Exchange rate exposure in the South African stock market before and during the COVID-19 pandemic

Abstract

We examine the nature of exchange rate exposure before and during the COVID-19 pandemic. Using a multifactor arbitrage pricing model and daily data from South Africa, we show that, as compared with sectors, industries have been more exposed to the exchange rate risk during than before the pandemic. We further show that exchange rate exposure mostly hurts the sectors and industries, although a few sectors, such as beverages, mining, personal goods, and tobacco, and industries, such as basic materials, consumer goods, and technology, may benefit from it. Our estimates survived robustness checks.

1. Introduction

In this paper, we examine the nature of exchange rate exposure before and during the COVID-19 pandemic. Our study focuses on the South African stock market and is motivated by the lack of literature on the influence of the COVID-19 pandemic on African financial markets. We focus on the South African stock market because it is the largest in Africa by market capitalization and is one of the largest in the world (see Schiereck et al., 2018).1 The South African currency, the rand, is one of the leading currencies in Africa and has experienced episodes of depreciation in recent times (Zerihun et al., 2020). The currency suffered even more during the pandemic. Figure 1 shows that the sharp depreciation in the rand during the pandemic is associated with a sharp decline in the composite stock price index of the Johannesburg stock exchange. Aside this, the country also has complete data on the economic factors at higher frequencies and thus offers the ideal setting to assess the dynamics of exchange rate exposure in Africa. Consistent with Figure 1, Table 1 shows that there are statistically significant differences in the means of stock prices, the exchange rate, and other key variables before and during the pandemic, further strengthening our motivation. In addition, exchange rate risk influences the value of firms (Dominguez and Tesar, 2006) and as a result firms allocate substantial resources to manage it (Jorion, 1991). However, empirical estimations of exchange rate exposure deliver fewer significant exposure coefficients, regardless of the approach or sample and have given rise to the “exchange rate exposure puzzle” (Bodnar and Bartram, 2007). Our analysis aligns with this literature, as we demonstrate that exposure, especially at the industry level, depends on extreme events like COVID-19.

Figure 1.

Exchange rate and stock price dynamics before and during COVID-19

The figure shows the rand/US dollar exchange rate and the stock price dynamics in South Africa before and during the COVID-19 pandemic. We observe a sharp depreciation of the exchange rate and a decline in the stock price index, following the COVID-19 outbreak in the country. The sample period is from 1 January 2019 to 19 November 2020.

Table 1.

Equality of means of stock prices, the exchange rate, and other variables

Variable	t-test	P-value
All Share index	12.713	0.000
Banks	74.276	0.000
Life Insurance	56.427	0.000
Auto and Parts	34.273	0.000
Beverages	39.863	0.000
Chemicals	37.564	0.000
Construction and Materials	45.790	0.000
Electronic and Electrical Equipment	63.472	0.000
Food and Drug Retailers	34.060	0.000
Food Producers	26.240	0.000
Forestry and Paper	10.107	0.000
Fixed Line Telecommunications	27.620	0.000
General Retailers	53.126	0.000
General Financials	61.362	0.000
General Industrials	65.220	0.000
Health Care Equipment and Services	33.157	0.000
Industrial Metals and Mining	-6.875	0.000
Industrial Transportation	30.403	0.000
Media	13.941	0.000
Mining	-15.919	0.000
Mobile Telecommunications	42.388	0.000
Nonlife Insurance	27.591	0.000
Personal Goods	1.344	0.180
Pharmaceuticals and Biotechnology	-6.878	0.000
Real Estate Investment Services	62.680	0.000
Real Estate Investment Trust	68.971	0.000
Software and Computer Services	-33.766	0.000
Support Services	24.314	0.000
Tobacco	-11.481	0.000
Travel and Leisure	62.577	0.000
Exchange rate (rand/US dollar)	-43.319	0.000
Oil price (Brent crude price)	43.130	0.000
Short-term interest rate	54.605	0.000

This table tests the equality of means of stock prices, the exchange rate, and other variables before and during COVID-19. There are clear differences in means before and during the pandemic, the exception being the stock price of the personal goods sector.

A new line of research shows that COVID-19 disrupted different financial markets. This line of research argues that the pandemic impacted, among others, oil markets (Devpura and Narayan, 2020; Iyke, 2020a; Prabheesh et al., 2020), stock markets (Mazur et al., 2020; Narayan et al., 2020; Topcu and Gulal, 2020), foreign exchange markets (Iyke, 2020b; Narayan, 2020), and insurance markets (Wang et al., 2020). Our study is closely related to this line of research. Unlike this literature, we connect foreign exchange markets to stock markets via exchange rate exposure before and during the pandemic. We are not aware of any study examining this issue across the pandemic period.

Our analysis uses daily data on sectoral and industrial stock returns, market returns, and economic factors from South Africa. Using a multifactor arbitrage pricing model, we show that sectors and industries have been more exposed to the exchange rate risk during the pandemic than before it. A few sectors, such as the beverages, mining, personal goods, and tobacco sectors, tend to benefit from this exposure. The exposure predominantly hurts other sectors. Likewise, the industry level estimates suggest that a few industries, such as basic materials, consumer goods, and technology, may benefit from the exchange rate exposure, while the others suffer. We find consistencies in our estimates by allowing the stock market to respond to good and bad news symmetrically and by using a two-factor arbitrage pricing model.

