Derivative Markets and Economic Growth: A South African Perspective

Abstract

It is well established that financial development and innovation promote economic growth through improving the allocation of capital, enhancing risk management, contributing to price discovery, and increasing market efficiencies. While a vast empirical literature is devoted to the nexus between financial development and economic growth, however, substantially less research has been done on the relationship between derivatives and growth, especially in the emerging-market context. Derivatives can be viewed as a specific category of financial innovation, which may advance economic growth through its specialised functions of risk management and price discovery. This paper contributes to bridging this gap in the literature by exploring the impact of exchange-traded futures derivatives on South African economic growth, output, and economic growth volatility. It employs ARDL bounds tests, Granger causality tests and GARCH volatility modeling to analyse the effects of exchange-traded futures derivatives on various measures of South African economic activity. The main result is that exchange-traded futures derivatives contribute positively to South African economic growth and economic activity. This may suggest that opportunities might exist in other emerging economies, with financial structures comparable to that of South Africa, to encourage the development of organised and well-regulated derivatives markets to unlock economic growth in these economies.

1. Introduction

Neoclassical growth theory posits that economic growth is driven by three factors, namely, capital, labour, and technological innovation (Solow 1956, 1957; Swan 1956). ‘Innovation’, however, need not be limited to technological advances only—it could also represent financial innovation. It has been argued that financial innovation is critical to accompany technological advances in order for economic development to occur, and that “the nexus of finance and innovation is… central to the process of economic growth” (King and Levine 1993b, p. 515). Similar to how technological innovations can make capital and labour inputs more productive to realise higher economic growth rates, financial innovation “ameliorates information and transactions costs”, which positively “influence saving rates, investment decisions, technological innovation, and hence long-run growth rates” (Levine 2005, p. 867).

Research shows that financial development and innovation are associated with improved economic performance. Levine (2005) observes that economies with better functioning banks and markets tend to grow faster. An efficient financial sector pools domestic savings and mobilises capital for productive investment, supporting “capital-raising efforts of large domestic corporations” (Bekaert and Harvey 1998, p. 34), thereby easing the “external financing constraints that impede firm and industrial expansion” (Levine 2005, p. 867), which in turn contribute to improved production and economic growth. Beck et al. (2000, p. 261) find that “financial intermediaries exert a large, positive impact on total factor productivity growth, which feeds through to overall GDP growth”, while better financial services “expand the scope and improve the efficiency of innovative activity… thereby accelerat[ing] economic growth” (King and Levine 1993b, p. 517).

However, while it has long been established that financial innovation is positively associated with economic growth and development, questions about the causal direction remain. The classical view on the relationship between finance and growth was that “where enterprise leads finance follows” (Robinson 1952, p. 86). Much of the subsequent empirical literature was therefore devoted to determining whether the relationship between finance and economic growth is ‘demand-following’, that is, whether “the evolutionary development of the financial system is a continuing consequence of the pervasive, sweeping process of economic development” (Patrick 1966, pp. 174–75), or ‘supply-leading’, where financial system development precedes the demand for its services. If evidence supports the supply-leading hypothesis, real economic growth could be induced by financial means (Patrick 1966). Encouraging, and indeed driving, financial development could, therefore, from a policymaker’s perspective, be an “active approach to promote economic growth” (Baluch and Ariff 2007, p. 9).

This paper explores the potential growth-enhancing characteristics of derivatives as one specific dimension of financial innovation. Derivatives are financial instruments that derive their value from underlying assets, securities or indices. Well-known derivative instruments include options—contracts giving the holder the right, but not the obligation, to buy or sell an asset at an agreed price before the option expires—futures and forwards—agreements to buy or sell an asset at a predetermined price at some specified future date—and swaps—contracts in which two parties agree to exchange underlying cash flows over a specified period. Derivatives are mainly used for the purposes of hedging, speculation, and arbitrage (Van Wyk et al. 2019). Hedging refers to actions taken to protect against the risk of an adverse outcome, while speculation refers to the taking of a risk position in a market purely for profit. Arbitrage is the exploitation of pricing anomalies between markets to realise low-risk profits. Derivatives are traded in two types of markets, namely exchange-traded and over-the-counter (OTC). Exchange-traded derivatives are standardised contracts traded on an organised exchange, while OTC derivatives take the form of tailored contracts between two counterparties.

Even though derivatives are not a new phenomenon—commodities futures, for example, were already traded in ancient times—in recent decades the inexorable rise of organised derivatives markets contributed substantially to financial innovation. Derivatives can enhance risk allocation and market liquidity, contribute to price discovery by ameliorating informational inefficiencies, and reduce transaction costs, thereby improving market efficiency and stimulating economic activity (Haiss and Sammer 2010). However, the misuse of derivatives could introduce additional risks which might increase financial market volatility and perversely constrain economic activity (Prabha et al. 2014).

Numerous studies have considered the nexus between finance and financial development, and economic growth. Seminal papers include King and Levine (1993a, 1993b), Rajan and Zingales (1998), Beck et al. (2000), and Levine (2005). However, despite the vast empirical literature on the relationship between finance and growth, the link between derivative markets, economic growth and macroeconomic factors is still insufficiently covered in the academic literature (Vo et al. 2020) and “comprehensive research in this field remains scarce” (Samarakoon et al. 2024, p. 187). Comparatively few studies have considered the relationship between derivatives and growth, and fewer still have considered the impact of derivatives on South African economic growth. There is thus currently a gap in the research investigating the relationship between derivatives and economic growth in South Africa specifically, and the potential growth-enhancing impact of derivatives in emerging markets more broadly. Due to derivatives being a relatively complex financial phenomenon, “derivatives markets in emerging economies remain small compared to those in advanced economies” (Mihaljek and Packer 2010, p. 44). Large and well-developed derivatives markets are relatively uncommon in emerging economies, which may explain the comparatively little attention paid to emerging economies in research related to derivatives. However, if policymakers are able to better understand how derivatives influence economic growth, derivatives could potentially be harnessed as a valuable tool for enhancing financial development and supporting or stimulating economic growth in emerging economies.

This paper employs Pesaran et al.’s (2001) auto-regressive distributed lag (ARDL) bounds test, Granger causality tests, and generalised auto-regressive conditional heteroskedasticity (GARCH) modeling to investigate the relationship between the organised derivatives market and South African economic growth. It contributes to the literature by analysing the long- and short-term effects of exchange-traded futures derivatives on economic growth, the level of output, and growth volatility, none of which have been done in the South African context before. The sparse domestic literature is partly due to the fact that formal exchange-traded derivatives markets have only been in existence in South Africa for about 30 years, while the bulk of the existing empirical literature concentrates on advanced economies with older and more sophisticated derivative markets. Furthermore, South Africa is an example of a small open emerging economy with very well-developed financial markets. Some of the lessons or salient features from the South African case could therefore potentially be generalised and extended to other emerging economies with comparable financial infrastructure, and who might be positioned to take advantage of the development of their own financial—and specifically derivatives—markets to stimulate domestic economic activity.

