This study investigates the relevant factors that drive house prices in South Africa with the aim of facilitating a better understanding of the dynamic relationship between house prices and key macroeconomic variables. This can serve as a prerequisite to the ability of policymakers to maximize the positive externalities associated with housing development, while implementing measures to reduce the unexpected effects.
The thesis consists of five independent papers corresponding to five chapters. The first chapter examines the economic sources underlying the comovement of real house prices across provinces in South Africa. First, a dynamic factor model is estimated on quarterly provincial-level data to disentangle the national component of real house price movements from the local (provincial or region-specific) component. Second, a Structural Vector Autoregressive (SVAR) model is applied to investigate the extent to which macroeconomic shocks are responsible to the common component of real house prices. Using theoretically motivated short run restrictions to identify macroeconomic with portfolio and monetary policy shocks playing greater roles. We also find evidence shocks, results indicate that comovement in real house prices is due to the combined effects of favourable and unfavourable structural shocks emanating from different sectors of the economy, of significant feedbacks from the housing sector to the real economy which theoretically channel through the wealth and /or collateral and balance sheet effects on consumption and investment, respectively.
The second chapter implements a Panel Vector Autoregression (PVAR) approach on provincial level data to analyse the role of house prices in determining the dynamic behaviour of consumption. Unlike individual regression, this approach accounts for individual heterogeneities characteristic of provincial housing markets. Based on the standard recursive identification, we find that house prices exhibit an asymmetric effect on consumption: a positive shock to house price growth has a positive and significant effect on consumption, while the negative impact of an anticipated house price causes an insignificant reduction in consumption.
Because consumption is a significant component of Gross Domestic Product (GDP), the effect of house prices on consumption serves as a key link between the housing market and economic activity. The third chapter, therefore, exploits panel time series methods to examine the impact of house price changes on economic growth across provinces. This framework offers a variety of tools designed to address econometric issues such as heterogeneity, endogeneity and spatial effects which have been found to be prominent in regional housing markets. Specifically, Fixed effect (FE) and Random coefficient (RC) models are used to address the issue of heterogeneity. The potential endogeneity is accounted for using SYSTEM-Generalised Method of Moments (SYS-GMM) while the Feasible Generalised Least Squared (FGLS) and the Seemingly Unrelated Regression (SUR) are used to control for spatial effects. Accounting for these above issues leads to a significant effect of house price changes on provincial economic growth in South Africa.
Since house prices affect the business cycle, monetary policy might not be neutral to house price movements. Moreover, one may expect asymmetric response of monetary policy to house price shocks giving the boom/bust nature of house price dynamics. In light of these considerations, the fourth chapter links South African housing market dynamics to the interest rate setting behaviour by relying on Markov-Switching Vector Autoregressive (MS-VAR). This technique allows identifying the bull and bear regimes in the South African housing market and therefore helps examining asymmetries in the impact of monetary policy shocks on the house prices during bull and bear regimes. The impact of the monetary policy on house prices is found to be larger in the bear regime than in the bull regime; indicating the role of information asymmetry in reinforcing the financial constraint of economic agent. Unsurprisingly, monetary reaction to a positive house price shock is found to be stronger in the bull regime. This suggests that central banker are more concerned in bull regime given the potential crisis related to the subsequent bust in house prices bubbles which are more prominent in bull markets.
Finally, changes in house prices induce an adjustment of consumption and investment decisions which could be reflected as a trade surplus or deficit. The fifth chapter characterises the dynamic relationship between house prices and the trade balance based on a Bayesian Vector Autoregressive (BVAR) approach with sign restrictions. The results indicate that 1 percent decline in house prices can improve the trade balance by 0.2 percent; suggesting that house prices represent an additional instrument for trade-balance adjustment besides the traditional exchange rate channel. Further, we find that the contribution of house price shocks to the historical path of the trade balance is less prominent in 2000s; possibly substantiating the effectiveness in the conduct of South African monetary policy, which has been shown to be incorporating house price movements in its interest rate setting behaviour.