2. Model and Data

2.1. Model—Arbitrage pricing theory and exchange rate exposure

To examine whether firms are exposed to exchange rate risk, we motivate our model by the arbitrage pricing theory of Ross (1976), which argues that expected returns on financial assets are a linear function of factors influencing the behaviour of these assets. The sensitivity of expected returns to the factors (i.e. the beta coefficients), indicates the risk posed by the factors to expected returns of financial assets. The arbitrage pricing theory shows that the price of a financial asset is the expected closing price discounted at the derived rate of return.

According to Chen, Roll, and Ross (1986), the arbitrage pricing theory can be tested by regressing expected returns on factors the researcher believes are priced in the financial market. Thus, following Chen, Roll, and Ross (1986), our model regresses expected returns on three factors—market returns (R^M), short-term interest rate (SR), and oil price growth/returns (OIL)—and the exchange rate returns (EX). Note that Chen, Roll, and Ross (1986) included six factors, namely market returns, industrial production growth, change in expected inflation, unexpected inflation, risk premium, and term structure in their model. Data on these variables, except market returns, are not available at daily frequencies making it impossible to include them in our model. However, we include the short-term interest rate to cater for inflation risk, risk premium, and term structure, consistent with Grossman and Shiller (1981). We also include oil price growth to cater for inflation risk and industrial production growth, because this indicator is a good predictor of these factors (see Sharma, Phan, and Iyke, 2019). The inclusion of the exchange rate factor is consistent with Jorion (1991). The rationale for including these factors is to reduce biasedness in the exchange rate exposure coefficients because stock prices and exchange rates are likely to be influenced by these factors.

Consider the two-factor arbitrage pricing model—the capital asset pricing model—containing the exchange rate. The arbitrage pricing theory implies that expected returns are linearly related to market risk and exchange rate exposure as follows

$$E\left(R_{i,t}\right)={\delta }_0+{\delta }_1{\beta }_i^M+{\delta }_{EX}{\beta }_i^{EX}$$ (1)

where $R_{i,t}=R_{i,t}^N-R_{f,t}$ is excess return, $R_{i,t}^N$ is nominal returns, R _f,t is the risk-free rate, δ′s are parameters of the model, and ${\beta }_i^M$ and ${\beta }_i^{EX}$ are market risk and exchange rate exposure, respectively. The market satisfies Equation (1), whereby ${\beta }_M^{EX}=0$ (i.e. the market exposure to exchange rate movements is zero), so that expected market returns are not sensitive to exchange rate movements.2 Thus, Equation (1) can be formulated as

$$E\left(R_{i,t}\right)={\delta }_0+\left[E\left(R_{M,t}\right)-{\delta }_0\right]{\beta }_i^M+{\delta }_{EX}{\beta }_i^{EX}$$ (2)

Empirically, this means that the return of an asset is composed of an expected component E(R _i,t) and an error/innovation component. In our case, this means

$$R_{i,t}=E\left(R_{i,t}\right)+{\beta }_i^M\left[R_{M,t}-E\left(R_{M,t}\right)\right]+{\beta }_i^{EX}{F}_{EX,t}+{\epsilon }_{i,t}$$ (3)

where F _EX,t, the exchange rate factor, is the zero mean residual term obtained by regressing the exchange rate returns, R _EX,t, on market returns, R _M,t. That is,

$$F_{EX,t}=R_{EX,t}-\left({\eta }_0+{\eta }_1{R}_{M,t}\right)$$ (4)

where η₀ and η₁ are parameters. The exchange rate factor is orthogonal to market returns. The orthogonalization of the exchange rate is necessary to isolate exchange rate risk/exposure from market risk because exchange rate returns and market returns are naturally correlated. Assuming rational expectations, we can replace expected returns in Equation (3) with Equation (2) and obtain

$$R_{i,t}={\alpha }_i+{\beta }_i^M{R}_{M,t}+{\beta }_i^{EX}{F}_{EX,t}+{\epsilon }_{i,t}$$ (5)

where ${\alpha }_i=\left[{\delta }_0\left(1-{\beta }_i^M\right)+{\delta }_{EX}{\beta }_i^{EX}\right]$. A multifactor extension of this model can be written as follows

$$R_{i,t}={\alpha }_i+\sum _{j=1}^3{\beta }_i^j{F}_{j,t}+{\beta }_i^{EX}{F}_{EX,t}^3+{\epsilon }_{i,t}$$ (6)

where ${\alpha }_i=\left[{\delta }_0+\sum _{j=1}^3{\delta }_j{\beta }_i^j+{\delta }_{EX}{\beta }_i^{EX}\right]$ and $F_{EX,t}^3=R_{EX,t}-\left({\eta }_0+\sum _{j=1}^3{\eta }_j{F}_{j,t}\right)$. F _j,t denotes the three factors, i.e. market returns, short-term rate, and oil growth, and $F_{EX,t}^3$ is the exchange rate factor obtained by regressing the exchange rate returns on the three factors. Firms are exposed to systematic risk and exchange rate movements if ${\beta }_i^M$ and ${\beta }_i^{EX}$ are statistically significant.

