Food Security in Less Developed Countries, 1970 to 1990 Author(s): J. Craig Jenkins and Stephen J. Scanlan Source: American Sociological Review, Vol. 66, No. 5 (Oct., 2001), pp. 718-744 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/3088955 Accessed: 23/07/2009 14:46

Your use of the JSTOR indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to American Sociological Review.

http://www.jstor.org FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990

J. CRAIG JENKINS STEPHEN J. SCANLAN The Ohio State University Universityof Memphis

Despite a globalfood surplus, almost half of the world's less developed countries suffer significantproblems concerning food. Most social science and policy discus- sions offood security make the "foodavailability" assumptionthat increasedfood supply is the key to reducinghunger. Critics argue, however,that increasedfood supply has little impacton hunger and that the primaryculprits are entrenched inequalityand militarism.A lagged panel analysis offood supply and child hunger rates (1970-1990) shows that thefood supply has only modest effects on child hun- ger rates and thatfood supply is structurallyrooted in developmentprocesses (do- mestic investment,urban bias, foreign capital penetration)while child hunger is politically based in arms imports,internal violence and political . Populationpressure, tappedby increased age dependency,undermines both the supply offood and the population'saccess to it, and culturaldualism magnifies the effects of populationpressure on child hunger.The effects of economic growth "trickledown" to affect bothfood supply and child hunger,and economic growth is also positively correlatedwith political democratization,suggesting there is no short-term"trade-off" between growth, democratization, and social equity.

OOD IS THE MOST basic of human (LDCs) suffered a decline in aggregate food needs and is central to the discussion of supply, and more than a quarter suffered an and social development increase in child hunger (Bongaarts 1996; (Brandt 1986; Dreze, Sen, and Hussain 1995; FAO 1996a, 1996b). In the mid-1990s, more Food and Agricultural Organization of the than 840 million, or about 20 percent of the [FAO] 1996a; Kutzner 1991). LDC population, lacked sufficient food to Despite the "green revolution" and the sig- meet basic nutritionalneeds (United Nations nificant growth in internationalfood aid and Population Fund 1999), and more than 200 assistance, between 1970 and 1990 almost million children, or almost one-third of those half of the world's less developed countries under age five, suffered from severe malnu- trition (United Nations International Direct correspondenceto J. Craig Jenkins, Children's Emergency Fund [UNICEF] Dept. of , 300 Bricker Hall, 190 N. 1998). Malnutritionis a major barrierto eco- OvalMall, Ohio StateUniversity, Columbus OH nomic and social development, leaving popu- 43210 ([email protected]).The authorsshare lations unable to maintain normal lives and equallyin the authorshipof this paper.We ben- to be economically and socially productive. efited greatlyfrom advice from William Form, Most social science and policy discus- David Edward Edward Jacobs, Kick, Crenshaw, sions, including the existing major early MarianneAbbott, and Lane Kenworthyas well models of assume that in- as from the ASRreviewers. This re- warning famine, anonymous creased food is the to search was supportedby the National Science supply key reducing Foundation(# SBR-9710958),the MershonCen- hunger and that the effects of economic ter for InternationalSecurity, Ohio StateUniver- growth will "trickle down" to reduce hunger sity, andthe Dean'sDistinguished Fellowship for (FAO 1996a; Foster and Leathers 1999:67; StephenScanlan from the GraduateSchool, Ohio Quinn and Kennedy 1994). Critics have StateUniversity. questioned both assumptions. In their his-

718 AMERICAN SOCIOLOGICAL REVIEW, 2001, VOL. 66 (OCTOBER:718-744) FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 719 torical and contemporary studies of famine, foreign investment is relatively less benefi- Sen (1981) and Dreze and Sen (1989) show cial than domestic investment, a position that resourceful households rarely go hungry that is compatible with mainstream develop- despite aggregate food shortages and that the ment theory (Firebaugh 1992, 1996). poor are often hungry even when food sup- is Economic im- ply plentiful. growth may WHAT IS FOOD SECURITY? prove aggregate food supply, but because of unequal control of economic resources the Food security has been promoted by the poor and disadvantaged remain hungry. A United Nations (FAO 1996a, 1996b; United second criticism is that economic growth Nations Development Programme [UNDP] does not address basic needs because of 1994) and analysts of hunger and famine sectoral disparities created by early industri- (Brandt 1986; Dr6ze et al. 1995; Foster and alization (Kuznets 1955), economic depen- Leathers 1999; Sen 1981) as the most basic dency (Bornschier and Chase-Dunn 1985; human need and as a central indicator of ab- London and Williams 1990), and urban bias solute and physical well-being. (Bates 1981; Lipton 1977; Weede 1996). A Food security refers not only to an adequate third critique focuses on militarism, arguing aggregate supply of food, but also means that contemporary hunger is due to internal that "all people at all times have both physi- repression, civil war and arms races (also cal and economic access to basic food. This known as the "military famine" thesis) (Ac- requires not just enough food to go around. tion Against Hunger 2000; Cheatham 1994; It requires that people have ready access to de Soysa and Gleditsch 1999; Macrae and food" (UNDP 1994:22). We use two indica- Zwi 1994). And finally, there is the neo- tors: (1) food supply is measured as the Malthusian argument (Cohen 1995; Ehrlich mean daily per capita supply of calories and and Ehrlich 1990; Smil 1994)-that sus- protein (FAO 1996b), and (2) the child hun- tained high fertility is overshooting ecologi- ger rate is measured by the percentage of cal limits, creating overcultivation, exces- children under age 5 who are undernour- sive fertilizer use, and soil degradation that ished (UNDP 1994).1 leads to a fall in food production and an in- Table 1 charts food security trends be- crease in hunger. tween 1970-1975 and 1990 in all LDCs with We address these critiques using a lagged populations over 1 million. For food supply, panel cross-national analysis of food supply we use the five-year means for 1968-1972 and child hunger. Past cross-national re- and 1988-1992 to eliminate annual volatil- search has examined food supply (often mis- ity. These are derived from the food balance leadingly referredto as "food consumption") sheets of the FAO (1972, 1974, 1992, 1994) and has focused on the modernization/de- and include estimates of marketed and non- pendency debate. This past research shows marketed food from all domestic and inter- that economic growth increases food supply national sources (including self-production (Firebaugh and Beck 1994) and that export and aid). This is the most reliable indicator dependence, foreign capital penetration, and of food supply available and has broader internationaldebt are detrimentalto the food temporal and country coverage than any supply (Bradshaw et al. 1993; Wimberley 1 1991; Wimberley and Bello 1992). This ne- "Food supply" has sometimes been mis- glects the more critical question of hunger labeled "food consumption"(Wimberly 1991; and ignores other explanations, such as de- Wimberly and Bello 1992), implying that aggre- mocratization, urban bias, population pres- gate supply reduces hunger. Tweeten and sure, the evolutionary advantages of agrar- McClelland (1997) refer to "availability" and "access"measures and note thatthere is also the ian density, and the role of militarism. Some micro of the of of this work (Bradshaw et al. 1993; question metabolizing ingested food. Both food supply and hunger rates could Wimberley 1991; Wimberley and Bello be seen as "access" in that the food has also the questions 1992) misinterpreted negative supply taps the access of LDCs as aggregate units effect of foreign capital penetration, seeing to the global food supply,while the hungerrate it as indicating an absolute decline in food tapsthe access of householdsand specific groups supply when in fact it merely indicates that to the country-levelfood supply. 720 AMERICAN SOCIOLOGICAL REVIEW

Table 1. Means and Standard Deviations for Measures of Food Supply and Child Hunger: Less Developed Countries, 1970 and 1990

Daily Calories Daily ProteinGrams Percentageof Hungry per Capita per Capita Childrenunder Age 5a World Region Circa 1970 Circa 1990 Circa 1970 Circa 1990 Circa 1975 Circa 1990 All less developed 2,227.47 2,397.43 57.62 60.91 28.88 23.42 countries(N = 88) (299.12) (416.82) (12.97) (14.47) (15.34) (13.81)

Asia and Oceania 2,184.0 2,426.32 54.89 60.51 46.29 38.24 (N = 21) (287.37) (377.67) (38.20) (13.74) (16.25) (13.70) Latin America 2,364.75 2,501.20 61.84 63.22 15.71 11.48 (N = 21) (345.77) (310.96) (14.55) (11.84) (7.19) (6.74) North and 2,357.65 3,004.76 64.13 81.18 18.20 10.70 the Middle (290.45) (316.66) (12.12) (9.96) (4.75) (2.83) East (N = 11) Sub-SaharanAfrica 2,130.27 2,126.96 54.70 53.40 31.52 27.24 (N = 35) (239.19) (269.24) (11.37) (10.91) (10.00) (8.74)

Final caloric supply 2,249.56 2,408.22 model (N = 55) (302.85) (383.09) Final proteinsupply 58.25 60.92 model (N = 55) (12.86) (13.10) Final hungrychildren - - - - 28.51 22.53 model (N = 63) (16.35) (14.64)

Countrieswith Food SecurityDecline Countriesthat lost 100 or more mean daily calories per capita (N = 33): Afghanistan,Angola, Argentina,, Burundi, , Cameroon, Central African Republic, Chad, Cuba, DemocraticRepublic of Congo, Guinea,Haiti, Kenya, Laos, Liberia, Madagascar,Malawi, Mongolia, Mozambique,Nicaragua, North Korea, Paraguay,Peru, Senegal, SierraLeone, Somalia, South Africa, Togo, Uganda, Uruguay,Zambia, Zimbabwe Countriesthat lost 3 or more grams of daily proteinper capita (N = 29): Afghanistan,Angola, Argentina,Burundi, Cambodia, Cameroon, Chad, Cuba, DemocraticRepublic of Congo, Guinea,Haiti, Iraq,Ivory Coast, Jamaica,Kenya, Laos, Liberia,Madagascar, Malawi, Mongolia, Nicaragua,North Korea, Rwanda,Sierra Leone, Somalia, Uganda,Uruguay, Zambia, Zimbabwe

