Climate variability and child height in rural Mexico

Economics and Human Biology 10 (2012) 54–73

Contents lists available at ScienceDirect

Economics and Human Biology journal homepage: http://www.elsevier.com/locate/ehb

Climate variability and child height in rural Mexico Emmanuel Skoufias 1,*, Katja Vinha The World Bank, MSN # MC4-415, 1818 H Str. NW, Washington, DC 20433, United States

A R T I C L E I N F O

A B S T R A C T

Article history: Received 10 January 2011 Received in revised form 12 June 2011 Accepted 13 June 2011 Available online 6 July 2011

We examine the impacts of weather shocks, defined as rainfall or growing degree days, a cumulative measure of temperature, more than a standard deviation from their respective long run mean, on the stature of children between 12 and 47 months of age in Mexico. We find that after a positive rainfall shock children are shorter regardless of their region or altitude. Negative temperature shocks have a negative impact on height in the central and southern parts of the country as well as in higher altitudes. Although on average there are no statistically significant impacts from positive temperature shocks, certain subpopulations – namely boys, children between 12 and 23 months at the time of measurement, and children of less educated mothers – in some of the regions are negatively impacted. The results also suggest that potentially both agricultural income and communicable disease prevalence contribute to the effects. ß 2011 Elsevier B.V. All rights reserved.

JEL classification: D13 I31 Q54 Q12 Keywords: Climate change Weather shocks Child height Mexico

1. Introduction While there is a great deal of uncertainty over the exact magnitudes of the global changes in temperature and precipitation, it is widely accepted that significant deviations of the variability of climate from its historical patterns are likely to occur (IPCC, 2007).2 Considering * Corresponding author. Tel.: +1 240 505 6758. E-mail address: eskoufi[email protected] (E. Skoufias). 1 Emmanuel Skoufias is a Lead Economist at the Poverty Reduction and Economic Management (PREM) Network at the World Bank. Katja Vinha is a Consultant for the World Bank. The findings, interpretations, and conclusions are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. 2 According to the Intergovernmental Panel on Climate Change (IPCC) a narrow definition of climate refers to the statistical description in terms of the mean and variability of quantities such as temperature, precipitation and wind over a period of time ranging from months to thousands of years. The norm is 30 years as defined by the World Meteorological Organization (WMO). Climate is different from weather which refers to atmospheric conditions in a given place at a specific time. The term ‘‘climate change’’ is used to indicate a significant variation (in a statistical sense) in either the mean state of the climate or in its variability for an extended period of time, usually decades or longer (Wilkinson, 2006). 1570-677X/$ – see front matter ß 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ehb.2011.06.001

that millions of poor households in rural areas all over the world are dependent on agriculture, there are increasing concerns that the change in the patterns of climatic variability is likely to add to the already high vulnerability of households, and especially of children, in rural areas of developing countries. Furthermore, the prevalence of diseases will most likely be affected by climatic changes adding to the potential welfare loss. In view of this impending threat of climate change upon the poor, it is critical to have a deeper understanding of the magnitude of the consequences and targeted measures that could mitigate any health effects from of erratic weather. With these considerations in mind, we carry out an analysis of the health impact from climatic variability on children between the ages of 12–47 months in the rural areas of Mexico, using the 1999 Encuesta Nacional de Nutricio´n (ENN, National Nutrition Survey) and meteorological data from the Instituto Mexicano de Tecnologı´a de Agua (IMTA). We quantify the extent to which unusual weather has any negative consequences on health outcomes, namely on height-for-age. Based on historical experience and the multiplicity of economic and institutional constraints faced, rural households in Mexico, as most rural households all

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

over the world, have managed to develop traditional strategies for managing climatic risk. Specific to Mexico, Eakin (2000) documents how smallholder farmers have adapted to climatic risk in the Tlaxcala region of Mexico, for example, by planting different crop varieties,3 altering fertilizer and pesticide use depending on the climatic conditions, and diversifying geographically by having plots of land different locations. Yet, quantitative evidence on how successful such strategies are at protecting household welfare is quite scarce.4 To the extent that the current riskcoping mechanisms are not very effective in protecting welfare from erratic weather patterns one can be quite certain that the change in the patterns of climatic variability associated with climate change is likely to reduce the effectiveness of the current coping mechanisms even more and thus increase household vulnerability further. On the one hand, erratic weather may affect agricultural productivity which, depending on how effective were the risk management strategies employed, may translate into reduced income and reduced food availability.5 On the other hand, both temperature and precipitation may affect the prevalence of vector borne diseases, water borne and water washed diseases, as well as determine heat or cold stress exposure (Confalonieri et al., 2007). Many parasitic and infectious species have very specific environmental conditions in which they survive and reproduce, and a slight change in precipitation or temperature could render previously uninhabitable areas suitable for a particular parasitic and infectious species. Specifically in Mexico, several studies have shown positive correlations between temperature, and vector- and food-borne illnesses (Ministry of Environment and Natural Resources, 2007). We focus on the impact of weather on health outcomes of children younger than 48 months living in rural areas. Early childhood health not only affects children’s current wellbeing but may determine their cognitive development as well as their quality of life and productivity as adults (for example, Doyle et al., 2009). Between the ages of zero and three the growth rates are faster than at any other point and thus any delays have a greater probability of affecting

3 For example, planting both corn that is fast maturing but has low yields and corn that is slow maturing but has higher yields, or planting different crops altogether such as wheat instead of corn depending on the prevailing weather (Eakin, 2000). 4 Other studies relying on the perceptions of respondents about the incidence of different types of shocks, such as floods, droughts, freeze, fires and hurricanes include Garcia Verdu (2002), Skoufias (2007) and De la Fuente (2010). None of these earlier studies, however, make use of actual meteorological data. 5 For example, households may undertake ex-ante income-smoothing strategies and adopt low return-low risk crop and asset portfolios (Rosenzweig and Binswanger, 1993). Households may use their savings (Paxson, 1992), take loans from the formal financial sector to carry them through the difficult times (Udry, 1994), sell assets (Deaton, 1993), or send their children to work instead of school in order to supplement income (Jacoby and Skoufias, 1997). These actions enable households to spread the effects of income shocks through time. Additional strategies include the management of income risk through ex-post adjustments in labor supply such as multiple job holding, and engaging in other informal economic activities (Morduch, 1995; Kochar, 1999). Baez (2006) provides a detailed summary of consumption smoothing mechanisms in developing countries.

55

overall growth (Martorell, 1999).6 In developing countries although the children when born are on average at the mean of standardized height-for-age there is a sharp decline in the height-for-age from ages zero months to 24 months and no subsequent catching up in the first five year of life (Shrimpton et al., 2001). However, there is some evidence that under the right conditions stunted children may be able to catch up later on in life such that they no longer are considered stunted (for the Philippines in Adair, 1999; and for Bolivia in Godoy et al., 2010).7 Furthermore, Martorell et al. (2010) find evidence that weight gain the first two years of life had a large effect on schooling outcomes whereas weight gain between two years and four years of age had a weaker one. Reviewing the existing literature on weather and welfare, Alderman (2010) emphasizes that weather-caused nutritional shocks experienced in the first years of life have lasting effects on productivity even if the household is able to overcome poverty later on. Height-forage and weight-for-age are strong predictors of school achievement and that stunting between 12 and 36 months of age is associated with poorer cognitive development (Victora et al., 2008). Malnutrition, from having insufficient food intake or as a byproduct of repeated diarrheal infections, can cause structural damage to the brain and impair motor development in infants which in turn affect the cognitive development of a child (Victora et al., 2008; Guerrant et al., 2008). Furthermore, Eppig et al. (2010) find a correlation between infectious diseases and IQ. They explain their findings as the competition between energy needs for the development of the brain and energy needs needed to fight off disease. They single out diarrheal diseases as potentially being the most energy consuming ones. Overall, childhood health has been found to have an impact on adult health, and employment (Case et al., 2005), cognitive abilities (Case and Paxson, 2008; Grantham-McGregor et al., 2007; Maluccio et al., 2009), educational outcomes (Alderman et al., 2006; Glewwe and Miguel, 2008; Maluccio et al., 2009), and productivity (Hoddinott et al., 2008). These findings underline the importance of focusing on the health outcomes for young children. The health consequences from climatic variability may depend, among other things, on the timing of the weather shock and key individual and household characteristics. For example, weather shocks during the period in which food is relatively scarce may have more of an adverse effect on child health in comparison to the impact of similar shocks during periods in which food availability is more abundant. Malnourished children are more likely to

6 Based on the WHO Child Growth Standards, in the first year the median length for boys increases by 25.5 cm and for girls by 24.9 cm. In the second year the median lengths increase by 12.1 cm and 12.4 cm for boys and girls, respectively. In the third year the median heights increase by 9.0 cm and 9.4 cm for boys and girls, respectively. 7 However, there is no consensus on the conditions for catch-up growth. Although, birth order and number of siblings appears to play a role (Adair, 1999; Godoy et al., 2010) the effect of economic conditions depends on the population studied. For example, Adair (1999) finds that with improved socio-economic conditions some Filipino children stunted at two were no longer considered stunted by 8.5 years of age, but interestingly Godoy et al. (2010) find that improved economic conditions are correlated with lower catch-up rates in the Tsimane’ in Bolivia.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

56

Other drivers

Other income

Environment: Precipitation and temperature

Other drivers

Agricultural income Prevalence of vector-, water- and food-borne illnesses

Coping mechanisms

Consumption Coping mechanisms Other drivers

Health outcomes (e.g.height)

Fig. 1. The channels of impact of climatic variability on different dimensions of household welfare.

become ill (Scrimshaw, 2003). We investigate this issue by examining the extent to which the timing of the climatic shock within the agricultural cycle matters for health. Considering that the agricultural cycle in Mexico is composed of a dry season from October to March next year, and a wet season from April to September, we distinguish among four types of precipitation and temperature shocks: precipitation and temperature shocks in the agricultural year and wet season (t1) prior to the health assessment, and precipitation and temperature shocks in the agricultural year and wet season two periods (t2) prior to the health assessment. It is also quite possible that the resilience and the ability to adapt to changes in weather and environmental conditions differ significantly across the spectrum of socio-economic characteristics in the population. For example, Rose (1999) finds differing effects on girls and boys from rainfall shocks. Behrman and Hoddinott (2005) find that in Mexico children who participate in PROGRESA – an anti-poverty program with a nutritional component – are taller. It may be possible that participation in such programs also protects children in the event of unusual weather. The education of the mother may also interact with weather shocks such that children of less educated mothers are impacted differently than children of more educated mothers. Rubalcava and Teruel (2004) find for Mexican children a positive correlation between mother’s education and cognitive abilities and a child’s height-for-age score. To this end, we estimate the effect of the climatic variability on child health as proxied by height, and interact the weather shocks with different individual characteristics, such as gender, educational attainment of the mother, or participation in supplemental nutrition programs. In order to examine geographically heterogeneous effects we separate the sample by regions and altitude. The rest of the paper is organized as follows: The next section gives an overview of past research on the impact of weather on consumption and on the prevalence of diseases,

which both affect health. Section 3 outlines our estimation strategy. Section 4 gives background on the timing of the shocks with respect to the survey and describes the data sources. Section 5 presents the results and Section 6 concludes. 2. Past research One could think of the environment, health and consumption as being part of a simple system (Fig. 1) where health and consumption are two important dimensions of welfare. Consumption, measured at the household level is influenced by the environment; and health, measured at the individual level, is influenced both by the environment and consumption.8 To see the interaction among the three facets, it is instructive to think of each of the impacts in isolation from the other two. The environment affects consumption in rural areas mainly through its effect on current agricultural production or income given that crop yields are a function of precipitation and temperature. Depending on the household’s ability to cope with income fluctuations, a negative income shock brought on by bad weather may translate into a reduction in consumption (Jacoby and Skoufias, 1998; Dercon and Krishnan, 2000). In addition, the intrahousehold allocation of resources may change after a weather shock possibly leading members to be differentially affected. Differences in the health outcomes of individuals within a household would be brought about if the food resources to a particular individual were reduced sufficiently for her/him to experience malnutrition or if her/his share of

8 There may also be some feedback from the health status of an individual to his/her wage earning capacity and ultimately to the consumption expenditures at the household level. For now, we do not explore this pathway. Also, health affects the consumption bundle directly in two ways—ex post (e.g. being sick requires buying medicines) and ex ante (e.g. preventive health care).

