migrant household’s knowledge of health and nutrition.Migration exposes migrants to a new type of diet,which may or may not be nutritionally bene?cial.The migration experi-ence can alter not just the quantity of food consumed but the com-position,potentially shifting a household’s nutritional habits. Although expected to be for the better,it may include poorer nutri-tion habits,such as greater consumption of high-calorie,and low-nutrient foods.Migration exposure can also supply information on child care practices and generally improve health knowledge.

The sending of a migrant means the loss or reduced presence of one or more members of the household.On the consumption side this clearly means fewer mouths to feed and to support in other ways.On the production side,migration means the loss of labor and,in fact,the negative consequences of migration on nutrition are likely to come through this labor loss.Migration,particularly if long term,reduces the labor endowment of the household,with potentially negative effects on food production and income gener-ation.Furthermore,migration may reduce the time and quality of child care,either because of a migrating mother,or because the mother,often the primary care taker,must take on additional responsibilities as a result of the migration of a spouse or other household member.If labor markets function perfectly,and the household is not credit constrained,the migrant-sending house-hold should be able to replace that labor through hiring and there-fore there should be little effect.Alternatively,in the event of labor or credit market imperfections whereby the household is unable to hire labor or that labor is not a perfect substitute for family labor, there may be negative consequences on household production. This may be particularly true for child care and home production for which labor market imperfection is the norm.

Although dif?cult to measure,there are also important psycho-logical and social problems brought on by migration,such as miss-ing parents,separated families,and deserted communities,which can also have potentially negative bearings on nutrition.Migration may cause a demographic shift both within the household and at the community level,resulting in increased female headship and higher dependency ratio,with the least productive members left behind.The observed increase in female headship in high-migration communities may actually result in a potentially posi-tive effect,as women,if in control of the household budget,are found to spend more on items such as food,education and health than their male counterparts(Quisumbing and McClafferty, 2006).If this is true,the observed shifts towards female headship, whether de jure or de facto,following the migration of a male spouse,may indeed bear positive effects on nutrition-related expenditures and possibly on nutritional outcomes.

Data and methodological challenges

Establishing the relationship between migration and nutrition is not a trivial undertaking due to complications in de?ning and measuring each variable as well as establishing a causal link between the two.Unless data are speci?cally collected for the

1See,for example,among many others,Behrman and Deolalikar(1987),Bouis and

Haddad(1992),and Subramanian and Deaton(1996).

2 A.Zezza et al./Food Policy36(2011)1–6

purpose of establishing this link,data limitations will affect the way variables are de?ned and the possible methodological ap-proaches.These issues are brie?y noted here as is the manner in which these are addressed in the papers in this special issue.

Measuring nutrition

To establish a link between migration and nutrition requires ?rst de?ning indicators that allow a measurement of both phe-nomena.Nutrition measures include outcomes,particularly anthropometric measures,and inputs,consisting of different mea-sures of kilocalories and micro-and macro-nutrient intake,as well as health,sanitation,care and feeding practices that have a bearing on those outputs.Anthropometric measures can be examined for both adults and children,but are more often measured for children since they are considered informative indicators of nutritional sta-tus as well as long-term health status(Strauss and Thomas,1998). Common indicators of nutrition include anthropometric measures of height-for-age,weight-for-age,weight-for-height or Body Mass Index(BMI)for age and related measures such as stunting,wasting and under-or overweight.Different measures are used because they can measure different dimensions of nutritional status.For example,height-for-age is considered a better long-term measure of nutrition while weight-for-age captures shorter term effects. When available multiple measures are then used.

Of the papers in the special issue?ve use some form of child anthropometric measure as an indicator of nutrition.Height-for-age measures are used in the studies in Tonga,El Salvador,Guate-mala and Tajikistan.The Guatemala paper also examined stunting which is measured has a height-for-age z-score that is more than two standard deviations below the median of the reference popu-lation and captures sustained episodes of malnutrition.The Tonga and El Salvador papers also use measures of weight-for-age and the El Salvador paper also explores weight-for-height measures.Final-ly,the Tonga paper examines BMI-for-age while the paper on China examines being overweight or underweight according to a BMI.Note that there is some variation across the papers in the age categories considered including those looking at children under three and those examining those under?ve.These differ-ences largely depend on the available data.