Countriesin which the percentageof hungrychild underage 5 increased 1 percentor more (N = 12): Afghanistan,Angola, Burundi,Madagascar, Malawi, Mozambique,Nigeria, Senegal, SierraLeone, Uruguay,Zaire, Zambia Note: Numbersin parenthesesare standarddeviations. a Weight-for-agemeasure (Foster and Leathers 1999:62-63). other available cross-national indicator (Fos- for-age" method, also known as the ter and Leathers 1999:66-73). However, it "Gomez system" (Foster and Leathers does not tap hunger (i.e., the differential ac- 1999:62-63), is the most reliable gauge of cess of households and particular groups to preschool children who are at risk for pri- adequate food). mary (i.e., from insufficient calories) and To measure hunger, we use the percent- secondary (i.e., from diarrhealinfection that age of children under 5 years of age whose depletes nutrients) undernutrition.2 It ad- body weight is more than two standard de- viations below the median weight-for-age 2 Other types of unhealthy weight are not cap- of the country's population. This "weight- tured by this weight-for-age measure (e.g., obe- FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 721 justs for the physical characteristics of the EXPLAINING FOOD SECURITY country's population and is available begin- ning in 1975 from country surveys con- We examine six explanations of food secu- ducted by the World Health Organization rity: (1) modernization, (2) economic depen- (FAO 1996b; UNDP 1994). Insofar as child dency, (3) urban bias, (4) neo-Malthusian undernutritionis correlated with that in the population pressure, (5) ecological evolu- population at large or among ethnic minori- tionary processes, and (6) militarism. Be- ties, women, and peasants (Bhuiya and cause there are few prior cross-national stud- Karim 1989; Kelly 1992), it serves as a ies of food security to serve as a benchmark, proxy for hunger in the general population. we draw on the findings of past studies of Because the measure is not available annu- social welfare, especially those on the physi- ally, we focus on the change from 1975 to cal quality of life and other basic needs. 1990, the longest time-period currently available. MODERNIZATION The overall trend among LDCs is toward increased food security (Table 1). Between emphasizes internal 1970 and 1990, caloric and protein supply sources of . It argues increased by 7.63 percent and 5.71 percent, that domestic investment and educational respectively, and mean child hunger rates growth create industrialization and cultural dropped from 28.9 to 23.4 percent. , modernization, which in turn contribute to however, has been regionally uneven with economic growth, societal integration, and the greatest improvements in North Africa stronger institutions for providing social and the , moderate change in welfare. In some formulations (Rostow Asia and Oceania and Latin America, and a 1960), economic savings and investment in decline in sub-Saharan Africa. Yet, by the physical capital are the key to an industrial weight-for-age measure, more than a third of "take-off," while for others (Inkeles and all children are estimated to be hungry in Smith 1974; Kerr 1969), including the Asia and Oceania in 1990, more than one- "new" or "endogenous growth" models quarterof children in sub-SaharanAfrica in (Barro 1991), investment in human capital 1990, and 10 percent or more in the other via increased education creates a labor force world regions. with modern motivation, greater skills, so- Table 1 lists the LDCs with significant cial mobility, and adaptability to modern negative changes and the means and stan- . While some argue that there is a dard deviations for the samples used in the short-termtrade-off between growth and eq- regression analysis below. For food supply uity (Arrow 1979; Letwin 1983), in the long we list all LDCs that lost 100 or more mean run, it is assumed that economic growth daily calories per capita and 3 or more grams "trickles down" to increase social welfare. of daily protein per capita; for child hunger International trade and investment are rates, we list all LDCs with an increase of 1 viewed as socially beneficial, stimulating percent or more. Over half of the countries economic development, the diffusion of new in sub-Saharan Africa suffered declining technology and practices, and improvements food supply. Estimated child hunger de- in social welfare. A significant body of re- clined everywhere except for 10 countries in search shows that economic development, sub-SaharanAfrica, civil war-torn Afghani- industrialization, education, and urbaniza- stan, and Uruguay. What drives these trends tion improve the physical quality of life and in food security? other social welfare measures (Bullock and Firebaugh 1990; Jackman 1975; Shen and Williamson 1997; Williamson 1987), and sity, iron deficiency) but, for a study of the that economic increases food of it is the best avail- growth supply, prevalence child hunger, infant and adult able measure. It the best overall mea- survival, longevity (Bullock provides and and Beck sure of both linear growth and body proportion Firebaugh 1990; Firebaugh and is thus superior to measures of stunting and 1994). wasting (de Onis et al. 1993; FAO 1996b:65-66; Modernization theorists disagree over the Foster and Leathers 1999:62-63). role of the state. Some focus solely on capi- 722 AMERICAN SOCIOLOGICAL REVIEW tal investment and industrialization (e.g., berley and Bello 1992), and contributes to Rostow 1960), ignoring political institutions higher child and infant mortality (Bradshaw and processes, while others treat state-build- et al. 1993; Firebaugh and Beck 1994). Eco- ing and national integration as central to nomic dependency also contributes to modernization (e.g., Huntington 1968; Kerr sectoral disarticulation, which in turn is as- 1969). Among the latter, one view is that in- sociated with higher child mortality, higher dustrialization creates new groups and po- crude death rates, and lower physical well- litical demands that have to be regulated to being (Breedlove and Armer 1997; Stokes sustain growth and political stability. Thus, and Anderson 1990). These effects of eco- there is a "growth-" trade-off nomic dependency should extend to child (Huntington 1987): Political hunger. are vulnerable to political instability, and In recent years, dependency theorists have "strong" authoritarianregimes are better at seen foreign investment as even more detri- regulating these new demands and introduc- mental to social welfare because of the glo- ing the economic interventions needed for bal shift in core-peripheryrelations since the industrialization (Apter 1965; Huntington 1960s (Bornschier and Chase-Dunn 1985; 1968; Janowitz 1977). Others argue that de- Frank 1967). Several processes are respon- mocratization boosts economic and social sible. First, multinational corporations development by allowing the lower classes (MNCs) repatriate most of their profits and to organize and press their interests, which discourage domestic firm formation and in- inhibits corruption, encourages growth-rein- vestment, thus reducing domestic economic forcing policies, including public invest- growth and social welfare benefits. Second, ments in nutrition, education, and health, foreign investment is more dependent on ad- and provides greater political stability vanced technology, absorbing less labor and (Goldsmith 1986; Goodell and Powelson creating a smaller number of high-paying 1982; Khohli 1986; Sorensen 1991). Studies jobs, which contributes to income inequality show that political democracy is positively and urbanpoverty. Third, foreign investment correlated with improved physical quality of is largely concentrated in export industries life (London and Williams 1990; Moon and and has weak marketties and economic mul- Dixon 1985; Wickrama and Mulford 1996), tipliers with domestically oriented produc- basic needs fulfillment, more balanced ur- tion. Fourth, MNCs oppose social programs ban-rural development, and lower income that benefit the lower classes, disguise prof- inequality (Crenshaw 1992; Goldsmith its that would otherwise be taxed, and pres- 1986; Sorensen 1991), which may also pro- sure governments to make infrastructurein- mote food security. vestments that divert funds from social pro- grams. Thus, although flows of foreign in- vestment create economic the ECONOMIC DEPENDENCY may growth, long-term impact of foreign capital penetra- The core of economic dependency/world tion should be negative. systems theory is the argument that interna- Several studies of the foreign capital pen- tional trade and investment impose negative etration (PEN) ratio have misinterpreted its social welfare effects. Drawing on theories negative effect as indicating an absolute de- of unequal exchange (Amin 1976; cline in economic growth and social welfare Emmanuel 1972), one argumentpoints to the (Bradshaw et al. 1993; London and Smith long-term decline in the international terms 1988; London and Williams 1990; of trade for primary versus processed goods Wimberley 1991; Wimberley and Bello and the sectoral disarticulation (or disparity) 1992). As Firebaugh (1992, 1996) notes, the associated with export dependence. Studies foreign capital penetrationratio is essentially have shown that export dependence lowers the ratio of foreign to total capital stock.3 economic growth and the physical quality of life (Delacroix and Ragin 1981; London and 3 This is true whether the indicator used is for- Williams 1990; Ragin and Bradshaw 1992), eign direct investment stock over total capital reduces aggregate food supply (Bradshaw et stock, GDP, or adds an adjustment for popula- al. 1993; Firebaugh and Beck 1994; Wim- tion size. FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 723

