Food Policy 67 (2017) 106–118 Contents lists available at ScienceDirect Food Policy journal homepage: www.elsevier.com/locate/foodpol Agricultural commercialization and nutrition revisited: Empirical evidence from three African countries Calogero Carletto a,⇑, Paul Corral b, Anita Guelfi c a Development Data Group, World Bank, United States b Poverty and Equity Global Practice, World Bank, United States c University of Rome ‘Tor Vergata’, Italy a r t i c l e i n f o a b s t r a c t Article history: The transition from subsistence to commercial agriculture is key for economic growth. But what are the Available online 3 November 2016 consequences for nutritional outcomes? The evidence to date has been scant and inconclusive. This study contributes to the debate by revisiting two prevailing wisdoms: (a) market participation by African small- JEL classifications: holders remains low; and (b) the impact of commercialization on nutritional outcomes is generally pos- 012 itive. Using nationally representative data from three African countries, the analysis reveals high levels of 013 commercialization by even the poorest and smallest landholders, with rates of market participation as Q10 high as 90%. Female farmers participate less, but tend to sell larger shares of their production, conditional Q12 Q18 on participation. Second, we find little evidence of a positive relationship between commercialization and nutritional status. As countries and international agencies prioritize the importance of nutrition-sensitive Keywords: agriculture, better understanding of the transmission channels between crop choices and nutritional out- Agricultural commercialization comes should remain a research priority. Cash crops Ó 2016 The World Bank. Published by Elsevier Ltd. This is an open access article under the CC BY IGO Nutrition license (http://creativecommons.org/licenses/by/3.0/igo/). Africa 1. Introduction market orientation.2 The concerns related especially to their food security and nutritional outcomes.3 While many of these studies dis- According to conventional wisdom, the transition from subsis- played a pronounced degree of ideology,4 they also highlighted the tence (or semi-subsistence) to commercial agriculture represents need to better understand the underlying linkages between crop a key ingredient for the economic development of low-income production, commercialization, income, consumption and nutrition countries. By exploiting comparative advantages, agricultural com- at the household level. mercialization enhances trade and efficiency, leading to economic Against this background, the International Food Policy Research growth and welfare improvement at the national level. This is fur- Institute (IFPRI) revisited the issue,5 using a more scientific and sys- ther expected to initiate a virtuous cycle which raises household tematic approach which consisted of three components: (i) the income, thus improving consumption, food security and nutri- development of a conceptual framework articulating the linkages tional outcomes inside rural households. Yet, this mainstream, beneficial view of agricultural commer- 2 For a quick overview of the several areas of debate of agricultural commercial- cialization has also been challenged several times since the ization over time, see for example Maxwell and Fernando (1989) or the more recent 1970s, with a large body of literature in the 1970s and the first half Wiggins et al. (2011). of the 1980s1 emphasizing the adverse effects on households’ wel- 3 Wiggins et al. (2011) mention that this may be somewhat overblown, since in fare and nutrition, especially on the poorest groups of the rural pop- most cases small farmers tend to prioritize growing their main staple food. 4 ulation and the most vulnerable individuals within the household Agricultural commercialization was often presented as the result of colonialist- type rural policies, favoring ‘‘cash crops” mainly for export reasons with minimal who are often considered unable to reap the benefits of increased advantages for the rural population. This line of argument was favored by researchers supporting the so-called ‘‘food-first” view. For a more detailed review of this line of argument against cash cropping and related sources, see Maxwell and Fernando (1989), Appendix A. ⇑ Corresponding author. 5 Other relevant research projects were also carried out in this period by the E-mail address: gcarletto@worldbank.org (C. Carletto). Department of Agricultural Economics of the Michigan State University. See for 1 See for instance Hernandez et al. (1974), Lappe (1977), as well as Dewey (1981). instance, Lev (1981). http://dx.doi.org/10.1016/j.foodpol.2016.09.020 0306-9192/Ó 2016 The World Bank. Published by Elsevier Ltd. This is an open access article under the CC BY IGO license (http://creativecommons.org/licenses/by/3.0/igo/). C. Carletto et al. / Food Policy 67 (2017) 106–118 107 between commercialization and nutrition; (ii) a better research household and crop levels. Section 5 descriptively explores the design to compare commercialized and non-commercialized house- relationship between agricultural commercialization and nutri- holds; and (iii) the use of a cross-country comparative approach tional outcomes. Then the section presents an econometric strat- based on six different but comparable country micro-level analyses6 egy and the main findings. Finally, conclusions are presented in carried out using a common research design. The IFPRI studies also Section 6. mitigated the traditional assumption of a dichotomy – and hence a necessary competition – between cash and staple crops, which had deeply influenced the way agricultural commercialization had been 2. Agricultural commercialization and nutrition: a brief conceived and measured in most of the previous literature.7 literature review Unlike many of the previous studies, the majority of IFPRI coun- try studies found generally a positive, though small, impact of agri- The empirical literature on the nutritional outcomes of agricul- cultural commercialization on the nutritional status of rural tural commercialization can be grouped into three strands: (i) a households, where the positive relationship was assumed to oper- wide and heterogeneous set of research projects carried out before ate primarily through the linkages between household income, the launch of the IFPRI agenda; (ii) the IFPRI work between 1986 household caloric intake, and child caloric intake. Nevertheless, and 1994; (iii) a few studies devoted to the topic starting from as the authors of the studies acknowledged, several limitations the early 1990s. remained: ‘‘Econometrically, a common practice is to estimate a A review of the first wave of studies fails to settle the debate on set of reduced form equations with an extended list of exogenous the linkages between agricultural commercialization and nutrition. explanatory variables that affect any of the structural relations. As shown in Table 1 (which reports the literature review carried This approach is not followed in this book, in part because of data out in von Braun and Kennedy, 198610), results are confusing and limitations (von Braun and Kennedy, 1994; Ch. 2, p. 24).” ambiguous, with the same crop having opposite effects both Since then, there has been little new empirical evidence8 on the between and within countries. Studies in this period usually lacked links between agricultural commercialization and nutrition,9 despite a proper conceptual framework, adopting instead a ‘‘black-box” the implementation of numerous expensive projects to promote approach which did not articulate the underlying channels leading market-oriented crops, based on the assumption of a beneficial to various outcomes. The main approach was a comparison of nutri- nutritional effect. tional outcomes between cash crop adopters