Essays in Macroeconomics: An Analysis of the South African Financial System

Abstract

This thesis consists of three empirical essays analyzing the financial systems in South Africa. Chapter 1 constructs a financial conditions index (FCI) for South Africa, over the period January 2000 to April 2017, that can be used to predict risks in the financial market emanating from both the domestic market and the global market. It uses a time-varying parameter factor-augmented vector autoregressive (TVP-FAVAR) model, proposed by Koop and Korobilis (2014) that allows the FCI to vary with time. Further, the results show that the TVP-FAVAR model outperforms the constant-loadings factor-augmented vector autoregressive (FAVAR) model and the traditional vector autoregressive (VAR) model in the out-of-sample forecasting of the inflation rate and the real gross domestic product (GDP) growth rate. The FCI leads the measure of GDP indicating that the FCI can be used as an early warning indicator of a financial crisis. The chapter shows that tighter financial conditions contract the real economy and are deflationary at the same time.

Chapter 2 analyzes how the relationship between bank capital and liquidity affects bank lending. It uses a non-linear fixed affected model that interacts banks' capital ratios with their liquidity levels, proposed by Kim and Sohn (2017), on 25 South African banks over the period January 2008 to July 2020. In addition, it runs the marginal effects of bank capital on lending at the mean of liquidity to interpret the interaction between bank capital ratios and liquidity levels. The results indicate that, at average liquidity levels, banks with higher capital ratios increase the volume of commercial loans and decrease the amount of retail loans, implying that banks prefer to lend to firms than to households. As banks strengthen their solvency position, they are better situated to increase the capacity of their "riskier" loan portfolio by raising their volume of enterprise loans. This then comes with the consequence of reducing household credit.

Chapter 3 examines the bank-sovereign nexus in South Africa. It uses a Markov regime-switching model over the period January 2008 to October 2020 to analyze how changes in bank exposure to the sovereign impact the sovereigns' risk position and how financial institutions' increased exposure to the sovereign impacts bank lending rates and volumes. The results indicate that during periods of economic stress, as banks increase their holding of government debt, the domestic sovereign finds it too costly to default as a default heightens the chances of a banking sector crisis. It also finds that higher bank exposure to the sovereign is linked with higher bank lending rates and lower loan volumes, again during periods of economic distress. Higher sovereign debt exposures tighten banks' capital constraints which impairs their lending abilities.