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

other resources, such as preventive or curative health related goods, was lower than in a typical year. Certain subpopulations, such as young children still growing, may be more likely to suffer negative consequences from worse than normal economic conditions (Woitek, 2003). Any effects from economic recession are more likely in poorer households (Sunder and Woitek, 2005). Furthermore, an environmental shock may also directly affect the health of an individual, for example, by changing the prevalence of communicable diseases or the risk of exposure to heat or cold stress. Assuming no changes in consumption choices, a change in the prevalence of communicable diseases itself has an impact on individual’s health depending on the individual’s characteristics and access to preventive measures. The final effect of a weatherrelated shock on health is an interplay among the indirect impact of weather on health through changes in income or production, the direct impact from changes in the environmental conditions, and the changes in the types of consumption that the household and individual are able to make. Studies on the consequences of weather shocks on individual welfare generally use some health outcome as the preferred measure. The evidence from other countries suggests that both gender and age matter. For example, Rose (1999) finds that in rural India a positive rainfall shock increases the survival probabilities of girls more than the survival probabilities of boys. Similarly, Hoddinott (2006) finds that there is a small but transient effect of drought on the BMI of women, but not on men’s. Also, the age of the individual at the time of the shock matters. For example, Hoddinott and Kinsey (2001) find that a drought experienced at 12 months to 24 months of age, had an impact on annual growth rate, and that it persisted for the four years of the study. No such effect was found for shocks experienced later in life. Maccini and Yang (2009) find that in rural Indonesia weather shocks experienced in the first year of life affect women’s outcomes. Namely, women who lived their first year in a locality where the rainfall was greater than the average rainfall of the locality are taller as adults, have completed more years of education, and live in wealthier households. They do not find any impacts on men’s outcomes or from shocks experienced later in life. A particular environmental shock may have a direct negative impact on health but a positive one indirectly through consumption. Both rainfall and temperature are important factors affecting crop yields and exhibit a concave relationship with agricultural productivity (Galindo, 2009). Whether increased precipitation or temperature is beneficial to the agricultural production process depends on the crop, region, and the season in which the change occurs. For example in Mexico, corn production is found to benefit from additional temperature in some regions where as it is found to decrease with additional temperature in others (Galindo, 2009). Similarly, Galindo finds the optimal levels of rainfall below and above which yields fall depend on the class of crops considered. In general within a normal range of rainfall and temperature, additional rainfall or warmer days should increase yields in temperate climates, but will most likely reduce yields in tropical climates. Both extremes of rainfall (drought or flood) and temperature

57

(extremely cold or extremely hot) negatively impact yields and thus, potentially, income and consumption as well. Thus the effects can be expected to differ pending on the underlying average climatic conditions. Malnutrition (and negative health outcomes) are possible, if food consumption is reduced as a result of a weather event especially if prior to the event the household or individual was barely consuming the required nutritional needs. The impact of changes in weather on health are even more complex.9 The prevalence and range of a particular pathogen, disease vector, or animal reservoir are determined by specific ranges of temperature, precipitation and humidity (Patz et al., 2003). Whether an unusually rainy or dry period increases the prevalence of a disease depends on the specific climate of a region. In regions bordering a pathogen’s habitat, even a small deviation from the normal climate, can make large areas susceptible to the infectious disease. That is, if a region is just too cold (or too hot) for a particular pathogen or vector then an unusually hot (or cold) year could make the region susceptible to the disease caused by the pathogen or carried by the vector. Evidence of the importance of climatic factors can be seen from the seasonality of many infectious diseases, such as influenza (to temperature), and malaria and dengue (to rainfall and humidity). In general, extreme temperatures are lethal to disease vectors. An increase in precipitation will in general improve breeding conditions. However, extremely high precipitation, i.e. floods, may, on one hand, reduce infectious diseases by eliminating breeding grounds but, on the other hand, may cause other vectors, such as rodents, to come in more frequent contact with humans. Extremely low precipitation, or droughts, may create stagnant pools of water from streams and rivers, which are good breeding grounds for vectors, thus increasing the prevalence of the diseases associated with such vectors. In addition, besides vectorborne pathogens, water- and food-borne pathogens (causing enteric infections) are also susceptible to precipitation and temperature. Unlike vector-borne illnesses, both heavy and low precipitation have been found to increase enteric infections. Furthermore, there is evidence of a positive relationship between temperature and diarrheal diseases. 3. Estimation strategy For the empirical analyses, we use cross sectional individual level data. We use the standardized z-score for height-for-age as our measure of health. The equation to be estimated can be written as: Hi;l;t ¼ a þ bW l;t þ g X i;l;t þ ml þ ei;l;t

(1)

where Hi,l,t is the height of individual i in locality l at time t, Wl,t is a vector describing the past weather shocks in locality l, at time t, Xi is a vector of other factors influencing the height of an individual, such as household and housing characteristics, ml are location fixed-effects, and ei,l,t is the error term. As long as the weather shock is exogenous, that is E(Wei,l,t) = 0, the coefficient estimate b is unbiased. A

9 The discussion on the impact of climate on health (this and the following paragraph) relies heavily on Patz et al. (2003).

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

58

Table 1 Average climate (1951–1985) and 2007 agricultural production in Mexico, by region. Variable

Annual rainfall (mm) Annual rainfall (std. dev.) Annual growing degree days (GDD) Annual GDD (std. dev.) Wet season rainfall (mm) Wet season rainfall (std. dev.) Wet season GDD Wet season GDD (std. dev.) Municipalities % of cultivated agricultural land. . . Rain-fed In seasonal crops (wet season) In corn (wet season)

North

Center

Pacific

Gulf and Caribbean

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

533 180 4444 307 402 147 2782 178

233 108 806 151 187 77 480 98

966 255 3998 308 788 217 2243 179

494 210 1082 182 345 165 600 101

1302 334 4763 401 1051 304 2564 227

787 181 1130 251 532 150 572 136

1565 371 5531 273 1110 295 3065 149

574 164 1184 169 384 133 603 104

396

919

766

354

0.68 0.76 0.26

0.82 0.71 0.50

0.96 0.51 0.43

0.97 0.30 0.25

Note: Authors calculations from ERIC III (IMTA) for years 1951–1985 for average weather and from Censo Agrı´cola, Ganadero y Forestal (INEGI, 2007) for agricultural production. The North includes municipalities from the states of Baja California, Baja California Sur, Chihuahua, Coahuila, Durango, Nuevo Leon, Sinaloa, Sonora, Tamaulipas, and Zacatecas; Center includes municipalities from Aguascaliente, Colima, Guanajuato, Hidalgo, Jalisco, Estado de Mexico, Michoacan, Morelos, Nayarit, Puebla, Queretaro, San Luis Potosi, and Tlaxcala; Pacific includes municipalities from Chiapas, Guerrero, and Oaxaca; and Gulf and Caribbean includes municipalities from Campeche, Quintana Roo, Tabasco, Veracruz, and Yucatan. The regional assignations are taken from Conroy (2009: p. 39).

locality has a negative or positive weather shock, neg Wl;t ¼ 1 or Wl;tpos ¼ 1, respectively, when the weather variable (such as rainfall) in period t is at least one standard deviation less than or more than the long run average weather in the locality. Given our definition of weather shocks, E(Wei,l,t) = 0 should hold. The height equation can be expanded to include interaction terms to test for the relevance of specific policy measures, thus Eq. (1) becomes, Hi;l;t ¼ a þ b0 W l;t þ b1 ðW l;t  P i;l;t Þ þ g 1 P i;l;t þ g 2 X i;l;t þ ml þ ei;l;t

(2)

where Pi,h,l identifies the type of individual or household. It could indicate, for example, the gender of the individual or whether or not the individual participates in a supplemental nutrition program. In the case of gender, if we set P = 1 for girls, b0 would measures the impact of the weather shock on boys and (b0 + b1) the impact of weather on girls. b1 measures the difference in the effect that a weather shock has on the two sexes. With this strategy we estimate the aggregate impacts from weather shocks on height. We cannot separate whether any negative (or positive) impact are due to changes in consumption such that the child is malnourished (or better nourished) or from more (less) frequent bouts of illnesses. However, as explained in the next section, given the timing of the survey used and by including weather shocks from the two prior agricultural years we gain some insights as to the potential channels through which the shocks are affecting health. 4. Background and data sources Mexico has a substantial population living in poverty. In 2005, the Consejo Nacional de Evaluacio´n de la Polı´tica de Desarrollo Social (CONEVAL, 2005) estimates that nationally 47% of the population lived in poverty, with 18% of the population living in extreme poverty. For all of Mexico, in

2006, 15.5% of 0–5 year-olds had a height-for-age z-score of less than 2 (stunted) and 3.4% of 0–5 year-olds had a weight-for-age z-score less than 2 (WHO). In rural areas the rates were slightly higher with the height-for-age and weight-for-age z-scores below 2 for 24.1% and 4.9% of the 0–5 year olds, respectively (WHO). Furthermore, in Mexico, about 82% of cultivated land is rainfed (INEGI, 2007), and thus susceptible to weather fluctuations. The dependence on rainfed agriculture varies by region, with the Pacific and Gulf and Caribbean regions relying most heavily on it (96% and 97%, respectively), but even in the North 68% of the cultivated agricultural land is rainfed (Table 1). Together these statistics suggest that a relatively large population of the country could be at risk from weather fluctuations. The agricultural year in Mexico runs from October to September. It is composed of a dry season, from October to the end of March, and a wet season, from April to the end of September. For all regions (except for the Gulf and Caribbean) more than 50% of cultivated land during the wet season is in seasonal crops. Corn is of special importance with more than 25% of cultivated land devoted to its production in during the wet season (Table 1) and it is used by many small-scale farmers not only as a source of income but also directly as a subsistence crop. Switching to other crops, such as wheat or barley, which have a shorter growth cycle but are not as useful for household consumption, is considered a last resort (Eakin, 2000). For the empirical analyses we use the Encuesta Nacional de Nutricı´on, ENN, (National Nutrition Survey) collected by the Instituto Nacional de Estadı´stica, Geografı´a e Informa´tica (INEGI, 1999) (National Institute of Statistics, Geography and Informatics) and the Secretarı´a de Salud de Me´xico (Secretariat of Health) in the last quarter of 1999 (or the beginning of the 2000 dry season).10 Table 2 depicts the

10 The ENN can be accessed at http://www.bdsocial.org.mx/, INEGI at http://www.inegi.org.mx/ and the Secretariat of Health at http:// www.salud.gob.mx/.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

59

Table 2 Timing of ENN and agricultural cycle in Mexico.