The remainder of this paper is structured as follows: Section 2 discusses the theoretical literature and describes how financial markets broadly, and derivatives in particular, could influence economic activity and growth. Empirical literature on the derivatives-growth nexus is also reviewed in this section. The data and econometric methodology are described in Section 3, followed by the presentation and discussion of the results in Section 4. Section 5 concludes.

2. Literature Review

2.1. Theoretical Overview

2.1.1. Financial Development and Economic Growth

Bekaert and Harvey (1998) assert that economic growth in a modern economy requires an efficient financial sector which pools savings and mobilises capital towards productive investment. Without effective financial institutions, “productive projects may remain unexploited” (Bekaert and Harvey 1998, p. 33). King and Levine (1993b, p. 513) argue that “financial systems are important for productivity growth and economic development”; in fact, “finance seems importantly to lead economic growth” (King and Levine 1993a, p. 730).

Levine (2005) identifies five broad functions (see also Merton and Bodie (1995)) provided by the financial system which contribute to economic growth: (i) Provision of information which improves the allocation of capital; (ii) Pooling and mobilisation of savings; (iii) Easing the exchange of goods and services; (iv) Facilitating the trading, diversification and management of risk; and (v) Improved corporate governance and monitoring of investments.

(i)

Individual savers may not have access to adequate information on possible investments. The large information costs associated with evaluating market conditions, firms and managers may thus prevent capital from flowing to its highest value use. Financial markets and intermediaries may, however, “reduce the costs of acquiring and processing information and thereby improve resource allocation” (Levine 2005, p. 871). Improved resource allocation fosters economic growth, which in turn provides the means to improve financial structures, giving rise to a virtuous development cycle (Greenwood and Jovanovic 1990).

(ii)

Mobilisation of savings is the “costly process of agglomerating capital from disparate savers for investment” (Levine 2005, p. 879). Effective pooling and mobilisation of savings enable capital projects and investments that can overcome investment indivisibilities and exploit economies of scale. Bekaert and Harvey (1998, p. 34) argue that the efficient allocation of capital is “by far the primary role” of the financial system, that is, “to allocate funds to the investment projects with the highest marginal product of capital”. Greenwood and Jovanovic (1990, p. 1076) argue that “financial intermediation promotes growth because it allows a higher rate of return to be earned on capital”. Capital markets thus have the ability to transform savers’ liquid assets into long-term capital investments, acting as “an engine for economic growth” (Bekaert and Harvey 1998, p. 33).

(iii)

Financial systems ease the exchange of goods and services through the provision of payment systems’ clearing and settlement facilities. This lowers transaction and information costs, facilitating trading, and fostering specialisation, innovation and growth.

(iv)

Asymmetric information and high transaction costs “may inhibit liquidity and intensify liquidity risk” (Levine 2005, p. 876). Financial intermediaries such as banks and insurers therefore often assume liquidity and credit risks on behalf of their clients (Van Wyk et al. 2019). Furthermore, although savers are generally risk-averse, projects with higher returns tend to be riskier than projects with lower returns (Levine 2005). It follows that financial systems with better risk diversification capabilities could channel resource allocation towards higher-risk-higher-return investments, which would contribute to higher production and growth. King and Levine (1993b) suggest that effective risk diversification also stimulates technological innovation, finding that “better financial systems improve the probability of successful innovation and thereby accelerate economic growth” (King and Levine 1993b, p. 513).

(v)

Effective corporate governance and monitoring incentivise managers to maximise firm value (see Jensen and Meckling’s (1976) and Boyd and Prescott’s (1986) seminal papers), improving the efficiency with which firms allocate resources, and making investors more willing to finance production and innovation (Levine 2005).

It follows that financial development—defined by The World Bank (2019, p. xv) as the “process of reducing the costs of acquiring information, enforcing contracts, and making transactions”—may then be equated to the progress and advances made by financial systems in their organisation and activities. Loayza and Ranciere (2006, p. 1069) argue that “financial development entails a deepening of markets and services that channel savings to productive investment and allow risk diversification”, which directly contributes to higher long-term economic growth. A better developed financial system can thus provide better information to more investors, improve the governance, ethics and behaviour of firms to which investment and credit have been allocated, and provide more viable options for risk mitigation and transfer. It improves the aggregation and allocation of savings, and facilitates more efficient transfer of goods and services. Financial development may also attract foreign direct investment (FDI), bringing benefits such as capital inflows, international technology transfer, the introduction of new processes, procedures and systems, inflow of skills and knowledge, and access to new markets and international production networks (Alfaro et al. 2004). Advanced financial tools foster “a better distribution of resources” (Vo et al. 2019, p. 2). Ultimately, well-functioning financial markets and intermediaries “ameliorate information and transaction costs to promote efficient resource allocation, and, hence, economic growth” (Şendeniz-Yüncü et al. 2018, p. 411).

2.1.2. Derivatives and Economic Growth

Derivative markets are but one cog in the larger financial system. As such, they cannot fulfill all the functions of the financial system described above, although they can support economic activity by specialising in some of these channels. Indeed, the absence of derivatives markets would substantially lower the efficiency of the financial system in performing some of these functions, which may well retard economic growth; the derivatives market is therefore a “vital part of a financial system” (Vo et al. 2020). Financial market inefficiencies restrict economic activity through high transaction costs, illiquidity, unfair prices, and incomplete information (Bekaert et al. 1995). Inefficiencies also distort investment decisions, causing “important projects to be turned down, leading to lower economic growth, while low-return projects are implemented” (Bekaert and Harvey 1998, p. 48).

(i)

Baluch and Ariff (2007, p. 16) argue that “countries with well-functioning derivatives markets should have a higher growth rate than countries without one.” Derivatives allow risk exposures to be “be better hedged against price and interest fluctuations” (Haiss and Sammer 2010, p. 7), thus fulfilling an important risk management function. Derivatives enable investors to “unbundle and transfer financial risk” (Adelegan 2009, p. 3), enabling the allocation of risk to those market participants willing and able to bear those risks. Derivatives are thus important “mechanisms to share risk and to allocate capital efficiently” (Şendeniz-Yüncü et al. 2018, p. 410).

(ii)

Hedging through derivatives can reduce uncertainty and volatility, particularly in sectors such as agriculture or commodities, or industries vulnerable to unexpected shocks or exposed to foreign exchange risk. By reducing the vulnerability of market participants to potential losses, derivatives contribute to financial stability, which could in turn “free up capital to invest in new value-enhancing and growth-driving projects” (Bekale et al. 2015, p. 5).

