Unravelling the devaluation puzzle: Empirical insights into the transmission channel on balance of payments and output in Ethiopia

2.1 Data

This study uses quarterly data obtained from the National Bank of Ethiopia (NBE) to investigate the effects of devaluation on the balance of payments and output. The dataset comprises seven variables: exchange rate, balance of payments, money supply, inflation, interest rates, foreign asset reserves, and real gross domestic product (GDP). To ensure consistency, the real GDP data initially available only annually from the NBE are interpolated to quarterly data using the quadratic-sum approach of E-views 12. The dataset covers a period of 92 observations from 2001Q1 to 2023Q4. The use of quarterly data facilitates the identification of structural shocks, reduces the likelihood of structural breaks, and captures important intra-year dynamics.³⁵

2.2 Theoretical framework and model specification

In line with established economic theories and empirical literature, this study builds a model that enables exploration of the transmission channels through which the exchange rate affects the balance of payments and output. By employing rigorous econometric techniques and a sound theoretical framework, this study provides valuable insights into the dynamics of exchange rate transmission in relation to the Ethiopian economy.

The model specification begins with the identification and determination of independent variables based on economic theory.³⁶ In accordance with the theoretical framework and empirical evidence, models can be specified to capture the impact of devaluation on the balance of payments and output growth by introducing structural shocks. The theoretical model examines the transmission channels of devaluation in the balance of payments and output. The model incorporates both the demand and supply sides of the economy to analyse the effects of devaluation.³⁷ On the demand side, the model considers the national income identity equation and aggregate supply.

The balance of payment channel: devaluation can improve the current account of an economy in equilibrium if the combined elasticities of foreign demand for exports and domestic demand for imports exceed unity, known as the Marshall-Lerner condition.^37–39 This indicates that devaluation has both a price effect, making exports more competitive and imports more expensive, and a volume effect, leading to increased export volumes and decreased import volumes. Moreover, devaluation can impact the current account balance by influencing the marginal propensity to absorb, income levels, and direct absorption. Furthermore, the imbalance in the money market, as explained by the monetary approach, contributes to the disequilibrium in the balance of payments. In addition, shocks in the money supply, central bank’s foreign exchange reserves, inflation, domestic interest rates, and real output levels influence the balance of payments, and are specified in Eq. (1).

Where, bop - is balance of payment, e - is nominal exchange rate, ${\mathit{M}}_{\mathit{s}}$ - is money supply (M2), inf - is inflation, I - is domestic interest rate, fxre - is foreign asset reserve, and rgdp - is real GDP.

The exchange rate channel: in a small open economy with free capital mobility, the domestic interest rate must equal the world interest rate in equilibrium.³⁷ Therefore, by equating the domestic interest rate with the world interest rate, we can determine the equilibrium level of the current account. However, in our economy, the interest rate is not determined through this mechanism but is instead fixed by the central bank. A higher domestic interest rate encourages domestic savings and reduces firms’ investment in physical capital. As a result, the domestic interest rate itself affects the current account by reducing exports. The relationship between the interest rate and the exchange rate, regardless of whether devaluation is expansionary or contractionary, is influenced by the type of shock.⁴⁰ This relationship between the domestic interest rate and investment demonstrates that the domestic interest rate has its own impact on the balance of payments.

This study also considers the inclusion of the price level in the balance of payment equation to account for the price environment of an economy. This is crucial in determining whether devaluation is inflationary or not, as inflation has a significant impact on output.⁴¹ An inflationary environment can have negative effects on output.^14,42 This is because it can lead to inefficient allocation of resources due to distortions in relative prices and increased administrative costs for firms^37,38; clearly indicate the interaction between the balance of payments and the exchange rate. According to the elasticity approach of exchange rate determination, the exchange rate is determined by the flow of currency through the exchange rate market affecting trade balance, which affects the balance of payments component. In the portfolio view of exchange rate determination; the change in asset prices in the stock market through changes in the price of bonds and money affects the exchange rate.