and non-adopters. The In the spirit of the other papers in this volume, this study revis- evidence was often anecdotal and based on country case studies, its two prevailing wisdoms. First, participation in market activities making it impossible to compare results both across and within by smallholders is low. Second, the impact of commercialization on countries. In most studies, the definition and measurement of com- nutritional outcomes is generally positive. In doing so, the paper mercialization was subjective (based on the adoption or non- reconsiders the quantification and characterization of agricultural adoption of a given list of cash crops). commercialization and provides new, systematic evidence on its Subsequently, the IFPRI studies also developed a conceptual relationship with nutritional outcomes in three Sub-Saharan coun- framework to articulate the complex set of linkages between the tries. In particular, it uses recent panel surveys from Malawi, Tan- process of agricultural commercialization and the nutritional and zania and Uganda conducted under the Living Standards health status at the household level. In particular, they examined Measurement Study-Integrated Surveys on Agriculture (LSMS- how agricultural commercialization affected each of the four key ISA) program. Unlike in most previous studies, these surveys are steps between national food production and individual nutritional nationally representative, which enables a more systematic com- outcomes, identified by Pinstrup-Andersen in the early 1980s,11 i.e. parison across different settings and also allows one to better con- ‘‘national/community food availability”, the ‘‘ability and desire of trol for a number of the endogeneity issues that arise in estimating households to obtain food”, ‘‘intrahousehold food distribution” and the impact of commercialization on nutritional outcomes. The ‘‘health and sanitary factors”. study further aims to capture the heterogeneity implicit in the First, the decision to adopt a market-oriented production sys- commercialization choices of different smallholder households. tem is expected to influence the degree of food availability at the For example, Duflo and Udry (2004) suggest that income from dif- national, community and household levels. Factors such as compe- ferent crops as well as income from different plot owners may tition among limited resources (such as land, labor and capital), the serve distinct purposes within the household and thus have differ- amount of food imports and aid, the degree of diversity of available ent impacts. Using individual-level crop data, we are able to differ- foods and the presence of seasonal and irregular fluctuations may entiate the impact of commercialization based on the gender of the be influenced by a rise in market orientation in smallholder farm- farmers and the type of crop mix grown and sold, which are both ers. Through this channel they may impact national or regional assumed to affect the relationship between commercialization food availability, which, by affecting food prices, may have relevant and nutritional outcomes. nutritional implications. However, national food sufficiency can be The paper is organized as follows. Sections 2 and 3 provide a a poor indicator of household nutritional status, as ‘‘food may be brief overview of the literature and a short description of the data, plentiful but the poor may still be unable to access it”.12 Thus, at respectively. Section 4 profiles commercialization in the three the household level, it is important to look at the ability of each countries by constructing an index of commercialization at the household to effectively obtain food.13 This ability varies depending 6 10 Previous literature was limited in its scope by the available data, since A wider literature review on studies conducted before the mid-1980s was carried information had been mainly available in aggregate form. out six years later by Randolph (1992). It showed similar results. 7 11 See in particular the empirical results provided in von Braun and Kennedy (1986). See in particular Pinstrup-Andersen (1983). 8 12 The IFPRI research agenda on agricultural commercialization and nutrition von Braun and Kennedy (1986). 13 stretched from the mid-1980s to the mid-1990s. In this paper, we use different anthropometric measures of children under five 9 The study by Wood et al. (2013) is a notable exception. Others (Carletto et al., years of age to compute measures of stunting, wasting and underweight as well as 1999, 2010, 2011) focused more on the determinants of the commercialization associated Z-scores capturing deviations of sample children from a reference process and its impact on poverty, as opposed to food security and nutrition. population. 108 C. Carletto et al. / Food Policy 67 (2017) 106–118 Table 1 Summary of micro-studies on income and nutritional effects of cash crop production reviewed in von Braun and Kennedy (1986). Source: von Braun and Kennedy (1986), table 16, page 47. Study Country Crop Effects on Family consumption Nutritional Status Lev (1981) Tanzania Coffee, Bananas Positive n.a. Hitchings (1982) Kenya Tea, Coffee, Cotton, Pyrethrum, Sugarcane n.a. Positive, Positive, Neutral, Neutral, Negative Rabeneck (1982) Kenya Coffee, Staples n.a. Positive Fleuret and Fleuret (1983) Kenya Coffee, Vegetables n.a. Positive Gross and Underwood (1971) Brazil Sisal Negative n.a. Dewey (1981) Mexico Cocoa, Sugarcane n.a. Negative Hernandez et al. (1974) Mexico Cocoa, Sugarcane n.a. Negative Lambert (1978) Papua New Guinea Coffee Negative n.a. Harvey and Heywood (1983) Papua New Guinea Coffee Positive Positive on the effects of the commercialization process on several factors, intake. Cash crop adoption generally increased real incomes, among which the most important one is household income.14 thereby stimulating a virtuous cycle whereby higher incomes If real income increases at the household level, it could then were used to increase food consumption, which benefited on stimulate a virtuous cycle through which smallholder farmers average both the household in general and the children in can enhance their level of food consumption. While necessary, particular. the rise in real income, is again not sufficient to improve household (ii). The effect of agricultural commercialization on nutrition will consumption. Indeed, households must have the ‘‘desire to obtain depend on a number of conditioning, complementary factors available (and nutritious) food”, a condition which is often not sat- both at the macro and micro level, making the adoption of isfied due to intra-household factors, as individual household commercial crops more or less remunerative and sustain- members are likely to have different income elasticities.15 Further- able. The complex set of linkages characterizing the com- more, even if additional income is spent on food, intra-household mercialization process and its impact on household food consumption could be heterogeneously distributed among fam- welfare and nutrition suggest that several different scenarios ily members, with children and women often relatively penalized can emerge depending on the factors dominating in each cir- compared with adult males. Furthermore, a high marginal propen- cumstance. According to the study, a key role must be sity to spend on food does not automatically imply a high marginal played by policies (both macro and micro19) aiming at propensity to consume calories. Households may choose to ‘‘diver- enhancing beneficial outcomes while minimizing adverse sify into a more varied, higher cost diet rather than simply using ones.20 the income to increase energy intake”.16 Finally, a crucial role is played by the potential impact of changes in income on health and Two decades after the publication of von Braun and Kennedy sanitary factors, as increased