1997 Oct

1998 …

Mar

Dry Season 1998

Apr

…

1999 Sep

Wet Season 1998

Weather shocks (t-2)

Oct

…

Mar

Dry Season 1999

Apr

…

Sep

Wet Season 1999

Oct

Nov

Dec

Dry Season 2000

Weather shocks (t-1) ENN survey conducted

Age at dry season of agr. year (t-2)

Age at dry season of agr. year (t-1)

Age cohorts (age at survey)

in utero

0 to 11 months

12 to 23 months

0 to 11 months

12 to 23 months

24 to 35 months

12 to 23 months

24 to 35 months

36 to 47 months

Note: ‘‘Agr. year’’ refers to agricultural year.

timing of the health survey and the range of dates used to determine the weather shocks. The survey interviewed 7180 rural households in 174 municipalities. The survey collected general information on all members of the household and more detailed information, including anthropometric measures and illnesses in the past 2 weeks, for females between 12 and 49 years of age, and for all children 12 years or younger.11 The climate data used in this paper come from Instituto Mexicano de Tecnologı´a del Agua—IMTA12 (Mexican Water Technology Institute). The IMTA has compiled daily weather data from more than 5000 meteorological stations scattered throughout the country. The data span a very large period of time – from as far back as the 1920s to 2007 – and contain information on precipitation, and maximum and minimum temperature. The meteorological stations register these variables on a daily basis and we use this information to interpolate daily values of these variables for a geographic centroid in each of the country’s municipalities.13 The centroid is determined as the simple average of the latitude and longitude coordinates of all the localities listed in INEGI’s 2005 catalogue corresponding to each municipality, which results in a locality-based centroid. We chose this method over a population-weighted average because that alternative would bias the interpolation towards urban rather than rural areas. The interpolation method used is taken from Shepard (1968), a commonly used method which takes into account relative distance and direction

11 From the ENN survey, we cannot determine whether or not a rural household is engaged in agricultural activity or what kind of agricultural practices are used. ENN is representative at the regional level and at the urban/rural level and should thus reflect the general population. Included in the sample there will households with small plots practicing subsistence farming as well as those who have irrigated lands and farm large areas of land. What we observe is an average impact over the whole rural population. In our sample, 86.6% of the rural children live in household without tapped water and 74.5% do not have access to an indoor toilet, suggesting that most of the children come from very modest means. 12 Instituto Mexicano de Tecnologı´a de Agua (IMTA). Eric III – extractor ra´pido de informacio´n climatolo´gica V. II. Instituto Mexicano de Tecnologı´a de Agua. The detailed rainfall data can accessed from the list of products available through the IMTA web site: http://www.imta.gob.mx. 13 We took INEGI’s 2005 catalogue of localities, which contained 2451 municipalities.

between the meteorological stations and the centroids.14 Given that some municipalities are extensive, the average municipal weather may not correspond precisely to the weather experienced by all households in the municipality. However, given that the weather measures are based on the average from the nearest stations, they should be highly correlated with actual weather experienced. We carry out an independent interpolation for every day between 1950 and 2007, for each municipality. Since not all meteorological stations existed throughout the entire period and given that during the time they were in operation they sometimes failed to report their records, each interpolation is based on a different number of data points—and indeed different weather stations. These problems, as well as the accuracy of the data, get worse for earlier years, which have a corresponding effect on our interpolations. Thus, interpolations for the year 1950 are less reliable than those for 2007. From these weather data, we calculate the total rainfall and cumulative growing degree days (GDD) for each agricultural year (October to September), and for each wet season (April to September).15 Instead of maximum of minimum temperatures we use GDD, a cumulative measure of temperature based on the minimum and maximum daily temperatures (see Eq. (3)). GDD measures the temperature degree contribution of each day to the maturation of a crop. Each crop, depending on the specific seed type and other environmental factors, has its own heat requirements for maturity. Different corn varieties, for example require between 2450 and 3000 GDDs to mature, whereas different wheat varieties only require between 1800 and 2000 GDDs.16

14 See Skoufias, Vinha and Conroy (2010) for a more detailed description of the procedure. 15 Given that the agricultural year runs from October to September, the first agricultural year that we use to calculate the average weather for municipalities is 1951, and thus we only use the last three months of the 1950 calendar year. 16 For other important crops in Mexico the required GDDs are 2400 for beans and 2200–2370 for sorghum. The GDD values are taken from The Institute of Agriculture and Natural Resources Cooperative Extension, University of Nebraska–Lincoln. Growing Degree Days & Crop Water Use. http://www.ianr.unl.edu/cropwatch/weather/gdd-et.html (accessed 22.07.10).

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

60

Table 3 Prevalence of weather shocks in Mexican municipalities between 1986 and 2002 from mean 1951 to 1985 weather. Type of shock

Standard deviations from mean

Negative shock

2 1 0 1 2

Positive shock

Annual rainfall

Wet season rainfall

Annual growing degree days

Wet season growing degree days

Freq.

%

Freq.

%

Freq.

%

Freq.

%

3487 13,925 53,475 6827 3951

4.3 17.1 65.5 8.4 4.8

3122 13,230 58,014 6804 3542

3.7 15.6 68.5 8 4.2

6102 10,961 45,515 12,045 7042

7.5 13.4 55.7 14.8 8.6

7312 12,174 48,346 11,198 5682

8.6 14.4 57.1 13.2 6.7

Note: Authors calculations from ERIC III (IMTA) weather data from 1951 to 1985 to construct the average weather and the shocks are based on annual and wet season weather from 1986 to 2002.

Furthermore, each crop has specific base and ceiling temperatures, Tbase and Tceiling, respectively, which contribute to growth. The base bound sets the minimum temperature required for growth and the ceiling temperature sets the temperature above which the growth rate does not increase any further. Thus, the contribution of each day, j, to the cumulative GDD is given by T j;min þ T j;max 2

 T base ¼ GDD j

(3)

where T j;min and T j;max are the minimum and maximum daily temperature truncated at the base and ceiling values. In other words, any daily temperature (minimum or maximum) below the base temperature is assigned the base temperature value and any daily temperature above the ceiling temperature is assigned the ceiling temperature value.17 To determine the cumulative GDD at any point in time for a specific cultivation the daily GDDs since planting are summed. Given the mixture of different crops grown in the survey areas, we use the generalized bounds of 8 8C and 32 8C (for example, used by Schlenker and Roberts, 2008). In our specific case, any daily minimum or maximum temperature below 8 8C is treated as being 8 8C and any daily minimum or maximum temperature above 32 8C is treated as being 32 8C. Thus a day with a minimum and maximum temperature of 8 8C or below will yield no GDDs, whereas a day with a maximum and a minimum temperature of 32 8C or above will yield 24 GDDs. The formula does not capture any potential decreases in yields from temperatures above the maximum. For our measures of weather shocks we first calculate the municipal historic mean rainfall and GDD between 1951 and 1985 for the agricultural year and for the wet season. We choose this time span in order to balance the need to use as many years of information, ideally at least 30 years, as possible to calculate the historic weather and to exclude both recent years which may have been affected by changing climate and the earlier years with less reliable data. Given that there is incomplete information for some months for some of our municipalities (i.e. none of the 20 closest weather stations reported data for 5 or more consecutive days), the average climate is based on 15–35

17 We use the Modified Growing Degree Days formula where the minimum and maximum temperatures are adjusted prior to taking the average. See for example Fraisse, Bellow, and Brown (2007).

years of information. 75% of the rural households in our sample live in municipalities with at least 30 years of complete weather information from 1951 to 1985. Our chosen measures of weather shocks, W, are based on the degree of deviation from the 1951–1985 average weather. A shock is defined by an indicator variable identifying those observations where the weather variable is more than one standard deviation from its long-run mean. A municipality is defined to have experienced a negative rainfall shock if the prior period’s rainfall was at least one standard deviation less than the average 1951–1985 rainfall; and a municipality is defined to have experienced a positive rainfall shock if the prior period’s rainfall was at least one standard deviation more than the average 1951–1985 rainfall. Thus, there are in total four measures describing a particular period’s weather. Table 3 shows the 1951–1985 average weather conditions by regions for Mexico. One standard deviation rainfall shock translates to an average of about 30% higher or lower rainfall during the year or wet season. One standard deviation of GDDs is, on average, about 8% from the mean. The climate in each of the regions is distinct and even within a region there is much variability. In general, however, the North is drier than the rest of the country and the Center is colder than the rest of the country. Comparing weather data from 1986 to 2002 with their historic means (from 1951 to 1985), there appears to be an increase in the number of temperature shocks (both negative and positive), but no similar increase in rainfall shocks in Mexico (Table 3). Given that the households were surveyed in the 2000 dry season, we use the weather during the 1999 agricultural year (from October 1998 to September 1999), and during the 1998 agricultural year (from October 1997 to September 1998) to build our set of weather shocks. The shocks experienced in the 1998 agricultural year, capture the weather shocks that would have affected the 1998 wet season harvests and thus agricultural income/production available to the household in the year prior to the household survey and weight measurements. The weather shocks experienced during the 1999 agricultural year would affect the wet season 1999 harvest and thus even if the production was low, the household would not yet feel the effects of low harvest in October/November 1999. Even after a poor harvest agricultural households do not face scarcity during the early months of the following dry season (Chambers et al., 1981). However, the weather shocks during the 1999 agricultural year capture the potential changes in the

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

61

Table 4 Weather shocks in ENN rural municipalities. Type of shock

Standard deviations from the mean

Rainfall

GDD

Annual

Wet season

Annual Municipalities

Wet season

Municipalities

% of sample

Municipalities

% of sample

% of sample

Municipalities

% of sample

Agricultural year t1 Negative shock 2 1 0 1 Positive shock 2

3 32 127 15 3

1.67 17.78 70.56 8.33 1.67

6 32 128 11 3

3.33 17.78 71.11 6.11 1.67

9 24 93 42 12

5.00 13.33 51.67 23.33 6.67

8 14 98 43 17

4.44 7.78 54.44 23.89 9.44

Agricultural year t2 Negative shock 2 1 0 Positive shock 1 2

3 33 136 7 1

1.67 18.33 75.56 3.89 0.56

5 39 127 8 1

2.78 21.67 70.56 4.44 0.56

14 24 106 23 13

7.78 13.33 58.89 12.78 7.22

14 26 115 15 10

7.78 14.44 63.89 8.33 5.56

Note: Authors calculations from ERIC III (IMTA) weather data from 1951 to 1985 to construct the average weather and the shocks are based on annual and wet season weather during the1998 and 1999 agricultural years (or t1 and t2 agricultural years prior to ENN). Based on the ENN rural municipalities with children aged 12 months to 47 months at the time of the survey.

prevalence of weather dependant communicable diseases in the year prior to the survey. Table 4 shows the distribution of rainfall and GDD shocks for the 169 municipalities in our sample. As is evident, there are more GDD shocks than rainfall shocks in our sample of municipalities both during the 1999 and the 1998 agricultural years. Furthermore, a very small number of municipalities (8 and 9) experienced a positive rainfall shock in the 1998 agricultural year and wet season, respectively. Therefore, any coefficient estimates for the positive rainfall shocks during the 1998 agricultural cycle need to be interpreted with caution, due to the small number of observations experiencing such a shock. 5. Results To analyze the average impact of weather shocks on child outcomes we estimate Eq. (1). We use the standardized height-for-age z-score for children between 12 and 47 months of age as our measure of health.18 These children were between zero and 35 months in the prior agricultural year (t1), and in utero to 23 months in the agricultural year two years prior (t2) to the health measurement. We are thus effectively measuring shocks experienced in the first 3 years of life on height. Unlike weight-for-age, height-for-age is not as sensitive to very short-term and immediate scarcities or illness, but would capture more chronic conditions.19 Given that all the data were collected during the 1999 dry season, the year and season terms in Eqs. (1) and (2) drop out. However, we include state level

18 We use, WHO Anthro for personal computers, version 3, 2009: Software for assessing growth and development of the world’s children. Geneva: WHO, 2009 (http://www.who.int/childgrowth/software/en/) for calculating the standardized height-for-age scores. 19 The measure does not capture any differences in mortality from unusual weather.

fixed effects. The decentralized decision process in Mexico gives states the responsibility, for example, of the delivery of health services, water supply and sewage, and rural development and extension services (Cabrero Mendoza and Martinez-Vazquez, 2000). The state level fixed effects control for the impact of the state-based policies on health outcomes as well as any general agro-climatic conditions that vary across states.20 Besides our measures of weather shocks, we also include information on household composition (the number of children, the number of adult males, and the number of adult females), on mother’s characteristics (mother’s education, height, and whether she speaks an indigenous language), information about the child (gender, if the child has an older sibling who was born alive within 2 years of the child’s birth, multiple birth, the birth order of the child, whether the child was characterized as very small at birth, and the age of the child at the time anthropometric measurement was taken), an asset index,21 housing characteristics (presence of indoor toilet, tap water, type of floor), and information about the child’s locality (altitude) as regressors in the analyses.22 Table 5 describes the variables used in the analyses. Given the differences in the average climate in the regions (Table 1), we separate the children into three regions and carry out the analyses separately. Furthermore, there is evidence that altitude and birth weight are related

20 We cannot introduce municipal level fixed effects, which would control more precisely for general agro-climatic conditions, given that our weather shocks are at the municipal level. 21 The asset index is based on the principal factor analysis of the household’s ownership of a radio, a television, a VCR, a telephone, a computer, a refrigerator, a washing machine, a stove, a heater, and motor vehicle. 22 Only when analyzing the effects by participation in a nutritional program, we also include nutritional program participation as a regressor.

