Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4, 186-198 2025 Publisher: Learning Gate DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate © 2025 by the authors; licensee Learning Gate History: Received: 27 January 2025; Revised: 21 March 2025; Accepted: 25 March 2025; Published: 3 April 2025 * Correspondence: Abdulkadir.nur@centralbank.gov.so Exchange rate dynamics and trade balance in Somalia: An ARDL approach Abdulkadir Mohamed Nur1*, Ali Yassin Sheikh Ali1 1Economist, Central Bank of Somalia, Somalia; Abdulkadir.nur@centralbank.gov.so (A.M.N.) 2Faculty of Economics SIMAD University, Somalia; profali@simad.edu.so (A.Y.S.A.) Abstract: This study looks at how Somalia's trade balance has been affected by variations in the exchange rate, addressing the gap in empirical research on the relationship between macroeconomic variables and trade performance in fragile economies. The study used time series data sourced from the World Bank, SESRIC, the Central Bank of Somalia, and the IMF. The study employed augmented Dickey-Fuller tests and Phillips-Perron tests to assess the stationarity of the data. The study utilized an autoregressive distributed lag (ARDL) model to identify both short-term and long-term effects of exchange rate volatility on Somalia’s trade balance. The study found that over time, GDP, along with the real effective exchange rate (REER), had positive impacts on trade balances, whereas inflation rates and foreign direct investment (FDI) negatively influenced Somalia's trade balance in the long run. In the short term, there is a significant positive correlation between GDP, FDI, and REER and the trade balance, whereas the Consumer Price Index (CPI) has a negative impact. These are clear indications that any slight change in economic performance, changes in the value of money circulating, or inflation levels shall be reflected accordingly in Somalia’s trade balances. In terms of trade politics, the government of Somalia should focus on economic diversification, enhancing export competitiveness, and maintaining price stability to achieve sustainable trade and economic growth. Keywords: ARDL, Consumer price index, Foreign direct investment, Gross domestic product, Real effective exchange rate, Trade balance. 1. Introduction Changes in output, market structure, levels, and fiscal and monetary policy all influence the exchange rate. Weerasinghe and Perera [1] showed that the volatility of a country's trade balance, which is the difference between exports and imports produced over a certain period, has become a significant challenge in developing nations. When exports outweigh imports, an economy has a positive trade balance. However, a trade imbalance occurs when imports exceed exports. A positive BOT indicates an improving economy. BOT surpluses can offset the balance of payments deficits. Similarly, Bhat and Bhat [2] noted that the influence of exchange rate fluctuations on an economy has long been recognized as a source of anxiety. However, with the introduction of the generalized floating paradigm, there has been a significant shift in attention to the matter. The fluctuation of the exchange value affects the country's exports and imports and has also impacted its trade balance. Moreover, exchange rate volatility causes trade deficits in Sub-Saharan Africa. The implementation of the Structural Adjustment Program (SAP) of the IMF in Sub-Saharan African nations in the 1980s and early 1990s caused considerable changes in currency rates. One of the SAP initiatives, international trade liberalization, has boosted currency rate volatility, as has the massive expansion in cross-border financial activities [3]. Leeson [4] says that in 1976 when Italy was still in charge of Somalia, a fixed exchange rate was made official. At that time, 1 Italian lira was equivalent to 8 Somali shillings. Somalia's central bank had the authority to regulate financial markets and interfere in the market. After the central government of 187 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate Somalia fell, the Somali shilling stopped caring about its history of being accepted by the public and being legal tender. However, the paper shillings have continued to be accepted and circulated at a positive value. Fiat money's self-fulfilling strategies allow people to accept paper shillings in the absence of a central government [5]. Similarly, the study by Nor, et al. [6] examined the fundamentals of macroeconomics and the exchange rate volatility in Somalia. The study discovered that the use of SOM shillings and the lack of government control in Somalia were to blame for the unstable exchange rate. Mohamud [7] says that there is an imbalance between supply and demand in the money market, which makes people more likely to use the US dollar as their main form of payment and less likely to use SOS shillings. Additionally, Somalia is implementing dollarized digital banking points of sale [8]. The Somalia Shilling witnessed a strong fixed exchange rate physical policy administering the exchange market during the civilian and revolutionary governments. After a decade of civil war, Somalia encountered poor economic development, resulting in trade deficits and exchange rate flotations with a crisis of balance of payments and high debt problems. Mohamud [7] and Nor, et al. [6] looked at Somalia's trade balance and the fluctuation of its currency rates. Both studies said that the shilling was used without any rules, and as far as the author knows, no studies have looked at how the exchange rate affects Somalia's trade balance. So, the goal of this study is to look at Somalia's short- and long-term exchange rates and trade balances in a way that is based on facts. This article uses GDP, the trade balance, the consumer price index, and the exchange rate to show how Somalia could improve its international trade strategies and policies. 