income from commercialization may (1994), the somewhat positive view on the impact of commercial- be invested towards improved water sources and/or sanitation both ization of agriculture on welfare outcomes is still prevalent. In fact, at the household and community level. since then, there have been only few new studies explicitly looking Two key results emerged from the IFPRI studies17: at the link between agricultural commercialization and nutrition and the evidence remains inconclusive.21 The present study is an (i). The impact of agricultural commercialization on the nutri- attempt to shed some light on this rather controversial, yet impor- tional status18 of rural households was found to be mostly tant, relationship using data from three countries in sub-Saharan positive, though rather small in magnitude. This positive rela- Africa. tionship mainly operated through the linkages between household income, household caloric intake and child caloric 3. Data description 14 To revisit the link between agricultural commercialization and Indeed, ‘‘where farmers are free to make their own production decisions, increased commercialization will occur only if the farmer does not perceive that nutrition, we use the data from the nationally representative another production option would more effectively achieve his goals within the household panel surveys collected in three countries (Malawi, Tan- constraints it faces. Thus, although higher income is only one of a set of possible goals, zania, Uganda) under the Living Standards Measurement Study – it is highly unlikely that a farmer would produce export crops unless he or she expects Integrated Surveys in Agriculture (LSMS-ISA) initiative.22 it to yield economic gains than any other realistic production option” (see von Braun and Kennedy, 1986). 15 19 Additionally it is possible that different household members have different At the macroeconomic level trade policies, market reforms, improvement in rural income elasticities among food products. infrastructure, as well as the development of legal and contractual environments are 16 Von Braun et al. (1989) also observe that in some cases malnutrition is endemic in seen as crucial to promoting an inclusive commercialization process. At the a given country; so that households are not always aware of the problem (‘‘their microeconomic level, instead, two important policy areas are identified in the setting children look like most other children in the community”). up of effective rural financial institutions and in the provision of extension services at 17 Von Braun et al. (1989) and von Braun and Kennedy (1994). The first round of the household level to help farmers to avoid crop management failures. Furthermore, country case studies, summarized in von Braun et al. (1989) included Guatemala, the development and promotion of community health and sanitation services is Kenya, Rwanda, the Philippines, and Gambia. In these countries, rural households had highly recommended to maximize the health and nutrition returns of increased recently undergone a change from semi-subsistence food production to increased income. Finally, promotion of technological change in food crops is advocated in order production of crops for sales, thereby representing an ideal study area of the impact of to enhance food security at the household level. 20 agricultural commercialization. These country studies were subsequently comple- For example when households adopt new crops and do not obtain the expected mented with 5 more field studies in India, Malawi, Papua New Guinea, Sierra Leone gains. This is detailed in von Braun and Kennedy (1994); an example is when there is and Zambia. inelastic demand, yet many producers enter the market and reduce crop prices. 18 21 To be more precise, a positive impact was found in all countries except for Kenya, See also Billah (2002) on the links between crop production and nutrition in where the effect was deemed neutral. A positive impact was recorded also in four of Bangladesh. 22 the five studies carried out in the early 1990s, with the exception of Sierra Leone, See http://www.worldbank.org/lsms. Other countries covered under the program where a deterioration of the nutritional status was detected. See Bellin (1994). include Ethiopia, Mali, Niger, Nigeria. C. Carletto et al. / Food Policy 67 (2017) 106–118 109 All households were administered a multi-topic household 4. Measuring agricultural commercialization in sub-Saharan questionnaire, and those involved in any agricultural activities Africa were also administered a detailed agricultural module. The surveys collect detailed crop and plot level information, as well as a rich set In the previous section, we demonstrated the high incidence of of socioeconomic characteristics and information on child anthro- participation in market activities among even the smallest of pometrics. Given agricultural production estimates at the plot level smallholders. However, how much do those farming households as well as the identification of the plot manager, the level of com- sell? Who are the individuals in the household most involved in mercialization can also be computed at the individual level. selling and what products do they sell the most? How best to The surveys was conducted throughout the year, though each define and measure the actual degree of agricultural commercial- household was only interviewed once. To adjust for this difference ization in a given country has been much debated in recent dec- in the timing of the interview, when calculating the commercial- ades. In this paper we use the Household Crop Commercialization ization index reported sales were annualized using imputation Index (CCI), introduced by Strasberg et al. (1999) and Govereth methods (see Appendix A for details).23 et al. (1999), which is defined as: Given our focus, our sample includes only farming households, defined as households who reported involvement in agricultural CCIi ¼ ½Gross v alue of crop saleshhi;year j = activities through ownership and/or cultivation of land in the most Gross v alue of all crop productionhhi;year j Š à 100 ð1Þ recently completed agricultural season.24 The descriptive analysis presented in this paper in Sections 3–5 are based on the full samples Though not without its own shortcomings,26 a measure on the of the baseline surveys which were carried out in 2010/11 in all output side is able to capture a ‘‘household’s ‘revealed’ marketing three countries. Our final sample at baseline thus consists of 9894 behavior,27” and can be seen as relatively easier to collect,28 while households in Malawi, 2074 households in Tanzania and 1788 lending ‘‘itself well to an empirical test within a regression households in Uganda. The panel component is introduced in Sec- framework29”. tion 5.1 in order to address some of the econometric challenges of According to this measure, the process of agricultural commer- the estimation. After excluding the non-panel and non-farming cialization can be represented by a continuum ranging from pure households, the final sample size of the panel used for the estima- subsistence (CCIi = 0) to a completely commercialized production tions in each country is 2222, 1744 and 1587 farming households system (CCIi = 100). Its main advantage is that it permits to go in Malawi, Tanzania and Uganda, respectively.25 Overall, sample beyond the traditional dichotomies of sellers versus non-sellers, attrition between the two waves was rather low. or between staple and cash crop producers. In fact, it adds an addi- As shown in Table 2, the great majority of households in our tional dimension to the discussion, i.e. how much of their harvest sample are male-headed, with the share of the female-headed households choose to sell - while still being relatively easy to households ranging from 25% (Malawi