Global Trade Model

Global trade modeling framework and household/spatial linkages and proposed uses in biofuels work

Building a realistic, policy-relevant model to track the impact of the emergence of biofuels is challenging because we want to pick up three types of effects: a) the effect on the agricultural sector (its crops, livestock, and interregional effects); b) the effect on individual households—both those in rural areas and those in urban areas, and those that engage in different types of enterprise (e.g., those that produce maize compared to those that produce rice); and c) global effects that measure the effect on one country from the action in another through international trade.

To do this, we propose taking three sets of existing models and adapting them so we can look at the three effects, and tying them together to get one set of consistent changes from the emergence of biofuels. We propose to do this across a set of case study countries so we can compare effects across the globe on agricultural production (at the sectoral level), the effects on households in the rural and urban economy, and the effects on international trade.

While much of the modeling framework below was developed in the context of China, it should be noted that modeling the impact of biofuels on China is not a main goal of our proposed work.? Work in China on biofuels modeling is in progress currently and will contribute both directly and indirectly to the project—as part of our effort to understand the rise of biofuels production and consumption in the world and impact pathways.? To shed light on our proposed modeling approach, we describe below the approach that CCAP has taken and the modeling approach embodied in CAPSiM (its analytical platform).? Our intention is to replicate the world General Equilibrium model approach, and to use the same technology and approach that is used by CAPSiM in China to assess some of the linkages inside our case study countries (India, Senegal and Mozambique) with national policy, supply, demand, and trade effects on households.?

Sub-Model 1: Examining the Impact of Biofuels on China’s Agricultural Sector Using CAPSiM

Introduction

China’s Agricultural Policy Simulation and Projection Model (CAPSiM) was developed by the Center for Chinese Agricultural Policy (CCAP) in the middle 1990s in response to the need to have a framework for analyzing policies affecting agricultural production, consumption, price and trade in China (Huang et al., 1999; Huang and Chen, 1999).? Since then CAPSiM has been periodically updated and expanded at CCAP to cover the impacts of policy changes at regional and household levels (Huang and Li, 2003; Huang et al., 2003).

 

 

Model Structure and Data

CAPSiM is a partial equilibrium model for 19 crop, livestock and fishery commodities, including all cereals (four categories), sweet potato, potato, soybean, other edible oil crops, cotton, vegetable, fruits, other crops, six livestock products, and one aggregate fishery sector, which accounted for more than 90% of China’s agricultural output.? CAPSiM is simultaneously run at national, provincial (31) and household (by different income groups) levels.? It is the first comprehensive model for examining the effects of policies on China’s and regional food economies as well as household income and poverty.

CAPSiM includes two major modules in terms of supply and demand balance for each of the 19 commodities. Supply includes production, import and stock changes. Demand includes food demand (specified separately for rural and urban consumers), feed demand, industrial demand, other demand, and export demand.? An example of a crop model is given in Figure 1 (page 33).?? Marketing clearing is reached simultaneously for each agricultural commodity and all 19 commodities (or groups).

Production equations, which are decomposed by area and yield for crops and total output for meat and other products, allow producer’s own- and cross-price market responses, as well as the effects of shifts in technology stock on agriculture, irrigation stock, three environmental factors—erosion, salinization, and the breakdown of the local environment—and yield change due to exogenous shocks of climate and others (Huang and Rozelle, 1998b; deBrauw et al., 2004).? Demand equations, which are decomposed by urban and rural, allow consumer own- and cross-price market responses, as well as the effects of shifts in income, population level, market development and other shocks (Huang and Rozelle, 1998a; Huang and Bouis, 2001; Huang and Liu, 2002).

Most of the elasticities used in CAPSiM were estimated econometrically by Rozelle and Huang using state-of-the-art econometrics and with assumptions that make our estimated parameters consistent with theory (e.g. that demand and supply elasticities change over time).? Recently, CAPSiM shifted its demand system from double-log to An Almost Ideal Demand System (AIDS, Deaton and Muellbauer, 1980), to make demand elasticities vary over time and across income groups.

CAPSiM generates annual projections for crop production (area, yield and production), livestock and fish production, demand (food, feed, industrial, seed, waste, etc), stock changes, prices and trade. The base year is 2001 and is currently updated to 2003.? The model is written in Visual C++.

Applications

CAPSiM has been frequently used by CCAP and its collaborators in various policy analyses and impact assessments.? Some of examples include China’s WTO accession and implications (Huang and Rozelle, 2003; Huang and Chen, 1999), trade liberalization and food security and poverty (Huang et al., 2003; Huang et. al., 2005a and 2005b), R&D investment policy and impact assessments (Huang et al., 2000), land use policy change and its impact on food prices (Xu et. al., 2006), China’s food demand and supply projection (Huang et. al., 1999; Rozelle et al., 1996; Rozelle and Huang, 2000), and water policy (Liao and Huang 2004).

Tracking Changes from the Sector down to the Household Level: CAPSiM-micro model interface module

?????? Because the analysis based on the original CAPSiM framework can only be done at the national level, we have to modify the original model in order to allow us to disaggregate the national impacts into household production, consumption and poverty effects that the emergence of biofuels will have on households in different income groups (and households that have different characteristics—e.g., ethnic status; those households with access to off-farm jobs; or those with certain cropping structure). To do so, we get access to the raw data from the China National Bureau of Statistics (CNBS) for 80,000 rural households (using 2001 data) and 30,000 urban households (also using 2001 data). The raw data is created into household modules that produce different types of crops. There characteristics are linked.

?????? Using a programming platform that allows interaction between the output of the CAPSiM model and the database itself, we can see how households are affected when the price of a set of crops change. The direct impacts on household income and consumption can be tracked. Then, we allow the households in an optimization routine to respond (using elasticities that are consistent with CAPSiM) and the effect on the household after the response to the shocks to the model can be tabulated.

After the response of each household is recorded, the total shifts can be summed for different groups. For example, we could sum the effects on all households under a certain income level (say, the poverty line) and compare it to the effect of the farmers in the middle decile and/or top decile. The effects on a certain minority group could also be tracked as could the effect on suburban or remote farms. This work (linking households to a sectoral and international trade model) is among the first to be carried out in a developing country.