An analysis of recent global economic development and GDP growth using Stein's Paradox, and South Africa's monetary and fiscal policy response.

Abstract

The economic crisis of 2007 has had debilitating effects on the global economy, affecting GDP growth, unemployment and trade to name a few. In response to these economic effects, numerous policy interventions were implemented. There are various existing time-series methods available to determine better estimates of GDP growth rates, one of which is Stein’s Paradox which uses observed averages to estimate unobservable quantities which are closer to the true unknown GDP growth rates or theta (θ) in order to determine better growth rates post the economic crisis. The resulting James-Stein estimator (z) is said to be better than the arithmetic average, and thus a closer approximation to the true GDP growth rates which are unobservable. This dissertation analyses the effects of the 2008 financial crisis on the global economy, with specific reference to South Africa and America, and their corresponding policy interventions to determine the growth trajectory after the crisis. The main objective is to determine if better estimates of GDP growth can be calculated using Stein’s Paradox, across a sample of 30 countries, using quarterly GDP growth for the period 2005 to 2008. Annual GDP data was also used for the period 2009-2011, and future GDP growth rates were forecasted for the period 2012 to 2016. To reinforce the Stein’s Paradox, the Monte Carlo study is undertaken. It is used to determine how the James-Stein estimates perform under different conditions using a common c or unique c, and to determine which condition will provide more accurate GDP growth rates (Muthen. 2002). Analysis of time series data across a sample of 30 countries using Stein’s Paradox provided better estimates of GDP growth rates than the individual average growth rates for each country based on the lower standard deviation and total squared error of estimation achieved. This shows that the results are closer to theta and have a smaller amount of error, particularly when a common c was used. The Monte Carlo results indicate that better GDP growth rates are achieved when using a common c instead of a unique c given that a smaller standard deviation and variance is derived. Therefore the Monte Carlo study aims to reinforce or verify Stein’s Paradox. The study also indicates that emerging and developing countries seem to be the driving forces of growth in the future, while developed countries seem to be lagging behind.