and Tanzania) to 30% compute. (Uganda). Significant differences emerge in terms of educational For the countries studied here, the CCI amounts to 17.6% in attainment: in Malawi about 78.8% of the rural households did Malawi, 26.3% in Uganda and 27.5% in Tanzania on average not receive any type of formal education, while this percentage (Table 3). When restricting the sample to farm households report- amounts to about 30.2 and 18.8% in Tanzania and Uganda, respec- ing any sales (conditional CCI), it rises slightly to 19.6, 40.4 and 33% tively. In the latter two countries, the large majority of rural house- respectively. Also, the degree of commercialization increases with holds attained at least a primary level of education. farm size, likely reflecting larger surpluses of edible crops and/or Another source of significant variation between the three coun- greater adoption of cash crops by farmers with larger landholdings. tries is the size of land available to farmers. Average land size is Relying on individual-level data, we are also able to compute slightly below 1 ha in Malawi, compared with 2.3 and 2.6 ha CCI measures separately for male and female farmers.30 At first respectively in Uganda and Tanzania. The three countries also look glance, female farmers appear to commercialize considerably less different in terms of crop diversification, with Malawian and Tan- of their harvest than their male counterparts. However, when focus- zanian families choosing to plant 2 types of crops on average, com- ing only on those individuals who report selling (conditional CCI), pared with around 4 in Uganda. the difference disappears or even reverses. In fact, among sellers, Table 2 also details the differences between selling households females appear to be more commercially oriented than their male and non-selling households. We find that close to 90% of house- counterparts both in Malawi (selling 31% of their production vs. 22 holds in Malawi engage in sales, compared to 80% in Uganda and among male farmers) and in Uganda (37 vs. 35). Meanwhile, in Tan- 68% in Tanzania. Not surprisingly, when broken down by land size zania both genders have virtually the same CCI, at 43%. quintiles, the incidence of households selling any crop is monoton- Breaking down our unconditional gender-disaggregated CCI ically increasing, with larger farmers selling greater shares. Overall, measure by farm size31 confirms a positive relationship between these preliminary figures suggest that in all three countries, the commercialization and landholdings for both male and female farm- majority of farm households sell part of their production for both ers. However, the gender gap in commercialization appears to staple and cash crops. In this respect, our data suggests that the increase, particularly in Malawi where among farmers with more vast majority of commercialized households are only producing (and selling) food crops, from a minimum of 79% in Uganda to a 26 See Appendix A. maximum of 91% in Tanzania. Most remaining commercializers 27 Randolph (1992), p. 11. 28 are growing and marketing both food and cash crops (9% in Tanza- In some case this choice was justified by the early stage of the process of nia, 21% in Uganda and 16% in Malawi), whereas those focusing agricultural commercialization in the country under review, with a negligible share of farmers resorting to purchased inputs. See for instance Randolph (1992). only on the production and sale of non-food crops represent less 29 Randolph (1992), p. 291. than 1% in each country. 30 In more detail, our dataset provides information as to who in the household decides what to do with the earnings from sales of a crop. This was used to determine male and female revenues from crop sales within the household. Surveys also provide information on who in the household manages a plot. This was used to determine the 23 Additional information and data from the imputation are available upon request. harvest value for each gender within the household. The CCI by gender was thus 24 In this study, livestock ownership has been excluded from the sample. computed as the percentage of each gender’s harvest (in monetary terms) which was 25 More detailed information on the surveys, including sample size and survey reported to have been sold. 31 design can be found at www.worldbank.org/lsms. Results by gender not reported and available upon request. 110 C. Carletto et al. / Food Policy 67 (2017) 106–118 Table 2 Main sample characteristics. Source: Own computations on LSMS -ISA. SD in parentheses. Sample characteristics Malawi Tanzania Uganda All Sellers Non sellers All Sellers Non sellers All Sellers Non sellers # of households 9894 8727 1167 2074 1335 739 1788 1415 373 of which: Male headed (%) 75.4 75.7 74.7 73.7 72.7 71.7 70.7 69.7 68.7 Female headed (%) 24.6 24.3 25.3 26.3 27.3 28.3 29.3 30.3 31.3 Education (%) – None 78.8 78.2 83.8 30.2 28.7 33.3 18.8 16.6 27.7 – Primary 9.1 9.4 6.9 62.9 65.4 57.5 58.2 59.0 55.2 – Secondary 11.0 11.2 8.7 6.8 5.8 9.1 18.2 19.4 13.7 – Tertiary 1.1 1.2 0.6 0.1 0.1 0.1 4.8 5.1 3.4 HH head age 43.1 43.7 48.6 47.3 51.4 47.0 46.7 48.1 (16.51) (16.42) (17.28) (15.76) (15.25) (16.45) (15.73) (15.51) (16.54) HH size 4.7 4.7 4.5 5.5 5.6 5.5 5.3 5.4 5.1 (2.17) (2.18) (2.08) (2.93) (2.93) (2.94) (2.62) (2.62) (2.63) CDR 0.8 0.8 0.8 0.7 0.7 0.7 0.8 0.8 0.8 (0.70) (0.70) (0.73) (0.68) (0.69) (0.67) (0.76) (0.75) (0.76) Distance market (Kms) 7.9 7.9 8.0 75.5 78.7 68.8 31.6 31.4 32.4 (5.31) (5.32) (5.25) (50.97) (53.09) (45.46) (17.79) (16.87) (21.02) Distance pop. Center (Kms) 36.0 35.8 38.2 51.2 54.0 45.2 25.2 24.1 29.8 (20.06) (20.03) (20.19) (39.00) (38.99) (38.36) (16.83) (15.02) (22.01) P.c. food expenditure (USD) 0.52 0.53 0.42 0.63 0.63 0.63 0.40 0.42 0.35 (0.34) (0.35) (0.29) (0.35) (0.34) (0.36) (0.26) (0.27) (0.25) P.c. kcal Consumption 2536 2554 2383 2044 2078 1972 2243 2317 1954 (2305.56) (2239.66) (2807.45) (867.95) (858.17) (884.73) (1567.90) (1573.86) (1512.01) Hired labor (days) 4.4 4.8 1.5 8.5 10.8 3.4 16.2 17.9 9.7 (17.27) (18.08) (6.47) (25.32) (29.57) (10.42) (35.31) (37.20) (25.63) Land owned (Ha) 0.9 1.0 0.6 2.6 3.0 1.7 2.3 2.5 1.5 (13.13) (13.87) (0.61) (4.59) (5.27) (2.34) (12.76) (14.15) (3.83) # crops harvested 2.2 2.3 1.5 2.1 2.3 1.6 3.8 4.2 2.4 (1.12) (1.09) (1.11) (1.12) (1.14) (0.85) (1.87) (1.80) (1.45) # crops sold 1.7 1.8 0.0 1.0 1.5 0.0 1.8 2.3 0.0 (1.02) (0.90) 0.00 (0.94) (0.76) 0.00 (1.49) (1.31) 0.00 HH Harvest value (USD) 269.92 292.28 76.91 244.65 314.20 96.63 215.03 254.46 60.06 (731.34) (768.50) (110.90) (500.71) (429.69) (599.61) (334.09) (361.59) (79.97) HH revenue (USD) 102.28 114.13 0.00 112.23 164.96 0.00 85.59 107.38 0.00 (542.64) (572.03) 0.00 (286.60) (334.75) 0.00 (217.84) (239.16) 0.00 AG Income (USD) 285.18 308.40 84.72 281.35 350.66 133.84 234.07 272.07 84.77 (738.08) (775.05) (130.67) (534.48) (472.76) (621.66) (354.33) (378.56) (167.05) HH Days worked 125.30 129.00 93.36 149.32 161.51 123.38 134.10 143.74 96.22 (116.28) (119.63) (74.64) (146.35) (150.35) (133.91) (106.10) (109.05) (83.57) Table 3 in the production and sale of tobacco (the main cash crop in Malawi), CCI by chosen characteristics. as well as constraints to accessing more land resources to allow for greater crop diversification. CCI To assess the degree of commercialization of these staple and Malawi Tanzania Uganda other food crops, we thus proceed to construct separate CCIs to Country average 17.6 27.5 26.3 reflect the degree of commercialization of food items versus non- Country average (conditional on sales) 19.6 40.4 33.0 edible items. As non-edible items are planted in most instances Female headed 10.8 20.3 20.7 Female headed (conditional on sales) 12.2 33.7 28.7 with the primary purpose of selling, it is not surprising to find Male headed 19.8 29.8 28.6 the CCI for households who plant these crops to be