Household nutritional effects of the DICONSA food subsidy program

• Název v anglickém jazyce: Household nutritional effects of the DICONSA food subsidy program
• Klíčová slova: Nutrition, food, subsidy, Mexico, survey, household, microeconomics
• Klíčová slova anglicky: Nutrition, food, subsidy, Mexico, survey, household, microeconomics
• Akademický rok vypsání: 2010/2011
• Typ práce: diplomová práce
• Jazyk práce: angličtina
• Ústav: Institut ekonomických studií (23-IES)
• Vedoucí / školitel: doc. PhDr. Michal Bauer, Ph.D.
• Řešitel: skrytý - zadáno vedoucím/školitelem
• Datum přihlášení: 10.06.2011
• Datum zadání: 10.06.2011
• Datum a čas obhajoby: 29.06.2012 00:00
• Místo konání obhajoby: IPS FSV UK, U Kříže 8, 158 00 Praha 5 - Jinonice
• Datum odevzdání elektronické podoby: 16.05.2012
• Datum proběhlé obhajoby: 29.06.2012
• Oponenti: doc. Petr Janský, M.Sc., Ph.D.

Seznam odborné literatury

Agricultural Resource Marketing Centre, “White Corn Profile”, 2011 (Available at: http://www.agmrc.org/commodities__products/grains__oilseeds/corn_grain/white_corn_profile.cfm)
Andreyeva, T., Long, M.W., and Brownell, K.D. “The Impact of Food Prices on Consumption: A Systematic Review of Research on the Price Elasticity of Demand for Food”,American Journal of Public Health, Vol 100, No. 2, 2010
Becker, S.O. and Inchino. A., "Estimation of Average Treatment Effects based on Propensity Scores." The Stata Journal 2(4):358-377, 2002
Blundell, R. and Costa Dias, M., “Alternative approaches to evaluation in empirical microeconomics”, The Institute for Fiscal Studies, Department of Economics, UCL, 2008
Conniffe, D., Gash, V. and O’Connell, P.J., “Evaluating State Programmes: “Natural Experiments” and Propensity Scores”, The Economic and Social Review, Vol. 31, No. 4, October, pp. 283-308, 2000
Cunha, J.M., “Testing Paternalism: Cash vs. In-kind Transfer in Rural Mexico”. Stanford University, 2010
Diconsa, Directorio de Tiendas, 2010
Food and Agriculture Organisation (FAO), World Food Price index, 2012 (Available at: http://www.fao.org/worldfoodsituation/ wfshome/foodpricesindex/en/)
Fox, J. A., “Accountability Politics: Power and Voice in Rural Mexico”, Oxford 2007
French, S.A, “Pricing Effects on Food Choices”, Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, American Society for Nutritional Sciences, 2003
George, P.S., “Costs and Benefits of Food Subsidies in India” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press, 1988
Gundersen, C., Yañez, M., Valdez, C. and Kuhn, B., “A Comparison of Food Assistance Programs in Mexico and the United States” Food Assistance and Nutrition Research Report No. 6., 2000
Hopkins, R.F., “Political Calculations in Subsidizing Food” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press, 1988
Iqbal, F., “Sustaining Gains in Povery Reduction in the Middle East and North Africa”, Orientations in Development Series, World Bank 2006, ch5, p57-70
Instituto de Nutrición de Centroamérica y Panamá (INCAP) and Organización Panamericana de la Salud (OPS), “Tabla de Composición de Alimentos de Centroamérica”, 2007
Instituto Nacional de Nutrición (INN), México, “Recomendaciones para el consumo de calorías, para individuos normales, con la dieta en las condiciones de México”, 1983
Jensen, R.T. and Miller, N.H, “Giffen Behavior: Theory and Evidence” Faculty Research Working Papers Series, Havard JFK School of Governance, 2007
Jensen, R.T. and Miller, N.H, “Do Consumer Price Subsidies Really Improve Nutrition?” CID Working Paper No. 160, 2008a
Jensen, R.T. and Miller, N.H., “Giffen Behavior and Subsistence Consumption “ (The American Economic Review, Vol. 98, No. 4, pp. 1553-1577, 2008b
Kochar, A., “Can Targeted Food Programs Improve Nutrition? An Empirical Analysis of India’s Public Distribution System” (Economic Development and Cultural Change, Vol. 54, No. 1, pp. 203-235, 2005
