The income-climate trap of health development: A comparative analysis of African and Non-African countries
Introduction
Malaria is a constant reminder of the impact climate could have on health and development. According to the World Health Organization (WHO, 2009), nearly 250 million people, mostly children, were affected by malaria in 2006 in Sub-Saharan Africa, and the disease causes an average loss of 1.3% annual economic growth in countries with intense transmission. Climate change is expected to raise the number of people vulnerable to malaria in the region dramatically (McMichael, Friel, Nyong, & Corvalan, 2008). Malaria or vector-borne diseases in general are not the only pathway through which climate impacts upon health; heatwaves, air pollution, cyclones, floods, and droughts are other commonly mentioned pathways (see, e.g. Greer, Ng, & Fisman, 2008). The growing concerns about global warming rightly put the climate-health relationship at the centre for the assessment of the impact of climate change on different regions (IPCC, 2007).
Although low income countries are at much greater risk than high income countries in the face of rapid climate change (Haines et al., 2006a, Haines et al., 2006b), previous research on the health effects of climate tended to focus on high income countries due to data availability. Amongst the low income countries, African countries, in general, have significantly larger proportions of populations exposed to hostile climatic conditions, greater dependency on agricultural production, lower income levels and lower life expectancy (Bloom and Sachs, 1998, Collier and Gunning, 1999). According to the IPCC Third Assessment Report (2001, Chapter 10), the African continent had the lowest conversion factor of precipitation to runoff, averaging 15%, and is highly vulnerable to the various manifestations of climate change. Tol, Ebi, and Yohe (2007) show that development is the preferred strategy to reduce infectious diseases in Sub-Saharan Africa in the face of climate change. In view of this, this paper aims to provide a comparative analysis of the health effects from different climatic conditions in African and non-African countries.
Besides geographical bias, previous studies have mostly focused on individual climate-sensitive health hazards in specific countries, such as thermal stress (i.e. heatwave or cold-wave), other extreme weather events (e.g. droughts and cyclones), and infectious diseases (e.g. malaria and cholera), with some extension to crop production and malnutrition (McMichael et al., 2006, Haines et al., 2006a). Research on the social, economic, and demographic effects of climate and its flow-on effects to health is in its infancy (Blashki et al., 2007, Bosello et al., 2006, Dell et al., 2008, Fankhauser and Tol, 2005, McMichael et al., 2006: Tol, 2008, Tol et al., 2007).
This paper adds to the evidence by examining the effect of climate on life expectancy (LE) using a cross-country data set of 158 countries. Although LE does not take morbidity into account, Pitcher, Ebi, and Brenkert (2008) show that LE is clearly positively related to the burden of disease as measured by years of life lived with disability. Cross-country data sets are useful for quantifying the differences in LE that are related to variations in climatic conditions across countries and, most importantly, how this relationship may vary with socioeconomic factors. Life expectancies across countries are routinely compared in many development studies, such as the Human Development Reports, and, thus, given the evidence on climate-sensitive health hazards, meaningful comparisons of LE may require controlling for climatic differences between countries. Furthermore, the returns to health expenditure or interventions may differ across climatic zones, so knowledge about the impact climate has on LE can be used to make cross-country comparisons of health performance and health system efficiency more meaningful.
Examining the climate-LE relationship is made challenging by a number of factors. First, climate (and geography in general) affects not only health but also economic development, especially via agricultural production (Gallup et al., 1999, Masters and McMillan, 2001, Sachs, 2001). At the same time, income (alongside other elements of development) is vital in determining health status. Therefore, the climate-LE relationship should not be examined in isolation from the income-health relationship. Second, while climate can adversely affect production and health, income and economic development can in turn provide protection against these adverse effects (Kinney et al., 2008, Woodward et al., 2000). Combined with the last factor, this implies that climate, income and LE could have complex interactions and multiple causalities (Campbell-Lendrum & Woodruff, 2006). Third, climate is multidimensional; a good climate for LE may not always be good for productivity, and vice versa. For instance, frost can help kill pests that damage crops as well as parasites that infect humans; whereas regular rainfall which is beneficial to crop production can also lead to mosquito breeding. Previous studies like Sachs (2001) and Masters and McMillan (2001) focused largely on the tropical-temperate climate zone division which may risk overly simplifying the situation. Also, Pitcher et al. (2008) develop a population health model that is rich in socioeconomic measures but only considers a single dimension of climate (temperature), which is not statistically significant in the presence of other control variables.