2.2. Data

We used daily data covering the period 1 January 2019 to 19 November 2020 for our estimations. The pre-COVID-19 period is from 1 January 2019 to 5 March 2020, while the COVID-19 period is from 6 March 2020 to 19 November 2020. The starting date for the pre-COVID-19 sample was chosen to align with the first case of COVID-19 recorded in South Africa.3 To construct the individual stock returns, we collected data on 30 FTSE/JSE sectoral and eight FTSE/JSE industrial stock indices from Datastream. The market return is based on data on the FTSE/JSE All Share Index from the same database. We calculated the stock returns as R _i,t = [ln(Stock  Index)_i,t/ln(Stock  Index)_i,t − 1]*100 for stock i in period t. We collected data on the exchange rate (South African rand per US dollar), short-term interest rate (treasury bill rate, 91 days), and oil price (Europe Brent Spot FOB U$/BBL) from Datastream as well. We calculated exchange rate returns and oil price growth/returns as logarithm changes in exchange rate and oil price multiplied by 100, respectively. Table A.1 in the appendix shows summary statistics of the variables used in our analysis. Table A.2 shows that, apart from the short-term interest rate, all the variables are stationary or do not have a unit root. Accordingly, we differenced the short-term interest rate before including it in our regressions.

Table A1.

Summary statistics

	Panel A: Sectoral returns
	Before COVID-19				During COVID-19				Full Sample
Variables	Mean	Max	Min	SD	Mean	Max	Min	SD	Mean	Max	Min	SD
Banks	-0.055	6.022	-5.382	1.470	-0.087	10.231	-16.132	3.741	-0.067	10.231	-16.132	2.567
Life Insurance	-0.079	4.174	-4.763	1.389	-0.081	16.541	-12.714	3.309	-0.080	16.541	-12.714	2.303
Auto and Parts	0.022	6.977	-4.546	1.657	-0.107	12.148	-21.447	3.693	-0.027	12.148	-21.447	2.613
Beverages	-0.032	5.213	-8.131	1.315	0.059	14.347	-12.865	2.861	0.002	14.347	-12.865	2.037
Chemicals	-0.269	9.952	-13.020	2.458	-0.152	30.354	-49.534	7.042	-0.225	30.354	-49.534	4.728
Construction and Materials	-0.137	4.636	-5.770	1.451	0.011	11.851	-7.754	2.201	-0.082	11.851	-7.754	1.770
Electronic and Electrical Equipment	-0.059	5.884	-6.172	1.735	-0.208	9.789	-15.260	3.650	-0.115	9.789	-15.260	2.622
Food and Drug Retailers	-0.020	4.481	-4.666	1.270	0.008	7.691	-9.968	2.020	-0.009	7.691	-9.968	1.592
Food Producers	-0.080	4.466	-4.149	1.081	0.036	7.189	-5.075	1.666	-0.036	7.189	-5.075	1.331
Forestry and Paper	-0.034	3.801	-5.310	1.572	0.030	10.969	-11.661	2.665	-0.010	10.969	-11.661	2.050
Fixed Line Telecommunications	-0.243	9.168	-10.982	2.528	0.082	16.145	-18.549	5.536	-0.121	16.145	-18.549	3.936
General Retailers	-0.137	4.688	-7.928	1.512	-0.020	8.929	-11.632	2.955	-0.093	8.929	-11.632	2.168
General Financials	-0.039	4.030	-4.702	1.104	-0.080	14.019	-12.190	2.548	-0.054	14.019	-12.190	1.787
General Industrials	-0.093	3.761	-4.077	1.223	-0.087	9.513	-9.266	2.970	-0.091	9.513	-9.266	2.058
Health Care Equipment and Services	-0.030	3.870	-4.216	1.290	-0.154	9.734	-12.700	2.713	-0.076	9.734	-12.700	1.949
Industrial Metals and Mining	-0.004	8.092	-8.726	2.673	0.284	14.182	-14.522	3.680	0.104	14.182	-14.522	3.090
Industrial Transportation	-0.113	4.720	-4.711	1.274	0.080	11.122	-12.911	3.265	-0.040	11.122	-12.911	2.240
Media	0.003	8.884	-6.408	1.808	0.161	11.004	-9.099	2.554	0.062	11.004	-9.099	2.118
Mining	0.076	3.678	-7.506	1.245	0.104	13.464	-15.891	3.087	0.086	13.464	-15.891	2.130
Mobile Telecommunications	-0.045	12.061	-6.165	1.614	-0.035	10.810	-11.976	3.293	-0.041	12.061	-11.976	2.385
Nonlife Insurance	-0.035	3.964	-4.611	1.402	-0.008	8.112	-7.101	2.327	-0.025	8.112	-7.101	1.804
Personal Goods	0.034	6.095	-6.131	1.527	0.115	10.387	-11.067	2.465	0.064	10.387	-11.067	1.932
Pharmaceuticals and Biotechnology	-0.074	8.891	-29.701	2.797	0.072	10.299	-12.188	2.964	-0.019	10.299	-29.701	2.859
Real Estate Investment Services	0.003	4.337	-3.159	0.865	-0.224	17.353	-23.667	3.743	-0.082	17.353	-23.667	2.394
Real Estate Investment Trust	-0.125	2.725	-3.612	0.853	-0.251	15.984	-16.380	3.583	-0.173	15.984	-16.380	2.295
Software and Computer Services	0.036	3.971	-8.619	1.685	0.116	9.543	-8.961	2.560	0.066	9.543	-8.961	2.056
Support Services	-0.033	4.766	-4.241	1.384	0.088	9.480	-12.178	2.668	0.013	9.480	-12.178	1.966
Tobacco	0.104	8.306	-5.480	1.724	-0.061	8.428	-9.066	2.254	0.042	8.428	-9.066	1.940
Travel and Leisure	-0.184	4.024	-3.291	1.230	-0.319	21.819	-20.132	4.421	-0.235	21.819	-20.132	2.876
	Panel B: Industry returns
	Before COVID-19				During COVID-19				Full Sample
Variables	Mean	Max	Min	SD	Mean	Max	Min	SD	Mean	Max	Min	SD
Basic Materials	0.030	3.483	-6.767	1.195	0.088	12.091	-15.665	3.004	0.052	12.091	-15.665	2.067
Consumer Goods	0.031	4.722	-3.887	1.161	0.072	7.236	-10.041	2.015	0.047	7.236	-10.041	1.537
Consumer Services	-0.041	3.563	-4.245	1.176	0.006	6.371	-9.309	2.009	-0.023	6.371	-9.309	1.541
Financials	-0.060	3.653	-4.299	1.095	-0.096	13.272	-13.096	3.090	-0.073	13.272	-13.096	2.080
Health Care	-0.045	4.908	-10.383	1.470	-0.068	6.506	-11.113	2.388	-0.054	6.506	-11.113	1.867
Industrials	-0.098	3.080	-3.835	1.034	-0.050	7.643	-9.723	2.570	-0.080	7.643	-9.723	1.773
Technology	0.036	3.971	-8.619	1.685	0.116	9.543	-8.961	2.560	0.066	9.543	-8.961	2.056
Telecommunications	-0.056	11.245	-6.060	1.577	-0.030	10.630	-11.739	3.289	-0.046	11.245	-11.739	2.367
	Panel C: Factors and exchange rate
Variables	Before COVID-19				During COVID-19				Full Sample
RM	0.001	2.133	-4.608	0.896	0.038	7.261	-10.227	2.071	0.015	7.261	-10.227	1.452
EX	0.027	2.499	-2.411	0.818	-0.007	3.983	-2.842	1.156	0.014	3.983	-2.842	0.958
OIL	0.005	11.070	-6.749	2.108	-0.099	41.202	-64.370	8.366	-0.034	41.202	-64.370	5.385
SR	7.008	7.610	6.200	0.279	4.159	6.430	3.430	0.842	5.937	7.610	3.430	1.491

