Forecasting the Probability of Recessions in South Africa: the Role of Decomposed Term Spread and Economic Policy Uncertainty

Abstract

This paper decomposes the term spread into the expectation and the term premium components using a fractional integration approach and subsequently uses same with the economic policy uncertainty index to forecast the probability of recession in South Africa. We use different specifications of the probit model and quarterly data from 1990:1 to 2012:1. Our out-of-sample results show that the model that incorporates the expectation component and economic policy uncertainty provides the best forecast of recession. All three recession periods in our sample were accurately dictated by the prediction models and the best forecast occurred at the four quarters ahead horizon. A robustness check with a longer sample from 1946q1 to 2017q4 but excluding the factors and economic policy uncertainty due to data limitation provided justification for decomposing the term spread as the model with the expected spread turned out to be the best. We draw the implications of these findings.

The effect of uncertainty on the information content of term spread and its components

“…Rosenberg and Maurer (2008) provide empirical evidence that the risk-neutral expectation is a leading indicator of recession, while the term premium is not. In addition, Aye et al (2019) find that the probit model incorporating the risk-neutral expectation and economic policy uncertainty (EPU) is the best forecasting model for recessions of South Africa in out-of-sample analysis. On the other hand, Hamilton and Kim (2002) show that the term premium and the risk-neutral expectation are relevant for predicting real GDP growth, but their respective contributions Information content differ depending on the predictive horizon.…”

Oil-price uncertainty and the U.K. unemployment rate: A forecasting experiment with random forests using 150 years of data

“…Because existence of in-sample impact (predictability) does not necessarily ensure out-of-sample forecasting gains (Rapach and Zhou 2013), our paper aims to provide a robust extension of the literature on oil-price uncertainty and its impact on economic activity by conducting an out-of-sample forecasting analysis of the predictive value of oil-price volatility for changes in the unemployment rate of the United Kingdom (UK) over the historical monthly period from 1859:10 to 2020:05. While our focus is oil-price uncertainty, to prevent omitted-variable bias, we incorporate a host of other macroeconomic and financial uncertainties (as well as the first-moments of these variables), the importance of which in forecasting macroeconomic aggregates has gained tremendous prominence in the wake of the "Great Recession" and the Global Financial Crisis (GFC) that followed thereafter (see for example, Karnizova and Li 2014, Balcilar et al, 2016, Junttila and Vataja 2018, Aye et al 2019a, 2019b, Pierdzioch and Gupta 2020. Note that the choice of the UK as our case study is driven by the availability of a long span of data involving economic activity (as captured by the unemployment rate in our case) and a wide-array of predictors.…”

Uncertainty and forecastability of regional output growth in the UK: Evidence from machine learning

“…Although the literature dealing with the influence of uncertainty on output primarily relies on in-samplebased structural analyses, more recently, quite a few studies (see e.g., Aye et al, 2019aAye et al, , 2019bBalcilar et al, 2016;Gupta et al, forthcoming;Junttila & Vataja, 2018;Karnizova & Li, 2014;Pierdzioch & Gupta, 2020;Salisu et al, 2022;Segnon et al, 2018) have also analyzed the role of uncertainty in forecasting economic activity (output growth and recessions) in out-of-sample analyses. This is an important line of research, because policymakers in general, and central banks in particular, would need accurate predictions of the future path of the economy following periods of heightened uncertainty while making their policy decisions.…”