Managing earnings volatility is an important aspect of firm value for several reasons. Earnings volatility reflects an inherent business risk that can arise from the firm’s operations as well as the result of market shocks (Ghasemzadeh, Heydari & Mansourfar 2021). Firms manage earnings volatility because earnings volatility can enhance a firm’s cost of capital (Minton, Schrand & Walther 2002), increase the possibility of the violation of debt covenants and cost of financial distress (DeFond & Jiambalvo 1994; Dichev & Skinner 2002; Smith & Stulz 1985), create agency problems (Morellec 2004; Stulz 1990), lead to underinvestment problems (Froot, Scharfstein & Stein 1993), lead to information disparity between informed and uninformed investors (Goel & Thakor 2003), and can increase the possibility of missing earnings targets which is associated with negative stock market reaction (Barton 2001; Kothari 2001).

The foundation of the modern financial theory is based on three premises: value creation, the risk versus return trade-off, and the no-arbitrage principle (Fatemi & Luft 2002). The firm value creation can be attributed to one of three sources: reducing the cost of financial distress (Smith & Stulz 1985), efficient and optimal tax payment (Graham & Smith 1999; Smith & Stulz 1985), and mitigating the effect of credit rationing by reducing the possibility that the firm may be forced to forego positive net present value projects because of insufficient internal funds (Froot et al. 1993). Alternatively, managerial risk aversion is based on the agency model (Tufano 1998).

Today, across the globe, most non-financial firms use derivatives; however, to determine whether the intent is hedging or timing the market (speculating) is not always obvious (Bartram 2019). Researchers across the globe have examined the use of derivatives by non-financial firms and the use of derivatives varies across countries. For instance, Bodnar and Gebhardt (1999) performed a comparative survey study on derivatives use in risk management by United States (US) and German non-financial firms. The study found that more German firms (78%) used derivatives than US firms (57%). The use of derivatives was for hedging purposes; however, firms in Germany hedged with derivatives to manage accounting earnings volatility, whereas US firms used derivatives to manage cash flow fluctuations. Based on the analysis of a survey conducted in 1996 on the sample data of 77 Canadian non-financial firms, Jalilvand (1999) found that 75% of the firms in the survey used derivatives for risk management. The study also found that companies with global presence engaged in derivatives.

Several studies have been undertaken across the European continent. Fatemi and Glaum (2000) presented a survey study of corporate hedging behaviour of non-financial firms listed on the Frankfurt Stock Exchange. The overall sample size included 71 firms and 89% indicated that they used derivatives for risk management. This rate is higher than the 78% reported by Bodnar and Gebhardt (1999). In Scandinavia, Hagelin (2003) examined 160 Swedish non-financial firms’ use of currency derivatives. Unlike Fatemi and Glaum (2000), this study adopted a combination of survey data and financial statements. The study found that Swedish firms hedged currency risk with derivatives. El-Masry (2006) surveyed 173 United Kingdom (UK) non-financial firms in 2001 and found that 67% of UK-based firms in the sample hedged with derivatives. The primary reasons for not using derivatives were cited as inadequate financial risk exposure and costs in setting up derivatives programmes.

Bodnar et al. (2013) conducted a web-based survey on risk management practices and use of derivatives by Italian non-financial firms. The analysis was on a sample of 86 firms during the period September 2007 to January 2008. The study found that Italian firms used derivatives to hedge foreign currency risk. Further, similar to Bodnar and Gebhardt (1999), some studies adopted an international perspective and comparison across countries. Examining 155 firms using a survey approach, Prevost, Rose and Miller (2000) found that derivatives use and trends in small open economies such as New Zealand were comparable and similar to more developed countries such as the UK, the US and Germany. Pramborg (2005) undertook a comparative survey study on the use of derivatives by Swedish and Korean non-financial firms. The sample was constituted of 163 firms, 60 from Korea and 103 from Sweden. The study found that firms in Sweden hedged with derivatives to manage earnings volatility and Korean firms primarily used derivatives to manage cash flow fluctuations. Bartram, Brown and Fehle (2009) conducted the first comprehensive global examination of risk management practices among non-financial firms. Annual reports were used as the source of data in this study. The overall sample included 7292 firms across 48 countries including the US. The results showed that 60% of the firms used derivatives for hedging purposes.