2. Literature Review Bussière, et al. [9] looked at how the exchange rate affects the amount and price elasticity of emerging and developing economies in the 21st century. The descriptive and OLS panel data showed that the exchange rate had a big effect on export price quantity elasticities. The exchange rate affects the trade deficit, but the trade balance got better when both exports and imports fell in value. Dogru, et al. [10] examined how the exchange rate affects the US tourism trade balance. ARDL, ECM, and cointegration were used. The study concluded that the international trade balance improves when the dollar falls and worsens when it rises. Arruda, et al. [11] looked at how real depreciations in the exchange rate affect Brazil's trade balance. The paper looked at the most important economic factors, such as consumer goods, capital goods, intermediate goods, and fuel. The study employed the VAR model and ADF. The study found that exchange rates have a positive influence on the trade balance, one of the most examined economic variables, while fuel prices are indicated to have a negative relationship with exchange rates. Additionally, Cristanto and Bowo [12] the economics of Indonesia's trade balance were examined. The study utilizes 2010–2019 time series data collected every quarter. The VECM model was used in the study. Investment passively affects short- and long-term trade balances. According to the study, both short- and long-term trade balances were unrelated to GDP, whereas the exchange rate affected both long- and short-term trade balances. Šimáková [13] used quarterly data from 1997 to 2013 to look at how the exchange rate affects the exports and imports of Slovakia's partners. J-curve, ADF, and cointegration tests were used. The J-curve does not affect Slovakia's aggregate exports and imports. Slovakia's trade balance improves significantly when the real exchange rate falls, but it has little effect when it rises. Furthermore, Rahman and Serletis [14] examined currency rate uncertainty's impact on exports using a multivariate context approach. Mean errors and vector autoregression were used. The study found that actual exchange rate uncertainty undermines US exports. The study also indicated that asymmetrical exchange rate shocks equalize actual exchange accounting uncertainty. However, Thanh [15] investigated the macroeconomic variables affecting Vietnam's trade balance. The study used ADF and VECM models. The trade balance and currency rate were unrelated. Capital mobility and foreign direct investment offset the trade imbalance. The study discovered that trade balances adapt over time to controllable shocks. Additionally, Kingia and Muba [16] analyzed Tanzanian balance of payments factors. OLS, ADF, and Pearson's correlation were employed. The 188 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate analysis found that FDI and the CPI affected Tanzania's balance of payments. Tanzania's balance of payments improves due to the currency rate and interest rate. Furthermore, Chiu and Sun [17] examined the trade balance, the real exchange rate, and savings. The study employed descriptive statistics, the ADF diagnostic test, Pearson's correlations, CIPS test statistics, and the PSTR model. Trade balance and other control variables are unrelated, according to empirical research. From 1960 to 2016, Nathaniel [18] looked at the trade balance of Nigeria and the effective currency rate. The study employed ARDL-bound test estimates. The analysis found that the actual currency value affects the trade balance and that devaluing the Nigerian naira during a trade balance deficit is not a good strategy. Nasir and Leung [19] also looked at the US trade balance using the nonlinear ARDL framework. They examined the fiscal deficit, surplus, savings, and real exchange rate. The study found that domestic savings, price stability, fiscal shortfall, and productivity affect the US trade balance in the long and short term. Hunegnaw and Kim [20] also looked at trade balances and the exchange rate in East Africa. The ADRL model was used to explain variables such as the real exchange value, export-to- import ratio, foreign domestic income, and real GDP. The study found that lowering the actual effective currency rate improves trade openness considerably. After real exchange and trade liberalization, the balance