(iii)

Derivatives can “drive the markets to one common price for a specific asset through the use of arbitrage opportunities” (Haiss and Sammer 2010, p. 20), thus playing an integral role in price discovery. Derivatives also reveal information to market participants (Prabha et al. 2014), thereby providing an “additional channel through which new information can be quickly incorporated into asset prices” (Rodrigues et al. 2012, p. 4). Derivatives enable easier access to capital markets for new firms, with such start-ups being vital contributors to economic growth and job creation, while in general lowering funding costs and enabling more diversified funding sources for firms and investors.

(iv)

Finally, derivatives can reduce transaction costs by increasing market liquidity, substituting cash market trades, transferring resources across time and space, and providing means to manage risks and provide pricing information, all of which contribute to the “efficiency with which economies combine capital and labour in production” (Haiss and Sammer 2010, p. 7).

2.1.3. Derivatives, Risk and Volatility

It was established above that derivatives could theoretically play an important role in financial markets’ ability to enhance economic growth. However, derivatives may also be associated with increased systemic risk and contagion effects which could cause, or exacerbate, financial crises (Aali-Bujari et al. 2016). There may be two broad reasons for this. First, unbridled innovation has led to the emergence of much more complicated and highly leveraged derivative instruments. These include instruments such as collateralised debt obligations (CDO) and credit default swaps (CDS). During the build-up to the 2007–2008 global financial crisis (GFC) these were (i) based on underlying assets of increasingly questionable quality, and (ii) packaged and repackaged to obscene levels. Indeed, poorly regulated derivative instruments “played a ‘catalytic role’ in the global financial crisis” (Haiss and Sammer 2010, p. i). Second, the proliferation of derivative markets can be associated with a “structural move from regulated to more unregulated market activity” (Haiss and Sammer 2010, p. 2). In this sense, although derivatives have been argued above to play a positive role in allocation and management of risks, derivatives could perversely increase financial fragility by attracting speculators (Aali-Bujari et al. 2016) and encouraging risk-taking behaviour. Because derivatives derive their value from the value of some underlying asset, this might then encourage further speculation in underlying assets (Tobin 1984), “diverting private and public resources from efficient allocation” (Haiss and Sammer 2010, p. 1).

However, it should be noted that derivatives’ two main uses—hedging and speculation—introduce very different risk characteristics into markets. Hedging is an activity designed to manage risk, and can therefore be expected to improve risk allocation and reduce systemic risk. Speculation, on the other hand, is a profit-seeking activity, and, given the complex and leveraged nature of derivatives, can likely introduce additional risk and volatility. While the speculator is, of course, the natural counterparty to the hedger (if one party wishes to reduce his risk exposure, another party is required to assume that risky position), excessive speculation could increase volatility and pose risks to financial stability.

Furthermore, the two types of markets in which derivatives are traded, organised exchange-traded and OTC, also have very different risk characteristics. OTC derivative contracts are often “privately negotiated and bilaterally arranged” (Adelegan 2009, p. 6); they therefore tend to be more flexible as they can be tailored to specific requirements, and can be cheaper and easier to access than exchange-traded derivatives. However, there is less transparency in these markets, contributing to information asymmetry, while they are often unregulated. OTC derivative contracts could also be more vulnerable to counterparty and credit risk (Sill 1997). Exchange-traded derivatives, on the other hand, take the form of standardised contracts trading on an organised and regulated exchange, offering a lot more transparency and oversight. This lowers the risk of abuse and alleviates information asymmetries. The downside to exchange-traded derivatives, however, largely has to do with their standardised, and therefore relatively rigid, nature. They are therefore not always suited to meet the specific needs of some participants. These instruments can also require large minimum contract sizes which may rule out smaller businesses or investors from using them.

From the discussion above, it should be clear that the potential undesirable effects of derivatives—arising from complexity and excessive speculation—could to some degree be alleviated through appropriate oversight and regulation to promote responsible use of derivatives.1 Despite the potential risks introduced by derivatives, Stulz (2004, pp. 190–91) argues that “on balance… the whole economy gains from the existence of derivative markets”. Exchange-traded derivatives introduce less risk and are therefore arguably healthier for financial development and the economy as a whole, and will thus be the focus of the remainder of this study.

2.2. Empirical Literature Review

While a vast empirical literature exists on the nexus between finance and economic growth, comparatively few studies explore the relationship between derivatives and growth. These studies’ approaches can be broadly grouped into two distinct categories, namely time series analyses on an individual country basis, and panel studies encompassing several countries.

Baluch and Ariff (2007) employ time series analyses and Granger causality tests to determine whether derivatives exhibit demand-following or supply-leading patterns of economic development in a sample of 11 countries over 1990–2006. They find mixed evidence on the relationship between derivatives and growth. Only for Brazil, the UK and Denmark do they establish unidirectional causality from economic growth to derivatives (the demand-following pattern), while they find no evidence to support the supply-leading hypothesis. Şendeniz-Yüncü et al. (2007) investigate the relationship between futures market development and economic growth, using both time series and dynamic panel methods. Their study considers both developed and emerging economies, and detects a positive relationship between futures market development and economic growth. Haiss and Sammer (2010) augment a production function model with derivatives and various financial variables to investigate the impact of derivatives markets on the finance-growth nexus in the US over 1990–2008. While they find evidence of a significant positive relationship between economic growth and indicators of financial development such as bank assets and stock market liquidity, they find only a small correlation between derivatives and growth. Prabha et al. (2014) assess the economic impact of derivatives in the US in a two-step framework: First, they estimate the impact of derivatives on bank lending and firm value, and then examine the effects of bank lending and firm value on economic growth. They find that derivatives improve banks’ ability to provide credit, and that derivatives have a positive impact on firm value. Increased bank lending and firm value, in turn, both lead to stronger economic growth. Şendeniz-Yüncü et al. (2018) conduct individual vector error-correction (VECM) analyses on a sample of 32 countries over 1982–2015. They find evidence supporting the supply-leading hypothesis in the form of Granger causality from the futures market to economic growth for middle-income countries, concluding that derivative markets stimulate economic growth in these economies. However, they detect the reverse effect for high-income countries, suggesting that in these countries economic growth leads to the development of futures markets; for high-income countries the evidence therefore supports the demand-following hypothesis. Vo et al. (2019) utilise the ARDL bounds approach of Pesaran et al. (2001), as well as Granger causality tests within a VECM framework, to investigate the relationship between derivatives and economic growth for the US, Japanese, Indian and Chinese economies. Their main result is that derivative markets have a positive influence on economic growth. They find mixed evidence on the impact of derivatives on growth volatility, as well as the directions of causality among the four countries. Lema and Grandes (2020) utilise a Spearman correlation approach, and find that “derivatives are positively correlated with economic growth” (Lema and Grandes 2020, p. 45).

