Vulnerability to currency crises: a contrast between south Africa and Zimbabwe in the 1990s.

Crosoer, David

Vulnerability to currency crises: a contrast between south Africa and Zimbabwe in the 1990s.

Master Thesis

2000

Abstract

This paper conducts a review of theoretical models of currency crises, distinguishing between first- and second-generation models. First generation models focus on the role of unsustainable government policy in precipitating crisis, whereas second generation models are more concerned with the credibility problem facing government could how private agents can push government into crisis. Second generation models also imply currency crises can be unpredictable. Despite this, within the literature on currency crises, two separate methodologies exist which might serve as 'early warning' based systems. The first methodology, the multivariate approach, assesses a country's 'vulnerability' to (say) a global shock and hence the conditional likelihood of currency attack. The second methodology, the signalling approach, argues that extreme behaviour of certain indicators is suggestive of an impending crisis. The multivariate approach is applied to South Africa and Zimbabwe during the Asian crisis of 1997. Their vulnerability is assessed in terms of a simple second-generation model developed by Sachs et al (1996). It is argued, however, that because Sacetal's (1996) model conceptualises the countries' vulnerability in terms of the 1994 lvfexican crisis, it works poorly for a subsequent crisis. The ex-post nature of much of the literature is stressed. The signalling approach is applied to South Africa and Zimbabwe by monitoring the behaviour of a number of indicators Kaminsky et al (1997) suggest as useful for anticipating currency crises. It is found that these indicators in the 1990s are not only crisis sensitive but can exhibit extreme behaviour in response to a number of events. Again, it is often only ex post that one can claim the extreme behaviour an indicator exhibits is indicative of an impending crisis. The paper argues that Zimbabwe has been subjec1 to a number of first-generation type crises largely precipitated by unsustainable government policies. It also suggests that more sophisticated second-generation models should be applied to South Africa, where an acute trade-off appears to exist between disciplined policies and deteriorating external conditions. South Africa's disciplined policies seem to have allowed it to weather the Asian crisis better than Zimbabwe. The impact of the Mexican and Asian crises for South Africa, however, is slight when compared with the 1996 land 1998 crises it experienced. The effect of the Mexican and Asian crises on Zimbabwe is also muted. Both South Africa and Zimbabwe appear to be more affected by 'local' rather than 'global' crises. Neither is simply victim of indiscriminate capital flows. The paper concludes by arguing that the literature on currency crises is generally deficient in its ability to predict crisis. The multivariate approach makes no attempt to predict the occurrence of a global shock, and furthermore is often ex post in its insights. The indicator approach is a-theoretical, and this paper argues unsuited to anticipate second generation crises of the kind that affected

Keywords

Social Sciences

Reference:

Collections

Masters

Full item page