The table shows summary statistics on the sectoral and industrial returns, the three factors (market returns (R^M), short-term interest rate (SR), and oil price growth/returns (OIL)), and exchange rate returns (EX). The statistics are reported for three periods: before COVID-19 (1 January 2019 to 5 March 2020), during COVID-19 (6 March 2020 to 19 November 2020), and the full sample period (1 January 2019 to 19 November 2020). Max, Min, and SD denote maximum, minimum, and standard deviation, respectively.

Table A2.

Unit root test results

	ADF (Constant)	P-value	ADF (Trend)	P-value
Variable	Panel A: Sectoral returns
Banks	-22.697	0.000	-22.675	0.000
Life Insurance	-24.146	0.000	-24.121	0.000
Auto and Parts	-22.065	0.000	-22.062	0.000
Beverages	-19.770	0.000	-19.761	0.000
Chemicals	-12.000	0.000	-11.990	0.000
Construction and Materials	-13.296	0.000	-13.295	0.000
Electronic and Electrical Equipment	-23.938	0.000	-23.915	0.000
Food and Drug Retailers	-24.291	0.000	-24.270	0.000
Food Producers	-23.017	0.000	-23.008	0.000
Forestry and Paper	-23.938	0.000	-23.924	0.000
Fixed Line Telecommunications	-23.474	0.000	-23.452	0.000
General Retailers	-21.817	0.000	-21.830	0.000
General Financials	-13.754	0.000	-13.741	0.000
General Industrials	-22.814	0.000	-22.791	0.000
Health Care Equipment and Services	-22.868	0.000	-22.843	0.000
Industrial Metals and Mining	-24.352	0.000	-24.341	0.000
Industrial Transportation	-9.297	0.000	-9.374	0.000
Media	-20.908	0.000	-20.886	0.000
Mining	-14.097	0.000	-14.095	0.000
Mobile Telecommunications	-24.211	0.000	-24.186	0.000
Nonlife Insurance	-23.559	0.000	-23.536	0.000
Personal Goods	-22.123	0.000	-22.100	0.000
Pharmaceuticals and Biotechnology	-17.938	0.000	-17.945	0.000
Real Estate Investment Services	-9.862	0.000	-9.875	0.000
Real Estate Investment Trust	-9.498	0.000	-9.489	0.000
Software and Computer Services	-22.256	0.000	-22.233	0.000
Support Services	-23.121	0.000	-23.161	0.000
Tobacco	-21.927	0.000	-21.940	0.000
Travel and Leisure	-11.708	0.000	-11.696	0.000
	Panel B: Industry returns
Basic Materials	-13.943	0.000	-13.932	0.000
Consumer Goods	-13.881	0.000	-13.868	0.000
Consumer Services	-23.087	0.000	-23.065	0.000
Financials	-14.090	0.000	-14.076	0.000
Health Care	-22.390	0.000	-22.372	0.000
Industrials	-23.201	0.000	-23.184	0.000
Technology	-22.255	0.000	-22.233	0.000
Telecommunications	-24.095	0.000	-24.071	0.000
	Panel C: Factors and exchange rate
RM	-9.416	0.000	-9.405	0.000
EX	-22.210	0.000	-22.200	0.000
OIL	-18.114	0.000	-18.097	0.000
SR	0.113	0.966	-1.365	0.870