Outside of the US and Europe, similar studies have been undertaken in New Zealand and Australia. Extracting data from the 2007 annual reports of 134 non-financial firms listed on the New Zealand Stock Exchange, Li, Visaltanachoti and Luo (2014) examined corporate benefits of derivatives use. The study found no evidence that corporate hedging with derivatives is associated with an increase in firm value measured by Tobin’s Q. Nguyen and Faff (2002) investigated the determinants of derivatives use among Australian listed firms. The study included a sample of 469 firm observations between 1999 and 2000. The study found that liquidity, size, and leverage are the determinants of derivatives use among Australian firms. Firms also used derivatives to reduce expected cost of financial distress and cash flow volatility.

Research on corporate derivatives use in emerging markets has gained attraction in recent years. However, research that has examined corporate hedging strategies in emerging markets has documented relatively low use of derivatives. Accordingly, Martin et al. (2009) adopted a survey method to examine corporate derivatives use by Peruvian non-financial firms. The analysis included 65 non-financial firms and was carried out during 2005. The study found that only a small fraction (33%) of the firms in the sample engaged in derivatives. This number is relatively low compared to developed markets studies. During the period 2004–2007, Afza and Alam (2011) examined the determinants of foreign currency derivatives among non-financial firms in Pakistan. The annual reports were used as the source of firm data of 86 non-financial firms listed on the Karachi Stock Exchange. The study found that firms in Pakistan hedged currency risk with derivatives. In South Africa, Correia, Holman and Jahreskog (2012) conducted a survey on derivatives use by 98 non-financial firms listed on the JSE in 2006. The study found that South African firms hedged with derivatives. In addition, as in some emerging markets such as Peru, high costs of setting up derivatives programmes were cited by South African firms as one of the reasons that hinder derivatives use.

Recently, there has been a growing interest in research that focuses on the relationship between corporate derivatives use and earnings volatility. Barton (2001) examined whether the use of derivatives and discretionary accruals can be used as alternatives to manage earnings volatility. The sample included Fortune 500 firms between 1994 and 1996 and the study found that firms that engaged in derivatives had lower earnings volatility. Similarly to Barton (2001), Pincus and Rajgopal (2002) examined US-based firms focusing on the Oil and Gas sector. The study also used annual reports data on a sample of 236 firms. The results showed that the choice of discretionary accrual and derivatives use is associated with lower earnings volatility. In contrast, Zhang (2009), found no evidence that suggests that derivatives use is associated with changes in earnings volatility. However, even though the time period in the analysis is similar to the Barton (2001) study, the sample size is much smaller. Beneda (2013) used a regression model to examine the use of derivatives by US non-financial firms over the period 2003–2010. Similarly to Barton (2001), and Pincus and Rajgopal (2002), the study found the use of derivatives by US non-financial firms is negatively associated with earnings volatility. Abdel-Khalik and Chen (2015) point out that the earlier mixed results can be attributed to the use of small total derivatives amounts or to insufficient derivatives after the implementation of SFAS No. 133.

Among Canadian firms, the evidence also showed mixed findings. Jalilvand (1999) examined survey data of 77 non-financial listed Canadian firms and found that corporate hedgers had higher earnings volatility compared to non-hedgers. Paligorova and Staskow (2014) found that the derivatives use by Canadian firms is associated with lower earnings volatility between 2005 and 2013. This contradicts the result of Jalilvand (1999). More recently, in an emerging market, Phua et al. (2021) examined the association between derivatives use and earnings volatility. The study found that derivatives use by Malaysian non-financial firms is positively associated with earnings volatility, suggesting that derivatives use did not lower earnings volatility as expected. These results did not concur with similar studies in the US and other developed parts of the world. Possible reasons for this are that the derivatives market in emerging markets remains relatively small compared to that in developed markets.