of trade is marginally more responsive to the real effective exchange rate, but it remains inelastic. Vural and Tunaer [21] used the J-Curve and cointegration to look at the balance of trade and the exchange value. The study used an error correction model. According to the study, when the Turkish Lira falls in value, the actual effective exchange value has a negative impact on the trade. Aliyu Jibrilla and Mohammed Tijjani [22] also examined the long-term through of the exchange value into the Nigerian balance of trade using threshold relationships and asymmetric error correction models. The study indicated that actual exchange rate devaluations reduce trade balances. Korbi and Banushaj [23] examined the trade balance, real effective exchange rate, FDI, and remittances. The research used a VAR model. VAR analysis found that there is no association between the Albanian trade balance and actual exchange rates. Chebbi and Olarreaga [24] explored how Tunisia's currency rate affects the agricultural net's external position. The Tunisian trade balance was examined using VECM and cointegration techniques. The Tunisian agriculture sector's net external position suffers from currency depreciation, according to the research findings. Olomola and Dada [25] examined Sub-Saharan Africa's real exchange rate and volatility. Moments were used. In Sub-Saharan Africa, the actual effective currency rate improves the trade openness. High real domestic GDP improves the balance of trade, as the study revealed. The study concluded that foreign income impacts the trade balance. Olayungbo [26] examined how the real effective currency rate, oil price, and reserves impact Nigeria's trade balance. The Granger causal test, ADF test, and VAR model estimated the empirical model. The analysis established a long-term association between Nigeria's trade balance and control factors. The results show that oil prices affect foreign reserves. The analysis shows no overall causal association between commerce, oil prices, or the actual exchange rate. Sakyi and Immurana [27] examined how seaport efficiency, household spending, the money supply, government expenditure, and the real effective exchange rate influenced the trade balances of 27 African nations. The general technique of moments showed that the explanatory factors improved long-term and short-term trade balances. Nyahokwe [28] used the J-curve to study South African agriculture, mining, and industry, the balance of trade in South Africa is examined using the VAR model. According to the findings, South Africa's trade did not improve during the period of real effective currency rate depreciation. Kallianiotis [29] used a VAR model to examine how the real currency rate, income from domestic, relative price level, and foreign investment affected seven nations balance of trade: the US, Canada, the Eurozone, the UK, Switzerland, Australia, and Japan. The study concluded that US currency depreciation raises import prices and lowers export prices. During a devaluation, imports outnumber exports. The analysis indicated that imports exacerbate the balance of trade in the short run but improve it in the long term. 189 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate Hunegnaw and Kim [30] examined East African trade balances using panel data. By using linear and nonlinear ARDL models, it was possible to determine how the real currency rate, domestic GDP, and foreign GDP affected the trade balance. The study indicated that the actual effective currency rate improves manufacturing and mining trade. Research indicates that the currency rate has an effect on agriculture's trade balance over time. Banerjee, et al. [31] examined China and other countries' currency rates, GDP growth rates, and real GDP through temporal trends using OLS. The study concluded that China's trade surplus boosts its non-price competitiveness by increasing exports of skill ladder items and decreasing imports. Eshetu [32] examined how the Birr depression and other control variables (import spending, export revenue, local currency price of exports, nominal exchange rate, export quantity, foreign currency price of imports, domestic income, foreign income, and import quantity) affected Ethiopia's trade balance from 1970 to 2014. VECM revealed that depreciating the birr worsens Ethiopia's balance of trade in the short run but enriches it in the long term. The results also demonstrated that the real effective exchange rate, GDP, money supply, and terms of trade affect the Ethiopian balance of trade in the short and long term. 3. Methodology 3.1. Theoretical Framework Goldstein and Khan [33] developed the underlying hypothesis of the imperfect substitute, which says that imports and exports are not ideal substitutes for domestic commodities. This assumption supported the argument that foreign and domestic goods are perfect substitutes and the evidence from the theoretical studies. Additionally, Rose and Yellen [34] proposed a different concept of short-run and long-run trade balance depreciation and recovery. Therefore, this study adopted the two-country model developed by Goldstein