Research: Policy questions and CGE answers Reply

Our Research posts are about the latest academic research being done in the School of Economics. This week:


You can download the eBook!

by Riaan Rossuw & Waldo Krugell

CGE models have become workhorses for policy analysis since they are particularly suited for answering “what if” questions: What if productivity in agriculture increased, what if foreign demand for exports fell, what if emissions were taxed? Policymakers are typically interested in the direct and indirect effects of specific policy measures, but often these effects are studied only in partial setting. The CGE models have the advantage that the possible effects of specific policy measures can be examined without the excessive simplification and aggregation of partial equilibrium analysis.

Since the 1990s, there has been a considerable increase in the use of computable general equilibrium (CGE) models in South Africa to study a variety of policy issues, specifically, development plans, agricultural programmes, tax policy and international trade. This was set against the backdrop of the issues of the day such as macroeconomic stability and the opening up of the economy. Subsequently, there has been a major shift both in the questions which models address and in the types of models used. Currently the challenges have shifted to the labour market, environmental questions and sub-national issues. A timely appraisal may therefore help to pave the way forward for general equilibrium modelling in South Africa.

In South Africa, early efforts started in 1993 with a small multi-sector model of Naudé and Brixen (1993)[1] and concluded with the operationalisation of the Industrial Development Corporation’s IDCGEM.

Naudé and Brixen (1993) used the above model in comparative static fashion to simulate the impact of a 10 percent increase in real government consumption, a 10 percent increase in export demand, a 10 percent increase in the world price for exports and imports, and a 50 percent reduction in import tariffs. Each of the four experiments listed above were repeated under four sets of closure rules. The results from the experiments were consistent on a number of points. First, the type of closure rule under which the experiments were performed was crucial for the results. Second, the special role played by the mining sector was distinguished. In many experiments the mining sector behaved differently from the other domestic sectors. Third, it was noticeable that the South African economy was particularly vulnerable to external shocks, no doubt due to the specific structure of the South African economy such as the dependence on foreign capital goods, coupled with the tendency for the government to expand expenditure when the gold price rises but not to cut back on programmes when the gold price falls. Fourth, the simulations showed that other aspects of traditional structural adjustment programmes, such as liberalising trade, should be considered. Fifth, the saving-investment closure, whereby investment is savings-driven, played an important role in the results that were obtained. Finally, although some of the simulated policies did have some effect in redressing the unequal distribution of income, these effects were marginal. It was found that the dominance of the government sector was part of the structural problem facing South Africa.

The IDC model was developed in 1993 with the aid of the Impact Research Group of Monash University in Australia. The model was developed with the objective of assisting policymakers in quantifying the impacts of proposed economic policy measures. It was used to evaluate the effects of increases in government spending under different financing methods; the implications of capital flows and trade policy on the industrial sector of the economy; the effects of an increase in government spending on provincial gross products; and the implication of the adoption of Uruguay Round trade liberalisation measures.

The results obtained from applying the model indicate that the income distribution effects of an increase in government spending financed by domestic borrowing and foreign capital inflows are relatively small. The results recommended the adoption of counter measures to offset the real exchange rate appreciation. At this time the IDCGEM was the largest and most frequently used CGE model in South Africa (Burrows, 1995)[2], and was used to inform South Africa’s GATT negotiations. The results indicated that the implementation of the GATT would stimulate both final and intermediate goods, and that by reducing domestic costs relative to international price levels it would stimulate exports.

Our eBook goes on to outline a range of models that extended and updated the early work:


4.1 The Development Bank of Southern Africa (DBSA) Model

4.2 Horridge, Parmenter, Cameron, Joubert, Suleman and De Jongh (1995)

4.3 Coetzee, Gwarada, Naudé and Swanepoel (1997)


Since 2000 there has been an upsurge in interest in CGE modelling in South Africa. This has been lead by international researchers from the World Bank and IFPRI, but local modellers have also been playing an active role in building new models to answer to new policy challenges.

5.1 Humphreys (2000)

5.2 The World Bank CGE model

5.3 International Food Policy Research Institute (IFPRI) models

5.4 De Wet and Van Heerden (2001)

5.5 Kearney and Van Heerden (2001)

5.6 McDonald and Punt (2001)

5.7 Pauw and Edwards (2003)

5.8 Van Schoor and Burrows (2003)

5.9 Van Heerden, Gerlagh, Blignaut, Hess, Mabugu, Chitiga and De Wet (2004)

5.10 Naudé and Coetzee (2004)

To conclude, CGE models are becoming more widely used by policymakers and academics in policy analysis and provide unique insights into the working of economies and on the possible effects of macroeconomic policies. CGE models have, however, only over the past twenty years received attention from economists in South Africa.

The first models were strongly influenced by the work of Naudé and Brixen (1993) and the structural approach to CGE modelling. Recent CGE models, though, are partly built on this approach although many are related. Part of the explanation for this may be found in their difference in focus: extending only a small part of the model such that the model may answer the specific policy questions the model builder intends to address. This has resulted in various models that are sophisticated with respect to small subsections, while staying highly stylised in other parts of the model (Thissen, 1998)[3]. In future work researchers may also want their models to reflect the structural changes in the South African economy that has been taking place, while not forgetting the lessons from earlier models.

The structure of a model is closely related to the parameters that are used. To increase the policy relevance of South African models it is of great importance that the models are also capable of addressing medium- to long-run policy simulations. Therefore, it is crucial that the reliability of parameter estimates increases.

Since building, applying, evaluating, and maintaining a CGE model of an economy is such a large responsibility and time consuming only fractions of individual careers have been devoted to the effort. Institutional effort and funding might be required, both for application and theory. Joint research programmes of modellers, econometricians, theoreticians, and computational economists might be necessary to sustain an effective large scale CGE model of the economy.

[1] W.A. Naudé, and P. Brixen, “A Provisional CGE model for South Africa”, South African Journal of Economic and Management Sciences, 10, 1993, pp 22-33

[2] L. Burrows, “The structure of a CGE model for South Africa”, Paper presented at the biennial conference of the Economic Society of South Africa, Johannesburg, 1995

[3] THISSEN, M. 1998. “Two Decades of CGE Modelling: Lessons from Models for Egypt”, SOM Research Report 99C02, University of Groningen, Groningen, The Netherlands.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s