Rodrigues et al. (2012) analyse the effect of the institutionalisation of derivatives markets on economic growth and growth volatility using a panel data set of 45 countries over 1971–2009. Their results show a robust “statistically and economically significant positive effect of the establishment and existence of a domestic derivatives exchange on economic growth” (Rodrigues et al. 2012, p. i). Aali-Bujari et al. (2016) employ a dynamic panel data model over 2002–2014 to assess the impact of derivatives on economic growth for six major world economies (the EU, the US, Japan, China, India and Brazil). Their main empirical result is that “derivatives markets have a positive influence on economic growth” (Aali-Bujari et al. 2016, p. 110). Vo et al. (2020) employ a panel vector auto-regression (PVAR) model and panel causality tests to analyse the relationship between derivatives and economic growth for a sample of 17 countries over 1993–2017. Controlling for macroeconomic factors such as inflation, trade openness and government expenditure, they find evidence of, respectively, bidirectional and unidirectional causality from economic growth to derivative markets for high-income and upper-middle-income economies. Consistent with Şendeniz-Yüncü et al. (2018), their findings support the ‘demand-following’ hypothesis for the higher-income countries, with Vo et al. (2020) concluding that economic growth in higher-income economies increases the demand for more complex financial systems and instruments. Similarly, Samarakoon et al. (2024) utilise a panel-ARDL framework to investigate the relationship between derivatives markets, measured as the volume of equity derivatives contracts traded, and real GDP for a panel of Asia-Pacific economies. They find evidence of a positive and significant long- and short-term relationship between derivatives and real GDP, as well as evidence of bidirectional causality between derivatives and GDP.

In the fledgling South African literature, Marozva (2014) employs the ARDL bounds testing approach and Granger causality tests to examine the nexus between derivatives and economic growth. While Marozva (2014) finds evidence that derivatives support capital market development, he finds no direct link between derivatives and economic growth. Bekale et al. (2015) employ a time series approach based on Rodrigues et al. (2012) to the South African economy, finding evidence of unidirectional causality from economic growth to derivatives. They also find that the establishment of a domestic derivatives exchange is associated with a reduction in economic growth volatility, and conclude that the existence of a formal derivatives exchange has had a stabilising effect on the economy. Mulei (2019) finds no significant relationship between derivatives and economic growth, and no impact of derivatives on growth volatility. Finally, Chikwira (2021) follows a similar approach to that of Prabha et al. (2014), finding that derivatives permit more private and public sector credit extension, which contribute to economic growth.

3. Data and Methodology

3.1. Data

This study follows Beck et al.’s (2000) broad paradigm to analyse the effects of finance on growth, where economic growth is expressed as a function of a financial development variable and a group of conditioning variables to “control for other determinants of growth” (Beck et al. 2000, p. 35). This approach is also followed in the empirical literature, where many studies express economic growth as a function of a financial development variable and a set of control or conditioning variables to isolate the effects of finance on growth (see Section 2.2). Financial development variables from the empirical literature include banking assets, credit extension, stock market capitalisation and various measures of derivatives traded, while the control variables may include wealth or income, capital investment, human capital, inflation and trade openness.

To analyse the impact of derivatives on South African economic growth, the value of domestic exchange-traded futures is chosen as key explanatory variable. This is in line with the approaches of Şendeniz-Yüncü et al. (2018) and Vo et al. (2019), who also use futures as proxy for derivatives market activity. The co-movement between real GDP and domestic exchange-traded futures since 1990 is illustrated in Figure 1. Exchange trading of derivatives has grown as the economy expanded. During the two notable recessions in the sample period, the 2008–2010 GFC and the 2020 COVID-19 shock, derivatives trading temporarily slowed down. This was especially pronounced during the GFC, likely due to the sharp global slowdown in financial derivatives trading during and shortly after the crisis. In fact, Figure 1 suggests that it took nearly six years for the domestic derivatives market to recover its value after the GFC. A smaller contraction in the derivatives market is noticeable between 2001–2003, potentially as a result of the burst of the dot-com bubble in 2000 and the rand currency crisis of 2001.

Following Levine et al. (2000) and Loayza and Ranciere (2006), capital investment, inflation and trade openness are included as macroeconomic control variables. Labour productivity is utilised as human capital control variable. The main dependent variable, economic growth, is calculated as the quarter-on-quarter log-difference of real GDP. Variables and their proxies are summarised in

Table 1. Model variables and data sources.

Variable	Code	Description	Source
Economic growth	Y	Log-difference of real GDP (base year 2015)	SARB
Derivatives	DER	Underlying value of exchange-traded futures	SARB
Capital investment	INV	Gross fixed capital formation	SARB
Trade openness	OPEN	(Exports + imports)/GDP	SARB
Inflation	INFL	q.o.q. % change in consumer price index (CPI)	StatsSA
Productivity	PROD	Labour productivity in the non-agricultural sectors	SARB

Data sources: South African Reserve Bank Quarterly Bulletin; Statistics South Africa. The quarterly value of futures derivatives is the summation of its monthly underlying values. The full dataset can be made available upon request.

. Data were sourced from the South African Reserve Bank (SARB) and Statistics South Africa (StatsSA) via Quantec’s EasyData portal. Quarterly data spanning 1990–2019 are utilised. The sample is deliberately cut off before 2020, so as to prevent the massive data volatility due to the impact of the COVID-19 pandemic on economic growth from biasing the estimations. Natural logarithms of real GDP, derivatives, and capital investment are used.

One important limitation arises in the data: Data on OTC derivatives trading are not regularly published, and in the South African context only data on domestic exchange-traded futures are available over a long enough sample to enable meaningful statistical analyses. This means that these analyses can only evaluate the relationships between formal exchange-traded futures derivatives and our various measures of economic activity. This is a notable caveat, given the fact that OTC derivatives represent a significant portion of the total volume and value of derivatives traded globally (Van Wyk et al. 2019). Moreover, because OTC markets represent a less transparent, less regulated, and therefore riskier, portion of derivatives trading (see Section 2.1.3), the absence of OTC data may therefore bias the results away from the inherent riskiness arising from this class of derivatives. The lack of comprehensive and reliable data on OTC derivatives means that the results below should be interpreted with this in mind.

The remainder of this section briefly describes the econometric techniques and tests employed in the analyses, with the results presented and interrogated in Section 4.

3.2. ARDL Bounds Test

The autoregressive distributed-lag (ARDL) bounds approach of Pesaran et al. (2001) is employed to test the impact of exchange-traded derivatives on South African economic growth. The benefits of the ARDL paradigm lie in its power in relatively short time series, while it can also be applied to models containing variables integrated of different order, which, as

Table 2. Unit root tests.