From the empirical evidence presented above, the theoretical framework states that firms primarily engage in derivatives to manage risks. This suggests that corporate hedging with derivatives can lead to lower earnings volatility (Beneda 2013; Smith & Stulz 1985). Previous research on the effect of derivatives use on earnings volatility has focused primarily on developed markets such as the US and Canada. However, findings from these studies offer mixed results, highlighting the need for further investigation, particularly in emerging markets such as South Africa. Based on the empirical findings and theoretical framework, the following hypothesis is developed:

The present study examined a sample of 135 firms listed on the JSE. Financial firms, such as insurance firms, banks, investment banks, and asset managers, were excluded because the intent of derivatives use could be profit-making rather than risk management (Batten & Hettihewa 2007). The research period covers the years 2005–2021. All the variables used in this study were collected from this period. In addition, derivatives data were available in the annual reports for the period chosen in this study. The data used in this study are firm-level data obtained from financial data depositories, Iress and Thomson Reuters Datastream, and annual financial statements. Previous studies relied on the survey questionnaire method to capture data on derivatives use. However, the International Financial Reporting Standards (IFRS) prescribes disclosure requirements on both quantitative and qualitative information regarding financial instruments such as derivatives. Hence, in this study, derivatives data were sourced from the annual reports using the Thomson Reuters Datastream. Relying on the annual reports for derivatives data is a better alternative to a survey method that places much reliance on the participatory rate.

The data used in this study can be characterised as panel data. It includes firm observations over a period of 17 years. As there were some missing data points over the sample period and across firms, an unbalanced panel design was adopted. Eviews11, a statistical package which allows for statistical analysis of panel data, was used. The analysis outcomes were descriptive statistics, Pearson correlation coefficients and panel regression analysis.

Regression modelling was conducted in Eviews11. In regression modelling, there are numerous assumptions regarding the model, namely autocorrelation, linearity, multicollinearity and homoskedasticity. Before determining the appropriate regression model, tests were performed to determine a statistically valid regression model for this study. The Hausman test was used to determine if a random or fixed effect model applied for this study. To account for the presence of autocorrelation and homoskedasticity, the panel estimated generalised least squares method (period seemingly unrelated regression) was suitable for this study instead of an ordinary least squares (OLS), random or fixed effect model. A white diagonal standard errors and covariance which is a robust standard error estimation was applied and thus ensured that the significant values were not influenced by heteroskedasticity.

It is common to have outliers in the financial data that can influence the results. Therefore, testing for normality is important. Normality test showed a skewness value which is within ±2 and is acceptable. Kurtosis results were also within an acceptable range of ±7. Where applicable, the presence of skewness and kurtosis were addressed by winsorisation. Because autocorrelation in a panel data set biases the standard errors and causes the result to be less efficient, this study used the Durbin Watson test to check for autocorrelation. The Durbin Watson test is between the acceptable thresholds of 1.5 and 2.5 and therefore indicates that autocorrelation was addressed. A Pearson correlation matrix was used to test for multicollinearity and no presence of multicollinearity was depicted.

In this section, descriptive statistics for the sample of non-financial firms included in this study are summarised and presented. The descriptive statistics of the variables are depicted in Table 1. The mean and standard deviation for earnings volatility (EARNVOL) were 0.050 and 0.036. This suggests that the earnings volatility of the sample firms in this study was not widely distributed. DEBTCAP as a measure of interest debt level has a mean of 0.498, indicating that on average firms included in the sample finance their assets by 50% of debt. The minimum value of DEBTCAP was 0.144 while the maximum value was 0.930.