There are two reasons to examine input indicators of nutrition directly.First,two of the data sets(Ghana and Vietnam)did not collect anthropometric measures.Second,three of the papers(Ton-ga,China and Tajikistan)explore the mechanisms by which changes in nutritional status occur.Measures of energy intake used in this set of papers include food expenditures per adult equivalent as well as calorie consumption per adult equivalent(in both cases in total and by food category).These measures are widely used since food consumption information tends to be available in mul-ti-topic household surveys such as the Living Standards Measure-ment Study surveys supported by the World Bank.Some papers also include other measures such as food diversity(Vietnam),time allocation to activities that might be linked to child nutrition(Chi-na)and breastfeeding(Tajikistan).Although on their own these measures cannot completely establish the link between migration and to nutrition,they are indicative of the type of effects that migration is expected to have and which would ultimately in?u-ence nutrition.

Measuring migration

With respect to migration measures,a key initial issue for the current study is whether it is migration or remittance receipt that should be measured since,as noted in the conceptual discussion, much of the effects of migration are expected to be through remit-tances.There are three problems with using remittances as the measure.First,as noted in the conceptual section,is that migration can have impacts beyond remittances including through informa-tion?ow and changes in family labor availability and consumption requirements.Second,remittances are dif?cult to measure accu-rately since they are often sent irregularly and through in-kind transfers.Even when cash is sent determining exact amounts is of-ten dif?cult.This often leads to substantial measurement error in remittance variables.Third,the lack of receipt of remittances does not mean that the household has never received remittances in the past,or will not receive in the future.Migration status,on the other hand,indicates a potential to receive remittances or other types of in?uence.

Given the focus on migration,the key issue for these studies is in de?ning what a migrant household is.Ultimately,the interest is in identifying households that could potentially be in?uenced by the existence of a migrant that is,or was,a household member or other close connection.Operationalizing this de?nition,though, is not always straightforward since there are several modes of migration and these modes are linked to different motivations and are likely to result in differential impacts.de Brauw and Carletto(2010)identify?ve possible characteristics for determin-ing which individuals could be classi?ed as migrants and the type of migration:place of birth,whether or not the individual resides in the place of birth,household membership,the duration of any stays away from the residence,and a time period of reference.As this special issue deals exclusively with emigrants(as opposed to immigrants)and the impact on the household of origin,some of these concepts are more relevant than others.

Using some of these characteristics,two key types of migrants are distinguished in the papers in this special issue:international versus domestic migrants and long-term versus short-term mi-grants.De?ning an international migrant is straightforward since it involves leaving the country,but in de?ning a domestic migrant a decision must be made over what type of move makes someone a migrant—that is,whether a person who moved to a nearby town is considered a migrant or only if those who have moved further away.Distinguishing long-and short-term migration may also be important since they may be motivated by different considerations and,more importantly,result in different returns.A short-term mi-grant may be seasonally migrating on a regular basis to supple-ment income or temporarily migrating to respond to some shock.

A long-term migrant is likely to be more permanent and the in?u-ence then may depend on the relationship the migrant maintains with the household.Distinguishing long-and short-term migration becomes tricky since it requires coming up with often arbitrary assumptions on the length of time associated with each type of migration.The distinction is particularly dif?cult for recent mi-grants,whose planned length of stay away from the household is unknown.Once this is de?ned,there is also an issue of the time period over which migration is to be considered.Is a relation who left ten years prior still likely to in?uence the household? The expectation is that more recent migrants are likely to have more of an in?uence than those who left years ago.On the other hand,migration events that are too recent may have yet to bear any signi?cant bene?ts to the household of origin.

In some cases,migration may also be distinguished based on the reason for https://www.wendangku.net/doc/fa6129443.html,ually the interest in making this dis-tinction is in identifying departures for employment purposes rather than for marriage,education or other reason.The motivation for making this distinction is the perception that migration for employment may have a greater in?uence on the household rela-tive to other forms of migration.

Differences in the approaches taken in the papers in this issue depend on these factors,as well on the way data were collected in each country.In the cases of Tonga,El Salvador,Guatemala and Tajikistan the focus of the study is on international migration

A.Zezza et al./Food Policy36(2011)1–63

while in China and Vietnam it is domestic migration and in Ghana mostly domestic migration.The period over which migration is studied is?xed over time in nearly all cases.For Tonga and Guate-mala,speci?c calendar periods are used(2002–2005and2005–2008respectively).For China and Vietnam,migration de?nitions use the panel nature of the data to identify who left between rounds of data collection and whether it was for employment. The Vietnam paper uses this information for long-term migration (more than12months)and de?nes separately short-term migra-tion from questions on how long someone was outside the house-hold during the year and for what reason(less than12months). The Tajikistan study combines information on current migration (individuals working abroad who would otherwise be considered part of the household)and previous migration(past migration epi-sodes longer than one month by current household members). Along similar lines,Ghana has a migrant module which includes information on current and recent migrants.Finally,El Salvador de?nes migrant household as those who have migrants which have kept in regular contact with the household.