as high as Male headed (conditional on sales) 22.0 42.3 34.5 91% for tobacco in Malawi32 or 87% for coffee in Uganda. Female farmers 9.0 19.1 23.0 Farm households, however, do not only sell traditional cash Female farmers (conditional on sales) 30.6 42.9 37.0 crops, i.e. crops grown almost exclusively for sale. Table 4 shows Male farmers 19.8 30.8 27.0 Male farmers (conditional on sales) 21.7 42.8 34.6 that households in all three countries to a large extent are also involved in the sale of traditional staple crops such as maize and/ By land size – Less than 0.5 ha 9.9 15.4 20.8 or cassava. However commercialization of most food crops remains – Between 0.5 and 1 ha 19.8 21.6 25.3 low although, with the exception of Malawi, households who do – Between 1 and 2 ha 28.8 26.2 28.5 choose to sell, sell a considerable portion of their harvest. In – 2 ha or more 34.8 34.8 30.7 Malawi, on the contrary, food crops like maize and cassava are sold by many households, but it is only done in small quantities. This relatively high incidence of small quantities of maize sales is the reason why the country’s CCI is low. than 1 ha, male unconditional CCI is almost double that of their female counterparts. This gender gap across farm sizes is not as stark in the other two countries. The larger gender gaps in Malawi may 32 Figures lower than 100% for tobacco in Malawi are likely to reflect accumulation reflect greater restrictions for female farmers on fully participating of stocks. C. Carletto et al. / Food Policy 67 (2017) 106–118 111 Table 4 The degree of HHs’ agricultural commercialization by type of crop. Crop % Planting CCI among planters % selling among planters CCI conditional on planting and selling Malawi Maize 97.4 5.0 84.2 5.9 Cassava 11.0 4.3 60.8 7.1 Ground Nut 27.1 29.1 88.1 33.1 Tobacco 14.8 90.5 95.1 95.2 Soya 5.6 43.0 76.8 56.0 Pigeon Peas 22.1 15.1 58.3 26.0 Beans 11.1 10.1 37.8 26.8 Food crops 99.7 9.9 88.1 11.3 Non-food crops 16.8 89.8 94.2 95.3 Tanzania Maize 78.3 15.6 53.8 29.0 Ground Nut 14.4 28.3 42.6 66.5 Paddy Rice 19.8 30.7 56.0 54.8 Beans 28.7 19.9 35.0 57.0 Sorghum 11.1 12.7 24.2 52.4 Sweet Potato 9.9 11.2 20.8 53.9 Cowpeas 6.8 19.4 26.9 72.0 Food crops 99.2 23.8 64.8 36.7 Non-food crops 9.3 85.9 88.6 97.0 Uganda Maize 58.4 20.0 54.7 36.5 Cassava 41.4 8.0 19.8 40.3 Ground Nut 26.3 21.2 61.5 34.5 Banana (food) 49.6 34.4 67.3 51.2 Sweet Potato 42.8 5.5 13.9 39.9 Coffee 18.8 86.7 87.6 98.9 Beans 65.0 13.2 33.6 39.3 Food crops 99.7 22.5 76.3 29.5 Non-food crops 21.2 92.1 94.6 97.4 In the countries studied, there is commercialization of staple items such as maize (5% in Malawi, 16 in Tanzania and 20 in Uganda) and beans (10% in Malawi, 20 in Tanzania and 13 in Uganda). Overall, the share of food crops sold is 10% in Malawi, 24% in Tanzania and 23% in Uganda. Looking at the shares of food crop sold by farm size, as expected, farmers with larger landhold- ings tend to sell larger shares of their food production, reflecting greater surpluses, although in countries like Malawi, the share remains rather low, at 14% even for farmers with more than 2 ha of land. Finally, as expected, those with greater harvests (measured in monetary value) tend to have higher levels of commercialization in all three countries. Graph 1, presents the average CCI by harvest value deciles, and illustrates the level of commercialization across the harvest value distribution. Even households with the lowest harvest values engage the market. Graph 1. Avg. agricultural commercialization by harvest value deciles. 5. Exploring the relationship between agricultural In terms of average caloric consumption, Tanzania exhibits average commercialization and nutrition per capita caloric consumption of 2044 kilocalories, compared with 2536 in Malawi and 2243 in Uganda. Across countries, there is no In this section, we investigate the nexus between the degree of clear relationship between the nutritional outcomes and the agricultural commercialization and the nutritional status of farm degree of commercialization (as proxied by the CCI quintiles) and households. Three indicators are used to measure household nutri- the different nutritional indicators, with the exception of stunting tional status: (i) children’s anthropometric measures (measured in Tanzania. Similarly, no clear trends emerge when the degree of both in terms of percentage of children stunted, wasted and under- agricultural commercialization is correlated with children’s weight, and through the computations of Z-scores), (ii) household anthropometrics as measured through Z-scores (see Graph 2). per capita food expenditure, and (iii) household per capita caloric The absence of a correlation with child anthropometry might be consumption. partially attributable to the smaller sample size of children, partic- Table 5 suggests high levels of malnutrition in all countries, ularly for Tanzania and Uganda. Pooling the three country samples with an incidence of stunting among children under five years and running a local polynomial non-parametric regression (with- old of about 42% in Tanzania, compared to 36% in Uganda and out any control variables) a slightly upward gradient with com- 31% in Malawi. Similarly, the share of children wasted amounts mercialization emerges for the height-for age (stunting) and to 6.2, 3.2 and 3.6% in Tanzania, Uganda and Malawi respectively. weight-for-height (wasting) measures, suggesting a some positive 112 C. Carletto et al. / Food Policy 67 (2017) 106–118 Table 5 CCI quintile breakdown of nutritional outcomes. Nutritional measure HAZ WAZ WHZ Stunted Wasted Underweight Food Expenditure ($) Kilo Calories Malawi CCI Quintile No Sales À1.31 À0.52 0.29 25.6 3.9 5.7 0.42 2418 1 À1.22 À0.48 0.28 25.2 3.0 4.7 0.47 2352 2 À1.53 À0.57 0.41 32.8 2.7 7.5 0.53 2546 3 À1.32 À0.41 0.46 30.3 3.6 5.5 0.54 2670 4 À1.40 À0.54 0.35 30.5 4.7 6.8 0.55 2538 5 À1.52 À0.51 0.47 36.5 3.8 7.7 0.57 2640 Country mean À1.39 À0.57 0.39 30.7 3.6 6.4 0.52 2536 Tanzania CCI Quintile No Sales À1.72 À0.95 0.02 42.6 5.7 14.4 0.63 1972 1 À1.81 À1.10 À0.15 43.4 7.7 24.1 0.66 2215 2 À1.85 À0.97 0.10 47.2 7.1 16.9 0.59 2051 3 À1.67 À1.02 À0.13 45.1 6.2 15.5 0.61 2004 4 À1.62 À0.88 0.03 40.1 5.5 12.1 0.62 2074 5 À1.58 À0.92 À0.06 32.4 5.5 14.0 0.64 2044 Country mean À1.71 À0.96 À0.02 41.9 6.2 15.6 0.63 2044 Uganda CCI Quintile No Sales À1.43 À0.83 À0.04 32.1 2.4 14.5 0.35 1954 1 À1.35 À0.58 0.26 36.8 1.7 9.4 0.40 2229 2 À1.85 À1.01 0.07 45.6 6.1 16.1 0.41 2299 3 À1.59 À0.78 0.18 37.0 5.1 12.3 0.44 2546 4 À1.57 À0.70 0.27 31.8 1.6 10.7 0.40 2362 5 À1.44 À0.58 0.31 34.4 2.4 7.9 0.44 2132 Country mean À1.53 À0.75 0.16 36.0 3.2 11.9 0.40 2243 Note: Food expenditure and kilo calorie data are per capita, and at the household level. Graph 2. Agricultural commercialization and nutrition: pooled sample. Source: Own computations on LSMS -ISA. C. Carletto et al. / Food Policy 67 (2017) 106–118 113 Table 6 Individual fixed effects specification for pooled sample. HAZ (1) (2) (3) (4) (5) ln(pc. Expenditure) 0.108 0.115 0.110 0.115 0.107 (0.0737) (0.0742) (0.0742) (0.0735) (0.0737) CCI 0.178 0.160 0.155 (0.123) (0.121) (0.122) Perc. Female 0.00186 (0.00118) Female CCI 0.298 (0.213) Perc. Food 0.00108 (0.000895) Food CCI 0.201 (0.134) Adjusted R2 0.097 0.099 0.097 0.099 0.097 WAZ (1) (2) (3) (4) (5) ln(pc. Expenditure) 0.0579 0.0502 0.0565 0.0639 0.0577 (0.0518) (0.0503) (0.0520) (0.0519) (0.0518) CCI 0.0620 0.0816 0.0409 (0.0871) (0.0872) (0.0863) Perc. Female À0.00195⁄⁄ (0.000864) Female CCI 0.00839 (0.138) Perc. Food 0.000950⁄ (0.000533) Food CCI 0.0827 (0.0954) Adjusted R2 0.097 0.099 0.097 0.099 0.097 WHZ (1) (2) (3) (4) (5) ln(pc. Expenditure) 0.0148 À0.000592 0.0118 0.0187 0.0152 (0.0828) (0.0809) (0.0835) (0.0825) (0.0826) CCI À0.0462 À0.00688 À0.0600 (0.133) (0.130) (0.132) Perc. Female À0.00391⁄⁄ (0.00151) Female CCI À0.199 (0.265) Perc. Food 0.000622 (0.000787) Food CCI À0.0396 (0.155) Adjusted R2 0.031 0.039 0.032 0.032 0.031 Observations 3140 3140 3140 3140 3140 *** ** * p < 0.01; p < 0.05; p < 0.1. correlation between child nutritional outcomes and household’s metric outcomes. The owner of the revenue from the sales could be commercialization. In the remainder of the paper, we explore these of importance33; specifications that take this into account are also relationships in more detail using a multivariate framework based included. In the first specification of each model (Column 1) we on the pooled sample. include the overall household commercialization index. In column 2, we add to the previous specification the share of the household 5.1. Empirical strategy and main results CCI accruing to female farmers within the household. In column 3, we introduce yet another specification of the gender CCI by using In order to analyze further how commercialization may affect the Female CCI. In a similar fashion, to account for the potentially nutritional outcomes, more directly through changes in food con- differential impact of commercialization of food commodities, in col- sumption and more indirectly through changes in income, we esti- umn 4 and 5 we introduce two variants of the Food CCI, first by add- mate a set of models first at the individual level to investigate how ing to the total CCI the share of total CCI deriving from the sale of CCI impacts child nutritional status and then at the household level food crops and then by replacing the total CCI with the Food CCI. with the aim of exploring how CCI correlates to household per cap- In each model and specification, the common correlates are: ita expenditure. gender of head, age of head, education of the head, natural loga- Specifically, in Table 6 we report selected findings of the esti- rithm of land holdings, natural logarithm of land holdings squared, mated impact of commercialization on child anthropometrics. In the natural logarithm of the household’s harvest value, annual this instance the sample has been pooled due to the reasons men- average rainfall in millimeters, and the child’s age in months as tioned in the previous section. Table 7 illustrates estimates of the well as the child’s gender. probability of a child being stunted, wasted, or underweight. For The key variables in this model (CCI and its variants) are in all each model, we use five different specifications based on different likelihood endogenous due to simultaneous causality between characterizations of household commercialization and thus how we introduce the CCI. The aim is to determine if increased sales 33 For example in Cote d’Ivoire, Duflo and Udry (2004) find that income shocks that of the household’s harvest could be related to observed anthropo- benefitted females were positively related to household food expenditure. 114 C. Carletto et al. / Food Policy 67 (2017) 106–118 Table 7 Random effects logit specification for pooled sample. Stunted (1) (2) (3) (4) (5) ⁄⁄ ⁄⁄ ⁄⁄ ⁄⁄ ln(pc. Expenditure) À0.242 À0.242 À0.243 À0.247 À0.244⁄⁄ (0.109) (0.109) (0.109) (0.110) (0.109) CCI 0.179 0.188 0.191 (0.239) (0.240) (0.239) Perc. Female À0.000817 (0.00228) Female CCI 0.165 (0.441) Perc. Food À0.00138 (0.00136) Food CCI À0.0405 (0.263) Wasted (1) (2) (3) (4) (5) ln(pc. Expenditure) À0.176 À0.185 À0.180 À0.183 À0.183 (0.189) (0.191) (0.190) (0.191) (0.189) CCI 0.161 0.284 0.171 (0.436) (0.434) (0.429) Perc. Female À0.0112⁄⁄ (0.00487) Female CCI À0.688 (1.097) Perc. Food À0.00279 (0.00221) Food CCI 0.526 (0.448) Underweight (1) (2) (3) (4) (5) ln(pc. Expenditure) À0.692⁄⁄⁄ À0.692⁄⁄⁄ À0.698⁄⁄⁄ À0.700⁄⁄⁄ À0.695⁄⁄⁄ (0.168) (0.168) (0.168) (0.169) (0.167) CCI 0.315 0.326 0.334 (0.362) (0.365) (0.361) Perc. Female À0.000952 (0.00402) Female CCI À0.238 (0.748) Perc. Food À0.00234 (0.00212) Food CCI 0.386 (0.395) Observations 3140 3140 3140 3140 3140 *** ** * p < 0.01; p < 0.05; p < 0.1. the dependent variables and commercialization. Additionally, it is and significant. This suggests that greater involvement by women possible that several common unobservable factors impact both may result in some negative effect for short-term nutritional out- kinds of outcomes. In order to address these potential endogeneity comes. However, in light of the rather small sample size of children issues, the panel component of the data is used by estimating fixed in the panel, these results should be taken with some caution. effects models. Naturally, time-varying covariates which are not Finally, the level of per capita expenditure in the household is also controlled for by the fixed-effect model still present a potential not significantly related to Z-scores. problem. The probability of a child being stunted, wasted, or underweight is modelled in Table 7. In line with the results presented in Table 6, the coefficients only show a significant and negative effect of 5.2. Anthropometrics with pooled sample of children greater commercialization by women on short-term nutritional indicators, likely a reflection of the potentially deleterious effect In order to analyze how commercialization relates to children’s of lower levels of child care on child nutritional status. Per capita nutritional outcomes, we focus on the pooled sample of children expenditure in this instance does seem to play a role, with an present in both waves, who were older than 6 months during the increase in expenditure negatively related to the child’s likelihood first round and younger than 60 months old by the second wave. of being stunted, and underweight. The relationship, however, is We run both an individual fixed effects linear model on Z-scores not significant for wasting. as well as a random effects logit model on the probability of being In Table 8 we analyze the relationship between commercializa- stunted, wasted or underweight. tion and per capita food expenditures. With the above mentioned The fixed effects results in Table 6 are quite consistent for total caveats on the potential endogeneity of some of the regressors,34 CCI and its variants, with the coefficients being largely not signifi- overall we fail to find any clear pattern, and the findings seem to cant. More explicitly, there is no relationship between anthropo- diverge only slightly across countries. For instance, looking at the metric outcomes and CCI, both as a total and when disaggregated by food and non-food products. The few exceptions are in the 34 Simultaneity is especially a strong concern, since we observe expenditures at a WAZ and WHZ models where, at equal levels of total household point when the household may not have completed its commercialization process commercialization, the share of CCI accruing to women is negative yet. C. Carletto et al. / Food Policy 67 (2017) 106–118 115 Table 8 Household fixed effects specification for nat. log of household’s per capita food expenditure. Malawi (1) (2) (3) (4) (5) ⁄⁄⁄ ⁄⁄⁄ ⁄⁄⁄ ⁄⁄⁄ ln(harvest value) 0.103 0.102 0.102 0.102 0.102⁄⁄⁄ (0.0186) (0.0185) (0.0175) (0.0185) (0.0176) CCI 0.0132 À0.00158 0.0172 (0.0842) (0.0842) (0.0824) Perc. Female 0.000628 (0.000522) Female CCI 0.202 (0.163) Perc. Food 0.000223 (0.000383) Food CCI 0.0791 (0.0824) Adjusted R2 0.347 0.348 0.348 0.347 0.347 Observations 4444 4444 4444 4444 4444 Tanzania (1) (2) (3) (4) (5) ln(harvest value) 0.0436⁄⁄⁄ 0.0439⁄⁄⁄ 0.0396⁄⁄⁄ 0.0469⁄⁄⁄ 0.0437⁄⁄⁄ (0.0121) (0.0121) (0.0115) (0.0131) (0.0119) CCI À0.0733 À0.0657 À0.0552 (0.0588) (0.0601) (0.0588) Perc. Female À0.000311 (0.000599) Female CCI À0.105 (0.104) Perc. Food À0.000399 (0.000366) Food CCI À0.0856 (0.0611) Adjusted R2 0.130 0.130 0.130 0.131 0.130 Observations 3488 3488 3488 3488 3488 Uganda (1) (2) (3) (4) (5) ln(harvest value) 0.0529⁄ 0.0527 0.0506 0.0498 0.0530 (0.0319) (0.0322) (0.0309) (0.0325) (0.0321) CCI À0.132⁄ À0.133⁄ À0.146⁄ (0.0754) (0.0744) (0.0764) Perc. Female 4.21eÀ05 (0.000575) Female CCI À0.202⁄⁄ (0.102) Perc. Food 0.000644 (0.000473) Food CCI À0.133 (0.0841) Adjusted R2 0.134 0.134 0.136 0.136 0.134 Observations 3174 3174 3174 3174 3174 *** ** p < 0.01; p < 0.05; * p < 0.1. first columns, we find little evidence of a relationship between CCI 6. Main conclusions and food expenditures, except for Uganda where the coefficient is negative and marginally significant. Also, in Uganda, the negative Despite the inconclusiveness of the available empirical evidence coefficient on female CCI is still marginally significant though some- to date, agricultural commercialization among poor smallholders what larger than for the total CCI. All other coefficients provide little continues to be heralded as an effective solution to reduce poverty, support to the existence of a relationship between commercializa- improve household food and nutrition security, and foster growth tion, in its different specifications, and food expenditures in any of in rural areas. Based on new comparable data from across sub- the countries analyzed. Saharan Africa which enables the calculation of commercialization In Table 9 we report the same coefficients by regressing house- indexes at the individual and crop level, this paper contributes to hold total per capita expenditure on the same set of regressors and the ongoing debate by investigating the relationship between various CCI specifications. The results we found for food expendi- increased agricultural commercialization and several nutritional tures for Malawi carry over to this specification with little evidence indicators in three African countries, differentiated by gender and of any impact of commercialization on total expenditures. This lack types of crops sold. of impact may be due to the fact that while commercialization is Against conventional wisdom, the data reveal very high levels widespread across farmers, sales often involve small amounts of commercialization by even the poorest and smallest land hold- which fail to have a significant impact on total household per cap- ers, with rates of market participation as high as 90% in Malawi. ita expenditures.35 Similarly, against common perceptions, a considerable portion of this market presence is driven by the sale of staple and other food crops and not necessarily by traditional cash crops. This is in part 35 due to the fact that the great majority of smallholders are still spe- We also run a specification where we include the household’s revenue from the cializing in the production of food crops (between 80 and 90% in sale of crops, instead of the CCI, and the coefficient is also not significant (results available upon request). the three countries analyzed), with only a relatively small share 116 C. Carletto et al. / Food Policy 67 (2017) 106–118 Table 9 Household fixed effects specification for nat. log of household total per capita expenditure. Malawi (1) (2) (3) (4) (5) ⁄⁄⁄ ⁄⁄⁄ ⁄⁄⁄ ⁄⁄⁄ ln(harvest value) 0.0977 0.0978 0.0997 0.0973 0.0990⁄⁄⁄ (0.0167) (0.0166) (0.0160) (0.0166) (0.0159) CCI 0.0566 0.0605 0.0600 (0.0686) (0.0687) (0.0674) Perc. Female À0.000163 (0.000425) Female CCI 0.0442 (0.137) Perc. Food 0.000185 (0.000306) Food CCI 0.120 (0.0728) Adjusted R2 0.444 0.445 0.444 0.445 0.445 Observations 4444 4444 4444 4444 4444 Tanzania (1) (2) (3) (4) (5) ln(harvest value) 0.0319⁄⁄⁄ 0.0330⁄⁄⁄ 0.0309⁄⁄⁄ 0.0357⁄⁄⁄ 0.0324⁄⁄⁄ (0.0105) (0.0105) (0.0100) (0.0112) (0.0103) CCI À0.0416 À0.0178 À0.0212 (0.0576) (0.0585) (0.0574) Perc. Female À0.000969⁄ (0.000507) Female CCI À0.163 (0.0994) Perc. Food À0.000450 (0.000321) Food CCI À0.0555 (0.0596) Adjusted R2 0.130 0.132 0.131 0.131 0.130 Observations 3488 3488 3488 3488 3488 Uganda (1) (2) (3) (4) (5) ln(harvest value) 0.0164 0.0162 0.0154 0.0130 0.0146 (0.0222) (0.0223) (0.0210) (0.0225) (0.0223) CCI À0.0539 À0.0552 À0.0704 (0.0805) (0.0799) (0.0809) Perc. Female 5.46eÀ05 (0.000463) Female CCI À0.0826 (0.0942) Perc. Food 0.000726⁄ (0.000421) Food CCI À0.0156 (0.0890) Adjusted R2 0.225 0.225 0.226 0.228 0.225 Observations 3174 3174 3174 3174 3174 *** ** p < 0.01; p < 0.05; * p < 0.1. cultivating both food and traditional cash crops. However, in most weak association between increased commercialization and cases, particularly in Malawi, market participation only involves improved food security and nutritional outcomes. the sale of relatively small quantities of own food production, resulting in low food CCI-- 10% for the entire sample and only Acknowledgement 14% among the largest farmers. Another important finding of the cross-country analysis is that The authors would like to acknowledge participants to two although female farmers appear to participate less in market activ- workshops organized by the Agriculture in Africa: Telling Myths ities, when they do participate, they tend to sell larger shares of the from Facts project. We would also like to thank the participants production under their control relative to their male counterparts. of the 2015 AAEA conference. We are particularly grateful to Luc In line with previous research, we also find little evidence of a Christiaensen for continuous inputs throughout the process. The relationship between increased commercialization and improved findings, interpretations and conclusions expressed in this paper nutritional status. The only exception is a weak negative relation- are entirely those of the authors and do not necessarily express ship between the portion of commercialization accruing to females the views of the World Bank and its affiliated organizations. All and short-term nutritional indicators, which could be the results of errors and omissions remain the responsibility of the authors. the negative effect of greater female market participation on time allocated to child care and homemaking. Nonetheless, these findings should still be taken with caution, Appendix A as we are still unable to fully control for the potential simultaneity of the CCI and total harvest value and other variables, despite the A.1. Adjustment of crop production sold use of panel data. That said, the arguably more robust and repre- sentative evidence presented here is in line with the bulk of evi- Obtaining figures for household commercialization is one of the dence to date, and yet another piece of empirical evidence of the aims of this study. Households in our samples were interviewed at C. Carletto et al. / Food Policy 67 (2017) 106–118 117 Table A1 Sample composition. Malawi Tanzania Uganda Maize Observations 9578 1510 1172 Obs. Selling 1520 489 441 Legumes Observations 4224 1172 256 Obs. Selling 1279 488 Tubers Observations 479 360 1832 Obs. Selling 200 101 Grains Observations 1104 541 764 Obs. Selling 126 214 241 Rice Observations 619 351 N/A Obs. Selling 378 221 Ground nut Observations 2466 N/A 494 Obs. Selling 855 193 Cassava Observations 1452 N/A N/A Obs. Selling 257 Banana Observations N/A N/A 1202 Obs. Selling 533 Beans Observations N/A N/A 1208 Obs. Selling 399 different points of the year, thus households in the sample differ in is necessary that the second derivative of the estimated model be the time span during which they could have sold their harvest. negative. Therefore, the adjustment is done only for the crops Additionally, different crops have different harvest periods where this holds true. The final sales value is the original sales throughout the season. These two facts complicate assessing value for all households for which the time span is greater than whether or not a household sells crop, and if so, how much would the optimal t. For households where the time span is less than it sell if it had sufficient time. To proceed, we utilize an adjustment the optimal t, the adjusted sales value is equal to the original sales procedure similar to that detailed in Kaminski and Christiaensen value and the difference between the expected sales