62

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

Table 5 Characteristics of the sample of rural children aged 12 months to 47 months. Variable Number of adult males in household Number of adult females in household Number of children (16 yrs or younger) in household Mother’s height (cm) Mother speaks an indigenous language Mother has not completed primary school Sex (1 = female) Birth order Multiple birth Categorized as very small at birth Has an older sibling less than 2 years apart Age: 12 months to 23 months Age: 24 months to 35 months Age: 35 months to 47 months Altitude of locality (km) Household asset score Floor of dirt No tap water to kitchen or bath No proper indoor toilet Region: North Region: Center Region: Pacific, Gulf and Caribbean

Mean 1.318 1.446 3.498 147.7 0.173 0.468 0.514 3.375 0.016 0.075 0.205 0.320 0.324 0.357 1.201 0.304 0.332 0.870 0.750 0.297 0.424 0.279

Std. dev. 0.919 0.885 1.847 23.3

2.477

0.880 0.711

Note: Authors calculations from the ENN. Included in the sample are all rural children between 12 and 47 months and with non-missing information on all covariates. There are a total of 1530 observations. For binary variables the mean indicates the percentage of the sample with the particular characteristic.

(Jensen and Moore, 1997; Yip et al., 1988; Wehby et al., 2010) and that the effects become significant at altitudes greater than 1500 m (Yip, Binkin and Trowbridge, 1988).23 To complement the regional results and to investigate whether the effects of weather shocks are different at different altitudes, we also analyze the impact for children living in low altitudes (less than 1500 m above sea level) and for children living in high altitudes (more than 1500 m above sea level). There are 2007 rural children between the ages of 12 months and 47 months in the ENN dataset and our sample consists of 1530 children. We only include the 1882 children whose mother has not moved in the past two years to ensure that the weather shocks used match what the child experienced. Other children are excluded because of missing height information (128 children), improbable z-scores (35 children),24 or incomplete information on the covariates (189 children). The children measured (and with probable z-scores) have mothers that are statistically significantly taller and more likely to speak an indigenous language, they are more likely to live in lower altitudes and less likely to have running water or indoor sanitation than

23 Furthermore, above 1500 m there are some physiological impacts on humans and specifically in Mexico it has been used as a cut-off altitude for some disease vectors. Besides the correlation between altitude and birth weight, above 1500 m ‘‘physiological changes due to hypobaric hypoxia are detectable’’ (p. 1) (Pollard and Murdoch, 2003). In addition, Herna´ndez-Avila et al. (2006) only use localities below 1500 m in their study on malaria in Oaxaca citing that cases above 1500 m are likely to have been imported. 24 That is, their height-for-age z-scores are less than 6 or more than 6.

those who are not measured. This poses a problem given that those children who were not measured are different, and they may be systematically different in other, unobserved, characteristics as well.25 Bearing these caveats in mind, Table 6 summarizes the average relationship between weather shocks and heightfor-age. The full results for the specification are included in Appendix A. A positive rainfall shock in the 1999 agricultural year or wet season is associated with lower height-for-age scores. This is true for both a positive annual and a positive wet season rainfall shocks. The statistically significant coefficient estimates between 0.87 and 0.32 points are non-trivial given that a z-score of 2 is indicative of stunting and the average height-for-age zscore for the children in the sample is 1.4. The biggest impact is from a positive rainfall shock during the wet season in the North. Children who experienced such a shock have, on average, z-scores 0.87 points lower than children who experienced an average amount of rain during the wet season. The statistically significant effects are negative but smaller in the Central region as well as in the Pacific, Gulf and Caribbean region. Dividing the sample by the altitude of the municipality yields similar negative results. The results from positive rainfall shocks in the 1998 agricultural year need to be interpreted with caution since in our sample there were only a few municipalities (less than 5% of the sample) with positive rainfall shocks in the 1998 agricultural year/wet season. No statistically significant impacts are found for the North or the in the Pacific, Gulf and Caribbean regions and the improbably large coefficient estimate in the Central region most likely is an artifact of few observations rather than a causal correlation. Negative rainfall shocks have a different effect depending on the region and altitude. Children living in the Central region are taller – they have a z-score 0.70 points higher – if the 1999 wet season was at least one standard deviation drier than average than if the wet season was within one standard deviation of the historic mean. In the North and in the Pacific, Gulf and Caribbean region the relationship is not statistically significant. However, in the Pacific, Gulf and Caribbean region a negative rainfall shock in the 1998 agricultural year is associated with taller children. Children living in high altitudes are 0.54 points taller if the 1999 agricultural year was drier than normal and 0.43 points taller if the 1999 wet season were drier than normal. Children living in low altitudes are 0.39 points shorter if the 1998 wet season was drier than normal. Whereas, negative GDD shocks during the 1999 agricultural year are positively correlated with height-for-age in the Central region as well as in high altitudes, negative GDD shocks during the 1998 agricultural year are negatively associated with height-for-age in the Central, and Pacific, Gulf and Caribbean regions as well as in high altitudes. The largest reduction is 0.72 points in the Center

25 If those who are not measured are more likely to be sick (and some of these illnesses are due to the weather), then the coefficient estimates of the weather shock variables is likely to provide a lower bound of the true impact of the weather shock.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

63

Table 6 Impact of weather on child’s height-for-age, by region and by altitude. Type

Variables

Mainly environmental

Negative rainfall shock (t1) Positive rainfall shock (t1) Negative GDD shock (t1) Positive GDD shock (t1)

Both agricultural and environmental

Negative rainfall shock (t2) Positive rainfall shock (t2)a Negative GDD shock (t2) Positive GDD shock (t2)

North

Center

Pacific, Gulf and Caribbean

Annual

Wet season

Annual

Wet season

Annual

Wet season

0.01 0.18 0.59** 0.25 0.02 0.26 0.07 0.17

0.25 0.24 0.87*** 0.3 0.38 0.36 0.19 0.17

0.45 0.31 0.38** 0.16 1.06*** 0.33 0.17 0.19

0.70*** 0.2 0.17 0.22 0.26 0.28 0.13 0.14

0.17 0.17 0.33 0.26 0.04 0.26 0.03 0.15

0.34 0.38 0.43** 0.21 0.23 0.25 0.03 0.19

0.1 0.11 0.01 0.59 0.72*** 0.21 0.03 0.22

0.03 0.15 5.10*** 0.38 0.07 0.21 0.03 0.27 654

0.27* 0.15 0.27 0.24 0.70*** 0.23 0.27 0.2

0.31 0.31

0.1 0.34 0.14 0.48 0.46* 0.27 0.09 0.24

Observations

0.29 0.19 0.14 0.18 0.09 0.22 0.13 0.21 449 Low altitude

High altitude Wet season

Annual Mainly Environmental

Negative rainfall shock (t1) Positive rainfall shock (t1) Negative GDD shock (t1) Positive GDD shock (t1)

Both Agricultural and Environmental

Negative rainfall shock (t2) Positive rainfall shock (t2)a Negative GDD shock (t2) Positive GDD shock (t2)

Observations

0.42** 0.19 0.26 0.17 427

Annual

Wet season

0.06 0.15 0.22 0.21 0.02 0.22 0.11 0.18

0.3 0.24 0.54*** 0.2 0.14 0.25 0.1 0.18

0.54 0.21 0.32* 0.17 0.59** 0.29 0 0.14

0.43** 0.16 0.06 0.19 0.26 0.26 0.12 0.13

0.16 0.16 0.33 0.32 0.2 0.23 0.14 0.19

0.39** 0.16 0.5 0.37 0.13 0.23 0.1 0.19

0.03 0.11 0.5 0.44 0.37** 0.17 0.04 0.21

0.04 0.14 0.52 0.38 0.25* 0.14 0.19 0.27

840

**

690

Note: Robust standard errors in parentheses, clustered by locality. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. Other independent variables included are: household composition (number of children in the household, number of adult males in the household, number of adult females in the household), characteristics of the mother (height, speaks an indigenous language and education), characteristics of the child (age, sex, birth order, multiple birth, small at birth, older sibling less than 2 years older), household assets and housing characteristics (presence of a kitchen, access to tapped water indoors, toilet, floor type), and altitude of locality. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. The regional assignations are taken from Conroy (2009: p.39). Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution. Only 1 child in the central region experienced such a shock. No children in the Pacific, Gulf and Caribbean region experienced such a shock during the 1998 wet season.

region. Unlike for most of Mexico, in the northern states negative annual GDD shocks are positively correlated with height with the average height-for-age being 0.46 points higher after such a shock than had the shock not occurred. Positive GDD shocks are not statistically significantly correlated with height-for-age, regardless of where the child lives or the timing of the shock. The result is consistent whether we separate the sample by geographic regions or by altitude.

Table 6 presents the average impact, but it is possible that not all children experience the same kind of health outcomes from weather shocks. Tables 7–10 present the results when weather shocks are interacted with the sex of the child, the age cohort of the child, the educational attainment of the mother, and the household’s participation in a supplemental nutrition program, respectively. We present the average result for all of Mexico as well as the average results for municipalities below 1500 m above sea

64

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

Table 7 Differential impact of weather shocks on height-for-age, by sex with boys the as comparison group. Variables

All of Mexico Annual

Wet season

Annual

Wet season

Annual

Wet season

Sex: Girl = 1

0.013 (0.143) 0.046 (0.179) 0.244 (0.230) 0.251 (0.154) 0.110 (0.167) 0.328 (0.217) 0.155 (0.253) 0.101 (0.170) 0.207 (0.261) 0.002 (0.138) 0.057 (0.209) 0.004 (0.257) 0.161 (0.279) 0.132 (0.213) 0.270 (0.289) 0.009 (0.164) 0.138 (0.231) 1530

0.059 (0.143) 0.284 (0.184) 0.197 (0.237) 0.293* (0.160) 0.022 (0.159) 0.261 (0.233) 0.308 (0.250) 0.277** (0.130) 0.426** (0.197) 0.039 (0.145) 0.287 (0.193) 0.446** (0.217) 0.466** (0.233) 0.144 (0.175) 0.051 (0.262) 0.056 (0.182) 0.119 (0.237) 1530

0.009 (0.197) 0.222 (0.178) 0.351 (0.270) 0.072 (0.218) 0.279 (0.221) 0.218 (0.293) 0.540* (0.281) 0.193 (0.227) 0.125 (0.249) 0.198 (0.222) 0.073 (0.373) 0.221 (0.316) 0.226 (0.335) 0.319 (0.307) 0.274 (0.301) 0.108 (0.213) 0.036 (0.249) 840