Thus, a negative effect merely means that Robinson, and Syrquin 1986), and lower foreign investment is relatively less benefi- physical quality of life (Breedlove and cial to economic growth than domestic in- Armer 1997; Stokes and Anderson 1990). vestment (Firebaugh 1992, 1996), not that it However, these inequities may simply be a produces an absolute decline in economic result of the economic backwardness and growth or social welfare. As Dixon and cultural dualism of traditional rural social Boswell (1996a) argue: "often the choice systems (Alderson and Nielsen 1995; [facing LDCs] is not between foreign capital Nielsen 1994). Because we lack direct mea- and domestic capital but between foreign sures of the political power of urban elites, capital and no capital at all" (p. 547). We re- we control simultaneously for the change in fer to this as the "relativebenefit" thesis. Fol- rural/urban disparity net of simultaneous lowing Dixon and Boswell (1996a, 1996b), controls for the initial level of sectoral dis- we test the interaction of the foreign capital parity, agricultural labor force, and the de- penetration ratio with the domestic invest- velopment level. We treat the change in dis- ment rate (to represent relative decapital- parity as indicating urban bias (i.e., under- ization) and economic growth (to represent investment in and exploitation of the rural relative distorted growth), but following sector) when it is significant net of these Firebaugh's (1996) argument, we interpret other controls. The initial level of rural/ur- interactions as reflecting the relative benefits ban disparity is treated as indicating a heri- of foreign versus domestic investment in- tage of cultural dualism. Some argue that stead of an absolute decapitalization or dis- dependency is responsible for urban bias tortion effect. (Gugler 1997; London and Smith 1988; Stokes and Anderson 1990), which we test the effect of economic URBAN BIAS by examining depen- dency on the change in sectoral disparity. A third approach is urban bias theory, which focuses on the of urban political power elites, NEO-MALTHUSIAN PRESSURE arguing that elites use the state to channel resources to the cities and exploit the peas- Neo-Malthusians argue that high fertility antry, who, although large in number, are and lead to an ecological geographically dispersed, disorganized, and "overshoot" in terms of overcultivation, ex- subject to the "free-rider" problem (Bates cess fertilizer use, deforestation, desert for- 1981; Lipton 1977; Weede 1996:65-87). mation, and soil degradation, all of which Studies show that overvalued currency, pro- should undermine food security (Cohen tective import tariffs, high agriculturaltaxes 1995; Ehrlich and Ehrlich 1990; Ehrlich, and export duties, and the creation of mo- Ehrlich, and Daily 1993; Smil 1994). These nopolistic crop procurementand export mar- variables are sometimes gauged by the den- keting boards discourage agriculturalinvest- sity of rurallabor force normed against mea- ment and contribute to sectoral disparity in sures of arable land, termed "agrarianadver- productivity, thereby undermining agricul- sity" (Firebaugh 1979). Increased age de- tural development (Lofchie 1997). The most pendency (essentially expanded household important symptom is sectoral disparity in size owing to high fertility) is especially det- investment, which means that the productiv- rimental in rural areas because this pressures ity of agricultural labor is about half of that rural households to put marginal land into in nonagriculture (Lipton 1977:189-215). production and intensifies land use, thus de- This bias is not only inequitable, producing grading the soil and reducing agricultural rural/urbanincome inequality and a signifi- output, thereby undermining food security. cant social welfare gap, but is also ineffi- There is considerable debate about such cient, undermining agriculturaldevelopment claims. As noted above, the "green revolu- and food production. tion" and the expansion of cultivated lands Studies show that rural/urbandisparity is have kept agricultural production ahead of associated with lower economic growth population growth in the LDCs (Bongaarts (Bradshaw 1987; Bradshaw and Noonan 1996). Elsewhere (Scanlan 2001), we show 1997), greater income inequality (Chenery, that rural and total population growth do not 724 AMERICAN SOCIOLOGICAL REVIEW affect food security, but slow downward imports, military instability, repression, and change in fertility and increased age depen- civil wars reduce the food supply and con- dency reduce food supply and contribute to tribute to child hunger (Scanlan and Jenkins hunger. We extend this observation here by 2001). Here we examine a broader range of examining the interaction between age de- violence measures and investigate how these pendency and sectoral disparity, testing the interact with development processes. argument that cultural dualism magnifies Peasants have an eco- population pressure. POSSIBLE INTERACTIONS AND nomic incentive to maintain high fertility, MEDIATING PROCESSES which leads to overcultivation, use of mar- ginal land, and agriculturaldecline. The factors identified by the theories dis- cussed above may also mediate and interact with each other. For modernization ECOLOGICALEVOLUTIONARY THEORY example, should produce a demographic transition, A rival interpretation to agrarian density is whereas economic dependency should slow advanced by ecological evolutionary theory. fertility decline (London 1987), thus alter- This theory contends that traditional "plow" nately countering and aggravatingpopulation agriculture and resulting high levels of pressures. If the "democraticinstability" the- agrarian density facilitate rapid industrial- sis (Huntington 1968) is correct, democrati- ization and thus improved social welfare zation should give rise to militarism and in- (Lenski, Lenski, and Nolan 1995). Studies ternal violence, thereby disrupting food se- have found that agrariandensity accelerates curity. Or alternatively, dependent develop- fertility decline, promotes rapid industrial- ment may lead to bureaucratic-authoritarian ization, and lowers income inequality (Cren- rule, military coups and internal repression shaw 1993; Crenshaw and Ameen 1993). (Cardoso and Faletto 1979), which should Thus, agrarian density may also improve also undermine food security. Thus milita- food security. rism should stem from underlying structural development processes associated with either modernization or economic MILITARISM dependency. If, however, military instability is not rooted in A final thesis is the "military famine" argu- these development processes, then it may fit ment that militarism is responsible for hun- a distinct "military famine" thesis according ger and that violent conflict, political repres- to which hunger is largely a political prob- sion, and structuralinequality, not the aggre- lem. There is also the long-standing debate gate food supply, is the major cause of hun- about rural/urbandisparity, which some at- ger (Action Against Hunger 2001; Cheatham tribute to foreign investment and export de- 1994; de Soysa and Gleditsch 1999; Macrae pendency (Gugler 1997), urban bias (Brad- and Zwi 1994; Poleman 1997). Militarism, shaw 1987; Lipton 1977), and long-standing refers to the use of military force to regulate cultural dualism (Alderson and Nielsen political conflicts (Ross 1987), including 1995; Nielsen 1995). Finally, there is the both civil and international relations. Mili- "food availability" question about whether tarism includes international arms races, food supply reduces hunger. If we find little which result in reduced public investment in impact of food supply on child hunger and basic needs such as nutrition, health, and different causal factors emerge for food sup- education programs (Brzoska and Ohlson ply and child hunger rates, then hunger must 1987), and civil and interstate wars, state re- be assumed to have distinct origins. Given pression, and military coups. Such internal these arguments, we examine the data with violence disrupts food production and distri- an eye to multiple explanations, possible bution, including international food assis- overlaps, and interactions. tance, because rival forces use food as a po- litical to control and ter- weapon populations METHOD AND MEASUREMENT ritory (de Soysa and Gleditsch 1999), often targeting ethnic minorities and the rural We use a lagged panel or conditional change poor. In earlier work, we found that arms model. We predict the value of the depen- FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 725 dent variable for food supply and child hun- To provide a substantial time lag, we pre- ger, (Yt2), net of its earlier value, Ytl, and in- dict food supply and child hunger for 1990 dependent variables measured at an earlier net of lagged controls for 1970-1975. Some time. This is the best method for capturing of our arguments are about long-standing structural trends that display considerable structural conditions (e.g., foreign capital stability over time in which values at an penetration), while others involve change early point have a strong effect on later val- (e.g., increased investment in human capi- ues (Allison 1990; Finkel 1995:7-11). In ef- tal). We therefore measure structural vari- fect, this model estimates the impact of in- ables circa 1970 and changes in variables dependent variables on change in a depen- between 1965-1970 and 1985-1990. To dent variable, which thus provides an assess- check the possibility that structural condi- ment of the longitudinal trend (Hannan tions might generate persisting patterns 1979). masked by our lagged panel, we also exam- This model also has several advantages ine cross-sectional models and report the over a cross-sectional and an unconditional small differences from the lagged panel change-score design. First, relative to a analysis. To maintain consistent signs in the cross-sectional design, it helps rule out re- models (i.e., positive signs indicate favor- ciprocal effects and possible spuriousness. able change in food security), we use the Second, it is more appropriatefor testing ar- percentage of healthy weight children to in- guments about structuraltrends that involve dex improvements in child hunger rates. a significant time lag. For example, the for- We draw on theory and past research to eign capital penetration and cultural dualism define how change in independent variables (rural/urbandisparity) arguments are about is represented. First, where the question is a structuraleffects that take several decades to simple structuraltrend, we use a first-differ- become fully manifest. Third, because it ence score (Xt2-Xtl). Second, where the time controls for the typically strong correlation period of exposure is critical (e.g., growth between the dependent variable at two points in age dependency) we use the average an- in time, including the possibility of serial nual percentage change (100x{[(Xt2-X,) / correlation of measurement errors, it pro- number of years] /Xtl}). Third, where the is- vides a conservative estimate of the effects sue is the amount of change relative to a of independent variables (Hannan 1979). base value, regardless of time exposure (e.g., Fourth, it improves on unconditional increased export dependency), we use the change-score models, which assume that the percentage change (100 x [(Xt2-Xtl) / Xtl]) lagged endogenous variable has no effect on Finally, where compounding should affect later values. This assumption is unrealistic the process (e.g., the rate of investment or with a phenomenon like food security that economic growth) we use one of several tends to remain stable over time with small compounding formulas. Recent studies of incremental changes. A lagged panel model the compounding of the rate of foreign in- is also superior to fixed-effects models or vestment (de Soysa and Oneal 1999; Dixon the "method of first difference" (Allison and Boswell 1996a, 1996b; Firebaugh 1992) 1990) insofar as it allows us to examine in- have used the following formula: variables that are dependent long-standing 1973K structural conditions as well as trends AnnualIr =10x r (1967K -1 , (Finkel 1995:7-8). Although it is more vul- nerable to specification bias than fixed-ef- where n is the number of years for the fects models, we test a broad set of theories, change, and 1973 and 1967 are the ending including dummy controls for region that and starting points for the time period for K. test for unmeasuredregional factors. We also This is the mathematical equivalent to an- checked all of our equations for hetero- nual compounding. To capture annual com- skedasticity by inspecting plots of residuals pounded economic growth, we use the natu- and using the Breusch-Pagan test in ral log of the ratio of GNP per capita in 1990 LIMDEP 6.0 (Breusch and Pagan 1979; to the natural log of GNP per capita in 1970 Judge et al. 1985) and found no evidence of or ln(Yt2/Ytl) (Jackman 1980:606). problems. The basic equation for our models is: 726 AMERICAN SOCIOLOGICAL REVIEW

Yt2=ao +al (Ytl)+ a2 (Xtl) tions as well as testing all variables sequen- - tially. We discuss below the few instances +a3(Zt2 Ztl )+...+ a, (1) of multicollinearity that we found. where Y is food supply or child hunger, X is a structural and Z is a condition, process rep- MEASUREMENT resented by a first-difference score. An additional substantive question man- Table 2 shows the measurement and data dates that food supply be examined in terms sources for our variables. Most of our mea- of change. Famine is not a sudden event but sures have been used before so we discuss a long-term development. Thus, it is impor- fully only those that are distinctive. As tant to capture structural trends in hunger. noted, the measure of food supply is based Anthropometric research (FAO 1996b:128- on the mean of the five years surrounding 42; Foster and Leathers 1999:66-73) also 1990 (i.e., 1988-1992) and 1970 (i.e., 1968- shows that there is no simple caloric or pro- 1972), while child hunger is measured by tein threshold for undernutrition. Healthy the percentage of children who are of food intake is always somewhat relative to healthy weight. These are the most recent the climate, culture, age and sex composi- food security measures currently available tion, and average body sizes of the popula- and allow us to test the food availability tion in question. Thus, North Korea, which thesis. had 2,867 mean daily calories per capita in Five measures tap modernization. First, 1990, experienced widespread famine five GDP per capita taps the level of economic years later with an 18 percent drop to 2,360 development and serves as a control variable mean daily calories per capita, despite the throughout.4It is highly skewed so we use fact that this was considerably higher than the natural log. Second and third, invest- the daily caloric supply in almost all of sub- ments in physical and human capital are SaharanAfrica. Thus, as a gauge of food se- gauged by the first-difference changes in do- curity, the change in food supply is often mestic investment and in secondary school- more relevant than the absolute level. ing rates. A fourth measure is political de- We present results using the maximum mocratization, gauged by the first-difference number of cases available for each model, change in political democracy (1965-1985). which reduces the vulnerability to sub- Finally, to tap political instability, we use sample differences. In Table 1, the means military instability based on the morale, loy- and standarddeviations for the countries in- alty, and level of coup activity by the mili- cluded in the final models are close to those tary during the 1970s and 1980s. We also in- for the full sample of LDCs for which there clude a set of internal violence and political is available data on caloric and protein sup- repression measures to evaluate the "demo- plies and child hunger rates; this suggests cratic instability" thesis (see militarism dis- that the regression results are representative cussion below). of the broader set of LDCs. We also tested Economic dependency is measured by: ex- these equations using a constant N, which port dependency based on the percentage produced similar results. We also conducted change in primaryproducts exports as a per- exhaustive tests for influential cases using cent of total exports;5 foreign capital pen- standard methods (Bollen and Jackman 1985) and found none. On the caloric analy- 4 Variablesmeasuring energy consumption per ses, Nigeria and Madagascar were negative capitaand industriallabor produce identical re- outliers, and on protein supply, Nicaragua sults, but they slightly reduce the available and Mauritania were negative outliers. In sample size. Thus, we use the GDP per capita measure. the child hunger analysis, and Zambia 5 had more child than We also testedthe 1970 level andpercentage hunger predicted. 1970to 1985in concentration These were not influential change commodity cases, however, basedon the threecommodities as a as them did not affect our results. top percent- removing age of totalexports. And we testedexport inten- Thus, we include them in the results shown. sity as measuredby the 1970 level and percent- To test for multicollinearity, we used vari- age change 1970-1985 in exports/GDPratio. ance inflation factors and bivariate correla- None of these were significant. The primary FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 727

Table 2. Measurement of Variables Used in the Analysis

Variable Descriptionand Data Source DEPENDENTVARAIBLES: Food supply The five-year mean calories per capita and grams of proteinper capita, 1988-1992 (FAO 1994, 1992) and lagged control, 1968- 1972 (FAO 1972). Percentageof healthy The percentageof childrenunder 5 who are of healthy weight in weight children 1990 and the lagged control, 1975. In the regressionmodels, positive signs indicate fewer childrenof unhealthyweight (UNDP 1994). INDEPENDENTVARIABLES:

Modernization Theory Level of economic Real gross domestic productper capita (1980 U.S. dollars), 1970 development(ln) and 1975 (Summersand Heston, 1991). Domestic investment Average annualpercentage change in real domestic investment (1980 U.S. dollars) over GDP, 1970-1980 (WorldBank 1994). Humancapital investment First-differencechange in secondaryschooling rate constructed as the ratio of total secondaryschool enrollmentto the population of the age group that officially correspondsto that level of education, 1970-1985 (WorldBank 1997). Political democratization First-differencechange in index of political democracy, 1965- 1985 (Gurr1989). Militaryinstability Measurebased on militarymorale and loyalty, coded for the 1980s as utterlyunstable = 4, unstable= 3, barely unstable= 2, not wholly stable = 1, stable = 0 (Kidronand Smith 1983 and supplementedby Morrisonet al. 1989).

