Josine Uwilingiye

Out-of-Sample Equity Premium Predictability in South Africa: Evidence from a Large Number of Predictors

ABSTRACT This article uses a predictive regression framework to examine the out-of-sample predictability of South Africa’s equity premium, using a host of financial and macroeconomic variables. We employ various methods of forecast combination, bootstrap aggregation (bagging), diffusion index (principal component), and Bayesian regressions to allow for a simultaneous role of the variables under consideration, besides individual predictive regressions. We assess both the statistical and economic significance of the individual predictive regressions, combination methods, bagging, principal components, and Bayesian regressions. Our results show that forecast combination methods and principal component regressions improve the predictability of the equity premium relative to the benchmark autoregressive model of order one (AR[1]). However, the Bayesian predictive regressions are found to be the standout performers with the models outperforming the individual regressions, forecast combination methods, bagging and principal component regressions, both in terms of statistical (forecasting) and economic (utility) gains.

Comparing South African Inflation Volatility Across Monetary Policy Regimes: An Application of Saphe Cracking

Recent empirical evidence on the direct link of inflation targeting and inflation volatility is at best mixed. However, comparing inflation volatility across alternative monetary policy regimes within a country based on conventional ways, used in previous studies, begs the question. The question is not whether the volatility of inflation has changed, but rather whether the volatility is different than it otherwise would have been. In such a backdrop, this paper uses the cosine-squared cepstrum to provide evidence that CPI inflation in South Africa has become more volatile since the first quarter of 2000, when the country moved into an inflation targeting regime, than it would have been had the South African Reserve Bank (SARB) continued with the more eclectic monetary policy approach pursued in the pre-targeting era.

Dynamic Relationship between Oil Price and Inflation in South Africa

The oil price-inflation relationship has been at the center of attention among economists and policy analysts, especially after 1970s oil shocks that resulted in a significant increase in the rate of inflation in number of countries around the world. However in the recent years, a number of empirical study, mostly in developed economies, has found that the effect of oil price shocks on inflation has weakened; mainly due to a reduction in oil intensity in production process and lower inflation environment. Moreover, some empirical evidences have shown the response of inflation from a negative and a positive oil shock to differ. This study aims to investigate the evolving relationship between oil price and inflation in South Africa, using time series data of inflation and oil price starting from January, 1922 to July, 2013. The study has a policy relevance on monetary policy reaction to oil shocks as South Africa is a small open economy with higher dependency on oil import and a floating exchange rate system. We fit both symmetric and asymmetric dynamic conditional correlation GARCH (DCC-GARCH) to the data. The results reveal the oil price to have a positive relationship with inflation, however the correlation is low and ranges between 0.07-0.08. The time-series patterns show a tendency of temporary upward shift in the pair-wise conditional correlations during predominant oil crisis. We also observe an upward shift in correlation in the year 1986, which can be attributed to South Africas oil embargo. Further, the asymmetric-DCC model, based on the DCC exponential GARCH (DCC-EGARCH) framework, which fits the data better than the symmetric DCC-GARCH, show that the positive shocks have higher effect on inflation than negative shocks of the same magnitude. Finally, we observe that the correlation has been decreasing gradually over time for both symmetric and asymmetric specification of DCC model. The weaker oil price-inflation relationship, observed in recent years, could be attributed to the South African Reserve Banks commitment to stabilize inflation expectation in the presence of external shocks. The study highlights the relative importance of oil shocks on inflation, and recommends that the Reserve Bank needs to remain attentive to oil price shocks, especially the positive ones.

The impact of oil price on South African GDP growth: a Bayesian Markov switching-VAR analysis

One characteristic of many macroeconomic and financial time series is their asymmetric behaviour during different phases of a business cycle. Oil price shocks have been amongst those economic variables that have been identified in theoretical and empirical literature to predict the phases of business cycles. However, the role of oil price shocks to determine business cycle fluctuations has received less attention in emerging and developing economies. The aim of this study is to investigate the role of oil price shocks in predicting the phases of the South African business cycle associated with higher and lower growth regimes. By adopting a regime dependent analysis, we investigate the impact of oil price shocks under two phases of the business cycle, namely high and low growth regimes. As a net importer of oil, South Africa is expected to be vulnerable to oil price shocks irrespective of the phase of the business cycle. Using a Bayesian Markov switching vector autoregressive (MS-VAR) model and data for the period 1960Q2 to 2013Q3, we found the oil price to have predictive content for real output growth under the low growth regime. The results also show the low growth state to be shorter-lived compared to the higher growth state. against standard forecasting models. U.S. inflation forecasts improve when controlling for persistence and economic policy uncertainty (EPU). Importantly, the VARFIMA model, comprising of inflation and EPU, outperforms commonly used inflation forecast models.

Using large data sets to forecast sectoral employment

We use several models using classical and Bayesian methods to forecast employment for eight sectors of the US economy. In addition to using standard vector-autoregressive and Bayesian vector autoregressive models, we also augment these models to include the information content of 143 additional monthly series in some models. Several approaches exist for incorporating information from a large number of series. We consider two multivariate approaches—extracting common factors (principal components) and Bayesian shrinkage. After extracting the common factors, we use Bayesian factor-augmented vector autoregressive and vector error-correction models, as well as Bayesian shrinkage in a large-scale Bayesian vector autoregressive models. For an in-sample period of January 1972 to December 1989 and an out-of-sample period of January 1990 to March 2010, we compare the forecast performance of the alternative models. More specifically, we perform ex-post and ex-ante out-of-sample forecasts from January 1990 through March 2009 and from April 2009 through March 2010, respectively. We find that factor augmented models, especially error-correction versions, generally prove the best in out-of-sample forecast performance, implying that in addition to macroeconomic variables, incorporating long-run relationships along with short-run dynamics play an important role in forecasting employment. Forecast combination models, however, based on the simple average forecasts of the various models used, outperform the best performing individual models for six of the eight sectoral employment series.

Temporal causality between budget deficit and interest rate : the case of South Africa

This paper investigates the direction of temporal causality between budget deficit and interest rate in South Africa using quarterly data for the period of 1961:02 to 2005:04, and also for annual data covering 1961 to 2005. Based on a multivariate Vector Error Correction Model (VECM), estimated using Johansen’s (1991, 1995) Maximum Likelihood Approach, we find that budget deficit Granger causes interest rate in the quarterly data. However, for the annual data, no causal relationship could be detected between the budget deficit and the Treasury bill rate. The two variables of interest are, however, positively cointegrated for both data frequency. Interestingly though, exactly the same results were obtained from the simple Granger causality tests based on a bivariate framework, comprising merely of budget deficit and interest rate.