0.050 (0.198) 0.252 (0.287) 0.063 (0.297) 0.522** (0.242) 0.078 (0.260) 0.048 (0.309) 0.384 (0.353) 0.329* (0.197) 0.417* (0.221) 0.269 (0.226) 0.249 (0.312) 0.704** (0.340) 0.539 (0.338) 0.112 (0.244) 0.436 (0.307) 0.174 (0.232) 0.076 (0.283) 840

0.025 (0.223) 0.540 (0.370) 0.056 (0.445) 0.369 (0.259) 0.103 (0.302) 0.613* (0.355) 0.075 (0.358) 0.257 (0.269) 0.474 (0.484) 0.076 (0.193) 0.176 (0.275) 1.100** (0.498) 1.041*** (0.370) 0.171 (0.226) 0.354 (0.338) 0.292 (0.297) 0.494 (0.463) 690

0.015 (0.208) 0.193 (0.254) 0.483 (0.505) 0.038 (0.230) 0.107 (0.277) 0.380 (0.353) 0.267 (0.320) 0.428* (0.220) 0.507 (0.361) 0.170 (0.189) 0.435 (0.292) 0.154 (0.369) 0.938** (0.387) 0.159 (0.202) 0.177 (0.302) 0.397 (0.295) 0.403 (0.478) 690

Negative rainfall shock (t1) . . .  girl Positive rainfall shock (t1) . . .  girl Negative GDD shock (t1) . . .  girl Positive GDD shock (t1) . . .  girl Negative rainfall shock (t2) . . .  girl Positive rainfall shock (t2)a . . .  girl ^ Negative GDD shock (t2) . . .  girl Positive GDD shock (t2) . . .  girl Observations

Low altitude

High altitude

Note: Robust standard errors in parentheses, clustered by locality. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. Other independent variables included are: household composition (number of children in the household, number of adult males in the household, number of adult females in the household), characteristics of the mother (height, speaks an indigenous language and education), characteristics of the child (age, sex, birth order, multiple birth, small at birth, older sibling less than 2 years older), household assets and housing characteristics (presence of a kitchen, access to tapped water indoors, toilet, floor type), and altitude of locality. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. ‘‘. . .  girl’’ indicates the interacted term between the shock above and an indicator variable for girl. Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution.

level and the average results for municipalities above 1500 m above sea level.26 Although, on average in this sample the girls’ and boys’ average height-for-age z-scores are not statistically significantly different, they are significantly different when the child experienced a positive GDD shock in the prior wet season (Table 7). Boys are shorter when the prior wet season was at least one standard deviation warmer than on average. The coefficient estimate is larger for boys living in higher altitudes than for boys living in lower altitudes.

26

As with the average impacts, there are some regional differences as to how the various shocks affect different subpopulation. These results are available from the authors upon request.

Girls are statistically significantly different from the boys in low altitudes, by 0.42 points. Among girls, regardless of altitude, there are no differences between those who experienced an unusually warm year from those who did not.27 However, a positive GDD shock experienced two agricultural years ago (t2), does not have a statistically significant effect on height-for-age for boys or for girls, suggesting that the impact of such shocks do not persist in time. In contrast, in the low altitudes girls are statistically

27 That is, we cannot reject the null hypothesis of b1 + g1 = 0, where b1 is the coefficient estimate on the weather shock and g1 the coefficient estimate on the interaction term of weather and the girl indicator variable.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

65

Table 8 Differential impact of weather shocks on height-for-age, by age cohort with 12–23 months as the comparison group. Variables

All of Mexico Annual

Wet season

Annual

Wet season

Annual

Wet season

Age cohort 2 (24–35 months)

0.321** (0.126) 0.283** (0.137) 0.282 (0.209) 0.012 (0.286) 0.297 (0.255) 0.419** (0.174) 0.074 (0.175) 0.197 (0.219) 0.256 (0.298) 0.079 (0.315) 0.147 (0.279) 0.321* (0.183) 0.421 (0.281) 0.563* (0.287) 0.254 (0.157) 0.334* (0.188) 0.494*** (0.186) 0.032 (0.324) 0.379 (0.308) 0.019 (0.421) 0.334 (0.253) 0.068 (0.287) 0.041 (0.289) 0.134 (0.216) 0.194 (0.315) 0.012 (0.264) 1530

0.309** (0.146) 0.344** (0.139) 0.357* (0.199) 0.135 (0.269) 0.193 (0.215) 0.331 (0.227) 0.054 (0.211) 0.180 (0.264) 0.191 (0.270) 0.319 (0.319) 0.519* (0.310) 0.304* (0.155) 0.310 (0.222) 0.428** (0.216) 0.031 (0.154) 0.108 (0.204) 0.269 (0.173) 0.206 (0.302) 0.025 (0.248) 0.098 (0.345) 0.051 (0.206) 0.013 (0.254) 0.190 (0.274) 0.159 (0.292) 0.414 (0.401) 0.049 (0.309) 1530

0.365** (0.164) 0.330** (0.144) 0.124 (0.232) 0.338 (0.331) 0.234 (0.320) 0.342 (0.291) 0.001 (0.234) 0.144 (0.317) 0.074 (0.258) 0.228 (0.344) 0.407 (0.259) 0.563** (0.234) 0.389 (0.360) 0.939** (0.360) 0.563** (0.258) 0.558* (0.308) 0.535** (0.268) 0.217 (0.361) 0.429 (0.325) 0.075 (0.459) 0.238 (0.243) 0.191 (0.408) 0.114 (0.303) 0.246 (0.278) 0.409 (0.414) 0.154 (0.326) 840

0.420** (0.203) 0.371** (0.164) 0.146 (0.284) 0.037 (0.312) 0.328 (0.229) 0.727** (0.288) 0.156 (0.235) 0.258 (0.407) 0.204 (0.326) 0.291 (0.368) 0.585 (0.391) 0.380* (0.212) 0.334 (0.279) 0.521* (0.294) 0.099 (0.225) 0.295 (0.308) 0.524** (0.242) 0.496 (0.399) 0.122 (0.212) 0.132 (0.399) 0.014 (0.290) 0.436 (0.399) 0.259 (0.367) 0.667** (0.255) 1.103** (0.433) 0.445 (0.322) 840

0.435** (0.191) 0.240 (0.296) 1.027*** (0.286) 0.594* (0.347) 0.811* (0.416) 0.540** (0.223) 0.322 (0.209) 0.367 (0.250) 0.654 (0.544) 0.126 (0.598) 0.224 (0.478) 0.269 (0.278) 0.792** (0.391) 0.037 (0.414) 0.160 (0.177) 0.097 (0.208) 0.352 (0.228) 1.044* (0.613) 0.341 (0.451) 0.849* (0.495) 0.466 (0.374) 0.260 (0.421) 0.134 (0.418) 0.076 (0.351) 0.347 (0.462) 0.444 (0.395) 690

0.319* (0.179) 0.438 (0.272) 0.694* (0.401) 0.412 (0.406) 0.331 (0.620) 0.023 (0.299) 0.331 (0.298) 0.303 (0.274) 0.058 (0.405) 0.690 (0.594) 0.318 (0.435) 0.460* (0.268) 0.642* (0.358) 0.298 (0.354) 0.011 (0.233) 0.031 (0.256) 0.020 (0.285) 1.495*** (0.551) 0.411 (0.596) 1.795*** (0.665) 0.228 (0.279) 0.046 (0.301) 0.011 (0.366) 0.405 (0.545) 0.482 (0.635) 0.136 (0.561) 690

Age cohort 3 (36–47 months) Negative rainfall shock (t1) . . .  age cohort 2 . . .  age cohort 3 Positive rainfall shock (t1) . . .  age cohort 2 . . .  age cohort 3 Negative GDD shock (t1) . . .  age cohort 2 . . .  age cohort 3 Positive GDD shock (t1) . . .  age cohort 2 . . .  age cohort 3 Negative rainfall shock (t2) . . .  age cohort 2 . . .  age cohort 3 Positive rainfall shock (t2)a . . .  age cohort 2 ^ . . .  age cohort 3 ^ Negative GDD shock (t2) . . .  age cohort 2 . . .  age cohort 3 Positive GDD shock (t2) . . .  age cohort 2 . . .  age cohort 3 Observations

Low altitude

High altitude

Note: Robust standard errors in parentheses, clustered by locality. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. Other independent variables included are: household composition (number of children in the household, number of adult males in the household, number of adult females in the household), characteristics of the mother (height, speaks an indigenous language and education), characteristics of the child (age, sex, birth order, multiple birth, small at birth, older sibling less than 2 years older), household assets and housing characteristics (presence of a kitchen, access to tapped water indoors, toilet, floor type), and altitude of locality. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. ‘‘. . .  age cohort’’ indicates the interacted term between the shock above and an indicator variable for the particular age cohort. Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution.

66

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

Table 9 Differential impact of weather shocks on height-for-age, by mother’s educational attainment with mother having completed primary school as the comparison group. Variables

Mother has not completed primary school Negative rainfall shock (t1) . . .  mother has not completed primary school Positive rainfall shock (t1) . . .  mother has not completed primary school Negative GDD shock (t1) . . .  mother has not completed primary school Positive GDD shock (t1) . . .  mother has not completed primary school Negative rainfall shock (t2) . . .  mother has not completed primary school Positive rainfall shock (t2)a . . .  mother has not completed primary school ^ Negative GDD shock (t2) . . .  mother has not completed primary school Positive GDD shock (t2) . . .  mother has not completed primary school Observations

All of Mexico

Low altitude

High altitude

Annual

Wet season

Annual

Wet season

Annual

Wet season

0.167 (0.150) 0.209 (0.154) 0.105 (0.239) 0.148 (0.186) 0.267 (0.255) 0.090 (0.229) 0.332 (0.249) 0.070 (0.116) 0.230 (0.202) 0.073 (0.131) 0.193 (0.181) 0.138 (0.288) 0.365 (0.286) 0.172 (0.188) 0.184 (0.233) 0.074 (0.149) 0.360* (0.212) 1530

0.193 (0.154) 0.379** (0.161) 0.057 (0.201) 0.218 (0.202) 0.231 (0.234) 0.153 (0.231) 0.501* (0.264) 0.112 (0.116) 0.169 (0.181) 0.100 (0.132) 0.113 (0.169) 0.100 (0.198) 0.350 (0.361) 0.012 (0.160) 0.182 (0.210) 0.193 (0.190) 0.482* (0.259) 1530

0.077 (0.177) 0.059 (0.170) 0.277 (0.268) 0.051 (0.242) 0.531* (0.290) 0.245 (0.234) 0.396 (0.261) 0.076 (0.174) 0.007 (0.240) 0.248 (0.225) 0.170 (0.255) 0.024 (0.331) 0.669** (0.334) 0.031 (0.224) 0.353 (0.237) 0.055 (0.210) 0.269 (0.264) 840

0.141 (0.195) 0.241 (0.260) 0.015 (0.257) 0.405** (0.200) 0.315 (0.224) 0.348 (0.341) 0.396 (0.339) 0.122 (0.168) 0.068 (0.241) 0.255 (0.201) 0.108 (0.258) 0.095 (0.320) 1.046*** (0.347) 0.424 (0.256) 0.551* (0.297) 0.124 (0.248) 0.605 (0.368) 840

0.460 (0.278) 0.463 (0.280) 0.120 (0.402) 0.334 (0.326) 0.015 (0.438) 0.418 (0.421) 0.406 (0.410) 0.283 (0.207) 0.620* (0.339) 0.039 (0.161) 0.068 (0.256) 0.335 (0.497) 0.357 (0.303) 0.422 (0.294) 0.111 (0.388) 0.227 (0.249) 0.392 (0.318) 690

0.324 (0.264) 0.430** (0.206) 0.053 (0.308) 0.067 (0.280) 0.087 (0.372) 0.101 (0.310) 0.754** (0.328) 0.276 (0.197) 0.345 (0.309) 0.029 (0.167) 0.060 (0.226) 0.277 (0.315) 1.217*** (0.300) 0.256 (0.196) 0.019 (0.284) 0.283 (0.345) 0.274 (0.386) 690