Dependency Theory Exportdependency Percentagechange in primaryproducts as a percentageof total exports between 1970 and 1985 (WorldBank 1997). Foreign capital penetration Total foreign stock of direct investmentin a countrydivided by ratio (PEN ratio) total capital stock, 1967 (Ballmer-Caoand Sheiddegger 1979). Foreign debt service load Percentagechange in the debt service ratio (debt payments divided by total exports) between 1970 and 1985 (United Nations Conferenceon Tradeand Development 1975-1988; WorldBank 1997). Economic growth The logged ratio of the startpoint/ endpointof real gross domestic productper capita (1980 U.S. dollars), 1970 to 1985 (Summersand Heston 1991). Also used as a test for the "food availability"thesis.

Urban Bias Theory Rural/urbandisparity ratio 1965 The ratio of percentnonagricultural GDP to percentagricultural (culturaldualism) GDP divided by the ratio of the percentof nonagriculturallabor force to percentagricultural labor force, 1965 (WorldBank 1997). Changein rural/urbandisparity The first difference change in the rural/urbandisparity ratio, 1965 to 1985 (urbanbias) ratio,1965-1985 (WorldBank 1997).

Neo-Malthusian and Ecological-Evolutionary Theories Age dependency Average annualpercentage change in the ratio of the proportion of the populationunder age 15 to those over age 15, 1970-1985 (WorldBank 1994). Agriculturaldensity The agriculturallabor force divided by the surface area of arable land in squarekilometers, 1970 (WorldBank 1997). (Continued on next page) 728 AMERICAN SOCIOLOGICAL REVIEW

(Table 2 continued) Variable Descriptionand Data Source

Militarism Theory Arms imports Averageannual percentage change in the valueof armsimports over real GDP between 1970 and 1985 (Brzoskaand Ohlson 1987). Ethnicpolitical discrimination Thesummed percentage of thecountry population that is a minorityexperiencing severe political discrimination in the form of neglect, social ostracism,and political exclusion and repression1970 to 1985 (Gurr1993). Internalviolence Dummy variablefor the presence of civil war or genocide/ politicide, coded yes = 1, 1970-1990 (Fein 1993; Harff and Gurr 1989; Kidronand Smith 1983, 1991; Singer and Small 1993). Presenceof civil war Dummy variablefor the presence of civil war, coded yes = 1, 1970-1990 (Kidronand Smith 1983, 1991; Singer and Small 1993). Presenceof genocide/politicide Dummyvariable for thepresence of genocide/politicide,coded yes = 1, 1970 (Fein 1993; Harff and Gurr1989). etration (foreign direct capital stock in 1967 gument (see dependency/world system dis- divided by total direct capital stock in cussion on p. 722). 1967);6 and foreign debt service load de- Urban bias is measured by the first-differ- fined as the percentage change in debt pay- ence change in Bradshaw's (1987) rural/ur- ments over total exports between 1975 and ban disparity index (1965-1985) net of si- 1985).7 Foreign investment stock is a struc- multaneous controls for the initial level tural factor so we use the initial stock ratio. (1965) of disparity (which captures cultural The key issue for exports and debt is the ex- dualism), the development level, and the tent of economic squeeze, making the per- percentage of the labor force in agriculture. centage change the most relevant measure. Because this is an indirect proxy for urban We use the interaction of the foreign capital bias, which is a political process, it is im- penetration (PEN) ratio with economic portant to control for the last three terms to growth and domestic investment (see Table make sure that we are tapping urban bias. 2 for definition) proposed by Dixon and This ensures that any effects are not simply Boswell (1996a) and, following Firebaugh due to development level or a large agricul- (1992, 1996), we interpret these as reflect- tural labor force, which could be construed ing the relative benefits of foreign versus as merely a low level of development. domestic investment rather than an absolute Population pressure is measured by the distortion of growth or decapitalization. We average annual rate of change in the age de- also include a set of measures for military pendency ratio between 1970 and 1985. We instability, internal violence, and repression use the ratio of the population under age 15 to evaluate the "authoritarianrepression" ar- to that over age 15 because less than 5 per- cent of the average LDC population is over productsvariable is more centralto the unequal 65. The effects of the more conventional age exchangethesis and was the only measurethat dependency ratio (population under age 15 was ever significant. plus over age 65 as a percentage of the total 6 We preferthis morestraightforward ratio, but population) were slightly weaker, suggesting alternativeversions directstock di- (e.g., foreign the "youth bulge" is critical. (We also used vided by GDP with and withoutthe controlfor the annual in the size and Chase-Dunn average change fertility population [Bornschier which similar but 1985:71-72]) identicalresults. rate, produced slightly produced weaker is inter- 7 We also tested the 1970 level and change results.) Agriculturaldensity 1970 to 1985 in internationaldebt load GDP and preted as representing rural adversity debtload percapita, but these were never signifi- (Firebaugh 1979) or, following ecological cant. evolutionary theory, the organizational ad- FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 729 vantages of traditional "plow" agriculture there is an equity versus growth trade-off- (Lenski et al. 1995).8 These theories predict that growth produces greater inequality, and opposite signs on agrariandensity. increased inequality (including democratiza- We tap militarism by: the average annual tion) undermines growth incentives (cf. rate of change in arms imports over GNP Huntington 1987 and Kenworthy 2000). Ad- (1970-1985); the intensity of ethnic political vancing a different argument with a similar discrimination;9 and internal violence as outcome, economic dependency theories measured as the presence of a civil war or a contend that foreign investment creates genocide/politicide between 1970 and growth without social welfare (or "distorted 1990.10The military instability measure (dis- growth"), which likewise "trades off" cussed above as a modernization and a de- against equity. pendency variable) can also be viewed as an Table 3 shows the results predicting for indicator of militarism insofar as it taps mili- change in mean daily calories per capita. tary coups and political interventions. The Because caloric deficiency is central to food arms import question is best indexed by a rate security (Foster and Leathers 1999:28), we of change, tapping the social burden created show the results for caloric supply and then by arms races, including budget squeezes on discuss the small differences when predict- governmental basic needs programs.We also ing supply of protein. Economic growth separately test the two components of the in- boosts food supply, improving the adjusted ternal violence measure. Because militarism R2-from .28 to .39 for the model including might be a result of underlying moderniza- the lagged term alone (Model 1). All mod- tion, economic dependency, urban bias, or ernization theory variables improve food age dependency, we also regressed all the supply (Model 2). This supports arguments militarism measures on these relevant struc- about the benefits of domestic investment tural development factors. None was signifi- and political democratizationas well the the- cant, indicating that militarism is indepen- sis about military instability. It also suggests dent of these factors. that growth and equity are compatible in terms of boosting food supply and, in view of the positive correlationbetween economic RESULTS growth and political democratization (r = that there is no versus FOOD SUPPLY .34), "democracy growth" trade-off that undermines food sup- Do the effects of economic growth "trickle ply. Political democratization is also weakly down" to yield increased food security? This or negatively correlated with the various po- question is critical to the modernization-eco- litical instability measures (r = .10 for mili- nomic dependency discussion, in which tary instability, -.24 for internal violence, some modernization theorists contend that -.27 for civil war, and -.11 for ethnic politi- cal repression), undermining the "political 8 To see if has populationpressure magnified instability" thesis. effects in dense agrariansystems, we also tested The "relative benefits" version of the de- the interaction between agricultural density about for- (1970) and the annualrate of changein fertility pendency/world system argument investment is also in that the rate (1965 to 1985) but these tests were not sig- eign supported nificant. PEN ratio interactions with economic 9 We also testedthe argumentthat ethnic het- growth and domestic investment are both erogeneityblocks developmentand social inte- significant (Model 3), thus indicating that gration,producing food insecurity.Ethnic hetero- foreign investment is less beneficial to ca- geneity (Sullivan 1991:252-59) reduces food loric supply than is domestic investment. supplyslightly but has no impacton hunger.The Note that this result fits the "relative ben- numberof ethnic groupsdid not have a signifi- efits" that investment is canteffect on eitherfood or child argument foreign supply hunger. less beneficial to food com- 10We also testedmeasures for the presenceof simply supply interstatewar derived from the Correlatesof War pared with domestic investment. In a simple project(Singer and Small 1993), but the variable equation (not shown) containing only the was not significantin predictingeither caloric/ lagged calorie term, development level and proteinsupply or child hungerrates. export dependence, export dependence is 730 AMERICAN SOCIOLOGICAL REVIEW

Table 3. Unstandardized OLS Regression Coefficients Predicting Daily Calories per Capita from Single Theories, 1970 to 1990

IndependentVariable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Controls Daily calories per capita, 1970 .66*** .32** .22* .40** .43** .45*** (.12) (.11) (.13) (.14) (.14) (.13) Economic growth, 1970-1985 262.73*** 404.62** (75.28) (159.93) ModernizationTheory Level of economic development, - 112.67** 182.39*** 85.67* 95.01* 153.30*** 1970 (In) (44.89) (46.78) (50.69) (53.68) (43.93) Average annualpercentage 11.28** 15.36** change in domestic investment, (4.42) (5.30) 1970-1980 Change in humancapital 7.71** investment, 1970-1985 (3.23) Changein political 20.55* democratization,1965-1985 (9.47) Militaryinstability -73.42** (26.41) Dependency/WorldSystems Theory Percentagechange in export -216.97 dependency, 1970-1985 (132.30) Foreign capital penetrationratio - 231.55** (PEN), 1967 (86.72) Average annualpercentage _- _~ -8.01* change in domestic investment, (3.96) 1970-1980 x foreign capital penetrationratio, 1967 Economic growth, 1970-1985 -178.49* x foreign capital penetration (94.44) ratio, 1967 Foreign debt service load, -.001 1970-1985 (.10) (Continuedon next page) also significant and negative, but once the the 1965 level or the 1965-1985 first differ- PEN measure is included, it loses signifi- ence change in the rural/urbandisparity ra- cance. This indicates that its effects are tied tio, indicating that urbanbias is a distinctive up with foreign investment. The debt service force in its own right. load is not significant here or in other pos- Increased age dependency reduces food sible alternative specifications. supply, while agriculturaldensity has no sig- Increased rural/urbandisparity reduces the nificant effect (Model 5). The effect of the food supply net of the initial development change in fertility rate was similar but level and our cultural dualism measure, the slightly weaker (not shown). We also tested 1965 level of rural/urban disparity ratio the interactions of age dependency with ag- (Model 4). To ensure that this measure taps ricultural density, the 1965 level of the ru- urban bias, we substituted the percentage of ral/urban disparity ratio and the 1965-1985 the labor force in agriculture for the devel- change in the rural/urbandisparity ratio to opment level and obtained similar results see if population pressure is magnified by (not shown). None of the dependency mea- these contexts, but none of these tests was sures was significantly correlated with either significant (not shown). FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 731

