Asymmetric effects of exchange rate volatility on trade flows in Nigeria

Purpose — This study assesses the symmetric and asymmetric effects of exchange rate volatility on trade flows in Nigeria. Method — The study employs quarterly data and covers the period 1995q1 to 2020q4. The data were sourced from International Financial Statistics (IFS) and Central Bank of Nigeria (CBN) websites. The paper applies both linear ARDL and non-linear ARDL (NARDL) models. These methods are employed to evaluate the symmetric and asymmetric effects of exchange rate volatility. Result — The results from linear ARDL model show that exchange rate volatility has only significant short-run effect on export while it has both short-run and long run effects on the imports. The findings from the non-linear ARDL suggest that exchange rate volatility has neither short run nor long run asymmetric effects on exports. However, the non-linear ARDL model reveals short run and long run asymmetric effects of exchange rate volatility on imports. The findings show that increase in volatility reduces imports while decrease in volatility boosts imports. Contribution — Previous studies have only investigated the symmetric effects of exchange rate volatility on trade balance in Nigeria. This study contributes to the literature by examining the symmetric and asymmetric effects of exchange rate volatility on trade flows, using the GARCH-based measure of exchange rate volatility.


INTRODUCTION
Since the adoption of the floating exchange rate system in 1973, economists have been concerned about the effects of exchange rate volatility on trade flows. Both the theoretical advances and empirical findings indicate that exchange rate volatility impact trade flows (Senadza & Diaba, 2018). However, results from the theoretical and empirical literature show that the effects of exchange rate volatility on trade flows can either be positive, negative or neutral. Theoretically, the impacts of exchange rate volatility will largely depend on the risk attitude of the traders or the degree of risk tolerated by the traders (Grauwe, 1988;Perée & Steinherr, 1989). For a trader that is risk averse, exchange rate volatility will have negative effects on trade flows while for a trader that is risk tolerant, exchange rate volatility will have positive effects on trade flows. This theoretical ambiguity has been supported by empirical findings. For instance, Senadza  The results show that exchange rate volatility negatively affect export in the short run but not in the long run. Yunusa (2020) examines the effects of exchange rate volatility on crude oil export to 7 Nigerian trading partners. The results indicate that the volatility of exchange rate significantly affect crude oil exports to the trading partners. Ayomitunde et al. (2020) evaluate the impact of exchange rate volatility on exports in Nigeria. The findings show negative effects of exchange rate volatility on exports.
This study contributes to the existing literature on the asymmetric effects of exchange rate volatility on trade flows by using the GARCH-based measure of exchange rate volatility. Similar to Bahmani-Oskooee & Aftab (2017) and Bahmani-Oskooee & Arize, (2020), we examine the asymmetric effects of exchange rate volatility on exports and imports in Nigeria using the non-linear ARDL model. Recent studies have shown that impacts of exchange rate volatility on trade flows might be asymmetric and not symmetric (Bahmani-Oskooee et al., 2020). This suggests that traders react differently to an increase and a decrease in volatility.
The theoretical proposition that exchange rate volatility has negative effects on trade can be traced to Ethier (1973). The model is based on decision making regarding trade by a risk averse firm when exchange rate is volatile. Based on this risk averse assumption, exchange rate volatility has adverse effects on trade. Similarly, Clark (1973) developed a model of risk averse that shows that exchange rate volatility decreases exports. However, the theory suggest that perfect forward market could reduce the impacts. Baron (1976) shows that sellers may be uncertain of how much exchange rate risk they will cover if the forward market is not fully developed.
Conversely, some theoretical studies have shown that exchange rate volatility has positive effects on trade flows. Franke (1991) shows that under a monopolistically competitive markets where firm are risk neutral, exchange rate volatility increases the volume of exports. Sercu (1992) concludes that by increasing the probability that the price a trader receives might be greater than the trade costs, exchange rate volatility can have positive impacts on trade. Sercu & Vanhulle (1992) hypothesize an increase in exchange rate volatility increases the value of exporting firm thereby increasing exports. The findings indicate that the asymmetric effects of exchange rate volatility are partner specific. Generally, the results suggest that increase in exchange rate volatility has significant effects on export and imports but decrease in volatility has no effects.
Chien et al. (2020) examine the asymmetric effects of exchange rate volatility on bilateral trade flows between Taiwan and Indonesia for 19 export and import industries. The findings reveal that the long run asymmetric effects of exchange rate volatility has much higher impacts on Taiwan export to Indonesia than imports from Indonesia. However, the short run asymmetric effects cause unstable change on exports and imports industries. Bahmani-Oskooee et al. Owing to these evolving literature and divergent results on the effects of exchange rate volatility and trade flows, this study empirically investigates the symmetric and asymmetric effects of exchange rate volatility on trade flows in Nigeria.

