This paper applies a factor‐augmented Markov‐switching model to the South African economy to provide an alternative classification of the business cycle and its turning points. In the principal components step, 123 variables are used to establish the aggregate cyclicality in all sectors of the economy with the number of factors chosen using a modified Bai and Ng method. By exploiting the rich nature of the dataset, we provide a model with well‐defined statistical properties that compares favourably with the South African Reserve Bank (SARB) dating points. Combining the results of the parametric approach followed in the Markov‐switching model and the non‐parametric approach followed by the SARB should allow for a robust turning point analysis. A Markov‐switching model of real gross domestic product is also estimated because this variable is commonly used in the literature and provides a benchmark for the factor models.

This paper constructs a number of possible core measures of inflation using singular spectrum analysis (SSA). Annual changes in monthly inflation are decomposed into its trend, oscillatory and noise components in order to develop an understanding of the trend and cyclicality in South African headline inflation. Three cyclical components with differing amplitude and frequency are identified. The trend and cyclical components of inflation are found to be a good approximation of core inflation, the inertial part of inflation. These core measures are compared with other candidate core measures based on the properties of a good core inflation measure. Generally, the SSA measures outperform commonly used measures of core inflation based on both in-and out-of-sample performance. JEL classification: C41, C14, E31, E37, N17

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

No abstract

Forecasting and estimating core inflation has recently gained attention, especially for inflation targeting countries, following research showing that targeting headline inflation may not be optimal; a Central Bank can miss the signal due to the noise. Despite its importance there is sparse literature on estimating and forecasting core inflation in South Africa, with the focus still on measuring core inflation. This paper emphasises predicting core inflation using large time-varying parameter vector autoregressive models (TVP-VARs), factor augmented VAR, and structural break models using quarterly data from 1981Q1 to 2013Q4. We use mean squared forecast errors (MSFE) and predictive likelihoods to evaluate the forecasts. In general, we find that (i) small TVP-VARs consistently outperform all other models; (ii) models where the errors are heteroscedastic do better than models with homoscedastic errors; (iii) models assuming that the forgetting factor remains 0.99 throughout the forecast period outperforms models that allow for the forgetting factors to change with time; and (iv) allowing for structural break does not improve the predictability of core inflation. Overall, our results imply that additional information on the growth rate of the economy and interest rate is sufficient to forecast core inflation accurately, but the relationship between these three variables needs to be modelled in a time-varying (nonlinear) fashion.JEL Classification: C22, C32, E27, E31.