	Levels		First Difference		Order of Integration
Variable	t-Stat	Prob.	t-Stat	Prob.
Augmented Dickey-Fuller (ADF)
GDP growth	−2.748	0.069	−3.771	0.043	I(0)/I(1)
Real GDP	−0.533	0.880	−5.465	0.000	I(1)
DER	−2.493	0.120	−13.073	0.000	I(1)
INV	−0.895	0.787	−7.397	0.000	I(1)
OPEN	−2.597	0.097	−12.905	0.000	I(0)/I(1)
INFL	−1.839	0.063	−9.283	0.000	I(0)/I(1)
PROD	−0.313	0.919	−9.734	0.000	I(1)
VOL	−4.426	0.003	-	-	I(0)
Phillips-Perron (PP)
GDP growth	−2.698	0.074	−6.845	0.001	I(0)/I(1)
Real GDP	−0.290	0.922	−5.444	0.000	I(1)
DER	−2.999	0.038	−12.975	0.000	I(1)
INV	−0.737	0.832	−7.572	0.000	I(1)
OPEN	−2.534	0.110	−13.220	0.000	I(1)
INFL	−1.978	0.046	−10.233	0.000	I(0)/I(1)
PROD	−0.337	0.915	−9.804	0.000	I(1)
VOL	−4.505	0.003	-	-	I(0)

below illustrates, is applicable here. It also enables the distinction between long- and short-run effects of derivatives on economic growth.

The ARDL bounds test consists of two steps. First, the presence or absence of a cointegrating relationship between the variables needs to be established. This involves testing the null hypothesis of no cointegration

$$H_0:{\theta }_0={\theta }_1={\theta }_2=\dots ={\theta }_j=0$$

against the alternative hypothesis

$$H_1:{\theta }_0\ne {\theta }_1\ne {\theta }_2\ne \dots \ne {\theta }_j\ne 0$$

based on the following regression framework which is then used to examine the theoretical specification:

$$\begin{array}{c}\hfill \Delta Y_t=\beta +\sum _{i=1}^n{\beta }_{0i}\Delta Y_{t-i}+\sum _{i=1}^n{\beta }_{1i}\Delta DE{R}_{t-i}+\sum _{i=1}^n{\beta }_{2i}\Delta IN{V}_{t-i}+\sum _{i=1}^n{\beta }_{3i}\Delta OPE{N}_{t-i}\\ \hfill +\sum _{i=1}^n{\beta }_{4i}\Delta INF{L}_{t-i}+\sum _{i=1}^n{\beta }_{5i}\Delta PRO{D}_{t-i}+{\theta }_0{Y}_{t-i}+{\theta }_{1}DE{R}_{t-i}+{\theta }_{2}IN{V}_{t-i}\\ \hfill +{\theta }_{3}OPE{N}_{t-i}+{\theta }_{4}INF{L}_{t-i}+{\theta }_{5}PRO{D}_{t-i}+{\epsilon }_t\end{array}$$

(1)

The dependent variable Y represents economic growth, and DER, INV, OPEN, INFL and PROD represent the explanatory variables as described in Table 1. ${\beta }_{ji}$ and ${\theta }_{ji}$ represent, respectively, the short- and long-term coefficients of the explanatory variables. $\Delta$ is the first difference operator, and ${\epsilon }_t$ is the residual error term. The long-term coefficients have to be jointly statistically significant (i.e., cointegration has to be present) to ensure that the model can be reliably estimated.

Each variable’s optimal lag (i) is determined by simulating various lag length specifications and then choosing the combination that yields the minimum Akaike information criterion (AIC). The cointegration hypothesis is then evaluated by comparing the model’s calculated F-statistic against Pesaran et al.’s (2001) critical values. For a given confidence interval and sample size these critical values contain clear upper and lower bounds. If the F-statistic exceeds the upper bound the null hypothesis can be rejected and cointegration can be established. If the F-statistic is smaller than the lower bound the null hypothesis cannot be rejected, while if the F-statistic falls within the bounds the test is inconclusive.

If cointegration is established, the second step is to estimate a short-term error-correction model (ECM). The ECM determines the model’s dynamics around its long-term trend, and can be represented as follows:

$$\begin{array}{c}\hfill \Delta Y_t=\beta +\sum _{i=1}^n{\beta }_{0i}\Delta Y_{t-i}+\sum _{i=1}^n{\beta }_{1i}\Delta DE{R}_{t-i}+\sum _{i=1}^n{\beta }_{2i}\Delta IN{V}_{t-i}+\sum _{i=1}^n{\beta }_{3i}\Delta OPE{N}_{t-i}\\ \hfill +\sum _{i=1}^n{\beta }_{4i}\Delta INF{L}_{t-i}+\sum _{i=1}^n{\beta }_{5i}\Delta PRO{D}_{t-i}+\delta {\epsilon }_{t-1}+{\mu }_t\end{array}$$

(2)

where ${\epsilon }_{t-1}$ denotes the lagged error-correction term from the long run specification (Equation (1)) and $\delta$ represents the cointegration coefficient. ${\mu }_t$ is the ECM’s residual error term.

3.3. Economic Growth Volatility

Following Mills (2019), a generalised auto-regressive conditional heteroskedasticity (GARCH) model is applied to the real GDP growth series to extract the time-varying volatility of economic growth. The GARCH(1,1) model is represented as

$${\sigma }_t^2=\omega +\alpha {\epsilon }_{t-1}^2+\beta {\sigma }_{t-1}^2$$

(3)

where ${\sigma }_t^2$ is the conditional variance of economic growth at time t. $\omega$ is a constant, and $\alpha$ and $\beta$ represent, respectively, the auto-regressive and moving average parameters. ${\epsilon }_{t-1}^2$ is the lagged squared error term, capturing ‘news’ about volatility in the previous period, while ${\sigma }_{t-1}^2$ is the forecasted variance from the previous period.

Figure 2 shows evidence of volatility clustering during episodes of uncertainty, notably around South Africa’s political transition during the early 1990s and the GFC. It is also noticeable that growth volatility is higher around periods where derivative market behaviour may have changed (see Figure 1): 1992–1994 saw a sharp growth in derivatives trading, while derivative market activity slowed markedly in 2008–2009 during the GFC. The economic growth volatility during 2020 and 2021 resulting from the COVID-19 shock dwarfs even the significant volatility observed during the GFC, substantiating the choice to exclude the COVID-19 data points from the sample.

3.4. Granger Causality

Granger causality implies that past values of a series X provide useful information in predicting future values of Y, beyond that which could be predicted using the past values of Y alone. The Granger causality test can be summarised by the following regression specification:

$$\begin{array}{c}\hfill \Delta Y_t={\alpha }_1+\sum _{i=1}^n{\beta }_{11}\Delta Y_{t-i}+\sum _{i=1}^n{\beta }_{12}\Delta X_{t-i}+{\mu }_t\\ \hfill \Delta X_t={\alpha }_2+\sum _{i=1}^n{\beta }_{21}\Delta X_{t-i}+\sum _{i=1}^n{\beta }_{22}\Delta Y_{t-i}+{\upsilon }_t\end{array}$$

(4)

where in this context X and Y represent, respectively, derivatives and economic growth. Granger causality between derivatives and both economic growth and the level of output, as well as economic growth volatility, are evaluated below.