METHOD
This study applied the Autoregressive Distributed Lag (ARDL) and Non-linear Autoregressive Distributed Lag (NARDL) as it is standard in the literature on symmetric and asymmetric effects. Similar to previous studies (Bahmani-Oskooee & Aftab, 2017; Bahmani-Oskooee & Arize, 2020), we specify the export and import models. These studies have also included real income, real exchange rate and a measure of exchange rate volatility. The study used quarterly data that covers the period 1995q1 to 2020q4.
We specify the export model as: Where is the Nigerian export, is a measure of the world income, is the real exchange rate, and is a measure of real exchange rate volatility based on GARCH. Eq. (1) shows the global determinants of Nigerian exports. If a rise in world income leads to an increase in export, the parameter 1 will be positive. In theory, a rise in real exchange rate, depreciation is expected to increase export. Hence, the estimate of 2 is expected to be positive. Lastly, since real exchange rate volatility could either have positive or negative effects on exports, the parameters 3 can either be positive or negative.
Estimating eq. (1) will only show the long run effects of the exogenous variables on exports. To identify the short run effects and distinguish it from the long run effects of exchange rate volatility on exports, we follow the literature and adopt  1) in error-correction model as: In eq. (2) , the short run effects are captured by the estimates of assigned to the first-differenced variables and long run effects are represented by the estimates of 2 − 4 normalized on 1 . According to Pesaran et al. (2001), the − is applied to establish joint significance of lagged variable as a sign of cointegration.
The above model in eq. (2) assumes that change in exchange rate volatility has symmetric effects on exports and imports. Recent literature has shown that exchange rate volatility might have asymmetric effects where increased volatility on trade flows might have different effects from decreased volatility. To test for asymmetric effects, we modify Eq.
Eq. (4)  to establish cointegration in the two models. After estimating eq. (4) with OLS, we can test the asymmetric hypothesis. A short run asymmetry is established if Δ takes a lag order different from Δ . We can also confirm the short run asymmetric effects if the estimate of 5 is different than the estimate of 6 . To establish the short run accumulative or asymmetric impacts of exchange rate volatility, the Wald test is applied to determine if the sum of short run estimates assigned to Δ − and Δ − are statistically different. i.e if ∑̂5 ≠ ∑ 6 . Lastly, the long run asymmetric effects of exchange rate volatility on export is established if 4|− 1 ≠ 5| −1 in Eq. (4). Also, the Wald test will be applied. Table 1 shows the results for the Augmented Dickey Fuller (ADF) and Phillips Peron (PP) unit root tests. The ARDL method requires variables to be a combination of I(0) and I(1) but not I(2), we apply the ADF and PP to test the stationarity levels. All the variables are non-stationary at levels but stationary at first differences.  Table 2 showed the estimates of Nigerian export linear demand model for both the short run and long run. The short run coefficient attached to the measure of exchange rate volatility suggests that Nigerian exports are affected by exchange rate volatility. The ∆lnvolat carries at least one significant lagged estimate. Further, the estimates show that world economic activities significantly determine Nigerian export demand in the short run. The estimates, however, indicate that real exchange rate has no significant effects on the demand for Nigerian exports in the short run. However, the short run effects of exchange rate volatility on demand for Nigerian export do not extend into the long run effects. The diagnostic tests are also reported in the table. The Lagrange Multiplier (LM) test is employed to detect autocorrelation and Ramsey RESET is used to detect misspecification. The two tests are insignificant showing that the residual is error free and the model does not suffer from misspecification.    Table 4 showed the estimates of non-linear model for export demand in the short run and long run. The estimates in the short run show that either increase in volatility (∆ ) or decrease in volatility (∆ ) carries at least one significant lagged estimate. Further, the table show that at a given lag , the estimates attached to ∆ − is different than the estimates attached to ∆ − , indicating short run asymmetric effects of exchange rate volatility. The cumulative short run asymmetric effects are presented in the Wald test reported in Table 6. Going by the Wald-test estimates, there is no short run evidence for the cumulative impact of asymmetric effect. Similarly, there is no    Table 5 show the estimates of non-linear model for import demand in the short run and long run. The estimates in the short run show that either increase in volatility (∆ ) or decrease in volatility (∆ ) carries at least one significant lagged estimate. Further, the table show that at a given lag , the estimates attached to ∆ − is different than the estimates attached to ∆ − , indicating short run asymmetric effects of exchange rate volatility. The cumulative short run asymmetric effects are presented in the Wald test reported in Table 6. Going by the Wald-test estimates, there is short run evidence for the cumulative impact of asymmetric effect. Moreover, as supported by the Wald estimates for the long run, there are long run asymmetric effects of exchange rate volatility on import. Table 6 shows the results for the effects of exchange rate volatility on export and import in the long run. The results support the short run and long run asymmetric effects of exchange rate volatility on imports only. There are neither short run asymmetric effect nor long run asymmetric effects of exchange rate volatility on exports.