(Table 3 continued) IndependentVariable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 UrbanBias Theory Rural/urban disparity ratio, 1965 - -38.69* (21.26) Change in rural/urban disparity, - - - -16.78** 1965-1985 (7.02) Neo-Malthusianand Ecological-EvolutionaryTheories Average annual percentage change - - -- 142.91** in age dependency, 1970-1985 (46.31) Agricultural density, 1970 - .83 (.75) MilitarismTheory Change in arms imports/GDP, - - -1.38*** 1970-1985 (.40) Ethnic political discrimination - - -58.26 (178.71) Internal violence - 35.38 (70.76)

Constant 1,047.44 1,085.26 1,207.15 1,188.93 668.93 539.46

Numberof cases 78 53 48 66 72 77

F-value 25.33 17.52 8.42 10.53 18.04 13.79 (F-probability) (.0001) (.0001) (.0001) (.0001) (.0001) (.0001) AdjustedR2 .39 .66 .59 .37 .49 .46 Note: Numbersin parenthesesare standarderrors. *p < .05 *p < .01 ***p< .001 (one-tailed tests)

Of the militarism variables, only the arms Table 4) we build combined models by treat- imports variable has significant effects in re- ing the significant modernizationvariables as ducing food supply (Model 6). Although in- the base and then adding significant variables ternal repression and violence create hunger from the other theories. Neither of the inter- (see below), they do not disrupt the aggre- action terms for decapitalization or distorted gate supply of food. Thus, in addition to growth are significant net of controls for military instability, arms imports may limit modernization (Models 1 and 2). In a sepa- food supply, presumably reflecting reduced rate equation (not shown) with the simple expenditures on food and health programs. foreign capital penetration (PEN) ratio, the Using the increment to adjusted R2 over a modernization terms plus the lagged caloric simple equation that includes only the control, foreign capital penetration was not lagged endogenous term for 1970 calorie significant, indicating that its drawbacks are supply (R2 = .28), the modernization vari- weaker than the modernization factors. In ables are the most important (increasing R2 Model 2, economic growth reduces all of the to .66 in Model 2), followed by dependency/ modernization terms below significance, world system variables in Model 3 (R2 = suggesting that the benefits of modernization .59), neo-Malthusian population pressure in are largely tied up with economic growth. Model 5 (R2 = .49), and urbanbias in Model Next we add the rural/urbandisparity ratio 4 (R2= .37). and change in rural/urbandisparity, and we Because of the theoretical centrality of find that both are weaker than the modern- modernization and its consistent results (in ization variables (Model 3). Age dependency 732 AMERICAN SOCIOLOGICAL REVIEW

Table 4. Unstandardized OLS Regression Coefficients Predicting Daily Calories per Capita from Combined Models, 1970 to 1990

IndependentVariable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Controls Daily calories per capita, 1970 .27* .17 .30** .34** .32 .24* (.12) (.12) (.12) (.13) (.11) (.11) Economic growth, 1970-1985 344.48* - 210.80** (155.17) (86.75)

ModernizationTheory Level of economic development, 126.08** 164.59** 83.85* 101.19* 109.08** 159.27** 1970 (ln) (47.76) (50.96) (49.91) (49.60) (43.98) (46.71) Average annualpercentage change 14.35** 6.40 11.32** 10.85* 10.55** 9.73* in domestic investment, 1970-1980 (6.00) (4.92) (4.60) (4.62) (4.36) (4.14) Change in humancapital investment, 6.83* 4.53 7.16* 7.18* 7.97** 5.01 1970-1985 (3.42) (3.48) (3.43) (3.66) (3.16) (3.23) Change in political democratization, 19.50* 7.80 22.51** 19.12* 19.40* 7.53 1965-1985 (9.95) (10.83) (9.96) (9.76) (9.28) (10.04) Militaryinstability -58.49* -50.17 -62.25* -75.34* -68.67** -38.67 (24.25) (30.59) (29.26) (33.94) (26.01) (27.62)

Dependency/WorldSystems Theory Foreign capital penetrationratio -93.31* -111.33* (PEN), 1967 (49.90) (55.39) Average annualpercentage change -4.19 in domestic investment 1965-1980 (3.52) x foreign capital penetration1967 Economic growth 1970-1985 - -124.92 x foreign capital penetration (86.15)

UrbanBias Theory Rural/urbandisparity, 1965 -9.08 (13.16) Change in rural/urbandisparity, -10.58 1965-1985 (1.46)

Neo-Malthusianand Ecological-EvolutionaryTheories Average annualpercentage change -38.35 in age dependency, 1970-1985 (51.82)

MilitarismTheory Change in arms imports/GDP, - -1.58* -2.10** 1970-1985 (.90) (.88)

Constant 1,198.29 1,311.11 1,339.77 1,060.32 1,123.97 1,125.05

Numberof cases 50 50 50 48 53 53

F-value 10.96 10.81 10.52 14.12 16.13 16.38 (F-probability) (.0001) (.0001) (.0001) (.0001) (.0001) (.0001) AdjustedR2 .62 .64 .61 .66 .67 .70 Note: Numbersin parenthesesare standarderrors. *p < .05 **p< .01 ***p< .001 (one-tailed tests) FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 733 added to the modernization variables is not children who are of healthy weight. First we significant, suggesting that it is counteracted test the "food availability" thesis (Model 1, by modernization processes (Model 4), and Table 5). Because our dependent variable is in Model 5, arms imports remains significant the percentage of healthy weight children, alongside modernization. positive regression coefficients indicate Finally, in Model 6 we add the control for progress in reducing child hunger rates. Al- economic growth, which causes all the mod- though increased food supply (change in ernization factors to loose significance, while daily calories per capita) reduces child hun- arms imports remain significant. This result ger, the increment to adjusted R2 is modest, reinforces our conclusion that modernization from .78 for an equation including only the improves caloric supply through its effects lagged term (not shown) to .79 in Model 1. on economic growth and that arms imports Model 2 adds economic growth, showing constitute an economic burden on food sup- that growth is positive and significant, re- ply. The standardizedcoefficients for Model ducing the significance of improved food 5 (with growth removed) show that modern- supply, indicating that it is more central. ization variables are the most important,with Model 3 removes the economic growth vari- human capital investment (.25), military in- able and controls for the modernizationvari- stability (-.24), and domestic investment ables, of which political democratization is (.23) being the strongest, followed by politi- the only significant factor. The change in ca- cal democratization (.17) and arms imports loric supply is not significant, and removing (-.14). Thus, modernization and militarism it does not alter this result. Rather than un- arguments are supported, and the idea that dermining political stability and creating economic dependency, urban bias, and a hunger, political democratization appears to rapid increase in age dependency indirectly be a strong positive force reducing child reduce food supply is supported as well. hunger and part of its benefits are associated We conducted four additional sensitivity with increased caloric supply. tests. Taking mean daily protein grams per Neither the economic dependency vari- capita as the dependent variable produces ables nor the urban bias variables affect the similar results except that human capital in- percentage of healthy weight children (Mod- vestment, political democratization,the PEN els 4 and 5). We also examined these and al- ratio x economic growth interaction, and the ternative measures of export and debt depen- urbanbias variables are not significant. Thus, dence (see notes 5 and 6), with simpler equa- protein supply appears to be less influenced tions controlling only for the lagged term and by structuraldevelopment processes but the change in caloric supply. None of these was other significant factors remain the same. significant (not shown). And we tested an in- Second, to test for underspecification, we teraction between the development level and added dummy variables for all world regions foreign capital penetration to see if foreign to Model 6 in Table 4, but none was signifi- investment was stronger among the more de- cant, suggesting that there are no major omit- veloped LDCs, but this interaction was not ted variables. Third, we ran these equations significant either (not shown). Removing the with a constant 48 countries with complete caloric supply variable does not alter any of measures for all variables, which produced these results. Export dependence and foreign the same pattern of significant factors. capital penetrationdo not appearto affect the Fourth, a simple cross-sectional model pre- prevalence of child hunger. dicting the 1990 caloric/protein supply also Increased age dependency boosts child found the same factors significant with hunger rates but agricultural density is not slightly stronger t-statistics, indicating that significant (Model 6). The change in total the trend and 1990 outcome values have the fertility rate has similar effects (not shown) same origins. but lower significance, so we show the age dependency effect. Arms imports and inter- nal violence increase rates of child CHILD HUNGER hunger, but ethnic political discrimination has no We turn now to our second measure of food significant impact (Model 7). We also tested security, the change in the percentage of separate dummy variables for genocide/ 734 AMERICAN SOCIOLOGICAL REVIEW

Table 5. Unstandardized OLS Regression Coefficients Predicting the Percentage of Healthy Weight Children from Single Theory Models, 1975 to 1990

IndependentVariable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Controls Percentageof childrenof .80*** .80*** .71*** .67*** .77*** .74*** .75*** healthy weight, 1970 (.06) (.05) (.08) (.08) (.06) (.06) (.05) Changein daily calories per .005* .003 .003 .003 .004* .004* .005** capita, 1970-1990 (.003) (.003) (.003) (.003) (.002) (.002) (.002) Economic growth, 1970-1985 6.17** 7.44 (1.78) (4.65)

ModernizationTheory Level of economic - - .42 3.15** .95 -.10 .96 development, 1970 (ln) (1.31) (1.22) (1.07) (1.11) (.90) Average annualpercentage -.07 -.09 change in domestic (.13) (.15) investment, 1970-1980 Changein humancapital .11 investment, 1970-1985 (.09) Changein political .60** democratization,1965-1985 (.25) Militaryinstability -.63 (.69)

Dependency/WorldSystems Theory Percentagechange in export - 1.24 dependency, 1970-1985 (3.35) Foreign capital penetration .54 ratio (PEN), 1967 (2.45) Average annualpercentage - ~-- .04 change in domestic investment, (.10) 1970-1980 x foreign capital penetration ratio, 1967 Economic growth, 1970-1985 ~- ~~- - ~.31 x foreign capital penetration (2.82) ratio, 1967 Change in foreign debt service - .001 load, 1970-1985 (.003) (Continuedon next page) politicide and civil war, but owing to thereby increases in child hunger rates. Civil multicollinearity between these two terms, wars are statistically significant only when we show only the "internal violence" mea- all terms other than the lagged term and sure that combines these two measures. In change in caloric supply are included, sug- simpler equations controlling only for the gesting that civil war is not a central source lagged term and arms imports, we found that of child hunger rates. Gauging relative im- ethnic political discrimination and genocide/ portance by increments to adjusted R2 over politicide had significant effects (not Model 1, the neo-Malthusian age depen- shown). Genocide/politicide is stronger than dency variable and the militarism variables ethnic political discrimination, reducing dis- have the strongest effects on child hunger, crimination to nonsignificance when they adding 6 to 8 percent to the explained vari- are included together. This suggests that eth- ance (.85 and .86 respectively); including nic political discrimination is an underlying political democratization increases the ad- problem that leads to genocide/politicide and justed R2 by only 1 percent (.80). FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 735

