Predicting Caloric Consumption at Mid-Century: Implications for Sustainability
(with Christophe Gouel and Thomas W. Hertel), 2019
The global growth in food consumption has raised concerns about the use of natural resources and the potential environmental impacts of increased agricultural production. The demand for calories from foods is a key contributor to growing global agricultural land use and greenhouse gas (GHG) emissions. For instance, the animal-based calories contribute to over two-thirds of GHGs from agriculture and the ratio becomes even higher (about three-fourths) for global farmland use (FAO, 2015; World Resources Institute, 2016). The daily supply of calories driven by growing demand has increased in both developed and emerging economies, albeit at very different rates across income regions. For example, according to FAO (2017), the average apparent consumption of foodstuffs measured in kilocalories (kcal)/capita/day in China has grown by a factor of two, from 1,453 kcal in 1960 to 3,044 kcal in 2010, while it has only increased by 27% in the United States, compared with increased in real GDP per capita increase of factors 24 and 2.8, respectively. This nonlinear relationship between food consumption and income per capita can be mostly explained by two economic mechanisms: the decrease of food budget share with income (Engel's law) and the eventual saturation of food demand, and the shift toward more nutritious food with income (Bennett’s law). Most analyses of future food demand and its environmental consequences account for these mechanisms only implicitly by following FAO’s demand projections. This has prevented these studies from analyzing the contribution of each potential source (i.e., population increase, income increase, diet shift) to future sustainability stresses.
In this paper, we seek to investigate the relationship between food consumption and sustainability by combining an estimated flexible caloric demand system with an endogenous price mechanism driven by both forces of supply and demand. Our model is a derivative of the partial equilibrium (PE) SIMPLE model of global agricultural supply and demand developed by Baldos and Hertel (2012). For this study, we update the non-homothetic demand system in SIMPLE to reflect the recent work of Gouel and Guimbard (2018) who estimate the Modified Implicitly Directly Additive Demand System (MAIDADS) that governs the allocation of caloric consumption. We follow their estimation methodology and estimate a demand system for SIMPLE using constrained maximum likelihood. We implement non-parametric bootstrap and generate 2,000 replicates of parameter estimates following Cranfield, Preckel, Eales and Hertel (2002) and Gouel and Guimbard (2018). We then run 2,000 simulations using this parameter distribution in our constructed PE model of global agriculture. This set of simulations produces a distribution of model outputs. For our purposes, we focus specifically on the relative contributions of income and population to global land use using the numerical decomposition technique of Harrison, Horridge and Pearson (2000).
In this paper, we seek to investigate the relationship between food consumption and sustainability by combining an estimated flexible caloric demand system with an endogenous price mechanism driven by both forces of supply and demand. Our model is a derivative of the partial equilibrium (PE) SIMPLE model of global agricultural supply and demand developed by Baldos and Hertel (2012). For this study, we update the non-homothetic demand system in SIMPLE to reflect the recent work of Gouel and Guimbard (2018) who estimate the Modified Implicitly Directly Additive Demand System (MAIDADS) that governs the allocation of caloric consumption. We follow their estimation methodology and estimate a demand system for SIMPLE using constrained maximum likelihood. We implement non-parametric bootstrap and generate 2,000 replicates of parameter estimates following Cranfield, Preckel, Eales and Hertel (2002) and Gouel and Guimbard (2018). We then run 2,000 simulations using this parameter distribution in our constructed PE model of global agriculture. This set of simulations produces a distribution of model outputs. For our purposes, we focus specifically on the relative contributions of income and population to global land use using the numerical decomposition technique of Harrison, Horridge and Pearson (2000).
Will Income or Population be the Main Driver of Food Demand Growth to 2050?
(with Christophe Gouel and Thomas W. Hertel), 2018
AgEcon Search, University of Minnesota; 21st Annual Conference on Global Economic Analysis, Cartagena, Colombia
There is considerable uncertainty about how much food will be produced and how much land and other resources will be required in 2050 (Hertel, Baldos and van der Mensbrugghe, 2016). Part of the challenge is that we still do not understand the relative importance of different drivers of demand. Of particular note is the relative role of population and income that drives future sustainability stresses. Over the past 50 years, population has been the main driver of food demand, as global population has risen rapidly from 3.6 billion in 1968 to 7.6 billion in early 2018. Furthermore, the FAO predicts that the world population is expected to grow over 2 billion in the next thirty-some years, while at the same time per capita incomes on a global scale are projected to be raised several-fold. Projections to the future food consumption however are mixed. Baldos and Hertel (2014) argue that income will overtake population as the key driver of food demand as population growth shifts to the poorer countries where per capita consumption is lower and less resource-intensive. Meanwhile strong income growth in the developing countries will lead to dietary upgrading and strong income-driven growth in demand. On the other hand, Gouel and Guimbard (2017) come to the opposite conclusion, arguing that population will persist as the main driver of global food demand to 2050. However, their analysis abstracts from the supply side of the puzzle, assuming perfectly elastic supplies of food and fixed prices. For this reason, the two studies are not comparable. A key factor in this debate is understanding how the income elasticities of demand for food will evolve as incomes rise. This paper seeks to shed further light on this debate by incorporating a new demand system (MAIDADS) into the SIMPLE model of global food supply and demand, using this revised framework to determine the relative importance of income and population growth in driving global food output to 2050.