Kumar, S.K. and Alderman, H. “Food Consumption and Nutritional Effects of Consumer-Oriented Food Subsidies” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press, 1988
Lustig, N., “Coping with Rising Food Prices: Policy Dilemmas in the Developing World”, Working Paper Number 164, March 2009
McKenzie, D., “Are tortillas a Giffen Good in Mexico?” , Economics Bulletin, Vol. 15, No. 1 pp. 1−7, 2002
Pinstrup-Andersen, P. "The Social and Economic Effects of Consumer-Oriented Food Subsidies: A Summary of Current Evidence" in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press. 1988a
Pinstrup-Andersen, P.,“The Effectiveness of Consumer-Oriented Food Subsidies in Reaching Rationing and Income Transfer Goals” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press. 1988b
Pufahl, A., Heinrich, J. and Weiss, C.R., “Evaluating the effects of farm programmes results from propensity score matching”, European Review of Agricultural Economics Vol 36 (1), pp. 79–101, 2009
Rogers, B.L., “Design and Implementation Considerations for Consumer-Oriented Food Subsidies” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press, 1988a
Rogers, B.L., “Pakistan's Ration System: Distribution of Costs and Benefits:” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press. 1988b
Senauer, B. “Household behaviour and nutrition in developing countries” in Food Policy, vol. 15, issue 5, pages 408-417, 1990
Tritah, A., “The Public Distribution System in India: Counting the poor from making the poor count” GREMAQ, Universit´e des Sciences Sociales, Toulouse, France, 2004
Universidad Autónoma Chapingo (authors not cited), “Evaluación Externa Programa de Abasto Rural, Diconsa 2006”, Universidad Autónoma Chapingo, Unidad Gestora de Servicios Tecnológicos, 2007
Valdis, A. “Explicit versus Implicit Food Subsidies: Distribution of Costs” in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Per Pinstrup-Andersen (ed.) John Hopkins Press, 1988
Victor, C., “The Role Of Diconsa In Promoting Food Security In Rural Mexico And Providing Assistance In Case Of Natural Disasters ” p121-152 “Border Lines: Journal of the Latino Research Centre”, Vol V (Special Edition), 2011
Wood, B., Nelson, C., Nogueria, L., “Measuring the Welfare Impact of Food Price Increases on Mexican Households”, University of Illinois at Urbana-Champaign, 2010
World Health Organization (Joint WHO/FAO Expert Consultation), “Diet, Nutrition and The Prevention Of Chronic Diseases”, WHO Technical Report Series, World Health Organization Geneva, 2002
Xuguang, G., Popkin, B.M., Mroz, T.A. and Fengying, Z., “Food Price Policy Can Favorably Alter Macronutrient Intake in China”, The Journal of Nutrition, Community and International Nutrition, 1999
Zazaig, S.M & Adams, R.H., “The Egyptian Food Subsidy System: Operation and Effects on Income Distribution”, World Development Vol. 24, No. 11, pp. 1777-1791, Elsevier Science Ltd, 1996

Předběžná náplň práce

The Diconsa program was created in 1976 as a means to distribute basic commodities (corn, beans, rice, sugar, corn flour, powdered milk and tortilla products) at subsidized prices to marginalized populations throughout Mexico in rural stores. Diconsa products retail at prices on average 7% lower than their open-market equivalents, a result of the Diconsa principle that transportation and operational costs are not passed on to the consumer.

The Diconsa program is essentially a government intervention in the private market for basic food commodities, providing a regulatory price anchor and increasing the targeted population’s food security by means of increasing its purchasing power and by subsidizing staple foods. This targeted population consists of rural households in highly marginalized areas, where prohibitive operating costs prevent the establishment of a sufficiently competitive private market.

