Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

The impact of non-traditional export agriculture on income and food availability in Guatemala: An intra-household perspective

Elizabeth Katz

Abstract

This paper assesses the impact of non-traditional agriculture exports (NTAE) in highland Guatemala on the level and distribution of household income, and the financing and purchase of food items. Heckman's two-step procedure was used to control for self-selection of NTAE producers and potential correlation of preferences. Farmers with greater extension of land, as well as younger Catholic farmers with higher-quality land, are most likely to adopt the new crops (broccoli, snow peas {mange-tout!), and NTAE crops are highly profitable for those who adopt them. Despite the widespread use of female labour in the production of NTAE crops, the extent of adoption is not negatively associated with women's ability to generate independent income. Land devoted to the new crops also exhibits a strong positive relationship with subsistence production of corn and beans. Weak evidence supports the hypothesis that distinct intra-household income transfer patterns obtained in NTAE crop-adopting households compared with non-adopting households: in the former, the level of the weekly food allowance provided to women by their husbands is less responsive to changes in both household and female income, making women's earnings a more important overall determinant of food expenditure. A simulation exercise showed that households that have not adopted new crops would use increases in income to augment food expenditures to a much greater extent than adopting households. Within adopting households, the impact of increases in women's income on food expenditures is nearly double that of increases in men's income. Women's flexibility with regard to the allocation of their labour time and income plays an important role in the relative success of non-traditional agriculture as a means of improving household food availability.

Improving the food consumption of poor households is often an important implicit or explicit goal of income-generating strategies such as the promotion of non-traditional agricultural export (NTAE) crops. However, a fair amount of controversy surrounds the strength of the relationship between increases in household income and greater access to and intake of nutritious foods. This is particularly the case when those increases in income are based on the displacement of subsistence crops and/or are biased in favour of men, who may have less propensity than women to spend marginal income gains on food for the family.

The nutritional impact of NTAE in highland Guatemala was assessed in a conceptual framework that accounts for the mediating influences of changes in household income, subsistence food production, and alterations in intra-household resource transfers. The data were drawn from a survey of over 300 rural households in five central highland communities between October 1990 and August 1991.

Market food acquisition in highland Guatemala

In rural Guatemala, as in many other parts of the world, purchasing and preparing food are fundamental female responsibilities. Indeed, the rhythm of a village woman's life is largely determined on a daily basis by the work involved in preparing almuerzo, the principal meal of the day, and on a weekly basis by the día(s) de mercado, when she makes the trip to the nearest market town to do the bulk of her food shopping. Items that nearly every household purchases on a regular basis are sugar, rice, bananas, noodles, bread, salt, tomatoes, onions, chili, and stewing beef. Approximately half of all households also regularly purchase oil, eggs, chicken, cheese, potatoes, oranges, soup mix, soda, and chips. Households that are not self-sufficient in corn and/or beans must purchase these staples, which can easily constitute a third of their monthly food bill.

With the exception of corn, purchasing food is by and large women's responsibility [1]. Women finance food expenditures with a combination of their own earned income (in the current sample, 90% of women earning income reported spending some portion of it on food, and 15% claimed that all of their earnings went for food), income transfers from their husbands, and contributions from their older children and other adults who form the household. In the sample, nearly 86% of women received a regular food allowance from their husbands. Of those who did not receive such an allowance, 40% were single heads of households, 30% had husbands who were in charge of all food purchases, 13% financed all food purchases from their own earnings, and the remainder had special circumstances such as an indigent husband or residence in their parents" home.

Of women who did receive a food allowance from their husbands, this allowance, known as el gasto, accounted for an average of 54.1% of total food expenditures. If corn purchases were excluded, husbands' mean contribution still did not exceed 63% of the total cost. The remainder was made up of women's earnings, contributions from working children and/or other household members, and any other purchases that may be made by the male head. The differences between communities with regard to the level of el gasto were striking. In the two communities where income-earning opportunities for women were worst, men's cash contributions to the food budget (excluding corn) were 91% and 81%, respectively, whereas in those villages where women were more likely to perform remunerated labour, men's contributions were just over 50%.