valued at the (2014), which was used to estimate post-harvest loss. optimal time and the expected sales value at the actual time. The analysis is conducted at the crop level. The sample is addi- The other portion of the analysis consists in determining if and tionally separated into crop groupings, depending on the country how much households that did not report sales at the time of inter- and crop type.36 Table A1 gives a breakdown of the crops considered view would have sold. The probability of selling crops is estimated by country. Vegetables and fruits are not considered due to their using a probit model given the reported time span, and utilizes the short shelf life. Also not considered are goods that are inherently same set of regressors as in Eq. (1). Given the estimated parame- commercial in nature and the household has produced for the mar- ters, once again it is necessary to find the time span that maximizes ket: these are all non-food items. Examples of non-food items the probability of selling the given crop. Therefore, it is necessary include cotton, tobacco and coffee. that the second derivative of the function with respect to time span For each crop, the sample is initially analyzed only for those be negative. If this holds, then the adjustment for that crop is done who reported sales of the crop by the time of interview. For these for households for which the time span is less than or equal to the observations, the goal is to determine how much they would have estimated optimal time span. For these households, the predicted sold given enough time, or if they would have sold more given probability of selling crops is equal to: more time. The following regression is estimated for each crop: Pðselli ¼ 1Þ ¼ P ðselli ¼ 1jX i ; tÃ Þ À Pðselli ¼ 1jX i ; t ¼ ti Þ ð3Þ y ¼ X b þ f ðt Þ þ dðHht Þ þ e ð 1Þ Using this probability, it is possible to obtain a predicted sales value In this specification, y is T  1 vector of the natural logarithm of for these households which is equal to: the kilos sold by the household, b is a K  1 vector of parameters to ½Eðyi jX i ; t à ; selli ¼ 0Þ À Eðyi jX i ; t i ; selli ¼ 0ފ  Pðselli ¼ 1Þ ð4Þ be estimated including a constant term, and X is a T  K matrix. Among the considered regressors are a set of household character- Once these values are obtained it is possible to obtain the istics, including characteristics for the head of the household. Time adjusted CCI for each household. The results from the estimations span between harvest and interview is considered in are available from the authors upon request. Hhtðharv ested kilos per capita  time spanÞ. Additionally, time is also considered in f ðt Þ which is a T  1 vector and is equal to: References f ðt Þ ¼ ht þ ct2 ð 2Þ Bellin, F., 1994. Smallholder tree crops in sierra leone: impacts on food consumption Finally, e is the random error term and is assumed to be normal, and nutrition. In: von Braun, J., Kennedy, E. (Eds.), Agriculture independent and identically distributed (i.i.d.) with mean zero. Commercialization, Economic Development, and Nutrition. The John Hopkins University Press, Baltimore and London. In order to get the adjusted sales value, the time span which Billah, K., 2002. Crop Production and Nutrition in Bangladesh Ph.D Dissertation. maximizes kilos sold must be obtained. For this to be the case, it Cornell University. May. Braun, J.V., Kennedy, E., 1994. Agricultural Commercialization, Economic Development, and Nutrition. Johns Hopkins University Press. 36 Malawi: Maize, ground nut, rice, legumes, tubers, grains, cassava. Tanzania: Von Braun, J., Kennedy, E., Bouis, H., 1989. Comparative Analyses of the Effects of Maize, paddy rice, legumes, grains, tubers. Uganda: Maize, ground nut, beans, banana, increased Commercialization of Subsistence Agriculture on production, legumes, tubers, grains. Consumption, and Nutrition, Final Report, IFPRI. 118 C. Carletto et al. / Food Policy 67 (2017) 106–118 Von Braun, J., Kennedy, E., 1986. Commercialization of Subsistence Agriculture: Hitchings, Jon Albert, 1982. Agricultural Determinants of Nutritional Status among Income and Nutritional Effects in Developing Countries. IFPRI Working Papers Kenyan Children. Stanford University. on Commercialization of Agriculture and Nutrition, n.1. Kaminski, J., Christiaensen, L., 2014. Post-harvest loss in sub-Saharan Africa—what Carletto, C., Kilic, T., Kirk, A., 2011. Nontraditional crops, traditional constraints. The do farmers say? Global Food Secur. 3 (3), 149–158. long-term welfare impacts of export crop adoption among Guatemalan Lambert J. N. ‘‘Does cash cropping cause malnutrition.” National Planning Office, smallholders. Agric. Econ. 42 (supplement), 61–75. 2, 5. Port Moresby, Papua New Guinea (1978). Carletto, C., Davis, B., Kirk, A., Winters, P., 2010. Globalization and smallholders: the Lappe F.M. (1977), ‘‘Food First: beyond the Myth of Scarcity”, Boston: Houghton adoption, dissemination and welfare impact of non-traditional export crops in Mifflin Guatemala. World Dev. 38 (6). Lev, L., 1981. The Effect of Cash Cropping on Food Consumption Adequacy among Carletto, C., de Janvry, A., Sadoulet, E., 1999. Sustainability in the diffusion of the Meru of Northern Tanzania, MSU Rural Development Series, Working Paper innovation: smallholder non-traditional agro-exports in Guatemala. Econ. Dev. No. 21. Cult. Change 47 (2). Maxwell, S., Fernando, A., 1989. Cash crops in developing countries: the issues, the Dewey, K.G., 1981. Nutritional consequences of the transformation from facts, the policies. World Dev. 17 (11), 1677–1708. subsistence to commercial agriculture in Tabasco, Mexico. Hum. Ecol. 9 (2), Pinstrup-Andersen, P., 1983. Export Crop Production and Malnutrition, Occasional 151–187. Paper Series, vol. 2, no. 10, Chapel Hill, University of North Carolina, reprinted Duflo, E., Udry, C., 2004. Intrahousehold Resource Allocation in Cote d’Ivoire: Social by IFPRI. Norms, Separate Accounts and Consumption Choices. Technical report, National Rabeneck, Sonya, 1982. The Determinants of Protein-Energy Malnutrition among Bureau of Economic Research. Preschool Children in Kenya With Respect to Cash Cropping and Self-Sufficiency Fleuret, P., Fleuret, A., 1983. Socio-economic determinants of child nutrition in in Staple Food Production. Cornell University. Aug. Taita, Kenya: a call for discussion. Cult. Agric. 19 (8), 16–20. Randolph, F., 1992. The Impact of Agriculture Commercialization on Child Govereth, J., Jayne, T.S., Nyoro, J., 1999, Smallholder commercialization, interlinked Nutrition: A Case Study of Smallholder Households in Malawi Ph.D markets and food crop productivity: cross-country evidence in eastern and Dissertation. Cornell University. May. Southern Africa, mimeo, MSU, June. Strasberg, P.J., Jayne, T.S., Yamano, T., Nyoro, J., Karanja, D., Strauss, J., 1999. Effects Gross Daniel, R., Underwood, Barbara A., 1971. Technological change and caloric of Agriculture Commercialization on Food Crop Input Use and Productivity in costs: sisal agriculture in northeastern Brazil. Am. Anthropol., 725–740 Kenya, MSU International Development Working Papers No. 71. Harvey Philip, W.J., Heywood, Peter F., 1983. Nutrition and Growth in Simbu. Wiggins, S., Argwings-Kodhek, G., Leavy, J., Poulton, C., 2011. Small farm Boroko: Papua New Guinea Institute of Applied Social and Economic Research, commercialization in Africa: Reviewing the issues, Future Agricultures 1983. Print. Research Paper No. 23, April. Hernandez, M., Perez, Hidalgo C., Ramirez, Hernandez J., Madrigal, H., Chavez, A., Wood, B., Nelson, C., Kilic, Talip, Murray, S., 2013. Up in the Smoke? Agriculture 1974. Effect on economic growth on nutrition in a tropical community. Ecol. Commercialization, Rising Food Prices and Stunting in Malawi, World Bank Food Nutr. 3 (4), 283–291. Policy Research Working Paper 6650, October.