4. Results and Discussion

4.1. Stationarity Tests

Unit root tests are presented in Table 2. The majority of the variables—real GDP, derivatives, capital investment, trade openness, and productivity—are non-stationary in levels, that is, the series are integrated of order one, and are stationary in first differences. Economic growth volatility—VOL, measured as the series of conditional variances (${\sigma }_t^2$) of real GDP growth, derived from the GARCH(1,1) model (Equation (3))—is stationary, i.e., $I\left(0\right)$. However, GDP growth and inflation can be interpreted as stationary at the 10% confidence level under both the ADF and PP unit root tests. The ambiguous evidence on the orders of integration on GDP growth and inflation, coupled with the fact that the variables under consideration are of mixed order of integration, substantiates the choice of the ARDL paradigm (Shrestha and Bhatta 2018).

4.2. The Impact of Derivatives on Growth

Table 3. Bounds test for cointegration—Real GDP growth.

Dependent Variable:	Real GDP Growth
F-Statistic:	6.89
	10%		5%		1%
Critical values	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
(Pesaran et al. 2001)	2.08	3.00	2.39	3.38	3.06	4.15

presents the results of the ARDL bounds test evaluating the impact of derivatives on economic growth. The calculated F-statistic of 6.89 exceeds the critical value of Pesaran et al.’s (2001) upper bound of 4.15, confirming that the variables in the model are cointegrated.

Having established cointegration, the ARDL model can be estimated. The optimal lag lengths are found at the combination that yields the minimum AIC, resulting in the choice of an ARDL(6,0,2,2,1,2) specification. Since quarterly data are utilised, this implies that contemporaneous growth rates can be influenced by its past values up to one-and-a-half years prior. The long and short run effect of derivatives on real GDP growth are reported in

Table 4. The long and short run effect of derivatives on real GDP growth.

Regressor	Coefficient	Std.Error	t-Statistic	Prob.
Long run estimates
DER	0.067 *	0.036	1.862	0.065
INV(−1)	−0.051	0.090	−0.561	0.576
INFL(−1)	−0.580 ***	0.207	−2.803	0.006
OPEN (−1)	−0.371	0.247	−1.499	0.137
PROD(−1)	−0.002 *	0.001	−1.793	0.076
C	0.329	0.403	0.818	0.415
Short run estimates
$\Delta Y$(−1)	0.306 ***	0.083	3.687	0.000
$\Delta Y$(−2)	0.081	0.074	1.090	0.278
$\Delta Y$(−3)	0.095	0.074	1.271	0.207
$\Delta Y$(−4)	−0.420 ***	0.074	−5.687	0.000
$\Delta Y$(−5)	0.162 **	0.078	2.078	0.040
$\Delta$INV	0.144 **	0.056	2.563	0.012
$\Delta$INV(−1)	0.109 *	0.061	1.796	0.076
$\Delta$INFL	−0.068 *	0.041	−1.675	0.097
$\Delta$INFL(−1)	0.132 ***	0.043	3.064	0.003
$\Delta$OPEN	0.032	0.031	1.042	0.300
$\Delta$PROD	0.003 ***	0.001	3.390	0.001
$\Delta$PROD(−1)	−0.001	0.001	−1.556	0.123
${\epsilon }_{t-1}$	−0.204 ***	0.029	−7.172	0.000

Dependent variable is real GDP growth. ARDL(6,0,2,2,1,2). ***, ** and * represent significance at 1%, 5% or 10%. Residual diagnostics confirming paramater stability are presented in Figure A1.

.

The estimated cointegration coefficient ${\epsilon }_{t-1}$ is statistically significant and falls between −1 and 0; the model is therefore dynamically stable (Loayza and Ranciere 2006, p. 1059). Economic growth is highly path dependent, as evidenced by the statistical significance of its own lags in the short run estimates.

Derivatives have a statistically significant positive long run relationship with economic growth at the 10% confidence level. This result is consistent with much of the empirical literature on the finance-growth nexus, which finds that finance, financial development and financial innovation have a positive impact on economic growth. In the context of derivatives as a specific type of financial innovation, these results are consistent with Samarakoon et al. (2024), Lema and Grandes (2020), Vo et al. (2020), Vo et al. (2019), Şendeniz-Yüncü et al. (2018), Aali-Bujari et al. (2016), Rodrigues et al. (2012), and Şendeniz-Yüncü et al. (2007), who all find a positive effect of derivatives on economic growth.

The positive long-term effect of derivatives on growth suggest that exchange-traded futures derivatives have a long run structural impact on South African economic growth. Moreover, the absence of a short-term effect substantiates the theoretical arguments that increased derivatives trading improves the long-term efficiency of financial markets. These particular derivative instruments improve risk management, assist in price discovery, and ameliorate information asymmetries. These all represent long-term structural improvements to financial markets, which in turn contribute to improved financial resource allocation and the lowering of transaction costs, ultimately enhancing productive activity and economic growth. It should, however, be acknowledged that this analysis considers exchange-traded futures derivatives only, which make up only a portion of total derivatives market activity, and also does not consider the more risky OTC derivatives.

Inflation negatively influences economic growth in the long run. The positive short-term impact of inflation suggests that a byproduct of faster economic growth may be temporarily higher inflation. Capital investment and trade openness are not significant drivers of long-term economic growth. Capital investment, however, has a positive relationship with economic growth in the short run, whereas trade openness plays no role in the short run.

The negative and significant long run coefficient on labour productivity is somewhat surprising. However, the magnitude of productivity’s long run impact is very small, while, in the short run, it has a relatively larger, significant and still positive impact on growth. This result may suggest that the South African economy is becoming ever more capital intensive, and therefore less sensitive to labour productivity. This is consistent with a growing literature on the increasing capital intensity of South African economic activity (CDE 2019; Mkhize 2019). Interrogating this dynamic further, however, falls beyond the scope of this paper.

4.3. The Impact of Derivatives on the Level of Output

A second ARDL specification, now with the level of real GDP as dependent variable, is estimated. This will arguably improve the robustness of the econometric results, while it enables the distinction between derivatives’ effects on growth rates and levels of output.

Table 5. Bounds test for cointegration—Real GDP.

Dependent Variable:	Real GDP
F-Statistic:	7.22
	10%		5%		1%
Critical values	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
(Pesaran et al. 2001)	2.08	3.00	2.39	3.38	3.06	4.15

and

Table 6. The long and short run effect of derivatives on real GDP.

