1. TERM as a tool for sub-national CGE modelling
Policy analysis increasingly concerns the regional dimension. Real GDP remains an
important indicator of economic growth. Simulations using single country CGE models
have been used since the late 1970s to examine the winners and losers from policy
changes or changes in global economic conditions (see Dixon et al, 1982). In the
Australia context, the resources boom and droughts of this millennium have contributed
to wide disparities in growth rates between regions.
The first objective of this paper is to provide a brief overview of various applications of
The Enormous Regional Model (TERM). Then two variants of TERM, representing the
economies of Australia and Indonesia, are used to analyse the impacts of the recent
improvement in the terms-of-trade faced by each country.
TERM is increasingly becoming the CGE model of choice for sub-national multi-
regional modelling. In Australian applications, variants of TERM have proliferated to
undertake different applications:
• The impact of the 2002-03 drought on Australia (Horridge et al., 2005)
[comparative static].
• Modelling of water trading scenarios (Wittwer, 2003; Peterson et al., 2005) and
future rural-urban water requirements (Young et al., 2006)) [comparative static
with water accounts, projections of the database ahead 25 years in the case of
Young et al. ].
• The economic impacts of improved weeds management (Wittwer et al., 2005),
plant disease outbreaks (Wittwer et al., 2006), natural disasters and
telecommunications upgrades (CoPS studies for clients) [dynamic].
We are undertaking further work at CoPS to enhance TERM’s capabilities in water-
related scenarios (Dixon et al., 2005). And a little known variant of TERM is able to
model economic scenarios that present changes in real income by federal electorate (see
http://monash.edu.au/policy/archivep.htm#tpmh0074).
2. Why TERM is quicker
The advantage of TERM over predecessors, notably MMRF (Naqvi and Peter, 1996), is
its ability to handle a greater number of sectors and regions while still being able to solve
relatively quickly. In TERM, the user, the regional source and the regional destination for
each commodity are not detailed in a single huge matrix. Rather, two much smaller
matrices separately provide data on (1) the user by regional destination and (2) the
regional source and regional destination for each commodity. A common-sourcing
assumption (i.e, all users of a given commodity source it in common proportions from