This paper discusses South Africa’s recent growth performance and its new growth targets. It analyzes the history of exchange rate volatility, compares it with other countries, and examines the relationship between volatility and trade flows in South Africa. It highlights facts on reserves holding and presents an empirical analysis of a model in assessing the adequacy of South Africa’s reserves. It also analyzes the cyclical balances to determine revenue performance and discusses the penetration of the South African financial conglomerates into State Security Agency and assesses potential vulnerabilities.

Abstract

This paper discusses South Africa’s recent growth performance and its new growth targets. It analyzes the history of exchange rate volatility, compares it with other countries, and examines the relationship between volatility and trade flows in South Africa. It highlights facts on reserves holding and presents an empirical analysis of a model in assessing the adequacy of South Africa’s reserves. It also analyzes the cyclical balances to determine revenue performance and discusses the penetration of the South African financial conglomerates into State Security Agency and assesses potential vulnerabilities.

II. Exchange Rate Volatility in South Africa1

A. Introduction

1. In defining its new Accelerated and Shared Growth Initiative for South Africa (ASGISA), the South African government group overseeing the initiative listed exchange rate volatility as a possible major impediment to higher growth. Over the past decade, South Africa has had several episodes of high exchange rate volatility. The 2006 budget speech mentions that the volatility of the exchange rate and its impact on exports deserves more analysis.

2. Theoretically, exchange rate volatility can affect growth through both trade and other channels.2 Increased trade can have a number of positive impacts on growth. According to Grossman and Helpman (1991), trade may cause a country to (i) use a larger variety of intermediate goods and capital equipment, which could enhance the productivity of its other resources; (ii) acquire technology developed worldwide, especially embodied in capital goods; (iii) increase the variety of consumer goods; and (iv) improve efficiency in the use of resources, which can induce a change in market structure and reduce markups. Jonsson and Subramanian (2000) find a robust significant positive effect of increased trade on total factor productivity growth in South Africa in the 1990s.

3. The general presumption that trade is adversely affected by high exchange rate volatility does not necessarily hold in all economic models; specific assumptions are required. For instance, as Clark and others (2004) point out, it is usually also assumed that:

  • There are no hedging possibilities either through the forward exchange market or through offsetting transactions. In South Africa, where the forward exchange market is well developed, capital account restrictions limit the extent to which firms can access it.

  • If firms can adjust factor inputs only with some lag, exchange rate volatility could negatively affect trade.

  • Engagement in international trade usually causes firms to incur fixed costs that may create some inertia in their decisions to enter and exit international markets. The degree of this inertia would increase with higher exchange rate volatility.

4. To understand the relationship between exchange rate volatility and trade in a general equilibrium framework it is necessary to understand the interaction of all major economic variables. Here a structural shock that affects the exchange rate may lead to changes in other macroeconomic variables that offset the effect on trade of the change in the exchange rate. Especially in countries that are often hit by significant terms of trade shocks, adjustments in the exchange rate, perceived as high volatility, may actually act as automatic stabilizers that help the economy adjust trade flows more smoothly. Broda (2001) shows that flexible exchange rates insulate economies from terms of trade shocks far better, but the benefit comes at the expense of a more volatile real exchange rate.

5. Studies have found only a weak link between trade flows and exchange rate volatility. Clark and others (2004) analyze the effect of exchange rate volatility on bilateral trade flows for a large group of countries (Box II.1). They confirm the findings of previous studies that exchange rate volatility has a small but statistically significant negative effect on trade. The effect, however, is not robust to alternative ways of controlling for factors other than exchange rate volatility.

The Impact of Exchange Rate Volatility on Bilateral Trade Flows

Clark and others (2004) use the gravity equation model to estimate the effects of exchange rate volatility on bilateral trade flows. This is the standard approach used in the empirical literature to estimate the effects of various trade costs on trade flows. It models trade between two countries as a function of the incomes of the two countries and the distance between them. It has been very successful in explaining bilateral trade flows in a variety of applications. Moreover, the gravity model has solid theoretical foundations (see Anderson, 1979; Deardorff, 1998; and Anderson and van Wincoop, 2003, for example). In addition to incomes and distance, the empirical specifications of the gravity model typically control for other factors that may enhance or reduce trade between two countries, such as cultural linkages, historical relations, and geographic locations.