Figure 1 depicts the use and non-use of derivatives by firms across sample years 2005–2021. As shown in Figure 1, more firms in the sample years did use derivatives.

Table 2 and Table 3 presents the Pearson correlation matrix for variables in this study. Correlation analysis examines the relationship between variables; however, correlation does not imply causation. Table 2 and Table 3 also show no evidence of multicollinearity between variables.

From Table 2, it can be observed that there is a statistically significant weak negative relationship between earnings volatility (EARNVOL) and derivatives use (DERDUM). Table 2 also shows a correlation between earnings volatility (EARNVOL) and the control variables. It can be observed that there is a statistically significant negative weak relationship between earning volatility and three control variables, interest-bearing debt level (DEBTCAP), market-to-book ratio (MKBK) and firm size (LNSIZE), but a positive weak relationship is found between earnings volatility (EARNVOL) and research and development (RD). Correlation analysis was also conducted between derivatives use (DERDUM) and control variables. From Table 2, it can be observed that there is a statistically significant positive weak relationship between derivatives use (DERDUM) and two control variables, interest-bearing debt level (DEBTCAP) and firm size (LNSIZE). Table 2 also shows a positive weak relationship between derivatives use (DERDUM) and research and development expense (RD) and, lastly, a negative weak relationship was observed with market-to-book-ratio (MKBK).

A statistically significant positive relationship between the use of derivatives (DERDUM) and interest-bearing debt level (DEBTCAP) supports the financial distress argument for corporate hedging. The result indicates that the use of derivatives increases as debt levels on the balance sheet increases. Corporate risk management theory argues that hedging with derivatives can lower earnings volatility and generate sufficient internal cash to fund investment opportunities. Therefore, firms that have growth opportunities will use derivatives to manage earnings volatility. However, the results of this study do not support that argument shown by a statistically negative weak relationship between derivatives use (DERDUM) and market-to-book ratio (MKBK). The reason for a negative relationship in this study could be that firms included in this study are matured firms that generate enough internal cash flow to fund growth opportunities. It is also expected for firms with high research and development costs to use derivatives (Froot et al. 1993). The present study found a statistically positive weak relationship between derivatives use (DERDUM) and research development (RD). Lastly, Table 2 shows a statistically significant positive weak relationship between derivatives use (DERDUM) and firm size (LNSIZE). The results indicate that larger firms in terms of their market capitalisation listed on the JSE are more likely to engage in derivatives. Because of economies of scale, large firms can also set up and maintain derivatives programmes (El-Masry 2006).

From Table 3 it can be observed that there is a negative weak relationship between earnings volatility (EARNVOL) and derivatives use (TOTDER). From Table 3, it can also be observed that there is a statistically significant positive weak relationship between derivatives use (TOTDER) and two control variables, market-to-book ratio (MKBK) and firm size (LNSIZE). Table 3 also shows a negative weak relationship between derivatives use (TOTDER) and interest-bearing debt level (DEBTCAP) and, lastly, a negative weak relationship was observed with research and development expense (RD).

The more leveraged a firm is, the more likely the firm is to engage in derivatives. This is because highly leveraged are more sensitive to changes in interest rates and high earnings volatility may make it difficult to service debt obligations. The results of this study using derivatives as a continuous variable do not support this rationale shown by a negative weak relationship between total amount of derivatives (TOTDER) and interest-bearing debt level (DEBTCAP). Firms with high growth prospects are more likely to use derivatives to lower the probability of underinvestment and ensure availability of funds for growth opportunities (Géczy, Minton & Schrand 1997). The result of this study confirms this expectation shown by a statistically significant positive weak relationship between total amount of derivatives (TOTDER) and market-to-book-ratio (MKBK). The expected relationship between derivatives use and research and development as a proxy for growth is positive. The results from Table 3 do not confirm that expected relationship shown by a negative weak relationship between total amount of derivatives (TOTDER) and research and development (RD). Lastly, a statistically significant positive weak relationship was found between total amount of derivatives (TOTDER) and firm size (LNSIZE).