and Khan [33] and Rose and Yellen [34]. The money demand functions may be used to express both domestic and international imports and exports as: 𝑀𝑑 = 𝑓1(𝑌𝑑, �̇�𝑑 , 𝑃𝑑), 𝜕𝑀𝑑 𝜕𝑌𝑑 > 0, 𝜕𝑀𝑑 𝜕�̇�𝑑 < 0, 𝜕𝑀𝑑 𝜕𝑃 > 0 (1) 𝑀𝑓 = 𝑓2(𝑌𝑓 , 𝑒, �̇�𝑓 , 𝑃𝑓), 𝜕𝑀𝑓 𝜕𝑌𝑓 > 0, 𝜕𝑀𝑓 𝜕�̇�𝑓 𝑒 < 0, 𝜕𝑀𝑓 𝜕𝑃 > 0 (2) In equations (1) and (2) where 𝑀𝑑 is notated domestic demand, 𝑌𝑑 is real domestic income, 𝑃𝑑 denotes domestic price and �̇�𝑑 is notated for the price of local currency paid domestically. Also, 𝑀𝑓 is notated foreign demand, 𝑌𝑓 is real foreign income, 𝑃𝑓 is denoted by overall price level of foreign currency, �̇�𝑓 is denoted by foreign currency paid domestically and 𝑒 is the exchange rate. The two demand functions, which are derivatives of the Marshallian demand function, result in a demand elasticity that is negatively correlated with price and positively correlated with income. 𝑃𝑑 = 𝑒 𝑃𝑓 ∗ 𝑃 = 𝑒 ( 𝑃𝑓 ∗ 𝑃 ) ( 𝑃∗ 𝑃∗ ) ≫ 𝑒 ( 𝑃𝑑 ∗ 𝑃 ) ( 𝑃∗ 𝑃∗ ) ≫ ( 1 𝑟 ) 𝑃𝑓 ∗ 𝑇ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒, 𝑃𝑑 = ( 1 𝑟 ) 𝑃𝑓 ∗ (3) Where r is exchange rate relative of the foreign price of goods to domestic goods and 𝑃𝑓 ∗ are export currency for foreign partners and 𝑃∗ are overall price both domestic and foreign. Similarly, the price of freeing partners imports is calculated as; 𝑃𝑓 ∗ = 𝑃𝑥𝑟 (4) Substituting (3) in to (1) and (4) in to (2) will find 𝑀𝑑 = 𝑀𝑑 (( 1 𝑟 ) 𝑃𝑓 ∗, 𝑌𝑑) (5) 𝑀𝑓 = 𝑀𝑓(𝑃𝑥 , 𝑌𝑓) (6) As stated, theoretical framework the hypothesis of assuming a perfect competitive market of two countries (Ds) is denoted by domestic supply and (Fs) is denoted by foreign supply and this could be written as function of export relative price. 190 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate 𝐷𝑠 = 𝐷𝑠(𝑃𝑠) 𝑎𝑛𝑑 𝐹𝑠 = 𝐹𝑠(𝑃𝑓) (7) 𝑀𝑑 = 𝐹𝑠 𝑎𝑛𝑑 𝑀𝑓 = 𝐷𝑠 (8) Making equation (8) as base and using (5) and (6), the trade balance is calculated as follows. 𝑇𝐵 = 𝐷𝑠 𝑀𝑑 ≫ 𝑇𝐵 = 𝑀𝑓 𝑀𝑑 = { 𝑀𝑓(𝑃𝑥𝑟, 𝑌𝑓) 𝑀𝑑 (( 1 𝑟 ) 𝑃𝑓 ∗, 𝑌𝑑) } (9) 𝑇𝐵 = 𝑀𝑓(𝑃𝑥𝑟, 𝑌𝑓) 𝑀𝑑 ( 𝑃𝑓 ∗ 𝑟 , 𝑌𝑑) (10) From equation (10) it can be derived trade balance from the export and import with their relative price ratio as. 𝑇𝐵 = 𝑇𝐵(𝑟, 𝑌𝑑 , 𝑌𝑓) (11) 3.2. Model Specification 𝑇𝐵 = 𝑓(𝑅𝐸𝐸𝑅, 𝐺𝐷𝑃, 𝐶𝑃𝐼, 𝐹𝐷𝐼) (12) Where 𝑇𝐵 = 𝐼𝑀𝑃/𝐸𝑋𝑃 if the ratio rises, the trade balance worsens; if it falls, it improves. It examines the trade balance in actual terms rather than in units. “The real effective exchange rate” uses imports as the numerator and exports as the denominator [35]. REER decreases a currency's actual value. To enhance the trade balance, the ratio of imports to exports must be established [36]. REER stands for "real effective exchange rate," GDP for "gross domestic products," CPI for "consumer price index," and FDI for "foreign direct investment." This is the long-term form of the model: 𝑙𝑛𝑇𝐵𝑡 = 𝛼𝑡 + 𝛽1𝑙𝑛𝑅𝐸𝐸𝑅𝑡 + 𝛽2𝑙𝑛𝐺𝐷𝑃𝑡 + 𝛽3𝑙𝑛𝐶𝑃𝐼𝑡 + 𝛽4𝑙𝑛𝐹𝐷𝐼𝑡 + 𝜇 (13) The following equation describes the ARDL model's short-run dynamics, with variables' co- integration in equations (13) and (14). 𝑙𝑛𝑇𝐵𝑡 = 𝛼 + ∑ 𝛥 𝑝1 𝑖=0 𝛽1𝑙𝑛𝑅𝐸𝑅𝑡−𝑖 + ∑ 𝛥 𝑝2 𝑖=0 𝛽2𝑙𝑛𝐺𝐷𝑃𝑡−𝑖 + ∑ 𝛥 𝑝3 𝑖=0 𝛽3𝑙𝑛𝐶𝑃𝐼𝑡−𝑖 + ∑ 𝛥 𝑝4 𝑖=0 𝛽4𝑙𝑛𝐹𝐷𝐼𝑡−𝑖 + 𝜆𝐸𝐶𝑀𝑡−𝑖 + 𝜇1 (14) Were, 𝛽1 𝑡𝑜 𝛽4 are the coefficients to be assessed, 𝛼 is the constant term, 𝛥 is the differential, ln is natural logarithms, 𝜇 is the error term, and 𝑝1 𝑡𝑜 𝑝4 is the lags. Table 1. Variables Definitions and Notations. Nature of the Variable Variable Definition Notatio n Expected Relation DV Trade Balance The trade balance is defined as 𝑇𝐵 = 𝐼𝑀𝑃 𝐸𝑋𝑃 . If the ratio goes up, it means that the trade balance is getting worse, and if it goes down, it means that the “trade balance” is getting better. TB IV Real effective Exchange rate The buying power of a currency with another currency is described by its exchange rate REER Negative (-) IV Gross Domestic Product All finished monetary value of produced goods and services within a nation GDP Positive (+) IV Consumer Price Index Inflation Rate CPI Negative (-) IV Foreign Direct Investment Foreign Domestic Income FDI Positive (+) 191 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate Table 2. Variable Justification. Variable Expected Relation Reason for expecting (+) or (-) relationship REER Negative (-) Exchange rate depreciation boosts exports and BOT. GDP Positive (+) BOT or BOP surpluses affect economic growth through trade. CPI Negative (-) In an export-driven economy, BOT surpluses undercut domestic prices and discourage imports. FDI Positive (+) Spillovers from FDI will increase the total factor productivity of domestic companies and reduce imports and exports, which will improve BOT. 3.3. Data Source This analysis uses secondary data from 1990 to 2021 from the World Bank, the IMF, SESRIC, and the Central Bank of Somalia. 