Attributing the in?uence of migration on food consumption and nutrition

The primary issue of identifying the in?uence of migration on nutrition comes down to establishing causality.Migration is not a random process,but is rather a choice that is in?uenced by a range of factors.In fact,it may be that nutritional outcomes,or concerns about such outcomes,may drive individuals to migrate or households to send a migrant;meaning there is a possibility that there is reverse causality in that nutritional outcomes drive migration rather than vice versa.Additionally,there could be traits of migrants,or their households,that in?uence the behavioral choices with respect to consumption and thus nutrition that also in?uence their choices with respect to migration.Any analysis of the link between migration and nutrition that shows a relationship may be capturing the fundamental characteristics of migrant households rather than the in?uence of migration.If these traits are observable,methods may be used to try to deal with these selection issues.However,if they are unobservable,there may re-main issues of omitted variable bias in any estimates of the relationship.

These issues of establishing causality are common in empirical analysis and different approaches may be used to address them.In four of the papers(Tonga,Vietnam,Ghana and Tajikistan),instru-mental variable(IV)approaches are used.In these cases,an instru-ment is found that is in principle linked to migration but not to the nutritional outcome of interest.Since in theory the instrument does not have the same endogeneity problems as the migration variable,the problem of establishing causality is addressed.Of course,the validity of this approach rests largely on the validity of the instruments.The most innovative of the papers in the use of instruments is the Tonga paper which takes advantage of the fact migrants to New Zealand are selected by a lottery.Since not all those selected take advantage of the opportunity to migrate,self selection is still an issue and thus a potential problem with endo-geneity.However,the random assignment of migrants via the lot-tery creates the perfect instrument since it is clearly linked to migration and exogenous.The other papers that use IV approaches use information on a combination of previous demographic changes and previous migration behavior.Details of these ap-proaches are found in each of the papers.

The other three papers use alternative methods to deal with concerns about causality.In the case of Guatemala,a double differ-ence approach that takes advantage of an understanding of child growth patterns is used.By using a double difference,the authors can compare trends in nutrition indicators in migrant and non-migrant households and thus‘‘difference out’’any unobservable differences between migrants and non-migrants.The paper on China has four rounds of data and is able to incorporate individual ?xed effects to control for unobservable differences in migrant and non-migrant households not addressed with other control variable. In the case of El Salvador,the author is careful to avoid claiming causality,but runs a series of robustness checks that support the legitimacy of the obtained results.

The papers in this special issue are therefore valuable as a set,in that they provide the reader with a range of possible methodolog-ical approaches to measure migration and nutritional outcomes, and to establish to the degree possible a causal relationship be-tween the two variables.

Synthesis of results

The papers in this special issue offer substantial new empiri-cal evidence to add to the(thin)existing body of evidence on the relationship between migration and nutrition and the chan-nels through which it operates.Since the case studies draw on data from very different settings,and look at very different types of migration(international,urban–rural,entire households,indi-vidual migrants,short-and long-term),it is not surprising that their?ndings are also diverse.To begin,nutrition outcomes linked to anthropometric measures are considered followed by inputs.

In the cases where measures of child anthropometry are avail-able,the evidence appears to point to improvements in child growth associated with https://www.wendangku.net/doc/fa6129443.html,rge impacts are found in the very poor,high malnutrition,high-migration communities in-cluded in the Guatemala study.Here households with migrants to the US have height-for-age z-scores(HAZ)that are0.5stan-dard deviations higher than other children,which translates into a6percentage points lower stunting rate.Signi?cant effects on HAZ are also found in the Tajikistan case study and in El Salva-dor,where migrant households were able to withstand the food price shock much better than non-migrant households,so that their children experienced a much more contained drop in HAZ scores following that shock.The Tonga study does not provide such clear results but suggests a difference in the impact of migration on movers(children who moved to New Zealand with their families)versus stayers(children left behind after the migration of other members of the household to New Zealand). Weight-for-age and height-for-age z-scores increase for movers but decreases for stayers.Unlike the other studies,there does not appear to be nutrition bene?ts for those that are linked to migrants.The China case study is somewhat different in that it looks at the impact of child growth in an older age group(2–12years old)using a different indicator(the Body Mass Index, BMI).The?ndings in this case study are somewhat mixed but overall positive:migration is associated with lower overweight rates in children in the2–6age group,and lower underweight rates in the7–12age group.