This table shows the unit root test results. These results are based on the Augmented Dickey–Fuller test. We consider test regressions with: (i) a constant and (ii) a constant and a trend, and these are denoted, respectively, ADF (Constant) and ADF (Trend).

3. Results

3.1. Main results

Studies show that financial markets respond to bad and good news differently (Iyke, 2020b; Iyke and Ho, 2020a). Hence, we estimate Equation (6) after addressing this feature of financial markets and report the systematic risk and exchange rate exposure coefficients in Table 2 . These estimates are based on stock returns from 30 sectors. The direction of exposure varies from negative to positive. Prior to the COVID-19 outbreak (Panel A), exchange rate exposure was significant in 16 sectors, of which 11 recorded a negative exposure, while five recorded a positive exposure. The estimates in Panel B tell a different story. Surprisingly, less sectors were exposed to exchange rate movements during the pandemic. Specifically, 13 sectors were exposed to exchange rate movements and of these 13 sectors, only four sectors gained from this exposure, while the remaining nine sectors were hurt by it. Over the full sample period (Panel C), 20 sectors recorded significant exchange rate exposure. Of these sectors, only four sectors benefitted from the exposure, whereas the remaining 16 sectors suffered.

Table 2.

Exchange rate exposure and systematic risk at the sector level based on EGARCH(1,1) model

	Panel A: Before COVID-19				Panel B: During COVID-19:				Panel C: Full Sample
Variable	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value
Banks	1.013	0.000	-0.552	0.000	1.330	0.000	-0.745	0.000	1.219	0.000	-0.641	0.000
Life Insurance	1.086	0.000	-0.350	0.000	1.220	0.000	-0.549	0.000	1.132	0.000	-0.452	0.000
Auto and Parts	0.188	0.016	-0.104	0.343	0.401	0.000	-0.383	0.000	0.381	0.000	-0.427	0.000
Beverages	0.466	0.000	0.492	0.000	0.869	0.000	-0.060	0.707	0.619	0.000	0.403	0.000
Chemicals	1.206	0.000	0.192	0.197	1.385	0.000	-0.303	0.367	1.691	0.000	0.158	0.277
Construction and Materials	0.306	0.000	-0.113	0.274	0.471	0.000	-0.189	0.198	0.439	0.000	-0.151	0.078
Electronic and Electrical Equipment	0.656	0.000	-0.279	0.011	0.558	0.000	-0.699	0.002	0.780	0.000	-0.334	0.000
Food and Drug Retailers	0.941	0.000	-0.211	0.004	0.601	0.000	-0.040	0.709	0.761	0.000	-0.161	0.005
Food Producers	0.598	0.000	-0.160	0.016	0.396	0.000	-0.130	0.170	0.529	0.000	-0.136	0.004
Forestry and Paper	0.911	0.000	0.121	0.154	0.980	0.000	0.149	0.201	0.945	0.000	0.064	0.299
Fixed Line Telecommunications	0.990	0.000	-0.086	0.573	1.191	0.000	-0.470	0.263	0.961	0.000	-0.141	0.307
General Retailers	1.010	0.000	-0.446	0.000	1.089	0.000	-0.508	0.000	1.121	0.000	-0.481	0.000
General Financials	0.941	0.000	-0.051	0.297	0.931	0.000	-0.250	0.014	0.935	0.000	-0.113	0.014
General Industrials	0.915	0.000	-0.340	0.000	0.869	0.000	-0.575	0.000	0.908	0.000	-0.388	0.000
Health Care Equipment and Services	0.866	0.000	-0.072	0.285	0.888	0.000	-0.150	0.318	0.866	0.000	-0.113	0.087
Industrial Metals and Mining	1.427	0.000	-0.211	0.211	1.109	0.000	-0.080	0.616	1.201	0.000	-0.059	0.588
Industrial Transportation	0.651	0.000	-0.138	0.067	0.876	0.000	-0.450	0.003	0.620	0.000	-0.144	0.011
Media	0.544	0.000	-0.062	0.626	0.467	0.000	0.095	0.416	0.546	0.000	-0.033	0.671
Mining	0.905	0.000	0.271	0.000	1.289	0.000	0.209	0.024	1.058	0.000	0.239	0.000
Mobile Telecommunications	0.964	0.000	0.041	0.616	1.071	0.000	-0.175	0.224	1.024	0.000	-0.138	0.020
Nonlife Insurance	0.588	0.000	-0.268	0.006	0.451	0.000	-0.116	0.483	0.530	0.000	-0.253	0.001
Personal Goods	0.860	0.000	0.410	0.000	0.875	0.000	0.335	0.001	0.792	0.000	0.401	0.000
Pharmaceuticals and Biotechnology	1.309	0.000	-0.652	0.000	0.725	0.000	0.031	0.839	0.798	0.000	-0.519	0.000
Real Estate Investment Services	0.476	0.000	0.200	0.000	1.073	0.000	-0.123	0.216	0.993	0.000	-0.019	0.684
Real Estate Investment Trust	0.471	0.000	-0.118	0.009	1.053	0.000	-0.563	0.000	0.547	0.000	-0.130	0.005
Software and Computer Services	1.400	0.000	-0.050	0.531	0.729	0.000	0.260	0.030	1.175	0.000	0.034	0.609
Support Services	0.253	0.002	0.051	0.164	0.339	0.000	0.050	0.771	0.278	0.000	0.040	0.627
Tobacco	0.533	0.000	0.603	0.000	0.381	0.000	0.511	0.000	0.409	0.000	0.621	0.000
Travel and Leisure	0.466	0.000	-0.135	0.118	0.752	0.000	-0.288	0.331	0.490	0.000	-0.047	0.534