Note: Robust standard errors in parentheses, clustered by locality. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. Other independent variables included are: household composition (number of children in the household, number of adult males in the household, number of adult females in the household), characteristics of the mother (height, speaks an indigenous language and education), characteristics of the child (age, sex, birth order, multiple birth, small at birth, older sibling less than 2 years older), household assets and housing characteristics (presence of a kitchen, access to tapped water indoors, toilet, floor type), and altitude of locality. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. ‘‘. . .  mother’s education’’ indicates the interacted term between the shock above and an indicator variable for mothers who have not completed primary school. Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution.

significantly shorter than boys, by 0.54 points, after a negative annual GDD shock during the 1999 agricultural year (t1). The coefficient estimates on the interaction term of girls with positive rainfall shocks in the 1998 agricultural year (t2) are also statistically significant; however, given the low number of children who experienced such shocks they must be interpreted with caution. The age of the child at the time of the weather shock also makes a difference (Table 8). Negative rainfall shocks in (t1) have a positive effect on the height of those children who at the time of the survey were 12–23 months and who live in high altitudes, but not on older children or those in low altitudes. In the low altitudes a negative annual

rainfall shock in the 1998 agricultural year is associated with taller children in the youngest cohort, and a negative rainfall shock in the 1998 wet season is associated with shorter children in the oldest cohort. There are no statistically significant differences from positive rainfall shocks in the 1999 cycle by age cohorts. Furthermore, there are differences as to the consequences of positive GDD shocks in the 1999 agricultural year or wet season. Such shocks negatively impact the youngest cohort (12 months to 23 months) but not the older children. Moreover, in lower altitudes, there is negative effect from positive GDD shocks in the 1998 wet season on the youngest cohort, but again not on the older ones.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

67

Table 10 Differential impact of weather shocks on height-for-age, by participation in nutrition supplement program with no participation as the comparison group. Variables

Household participates in a nutrition supplement program Negative rainfall shock (t1) . . .  nutrition supplement Positive rainfall shock (t1) . . .  nutrition supplement Negative GDD shock (t1) . . .  nutrition supplement Positive GDD shock (t1) . . .  nutrition supplement Negative rainfall shock (t2) . . .  nutrition supplement Positive rainfall shock (t2)a . . .  nutrition supplement Negative GDD shock (t2) . . .  nutrition supplement Positive GDD shock (t2) . . .  nutrition supplement Observations

All of Mexico

Low altitude

High altitude

Annual

Wet season

Annual

Wet season

Annual

Wet season

0.336**

0.278*

0.492***

0.405**

0.109

0.220

(0.147) 0.283* (0.165) 0.251 (0.226) 0.184 (0.140) 0.323** (0.160) 0.353 (0.218) 0.123 (0.211) 0.066 (0.127) 0.109 (0.226) 0.042 (0.105) 0.193 (0.161) 0.273 (0.336) 0.471 (0.336) 0.249 (0.191) 0.114 (0.203) 0.027 (0.152) 0.253 (0.265) 1530

(0.142) 0.447*** (0.162) 0.154 (0.236) 0.131 (0.146) 0.447** (0.186) 0.189 (0.206) 0.110 (0.226) 0.020 (0.118) 0.143 (0.189) 0.004 (0.139) 0.411** (0.176) 0.395 (0.252) 0.381 (0.329) 0.178 (0.159) 0.138 (0.197) 0.042 (0.199) 0.138 (0.270) 1530

(0.163) 0.043 (0.236) 0.228 (0.285) 0.049 (0.285) 0.430* (0.238) 0.146 (0.269) 0.209 (0.283) 0.017 (0.211) 0.220 (0.308) 0.063 (0.207) 0.012 (0.214) 0.496 (0.402) 0.418 (0.417) 0.033 (0.264) 0.387 (0.330) 0.047 (0.210) 0.214 (0.329) 840

(0.181) 0.463 (0.285) 0.342 (0.310) 0.263 (0.226) 0.567** (0.250) 0.174 (0.316) 0.040 (0.401) 0.020 (0.214) 0.274 (0.248) 0.463** (0.229) 0.058 (0.266) 0.573 (0.385) 0.142 (0.451) 0.241 (0.235) 0.173 (0.344) 0.171 (0.218) 0.244 (0.332) 840

(0.304) 0.562** (0.254) 0.119 (0.433) 0.330** (0.160) 0.045 (0.359) 0.610 (0.389) 0.023 (0.371) 0.034 (0.175) 0.091 (0.287) 0.060 (0.134) 0.269 (0.249) 0.268 (0.488) 0.447 (0.278) 0.405 (0.270) 0.008 (0.367) 0.201 (0.241) 0.436 (0.356) 690

(0.243) 0.456** (0.200) 0.004 (0.351) 0.078 (0.205) 0.388 (0.420) 0.399 (0.295) 0.176 (0.322) 0.183 (0.178) 0.101 (0.289) 0.211 (0.142) 0.726*** (0.213) 0.151 (0.413) 1.012* (0.581) 0.366* (0.184) 0.249 (0.266) 0.437 (0.394) 0.492 (0.466) 690

Note: Robust standard errors in parentheses, clustered by locality. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. Other independent variables included are: household composition (number of children in the household, number of adult males in the household, number of adult females in the household), characteristics of the mother (height, speaks an indigenous language and education), characteristics of the child (age, sex, birth order, multiple birth, small at birth, older sibling less than 2 years older), household assets and housing characteristics (presence of a kitchen, access to tapped water indoors, toilet, floor type), and altitude of locality. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. ‘‘. . .  nutrition supplement’’ indicates the interacted term between the shock above and an indicator variable for households participating in a nutritional supplement program. Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution.

The educational attainment (as measured by the completion of primary school) of the mother, on average, does not have an effect on children’s height-for-age scores (Table 9). However, when faced with a weather shock, the mother’s educational attainment is associated with a child’s height-for-age. In low altitudes, children from less educated mothers are shorter after a positive annual rainfall shock in 1999 or a negative wet season GDD shock in 1998 than children with a more educated mother. In the higher altitude municipalities, children from less educated mothers are taller after a negative wet season GDD shock than children from mothers with higher educational attainment.

Another household characteristic that may affect the impact from weather on health outcomes is the household’s participation in some type of social protection or assistance program. Supplemental nutrition programs (such as PROGRESA and LICONSA in Mexico) attempt to improve childhood nutrition in the poorest households. Households participating in such targeted programs are from the poorest households in the country who may have fewer resources available to cope with weather shocks. It is interesting to note that, in our sample, children in households participating in a supplemental nutrition program are statistically significantly taller in the low altitudes than children not benefitting from such programs

68

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

(Table 10). However, when faced with certain weather shocks the health of children living in households receiving supplemental nutrition is statistically significantly worse than the health of children not in such programs. Since program participation is not random (that is, the participants come from the most impoverished households), the results do not suggest that participation in such programs is disadvantageous to children. More likely the results suggest that participation in a supplemental nutrition program does not fully level the playing field in terms of child health outcomes after certain weather shock.28 In the low altitudes, annual and wet season positive rainfalls shocks in 1999 are associated with 0.43 and 0.57 point, respectively, decreases in the z-scores of children in nutritional programs when compared to those not in such program. In the high altitudes negative rainfall shocks in 1998 are associated with statistically significantly shorter children when the child’s household participates in a nutritional supplement program than when it does not.

6. Discussion and conclusions Weather-related events can have an impact on the welfare of individuals either through changes in agricultural production and therefore potentially on consumption, and/or through changes in the prevalence of certain types of diseases and ailments associated with different weather conditions. Exploring the consequences of weather on the health of a group of vulnerable individuals – rural children in Mexico between the age of one and four – we find some evidence of both unusual rainfall and unusual temperature affecting child’s height-for-age, and thus potentially their short and long term health and productivity. We cannot determine whether the effects derive from changes in consumption or from changes in the prevalence of diseases and ailments associated with different weather conditions, but the results suggest that potentially both pathways are important. We observe three consistent results. First, positive rainfall shocks in the prior agricultural year (1999) have a negative impact on the average height-for-age regardless of region and altitude. Although all of the coefficient estimates are negative, the statistical significance and magnitude of the impact vary spatially and temporally. In the Central region municipalities and in municipalities in altitudes above 1500 m positive precipitation shocks in the prior agricultural year are statistically significantly associated with lower height-for-age, whereas in the Pacific\Gulf\Caribbean region and in low altitudes it is the wet season precipitation that matters. In the North, both the annual and wet season shocks negatively affect height. Given that any potential food scarcities from a bad harvest in the 1999 agricultural year will most likely be experienced towards the end of the 2000 agricultural year and

28 In order to determine the causal impact of the program (and the interaction of weather shocks with program participation) we would need to determine the counterfactual, that is, the health outcomes after a weather shock for children who participated in such programs had they not benefitted from the programs.

the height measurements were taken in the beginning of the 2000 cycle, these health effects are more likely due to changes in the prevalence of communicable diseases than from under nutrition as a result of a bad harvest.29 The fact that negative rainfall shocks in 1999 are associated with taller children (statistically significantly in the Center region and in high altitudes) suggests that on average in rural Mexico weather-related illnesses are more prevalent with increases in precipitation. Furthermore, in municipalities below 1500 m after a positive rainfall shock children whose families are beneficiaries of a nutritional supplement program are statistically significantly shorter than those who do not. In fact, children in low altitudes who do not benefit from a nutritional supplement are not affected at all by such a shock. Thus, in 1999, participation in a nutritional supplement program did not protect children from the effects of unusual weather. Given that in general only the poorest households are beneficiaries of such programs, and thus participation is non-random, the results suggest that poorer families may not have the resources to protect their children from the effects of changes in the prevalence of diseases to the same degree as wealthier households. Such effects are not observed for the sample of children in the high altitude municipalities potentially due to a smaller percentage of sample households receiving supplements (10 percent versus 20 percent in the low altitude municipalities), or due to differences in how such shocks affect disease prevalence at different altitudes. Since only a few of the municipalities in our sample experienced a positive rainfall shock in the 1998 cycle, we cannot determine whether a positive rainfall shock also potentially affects health through the consumption channel nor whether the observed effects are short term rather than longer term ones. Second, negative GDD shocks, a measure based on cumulative temperature, experienced during the 1998 agricultural cycle have a negative impact on height measurements at the beginning of the 2000 agricultural year. We observe statistically significant decreases in the average height-for-age in both the central and southern parts of the country as well as in high altitudes.30 The negative effects from the 1998 shock suggests that households may not be able to protect themselves from income fluctuations brought on by the colder than usual weather in these regions. Furthermore, in the Central region as well as in high altitudes we observe a positive correlation between a negative GDD shock during the 1999 agricultural year and height-for-age. Together these results suggest that although the immediate effects from a negative GDD shock may be positive, potentially due to lower prevalence of communicable diseases, a year later

29 However, it is possible that part of the impact comes from reduced food availability; for example, if the unusually high precipitation is accompanied by floods that greatly reduce harvest food availability could be affected by the following mid-dry season. 30 In the North, a negative GDD shock in 1998 is associated with taller children suggesting that colder weather improves agricultural production in this region and a negative GDD shock in 1999 is not statistically significantly associated with child height-for-age suggesting that the shock does not affect the prevalence of diseases.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