While in a majority of cases the evidence points towards im-proved overall nutrition outcomes for individuals or households following the migration of one or more family members,the evi-dence on the impact on the inputs is more diverse.The input that is analyzed in most of the studies is calorie consumption,with some authors looking also at changes in the composition of the diet.The studies that can trace a positive impact of migration expo-sure on calorie consumption are those on China(but coupled with a lower share of protein-rich foods in the diet),Vietnam(limited to short-term migration),Tajikistan and possibly Guatemala.In Ghana there does not appear to be any signi?cant difference in food expenditure patterns between migrant and non-migrant households,as is the case for households with long-term migrants

4 A.Zezza et al./Food Policy36(2011)1–6

in Vietnam(where there are actually hints of a possible slightly negative impact).

In El Salvador the analysis suggests that following a sharp in-crease in food prices non-migrant households have been able to keep the level of kilocalorie consumption constant but have had to shift to lower quality calories more than migrant households did.Such impacts on dietary quality are not uncommon.The Tonga study is particularly interesting in this respect in its comparison of movers and stayers in a country in which obesity is of particular concern.For the children who have migrated,the move is accom-panied by a shift towards more meat and milk consumption,which is the opposite of what happens to the stayers group.Evidence of a shift towards poorer diets is also found in the Ghana(more sugars, beverages and food eaten outside)and China(lower protein intake, despite more kilocalories).The Vietnam study does not report any signi?cant change in dietary quality as measured by dietary diver-sity indicators.

In terms of the other channels determining nutritional out-comes,issues related to time allocation,and the subsequent changes in child care,meal preparation,and other adult and child household work patterns,appear to play a role in some instances. The apparently contradictory?nding of the China case study that we have just recalled(lower overweight in2–6,lower underweight in7–12years old)is most likely due to the combination of less adult time devoted to meal preparation,combined with the greater involvement of older children in the carrying out of household chores.A related interesting?nding from the China case study con-cerns the impact on child nutritional status of the quality of the surrogate care they receive.For instance,children of migrant par-ents who are left with their grandparents are less likely to be over-weight than those left in other circumstances.

The Guatemala study also contends that the decline in vaccina-tion coverage observed in children of migrant households may be linked to the increased opportunity cost of time for remaining household members associated with migration.On the other hand, their analysis seems to point to lower incidence of diarrhea which is a common deterrent to normal child growth.This may be due to an improved home environment and sanitation and it may be an additional mechanism besides increased expenditure by which migration in?uences child growth.Their descriptive statistics also suggest migrant households are able to spend more towards health care.

What we have learned and what remains to be understood Taken together,the evidence presented in the papers in this special issue points to a broadly positive impact of migration on nutrition.Within this overall picture,however,the cases studied also raise some?ags on issues that are of possible concern regard-ing aspects such as worsening dietary habits,and increased scar-city of adults’time,resulting in less time potentially devoted to some health and care activities,or in children being increasingly burdened with household chores.The self-evident implication of such?ndings for decision makers concerned with food security and nutrition issues in high migration areas is therefore to attempt to maximize the positive impacts of migration while minimizing its possible negative consequences.This is clearly not very helpful and in what follows we try to articulate the implications of this re-search into more meaningful considerations.Generalizations are, however,especially dif?cult in this case given the diversity of con-texts and migration phenomena covered by the special issue.

Several of the case studies in this issue point towards markedly higher growth for children in migrant households.Migration should therefore be perceived as an opportunity rather than a threat,and policies should be directed at facilitating remittance ?ows and reducing the costs of migration so as to maximize these positive impacts.Furthermore,while migration is obviously no substitute for effective food and nutrition security policies and pro-grams,the latter could work creatively to?nd ways to create syn-ergies between the?nancial,capital and knowledge resources mobilized by the programs and similar?ows connected to migration.

Migrants appear also to be better able to withstand food related shocks,such as a sudden increase in food prices.While based on just one of the cases presented in this issue,this is a?nding of potentially much broader relevance.Where possible,the scarce re-sources available to support vulnerable households in the event of such shocks may be better allocated by taking into consideration the fact that migrant households are likely to be already at least partly‘self-insured’against such shocks.

The?ndings also suggest that often the exposure to different, and not necessarily better,dietary habits associated with migra-tion,coupled with a reduction in the time available for food prep-aration,can lead to a worsening in the composition of the diets, for instance via a shift towards foods containing higher levels of fats and sugar,or lower levels of fresh fruit and vegetable con-sumption.In such case,particularly where the incidence of being overweight and obesity are on the rise,nutrition education pro-grams in high migration areas should exercise particular care in taking into account the possible negative signals being channeled by migrants.