The results from the regression analysis models are presented in Table 4 and Table 5. The findings from Table 4 revealed a positive association between derivatives use measured by a binary value and earnings volatility. However, the association is not significant. The findings from Table 5 revealed a weak association between derivatives use measured as a continuous variable and earnings volatility as indicated by coefficient of 0.000.

The formulated hypothesis of this study is that effective hedging with derivatives should lower earnings volatility. Corporate risk management theory argues that non-financial firms engage in derivatives to manage earnings volatility since earnings volatility affects firm value. Earnings volatility affects firm value as it leads to costs of financial distress and bankruptcy (Smith & Stulz 1985), reliance on costly external financing (Froot et al. 1993), underinvestment problems and agency costs of debt (Myers 1977), and costs of managerial risk aversion (Fatemi & Luft 2002). Prior research, for example by Bodnar and Gebhardt (1999), Pramborg (2005), El-Masry (2006), Bartram et al. (2009), and Bodnar et al. (2013), cites that one of the primary reasons for corporate derivatives use is to manage earnings volatility.

The main findings of this study suggest that derivatives use (measured by a binary value) is positively but not significantly associated with earnings volatility. The findings of this study contradict the findings of Beneda (2013), Paligorova and Staskow (2014), and Abdel-Khalik and Chen (2015) who found that the use of derivatives is negatively associated with earnings volatility. However, the findings of this study confirm the findings of Jalilvand (1999) and Phua et al. (2021) who found a positive association between the use of derivatives and earnings volatility. When compared to studies that measured derivatives use by total notional amount, the main findings of this study suggest that derivatives use is positively, yet not significantly, associated with earnings volatility. The results of the current study contradict the findings of Barton (2001), and Pincus and Rajgopal (2002) who found a negative association between the use of derivatives and earnings volatility. However, the results of this study are similar to the findings by Choi, Mao and Upadhyay (2015) who found a positive association between derivatives use and earnings volatility.

The reasons why the present study’s results might be different to studies such as those by Beneda (2013), and Paligorova and Staskow (2014) is because this study examined the relationship from an emerging market’s perspective. Correia et al. (2012) found that high costs of setting up derivatives programmes were cited by South African firms as one of the reasons that hinder the use of derivatives. Other reasons cited were negative perceptions on derivatives by investors, strict regulatory environment as well as perceived credit risk (Correia et al. 2012). Therefore, this means that JSE-listed firms, as indicated by Barton (2001), might be using other alternatives such as discretionary accruals to manage earnings volatility. Adding to this, the present study examined the use of derivatives measured by a continuous variable, whereas previous studies examined corporate derivatives use by a dichotomous variable. Therefore, different types of derivatives instruments could have a different effect on earnings volatility. Future studies might pursue this avenue of research.

The present study rejects the null hypothesis based on the findings from the regression models. The findings suggest that derivatives use by JSE-listed firms had no impact on smoothing earnings volatility. However, these findings do not indicate that JSE-listed firms are ineffective in managing risks. Based on previous research, corporate derivatives use in risk management is effective in hedging currency risk and interest rate risk.

The limitations encountered in this study are associated with the availability of data and the sample period. For instance, to capture the use of derivatives, this study relied only on the information that was available in the financial statements. The period covered in this study included the global financial crisis. The global financial crisis was a period of extreme stress in global financial markets and could have created volatility in variables used in this study.

Further research may examine the impact structural breaks or changes in accounting standards on derivatives and earnings volatility. Future research could also examine how periods of heightened volatility like the global financial crisis impact the use of derivatives. This study measured earnings volatility over a five-year period; therefore, future research could investigate measuring earnings volatility over a two-year period. This time span might provide a better match for derivatives use. Finally, other variables that could influence earnings volatility can be included in future regression models.