69

mother from a 1998 GDD shock may be their inability to smooth consumption from changes in agricultural production brought on by warmer weather as easily as their more educated peers. Although we cannot determine whether or not households would be able to change their behavior sufficiently were the increased temperatures permanent, the results do suggest that there are three specific sub-populations – namely boys, children between 12 and 23 months at the age of the survey, and children of less educated mothers – whose caregivers are not currently able to safeguard their children from the effects of warmer weather. Considering the available evidence to date linking childhood health to various aspects of adult wellbeing,31 these results warrant further research into the welfare impact of weather shocks and effective policy options to reduce any negative impacts from unusual weather. Although we cannot say how households will adapt if the temperature and rainfalls now considered shocks become permanent, the results suggest that certain populations may need additional resources to counter potential negative effects at least in the transition phase to the new climatic equilibrium. The results also raise the question whether a ‘‘tailored’’ approach to designing programs aimed at decreasing the sensitivity and increasing the capacity of rural households is likely to be more successful than a uniform program.

such positive gains may have been lost due to decreased food availability in the household. The fact that the negative impact from the 1998 shocks are observed in the Central region, as well as in high altitude municipalities, may reflect the lower average temperatures of these sets of municipalities. On average these municipalities may thus be more likely to experience freezing temperatures with such a degree of damage to crops that households are unable to protect their consumption in the following year. Third, positive GDD shocks, both during the 1999 and the 1998 agricultural cycle do not appear to have an impact on the health of an average child in Mexico. That is, for our sample of municipalities and children, unusually warm weather in the prior two years does not, on average, have statistically significant effects on health as measured by height-for-age. Any changes in the prevalence of diseases (captured by the 1999 shock measure, as well as in the 1998 one) or agricultural income (captured by the 1998 shock measure) are sufficiently small and households are able to mitigate their consequences such that no adverse effects on height are observable. The result suggests that in 1999 an average households was able to cope with intermittent higher temperatures. However, interacting the positive GDD shock with various characteristics of the child yields a more varied panorama. A positive GDD shock in the 1999 agricultural cycle negatively affects the height of boys and of children between the ages of 12 and 23 months at the end of 1999, and a positive GDD shock in the 1998 agricultural cycle negative affects children of less educated mothers. One possible explanation for a negative impact on boys is the difference in morbidity rates between girls and boys especially when marginally malnourished (Wells, 2000) and similarly the negative effect on the youngest cohort may be their greater susceptibility to illnesses such as diarrheal diseases (Kosek et al., 2003) which may increase with temperature (Ministry of Environment and Natural Resources, 2007). The statistically significant decrease in the nationally averaged height-for-age of children of less educated

Acknowledgements We wish to thank Mariano Rabassa, Abla Safir, and the four referees for useful comments and Hector V. Conroy for interpolating the weather data.

Appendix A Table A1 Table A2

Table A1 Impact of weather shocks on height-for-age, by region. Variables

North Annual

Negative rainfall shock (t1) Positive rainfall shock (t1) Negative GDD shock (t1) Positive GDD shock (t1) Negative rainfall shock (t2) Positive rainfall shock (t2)a Negative GDD shock (t2)

0.01 0.18 0.59** 0.25 0.02 0.26 0.07 0.17 0.1 0.34 0.14 0.48 0.46* 0.27

Center Wet 0.25 0.24 0.87*** 0.3 0.38 0.36 0.19 0.17 0.29 0.19 0.14 0.18 0.09 0.22

Annual 0.45 0.31 0.38** 0.16 1.06*** 0.33 0.17 0.19 0.1 0.11 0.01 0.59 0.72*** 0.21

South Wet

Annual ***

0.70 0.2 0.17 0.22 0.26 0.28 0.13 0.14 0.03 0.15 5.10*** 0.38 0.07 0.21

0.17 0.17 0.33 0.26 0.04 0.26 0.03 0.15 0.27* 0.15 0.27 0.24 0.70*** 0.23

Wet 0.34 0.38 0.43** 0.21 0.23 0.25 0.03 0.19 0.31 0.31

0.42** 0.19

31 Childhood health has been shown to have an impact on employment (Case, Fertig and Paxson, 2005), cognitive abilities (Case and Paxson, 2008; Grantham-McGregor et al., 2007), educational outcomes (Glewwe and Miguel, 2008), and productivity (Hoddinott et al., 2008).

70

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

Table A1 (Continued ) Variables

Positive GDD shock (t2) Number of adult males in household Number of adult females in household Number of children (<17 yrs) in household Mother’s height Mother speaks an indigenous language Education mother: primary Sex Birth order Multiple birth Categorized as very small at birth Has an older sibling less than 2 years apart Age: 24–35 months Age: 36–47 months Altitude of locality (km) Household asset score Floor of dirt No piped water to kitchen or bath No proper indoor toilet Constant Observations R-squared

North

Center

South

Annual

Wet

Annual

Wet

Annual

Wet

0.09 0.24 0.08 0.07 0.07 0.09 0.02 0.08 0 0 0.09 0.3 0.09 0.15 0.04 0.12 0.01 0.04 0.85* 0.46 0.31 0.25 0.11 0.15 0.42** 0.16 0.23 0.17 0.29 0.21 0.66*** 0.11 0.14 0.16 0.14 0.22 0.32 0.22 0.84 0.8 449 0.2

0.13 0.21 0.09 0.07 0.09 0.09 0.02 0.08 0 0 0.18 0.32 0.11 0.16 0.03 0.12 0.01 0.05 0.87* 0.45 0.29 0.24 0.12 0.15 0.41** 0.16 0.24 0.17 0.24 0.18 0.69*** 0.12 0.14 0.16 0.1 0.21 0.27 0.2 0.8 0.78 449 0.2

0.03 0.22 0.05 0.07 0.06 0.08 0.11** 0.05 0 0 0.31 0.22 0.31** 0.12 0.06 0.13 0.03 0.03 0.36 0.29 0.70*** 0.2 0.2 0.14 0.08 0.15 0.17 0.13 0.49*** 0.13 0.64*** 0.13 0.09 0.19 0.34 0.25 0.16 0.18 0.24 0.47 654 0.24

0.03 0.27 0.06 0.07 0.04 0.08 0.11** 0.05 0 0 0.24 0.23 0.35*** 0.12 0.06 0.13 0.04 0.03 0.23 0.31 0.71*** 0.19 0.23 0.14 0.06 0.15 0.22* 0.13 0.49*** 0.14 0.56*** 0.13 0.05 0.19 0.27 0.28 0.18 0.18 0.2 0.44 654 0.24

0.27 0.2 0.09 0.11 0.03 0.1 0.04 0.05 0 0 0.54*** 0.17 0.18 0.17 0.09 0.18 0.04 0.04 0.31 0.43 0.24 0.35 0.13 0.2 0.14 0.16 0.26* 0.14 0.05 0.12 0.40** 0.16 0.01 0.16 0.03 0.34 0.23 0.16 0.03 0.57 427 0.27

0.26 0.17 0.06 0.12 0.01 0.1 0.05 0.05 0 0 0.54*** 0.19 0.15 0.17 0.14 0.18 0.04 0.04 0.28 0.48 0.19 0.36 0.13 0.21 0.15 0.16 0.23 0.14 0.05 0.13 0.42*** 0.15 0.01 0.16 0.1 0.33 0.2 0.18 0.04 0.56 427 0.26

Note: Robust standard errors in parentheses, clustered by locality. These coefficient estimates should be interpreted with much caution. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. The regional assignations are taken from Conroy (2009: p.39). Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution. Table A2 Impact of weather shocks on height-for-age, by altitude. Variables

Low altitude Annual

Negative rainfall shock (t1) Positive rainfall shock (t1) Negative GDD shock (t1) Positive GDD shock (t1) Negative rainfall shock (t2) Positive rainfall shock (t2)a

0.06 0.15 0.22 0.21 0.02 0.22 0.11 0.18 0.16 0.16 0.33

High altitude Wet 0.3 0.24 0.54*** 0.2 0.14 0.25 0.1 0.18 0.39** 0.16 0.5

Annual **

0.54 0.21 0.32* 0.17 0.59** 0.29 0 0.14 0.03 0.11 0.5

Wet 0.43** 0.16 0.06 0.19 0.26 0.26 0.12 0.13 0.04 0.14 0.52

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

71

Table A2 (Continued ) Variables

Negative GDD shock (t2) Positive GDD shock (t2) Number of adult males in household Number of adult females in household Number of children (<17 yrs) in household Mother’s height Mother speaks an indigenous language Education mother: primary Sex Birth order Multiple birth Categorized as very small at birth Has an older sibling less than 2 years apart Age: 24–35 months Age: 36–47 months Altitude of locality (km) Household asset score Floor of dirt No piped water to kitchen or bath No proper indoor toilet Constant Observations R-squared

Low altitude

High altitude

Annual

Wet

Annual

Wet

0.32 0.2 0.23 0.14 0.19 0 0.07 0.03 0.07 0.11** 0.05 0 0 0.28 0.2 0.2 0.12 0.02 0.11 0 0.03 0.80*** 0.27 0.53*** 0.2 0.24* 0.13 0.16 0.13 0.16 0.11 0.26 0.17 0.52*** 0.12 0.06 0.14 0.19 0.24 0.09 0.15 0.47 0.6 840 0.33

0.37 0.13 0.23 0.1 0.19 0 0.07 0.04 0.07 0.10** 0.05 0 0 0.26 0.21 0.18 0.12 0.06 0.11 0.01 0.03 0.80*** 0.27 0.55*** 0.19 0.25** 0.13 0.18 0.13 0.16 0.11 0.3 0.18 0.55*** 0.12 0.08 0.13 0.17 0.23 0.06 0.15 0.45 0.6 840 0.33

0.44 0.37** 0.17 0.04 0.21 0.01 0.07 0 0.08 0.02 0.05 0.00** 0 0.53*** 0.19 0.22* 0.12 0.06 0.12 0 0.03 0.08 0.21 0.40* 0.21 0.1 0.15 0.18 0.12 0.29** 0.13 0.25 0.23 0.49*** 0.11 0.08 0.17 0.06 0.2 0.30* 0.16 0.72 0.62 690 0.27

0.38 0.25* 0.14 0.19 0.27 0.01 0.07 0.01 0.08 0.02 0.05 0.00** 0 0.35** 0.17 0.22* 0.12 0.05 0.12 0 0.03 0.1 0.24 0.41** 0.2 0.1 0.15 0.19 0.13 0.30** 0.13 0.38 0.25 0.51*** 0.11 0.07 0.17 0.05 0.19 0.24 0.15 0.47 0.66 690 0.26

Note: Robust standard errors in parentheses, clustered by locality. These coefficient estimates should be interpreted with much caution. (t1) refers to the 1999 agricultural year and (t2) refers to the 1998 agricultural year. Calculated using ENN with state fixed effects. A negative/positive weather shock identifies those municipalities which had at least 1 standard deviation less/more rain (or GDD) than in an average year. The regional assignations are taken from Conroy (2009: p.39). Low altitude refers to municipalities less than 1500 m above sea level and high altitude refer to those above 1500 m. GDD stands for growing degree days. * p < 0.1. ** p < 0.05. *** p < 0.01. a Only 5% of the sample experienced a positive rainfall shock in the 1998 agricultural year. These coefficient estimates should be interpreted with much caution.

References Adair, L., 1999. Filipino children exhibit catch-up growth from age 1 to 12 years. The Journal of Nutrition 129, 1140–1148. Alderman, H., 2010. Safety nets can help address the risks to nutrition from increasing climate variability. Journal of Nutrition 140, 148S– 152S. Alderman, H., Hoddinott, J., Kinsey, B., 2006. Long term consequences of early childhood malnutrition. Oxford Economic Papers 58, 450–474. Baez, J.E., 2006. Income volatility, risk-coping behavior and consumption smoothing mechanisms in developing countries: a survey. Desarrollo y Sociedad 58, 37–83.