Furthermore,a non-negligible portion of the?ndings points to-wards negative impacts of migration on the food security and nutrition equation.For instance,there is some evidence(Tonga) of increased obesity threats among migrant children,of reduced rates of vaccination among children of families with migrants (Guatemala),or of increased demand on children’s time for house-hold chores(China).These phenomena are so diverse that it is impossible to propose one speci?c action to address them,but do point to the need for health and nutrition advisory services in high migration area to be alert about the possible emergence of such threats.

Concerning the data and methodological challenges outlined above,the papers included in this issue all had to deal with data limitations that somehow complicated the task of identifying the causal relationship between migration and nutrition.Also,data availability was a major constraining factor for the analysis of the possible mechanisms through which the impacts of migration play out.Although the availability of migration related micro data has greatly increased in the last decade,these limitations clearly point to the need to still?ll several outstanding gaps.

Sub-optimal approaches to de?ning and identifying migrants are still found in many survey instruments,despite migration speci?c modules having been introduced in multi-topic surveys for some years now,and guidelines for data collection having been codi?ed and documented.It would seem advisable that at least a standard short module to identify migration patterns be introduced in integrated household surveys in all countries were migration constitutes a sizeable social and economic phenomenon.

The more thorny methodological challenges are,however,those around the identi?cation of causal relationships and the unpacking of the mechanisms through which the impacts of migration mate-rialize.Unlike other development issues,the potential for experi-mental survey design to contribute in the?eld of migration studies is limited as migration cannot be randomized,except for the exceptional cases in which it is linked to a lottery(as in the Tonga paper in this issue).The increasing availability of panel data holds promise of helping researchers push the agenda further con-cerning the identi?cation of causal relationships,but it will not re-solve all identi?cation problems.Unpacking the mechanisms

A.Zezza et al./Food Policy36(2011)1–65

through which migration operates requires expanding the cover-age of the topics included in survey modules,which may not al-ways be practical in nationally representative surveys,but may be a task for purposely built surveys.

Acknowledgements

The authors would like to thank Manuela Angelucci,Kathleen Beegle,Philippe De Vreyer,Francois Kabore,David Parker,and Valeria Perotti for their participation in the workshop‘‘Assessing the Impact of Migration on Food and Nutrition Security’’held at the FAO in Rome on22,23June2009.The intellectual,?nancial and logistic contribution of the FAO’s Agricultural Development Economics Division,under the leadership of Kostas Stamoulis,is gratefully acknowledged.The support of Giorgio D’Amore,Silvana Leone and Eliana Marino at ICstat was key to the successful com-pletion of this project.We would also like to thank Colin Poulton for his patience during the entire editorial https://www.wendangku.net/doc/fa6129443.html,st but not least,we would like to thank‘‘They-Who-Must-Not-Be-Named’’(i.e.the anonymous reviewers)for punctual and careful feedback on the papers.The views expressed in this special issue are the authors’alone and should not be attributed to their respective organizations.References

Behrman,J.R.,Deolalikar,A.B.,1987.Will developing country nutrition improve with income?A case study of rural South India.Journal of Political Economy95, 108–138.

Bouis,H.,Haddad,L.J.,1992.Are estimates of calorie-income elasticities too high?A recalibration of the plausible range.Journal of Development Economics39, 333–364.

de Brauw,A.,Carletto,C.,2010.Improving the Measurement and Policy Relevance of Migration Information in Multi-Topic Household Surveys.Manuscript. FAO,2010.The State of Food Insecurity in the World.Addressing Food Insecurity in Protracted Crises.Food and Agriculture Organization of the United Nations, Rome.

Quisumbing,A.,McClafferty, B.,https://www.wendangku.net/doc/fa6129443.html,ing Gender Research in Development.

International Food Policy Research Institute,Washington,DC.

Ratha,D.,2003.Worker’s remittances:an important and stable source of external development?nance.In:World Bank,Global Development Finance:Striving for Stability in Development Finance,vol.I:Analysis and Statistical Appendix,The World Bank,Washington,DC,pp.157–175.

Ratha,D.,Mohapatra,S.,Silwal,A.,2010.Outlook for Remittance Flows2010–11.

Migration and Development Brief No.12,July.The World Bank,Washington, DC.

Stark,O.,Bloom,D.E.,1985.The new economics of labour migration.American Economic Review75(2),173–178.

Stark,O.,Levhari,D.,1982.On migration and risk in LDCs.Economic Development and Cultural Change31(1),191–196.

Strauss,J.,Thomas,D.,1998.Health,nutrition,and economic development.Journal of Economic Literature36(2),766–817.

Subramanian,S.,Deaton,A.,1996.The demand for food and calories.Journal of Political Economy104(1),133–162.

6 A.Zezza et al./Food Policy36(2011)1–6