3.4. Data Estimation Methods To examine the effect of exchange rate changes on the trade balance of Somalia, the study employed the ARDL econometric estimation model. ADF technique was used in this study. The application of Johansen [37] maximum likelihood technique, according to Holden and Perlman [38], eliminates the necessity for unit root tests because the presence of a cointegrating connection between the variables guarantees the existence of unit roots [39]. ARDL was used in the study to assess the co-integration approach for estimating long- and short-run associations as well as dynamic interaction between the research variables. This study uses the ARDL model due to the different integration orders of its variables, with most of them being I (1) while the Consumer Price Index is I(2). The ARDL method is even more appropriate given that there is co-integration among the variables as shown by the Johansen test hence it can capture both short-term and long-term relationships effectively. The four optimal periods of delay according to specific criteria enhance the ability of this model to accurately capture dynamic interactions. For example, it’s advantages in small samples guarantee strong estimates of the relationship among Trade Balance, Foreign Direct Investment, GDP (gross domestic product), Consumer Price Index (CPI) & Real Effective Exchange Rate (REER). The data was analyzed using STATA 17. 4. Empirical Results and Discussions 4.1. Descriptive Statistics of the Data Table 3 Presents a summary of measures of selected economic indicators for Somalia between 1990 and 2021 are provided. The trade balance (TB) exhibited low variability averaging 5.59 USD (σ = 0.457). The foreign direct investment (FDI) mean was 130.194 USD, displaying high variability (σ = 93.203). GDP per capita (σ = 49.596) while the consumer price index (CPI) was 93.14 USD (σ = 43.894). The real effective exchange rate (REER) with an average of 16,004.244 which indicated an alarming degree of exchange rate instability as vouched by (σ = 8,766.053) of the rate. Table 3. Descriptive Statistics of the Data. Variable Obs. Mean Std. dev. Min. Max. TB 32 5.59 0.457 4.98 6.76 FDI 32 130.194 93.203 9.44 349.31 GDP 32 150.43 49.596 90.73 239.05 CPI 32 93.14 43.894 6.22 143.42 REER 32 16004.244 8766.053 1749.17 31558.91 192 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate 4.2. ADF and PP Tests Table 4 Presents the results of the stationarity tests using the binomial Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) test methods alongside their respective results. The trade balance (TB), foreign direct investment (FDI), GDP, and real effective exchange rate (REER) are integrated of order I(1), indicating stationarity after first differencing. The Consumer Price Index (CPI) is integrated of order I(2), meaning that second differencing is required to achieve stationarity. These findings imply that most variables appear to have unit roots in levels but, after differencing, become stationary which supports the necessity of more advanced time series modeling methods. Table 4. Augmented Dickey–Fuller and Phillip–Perron. Augmented Dickey–Fuller Phillip–Perron I. Order Variables Level Difference Level Difference TB -3.578 -4.828 *** -3.398 -5.663*** I (1) FDI -2.201 -4.139*** -3.766 -8.540*** I (1) GDP -2.653 -4.022** -2.380 -4.777*** I (1) CPI -2.202 -5.267** -1.226 -12.184*** I (2) REER -1.657 -3.849** -2.654 -4.723*** I (1) Note: *** p<.01, ** p<.05, * p<.1. 4.3. Lag Selection Criteria Table 5 presents the results of the criteria-based selection of lags. The optimal length of lag is given at the minimum value of the Final Prediction Error (FPE), Akaike Information Criterion (AIC), Hannan-Quinn Information Criterion (HQIC), and Schwarz Bayesian Information Criterion(SBIC). The FPE scores indicate that lag 4 is optimal because it minimizes FPE (0.00) while AIC (-12.5928), HQIC (-11.0655), and SBIC (-7.59701) also suggest blindly accepting the four-lag model. That is, a four-lag model is best for the dynamics in the data set. Table 5. Lag Selection Criteria. Lag LL LR df p FPE AIC HQIC SBIC 0 51.5951 0.000 -3.328 -3.256 -3.090 1 149.432 195.67 25 0.000 0.000 -8.531 -8.095 -7.104 2 173.077 47.29 25 0.005 0.000 -8.434 -7.634 -5.817 3 197.356 48.557 25 0.003 0.000 -8.383 -7.219 -4.576 4 281.299 167.89* 25 0.000 0.000* -12.5928* -11.0655* -7.59701* Note: *optimal lag Endogenous: lnTB lnFDI lnGDP lnCPI lnREER Exogenous: _cons. 4.4. Johansen Cointegration Test Table 6 presents the Johansen cointegration test results with their respective trace and maximum eigenvalue statistics. The trace test result indicates one cointegrating relationship at the 5% level as the trace statistic (43.3974) is lower than the critical value (47.21) at rank 1. Also, the maximum eigenvalue test confirms one cointegrating vector, being the test statistic (22.7791) below the critical value (27.07). These results demonstrate the existence of a long-run equilibrium relationship among the variables, implying that cointegration exists in the model. 