Behrman, J.R., Hoddinott, J., 2005. Programme evaluation with unobserved heterogeneity and selective implementation: The Mexican ‘‘PROGRESA’’ impact on child nutrition. Oxford Bulletin of Economics and Statistics 67, 547–569. Cabrero Mendoza, E., Martinez-Vazquez, J., 2000. Assignment of spending responsibilities and service delivery. In: Giugale, M.M., Webb, S.B. (Eds.), Achievements and Challenges of Fiscal Decentralization: Lessons from Mexico. The World Bank, Washington, DC, pp. 139–176. Case, A., Fertig, A., Paxson, C., 2005. The lasting impact of childhood health and circumstance. Journal of Health Economics 24, 365–389. Case, A., Paxson, C., 2008. Stature and status: height, ability, and labor market outcomes. Journal of Political Economy 116, 499–532.

72

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73

Chambers, R., Longhurst, R., Bradley, D., Feachem, R., 1981. 8.1 Seasonality in rural experience. In: Chambers, R., Longhurst, R., Pacey, A. (Eds.), Seasonal Dimensions to Rural Poverty. Allanheld, Osmun & Co. Publishers, Inc, Totowa, NJ, pp. 218–224. CONEVAL, 2005. Mapas de Pobreza por Ingresos y Rezago Social. http:// www.coneval.gob.mx/ (accessed 22.08.10). Confalonieri, U., Menne, B., Akhtar, R., Ebi, K.L., Hauengue, M., Kovats, R.S., Revich, B., Woodward, A., 2007. Human health. In: Parry, M.L.,Canziani, O.F., Palutikof, J.P., van der Linden, P.J., Hanson, C.E. (Eds.), Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK, pp. 391–431. Conroy, H.V., 2009. Sources of welfare disparities across regions in Mexico. In: Skoufias, E., Lopez-Acevedo, G. (Eds.), Determinants of Regional Welfare Disparities within Latin American Countries. The World Bank, Washington, DC, pp. 35–96. De la Fuente, A., 2010. Remittances and vulnerability to poverty in rural Mexico. World Development 38, 828–839. Deaton, A., 1993. Understanding Consumption. In: Carendon Lectures in Economics, Oxford University Press. Dercon, S., Krishnan, P., 2000. Vulnerability, seasonality, and poverty in Ethiopia. Journal of Development Studies 36, 25–53. Doyle, O., Harmon, C.P., Heckman, J.J., Tremblay, R.E., 2009. Investing in early human development: timing and economic efficiency. Economics & Human Biology 7, 1–6. Eakin, H., 2000. Smallholder maize production and climatic risk: a case study from Mexico. Climatic Change 45, 19–36. Eppig, C., Fincher, C.L., Thornhill, R., 2010. Parasite prevalence and the worldwide distribution of cognitive ability. Proceedings of the Royal Society B doi:10.1098/rspb.2010.0973 (published online before print June 30, 2010). Fraisse, J., Bellow, J., Brown, C., 2007. Degree Days: Heating, Cooling, and Growing, http://edis.ifas.ufl.edu/ae428 (accessed 15.02.11). Galindo, L.M., 2009. La economı´a del cambio clima´tico en Me´xico: Sı´ntesis. http://www.semarnat.gob.mx/informacionambiental/Publicacion/Sintesis2009cambioclimatico.pdf (accessed 16.08.10). Garcia Verdu, R., 2002. Risk-Sharing in Rural Mexico. Unpublished manuscript, Banco de Mexico, Mexico, DF. Glewwe, P., Miguel, E.A., 2008. The impact of child health and nutrition on education in less developed countries. In: Schultz, T.P., Strauss, J.A. (Eds.), Handbook of Development Economics. North Holland Press, Amsterdam, pp. 3561–3606. Godoy, R., Nyberg, C., Eisenberg, D.T.A., Magvanjav, O., Sinnar, E., Leonard, W.R., Gravlee, C., Reyes-Garcı´a, V., McDade, T.W., Huanca, T., Tanner, S., Bolivian TAPS Study Team, 2010. Short but catching up: statural growth among native Amazonian Bolivian children. American Journal of Human Biology 22, 336–347. Grantham-McGregor, S., Cheung, Y.B., Cueto, S., Glewwe, P., Richter, L., Strupp, B., the International Child Development Steering Group, 2007. Developmental potential in the first 5 years for children in developing countries. The Lancet 369, 60–70. Guerrant, R.L., Oria´, R.B., Moore, S.R., Oria´, M.O.B., Lima, A.A.M., 2008. Malnutrition as an enteric infectious disease with long-term effects on child development. Nutrition Review 66, 487–505. Herna´ndez-Avila, J.E., Rodrı´guez, M.H., Betanzos-Reyes, A.F., DanisLozano, R., Me´ndez-Galva´n, J.F., Vela´squez-Monroy, O.J., Tapia-Conyer, R., 2006. Determinant factors for malaria transmission on the coast of Oaxaca State, the main residual transmission focus in Mexico. Salud Pu´blica de Me´xico 48, 405–417. Hoddinott, J., 2006. Shocks and their consequences across and within households in rural Zimbabwe. Journal of Development Studies 42, 301–321. Hoddinott, J., Kinsey, B., 2001. Child growth in the time of drought. Oxford Bulletin of Economics and Statistics 63, 406–436. Hoddinott, J., Maluccio, J., Behrman, J., Flores, R., Martorell, R., 2008. Effect of a nutrition intervention during early childhood on economic productivity in Guatemalan adults. The Lancet 371, 411–416. INEGI, 2007. Censo Agrı´cola, Ganadero y Forestal. http://www.inegi.org.mx/sistemas/TabuladosBasicos/Default.aspx?c=17177&s=est (accessed 14.06.10). INEGI, Secretarı´a de Salud de Me´xico, 1999. Encuesta Nacional de Nutricı´on, http://www.bdsocial.org.mx/index.php (accessed 04.10.10). The Institute of Agriculture and Natural Resources Cooperative Extension, University of Nebraska–Lincoln. Growing Degree Days & Crop Water Use, http://www.ianr.unl.edu/cropwatch/weather/ gdd-et.html (accessed 22.07.10). IPCC, 2007. Summary for policymakers. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribu-

tion of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom. Jacoby, H., Skoufias, E., 1998. Testing theories of consumption behavior using information on aggregate shocks: income seasonality and rainfall in rural India. American Journal of Agricultural Economics 80, 1–14. Jacoby, H., Skoufias, E., 1997. Risk, financial markets, and human capital in a developing country. Review of Economic Studies 64, 311–335. Jensen, G.M., Moore, L.G., 1997. The effect of high altitude and other risk factors on birthweight: independent or interactive effects? American Journal of Public Health 87, 1003–1007. Kochar, A., 1999. Smoothing consumption by smoothing income: hoursof-work responses to idiosyncratic agricultural shocks in rural India. The Review of Economics and Statistics 81, 50–61. Kosek, M., Bern, C., Guerrant, R.L., 2003. The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bulletin of the World Health Organization 81, 197–204. Maccini, S., Yang, D., 2009. Under the weather: health, schooling, and economic consequences of early-life rainfall. American Economic Review 99, 1006–1026. Maluccio, J., Hoddinott, J., Behrman, J.R., Martorell, R., Quisumbing, A.R., Stein, A.D., 2009. The impact of nutrition during early childhood on education among Guatemalan adults. The Economic Journal 119, 734–763. Martorell, R., 1999. The nature of child malnutrition and its long-term implications. Food and Nutrition Bulletin 20, 288–292. Martorell, R., Horta, B.L., Adair, L.S., Stein, A.D., Richter, L., Fall, C.H.D., Bhargava, S.K., Biswas, S.K.D., Perez, L., Barros, F.C., Victora, C.G., Consortium on Health Orientated Research in Transitional Societies Group, 2010. Weight gain in the first two years of life is an important predictor of schooling outcomes in pooled analyses from five birth cohorts from low- and middle-income countries. The Journal of Nutrition 140, 348–354. Ministry of Environment and Natural Resources, 2007. Mexico’s Third National Communication to the United Nations Framework Convention on Climate Change. National Institute of Ecology, Mexico, Mexico, DF. Morduch, J., 1995. Income smoothing and consumption smoothing. Journal of Economic Perspectives 9, 103–114. Paxson, C., 1992. Using weather variability to estimate the response of savings to transitory income in Thailand. American Economic Review 82, 15–33. Patz, J.A., Githeko, A.K., McCarty, J.P., Hussein, S., Confalonieri, U., de Wet, N., 2003. Climate change and infectious diseases. In: McMichael, A.J., Campbell-Lendrum, D.H., Corvala´n, C.F., Ebi, K.L., Githeko, A.K., Scheraga, J.D., Woodward, A. (Eds.), Climate change and human health: risks and responses. WHO, Geneva, pp. 103–137. Pollard, A.J., Murdoch, D.R., 2003. The High Medicine Handbook, third edition. Radcliffe Medical Press Ltd, Abingdon, UK. Rose, E., 1999. Consumption smoothing and excess female mortality in rural India. The Review of Economics and Statistics 81, 41–49. Rosenzweig, M.R., Binswanger, H., 1993. Wealth, weather risk, and the composition and profitability of agricultural investments. Economic Journal 103, 56–78. Rubalcava, L.N., Teruel, G.M., 2004. The role of maternal cognitive ability on child health. Economics & Human Biology 2, 439–455. Schlenker, W., Roberts, M., 2008. Estimating the impact of climate change on crop yields: the importance of nonlinear temperature effects. National Bureau of Economic Research, Inc.: NBER Working Paper, number 13799, Cambridge, MA. Scrimshaw, N.S., 2003. Historical concepts of interactions, synergism and antagonism between nutrition and infection. Journal of Nutrition 133, 316S–321S. Shepard, D., 1968. A two-dimensional interpolation function for irregularly-spaced data. In: Proceedings of the ACM National Conference, ACM, New York, NY, pp. 517–524. Shrimpton, R., Victora, C.G., de Onis, M., Lima, R.C., Blo¨ssner, M., Clugston, G., 2001. Worldwide timing of growth faltering: implications for nutritional interventions. Pediatrics 107, e75. Skoufias, E., 2007. Poverty alleviation and consumption insurance: evidence from PROGRESA in Mexico. Journal of Socio-Economics 36, 630–649. Skoufias, E., Vinha, K., Conroy, H.V., 2010. The impacts of climate variability on household welfare in rural Mexico. World Bank Policy Research Working Paper, Number 5555, Washington, DC. Sunder, M., Woitek, U., 2005. Boom, bust, and the human body: further evidence on the relationship between height and business cycles. Economics & Human Biology 3, 450–466.

E. Skoufias, K. Vinha / Economics and Human Biology 10 (2012) 54–73 Udry, C., 1994. Risk and insurance in a rural credit market: an empirical investigation in northern Nigeria. Review of Economic Studies 61, 495–526. Victora, C., Adair, L., Fall, C., Hallal, P., Martorell, R., Richter, L., Sachdev, H., 2008. Maternal and child undernutrition: consequences for adult health and human capital. The Lancet 71, 340–357. Wehby, G.L., Castilla, E.E., Lopez-Camelo, J., 2010. The impact of altitude on infant health in South America. Economics & Human Biology 8, 197–211. Wells, J.C.K., 2000. Natural selection and sex differences in morbidity and mortality in early life. Journal of Theoretical Biology 202, 65–76.

73

WHO Anthro for personal computers, version 3, 2009: Software for assessing growth and development of the world’s children. WHO, Geneva. Accessed from: http://www.who.int/childgrowth/software/en/. WHO. Global database on child growth and malnutrition, http:// www.who.int/nutgrowthdb/database/countries/who_standards/ mex.pdf (accessed 05.07.10). Wilkinson, P. (Ed.), 2006. Environmental Epidemiology. Open University Press, Maidenhead, England. Woitek, U., 2003. Height cycles in the 18th and 19th centuries. Economics & Human Biology 1, 243–257. Yip, R., Binkin, N.J., Trowbridge, F.L., 1988. Altitude and childhood growth. Journal of Pediatrics 113, 486–489.