The aim of this thesis is to assess more precisely the program’s contribution to what must be considered a fundamental motivation for its implementation – improving the nutrition levels in low-income households. This is potentially an informative analysis because substitution for less-nutritious foods has previously been empirically observed in the general case of decreasing relative prices, and in particular Jensen & Miller’s study (2008) shows that there is no evidence for an increase in nutrition in the case of specific food subsidies for the very poor, and some for an actual decrease. Given the costs of maintaining food subsidies, I believe this is an important area of policy research.

I believe my study will provide a useful test of these results in a different setting, a different program, a different type of subsidy, and within a much larger and demographically more varied sample. This sample data is contained in the 2010 Encuesta Nacional de Ingresos y Gastos de los Hogares (ENIGH), or ‘National Survey of Household Incomes and Expenditures’, by the National Mexican Institute for Geography and Statistics’ (INEGI). This survey is taken over 7 days from a random sample of over 27,000 households throughout Mexico.

Working hypotheses


1) The particular pattern of multiple commodity food subsidies provided through Diconsa does not significantly improve household nutrition levels amongst the targeted population of the rural poor in Mexico.
2) The consumer response to the subsidy is to alter diet composition without increasing the overall nutritional content of the household food basket
3) The subsidization of staple commodities results in significant income effects which increase overall household food consumption
4) Consumer response to the subsidy is at least partly dependent on specific household characteristics such as income and demographic composition

Methodology

The main body of this thesis will draw on econometric analysis of the Mexican household survey data. I will first have to establish a control group of households – non-participants – using the propensity score matching method. Participation is (initially) defined as the purchase of at least one food product in Diconsa during the week long survey. Specifically, what I will then be looking for is:
E(N_1^*- N_0^* ┤|X,d=1)=ATE_1

where N_0^* and〖 N〗_1^* are the level of nutrition with and without treatment respectively, d is the dummy variable for Diconsa participation. X is a set of control variables, such as household income, etc, which is determined by the propensity score matching method.
The propensity score is essentially a probability, for a given household, of participation in the Diconsa program given the set of X variables. I match treatment households (participants) with control households (non-participants) according to propensity scores – matching like with like. I am then able to compare the two groups in order to see whether Diconsa participation has a statistically significant effect on nutrition levels once the X variables are controlled for, and to determine the sign of this effect – whether it is positive or negative.
I will then attempt to confirm my results by carrying out a number of robustness checks. This will involve modifying the set of X variables, as well as restricting the control group to households located in municipalities without a Diconsa store and comparison with OLS regressions with varying specifications at the household and municipal level. If the robustness checks fail, I will attempt to explain which significant unobservable characteristics might be absent from the data.
Given the methodological approach, I will also be able to identify which of the X variables – income, etc – are most important in determining participation, and, by looking at different propensity scores, I will be able to ascertain how the nutritional effect of participation (be it positive or negative) varies with different X characteristics. Furthermore, the detailed food expenditure data contained in the survey will allow me to uncover the specifics of the consumer response to the subsidy program and the degree to which it occurs.

Provisional Outline

Introduction
Food subsidy programs as policy tools
The Diconsa program and its specified goals
Review of literature
General studies of food subsidy programs
Econometric studies of nutritional and consumption effects
Experimental studies
Comparison of these empirical analyses with my own
Methodological aspects of previous studies: appropriateness for the current research
Results of previous empirical studies
Description of data
Details of ENIGH survey and dataset
Calorie counts: how to calculate household nutrition
Description of methodology
Propensity score matching approach (PSM)
OLS regression approach
Econometric results
PSM results and analysis
OLS results and analysis
Consumer behaviour at the household level
Analysis and comments on initial results
Comparison with the results of previous studies
Confirmation of hypotheses
Robustness checks
PSM approach: robustness checks
OLS approach: robustness checks
Implications of study
Policy implications
Suggestions for further research
Conclusions and summary

Předběžná náplň práce v anglickém jazyce

The Diconsa program was created in 1976 as a means to distribute basic commodities (corn, beans, rice, sugar, corn flour, powdered milk and tortilla products) at subsidized prices to marginalized populations throughout Mexico in rural stores. Diconsa products retail at prices on average 7% lower than their open-market equivalents, a result of the Diconsa principle that transportation and operational costs are not passed on to the consumer.