Therefore, we include BOP in the exchange rate model. We have also understood that an increase in the money supply lowers interest rates, reduces borrowing costs, and promotes investment, which might enhance domestic output. In addition, a higher money supply will reduce the value of the currency. This relationship shows that money supply affects investment and output.⁴⁰ As in the quantity theory of money, the nominal exchange rate is determined by the quantity of money printed by the central banks that is money supply. This is seeing the exchange rate from the viewpoint of the monetary approach; interest rate and money supply influence exchange rate determination. There is a long run relationship between foreign reserves and exchange rate.⁴³ Moreover; any changes in foreign reserves would lead to fluctuations in the exchange rate but not vice versa. On the basis of these theoretical foundations, we can develop the following exchange rate model, which is specified in Eq. (2):

From the above theoretical framework, we can develop the model that fulfills one of the objectives of this paper, and is specified in Eq. (3):

The existence of exchange rate devaluation has both contractionary and expansionary effects on output through the employment effect and the terms of trade balance effects.³⁸ To show which effect will happen when the country devalues its currency; adding nominal exchange rate into the model is necessary. An increase in money supply lowers interests, reduces borrowing costs, and promotes investment which might enhance domestic output. The improvement in investment leads to increase output which intern promotes a country to export more. In addition, a higher money supply will reduce the value of currency. This relation shows money supply affects the balance of payment.⁴⁰ Thus, we include the money supply in the model. Table 1 summarizes the contractionary effect of devaluation on the basis of various theoretical and empirical studies, highlighting the transmission channels.

2.2.1 Structural Vector Autoregressive model (SVAR)

The Structural Vector Autoregressive (here after, SVAR) model is a widely used tool for analysing the monetary transmission mechanism⁴⁵ and for empirically studying the impact of structural shocks empirically.^46,47 The SVAR model is constructed from the reduced form of the VAR (p) model, which represents the data generated by the structural VAR (p) model, as specified in Eq. (4);

Where, ${\mathit{\alpha }}_{\mathit{i},\mathit{j}}$ is an ($\mathit{n}\times 2$) matrix constants and linear trends, ${\mathit{B}}_{\mathit{i}}$, $\mathit{i}=1,\dots ,\mathit{p},$ is a $\mathit{K}\times \mathit{K}$ matrix of autoregressive slope coefficients, the $\mathit{K}\times \mathit{K}$ matrix ${\mathit{B}}_0$ reflects the instantaneous relations among the model variables, and the $\mathit{K}\times 1$ vector of mean zero structural shocks ${\mathit{w}}_{\mathit{t}}$ is serially uncorrelated with a diagonal covariance matrix $\mathit{\Sigma w}$ of full rank such that the number of shocks coincides with the number of variables. The endogenous variable,${\mathit{y}}_{\mathit{t}}$, that we include the VAR is the first difference of nominal exchange rate ($∆{\mathit{e}}_{\mathit{t}}$), output growth ($∆{\mathit{y}}_{\mathit{t}}$), consumer inflation ($∆{\mathit{p}}_{\mathit{t}}$), broad money supply ($∆{\mathit{ms}}_{\mathit{t}}$), nominal interest rate ($∆{\mathit{i}}_{\mathit{t}}$), foreign exchange reserve ($∆\mathit{\text{fxre}}$), and balance of payment ($\mathit{bop}$). These variables are assumed to be a covariance stationary vector process. For ($∆{\mathit{e}}_{\mathit{t}}$,$∆{\mathit{y}}_{\mathit{t}}$,$∆{\mathit{p}}_{\mathit{t}}$, $∆{\mathit{ms}}_{\mathit{t}}$, $∆{\mathit{i}}_{\mathit{t}}$, and $∆\mathit{\text{fxre}}$, we can reject the hypothesis of the existence of a unit root at the 10% level (using Augmented Dickey-Fuller and Phillips-Perron tests). We can, however, not reject the hypothesis of a unit root in$\mathit{bop}$. Given the low power of these tests in relatively small data sets, we follow^48–50 and assume that $\mathit{bop}$ is stationary since the balance of payment cannot have a unit root at level.

The reduced-form representation of the model in Eq. (5) can be obtained by pre-multiplying both sides of Eq. (4) by${{\mathit{B}}_0}^{-1}$, results in the model.^46,51,52

Where; ${\mathit{A}}_{\mathit{i}}$ = ${{\mathit{B}}_0}^{-1}{\mathit{B}}_{\mathit{i}}$and${\mathit{u}}_{\mathit{t}}={{\mathit{B}}_0}^{-1}{\mathit{w}}_{\mathit{t}}$, $\mathit{\lambda }$ = ${{\mathit{B}}_0}^{-1}\mathit{\alpha }$ and it can be estimated by maximum likelihood (ML) estimation methods.

