The impact of higher leverage ratios on the South African economy

Abstract

We employ a micro-founded, stock-flow consistent computable general equilibrium model to study the impact of increases of the leverage ratio on the South African economy. The model provides for a richer representation of institutional balance sheets than existing models and captures the important relationship identified in the literature between bank capital, lending spreads and economic activity. The financial accelerator mechanism operates through the balance sheets of all institutions in the economy. The move to a higher leverage ratio for banks is likely in the short-run to generate negative economic impacts that depend on the banks’ choice of adjustment strategy. The negative GDP effect is greatest if the financial sector reduces leverage by reducing the value of its assets rather than raising its liabilities. The shock also leads to the financial sector changing its perceptions of risk, which reduces the size of the money multiplier and increases lending spreads. The transition to a higher leverage ratio also affect the transmission of monetary policy. Executing monetary policy effectively thus requires understanding how the financial sector is likely to meet the new requirements and how its perceptions of risk are affected.

Keywords:

• Basel III leverage ratio
• stock-flow consistent model
• macro-prudential policy
• computable general equilibrium

JEL CLASSIFICATION:

• E44
• E47
• E50
• G20

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Our interest is in the short-run transition to higher capital requirements. FSB (2010) provides estimates of the long-run net benefits. They find that the net benefit of the proposed capital requirements in Basel III is around 2% of GDP on average across countries. The main channel is that a well-capitalised banking system will lead to a reduction in the risk and cost of financial crises and macroeconomic volatility. Caggiano and Calice (2011) find similar results for African economies.

2 As highlighted in the literature review, the presence of capital buffers does not imply that regulatory capital requirements are not binding as banks may want a buffer over the regulatory requirement to reduce the risk of anticipating difficulties with raising equity in the future or as a precaution against shocks that may hinder their future lending.

3 See for example Blum (1999), Repullo (2004) and Kim and Santomero (1988).

4 See for example MAG (2010b), Slovik and Cournède (2011), IIF (2011) and EU (2011); De Marco and Wieladek (2015) identify three conditions that need to be satisfied for higher capital requirements to affect loan supply: the cost of bank equity must exceed the cost of debt; capital requirements must be binding on a bank’s choice of capital; and the sources of funding of borrowers must be limited.

5 Cohen and Scatigna (2016) explain how the different options are operationalised, and the advantages associated with each.

6 See also Hollander and Van Lill (2019) and Hollander and Havemann (2021) for a review of macroprudential policy in the post-apartheid period.

7 The current systemically important banks are Absa, Capitec, First National Bank, Nedbank, Investec and Standard Bank.

8 While financial institutions have increased their Basel III leverage ratio, they have started from a point above the regulatory minima. Following the results of Acosta-Smith, Grill, and Lang (2020), this suggests that higher leverage ratio in South Africa has been accompanied by lower risk taking.

9 See for example Merven, Hartley, and Schers (2020).

10 The relationship between Tobin’s Q and investment is time and frequency varying. Cash flow tends to be a good predictor of investment in the medium to long-run while Tobin’s Q tends to be a good predictor at business cycle frequencies (Verona, 2020). There are also other factors that impact investment at different times. These include for example uncertainty or trade openness (Binge & Boshoff, 2020; Jadhav, 2012; Redl, 2018). The multiplicity of different drivers their time and business cycle variation make modelling investment a difficult task. Our simple specification improves the model tractability but it is also an area for future analysis.

11 For discussion of Taylor rules in South Africa see Bold and Harris (2018).

12 Hommes (2011) provides a review of the literature on bounded rationality. The theory of bounded rationality originates in the seminal work of Simon (1955).

13 This approach captures only the capital threshold effect identified by Borio and Zhu (2012). However, we can capture the framework effect via the model solution process. Different frameworks can be introduced between iterations.

14 The empirical evidence supporting the risk channel are mixed. For example, DellʼAriccia, Laeven, and Marquez (2014) show that risk taking depends on the capital structure, leverage and competition. Similarly, Valencia (2014) finds that the channel depends on the size of the shock and the ability of banks to issue equity; While we introduce a rich financial market representation, the current specification requires further disaggregation to capture fully the heterogenous and systemic risks, which has been identified by Borio and Zhu (2012) and Duca and Muellbauaer (2014). This is a limitation of our study.

15 The use of book values rather than market values simplifies the model solution. The impact of market value dynamics on the overall results depends on the composition of assets and consequently on how changes in market prices affect the numerator of the leverage ratio vis-à-vis the denominator. This is an item for further research.

16 The relationship between balance sheet strength, credit extension and lending spreads is well established in literature. In addition to the references in this paper, the reader can refer to Kapan and Minoiu (2018). For South Africa, Loate and Viegi (2021) illustrate how monetary shocks affect lending via the bank balance sheet. They distinguish between small and big impacts.

17 For example local, provincial and national government are grouped into Government.

18 See for example Makrelov et al. (2021)

19 We classify Households’ interest in retirement and life funds, which is a liability of the financial sector, as equity in terms of our classification framework presented in Makrelov et al. (2020).

20 Our shock is to leverage ratio whereas the shocks by Grobler and Smit (2014) and Havemann (2014) are to the capital adequacy ratio. The shocks are the same in relative terms given the intial value of the respective ratios. Grobler and Smit (2014) and Havemann (2014) do not model risk weights, similar to us. The absence of risk weights in the adjustment assumes that the compositional effects are less important and allows some comparison between the two sets of results.