Future Scenarios for Potato Demand, Supply and Trade in South America to 2030

Abstract

This paper presents estimates for potato demand, supply and trade in South America to the year 2030 according to three scenarios: baseline, high demand and limited supply. The results highlight the importance of Brazil with its massive population and low per capita consumption of potato as a key driver of regional outcomes. According to the baseline and high demand scenarios, improved productivity in Andean countries such as Ecuador and Colombia will influence output and consumption increases in those locations as well. The potential adverse effects of the advent of climate change on the potato sector in more vulnerable growing areas in the region will result in much more modest increases in output in those locations according to the low supply scenario. While domestic potato marketing will continue to expand, foreign trade in potatoes remains small in absolute terms and as a percentage of national and sub-regional output. The findings call attention to opportunities for agribusiness initiatives in input markets as well as for both fresh and processed potato products for human consumption in the decades ahead.

Appendix

The food module of the IMPACT model is a partial equilibrium representation of the global agricultural sector. It comprises 45 agricultural commodities and distinguishes 115 geopolitical regions and 126 water basins, which combine to 281 food production units (FPUs), thus offering a disaggregation of the analysis at the sub-national level. Crop production takes place at the level of the FPUs. For each period t, FPU n and commodity i, agricultural production is depicted by isoelastic functions for area AC and yield YC:

$$ {\mathrm{AC}}_{tni}={\alpha}_{tni}\times {\left({\mathrm{PS}}_{tni}\right)}^{\varepsilon iin}\times {\prod}_{j\ne i}\left({\mathrm{PS}}_{tnj}\right)\times \left(1+{\mathrm{gA}}_{tni}\right) $$

(1)

$$ {\mathrm{YC}}_{tni}={\beta}_{tni}\times {\left({\mathrm{PS}}_{tni}\right)}^{\gamma iin}\times {\prod}_k{\left({\mathrm{PF}}_{tnk}\right)}^{\gamma ikn}\times \left(1+{\mathrm{gCY}}_{tni}\right) $$

(2)

The two functions together represent QS, the supply of each commodity for each region:

$$ {\mathrm{QS}}_{tni}={\mathrm{AC}}_{tni}\times {\mathrm{YC}}_{tni} $$

(3)

As shown by Eqs. (1)–(3), agricultural production is assumed to be a function of input prices PS and output prices PF. Specific shifters gA and gCY incorporate intrinsic changes in area and yields and allow capturing anticipated trends in both variables. The production of livestock is represented by a similar set of functions for the number of animals and yields per head.

On the demand side, for each region, a set of separate functions is used to represent different demand components, namely food, feed, biofuels, crush demand for oilseeds and other uses, which add up to total demand. In this system, food demand QDF, again represented by an isoelastic function, is a function of own and cross prices PD, income INC and population size POP (see Appendix Table 8 for the key parameter estimates used for this study):

$$ {\mathrm{QDF}}_{tn i}={\alpha}_{tn i}\times {\left({\mathrm{PD}}_{tn i}\right)}^{\varepsilon iin}\times {\prod}_{j\ne i}{\left({\mathrm{PD}}_{tn j}\right)}^{\varepsilon ijn}\times {\left({\mathrm{INC}}_{tn}\right)}^{\eta_{in}}\times {\mathrm{POP}}_{tn}. $$

(4)

In which remaining components of demand are modelled as follows: feed demand is a fixed share of total supply, adjusted for price effects and technological progress in feed efficiency. Demand for biofuels is a function of government blending mandates, energy prices and policy support provided to producers. Crush demand for oilseeds derives from specific oil and meal processing ratios and the prices of oil and meal by-product and the oilseed commodity. For additional details on the demand components other than QDF, the interested reader should refer to Rosegrant et al. (2012).

Table 8 Key IMPACT parameters used for this study

The individual regions for which supply and demand is calculated are connected to each other via trade. Net trade adds to domestic supply and stocks to equilibrate domestic supply and demand. Global demand and supply for each commodity is brought into equilibrium for each subsequent year by an iterative process to calculate an endogenous world market price, which clears the world market and determines the domestic producer and consumer prices for all commodities and all regions. When an exogenous shock is applied to the model, such as the HD scenario presented in this study, the world market price adjusts to those changes to establish new market equilibrium with a new set of country level prices, demand, supply and trade. As a result, estimated annual series of projected market-clearing prices, commodity and country-specific levels of crop utilization, production area, yield and production levels by crop and country, and net trade flows are produced (Scott et al. 2000).

In the HD scenario, the increase in food demand for potatoes is introduced into the model from 2015 onwards as a 1% shift in the demand functions (Eq. 4) of the countries and sub-regions under consideration. This shift is increased over the subsequent years by 1% steps to reach its maximum of 15% by 2029.

For the LS scenario, slower productivity growth is introduced by reducing the intrinsic yield growth rates (Eq. 1) of the baseline scenario to obtain 1% lower yields in Argentina and Chile, 2% lower yields in Brazil and 6.9% in the other Andean sub-region (e.g. Ecuador, Colombia, Venezuela). For Peru, yields remain unchanged. Finally, it is assumed that not only potato production in SAM is negatively affected in that scenario but that the rest of the world also suffers from adverse supply and demand conditions. Accordingly, a reduction in potato yields in 2030 compared to the baseline scenario is introduced for the rest of the world. This, again, is implemented by the mean of lower intrinsic yield growth rates in Eq. 1.