This table reports estimates of the systematic risk and exchange rate exposure coefficients obtained from the four-factor asset pricing model, Equation (6). β^EX is the exchange rate exposure coefficient derived from projecting exchange rate returns on the three factors, namely market returns (R^M), short-term interest rate (SR), and oil price returns (OIL). F^EX is the exchange rate component that is orthogonal to the three factors. Panels A, B, and C focus on the period before COVID-19, during COVID-19, and the full sample, respectively.

In Table 3 , we report the exchange rate exposure dynamics at the industry level. Five industries were exposed to exchange rate movements before COVID-19 (Panel A). Of these industries, three (consumer services, financials, and industrials) recorded a negative exposure, while two (basic materials and consumer goods) recorded a positive exposure. During COVID-19 (Panel B), six industries recorded significant exposures, of which three, namely consumer services, financials, and industrials experienced a negative exposure, whereas the other three, namely basic materials, consumer goods, and technology benefitted from the exposure. Similarly, for the full sample period (Panel C), six industries recorded significant exposures, of which only two, namely basic materials and consumer goods benefitted from the depreciation of the rand.

Table 3.

Exchange rate exposure and systematic risk at the industry level based on EGARCH(1,1) model

	Panel A: Before COVID-19				Panel B: During COVID-19				Panel C: Full Sample
Variable	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value
Basic Materials	0.976	0.000	0.256	0.000	1.282	0.000	0.169	0.050	1.077	0.000	0.240	0.000
Consumer Goods	0.744	0.000	0.383	0.000	0.688	0.000	0.407	0.000	0.693	0.000	0.380	0.000
Consumer Services	0.962	0.000	-0.264	0.000	0.750	0.000	-0.144	0.087	0.792	0.000	-0.218	0.000
Financials	0.863	0.000	-0.353	0.000	1.136	0.000	-0.538	0.000	0.993	0.000	-0.363	0.000
Health Care	0.904	0.000	-0.046	0.593	0.696	0.000	-0.108	0.384	0.870	0.000	-0.095	0.203
Industrials	0.805	0.000	-0.291	0.000	0.802	0.000	-0.475	0.000	0.800	0.000	-0.327	0.000
Technology	1.266	0.000	-0.124	0.116	0.729	0.000	0.260	0.030	1.175	0.000	0.034	0.609
Telecommunications	0.977	0.000	0.044	0.621	1.070	0.000	-0.229	0.139	1.028	0.000	-0.161	0.008

This table reports estimates of the systematic risk and exchange rate exposure coefficients obtained from the EGARCH(1,1) specification of the three-factor asset pricing model. β^EX is the exchange rate exposure coefficient derived from projecting exchange rate returns on the three factors, namely market returns (R^M), short-term interest rate (SR), and oil price returns (OIL). F^EX is the exchange rate component that is orthogonal to the three factors. Panels A, B, and C focus on the period before COVID-19, during COVID-19, and the full sample, respectively.