193 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate Table 6. Johansen Cointegration Result. Maximum rank Params LL Eigenvalue Trace statistic Critical value 5% 0 5 83.986942 104.8761 68.52 1 14 114.72627 0.86237 43.3974* 47.21 2 21 126.11582 0.52040 20.6183 29.68 3 26 132.80684 0.35058 20.6183 15.41 4 29 135.26529 0.14667 2.3194 3.76 5 30 136.42497 0.07209 Maximum rank Params LL Eigenvalue Maximum Critical value 5% 0 5 83.986942 61.4787 33.46 1 14 114.72627 0.86237 22.7791 27.07 2 21 126.11582 0.52040 13.3820 20.97 3 26 132.80684 0.35058 4.9169 14.07 4 29 135.26529 0.14667 2.3194 3.76 5 30 136.42497 0.07209 Note: * selected rank 4.5. Bound Test Table 7 presents the bound test for cointegration output. From the results, it appears that the F- statistic of 16.02 is greater than the parameters set for both I(0) and I(1) at the 5% significance, which indicates the possibility of the existence of a long-run relationship amongst the variables under consideration. More precisely, the F-statistic exceeds the parameter for I(1) on a 5% level (3.49), meaning that the variables are cointegrated and are inclined to a significant long-term equilibrium relationship. Table 7. Results of Bounds test. Significant Bound test Test statistic Value K I (0) I (1) F-statistics 16.02228* 4 10% 2.2 3.09 5% 2.56 3.49 2.5% 2.88 3.87 Note: *leve relationship. 4.6. Empirical Estimation Table 8 demonstrates the result of both long-run and short-run. Over the long term, there is no significant correlation between any of the variables—FDI, GDP, CPI, or REER—with the trade balance (TB). This means that variation within these variables does not influence the trade balance, and other unknown factors are likely to be at play concerning the trade balance of Somalia. In the short term, FDI does have a positive significant correlation with trade balance (coefficient = 0.030, p = 0.022), implying that increased FDI inflows tend to worsen the trade balance. Likewise, CPI also has a significant negative relationship (coefficient = -0.575, p = 0.048), which means that higher inflation reduces the competitiveness of exports while increasing the volume of imports, thus damaging the trade balance. During the short term, GDP shows some positive but faint relationship (coefficient = 0.540, p = 0.060) which implies that it is plausible for growing economies to tighten the trade balance. Yet, this correlation is only rather modestly significant, and so the result should be taken with a large grain of salt. REER, on the other hand, does not have any remarkable effect on the short-term trade balance which indicates that movements in the exchange rate may not be a contributing factor impacting the trade balance in Somalia. 194 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate The parameter of the trade balance from the previous period (L1.lnTB) in the adjustment equation is -0.652, which shows that the shortfall from equilibrium is eliminated by 65.2% in the subsequent period. This indicates a quite high adjustment speed where the trade balance can attain the long-term equilibrium level faster after being disturbed. Table 8. ADRL Long-run and Short-run Result. D.lnTB Coefficient Std. errs t P>t [95% CI] ADJ lnTB L1. -0.652 0.167 -3.900 0.001 -1.000 -0.304 LR lnFDI -0.022 0.020 -1.110 0.278 -0.064 0.019 lnGDP 0.292 0.428 0.680 0.502 -0.597 1.181 lnCPI -0.348 0.330 -1.050 0.305 -1.035 0.339 lnREER 0.349 0.425 0.820 0.421 -0.536 1.234 SR lnFDI D1. 0.030 0.012 2.470 0.022 0.005 0.056 lnGDP D1. 0.540 0.272 1.990 0.060 -0.026 1.106 lnCPI D1. -0.575 0.274 -2.100 0.048 -1.145 -0.005 lnREER D1. 0.438 0.268 1.630 0.118 -0.120 0.996 _cons -0.921 3.260 -0.280 0.780 -7.700 5.859 4.7. Residual Diagnostic Tests 4.7.1. Serial Correlation Results Table 10 presents serial correlation results confirming that the F-statistic of 5.33 and associated p- value of 0.03 imply that the model is statistically significant at the 5% significance level. This means that the independent variables’ combination has a significant effect on the dependent variable and that these variables explain a considerable amount of variation in the dependent variable. Along with the above, the Obs R squared statistic of 15.18 with a p-value of 0.00 also indicates the significance of the model. This shows that the model fits the data well, and the null hypothesis which states that there is no relationship among the variables is rejected at a 1% significance level. Table 9. Serial Correlation Results. F-statistic 5.325338 Prob. F (2,9) 0.0298 Obs*R2 15.17601 Prob. 𝜒2(2) 0.0005 4.7.2. Heteroscedasticity Test for the Model Table 10 illustrates the results of the heteroscedasticity test suggesting an F-statistic of 1.53 with a p-value associated with 0.24 which indicates we do not reject the hypothesis of homoscedasticity. Furthermore, this shows that the model does not possess significant heteroscedasticity. In addition, the Obs*R² statistic (p-value = 0.25) as well as the Scaled Explained SS (p-value = 1.00) reinforce this claim. Thus, there is no evidence of heteroscedasticity in the model, concluding that the error terms do not vary in magnitude. 195 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate Table 10. Heteroskedasticity Result. F-statistic 1.532476 Prob. F (16,11) 0.2390 Obs*R2 19.32873 Prob. 𝜒2 (16) 0.2520 Scaled explained SS 2.050316 Prob. 