Following the panel specification in Rose (2004), Clark and others (2004) regress the log of aggregate bilateral trade on numerous explanatory variables, including the log of income of the bilateral pair, the log distance between the pair, and a measure of exchange rate volatility. The estimation is conducted using ordinary least squares with robust standard errors. Their data set covers the bilateral trade of 178 IMF member countries for every fifth year from 1975 to 2000. As a measure of exchange rate volatility, Clark and others use the standard deviation of log changes of the monthly bilateral real exchange rate calculated over five years using IFS data on nominal exchange rate and consumer prices. The study also considers alternative measures of volatility, such as nominal exchange rate volatility and the conditional volatilities estimated from GARCH regressions, to check the robustness of the results.

Clark and others (2004) confirm the findings of previous studies that exchange rate volatility has a statistically significant small negative effect on trade: If exchange rate volatility were to rise by one standard deviation, bilateral trade would fall by 7 percent. However, this effect does not hold up against variations in specifications. Different measures of exchange rate volatility do not yield robust results, either. Clark and others conducted separate analyses for homogeneous products and differentiated products and found a negative effect of exchange rate volatility on trade in differentiated products. However, this effect also does not stand up to the ways of controlling for variables other than exchange rate volatility. They conclude that there is no strong rationale for taking policy measures to restrict exchange rate movements in an attempt to promote trade flows.

6. This chapter looks at the history of exchange rate volatility in South Africa, compares it to volatility in other countries, and examines the relationship between volatility and trade flows in South Africa using disaggregate import and export data.

B. Recent History

Measuring Exchange Rate Volatility

7. Exchange rate volatility is defined as the standard deviation of changes in the log of the monthly real effective exchange rate. Exchange rates measured in real terms are appropriate on theoretical grounds, but nominal exchange rate is also sometimes used, given the stickiness of domestic prices (at least in the short run). The use of the standard deviation of percentage changes implies that a constant trend in the exchange rate does not affect this measure. Other measures of exchange rate volatility that have been employed include the standard deviation of the level of the nominal exchange rate, the standard deviation of percentage changes of the nominal exchange rate, and measures that try to isolate the unforecastable component of exchange rates by conditioning volatility on historical information. We use the real effective exchange rate to calculate the benchmark measure of exchange rate volatility.

8. Over the past decade, South Africa experienced several episodes of high exchange rate volatility (Figure II.1). In 1998, the rand came under severe pressure, and by the end of the year, the real effective exchange rate had depreciated by 15 percent; similar pressure in 2001 brought it down by 25 percent. While the 1998 drop was associated with the Asian crisis, the causes for the drop in 2001 are unclear. In both years, the sharp depreciation also led to a sizable increase in volatility, though since 2001, volatility has somewhat abated. A similar pattern was observed for alternative measures of volatility using nominal and real exchange rates against the US dollar (not reported here).

Figure II.1.
Figure II.1.

Volatility of the Real Effective Exchange Rate in South Africa, 1990-20053

Citation: IMF Staff Country Reports 2006, 328; 10.5089/9781451841046.002.A002

9. Exchange rate volatility in South Africa has been relatively high compared to other countries. Two groups of countries are used as comparators: (i) inflation targeters, and (ii) other emerging markets. South Africa’s real effective exchange rate has been more volatile than the currencies of other inflation-targeting countries, especially from 2000 to 2005. Compared to other major commodity exporters in this group, such as Australia, Chile, or New Zealand, South Africa’s exchange rate volatility also looks quite high. There are, however, other emerging market economies—Brazil, Indonesia, and Turkey, for instance—where exchange rate volatility is higher.

Table II.1.

Volatility of the Real Effective Exchange Rate in Inflation-Targeting Countries

article image
Note: The numbers in parentheses indicate years in which an inflation target was adopted. Volatility is measured as the standard deviation of changes in the log level of the monthly real effective exchange rate for each year.

10. High exchange rate volatility does not seem to have had any systematic effect on trade in South Africa. Figure II.2 shows the volatility of the real effective exchange rate, with total trade defined as the sum of the volumes of exports and imports.4 While a simple data plot does not necessarily reveal a causal relationship between the two variables, no clear decline in trade is observed in the major episodes of heightened volatility in 1998 and 2001.

Figure II.2.
Figure II.2.

REER Volatility and Total Trade

Citation: IMF Staff Country Reports 2006, 328; 10.5089/9781451841046.002.A002

Table II.2.