Estimation of the matrix ${\mathit{B}}_0$ requires additional restrictions on the data generating process (DGP) based on economic theory. If the matrix ${\mathit{B}}_0$can be solved for, given these restrictions and the data, we say that the structural VAR model parameters, $({\mathit{B}}_0,{\mathit{B}}_1,\dots ,{\mathit{B}}_{\mathit{p}},\mathit{\Sigma w}),$ are identified or, equivalently, that the structural shocks $\mathit{wt}={\mathit{B}}_0{\mathit{u}}_{\mathit{t}}$ are identified. In compact form, an SVAR system relates to the following relations as specified in Eq. (6):

The equation (6) is known as the AB model⁵³; Where ${\mathit{A}}_0$is $(\mathit{n}\times \mathit{n})$matrix of contemporaneous relations between endogenous variables, B is $(\mathit{n}\times \mathit{n})$matrix that linearly relates the SVAR residuals to the structural innovations, ${\mathit{u}}_{\mathit{t}}$is vector of reduced-form residual, and ${\mathit{w}}_{\mathit{t}}$is vector of structural shocks. The residual ${\mathit{u}}_{\mathit{t}}$in the reduced form is presumed to be white noise. Therefore, we can estimate the AB model by OLS (ordinary list square).

The structural innovations ${\mathit{w}}_{\mathit{t}}$ can be derived from errors ${\mathit{u}}_{\mathit{t}}$ of the reduced form, but certain restrictions must be placed on the system. In details,$\frac{\mathit{n}(\mathit{n}-1)}{2}$ [1] Where n is the number of variables in the model; restrictions must be imposed on A0 matrix to be able to identify the structural shocks.⁵⁴

2.2.2 Recursively identified structural VAR model

In this study, based on the work of,⁴⁶ the response of economic variables to temporary shocks was estimated using a recursively identified structural model. This approach assumes that the current structural shocks are not influenced by the preceding ordering variables. Instead, it is assumed that the variables are affected by a sequential chain of shocks, or alternatively, the matrix ${\mathit{A}}_0$ is diagonal, indicating that the structural shocks are orthogonal. Specifically, the matrix ${\mathit{A}}_0$ takes the form of a lower triangular matrix, as illustrated below:

Where, $({{\mathit{u}}_{\mathit{t}}}^{\mathit{e}},{{\mathit{u}}_{\mathit{t}}}^{\mathit{i}},{{\mathit{u}}_{\mathit{t}}}^{\mathit{ms}},{{\mathit{u}}_{\mathit{t}}}^{\mathit{inf}},{{\mathit{u}}_{\mathit{t}}}^{\mathit{\text{fxre}}},{{\mathit{u}}_{\mathit{t}}}^{\mathit{bop}}\text{and}{{\mathit{u}}_{\mathit{t}}}^{\mathit{\text{rgdp}}})$, are the structural disturbance, that is, exchange rate shocks, interest rate shocks, money supply, the price shocks, foreign asset reserve shocks, the balance of payment shocks, and output shocks respectively; and (${{\mathit{w}}_{\mathit{t}}}^{\mathit{e}},{{\mathit{w}}_{\mathit{t}}}^{\mathit{i}},{{\mathit{w}}_{\mathit{t}}}^{\mathit{ms}},{{\mathit{w}}_{\mathit{t}}}^{\mathit{inf}},{{\mathit{w}}_{\mathit{t}}}^{\mathit{\text{fxre}}},{{\mathit{w}}_{\mathit{t}}}^{\mathit{bop}},\text{and}{{\mathit{w}}_{\mathit{t}}}^{\mathit{\text{rgdp}}})$, are the residuals in the reduced form equations, which represent unexpected movements (given information in the system) of each variable.

In certain scenarios where a fully developed theoretical model is unavailable, the process of identification can be achieved through the incorporation of extraneous information and the selective utilization of insights from economic theory. In line with,⁴⁶ we present the construction of a recursive form of the SVAR (Structural Vector Autoregressive) model, as demonstrated above.

Once the structural shocks have been identified, a rigorous analysis and interpretation of these macroeconomic shocks become imperative. This analysis is conducted within the framework of structural VAR models and involves the examination of structural impulse responses as well as the implementation of forecast error variance decompositions. These analytical techniques provide valuable insights into the dynamics and impacts of the identified shocks.