(Table 5 continued) IndependentVariable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 UrbanBias Theory Rural/urban disparity, 1965 - .39 (.31) Change in rural/urban disparity, - .07 1965-1985 (.02) Neo-Malthusianand Ecological EvolutionaryTheories Annual average percentage change -- -2.81** - in age dependency, 1970-1985 (.92) Agricultural density, 1970 ------.02 (.02) MilitarismTheory Change in arms imports/GDP, - - - --.01* 1970-1985 (.01) Ethnic political discrimination - -4.89 (3.68) Internal violence - -3.59** (1.27)

Constant 19.47 22.99 23.09 14.01 17.38 21.43 20.09

Numberof cases 49 49 49 47 63 69 73

F-value 92.70 80.59 29.26 22.42 52.00 75.71 72.40 (F-probability) (.0001) (.0001) (.0001) (.0001) (.0001) (.0001) (.0001) AdjustedR2 .79 .83 .80 .82 .80 .85 .86 Note: Numbersin parenthesesare standarderrors. *p < .05 **p< .01 ***p< .001 (one-tailed tests)

The combined models in Table 6 include mocratization and the presence of genocide/ all significant terms from Table 5 and treat politicide are significant, but the arms im- change in caloric supply and development port variable is not. We reintroduced the level as controls in addition to the endog- genocide/politicide measure here (not enous lagged term. As in the analysis of ca- shown) because it proved stronger than the loric supply, we treat the significant mod- combination internal violence measure. ernization variables as the base model and Genocide/politicide seems to be more cen- then introduce the other variables, begin- tral than civil wars as a source of child hun- ning first with age dependency. We also ger, although both are contributing factors. control for the interaction between age de- Model 3 combines variables from all three pendency and the 1965 level of rural/urban theories and shows that all earlier results disparity ratio to see if age dependency is work additively. Finally, we add economic magnified by cultural dualism. Political de- growth (Model 4), which reduces child hun- mocratization is positive and significant ger and appears to account for the benefits and rural/urbandisparity does magnify the of political democratization. Standardized influence of increased age dependency coefficients for the variables in Model 3 (Model 1), suggesting that the influence of with the age dependency x rural/urbandis- rapid increase in age dependency is greater parity interaction removed showed that the in dualistic societies. We also tested for the genocide/politicide (-.15) and age depen- interaction of rural/urbandisparity with ag- dency (-.14) variables are the strongest, ricultural density, but it was never signifi- closely followed by political democratiza- cant (not shown). In Model 2, political de- tion (-.11) and change in caloric supply 736 AMERICAN SOCIOLOGICAL REVIEW

Table 6. Unstandardized OLS Regression Coefficients Predicting the Percentage of Healthy Weight Children from Combined Models, 1975 to 1990

IndependentVariable Model 1 Model 2 Model 3 Model 4 Controls Percentageof healthy .73*** .73*** .72*** .71*** weight children, 1970 (.06) (.06) (.06) (.05) Change in daily calories .001 .004* .001 .0005 per capita, 1970-1990 (.003) (.002) (.002) (.025) Economic growth, 1970-1985 5.33** (1.97) ModernizationTheory Level of economic development, .18 .67 -.14 1.33 1970 (In) (1.06) (1.02) (1.00) (1.08)

Changein political .53** .43* .39* .21 democratization,1965-1985 (.20) (.20) (.19) (.19) UrbanBias Theory Rural/urbandisparity, 1965 -.57 -.66* -.64* (.40) (.37) (.35) Neo-MalthusianTheory Average annualpercentage -.43 .28 1.55 change in age dependency, (1.43) (1.33) (1.34) 1970-1985 Rural/urbandisparity, 1965 x average -.83* -1.00** -1.01** annualpercentage change in (.44) (.42) (.39) age dependency, 1970-1985 MilitarismTheory Change in arms imports/GDP, -.01 -.01 -.01 1970-1985 (.01) (.01) (.01) Presence of genocide/politicide -4.98** -4.97** -4.08** (1.67) (1.57) (1.51)

Constant 23.21 22.91 28.41 24.74

Numberof cases 58 58 58 58 F-value 47.19 54.49 43.71 45.28 (F-probability) (.0001) (.0001) (.0001) (.0001) AdjustedR2 .85 .85 .87 .89 Note: Numbersin parenthesesare standarderrors. *p < .05 **p< .01 **p < .001 (one-tailed tests)

(-.08).11 When economic growth is added pansion of political and civil rights associ- to the model, this rank order remains intact, ated with political democratization, and the but economic growth has the strongest stan- reduction of rapid population growth in tra- dardized coefficient. This suggests that the ditional dualistic societies. keys to reducing child hunger are economic We also conducted four sensitivity tests growth, the prevention of violence, the ex- to check these results. Substituting change in daily protein grams per capita for caloric 1 We omitted the interaction term from this supply produces results similar to those for model because standardized coefficients for in- caloric supply except that protein is consis- teraction terms are not directly interpretable. tently significant throughout and washes FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 737

MilitaryInstability

Domestic Investmentin Lagged Caloriesper Capita Physicaland HumanCapital

FOODSUPPLY Foreign Capital Penetration - - - -> (DailyCalories per Capita)

, 4 Rural/UrbanDisparity , / i \ Ratio, 1965 \ I / \ / / /~/\|~ IEconomic Growthand RA/ I I Development Level Change in Rural/Urban / DisparityRatio, 1965-1985 /

I I I 9 PoliticalDemocratization CHILD HUNGER

l, "(Percentage of Childrenof ,d + Healthy Weight) Age Dependency If 1r/, o~~~/ \ Lagged Percentage of / ,/ Children of Healthy Weight Internal Violence ' / I /

