Finance-led growth discourses consider the availability of credit an important driver of economic growth. With the hegemony of neoliberal economic institutions, wage-led growth has become a distant reality that was only true during the heydays of Keynesianism (Skidelsky and Craig 2016; Stockhammer 2008). Global forces channelled through the financialisation of economies have displaced the wage from the centre of the growth circle and replaced it with profit and corporate welfare (Skidelsky and Craig 2016; Palley 2016). The decline in wage shares has seen a rise in household reliance on credit to finance consumption becoming a major growth determinant. Neoclassical theory hypothesises that household debt positively affects economic growth (Lombardi, Mohanty, and Shim 2022). Since credit helps households to smoothen consumption, the growth of credit to households is a sign of a prospering economy, according to this view.
Financial development, often measured as the share of credit to the private sector (firms and households) in national income, is considered an indicator of how finance drives growth by overcoming credit constraints in the economy. Lombardi, Mohanty, and Shim (2022) and Mian, Sufi, and Verner (2017) have challenged the simplistic understanding of the effect of household debt on economic growth showing that household debt positively affects economic growth in the short run only but negatively in the medium and long runs. The critique is in line with what has been called the ‘vanishing effect of financial development’ (Arcand, Berkes, and Panizza 2015), which means it only has a positive effect in the short run. The vanishing effect of finance has also been associated with stylised facts of ‘growthless credit expansion’ and ‘creditless economic recoveries’ (Juselius and Drehmann 2020).
Although there is a large body of literature on determinants of household debt, examining the effect of household debt and debt service on economic growth remains an under-researched issue, especially in developing countries because of lack of data or fragmentary data. Further, in spite of an unambiguous theoretical postulation of the relationship between household debt and growth, the empirical literature has focused more on the determinants of household debt than on its effects on growth (Lombardi, Mohanty, and Shim 2022). Household debt often is the source of growth instability through high debt default rates leading to financial crisis or abrupt deleveraging leading to sharp expenditure cascades (Eggertsson and Krugman 2012; Kapeller and Schütz 2014; Stockhammer, Engelbert Richard and Wildauer 2018).
In the case of South Africa, the South African Reserve Bank (SARB) has been condemned by Trade Unions and other left-leaning organisations for policy rate hikes (Marire 2023a). The outcry largely has been on the implication of policy rate hikes on the household debt service burden. The general sentiment is that the SARB’s interest rate policy is a blunt instrument that tames inflation while worsening business cycles as aggregate demand falls because households have more of their disposable income redirected to debt servicing. Despite this sentiment, empirical studies in (South) Africa have not explored the effect of household debt service, not to mention the effect of household debt, on economic growth. Yet, in a group of seventeen countries, Tunc and Kilinc (2023) established the existence of a household debt service channel to economic growth. The channel plays a much larger role than household debt in influencing economic growth. Tunc and Kilinc (2023) establish that the debt service channel carries more weight than the debt channel in influencing economic growth so much so that monetary authorities must be careful about their interest rate policy decisions. The negative effects of the debt service burden on economic growth are more prolonged than those of debt levels (Tunc and Kilinc 2023).
The aim of the paper is to examine the effect of the growth rates of debt ratios rather than their levels on future economic growth. The paper is concerned with what happens to future GDP growth following periods of fast rising (falling) debt and debt service ratios. The paper contributes to the literature by examining three questions.
First, does the growth rate of the household debt ratio in South Africa affect future economic growth in the short and long runs? Fig. 1 (Panel A) seems to depict the two phenomena of ‘growthless credit’ and ‘creditless growth’ as explained by (Juselius and Drehmann 2020) and an additional one, credit-driven growth. Creditless growth does not mean there is no credit but rather that national output recovers without significant pumping in of credit to stimulate the recovery, while growthless credit implies that significant extension of credit yields little to no response from growth. For example, in the 1970s, the household debt ratio was declining, while GDP was increasing (creditless growth), and the same is true for later periods such as 1997–2002, and 2008–2018. Then there are periods when the household debt ratio rose sharply while GDP rose very slowly (growthless credit), for example, 1980–1987, 1988–1996, 2003–2008. Periods of significant decreases in the debt ratio include the aftermath of the oil price crisis of the early 1970s, hawkish interest rate policy following the rand crises of 1996–1997 and 2001, as well as a hardnosed monetary policy stance after the global financial crisis (Marire 2023b). Periods of significant increases in the household debt ratio include periods of significant financial liberalisation and credit accessibility in the 1990s, the dotcom bubble in the mid-1990s, and the commodity supercycle of the 2000s (Marire 2023b; Aron and Muellbauer 2000; Prinsloo 2002). Lastly, during the COVID-19 pandemic, the debt ratio rose sharply, but this is not a reflection of new debt but the effect of the contraction of GDP and, ultimately, household disposable income (the denominator in the debt ratio).
Second, does the growth rate of the household debt service ratio influence economic performance in South Africa in the short and long runs? This question is particularly important in South Africa where the central bank policy rate decisions have become the single most contested policy issue, mainly regarded as “anti-poor, anti-labour, and pro-capital” (Marire 2023a: 58). If the debt service channel holds for South Africa, the findings will have important implications for the conduct of monetary policy.
The pattern of the relationship between the debt ratio and growth on the one hand, and the debt service ratio and growth on the other is clearer when changes in the growth rates of debt and debt service ratios are considered (Fig. 2, Panels A-D and Fig. 3, Panels A-D). The first observation is that debt and debt service growth rates are much more pronounced, and they tend to be followed by GDP growth. This shows that debt and debt servicing costs significantly influence business cycles. The second observation is that there are periods when the growth rate of the household debt ratio coincides with decreasing economic performance and also the decline in the growth rate of the household debt ratio coincides with rising GDP growth (Fig. 2, Panels A-D), suggesting what Juselius and Drehmann (2020) called ‘growthless credit’ and ‘creditless growth’. The period 1998–2005/6 has a significant divorcing of household debt ratio growth rate and GDP growth rate. This suggests that growth was driven by good fiscal governance and the commodity supercycle of the 2000s. The debt households accumulated had a limited effect on economic performance. Figure 3 (Panels A-D) shows that periods of fast-rising (fast-falling) debt service ratio are followed by economic decline (growth).
Third, given the sharp swings in the growth rate of the debt and debt service ratio, do their effects on future GDP growth vary along its distribution? The question is crucial to examine because periods of significant leveraging and deleveraging have substantial implications for economic stability (Teulings, Wouterse, and Ji 2023; Stockhammer and Wildauer 2018; Eggertsson and Krugman 2012). At present, the South African literature has yet to explore the possibility that the effects of the growth rates of household debt and debt service ratios might vary over the future GDP growth distribution. The extent to which the effects of debt service costs vary on the economic growth distribution has implications for interest rate policy because monetary policy tightening might trigger an economic recession while attempting to control another variable, inflation.
The paper finds that the growth rate of the household debt ratio has a positive effect on future growth and the effect is sustained into the 20-quarter growth horizon. Debt service has negative effects on future growth and the effect is sustained even into the 20-quarter growth horizon. More importantly, the quantile effects of debt and debt service ratio growth are larger in the lower half of the conditional future growth distribution than in the upper half of the conditional future growth distribution. This means that at lower levels of growth, fast-rising debt improves economic growth, but fast-rising debt service burden undermines growth significantly. The paper is organised as follows. Section 2 reviews the literature. Section 3 outlines the method and data issues. Section 4 presents the findings. Section 5 concludes the paper.