Trade of South Africa with the rest of the world has changed substantially since the mid-1990s.
The United States (US), which used to be the main trading partner of South Africa, is now only
the third largest trading partner of the country. South African trade with Germany, Japan and
the United Kingdom (UK) are also lower. The key reason is the emergence of China in the
world economy. South Africa did not trade with China before 1993, but from 2009 China
became the main trading partner of the country. Globalisation and China’s emergence have
influenced the trade linkages of many other countries in the world. To incorporate the changes
in global trade linkages, the foreign variables of all the models in the study are compiled with
trade-weighted three-year moving average data.
The foremost objective of the thesis is to determine how the changes in trade linkages affect the
transmission of economic shocks originating in the rest of the world on South Africa. The global
vector autoregression (GVAR) approach is used since one of its advantages is the incorporation
of global trade linkages, which facilitates the analysis of the transmission of shocks from one
country to another.
As a GVAR model combines many individual country models, the study first estimates such a
country-specific model for South Africa to determine whether it displays the expected impact of domestic shocks on the economy. This type of model is known as a vector error correction
model (VECM) with domestic variables and weakly exogenous (X) foreign (*) variables, denoted
by VECX*. The results from the VECX* for South Africa are in line with expectations, showing
the effective transmission of monetary policy.
The study then examines the impact of international shocks on the South African economy with
a GVAR model. The GVAR, which incorporates country-specific VECX* models for 33
countries, is solved for all 33 countries using global trade weight matrices at different dates. The
results indicate that over time South Africa is much more vulnerable to GDP shocks to the
Chinese economy, and less vulnerable to GDP shocks to the US economy. These trends are
however not confined to South Africa, and as such highlights the increased risk to the South
African economy and many other economies, should China experience slower GDP growth.
Finally, the thesis determines whether the forecasting performance of GVAR models is superior
to that of a country-specific VECX* model. The study compares the out-of-sample forecasts of
two key South African variables (real GDP and inflation) for five types of models: a VECX*, a
customised small GVAR for South Africa, the more general 33-country GVAR, simple
autoregressive models and random walk models. Better forecasts of both the GVAR models
compared to the VECX* model at forecast horizons of more than four quarters show that,
despite the complicated nature of the GVAR model with the inclusion of many countries and
global trade linkages, the additional information is useful for forecasting domestic variables