I / AI,/I--^t' Positiveeffect - - >- Negative effect

Figure 1. Findings on Variables Affecting Food Supply and Child Hunger in Less Developed Countries, 1970 to 1990 out economic growth and the age depen- if the question is predicting the static 1990 dency x rural/urban disparity interaction. level of child hunger. Protein supply is more central to hunger re- Overall, the internal political security af- duction and is closely tied to economic forded by political democratization and growth. None of the world region dummy freedom from internal violence plus popula- variables added to Models 3 and 4 in Table tion pressure in traditional dualistic societ- 6 were significant, and models for the con- ies are the strongest sources of increased stant 49 LDCs with complete data produce child hunger. Food supply has a positive but the same pattern of significant factors (not small impact, suggesting that critics of the shown). In the cross-sectional analysis pre- "food availability" thesis who contend that dicting the percentage of 1990 healthy food supply is irrelevant have overstated weight children, the only difference from their argument. They are, however, correct the lagged panel results is that arms imports that increased food supply alone is not suf- and the age dependency x rural/urbandis- ficient to reduce hunger. In poor dualistic parity interaction are not significant. This societies, problems associated with in- suggests that arms imports and population creased rapid population growth plus more pressure in dualistic societies do increase general restrictions on political freedoms child hunger, but that they are not important are more important than increasing the food 738 AMERICAN SOCIOLOGICAL REVIEW supply. Second, economic growth does dernutrition that leaves populations vulner- "trickle down" to increase food supply and able to disease and their members unable to thereby reduce child hunger rates (Model 4 lead active and productive lives. in Table 6) and is associated with the ben- Our most important finding is the rela- efits of political democratization. Thus food tively weak impact of food supply on child supply and child hunger have some shared hunger rates. Most researchershave assumed as well as divergent origins. this "food availability" thesis (see Foster and Figure 1 summarizes our major findings, Leathers 1999; Osmani 1995), and policy allowing us to explore these conclusions discussions have focused on ways to im- more closely. We show only simple variables prove agriculturalproduction and food aid in in the figure and discuss the interaction emergencies. Despite a global food surplus terms in the text. Political democratization, and increased international trade and food age dependency, and arms imports influence assistance, child hunger persists in LDCs both major aspects of food security-the and is only modestly reduced by recent in- supply of food and the access to food for creases in the supply of food. The child hun- children-as do the effects of economic ger problem is not simply a question of growth, which "trickle down" to increase growing more food, but of distributing it so food supply and reduce hunger. In terms of that disadvantaged children, minorities, their relative explanatory power, internal women, and rural households have secure violence, age dependency-linked population access to this supply. Our evidence suggests pressure, and political democratization are that entrenched inequality, especially that the most important variables affecting child associated with ethnic political discrimina- hunger, while modernization processes and tion and internal violence against minority age dependency are the most central vari- and other groups, is key to the persistence ables affecting food supply. This points to and increase in child hunger rates. Hunger significant differences between the two mea- is, in this sense, a political problem and must sures of food security. Food supply is more be addressed through political change-es- rooted in the structural development pro- pecially political democratization, restric- cesses associated with domestic and foreign tions on arms trade, and the reduction of investment and urban bias, while child hun- generalized violence. ger is shaped primarilypolitically, especially Our results suggest several policies that by political democratization, internal vio- might contribute to improved food security. lence, and arms imports. Increased food sup- First is the reduction of internalviolence, the ply does reduce child hunger, but its impact promotion of political democratization and is relatively small. restrictions on internationalarms trade. Sup- porting "military famine" arguments de and Gleditsch CONCLUSIONS (Cheatham 1994; Soysa 1999; Macrae and Zwi 1994; Poleman 1997) Food is a fundamentalbasic need and should about the hazards of militarism and internal be treated as a universal human right. Food violence, we found that genocide/politicide, also has significant implications for the po- civil war, arms imports, and underlying po- tential economic and social development of litical discrimination against targeted ethnic both the LDCs and the developed world. groups are central to child hunger. We also People without secure access to food are un- found that political democratization is a key likely to progress economically or to con- source of improving child hunger rates. This tribute indirectly to the welfare of other points to the importance of monitoring in- populations througheconomic trade, cultural ternational human rights, early warning and exchange, or social interaction. Thus, allevi- prevention efforts targeted on internal vio- ating hunger and poverty in LDCs is in the lence, and restraints on international arms long-run interest of the developed world as trade. Some have argued that democratiza- well as that of the LDCs. As Sen (1981) and tion in LDCs is politically destabilizing and others (Dreze and Sen 1989, 1995) argue, undermines incentives to invest and engage world hunger is generally not a question of in productive effort (Apter 1965; Hunting- sudden starvation but rather of chronic un- ton 1968). Our evidence contradicts this FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 739 view. We do not find support for either the on social welfare. It is also possible that our "political instability" thesis or either version indirect measurement of urban bias, which of the "growth-equity" trade-off idea. In- relies on change in sectoral disparity in la- stead, political democratization reduces bor productivity net of initial dualism and child hunger rates and improves food sup- development level, underestimates urban ply. It is also positively correlated with eco- bias. Further work on "biased" policies and nomic growth, suggesting that political de- political institutions is needed to further mocratization encourages economic growth evaluate this. and improvements in basic needs. This fits Building on recent discussions of foreign the results of other recent studies capital penetration (Dixon and Boswell (Kenworthy 2000; King 1998), suggesting 1996a, 1996b; Firebaugh 1992, 1996), we that the political problem in LDCs is not found that foreign investment is relatively enough democracy, not its premature or de- less beneficial for food supply than is do- stabilizing introduction. mestic investment but does not affect child Our second policy suggestion focuses on hunger. In contrast to previous dependency the population pressures created by persis- researchers who interpreted the negative ef- tent high fertility and the rapid growth in age fects for the foreign capital penetration ratio dependency. Scanlan (2000) showed that as indicating an absolute decline in social persistent high fertility and the resulting welfare (Bradshaw et al. 1993; Wimberley rapid growth in age dependency create an and Bello 1991), the ratio merely indicates ecological "overshoot"that undermines both the "relative benefits" of foreign versus do- aspects of food security. We have shown that mestic investment. Our negative effects, in- this is magnified in traditional dualistic so- cluding the foreign capital penetration ratio cieties in which households have an eco- interactions with economic growth and do- nomic incentive for maintaining high fertil- mestic investment, show merely that foreign ity, which leads to the use of marginal lands, investment is less beneficial than domestic overcultivation, and soil degradation. investment. As Dixon and Boswell (1996a) Scanlan (2000) also found that increased age argue, "often the choice (in LDCs) is not be- dependency lowered food security and was tween foreign capital and domestic capital associated with lower women's status and but between foreign capital and no capital at higher fertility. This suggest that improving all" (p. 547). women's status through education and em- In supplemental analyses (not shown, but ployment opportunities should be central to available on request from the authors), we improving both food supply and child hun- explored this furtherby examining other fac- ger rates. tors often linked with dependency, namely A third policy issue is balanced growth. land tenure inequality, income inequality, Development experts have long debated the and the intensity of ethnic economic dis- wisdom of balanced versus unbalanced crimination. None of these was significant in growth. Studies show that urban bias in predicting food supply or child hunger, and terms of protective tariffs, overvalued cur- at least in cross-sectional analyses, were not rency, state monopolies on agricultural ex- associated with economic dependency. We porting, heavier rural taxation and "price- also examined the growth of international twisting," and subsidies for urban industries food imports, including international food may contribute to urban-centered (or "un- aid, which some have argued is part of the balanced") growth, but that this urban bias "new international food order" that is dis- also creates sectoral disparities in produc- rupting food security in LDCs (Friedman tivity and investment, social welfare differ- 1982, 1993; McMichael 1996). We found no ences, and lowers overall economic growth support for this idea. In fact, international (Chenery et al. 1986; Krueger, Schiff, and food imports boost the food supply while Vales 1991; Lofchie 1997). We found that having no impact on child hunger. Thus, increased rural/urbandisparity reduces food while foreign trade and investment, includ- supply but does not affect child hunger ing increased food imports, may not reduce rates. This suggests that urban bias has hunger, these factors also do not appear to more impact on agriculturalproduction than be responsible for child hunger and may ac- 740 AMERICAN SOCIOLOGICAL REVIEW tually boost the food supply. We also found improving women's status in LDCs. Insofar that the "relative disadvantage" of foreign as these factors are influenced by structural capital penetrationwas not significant, net of modernization, they point to the need for an modernization controls, suggesting that at enlarged and revised modernization theory most it is a secondary factor for food supply. frameworku. Such a framework will need to In sum, this suggests that the traditional de- recognize the independent importance of po- pendency/world system advice to interna- litical democracy and conflict regulation as tionally "delink" is, at best, irrelevant, and well as the importance of incorporatinguse- that delinking may even be counterproduc- ful insights from ecological and urban bias tive if we take into consideration the ques- theories and the "relative benefits" of do- tion of food imports. At the same time, the mestic versus foreign investment. failure of food imports to reduce child hun- ger suggests that international food imports J. Craig Jenkins is Professor of Sociology and Political Science and Associate at the must be better targeted to address underly- Faculty Mershon Centerfor International Security, Ohio ing hunger problems. State His current work on Our overall results a structural University. focuses support comparative studies of political instability and modernization view that focuses on invest- violence, humanitarian disasters, and social ment in human and physical capital and is movement outcomes. He is working with Charles expanded to include political democratiza- Lewis Taylor on the World Handbook of Politi- tion. This view is limited, however, by the cal and Social Indicators IV. fact that these investments are more central to food The economic and Stephen J. Scanlan is Assistant Professor of So- improving supply. His research in- of structural moderniza- ciology, University of Memphis. population aspects terests are in the areas of comparative social tion seem to be critical to economic growth change, , , and but do not appear to "trickle down" to re- social stratification and inequality. Current re- duce hunger. Hunger is largely a political search projects include the cross-national deter- question, and insofar as political democrati- minants of child hunger in less industrialized so- zation boosts economic growth, and thereby cieties, the politics of international food assis- reduces child hunger, economic growth is a tance and the arms trade, food security and de- critical concern. Overall, our evidence mocratization, and hunger and poverty in Appa- sug- lachia. gests that it is domestic politics, especially political democratization and freedom from severe political repression coupled with REFERENCES population pressure in traditional dualistic Action Against Hunger. 2000. The Geopolitics of societies, that is central to this growth pro- Hunger, 2000-2001: Hunger and Power. cess. Further research using a time-series Boulder,CO: Lynne Reinner. analysis with simultaneous equations could Alderson, Arthurand FranqoisNielsen. 1995. assess the interaction of political democrati- "IncomeInequality, Development, and Dual- zation, internal violence, and economic ism." American Sociological Review 60:674- in their on food 701. growth impact security. Allison,Paul. 1990. Scoreas This to the need to discus- "Change Dependent points expand Variablesin RegressionAnalysis." Sociologi- sions of food traditional security beyond cal Methodology 20:93-114. concerns-agricultural development, do- Amin, Samir. 1976. Unequal Development. New mestic investment, and basic needs invest- York:Monthly Review. ments-by addressing basic political secu- Apter, David. 1965. The Politics of Moderniza- rity. Food supply alone is not the central is- tion.Chicago, IL: University of ChicagoPress. sue in reducing hunger. Hunger is also a dis- Arrow, Kenneth J. 1979. "The Trade-Offbe- tweenGrowth and in tributionalproblem, and the obstacles to im- Equity."Pp. 1-11 Theory Economic edited H. I. proved distribution are primarily political. for Efficiency, by Greenfieldand A P. Lerner.Cambridge, MA: Conflict regulation, violence prevention, the MITPress. reduction of internationalarms and the trade, Ballmer-Cao,Thanh-Huyen and Juerg Scheid- protection of civil and political rights should degger. 1979. Compendiumof Data for World- be central to policies that address hunger. A SystemAnalysis. Zurich, Switzerland: Univer- second focus should be reducing fertility and sity of Zurich, Sociological Institute. FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 741

Barro, Robert J. 1991. "Economic Growth in a Moises Syrquin. 1986. Industrialization and Cross Section of Countries." Quarterly Jour- Growth. New York: Oxford University Press. nal of Economics 106:407-43. Cohen, Joel E. 1995. How Many People Can the Bates, Robert. 1981. Markets and States in Earth Support? New York: Norton. Tropical Africa. Berkeley, CA: University of Crenshaw, Edward M. 1992. "Cross-National California Press. Determinants of Income Inequality." Social Bhuiya, Abbas Wojtyniak and Rezaul Bogdan Forces 71:339-63. Karim. 1989. "Malnutritionand Child Mortal- . 1993. "Polity, Economy, and Techno- ity." Journal of Bioscience 21:357-64. ecology." Social Forces 71:807-16. Bollen, Kenneth A. and Robert W. Jackman. Crenshaw, Edward M. and Ansari Ameen. 1993. 1985. "Regression Diagnostics." Sociological "Dimensions of Social Inequality in the Third Methods and Research 13:510-42. Word." Population Research and Policy Re- Bongaarts, John. 1996. "Population Pressure and view 12:297-313. the Food Supply System in the Developing Delacroix, Jacques and Charles C. Ragin. 1981. World." Population and Development Review "StructuralBlockage: A Cross-National Study 22:483-503. of Economic Dependency, State Efficacy, and Bornschier, Volker and ChristopherChase-Dunn. ."American Journal of So- 1985. Transnational Corporations and Under- ciology 86:1311-47. development. New York: Praeger. de Onis, M., C. Monteiro, J. Akre, G. Clughorn. Bradshaw, York W. 1987. "Urbanizationand Un- 1993. "The Worldwide Magnitude of Protein- derdevelopment." American Sociological Re- Energy Malnutrition." Bulletin WHO 71(6): view 52:224-39. 703-12. Bradshaw, York and Rita Noonan. 1997. "Urban- de Soysa, Indra and Nils Petter Gleditsch. 1999. ization, Economic Growth, and Women's La- To Cultivate Peace: Agriculture in a World of bor-Force Participation."Pp. 6-22 in Cities in Conflict. Oslo, Norway: International Peace the Developing World, edited by J. Gugler. Research Institute Oslo. New York: Oxford University Press. de Soysa, Indra and John R. Oneal. 1999. "Boon Bradshaw, York, Rita Noonan, Laura Gash, and or Bane? Reassessing the Productivity of For- Claudia Buchmann Sershen. 1993. "Borrowing eign Direct Investment." American Sociologi- Against the Future: Children and cal Review 64:766-82. Indebtedness." Social Forces 71:629-56. Dixon, William J. and Terry Boswell. 1996a. Brandt, Willy. 1986. WorldArmament and World "Dependency, Disarticulation, and Denomina- Hunger: A Call for Action. London, England: tor Effects." American Journal of Sociology Victor Gollancz. 102:543-62. Breedlove, William L. and J. Michael Armer. Dixon, William J. and Terry Boswell. 1996b. 1997. "Dependency, Techno-Economic Heri- "Differential Productivity, Negative Externali- tage, Disarticulation, and Social Development ties, and Foreign Capital Dependency: Reply in Less Developed Nations." Sociological Per- to Firebaugh."American Journal of Sociology spectives 40:661-80. 102:576-84. Breusch, Trevor S. and Adrian Pagan. 1979. "A Dreze, Jean and Amartya Sen. 1989. Hunger and Simple Test for Heteroskedasticity and Ran- Public Action. New York: Oxford University dom Coefficient Variation." Econometrica Press. 47:1287-94. Dreze, Jean, Amartya Sen, and Athar Hussein. Brzoska, Michael and Thomas Ohlson. 1987. 1995. The Political Economy of Food. New Arms Transfers to the Third World.New York: York: Oxford University Press. Oxford University Press. Ehrlich, Paul R. and Anne H. Ehrlich. 1990. The Bullock, Brad and Glenn Firebaugh. 1990. "Guns Population Explosion. New York: Simon and and Butter? The Effect of Militarization on Schuster. Economic and Social Development in the Ehrlich, Paul R., Anne H. Ehrlich, and Gretchen Third World." Journal of Political and Mili- Daily. 1993. "Food Security, Population, and tary Sociology 18:231-66. Environment." Population and Development Cardoso, Fernando H. and Enzo Faletto. 1979. Review 19:1-32. Dependency and Development in Latin Emmanuel, Arghiri. 1972. Unequal Exchange. America. Berkeley, CA: University of Califor- New York: Monthly Review. nia Press. Fein, Helen. 1993. "Accounting for Genocide Cheatham, Marcus. 1994. "War, Military Spend- After 1945: Theories and Some Findings." In- ing, and Food Security in Africa." Pp. 229-53 ternational Journal on Group Rights 1:79- in Seeking Security and Development, edited 106. N. A. Graham. Boulder, CO: Lynne Reinner. Finkel, Steven L. 1995. Causal Analysis with Chenery, Hollis B., Sherman Robinson, and Panel Data. Newbury Park, CA: Sage. 742 AMERICAN SOCIOLOGICAL REVIEW