Volatility of the Real Effective Exchange Rate in Selected Emerging Markets

article image
Note: Volatility is measured as the standard deviation of changes in the log level of the monthly real effective exchange rate for each year. Selected countries all belong to the JP Morgan EMBI. The list does not include Eastern European countries that were in transition throughout most of the 1990s.

C. A Disaggregated View

11. Changes in exchange rate volatility have been negatively correlated with the growth rate of trade flows in South Africa. Figures II.3 and II.4 plot the volatility of the real effective exchange rate with the growth rates of both exports and imports. These figures show that growth in trade flows tends to be lower when exchange rate volatility is higher. This tendency is more pronounced in exports; the correlation between exchange rate volatility and export growth is −0.66, while the correlation for import growth is −0.13. To examine whether this negative relationship between trade growth and exchange rate volatility is statistically significant, we conducted a statistical analysis.

Figure II.3.
Figure II.3.

REER Volatility and Export Growth

Citation: IMF Staff Country Reports 2006, 328; 10.5089/9781451841046.002.A002

Figure II.4.
Figure II.4.

REER Volatility and Import Growth

Citation: IMF Staff Country Reports 2006, 328; 10.5089/9781451841046.002.A002

12. Though a regression analysis that uses disaggregated data confirms the negative relationship between exchange rate volatility and growth in trade flows, it also shows that the effect is not statistically significant.5 Tables II.3 and II.4 summarize the regression results of three different specifications for export and import growth. The dependent variable is percentage changes in exports and imports, and the variables were regressed on the volatility of the real effective exchange rate (see Box II.2 for details on data and methodology).

Table II.3.

Regression Results for Export Growth

article image
Notes: Standard errors are in parentheses. The dependent variable is percentage changes in exports.

indicates statistical signiticance at the 10 percent level.

External demand growth is measured as the weighted average of growth rate of trade partners.
Table II.4.

Regression Results for Import Growth

article image
Notes: Standard errors are in parentheses. The dependent variable is percentage changes in imports.

indicates statistical significance at the 10 percent level.

indicates statistical significance at the 1 percent level.

Domestic demand growth is measured as real GDP growth rate.

13. Specification (1) is a pooled regression. Both export and import equations show a large and negative effect of exchange rate volatility on the growth rates of exports and imports, but the estimates are not statistically significant.

14. Specification (2) introduces additional explanatory variables, namely external demand growth in the export equation and domestic demand growth in the import equation. Coefficients for external and domestic demand are both expected to be positive. In the export equation, external demand growth is negative but insignificant. In the import equation, on the other hand, the coefficient for domestic demand is positive and statistically significant—but the estimate on exchange rate volatility is still insignificant.

15. Specification (3) controls for lagged exchange rate volatility. In the export equation, lagged volatility is positive but not statistically significant. The estimate of exchange rate volatility becomes somewhat smaller but remains insignificant. In the import equation, domestic demand is significant, but exchange rate volatility turns positive with no statistical significance. In general, the relationship between exchange rate volatility and trade growth seems at best unstable.

Data and Methodology

In the regression analysis, exchange rate volatility is measured as the standard deviation of monthly log changes in the real effective exchange rate over one year. The data on the real effective exchange rate are from the South African Reserve Bank. The sample period is 1998 to 2005. Sectoral data on South African exports and imports were obtained from Datastream. After the export and import data are deflated, the real trade growth rates are calculated. Sectors whose share in total trade exceeds 1 percent are included in the regression analysis (see the Data Appendix for the list of sectors in the sample). In the export equation, external demand growth is measured as the weighted average of the real GDP growth rates of all the export partners. In the import equation, domestic demand growth is measured as South African real GDP growth rate. For both equations the data are taken from the World Economic Outlook database.

The regression analysis was conducted using ordinary least squares. Despite the panel structure of our data set, we did not pursue panel data estimation because controlling for sector-specific effects does not affect our estimate of exchange rate volatility. Because exchange rate volatility does not vary by sector, taking account of sector-specific effects will not affect the estimated coefficient for exchange rate volatility even though they may account for different intercepts across sectors. The estimation results of a fixed effects and a random effects model (not reported) indeed show that both produce results identical to those from specification (1) in Table II.3 and II.4. Since we are interested in the estimate of coefficient for exchange rate volatility, we adopt alternative specifications, as in the tables, not the panel specifications.