3.1 Stationarity test

To enhance the robustness of the test, the presence of a unit root was assessed using the Augmented Dickey–Fuller (ADF) and Phillip-Perron (PP) tests. The results of the unit root tests conducted using the Augmented Dickey–Fuller and Phillip-Perron methods are summarized in Table 2. Based on the results of the ADF and PP tests, it is evident that all variables in the VAR model, except bop, are nonstationary at this level. This implies that the null hypothesis of a unit root, including both trend, and trend and intercept, cannot be rejected for these variables. However, upon considering their first differences, these variables exhibit stationarity.

3.2 Lag Length and stability check

o determine the lag length of VAR/SVAR, the study employs different lag-length selection criteria, including the likelihood ratio test statistic (LR), Final Prediction Error (FPE), Akaike Information Criteria (AIC), and the Hannan-Quinn Information Criterion (HQ). In Table 3 the lag length selection criterion is tabulated. The AIC lag length section test indicates that the appropriate lag length for the VAR model is three (3).

Prior to estimating the parameters of the SVAR model with the optimal lag length, it is essential to assess the stability conditions of the VAR model by examining the AR roots. The obtained results, which are reported in Table 4, indicate that all eigenvalues in the proposed model lie within the unit circle, signifying values that are either less than one or equal to unity. Consequently, it can be concluded that the VAR/SVAR model satisfies the requisite stability condition.

3.3 Co-integration test and post estimation test

The unit root test reveals that the model in our study consists of a mixture of I(1) and I(0) variables. To investigate whether these variables are cointegrated at the same order (only I(1) variables) and at different orders (I(0) and I(1)), we employ Johansen’s cointegration technique.³⁶ Our analysis is based on an SVAR model with a mixture of I(1) and I(0) variables, as described previously.^48,55,56 The results of Johansen’s cointegration test in Table 5 indicate the presence of multiple cointegrating equations and vectors in the case of I(1) variables, whereas for the mixture of I(0) and I(1), evidence suggests the presence of three cointegrating relationships. These findings contribute to a deeper understanding of the long-run dynamics and interdependencies among the macroeconomic variables studied.

The study investigated the presence of serial autocorrelation (Table 6) before discussing the impulse response function. The selected lags in the VAR model exhibit no significant serial autocorrelation at the 5% level, and the parameter estimates remain stable. As a result, the SVAR models with three lags satisfy the stability condition and can be reliably estimated.

3.5 Impulse response and variance decomposition

3.5.1 The response of the balance of payment to shocks

Figure 1 illustrates positive shocks on the exchange rate and other variables included in the balance of payment channel in the long run. In both short run and long run, devaluation negatively affects the balance of payment. Contrary to initial expectations, our findings indicate that devaluation does not have a significant long-run impact. Therefore, we reject the null hypothesis suggesting that devaluation enhances the balance of payments in the long run. Interestingly, we discover that a positive shock exchange rate actually has a negative effect on the balance of payments. The success of devaluation as a strategy to improve the balance of payments depends on the elasticities of exports and imports. In the context of Ethiopia, where elasticity is significantly low, devaluation is not expected to lead to an improvement in the balance of payments. These findings align with those of previous studies conducted by.^57,58

Figure 1. The response of the balance of payment to exchange rate and other macroeconomic shocks in the long run.

Additionally, we find that an initial positive shock in interest rates improves the balance of payments, but this effect diminishes after the fourth quarter in both the short and long run. On the other hand, in the short run, a positive innovation in money supply negatively affects the balance of payments up to the fourth quarter, and after the fourth quarter, its effect is positive. However, in the long run, a positive shock in the money supply initially enhances the balance of payments, but the effect diminishes after the fourth quarter in the long run. Furthermore, the response of the balance of payments to a positive innovation in inflation is negative in the short run and positive up to the third quarter. However, after the fourth quarter, the balance of payments worsens in the long run, which contradicts the findings of.⁵⁹ This divergence highlights the need for further investigation into the dynamics of inflation and its impact on the balance of payments in Ethiopia. Lastly, a positive shock in foreign asset reserves affects negatively up to the third quarter and positively after the third quarter in both the short and long run. Initially, a positive shock on real output has a negative impact on the balance of payments both in the short and long run up to the second quarter; thereafter, its impact is positive.