Regressor	Coefficient	Std.Error	t-Statistic	Prob.
Long run estimates
DER	0.037 *	0.021	1.762	0.081
INV(−1)	0.193 ***	0.070	2.743	0.007
INFL(−1)	−0.471 **	0.237	−1.990	0.049
OPEN(−1)	−0.212	0.151	−1.403	0.163
PROD(−1)	0.003 ***	0.001	3.186	0.002
C	5.108 ***	0.309	16.536	0.000
Short run estimates
$\Delta$Y(−1)	0.296 ***	0.078	3.794	0.000
$\Delta$INFL	0.040 **	0.017	2.326	0.022
$\Delta$INFL	−0.011	0.013	−0.898	0.371
$\Delta$INFL(−1)	0.040 ***	0.013	3.062	0.003
$\Delta$OPEN	0.014	0.009	1.551	0.124
$\Delta$PROD	0.001 ***	0.000	3.982	0.000
$\Delta$PROD(−1)	−0.001 **	0.000	−2.011	0.047
${\epsilon }_{t-1}$	−0.073 ***	0.010	−7.310	0.000

Dependent variable is real GDP. ARDL(2,0,1,2,1,2). ***, ** and * represent significance at 1%, 5% or 10%. Residual diagnostics confirming paramater stability are presented in Figure A2.

present, respectively, the ARDL bounds test and the long and short run estimates of the impact of derivatives on the level of output. Again, cointegration is established as the calculated F-statistic of 7.22 exceeds the critical value of Pesaran et al.’s (2001) upper bound of 4.15.

Consistent with the positive relationship between derivatives and economic growth established above, derivatives also have a statistically significant positive long run relationship with the level of real GDP (Table 6). Capital investment and labour productivity are significant positive drivers of long-term output, while inflation drags down output in the long run.

Similar to the absence of a short-term relationship between derivatives and growth (Table 5) no short-term effect of derivatives on the level of output is detected, again emphasising the long-term nature of the channels between derivatives, financial market efficiencies and economic activity theorised above. Positive relationships are detected between inflation and output and productivity and output in the short run. Finally, the coefficient of the error-correction term is negative and statistically significant, with $-1<{\epsilon }_{t-1}<0$.

4.4. Derivatives and Growth Volatility

The ARDL bounds test procedure is employed again, this time with the growth volatility series derived from the GARCH(1,1) model (Equation (3)) as dependent variable. Inflation and trade openness, which could potentially amplify economic shocks or volatility, are included as control variables.

Table 7. Bounds test for cointegration—GDP growth volatility.

Dependent Variable:	GDP Growth Volatility
F-Statistic:	5.02
	10%		5%		1%
Critical values	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
(Pesaran et al. 2001)	2.37	3.20	2.79	3.67	3.65	4.66

confirms that the variables in the volatility model are cointegrated. The long and short run estimates are presented in

Table 8. The long and short run effect of derivatives on GDP growth volatility.

Regressor	Coefficient	Std.Error	t-Statistic	Prob.
Long run estimates
DER(−1)	0.000	0.000	0.544	0.588
INFL(−1)	0.001 **	0.000	2.431	0.017
OPEN(−1)	0.000	0.000	0.288	0.774
C	0.000	0.000	−0.959	0.340
Short run estimates
$\Delta Y$(−1)	−0.011	0.104	−0.107	0.915
$\Delta Y$(−2)	0.133	0.104	1.282	0.203
$\Delta Y$(−3)	−0.035	0.104	−0.339	0.735
$\Delta Y$(−4)	0.106	0.104	1.026	0.308
$\Delta Y$(−5)	0.292 ***	0.099	2.930	0.004
$\Delta$DER	0.000	0.000	1.405	0.164
$\Delta$DER(−1) **	0.000	0.000	2.391	0.019
$\Delta$DER(−2)	0.000 **	0.000	−2.303	0.024
$\Delta$INFL	0.000	0.000	−0.982	0.329
$\Delta$INFL(−1)	−0.001 **	0.000	−2.174	0.032
$\Delta$INFL(−2)	0.000	0.000	−0.170	0.865
$\Delta$INFL(−3)	−0.001 ***	0.000	−2.817	0.006
$\Delta$INFL(−4)	−0.001 ***	0.000	−2.715	0.008
$\Delta$OPEN	0.000	0.000	−0.079	0.937
$\Delta$OPEN(−1)	0.000	0.000	−0.805	0.423
$\Delta$OPEN(−2)	0.000	0.000	0.956	0.341
$\Delta$OPEN(−3)	0.000 **	0.000	2.445	0.016
${\epsilon }_{t-1}$	−0.428 ***	0.084	−5.125	0.000

Dependent variable is real GDP growth volatility. ARDL(6,3,5,4). *** ** represent significance at 1% or 5%. Residual diagnostics confirming paramater stability are presented in Figure A3.

, and indicate that exchange-traded futures derivatives have no impact on growth volatility whatsoever. Although some of the explanatory variables are statistically significant, notably inflation, the size of the coefficients are so small that their impact on volatility is virtually indistinct from zero. It would appear that past growth volatility shocks are the significant drivers of contemporaneous volatility, which supports the notion of volatility clustering observed above (Figure 2).

As a robustness check on the influence of derivatives on economic growth volatility, Equation (3) is augmented to the following multivariate GARCH (MGARCH) specification:

$${\sigma }_t^2=\omega +\alpha {\epsilon }_{t-1}^2+\beta {\sigma }_{t-1}^2+DE{R}_t$$

(5)

The estimates from Equation (5) are presented in

Table 9. MGARCH(1,1) volatility estimation.

Term	Coefficient	Std.Error	z-Statistic	Prob.
C	0.000	0.000	0.838	0.402
${\epsilon }_{t-1}^2$	0.358	0.181	1.982	0.047
${\sigma }_{t-1}^2$	0.450	0.204	2.206	0.027
DER	0.000	0.000	−0.592	0.554

. The result is consistent with the result from Table 8, confirming that exchange-traded derivatives have no impact on the volatility of South African economic growth. Therefore, this specific class of derivatives seemingly does not introduce risk-taking behaviour, which could manifest in increased economic growth volatility. This could tentatively support the argument that formal exchange-traded derivatives introduce less volatility compared to OTC derivatives, as was discussed in Section 2.1.3 above.2 The sum of the ARCH and GARCH coefficients, $\alpha +\beta =0.808$, suggests relatively high persistence in economic growth volatility, implying that economic shocks can take a long time to dissipate. This is consistent with Table 8, which shows that volatility shocks from up to five quarters ago can still influence contemporaneous volatility, and the volatility clustering observed in Figure 2.