Econometric estimation of the nutritional impact of non-traditional agriculture

Food adequacy and diversity can be interpreted as outcomes of household and intrahousehold resource allocation. As such, they are subject to a series of (analytically and temporally) prior phenomena, chief among these the generation of household income, the distribution of that income among household members, and the allocation of that income to food acquisition. The following model of food expenditures and consumption seeks to build on the work of those who have explored the relationship between income and nutrient demand by not only allowing for gender differences in income control, but also by explicitly recognizing the role of intra-household income transfers in food acquisition. The measurement of these transfers, and their incorporation into the model, allow for the investigation of the nutritional impact of NTAE crop adoption as mediated in part by shifts in the intra-household "bargain. "

A unified model of the impact of NTAE crop adoption on food consumption is depicted in figure 1 (see FIG. 1. Unified model of the impact of NTAE crop adoption on food consumption). At the top of the flow chart are household members' resource allocation decisions, most important for present analytical purposes the allocation of land and labour to NTAE crops (broccoli, snow peas [mange-tout]), and women's time devoted to remunerated activities. The decision whether and to what extent to adopt NTAE crops, together with other resource allocation choices and stochastic phenomena (the "state of nature"), determines the male head of household's ability to generate in kind and cash income, as well as the female head's cash earnings in so far as they place competing claims on her land and/or labour time. In the model, NTAE crop adoption potentially affects both the level and the distribution of household income among men and women, cash and in kind, before any intra-household transfers have taken place.

This level and distribution of household income govern the choice of the weekly food allowance provided to most women by their husbands. In addition to its indirect effects by the income variables, NTAE crop adoption may have a direct effect on intrahousehold transfers in so far as it represents an enhancement of male bargaining power. Therefore, depending on the preferences of a given household, it may reduce female claims on male income, of which el gasto is a major component.

The bottom half of figure I shows how the food allowance determines the level of cash expenditures on food, in combination with independent female income, the availability of own produced food, and any additional purchases the male head may make. Once again, NTAE crop adoption is depicted as having a direct as well as indirect impact on food expenditures: in so far as it represents a shift in the overall pattern of household members' economic roles and responsibilities, the relative weight of each of the contributions to food acquisition—male income, female income, and el gasto—may differ among adopters and non-adapters. Food expenditure in turn serves as an input, together with subsistence production, to household and individual food availability, adequacy, and diversity.

Impact of NTAE crop adoption on household income, women's income, and subsistence production

The analysis begins by examining the effect of NTAE crop adoption on household income levels and pre-transfer distribution. A common approach to evaluating the so-called treatment effect of participation in a given project or scheme, and that generally used in the literature on cash cropping, is to consider selection into the treatment to be exogenously determined. However, adoption of NTAE crops is a voluntary phenomenon, and there are strong reasons to believe that some of the same characteristics that influence it also influence the household's ability to generate income from other sources and the distribution of that income among household members. Any estimation of household income and distribution must therefore control for the self-selection of NTAE producers and the potential correlation of preferences. Heckman's two-step procedure is therefore used at various stages of the model.

The probit estimation of NTAE crop adoption (not shown here), the first step in the Heckman procedure, implies that farmers with greater extensions of owned land are more likely to adopt the new crops, but decreasingly so at the margin. Younger farmers are more likely to adopt than older ones, as are households with higher-quality land. Evangelical Protestants, who are often assumed to be less traditional farmers than their Catholic counterparts [2], are significantly less likely to grow export vegetables, controlling for land size and quality.

The results of this relatively simple specification of NTAE crop adoption are used to calculate the inverse Mills ratio (landa), which is included as an additional independent variable in several of the following estimations.

TABLE 1. Heckman two-step estimation of household income

N = 281; R2 = 34

Dependent variable:
HHINC = total annual household income.
Mean = 4407.70; standard deviation = 3812 16.

Independent variables:

TPROP = owned land, in cuerdas (1 cuerda = 0.11 hectares).
TNTAE = land planted in NTAE crops, in cuerdas.
ADLTEQV = number of adult equivalent household members.
AGEHOH = age of head of household, in years.
AGEJEFEM = age of female head of household, in years.
FEMJEF = 1 if single female head of household.
NOCH05 = number of children under the age of six years.
NODT11UP = number of daughters over the age of six years.
EDMHOH = formal education of male head of household, in years.
EDFHOH = formal education of female head of household, in years.
EVEN = 1 if head of household is evangélico (Protestant).
COMUN1 = 1 if household resides in Las Canoas.
COMUN2 = 1 if household resides in El Tablón.
COMUN3 = 1 if household resides in Santo Domingo El Rosario.
COMUN4 = 1 if household resides in Xejolón.
LAMBDA = inverse Mills ratio calculated from probit estimation of NTAE crop adoption (not shown here).