The Diconsa program is essentially a government intervention in the private market for basic food commodities, providing a regulatory price anchor and increasing the targeted population’s food security by means of increasing its purchasing power and by subsidizing staple foods. This targeted population consists of rural households in highly marginalized areas, where prohibitive operating costs prevent the establishment of a sufficiently competitive private market.

The aim of this thesis is to assess more precisely the program’s contribution to what must be considered a fundamental motivation for its implementation – improving the nutrition levels in low-income households. This is potentially an informative analysis because substitution for less-nutritious foods has previously been empirically observed in the general case of decreasing relative prices, and in particular Jensen & Miller’s study (2008) shows that there is no evidence for an increase in nutrition in the case of specific food subsidies for the very poor, and some for an actual decrease. Given the costs of maintaining food subsidies, I believe this is an important area of policy research.

I believe my study will provide a useful test of these results in a different setting, a different program, a different type of subsidy, and within a much larger and demographically more varied sample. This sample data is contained in the 2010 Encuesta Nacional de Ingresos y Gastos de los Hogares (ENIGH), or ‘National Survey of Household Incomes and Expenditures’, by the National Mexican Institute for Geography and Statistics’ (INEGI). This survey is taken over 7 days from a random sample of over 27,000 households throughout Mexico.

Working hypotheses


1) The particular pattern of multiple commodity food subsidies provided through Diconsa does not significantly improve household nutrition levels amongst the targeted population of the rural poor in Mexico.
2) The consumer response to the subsidy is to alter diet composition without increasing the overall nutritional content of the household food basket
3) The subsidization of staple commodities results in significant income effects which increase overall household food consumption
4) Consumer response to the subsidy is at least partly dependent on specific household characteristics such as income and demographic composition

Methodology

The main body of this thesis will draw on econometric analysis of the Mexican household survey data. I will first have to establish a control group of households – non-participants – using the propensity score matching method. Participation is (initially) defined as the purchase of at least one food product in Diconsa during the week long survey. Specifically, what I will then be looking for is:
E(N_1^*- N_0^* ┤|X,d=1)=ATE_1

where N_0^* and〖 N〗_1^* are the level of nutrition with and without treatment respectively, d is the dummy variable for Diconsa participation. X is a set of control variables, such as household income, etc, which is determined by the propensity score matching method.
The propensity score is essentially a probability, for a given household, of participation in the Diconsa program given the set of X variables. I match treatment households (participants) with control households (non-participants) according to propensity scores – matching like with like. I am then able to compare the two groups in order to see whether Diconsa participation has a statistically significant effect on nutrition levels once the X variables are controlled for, and to determine the sign of this effect – whether it is positive or negative.
I will then attempt to confirm my results by carrying out a number of robustness checks. This will involve modifying the set of X variables, as well as restricting the control group to households located in municipalities without a Diconsa store and comparison with OLS regressions with varying specifications at the household and municipal level. If the robustness checks fail, I will attempt to explain which significant unobservable characteristics might be absent from the data.
Given the methodological approach, I will also be able to identify which of the X variables – income, etc – are most important in determining participation, and, by looking at different propensity scores, I will be able to ascertain how the nutritional effect of participation (be it positive or negative) varies with different X characteristics. Furthermore, the detailed food expenditure data contained in the survey will allow me to uncover the specifics of the consumer response to the subsidy program and the degree to which it occurs.

Provisional Outline

Introduction
Food subsidy programs as policy tools
The Diconsa program and its specified goals
Review of literature
General studies of food subsidy programs
Econometric studies of nutritional and consumption effects
Experimental studies
Comparison of these empirical analyses with my own
Methodological aspects of previous studies: appropriateness for the current research
Results of previous empirical studies
Description of data
Details of ENIGH survey and dataset
Calorie counts: how to calculate household nutrition
Description of methodology
Propensity score matching approach (PSM)
OLS regression approach
Econometric results
PSM results and analysis
OLS results and analysis
Consumer behaviour at the household level
Analysis and comments on initial results
Comparison with the results of previous studies
Confirmation of hypotheses
Robustness checks
PSM approach: robustness checks
OLS approach: robustness checks
Implications of study
Policy implications
Suggestions for further research
Conclusions and summary