The estimates suggest that both the sectors and industries are largely exposed to exchange rate movements whether before or during the pandemic. It appears that the significant drop in the market index during the pandemic, as displayed in Figure 1, is driven by industrial rather than sectoral exposure to rand depreciation. Moreover, much of the resilience of the South African stock market to exchange rate movements appears to be driven by sectors like beverages, mining, personal goods, and tobacco and industries like basic materials and consumer goods. This is unsurprising because South Africa is a dominant player in the Southern African export market in these sectors and industries, particularly in mining, tobacco, and beverages, but the country tends to significantly import electronic/electrical equipment, automobile, and car parts.4 Generally, sectors and industries that are import-dependent (export-dependent) experience a decline (an increase) in stock returns if the local currency depreciates (see Jorion, 1990), which is consistent with our findings.

3.2. Robustness checks

We examine the robustness of our results by estimating Equation (6) using the standard ordinary least squares (OLS) method, which assumes that bad and good news affect the stock market symmetrically. Following prior studies (e.g. Iyke and Ho, 2020b), we controlled for serial correlations and heteroskedasticity. Table A.3 in the appendix, which reports the sectoral-level estimates, suggests that 14 sectors were exposed to exchange rate movements before the pandemic. Nine of these sectors are negatively exposed, while five are positively exposed to exchange rate risk (Panel A). During the pandemic, 15 sectors were exposed to exchange rate risk, of which 11 recorded a negative exposure, while four recorded a positive exposure (Panel B). Over the full sample period (Panel C), 17 sectors recorded significant exposures, of which 13 reported negative exposures, while only four reported positive exposures.

Table A3.

Exchange rate exposure and systematic risk at the sector level

	Panel A: Before COVID-19				Panel B: During COVID-19				Panel C: Full Sample
Variable	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value
Banks	1.037	0.000	-0.537	0.000	1.280	0.000	-0.765	0.000	1.227	0.000	-0.668	0.000
Life Insurance	1.061	0.000	-0.364	0.000	1.210	0.000	-0.542	0.000	1.180	0.000	-0.466	0.000
Auto and Parts	0.214	0.101	-0.161	0.114	0.655	0.000	-0.131	0.606	0.547	0.000	-0.163	0.233
Beverages	0.578	0.000	0.517	0.000	0.871	0.000	0.162	0.401	0.795	0.000	0.327	0.003
Chemicals	1.223	0.000	0.233	0.167	1.784	0.000	-1.168	0.014	1.654	0.000	-0.512	0.065
Construction and Materials	0.364	0.008	-0.178	0.107	0.491	0.000	-0.026	0.827	0.463	0.000	-0.108	0.169
Electronic and Electrical Equipment	0.657	0.000	-0.319	0.015	0.728	0.000	-0.674	0.001	0.712	0.000	-0.492	0.000
Food and Drug Retailers	0.904	0.000	-0.197	0.006	0.632	0.000	-0.087	0.332	0.699	0.000	-0.128	0.031
Food Producers	0.606	0.000	-0.123	0.067	0.403	0.000	-0.048	0.640	0.454	0.000	-0.081	0.204
Forestry and Paper	0.900	0.000	0.130	0.275	0.974	0.000	0.137	0.246	0.961	0.000	0.110	0.189
Fixed Line Telecommunications	1.079	0.000	0.061	0.733	1.124	0.000	-0.515	0.111	1.109	0.000	-0.245	0.203
General Retailers	1.043	0.000	-0.443	0.000	1.055	0.000	-0.570	0.000	1.054	0.000	-0.505	0.000
General Financials	0.927	0.000	-0.081	0.151	0.988	0.000	-0.254	0.008	0.976	0.000	-0.179	0.001
General Industrials	0.917	0.000	-0.309	0.000	0.879	0.000	-0.444	0.000	0.892	0.000	-0.377	0.000
Health Care Equipment and Services	0.846	0.000	-0.066	0.350	0.721	0.000	-0.082	0.554	0.755	0.000	-0.077	0.328
Industrial Metals and Mining	1.363	0.000	-0.288	0.161	1.043	0.000	0.104	0.529	1.127	0.000	-0.087	0.533
Industrial Transportation	0.653	0.000	-0.076	0.323	0.877	0.000	-0.475	0.005	0.821	0.000	-0.285	0.004
Media	0.590	0.000	-0.069	0.555	0.413	0.004	0.222	0.249	0.461	0.000	0.075	0.520
Mining	0.941	0.000	0.251	0.005	1.272	0.000	0.187	0.054	1.187	0.000	0.209	0.002
Mobile Telecommunications	0.891	0.000	-0.057	0.654	0.996	0.000	-0.295	0.059	0.973	0.000	-0.186	0.056
Nonlife Insurance	0.596	0.000	-0.272	0.002	0.451	0.000	-0.116	0.439	0.488	0.000	-0.185	0.038
Personal Goods	0.858	0.000	0.429	0.000	0.791	0.000	0.290	0.030	0.801	0.000	0.369	0.000
Pharmaceuticals and Biotechnology	1.121	0.000	-0.010	0.954	0.551	0.000	0.273	0.271	0.695	0.000	0.141	0.349
Real Estate Investment Services	0.476	0.000	0.200	0.000	1.243	0.000	-0.312	0.013	1.050	0.000	-0.085	0.287
Real Estate Investment Trust	0.529	0.000	-0.119	0.045	1.235	0.000	-0.526	0.000	1.062	0.000	-0.356	0.000
Software and Computer Services	1.261	0.000	-0.024	0.771	0.820	0.000	0.226	0.071	0.927	0.000	0.129	0.103
Support Services	0.273	0.003	0.045	0.664	0.408	0.000	-0.010	0.957	0.377	0.000	0.005	0.965
Tobacco	0.613	0.000	0.657	0.000	0.468	0.000	0.603	0.002	0.495	0.000	0.646	0.000
Travel and Leisure	0.440	0.000	-0.132	0.107	1.044	0.000	-0.445	0.117	0.893	0.000	-0.316	0.044