The results of an endogenous switching model of household income (table 1) firmly supports the hypothesis that NTAE crops are highly profitable for those who adopt them. The regression coefficient indicates that every additional cuerda (1 cuerda = 0.11 hectare) of land placed in NTAE crops yields over Q550 (US$110) in additional income on an annual basis. Evaluated at the mean household income of Q4,408, this represents 12.5% of the total income generated over a 12-month period. The regression coefficient on owned land is positive, but less than a tenth of the value of that on land to non-traditionals, and statistically not significantly different from zero, implying substantially lower returns to other crops [1].

The coefficients on male and female education are both positive, but only the man's years of schooling appear to have a statistically significant impact on total household income: an increase of 5% for every additional year, evaluated at the mean level of income.

The age of the head of household is weakly negatively associated with total income, and the age of the female head has a small and statistically insignificant positive relationship with the dependent variable. Households with single female heads can be expected to have substantially lower incomes, although this result is not statistically significant.

Protestants, although less likely to adopt NTAE crops, do on average have higher incomes, perhaps owing to their greater tendency to diversify their income-generation strategies [2]. Again, however, this result is not statistically significantly different from zero.

Of interest is the fact that the number of young children in a household is quite strongly negatively associated with income, even after controlling for adult equivalents. This may reflect the fact that not only do small children increase the consumer: worker ratio, but they also place demands on other potential income-earners, especially mothers and older daughters who are primarily responsible for their care.

Next, the hypothesis that NTAE crop adoption is a significant determinant of women's income is tested with a Tobit maximum likelihood model that corrects for self-selection. The regression results (not shown here) suggest that the extent of adoption of NTAE crops is not negatively associated with a woman's ability to generate independent income. The magnitude of the regression coefficient on the number of cuerdas devoted to NTAE crops is quite small (representing less than 1% of the mean annual income of a female head of household) and not statistically significantly different from zero. Younger women, women from larger and wealthier families, those with fewer young children, and single female household heads are all likely to have higher incomes. The age of the household head is also positively associated with the level of women's income. Religion, formal female or male education, number of older daughters, and community of residence do not appear to have an independent impact on women's income generation.

The third and final component of pre-transfer income to be estimated is the level of subsistence agricultural production, using an endogenous switching model and the same set of regressors as the previous two estimations. The dependent variable is the imputed value (in quetzales) of that portion of annual corn and bean production set aside for the family's own consumption. The estimation results (not shown here) indicate that, controlling for wealth held in land (which has a positive but quantitatively small positive association with the value of subsistence production), and correcting for the self-selection of households into non-traditional agriculture, the extent of NTAE crop adoption has a strong positive linear relationship with subsistence production of corn and beans. The regression coefficient implies that every additional cuerda of land devoted to NTAE crops is associated with a 9% increase in the value of milpa crops dedicated to household consumption. This lends strong support to work that suggests that cash-crop production does not necessarily have a deleterious effect on subsistence production, if land is sufficient and/or there are spillover effects such as the nitrogen-fixing properties of snow peas or the increased availability of chemical inputs [1]

With regard to the level and distribution of house-hold income, it appears that whereas the extent of production of NTAE crops is associated with substantial increases in both the total amount of income earned by a household and the value of subsistence agricultural production, there is neither a negative nor a positive statistical relationship between the amount of land devoted to export vegetables and the level of female earnings. These distinct effects should be kept in mind in the following discussion that considers the determinants of intra-household income transfers and food expenditures.

Impact of NTAE crop adoption on intra-household income transfers and food expenditures

The results of a Tobit estimate of el gasto (not shown here) indicate that, for the sample as a whole, the mean elasticity of the weekly food allowance with respect to household income is 0.32, and that with respect to female income is —0.03. Both of these results are statistically significant at less than the 1% level. This implies that for a household generating the mean total and female incomes, and supplying the average level of food allowance, doubling total weekly income to Q170 would increase el gasto by less than Q6, whereas a 100% rise in women's weekly income to Q15 would reduce the allowance by just over 50 centavos. The income transfer does not appear to be sensitive to the level of subsistence agricultural production, which is not surprising given that corn expenditures generally fall outside the woman's sphere of responsibility. Wives of Protestants receive on average Q3.67 more per week than their Catholic counterparts. As expected, single female heads of household are much less likely to receive any such allowance at all, although older children do tend to remit regular amounts of money that serve a similar purpose. Surprisingly, household size appears to have a quite small and statistically insignificant effect on the level of el gasto.