The current study's objective is to address these challenges both theoretically and empirically. Although this study aims to provide a theoretical framework to understand the dynamic interactions between health, economic growth and climate, the econometric analysis, due to data limitations, largely focuses on the current long-run outcome of these interactions (though it also involves analysing changes in cross-sectional results over time). In the absence of time-series data, simulations may be better suited to explore the potential short-run dynamics. For example, Fankhauser and Tol (2005) use macroeconomic models to examine the long-term effects of climate change on growth through its negative impacts on the savings rate and capital accumulation. Using a multi-country computable general equilibrium model (GTAP-E), Bosello et al. (2006) simulate the impacts of climate change on labour productivity and demand for health care and once again find the investment channel crucial. Furthermore, Tol (2008) uses the integrated assessment model “FUND” to show that economic growth is important in determining the incidence of malaria morbidity and mortality.
To help understand the complex relationship between climate, income and LE, we first formulate, in Section 2, a coherent “income-climate trap” model of development to explain the vastly different income levels and LE outcomes between countries. A number of hypotheses on the interconnection between climate, income and LE are drawn from the model. In Section 3, we use regressions methods to test these hypotheses in the context of African and non-African comparisons. The empirical findings are reported in Section 4. Lastly, Section 5 concludes the paper with a discussion of the policy implications of the findings.
Section snippets
The basic model
Fig. 1 is a schematic representation of the theoretical model that demonstrates the possible underlying relationships between climate, income (proxy for development) and LE (proxy for health status). Path (a) indicates the direct effect of climate on LE, which could come from the spread of tropical diseases to natural disasters or from the availability of rain water to extreme temperatures (e.g. Blashki et al., 2007). Path (b) indicates the direct effect of climate (e.g. rainfall or drought) on
Data and methodology
While longitudinal data are desirable in testing the above hypotheses, consistent and reliable longitudinal data on health status, climate and other factors do not exist (Patz, Campbell-Lendrum, Holloway, & Foley, 2005). In this paper, a cross-sectional data set is compiled instead to test the above hypotheses. Cross-sectional data allows the examination of the extent to which cross-country income differences can account for differences in the effect of climate on LE across countries. Although
Variation partition
Table 1 reports the results of model (1) for the full sample of 158 countries. As expected, LGDPPC has a positive sign and is highly significant. Climate variables differ in the direction of their effects, and not all climate variables are individually significant. This may be related to the high correlation between certain (but not all) climate variables which can result in abnormally high standard errors and thus potentially wrong signs for the estimated individual climate variable
Discussion and conclusion
The current study represents an effort towards a better understanding of the inter-relationships between climate, income and LE. A caveat worth mentioning is that the analysis relies on the use of cross-sectional data to examine the relationship between climate, income and LE across countries. As such, establishing causation between variables is difficult as relationships may be due to important variables that are correlated but not caused by either climate or income, being omitted from the
References (49)
- et al.
Economic growth and greenhouse gas emissions
Ecological Economics
(2002) - et al.
Economy-wide estimates of the implications of climate change: human health
Ecological Economics
(2006) - et al.
On climate change and economic growth
Resource and Energy Economics
(2005) - et al.
Climate change and human health: impacts, vulnerability, and mitigation
The Lancet
(2006) - et al.
Climate change and human health: impacts, vulnerability and public health
Public Health
(2006) - et al.
Mortality decline, human capital investment, and economic growth
Journal of Development Economics
(2000) - et al.
Approaches for estimating effects of climate change on heat-related deaths: challenges and opportunities
Environmental Science & Policy
(2008) Effective aid: how democracy allows development aid to improve the quality of life
World Development
(2003)- et al.
Climate change and human health: present and future risks
The Lancet
(2006) - et al.
The weakest link hypothesis for adaptive capacity: an empirical test
Global Environmental Change
(2007)