This table reports estimates of the systematic risk and exchange rate exposure coefficients obtained from the four-factor asset pricing model, Equation (6). β^EX is the exchange rate exposure coefficient derived from projecting exchange rate returns on the three factors, namely market returns (R^M), short-term interest rate (SR), and oil price returns (OIL). F^EX is the exchange rate component that is orthogonal to the three factors. Panels A, B, and C focus on the period before COVID-19, during COVID-19, and the full sample, respectively. We estimate Equation (6) by OLS and address potential serial correlation and heteroskedasticity in the error terms using heteroskedasticity and autocorrelation consistent standard errors.

In Table A.4 in the appendix, we report the industry-level estimates. Consistent with the main estimates, we found that five industries recorded exchange rate exposures before the pandemic. Three of these industries, namely consumer services, financials, and industrials, experienced negative exposures, while two, namely basic materials and consumer goods, experienced positive exposures before the pandemic (Panel A). Similarly, consistent with the baseline results, six industries reported significant exposures during the pandemic—four of these (i.e. consumer services, financials, industrials, and telecommunications) reported negative exposures, whereas only two (consumer goods and technology) reported positive exposures (Panel B). Lastly, Panel C shows that six industries were significantly exposed to exchange rate movements—two, namely basic materials and consumer goods, benefitted from the depreciation, while the remaining four industries suffered as a result.

Table A4.

Exchange rate exposure and systematic risk at the industry level

	Panel A: Before COVID-19				Panel B: During COVID-19				Panel C: Full Sample
Variable	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value	βM	P-value	βEX	P-value
Basic Materials	0.979	0.000	0.239	0.003	1.264	0.000	0.138	0.111	1.193	0.000	0.175	0.004
Consumer Goods	0.769	0.000	0.404	0.000	0.688	0.000	0.324	0.002	0.702	0.000	0.373	0.000
Consumer Services	0.935	0.000	-0.244	0.000	0.743	0.000	-0.206	0.003	0.792	0.000	-0.218	0.000
Financials	0.906	0.000	-0.317	0.000	1.175	0.000	-0.538	0.000	1.113	0.000	-0.444	0.000
Health Care	0.938	0.000	-0.044	0.564	0.664	0.000	0.045	0.746	0.734	0.000	0.002	0.981
Industrials	0.806	0.000	-0.271	0.000	0.817	0.000	-0.399	0.000	0.816	0.000	-0.337	0.000
Technology	1.261	0.000	-0.024	0.772	0.820	0.000	0.226	0.071	0.927	0.000	0.129	0.103
Telecommunications	0.901	0.000	-0.055	0.655	0.999	0.000	-0.304	0.049	0.978	0.000	-0.189	0.047

This table reports estimates of the systematic risk and exchange rate exposure coefficients obtained from the four-factor asset pricing model, Equation (6). β^EX is the exchange rate exposure coefficient derived from projecting exchange rate returns on the three factors, namely market returns (R^M), short-term interest rate (SR), and oil price returns (OIL). F^EX is the exchange rate component that is orthogonal to the three factors. Panels A, B, and C focus on the period before COVID-19, during COVID-19, and the full sample, respectively. We estimate Equation (6) by OLS and address potential serial correlation and heteroskedasticity in the error terms using heteroskedasticity and autocorrelation consistent standard errors.

These estimates may also be driven by nuisance parameters. Accordingly, we estimated the two-factor arbitrage pricing model, i.e. Equation (5), to sidestep this problem. The estimates, which are unreported due to limited space, show differential exposures over the sample period. The industries were more exposed to the exchange rate risk during than before the pandemic and appeared to have driven the sharp decline in the composite stock market index, shown in Figure 1.

Our finding aligns with the growing literature highlighting the unprecedented exposure of economies and financial markets to the COVID-19 pandemic (Devpura and Narayan, 2020, Iyke, 2020a, Iyke, 2020c, Iyke and Ho, 2021, Mazur et al., 2020, Narayan et al., 2020, Sharma, 2020, Topcu and Gulal, 2020, Iyke, 2020b, Wang et al., 2020).

4. Conclusion

We examined the nature of exchange rate exposure before and during the COVID-19 pandemic. Our analysis uses daily data from South Africa and a multifactor arbitrage pricing model. We found that, as compared with sectors, most industries were more exposed to the exchange rate risk during than before the pandemic. Exchange rate exposure predominantly hurts the sectors and industries, although a few sectors, such as beverages, mining, personal goods, and tobacco, and industries, such as basic materials, consumer goods, and technology, may benefit from it. Our findings suggest that these few sectors and industries may be good for diversifying risks during bad times.