An Assessment Of The Costs, Benefits And Optimality Of The Common Monetary Area In Southern Africa

Abstract

In this paper we aim to answer the following two questions: 1) has the Common Monetary Area in Southern Africa (CMA) been an optimal currency area (OCA)Π2) What are the costs and benefits of the CMA for its participating countriesΠTo answer these questions, we carry out a two-step econometric exercise based on the theory of generalised purchasing power parity (G-PPP). The first exercise tests whether the CMA (but also Botswana as a de facto member) forms an OCA by detecting the existence of common long-run trends in their bilateral real exchange rates. The second exercise identifies the determinants of deviations from G-PPP in the CMA as measured through the degree of pairwise price correlation. These determinants stand for the costs and benefits of joining a monetary union, namely the degree of trade openness, the degree of production diversification (and the interaction between these two), the synchronicity of business cycles and the kind of gross capital inflows received by the member countries.

Data sources

Abbreviations: AFDI (African Development Indicators, from World Bank), CBB (Central Bank of Botswana), CBN (Central Bank of Namibia), CSO (Central Statistical Office, Botswana), DS (Datastream), IFS (International Financial Statistics from IMF), Statass (Statistical Agency South Africa), WBDI (World Bank Development Indicators)

Table

Notes

1 This literature points out that a monetary union can become an optimal one without complying ex-ante with the traditional OCA criteria. The endogeneity approach postulates that increasing monetary integration may drive more similar trade specialization patterns, hence more correlated GDP cycles and higher trade openness.

2 During the 1960's those countries became independent and started running their own monetary institutions around South Africa's.

3 In fact, Swaziland, Lesotho and Namibia introduced their own national currencies after becoming independent states (the lilangeni, the loti, dollar and in 1974, 1980 and 1993, respectively), but their exchange rates have remained fixed at parity with the Rand.

4 Only Swaziland, a very small country, does export significantly to SACU countries.

5 The US dollar real exchange rate may be a misleading measure of that kind, for a simple reason: even assuming the shares of exports to the US equalises across CMA countries (including Botswana), the import sides are clearly unequal. While South Africa imports very little from its partners, the latter import near 90% of their total imports from South Africa. Given the high tradability of the consumer price indices in all these countries, any inflation surge in South Africa derived from a depreciation of the rand against the U$S will be almost fully transferred to their own inflation. Despite having brought inflation down, South Africa-US inflation differentials have not been of minor importance over the last decade (1990's).

6 We are not able to disentangle, however, the part that is due to common policy responses, often seen in sub-regional integration processes where macroeconomic coordination is carried out.

7 We obtain 2 cointegrating vectors. See econometric appendix

8 In other words, we want to identify the sources of strain that would ultimately push bilateral real parities away from its PPP levels, i.e. when inflation differentials widen or other factors are in operation. Therefore, it would be redundant to include the inflation differential on the RHS of the equation when our dependent variable is already a measure related to it.

9 For example, CORRELCPI between Namibia and South Africa in 1996 results from calculating the linear correlation coefficient spanning January to December observations (12) for that year.

10 Good surveys of the OCA literature can be found in De Grauwe (2000), Laffrance and St-Amant (1999) or Kenen (2000), among others. See the introductory section.

11 Corden (1972) argues the openness criterion applies only to microeconomic demand changes in the domestic economy and does not apply to macroeconomic disturbances that occur abroad. He argues to the extent the latter has been the primary cause of payments disequilibrium then the economy should be insulated by flexible exchange rates (specially so a large economy).

12 As we mentioned above, the bilateral trade intensities in the CMA are very low and don't increase significantly over time. Therefore, we took the total trade flows over GDP as an indicator of the degree of openness. Curiously, the relevant trade flows for member countries are other than intra-zone exchanges, with the exception of the imports from South Africa, which account up to a 90% in some cases. However, the weights the export markets had in each country showed a relatively similar structure. In other words, the optimal trade-weights that would stabilise the REER for SACU as a whole should be (calculations made from IMF DOTS, 2000): 41.7% in European currencies (predominantly the euro and sterling), 11% in US dollar, 6.4% in Japanese yens and the remaining 40% in diverse Asian and African currencies.

13 When yields a negative value (i.e. the denominator is greater than the numerator) we take its absolute value. Ultimately, we are interested in relative GDP variations away from 0 irrespective of its sense.