From the empirical literature, Bekale et al. (2015) find that the establishment of a formal derivatives exchange has lowered South African growth volatility, concluding that exchange-traded derivatives have had a stabilising effect on the economy. Similarly, Rodrigues et al. (2012) find that the establishment of derivatives exchanges can lower GDP growth volatility. However, the econometric evidence here finds no evidence of any relationship whatsoever—neither stabilising, nor exacerbating—between exchange-traded futures derivatives and the volatility of economic growth. Unlike these authors, we therefore cannot draw the strong conclusion that exchange-traded futures derivatives have stabilised South African economic growth volatility. However, we can conclude that exchange-traded futures have not contributed to higher economic growth volatility.

4.5. Granger Causality Tests

The preceding ARDL analyses provide evidence of a positive and significant relationship between derivatives trading and economic activity. Granger causality tests are next employed to further evaluate the strength and significance of the relationships between these variables. The calculated F-statistics are evaluated against the null hypothesis that series X does not Granger-cause series Y.

The results of the Granger causality tests support the central hypothesis of this study.

Table 10. Granger causality tests: Derivatives and economic activity.

Relationship	Direction	F-Statistic	Prob.	Result
Growth	DER → GDPGR	2.357	0.016	Reject $H_0$. DER Granger-causes GDPGR.
	GDPGR → DER	0.957	0.487	Cannot reject $H_0$.
Real GDP	DER → GDP	5.966	0.004	Reject $H_0$. DER Granger-causes GDP.
	GDP → DER	1.308	0.275	Cannot reject $H_0$.
Volatility	VOL → GDP	1.741	0.180	Cannot reject $H_0$.
	GDP → VOL	0.278	0.758	Cannot reject $H_0$.

GDPGR—economic growth; GDP—level of output; VOL—economic growth volatility. Derivation of the growth volatility series was described in Section 3.3.

suggests unidirectional causality from exchange-traded futures derivatives to both measures of economic activity—growth and output. This result for the South African economy is consistent with Şendeniz-Yüncü et al. (2018)’s broader confirmation of the supply-leading hypothesis for middle-income economies.

These results, however, contrast to those of Marozva (2014), whose own Granger causality tests found no causality between derivatives and South African economic growth, and Bekale et al. (2015), who find causality in the opposite direction from growth to derivatives. However, Marozva (2014) and Bekale et al. (2015) both utilise annual data over different samples, and define their derivative proxies differently: The former utilises the trading volume of all derivatives for 1994–2012, while the latter employ a dummy variable approach for the existence/absence of a formal domestic derivatives exchange between 1971–2012. Therefore, the contrasting results may suggest that the direction of causality detected might be sensitive to the research paradigm and data.

4.6. Summary of Results

The results from the empirical analyses are largely consistent with the theoretical and empirical literature. Derivatives may contribute to a more efficient financial system by improving risk management, price discovery, and resource allocation, thereby supporting productive economic activity and growth. The ARDL analyses establish a positive long-term relationship between exchange-traded futures derivatives and both economic growth and the level of output in South Africa. This echoes the results from Şendeniz-Yüncü et al. (2007), Rodrigues et al. (2012), Aali-Bujari et al. (2016), Vo et al. (2019) and Samarakoon et al. (2024). Furthermore, Granger causality from derivatives to both GDP growth and the level of output is detected. This is consistent with Şendeniz-Yüncü et al. (2018)’s result that the supply-leading hypothesis holds for middle-income countries. While one cannot conclude that derivatives definitively cause growth or output, however, this paper finds evidence of derivatives leading South African growth and output over the sample period 1990–2019. This supports the notion that “well-functioning financial market development can promote economic growth” (Ncanywa and Mabusela 2019, p. 1), giving credence to the supply-leading hypothesis in the South African economy.

This paper finds no evidence suggesting that derivatives exacerbate or mitigate volatility in South African economic growth. The Granger causality tests support the results obtained in Section 4.4, confirming that exchange-traded futures derivatives do not influence economic growth volatility. However, an important caveat is that these results originate from activity in regulated exchange-traded futures derivative markets only. Due to data availability, this paper considers exchange-traded futures only, and not other exchange-traded derivatives such as swaps and options, thus encompassing only one segment of total exchange-traded derivatives activity in South Africa. Moreover, unregulated OTC derivative markets—which make out a substantial portion of derivatives traded globally—are not evaluated here as regular data on these markets are not formally published. This delimitation could bias the results away from riskiness and potential growth volatility which might be introduced by OTC derivatives. However, exchange-traded derivatives, the focus of this paper, arguably introduce substantially less risk and volatility than OTC derivatives do, by virtue of there being more regulation, oversight, standardisation and transparency in exchange-traded derivative markets. The impact of OTC derivatives on economic growth and volatility remains an important area for future research, provided adequate data on OTC markets can be obtained.

5. Conclusions and Policy Implications

This paper considers the impact that exchange-traded derivative market activity could have on South African economic growth and output. Derivatives, as a specific measure of financial innovation, theoretically influence economic growth and production through its role in managing risk, contributing to price discovery, and generally improving financial market efficiency. A body of literature subsequently argues that derivatives could positively affect economic growth, and that derivatives may, under certain conditions, contribute to unlocking economic growth.

Financial innovation is measured here by the underlying value of futures derivatives domestically traded on exchange. The econometric analyses find a statistically significant positive effect of exchange-traded futures derivatives on South African economic growth and the level of output. Unlike Robinson’s classic adage that finance follows where industry leads, financial innovation through derivatives is shown to, in fact, lead and potentially drive economic activity in the South African context. The significant positive impact of finance on economic growth detected here suggests that the supply-leading hypothesis is indeed applicable to South Africa, and that financial innovation and development can be an important driver of domestic economic growth.

These results have some important policy implications. Considering the positive impact of exchange-traded futures derivatives on economic growth, policies promoting and strengthening exchange-traded derivative markets may stimulate domestic economic growth. Therefore, strategies for enhancing and promoting further development of formal derivatives exchanges should be encouraged. However, even though we find no evidence of exchange-traded derivatives exacerbating domestic growth volatility, it remains important to establish and maintain a strong regulatory framework to avoid potential adverse consequences of increased derivatives trading, especially in unregulated OTC markets.

These results may be unique to South Africa, as a small emerging economy with a sophisticated financial infrastructure, including a well-developed banking sector and stock market, deep capital markets and relatively strong regulatory frameworks. While the evidence from the empirical literature on the relationship between derivatives markets and economic growth in the emerging economy context is mixed, the observed muted response of growth to derivatives detected in the literature might be due to the small size of derivatives markets in these economies, or their comparatively less developed financial systems; South Africa may therefore be an outlier in this regard. Further research on an individual-country basis would be needed to determine whether this paper’s main result—that organised derivatives market activity could enhance economic growth—can be generalised to other emerging economies, and the extent to which derivatives markets may contribute to economic growth and development elsewhere.