African Journals Online
Studies in Economics and Econometrics

Volume 26, Issue 2, 2002
Abstracts

Labour costs and inflation in South Africa : an econometric study
Akinboade, O.A.Niedermeier, E.W. 1-17

Abstract: The article develops a model which relates the labour market to domestic inflation in South Africa. The study includes a Granger causality test used to determine the direction of causality between labour costs and the consumer price index. Past wage growth in excess of inflation is positively correlated with prices in the short run. In the long run, rising labour costs have contributed significantly to inflation in South Africa. The predominant source of variation in domestic inflation's forecast errors is own shocks. Other sources include innovations from labour costs and the nominal effective exchange rate. 

Modelling the demand behaviour of Spanish consumers using parametric and non-parametric approaches
Molina, J.A. 19-36

Abstract: This article models the demand behaviour of consumers using the parametric and non-parametric approaches. To that end, we consider Spanish time-series from 1964 to 1995 relative to the quantities and prices of five aggregated consumption goods, namely Food, Clothing, Energy, Transport and Miscellaneous Goods. With respect to the non-parametric results, we find that there is a stable demand system underlying the personal preferences structure which explains the observed quantities of goods and is equivalent to the existence of a well-behaved utility function. The parametric results show that the homogeneous and symmetric version of the Rotterdam model satisfies the econometric and theoretical requirements, and hence can properly be used for modelling the demand behaviour of Spanish consumers from 1964 to 1995. 


Axiomatic analysis of choosing the second best
Lahiri, S. 37-50

Abstract: Three new axiomatic characterizations of the decision rule which invariably selects the second best alternative are provided. Subsequently it is shown that apparently non-optimal decisions can always be rationalized by embedding the set of alternatives in an augmented space. 


Business confidence and the South African business cycle
Pellissier, G.M. 51-67

Abstract: This article endeavours to compare the two business confidence indicators in South Africa (that of the BER and SACOB) in terms of their respective relationships as business cycle indicators. Although bearing the same name of business confidence indicator, totally different economic variables are measured in compiling the indices. The BER's business confidence indicator is a psychological measurement of the level of satisfaction with prevailing business condition, while SACOB's business confidence indicator is basically a composite coincident/leading indicator of the business cycle and by implication an indicator of business confidence. Both business confidence indicators correlate highly between themselves and show signs of having leading indicator capabilities of the South African business cycle. However, the leading indicator characteristics of both indices are not better than the official leading indicator and are also of a declining order. In terms of causality, both the indicators seem to be moving towards a coincident relationship rather than a leading one with the business cycle. Only the BER business confidence indicator displays comparable cyclical turning point attributes. 


On the market risk premium
Firer, C.Bradfield, D. 69-80

Abstract: A perspective is given on the dynamic nature, reliability, and the estimation of the market risk premium, as well as some implications concerning its current level. The analysis is based on a data set spanning some 76 years. An historical 'best estimate' of 7,5 percent is suggested for practical use. Furthermore, graphical insights on the dynamic nature of the market risk premium using a rolling estimation approach, reveal a slow decline in the market risk premium. 


Modelling exchange rate variations using principal components analysis : a note
Biekpe, N. 81-85

Abstract: This paper examines cross-country currency variations using principal components and a dynamic linear model (DLM). A combination of the principal components analysis and DLM (which is time-dependent) ensures that the variation explained by the analysis is non-static. Normally, ordinary principal components calculations produce static values that are not time-dependant. However, the dependence of spot and forward rates on time requires that a time-dependent modeling approach be adapted. The main argument in this analysis is that if countries have similar growth trends, then they are more likely to share common dominant growth factors. These factors could include, for instance, common inflation or currency risks. If this risk is significant enough then principal components analysis together with an appropriate DLM should capture it.