Firebaugh, Glenn. 1979. "Structural Determi- Harff, Barbaraand Ted Robert Gurr. 1989. "Vic- nants of in Asia and Latin tims of the State: Genocides, Politicides, and America, 1950-1970." American Sociological Group Repression Since 1945." International Review 44:199-215. Review of Victimology 1:23-41. . 1992. "Growth Effects of Foreign and Huntington, Samuel P. 1968. Political Order in Domestic Investment." American Journal of Changing Societies. New Haven, CT: Yale Sociology 98:105-30. University Press. . 1996. "Does Foreign Capital Harm Poor .1987. "The Goals of Development." Pp. Nations? New Estimates Based on Dixon and 6-32 in Understanding Political Development: Boswell's Measures of Capital Foreign Pen- An Analytical Study, edited by S. P. Hunting- etration." American Journal of Sociology ton and M. Weiner. Boston, MA: Little, 102:563-75. Brown. Firebaugh, Glenn and Frank Beck. 1994. "Does Inkeles, Alex and David Smith. 1974. Becoming Growth Benefit the Masses? Growth, Depen- Modern. New York: HarvardUniversity Press. dence, and Welfare in the Third World." Jackman, Robert W. 1975. Politics and Social American Sociological Review 59:631-53. Equality. New York: Wiley. Food and Agriculture Organization of the United . 1980. "A Note on the Measurement of Nations (FAO). 1972, 1974, 1992, 1994. FAO Growth Rates in Cross-National Research." Yearbook.Rome, : FAO. American Journal of Sociology 86:604-17. , Committee on World Food Security. Janowitz, Morris. 1977. Military Institutions and 1996a. Rome Declaraion on World Food Se- Coercion in the Developing Nations. Chicago, curity. Rome, Italy: FAO. IL: The University of Chicago Press. . 1996b. The Sixth World Food Survey. Judge, George G., W. E. Griffiths, R. CarterHill, Rome, Italy: FAO. Helmut Lutkepohl, and Tsoung-Chao Lee. Foster, Philips and Howard D. Leathers. 1999. 1985. The Theory and Practice of Economet- The World Food Problem. 2d ed. Boulder, CO: rics. 2d ed. New York: Wiley. Lynne Reinner. Kelly, Marion. 1992. "Anthropometryas an In- Frank, Andre Gunder. 1967. Capitalism and Un- dicator of Access to Food in Populations Prone derdevelopment in Latin America. New York: to Famine." Food Policy 17:443-54. Monthly Review Press. Kenworthy, Lane. 2000. "The Effects of Income Friedman, Harriet. 1982. "The Political Economy Inequality on Economic Growth." Department of Food: The Rise and Fall of the Postwar In- of Sociology, Emory University, Atlanta, GA. ternational Food Order."American Journal of Unpublished . Sociology 88 (supp.):S248-86. Kerr, Clark. 1969. Industrialism and Industrial Friedman, Harriet. 1993. "The Political Economy Man. New York: Oxford University Press. of Food." New Left Review 197:29-37. Kidron, Michael and Dan Smith. 1983. The War Goldsmith, Arthur. 1986. "Democracy, Political Atlas. New York: Simon and Schuster. Stability, and Economic Growth in Develop- . 1991. The New State of War and Peace. ing Countries." Comparative Political Studies New York: Simon and Schuster. 18:517-31. Khohli, Atul. 1986. "Democracy and Develop- Goodell, Grace and John P. Powelsen 1982. "The ment." Pp. 153-82 in Development Strategies Democratic Prerequisites of Development." Reconsidered, edited by J. Lewis and V. Pp. 167-76 in Freedom in the World, edited Kallab. New Brunswick, NJ: Transaction. by R. Gastil. New York: Freedom House. King, Loren A. 1998. "Economic Growth and Gugler, Josef. 1997. Cities in the Developing Basic Human Needs." International Studies World. New York: Oxford University Press. Quarterly 42:385-400. Gurr, Ted Robert. 1989. Polity II: Political Krueger, Anne, Maurice Schiff, and Alberto Structures and Regime Change, 1800-1986 Valdes. 1991. The Political Economy of Agri- [MRDF]. Boulder, CO: Center for Compara- cultural Pricing Policy. Baltimore, MD: John tive Politics [producer]. Ann Arbor, MI: Inter- Hopkins University Press. university Consortium for Politics and Social Kutzner, Patricia L. 1991. World Hunger: A Ref- Research [distributor]. erence Handbook. Santa Barbara, CA: AFC- . 1993. Minorities at Risk. Washington, CLIO, Inc. DC: U.S. Institute of Peace Press. Kuznets, Simon. 1955. "Economic Growth and Hannan, Michael T. 1979. "Issues in Panel Income Inequality." American Economic Re- Analysis of National Development: A Method- view 45:1-28. ological Overview." Pp. 17-33 in National Lenski, Gerhard,Jean Lenski, and Patrick Nolan. Development and the World System, edited by 1995. HumanSocieties. New York: Macmillan. J. Meyer and M. Hannan. Chicago, IL: Univer- Letwin, William. 1983. Against Equality. Lon- sity of Chicago Press. don, England: Macmillan. FOOD SECURITY IN LESS DEVELOPED COUNTRIES, 1970 TO 1990 743

Lipton, Michael. 1977. Why Poor People Stay The Ohio State University, Columbus, OH. Poor. Cambridge, MA: Harvard University . 2001. "Food Availability and Access in Press. Lesser Industrialized Societies: A Test and In- Lofchie, Michael F. 1997. "The Rise and Demise terpretation of Neo-Malthusian and Techno- of Urban-Biased Development Policies in Af- Ecological Theories." Sociological Forum rica." Pp. 23-39 in Cities in the Developing 16:231-62. World, edited by J. Gugler. New York: Oxford Scanlan, Stephen J. and J. Craig Jenkins. 2001. University Press. "Military Power and Food Security: A Cross- London, Bruce. 1987. "Dependence, Distorted National Analysis of Less Developed Coun- Development, and Fertility Trends in Noncore tries, 1970-1990." International Studies Quar- Nations." American Sociological Review terly 45:159-87. 53:606-18. Sen, Amartya. 1981. Poverty and Famines. New London, Bruce and David A. Smith. 1988. "Ur- York: Oxford University Press. ban Bias, Dependence, and Economic Stagna- Shen, Ce and John B. Williamson. 1997. "Child tion in Noncore Nations." American Sociologi- Mortality, Women's Status, Economic Depen- cal Review 53:454-463. dency, and State Strength." Social Forces London, Bruce and Kirk Williams. 1990. "Na- 76:667-700. tional Politics, International Dependency, and Singer, J. David and Melvin Small. 1993. The Basic Needs Provision." Social Forces Correlates of War Project: International and 69:565-84. Civil War Data, 1816-1992 [MRDF]. Ann Ar- Macrae, Joanna and Anthony Zwi, eds. 1994. bor, MI: Department of Political Science, Uni- War and Hunger. London, England: Zed versity of Michigan [producer]. Ann Arbor, Books. MI: Inter-university Consortium for Political McMichael, Philip. 1996. Development and So- and Social Research [distributor]. cial Change. Thousand Oaks, CA: Pine Forge. Smil, Vaclav. 1994. "How Many People Can the Moon, Bruce E. and William J. Dixon. 1985. Earth Feed?" Population and Development Re- "Politics, the State, and Basic Human Needs." view 20:255-92. American Journal of Political Science 29: Sorensen, Georg. 1991. Democracy, Dictatorship 661-94. and Development. London, England: Morrison, Donald, R. C. Mitchell, J. N. Paden, Macmillan. and H. M. Stevenson. 1989. Black Africa: A Stokes, Randall and Andy Anderson. 1990. "Dis- Comparative Handbook. 2d ed. Irvington, NY: articulation and Human Welfare in Less De- Irvington Press. veloped Countries." American Sociological Nielsen, Francois. 1994. "Income Inequality and Review 55:63-74. Industrial Development: Dualism Revisited." Sullivan, Michael J. 1991. Measuring Global American Sociological Review 59:654-77. Values. Westport, CT: Greenwood. Osmani, Siddiq. 1995. "The Entitlement Ap- Summers, Robert and Alan Heston. 1991. "The proach to Famine." Pp. 253-94 in Choice, Wel- Penn World Tables (Mark 5)." Quarterly Jour- fare and Development, edited by K. Basu, P. nal of Economics 106:327-68. K. Paltanaih and K. Suzumena. New York: Tweeten, Luther G., and Donald G. McClelland, Oxford University Press. eds. 1997. Promoting Third-World Develop- Poleman, Thomas D. 1997. "Recent Trends in ment and Food Security. Westport, CT: Food Availability and Nutritional Wellbeing." Praeger. Population and Development 19:145-65. United Nations Conference on Trade and Devel- Quinn, Victoria J. and Eileen Kennedy. 1994. opment. 1975-1988. UNCTAD Commodity "Food Security and Nutrition Monitoring Sys- Yearbook. New York: United Nations. tems in Africa." Food Policy 19:234-54. United Nations Development Programme Ragin, Charles and York Bradshaw. 1992. "In- (UNDP). 1994. Human Development Report ternational Economic Dependence and Human 1994. New York: Oxford University Press. Misery, 1938-1980." Sociological Perspec- United Nations International Children's Emer- tives 35:217-47. gency Fund (UNICEF). 1998. State of the Ross, Andrew I. 1987. "Dimensions of Militari- World's Children. New York: Oxford Univer- zation in the Third World." Armed Forces and sity Press. Society 13:561-78. United Nations Population Fund. 1999. Foodfor Rostow, Walter W. 1960. The Stages of Eco- the Future. New York: United Nations. nomic Growth. New York: Cambridge Univer- Weede, Erich. 1996. Economic Development, So- sity Press. cial Order, and World Politics. Boulder, CO: Scanlan, Stephen J. 2000. Globalization and Lynne Reinner. Food Security in Less Industrialized Societies. Wickrama, K. A. S. and Charles L. Mulford. Ph.D. Dissertation, Department of Sociology, 1996. "Political Democracy, Economic Devel- 744 AMERICAN SOCIOLOGICAL REVIEW

opment, Disarticulation, and Social Well-Be- "Effects of Foreign Investments, Exports, and ing in Developing Countries." The Sociologi- Economic Growth on Third World Food Con- cal Quarterly 37:375-90. sumption." Social Forces 70:895-922. Williamson, John B. 1987. "Social-Security and World Bank. 1994. WorldData 1994 [CD ROM]. Physical Quality of Life in Developing Na- Washington, DC: The International Bank for tions: A Cross-National Analysis." Social In- Reconstruction and Development [producer, dicators Research 19:205-27. distributor]. Wimberley, Dale W. 1991. "TransnationalCor- . 1997. World Development Indicators porate Investment and Food Consumption in [CD ROM]. Washington, DC: The Interna- the Third World." Rural Sociology 56:406-31. tional Bank for Reconstruction and Develop- Wimberley, Dale W. and Rosario Bello. 1992. ment [producer, distributor].