Discussion
The research investigated how exchange rate volatility affects trade flows in Nigeria in both symmetric and asymmetric ways. The study found that in the short term, exchange rate volatility has an impact on trade flows in Nigeria, but the effects vary between imports and exports in the long term. Specifically, the study found that the short-term effects of exchange rate volatility on imports persisted into the long term, while the short-term effects on exports did not. This suggests that exchange rate volatility has a more enduring effect on imports than on exports.
Additionally, the study found that exchange rate volatility has no asymmetric effects on demand for exports, both in the short and long term. However, exchange rate volatility has asymmetric effects on demand for imports in both the short and long term. The short-term asymmetric effects of exchange rate volatility on imports continue into the long term. This means that an increase in volatility has a negative impact on imports, while a decrease in volatility has a positive impact on imports.
The findings also reveal that world income has a significant positive effect on the demand for Nigerian exports. This suggests that as global income increases, there is a corresponding increase in demand for Nigerian exports. On the other hand, the study found that Nigerian income has a positive but insignificant effect on the demand for imports. This means that although Nigerian income influences the demand for imports, the effect is not statistically significant, and other factors likely have a more significant impact on import demand.
Overall, the study's findings provide valuable insights into the factors that influence trade flows in Nigeria, including exchange rate volatility and income levels. These findings can inform policymakers and businesses in Nigeria when making decisions related to trade and economic development.

CONCLUSION
The study examined the symmetric and asymmetric effects of exchange rate volatility on trade flows in Nigeria. Previous studies that investigated the effects of exchange rate volatility on trade in Nigeria has assumed a form of symmetric effects. Generally, the results show that exchange rate volatility affects trade flows in Nigeria and seems to affect demand for imports more than the demand for exports. The linear ARDL results indicate that exchange rate volatility has both short run and long run effects on imports but only short run effects on exports. Further, the non-linear ARDL results suggest that exchange rate volatility has both short run and long run asymmetric effects on demand for imports but no asymmetric effects on demand for exports. Since Nigerian domestic firms largely depend on imported intermediate inputs for domestic production, the practical implication of the results include domestic output volatility, unstable manufacturing production and declining domestic investment.
The study thus recommends that: (i) The Central Bank of Nigeria (CBN) should formulate appropriate policies to ensure the stability of exchange rate. (ii) The CBN should periodically intervene in the foreign exchange market to minimize exchange rate volatility and instil investors' confidence in the market (iii) The CBN should deploy monetary instruments to minimize the passthrough effects of exchange rate volatility to domestic prices.
Further studies can examine the effects of exchange rate volatility at the industry level. This is to evaluate whether the effects of exchange rate volatility on trade flows are industry-specific.