3.5.2 The response of real GDP (output growth) to shocks

Figure 2 presents empirical evidence on the contemporaneous response of output to various macroeconomic shocks, including positive innovations in the nominal exchange rate, interest rate, money supply, inflation, foreign asset reserves, and balance of payments in the long run. Our analysis reveals that the nominal exchange rate has a negative impact on output in both short run and long run, indicating a contractionary effect. These findings are consistent with prior research by,^15,32,60 providing further support for the hypothesis that positive shocks to the exchange rate adversely affect output. Additionally, we find empirical evidence supporting the null hypothesis, confirming that devaluation has a contractionary effect on output.

Figure 2. The response of output growth to exchange rate and other macroeconomic shocks in the long run.

Furthermore, our study reveals that positive innovations in interest rates have a negative effect on output both in the short and long run, providing empirical support for the monetary theory. This finding strengthens the existing literature by confirming the negative relationship between output and interest rates. Moreover, our analysis reveals that a positive shock in the money supply leads to a positive response in output. This observation suggests that an increase in money supply stimulates output, thereby contributing to economic growth in both the short and long run.

In line with the research conducted by⁵⁹ our study highlights the adverse impact of inflation on output against the findings of.¹² Specifically, we find that an increase in inflation significantly and negatively affects output in both the short and long run, emphasizing the detrimental consequences of rising inflation for overall economic performance. Regarding foreign asset reserves, our analysis indicates that an initial positive shock in these reserves adversely affects output in both the short and long run. However, after three quarters in the long run, the effect becomes positive, suggesting a gradual recovery and the potential for long-term benefits. This dynamic nature of the relationship between foreign asset reserves and output in the long run underscores the importance of considering the timing and persistence of shocks when assessing their effects. In addition, our study reveals a positive contemporaneous response of output to positive shocks in the balance of payments. Specifically, a surplus in the balance of payments improves the level of output, indicating a favourable impact on the economy.

In summary, the findings confirm the contractionary effect of positive shocks in the nominal exchange rate, the inverse relationship between output and interest rates, and the positive impact of money supply shocks on output. Furthermore, our study emphasizes the negative effects of inflation on output and highlights the dynamic nature of the relationship between foreign asset reserves and output.

3.5.3 The response of money supply to shocks

Figure 3 illustrates the response of the money supply (M2) to contemporaneous positive shocks in macroeconomic variables. Specifically, we examine the response of the money supply to positive innovations in the nominal exchange rate, interest rate, inflation, foreign exchange reserves, balance of payments, and output. In the case of a positive innovation in the nominal exchange rate, our findings reveal that the response of the money supply is positive during the first three quarters in both the short and long run. This implies that an initial positive shock to the nominal exchange rate leads to an increase in the money supply in Ethiopia. However, after the third quarter, the response of the money supply declined, although it remained positive. This suggests a diminishing effect over time. Consistent with economic theory, our analysis supports the inverse relationship between the interest rate and money supply in Ethiopia. A positive innovation in the interest rate leads to a decrease in the money supply, aligning with the expected relationship.

Figure 3. The response of money supply (M2) to exchange rate and other macroeconomic shocks in the long run.

Furthermore, our findings indicate that a positive innovation in inflation has a significant effect on increasing the money supply in Ethiopia. This suggests a positive relationship between inflation and money supply, highlighting the impact of inflation on the monetary dynamics of the Ethiopian economy. Regarding foreign exchange reserves, our analysis reveals a dynamic response of money supply. A positive shock in foreign exchange reserves initially leads to an increase in money supply. However, similar to the response to the nominal exchange rate, the effect gradually declines over time. Moreover, our study finds that a positive shock in the balance of payments leads to an increase in money supply in Ethiopia. This implies a positive relationship between the balance of payments and money supply, indicating the influence of external factors on the domestic money supply dynamics in the Ethiopian context in both short run and long run.

Finally, our analysis demonstrates a positive response of money supply to positive shocks in output. This implicates that a positive shock to output leads to a rise in money supply in Ethiopia, highlighting the relationship between output and money supply dynamics in the country.

3.5.4 The response of inflation

Figure 4 presents the impulse response analysis of inflation to exogenous shocks in the innovations of other variables within the system. Specifically, we examine the effects of a positive shock in the exchange rate, interest rate, and money supply on inflation dynamics. Our findings reveal that a positive shock in the exchange rate, indicative of devaluation in the domestic currency, exerts a positive impact on inflation in both the short and long run. This outcome suggests the presence of inflationary pressures resulting from exchange rate movements, whereby devaluation in the domestic currency contributes to higher import costs and subsequently increases domestic prices.⁶ In contrast, the response of inflation to a positive random shock in the innovation of the interest rate demonstrates a negative relationship. This observation implies that an increase in interest rates reduces inflationary pressures within the economy. Moreover, a positive random shock in the money supply is associated with an increase in inflation. This result suggests that an expansionary monetary policy, as reflected by an increase in the money supply, contributes to higher inflationary pressures.