16. Evidence from disaggregated sectoral data seems to confirm that there is no significant linkage between exchange rate volatility and trade flows in South Africa. We have plotted exchange rate volatility and trade growth for major export and import sectors (not reported here) to further investigate how exchange rate volatility affects trade flows at the disaggregated sector level.6 The simple sector data plots are consistent with the findings from the regression analysis. The plots suggest that the two variables appear to have a loose negative correlation, but it is difficult to detect a systematic pattern.

D. Conclusion

17. During several episodes of high exchange rate volatility over the past decade, South Africa’s exchange rate has been more volatile than that of other economies. That has raised concerns about possible adverse effects on growth. While exchange rate volatility may affect growth in a number of ways, this section looked at the possible link through trade. We found no clear linkage in the South African data between exchange rate volatility and trade. An implication of this finding is that there is no strong rationale for taking policy measures to restrict exchange rate movements in an attempt to promote trade flows. Exchange rate volatility in itself is not necessarily harmful to an economy. On the contrary, exchange rate movements can help the economy adjust more smoothly to significant terms of trade shocks. As far as trade is concerned, it is important to take this general equilibrium consideration into account in assessing the impact of exchange rate volatility.

Appendix

Sectors included in export growth regressions:

Live animals and animal products

Base metals and base metal articles

Chemical and allied industries products

Machinery, mechanical, and electrical appliances

Mineral products

Plastic and rubber articles

Prepared foodstuffs, beverages, and tobacco

Semiprecious stones and jewelry

Textiles and textile products

Vegetable products

Vehicles and other means of transport

Wood and wood products

Wood pulp, paper, and paper products

Sectors included in import growth regressions:

Base metals and base metal articles

Chemical and allied industry products

Machinery and mechanical and electrical appliances

Mineral products

Optical and precision instruments, watches and clocks

Plastic and rubber articles

Prepared foodstuffs, beverages, and tobacco

Semiprecious stones and jewelry

Stone, plaster, cement, ceramics, and glass products

Textiles and textile products

Vegetable products

Vehicles and other means of transport

Wood pulp, paper, and paper products

References

  • Anderson, J., 1979, “A Theoretical Foundation for the Gravity Equation,” American Economic Review, Vol.69 (1), pp.106-16.

  • Anderson, J., and E. van Wincoop, 2003, “Gravity with Gravitas: A Solution to the Border Puzzle,” American Economic Review, Vol.93 (3), pp.170-92.

    • Search Google Scholar
    • Export Citation
  • Broda, C., 2001, “Coping with Terms-of-Trade Shocks: Pegs Versus Floats,” American Economic Review, Vol. 91 (2), pp.376-80.

  • Clark, P., N. Tamirisa, A. Sadikov, S. Wei, and L. Zeng, 2004, “A New Look at Exchange Rate Volatility and Trade Flows,” Occasional Paper 235 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Deardorff, A., 1998, “Determinants of Bilateral Trade: Does Gravity Work in Neoclassical World?” in The Regionalization of the World Economy, ed. by J. Frankel (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation
  • Grossman, G., and E. Helpman, 1991, Innovation and Growth in the Global Economy (Cambridge, Massachusetts, and London: MIT Press).

  • Jonsson, G., and A. Subramanian, 2000, “Dynamic Gains from Trade: Evidence from South Africa,” Working Paper 00/45 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Kiyota, Kozo, and Shujiro Urata, 2004, “Exchange Rate, Exchange Rate Volatility and Foreign Direct Investment,” World Economy, Vol.27 (10), pp.1501-36.

    • Search Google Scholar
    • Export Citation
  • Rose, A., 2004, “Do We Really Know that the WTO Increases Trade?” American Economic Review, Vol.94 (1), pp.98-114.

1

Prepared by Thomas Harjes and Kazuko Shirono.

2

For instance, exchange rate volatility may affect growth through its impact on foreign direct investment, but empirical evidence is mixed. See Kiyota and Urata (2004).

3

Volatility is measured as the standard deviation of monthly changes in the real effective exchange rate. Colored columns show exchange rate volatility calculated over one year, and white columns show volatility for overlapping two-year rolling periods (t and t+1).

4

Data on trade, import index, and export index were obtained from Datastream. Trade data are in the local currency and deflated by export/import deflators.

5

The regression analysis is based on annual exports and imports data from 1998 to 2005 of thirteen sectors (see Appendix).

6

Since there are only seven data points at the sectoral level, we did not pursue a regression approach.

South Africa: Selected Issues
Author: International Monetary Fund