To test the hypothesis that the determinants of el gasto vary among adopters and non-adapters of NTAE crops, an endogenous switching model is estimated that allows the regression coefficients on each of the income variables to differ according to the NTAE adoption status of the household, while correcting for the common preference structure that influences both the adoption decision and the negotiated level of the food allowance (table 2).

Non-adopters. despite the fact that they are on average poorer than adopters (mean annual income Q3,327.47 v. Q5,273.28), provide their wives with a slightly higher mean weekly allowance (Q17.32 v. Q16.97) for food expenditures. Moreover, increases in total household income are to a substantially greater extent transferred to women for such expenditures in non-adopting households, although a large standard error on the interaction term of NTAE and the log of average weekly household income renders the regression coefficient statistically insignificant. Evaluated at the mean of each of the income quartiles for the sample as a whole, a Q10.00 increase in non-adopting household income translates into a Q2.86 increase in el gasto in the poorest households, a Q1.27 increase in quartile 2 households, a Q0.71 increase in quartile 3 households, and a Q0.33 increase in the wealthiest quarter of the sample. The corresponding figures for NTAE crop-producing households are Q1.43 for quartile 1, Q0.64 for quartile 2, Q0.36 for quartile 3, and Q0.17 for quartile 4.

Although the level of el gasto is moderately sensitive to changes in female income in non-adopting households every Q10 increase in a woman's weekly earnings decreases her allowance by Q0.59 at the mean NTAE crop adopters on average only alter el gasto with growth in female income levels by 40% as much, Q0.17 per Q10 increase in female earnings. As is the case with differences in the household income slopes, large standard errors render this contrast statistically insignificant.

The subsistence production component of income is weakly negatively associated with the level of el gasto in non-adopting households, but it is positively and statistically significantly related to the weekly food allowance in adopting households. The regression results indicate that a 10% increase in the weekly value of milpa production corresponds to a decline in el gasto of approximately 1% in non-adopting households, but an increase of the same proportion in adopting households. Why the relationship should differ between the two subsamples is not clear.

There is thus weak evidence (given the large magnitudes but also relatively large standard errors of the interaction terms) that distinct intra-household income transfer patterns obtain in adopting versus non-adopting households. In non-adopting households, increases in income are more readily transferred to women in the form of the weekly gasto, but women must also sacrifice a portion of these transfers in so far as they earn an independent income. In adopting households, el gasto is less responsive to changes in both household and female income, making women's earnings a potentially more important overall determinant of food expenditure levels.

TABLE 2. Tobit maximum likelihood estimation of el gasto with endogenous switching

N = 278

Dependent variable:.

GASEM = gasto per week received by female head of household.
Mean = 17.19; standard deviation = 13.29.

Independent variables:

LNHHYWK = log of weekly household income.
LNFMYWK = log of weekly income of female head of household.
LNVMILWK = log of weekly value of subsistence production of corn and beans.
ADLTEQV = number of adult equivalent household members.
FEMJEF= 1 if single female head of household.
EVAN = 1 if head of household is evangélico (Protestant).
COMUN1 = 1 if household resides in Las Cannoas.
COMUN2 = 1 if household resides in El Tablón.
COMUN3 = 1 if household resides in Santo Domingo El Rosario.
COMUN4 = 1 if household resides in Xejolón.
NTAE = 1 if adopting household, 0 if non-adopting household.
NTXLNYWK = NTAE x LNHHYWK.
NTXLNFW = NTAE X LNFMYWK.
NTAELSW =NTAE X LNVMILWK.

Since intra-household resource transfers in general, and the level of el gasto in particular, are choice variables for the household, and clearly associated with household and female income status and more weakly with NTAE crop adoption, it is appropriate to treat el gasto as an endogenous variable in the estimation of weekly food expenditures, the subsequent "block" of the model. The results of this estimation (table 3), which account for the endogeneity of el gasto and of NTAE crop adoption but treat household, female, and in kind income as exogenous, indicate that the mean elasticity of food expenditures with respect to total household income is 0.06 in non-adopting households and —0.01 in adopting households. Neither of the regression coefficients on total household income is statistically significantly different from zero. however.

TABLE 3. Heckman two-step estimation of weekly food expenditures

N= 273; R2 =.39

Dependent variable:

WKFOEXP = weekly cash food expenditures.
Mean = 43.51; standard deviation = 26.74.