Figure 4. The response of inflation to exchange rate and other macroeconomic shocks in the long run.

Overall, these findings underscore the interconnectedness and dynamics of inflation with respect to exogenous shocks in the exchange rate, interest rate, and money supply. The implications of these relationships provide valuable insights for policymakers in managing inflationary pressures and devising appropriate monetary and exchange rate policies.

3.5.5 The response of foreign exchange reserve

Figure 5 presents the impact of devaluation and other macroeconomic variables on foreign asset reserves. Devaluation of the domestic currency leads to a decline in foreign asset reserves. The response of reserves to positive innovations in the interest rate initially exhibits a negative relationship, but after a few quarters of adjustment, it results in an increase in reserves. Moreover, the response of foreign asset reserves to changes in the money supply is positive, indicating that an expansionary monetary policy contributes to an increase in reserves.

Figure 5. The response of foreign asset reserve to exchange rate and other macroeconomic shocks in the long run.

This study also highlights the positive relationship between inflationary shocks and foreign asset reserves. When inflationary pressures arise, the central authority takes measures to collect funds to mitigate inflation. Consequently, the reserve in banks increases. Furthermore, lagged positive innovations have a significant impact on the current performance of foreign asset reserves. These findings shed light on the dynamics and interplay among currency devaluation, interest rate innovations, money supply changes, inflationary shocks, and foreign asset reserves. The implications of these relationships provide valuable insights for policymakers in managing and optimizing foreign asset reserves in response to various economic shocks and policy interventions.

3.5.6 The response of interest rate

Figure 6 shows the responses of the interest rate to positive shocks in the innovations of other variables in the system. Initially, positive shocks on the exchange rate have a positive effect, but the effect declines after the second quarter. The response of the interest rate to the money supply is negative. An increase in inflation leads to a decline in the interest rate, and the effects of foreign asset reserves initially increase the interest rate, but the interest rate declines a few quarters later. The balance of payment surplus influenced the domestic interest rate to decline. Responses to positive innovation in output led to an increase in the lending interest rate.

Figure 6. The response of interest rate to exchange rate and other macroeconomic shocks in the long run.

3.5.7 Estimation results of Variance decomposition

The variance decomposition analysis of the balance of payment (bop) and output (lrgdp) is used to examine the strengths of various devaluation channels. The study includes forecast horizons ranging from 1 to 10 quarters. The first column of the results displays the forecast periods, while the second column represents the standard error (SE), indicating the forecast error at different quarters. Shocks denoted as Shock1 (e), Shock2 (i), Shock3 (lnms), Shock4 (inf), Shock5 (lnfxre), Shock6 (bop), and Shock7 (lrgdp) are used to represent each variable. Furthermore, the decomposition values for the 1st to 10th forecast horizons are presented in Tables 7 and 8.

The results of our variance decomposition analysis in the short-run shocks on the exchange rate significantly affect the balance of payment next to the foreign exchange reserve. In the long run, foreign exchange rate reserve shocks (Shock 5) have more significant explanatory power in accounting for balance of payment (bop) fluctuations than nominal exchange rate shocks (Shock 1) and output shocks (Shock 7). Thus, the foreign exchange rate reserve channel emerges as a dominant factor in the context of exchange rate policy and its influence on balance of payment dynamics. In contrast, the impact of the exchange rate channel alone appears to be ineffective in the long run, as indicated by its contribution to the balance of payment fluctuations in Table 7. However, the exchange rate channel indirectly affects the balance of payment through other channels, such as the foreign exchange reserve, output, and money supply channels. This highlights the interdependencies and interconnectedness of these channels in shaping balance of payment dynamics in the long run.

The results of our variance decomposition analysis in Table 8 indicate that balance of payment shocks (Shock 6), interest rate shocks (Shock 2), exchange rate shocks (Shock 1), and money supply shocks (Shock 3) have a significant impact on output variations in the long run. Among these shocks, the exchange rate channel emerges as a crucial factor that interacts with the balance of payment and interest rate channels, shaping the dynamics of output fluctuations.