Modelling the capital structure of manufacturing, mining and retail firms listed on the Johannesburg Stock Exchange

Abstract

This thesis examines three aspects of capital structure of manufacturing, mining and retail firms listed on Johannesburg Securities Exchange (JSE). Firstly, it tests for the validity of the pecking order, the static trade-off and the dynamic trade-off theories in the context of South African manufacturing, mining and retail firms. The study used data from 42 manufacturing, 24 mining and 21 retail firms with complete data for four or more consecutive years during 2000-2010 (panel 1) to test the validity of these theories. The research hypotheses were formulated and tested using generalised least squares (GLS) random effects, maximum likelihood (ML) random effects, fixed effects, Prais-Winsten regression, Arellano and Bond, Blundell and Bond and the random effects Tobit models. Secondly, the thesis examines the impact of the firm’s key financial performance variables on firm leverage and speed of target adjustment. A panel of 49 manufacturing, 24 mining and 23 retail firms with complete data for two or more consecutive years during the period 2005-2010 (panel 2) was constructed and used in this test. The research hypotheses were formulated and tested using the same regression models used in panel 1. Lastly, the thesis examines the existence of the discounted value premium in manufacturing, mining and retail firms listed on the JSE. This study was done using panel of 47 manufacturing, 31 mining and 20 retail firms with complete data for four or more consecutive years during the period 2006-2010. A simple t-test was used to evaluate the significance of the sample’s discounted value premium. The study documents that firm growth rate, non-debt tax shields, financial distress, profitability, capital expenditure, asset tangibility, price earnings, ordinary share prices and changes in working capital were significant predictors of firm leverage. Dividend paid, capital expenditure, firm growth rate, profitability, cash flow from operations and economic value added were positively correlated to leverage. Asset tangibility, firm profitability, non-debt tax shields, financial distress, liquidity, price earnings, share price and retention rate were negatively correlated to leverage. Asset tangibility, financial distress, firm growth, non-debt tax shields, and long-term debt repaid were negatively correlated to changes in debt issued, whilst profitability, actual dividend paid, capital expenditure and changes in working capital were positively correlated. These results confirm the complementary nature of the trade-off and pecking order theories. Furthermore, the firms had positive and significant speeds of adjustment. In panel 1, the true speed of adjustment for the sample was 57.64% (0.81 years) for book-to-debt ratio (BDR) and 42.44% (1.25 years) for market-to-debt (MDR). The speed for manufacturing firms was 45.08% (1.16 years) for BDR and 44.59% (1.17 years) for MDR; for mining firms, 72.07% (0.54 years) for BDR and 56.45% (0.83 years) for MDR; and for retail firms, 28.42% (2.07 years) for BDR and 42.48% (1.25 years) for MDR. In panel 2, the true speed of adjustment for the sample was 64.20% for book-to-debt ratio (BDR) and 28.11% for market-to-debt ratio (MDR). The true speed for manufacturing firms was 34.42% for BDR and 30.56% for MDR; for mining firms, 69.59% for BDR and 45.77% for MDR; and for retail firms, 9.34% for BDR. These results confirm the validity of the dynamic trade-off theory. Finally, manufacturing, mining and retail firms had a positive discounted value premium. This ranged from 5.16% to 9.48% (on perpetual growth), with mining firms having the largest (9.48%), followed by manufacturing (8.54%) and retail firms (5.16%). Of the observations for the full sample, 92.23% showed a positive discounted value premium. This evidence on the speed of adjustment and discounted value premium suggests the existence of a target capital structure different from the theoretical optimal capital structure hypothesised by the static trade-off theory.