Independent variables:

LNHHYWK = log of weekly household income.
FEMYWK = weekly income of female head of household.
LNVMILWK = fog of weekly value of subsistence production of corn and beans.
PRESTGAST = predicted value of el gasto.
ADLTEQV = number of adult equivalent household members.
EVAN = 1 if head of household is evangélico (Protestant).
FEMJEF = 1 if single female head of household.
COMUN1 = 1 if household resides in Las Canoas.
COMUN2 = 1 if household resides in El Tabló n.
COMUN3 = 1 if household resides in Santo Domingo El Rosario.
COMUN4 =1 if household resides in Xejolón.
NTAE = 1 if adopting household, 0 if non-adopting household.
NTXLNYWK = NTAE x LNHHYWK.
NTAEFYWK = NTAE x FEMYWK.
NTAELNSW = NTAE x LNVMILWK.
NTAEPRGS = NTAE x PREDGAST.

In the case of female income, disaggregating the sample reveals that the independent positive contribution of this variable to food expenditures is much stronger in adopting households than among non-adopters; this result is significant at the 3% level. Indeed, once distinct slopes are allowed for, the hypothesis that female income has no autonomous effect on food expenditures in non-adopting households cannot be rejected (p=.48), whereas every additional Q10 of income in the hands of the female head of household translates into a Q5 increase in the weekly food budget for households that have adopted the new crops. That is, in crop-adopting households, half of women's marginal earnings go toward increasing food expenditures, whereas women in non-adopting households do not devote an amount of their marginal income gains to food that is statistically different from zero.

Thus, in crop-adopting households, female income makes a positive contribution to the level of food expenditures over and above its contribution through the total household income effect. In non-adopting households, where el gasto is more responsive to increases in both total household and female income, no independent effect of female income on food expenditures can be discerned.

The value of self-produced food has a negative association with the level of cash food expenditures, more so in adopting than in non-adopting households, although neither regression coefficient is statistically significantly different from zero. The predicted value of el gasto has a strong and statistically significant linear relationship with food expenditures, as would be expected; this relationship is not significantly different among the subsamples. Weekly food expenditures increase by about Q3.50 per adult equivalent, are somewhat higher among evangelical Protestants, and are much higher in households with single female heads.

To evaluate the net relationship of NTAE crop adoption on market food acquisition, and compare this effect with several counterfactual situations, a simple experiment is presented in which the income generated from the new crops is distributed in alternative ways within and between households. Consider the marginal profit from a single cuerda of land devoted to broccoli or snow peas, estimated to be Q550. The model suggests that household income increases are associated with food expenditures in both a direct and indirect (by el gasto) manner. Similarly, increases in women's income can be expected to affect market food acquisition directly and indirectly by any induced changes in the weekly allowance. Table 4 summarizes the net effect of the profit from a cuerda of land planted in NTAE crops under four different circumstances: that the income accrues to the male head of a crop-adopting house hold; that the income accrues to the male head of a non-adopting household; that the income accrues to the female head of an adopting household; and that the income accrues to the female head of a non-adopting household.

Under the circumstance, which, it must be emphasized, is the one that actually happens, the regression results suggest that Q9.31 of the annual income increase of Q550 (an average weekly increase of Q21.15) goes toward increasing the weekly gasto. Drawing on the results of the estimation of weekly food expenditures, this Q9.31 translates into a Q14.46 rise in weekly food expenditures. However, the negative regression coefficient on household income for adopting households implies that Q1.18 must be netted out from this indirect increase, with the result that the profit from a cuerda of land in NTAE crops can be expected to increase weekly food expenditures by Q13.28 in the households that actually do adopt the new crops. Repeating this simulation for non-adopting households, the substantially larger regression coefficient on household income in the estimation of el gasto suggests that Q18.68 of the additional income would be transferred to the woman, and that the net effect on food expenditures would be Q35.12 almost 40% more than adopting households.

However, the reality is that non-adopters cannot and do not earn the substantial profits associated with the new crops. Thus, if we evaluate the expected level of el gasto at the mean income levels of each subsample, controlling for a series of other factors, adopters give an average of Q8.90 compared with non-adopters' Q6.26 per week. In the case of food expenditures, the predicted level for adopters, again evaluated at the subsample income mean and controlling for all the variants, is Q42.46, compared with non-adopters' Q46.68. This implies that despite the lower marginal propensity of male heads of crop-adopting households to transfer income gains to their wives, and despite the slightly negative income elasticity of food expenditures in adopting households, the absolute income increases associated with non-traditional agriculture, in combination with distinct values on in-kind and female income, are generally sufficient to compensate for these adverse effects.

Consider a distribution of that same Q550 such that it accrues to the female head of household. Net of the negative effects that such an income increase would have on el gasto, a boost of Q550 in women's earnings would raise food expenditures by Q35.50 in non-adopting households and by Q25.75 in adopting households. Although the difference relative to the male situation is not great in the case of non-adopters, it represents an increase of 94% for households producing NTAE crops. It is this latter group of women who devote over half of their marginal earnings to food expenditures, and it is in their hands that the most dramatic effect of an intra-household redistribution of income is realized.

TABLE 4. Predicted income effect on household food expenditures by gender and NTAE crop adoption

Conclusions

To assess the question of whether the policy goal of improving nutrition with increased access to income is achievable in the context of non-traditional export agriculture in Guatemala, the positive and negative, direct and indirect effects of the new crops on food acquisition must be evaluated together. As suggested in the literature, the central consideration is whether the increases in income brought about by the introduction of the new crops are substantial enough to overcome three principal barriers to nutritional improvement: substitution into status foods, displacement of subsistence crops, and adverse intrahousehold effects.

Substitution into higher-priced nutrients

Evidence shows that wealthier families diversify their food expenditures into luxury items such as meat, and obtain fewer of their calories and protein from maize; however, this is no more the case for NTAE crop adopters than for non-adopters. Depending on the intra-household distribution of these more expensive foods, which there is some reason to believe does not favour women and children [3], diversification could have positive or negative nutritional consequences.

Displacement of subsistence crops

The evidence presented here suggests that NTAE crop production has not had a negative impact on self-sufficiency in basic grains. Indeed, among some income groups, market acquisition of corn and beans is even greater for non-adopters than for adopters, and additional land planted in NTAE crops is associated with a significantly higher value of subsistence crop production. Whether this is attributable to greater agricultural productivity among farmers who have chosen to adopt the new crops, as others suggested [1], or to other factors not captured in the model, is beyond the scope of this analysis.

Adverse intra-household effects

It is in the realm of intra-household resource allocation, and in particular intra-household labour and income transfers, that adopters and non-adopters arguably demonstrate their most significant differences. With regard to the demand for female labour in the new crops, the apparent lack of adverse effects on women's ability to generate independent income suggests that these women may be absorbing the increased claims on their time by sacrificing activities other than those for which they receive some remuneration. Unfortunately, without complete time allocation data, it is impossible to test this hypothesis here.

At the same time, the lower marginal increases in el gasto with respect to total household income in adopting households may lie behind these women's observed greater contribution to food acquisition; that is, women absorb declining proportional male transfers by sacrificing other sorts of female expenditures, such as domestic technology, to maintain a desired level of food availability, which may also increase with higher incomes. The flexibility of these women with regard to the allocation of both their labour time and their income, combined with the sizeable profits that can be earned from NTAE crops, and continued (if not increased) self-provisioning in basic grains, appears to have largely compensated for the potentially adverse intra-household and food substitution impacts of the new crops. Women's adaptability to increased claims on their labour time and decreased marginal contributions to their expenditure budget, and their willingness to absorb the costs of participation in non-traditional agriculture thus play an important role in the relative success of non-traditional agriculture as a means of improving household food availability.

Acknowledgements

The financial support of the Inter-American Foundation, the Midwest Universities Consortium for International Activities, the John D. and Catherine T. MacArthur Scholars Program at the University of Wisconsin-Madison, the Tinker Foundation, and the University of Wisconsin-Madison Graduate School is gratefully acknowledged. The research for this paper was carried out while the author was a research associate at the Institute of Nutrition of Central America and Panama (INCAP).

References

1. von Braun J. Hotchkiss D, Immink M. Nontraditional export crops in Guatemala: effects on production, income, and nutrition. Washington, DC: IFPRI, 1989.
2. Annis S. God and production in a Guatemalan town. Austin, Tex, USA: University of Texas Press, 1987.
3. Nieves 1, Engle PL. Conductos de distribución intra-hogareña de alimentos en comunidades periurbanas de Guatemala: lecciones pare los programas de distribucion de alimentos. Arch Latinoam Nutr 1989;39(3):339-56.

Contents - Previous - Next