𝜒2 (16) 1.0000 4.8. Stability Diagnostic Test Table 11 presents the results from the RESET test indicating a t-statistic value of 1.02 and an F- statistic value of 1.04, both having a p-value of 0.33, which means there is no proof towards model misspecification. Moreover, the likelihood ratio which is 2.78 and a corresponding p-value of 0.10 further confirms the model is correctly specified. Hence, we do not reject the null hypothesis which indicates the model is free from functional form problems. Table 11. Ramsey RESET Test Results. Value df Probability t-statistic 1.021648 10 0.3310 F-statistic 1.043765 (1, 10) 0.3310 Likelihood ratio 2.779866 1 0.0955 4.9. Discussion Exchange rate depreciation tends to enhance export competitiveness although it deteriorates the balance of trade (TB), a case in point being the negative relationship between the real effective exchange rate (REER) and the latter. This is in line with previous research by Bussière, et al. [9] and Arruda, et al. [11] which emphasized how exchange rates shape trading patterns, especially in less developed countries. Hence, it can be put that one way in which Somalia’s trade deficit could be reduced through monetary policy is by focusing on stabilizing its currency. The trade balance and gross domestic product (GDP) exhibit positive relationships. Economic growth (measured by GDP) affects the trade balance positively perhaps by increasing the production capacity leading to high export volumes. This statement concurs with trade surpluses as harbingers of economic growth observations in Thanh [15] and Vural and Tunaer [21]. Therefore, encouraging economic growth in Somalia may improve the country’s trade outcomes. This showcases how the consumer price index (CPI) influences trade balance through rising domestic prices that reduce export competitiveness but increase imports thus leading to a negative relationship between the two economic variables. This implies that inflation should be controlled if we are to have favorable trade balances. This is in line with the research results that indicate that inflation is detrimental to trade performance which were carried out by Cristanto and Bowo [12]. Thus, higher levels of foreign direct investment (FDI) are often associated with a better trade balance. Concerning the findings of Kingia and Muba [16] boosting trade activities is one of the positive effects associated with foreign direct investments. In this regard, policymakers need to create an environment that is favourable for foreign investments. 5. Conclusion and Policy Implication This research analyzes the trade balance, foreign direct investment (FDI) activities, exchange rate fluctuations, inflation rates, and the economic growth of Somalia between 1990 and 2021 through the application of econometric analysis such as the augmented Dickey-Fuller unit root test, Johansen cointegration test, and error correction models. The focus was on determining the impact of these variables on Somalia's economy over a longer and shorter time frame. Results show FDI and trade balance to be of greater weight to the economic growth of Somalia than other factors over a longer duration, having a lesser bearing on the CPI and exchange rate. Conversely, exchange rates alongside FDI appear to be more beneficial in the shorter duration in extending the 196 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate economic growth, whereas, there is a negative impact from the CPI. It was also found that an economy is relatively responsive to shocks, adjusting back to equilibrium after an economic shock at a rate of approximately 65.2%. The results from heteroscedasticity and serial correlation tests prove that the model is strong, having no significant problems of variance or autocorrelation. The model also passes the Ramsey RESET test which indicates that functional form misspecification does not occur. These results are relevant to the economic policies of Somalia. Policymakers need to design a macroeconomic policy framework that encourages FDI and incorporates suitable policies regarding the exchange rate. At the same time, trade policy and inflation policy aimed at managing the economy should be formulated to increase economic activity in Somalia in both the short run and long run. Somalia is encouraged to focus on the development of infrastructure and institutional framework to bring in more foreign investment, especially in rapidly growing areas. Somalia can achieve sustained economic stability and grow in the long run by improving the investment climate. Transparency: The authors confirm that the manuscript is an honest, accurate, and transparent account of the study; that no vital features of the study have been omitted; and that any discrepancies from the study as planned have been explained. This study followed all ethical practices during writing. Copyright: © 2025 by the authors. This open-access article is distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). References [1] E. Weerasinghe and T. R. Perera, "Determinants of balance of trade in the Sri Lankan economy," International Journal of Trade, Economics and Finance, vol. 10, no. 1, pp. 17-24, 2019. https://doi.org/10.18178/ijtef.2019.10.1.631 [2] S. A. Bhat and J. A. Bhat, "Impact of exchange rate changes on the trade balance of India: An asymmetric nonlinear cointegration approach," Foreign Trade Review, vol. 56, no. 1, pp. 71-88, 2021. https://doi.org/10.1177/0015732520961328 [3] D. J. Temitope, "Exchange rate volatility and trade balance in Sub-Saharan African countries: A causal relationship," American Journal of Business, Economics and Management, vol. 5, no. 5, pp. 68-73, 2017. [4] P. T. Leeson, "Better off stateless: Somalia before and after government collapse," Journal of Comparative Economics, pp. 689-710, 2007. https://doi.org/10.1016/j.jce.2007.10.001 [5] W. J. Luther and L. H. White, "Positively valued fiat money after the sovereign disappears: The case of Somalia," GMU Working Paper in Economics No. 11-14, 2011. [6] M. I. Nor, T. A. Masron, and T. T. Alabdullah, "Macroeconomic fundamentals and the exchange rate volatility: Empirical evidence From Somalia," SAGE Open, pp. 1-12, 2020. https://doi.org/10.1177/2158244019898841 [7] I. A. Mohamud, "Exchange rate volatility: Empirical evidence from Somalia in 2010," Journal of Distribution Science, vol. 12, no. 3, pp. 99-103, 2014. https://doi.org/10.13106/jds.2014.vol12.no3.99 [8] A. M. Nur, "The role of digital banking services on commercial banks performance in Somalia: A descriptive and OLS approach," International Journal of Financial Research, vol. 13, no. 4, pp. 16-25, 2022. https://doi.org/10.5430/ijfr.v13n4p16 [9] M. Bussière, G. Gaulier, and W. Steingress, "Global trade flows: Revisiting the exchange rate elasticities," Open Economies Review, vol. 31, pp. 25-78, 2020. https://doi.org/10.1007/s11079-019-09573-3 [10] T. Dogru, C. Isik, and E. Sirakaya-Turk, "The balance of trade and exchange rates: Theory and contemporary evidence from tourism," Tourism Management, vol. 74, pp. 12-23, 2019. https://doi.org/10.1016/j.tourman.2019.01.014 [11] E. F. Arruda, A. C. de Brito, and P. U. de Carvalho Castelar, "Exchange rate and trade balances in Brazil: A Disaggregated Analysis by Major Economic CategorieS," International Journal of Economics and Finance, vol. 14, no. 6, pp. 1-62, 2022. https://doi.org/10.5539/ijef.v14n6p62 [12] F. A. Cristanto and P. A. Bowo, "Determinants of Indonesian trade balance: A Vecm analysis approach," Economics Development Analysis Journal, vol. 10, no. 4, pp. 463-474, 2021. https://doi.org/10.15294/edaj.v10i4.45909 [13] J. Šimáková, "The effects of exchange rate change on the trade balance of Slovakia," European Financial and Accounting Journal, vol. 9, no. 3, pp. 50-66, 2014. https://doi.org/10.18267/j.efaj.124 https://creativecommons.org/licenses/by/4.0/ https://doi.org/10.18178/ijtef.2019.10.1.631 https://doi.org/10.1177/0015732520961328 https://doi.org/10.1016/j.jce.2007.10.001 https://doi.org/10.1177/2158244019898841 https://doi.org/10.13106/jds.2014.vol12.no3.99 https://doi.org/10.5430/ijfr.v13n4p16 https://doi.org/10.1007/s11079-019-09573-3 https://doi.org/10.1016/j.tourman.2019.01.014 https://doi.org/10.5539/ijef.v14n6p62 https://doi.org/10.15294/edaj.v10i4.45909 https://doi.org/10.18267/j.efaj.124 197 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate [14] S. Rahman and A. Serletis, "The effects of exchange rate uncertainty on exports," Journal of Macroeconomics, vol. 31, pp. 500–507, 2009. https://doi.org/10.1016/j.jmacro.2008.12.005 [15] T. T. Thanh, "Determinants of trade balance in vietnam during 1997-2015," International Journal for Applied Information Management, vol. 2, no. 1, pp. 13-25, 2022. https://doi.org/10.47738/ijaim.v2i1.23 [16] M. E. Kingia and D. S. Muba, "Assessment of the determinants of balance of payment in Tanzania," East African Journal of Business and Economics, vol. 4, no. 1, pp. 62-75, 2021. https://doi.org/10.37284/eajbe.4.1.485 [17] Y.-B. Chiu and C.-H. D. Sun, "The role of savings rate in exchange rate and trade imbalance nexus: Cross-countries evidence," Economic Modelling, vol. 52, pp. 1017-1025, 2016. https://doi.org/10.1016/j.econmod.2015.10.040 [18] S. P. Nathaniel, "Does exchange rate have asymmetric impact on trade balance? Fresh insights from combined cointegration," Studies in Business and Economics, vol. 15, no. 1, pp. 259-269, 2020. https://doi.org/10.2478/sbe-2020- 0019 [19] M. A. Nasir and M. Leung, "US trade deficit, a reality check: New evidence incorporating asymmetric and nonlinear effects of exchange rate dynamics," The World Economy, vol. 44, no. 3, pp. 818-836, 2019. http://dx.doi.org/10.2139/ssrn.3439302 [20] F. B. Hunegnaw and S. Kim, "Foreign exchange rate and trade balance dynamics in East African countries," The Journal of International Trade & Economic Development, vol. 26, no. 8, pp. 979-999, 2017. https://doi.org/10.1080/09638199.2017.1327611 [21] B. Vural and M. Tunaer, "Effect of real exchange rate on trade balance: Commodity level evidence from turkish bilateral trade data," Procedia Economics and Finance, vol. 38, pp. 499-507, 2016. https://doi.org/10.1016/S2212- 5671(16)30221-0 [22] A. Aliyu Jibrilla and S. Mohammed Tijjani, "Asymmetric cointegration between exchange rate and trade balance in Nigeria," Cogent Economics & Finance, vol. 3, no. 1, p. 1045213, 2015. [Online]. Available: https://doi.org/10.1080/23322039.2015.1045213 [23] A. Korbi and B. Banushaj, "Relationship between exchange rate and trade balance: "The Case of Albania"," European Journal of Economics and Business Studies, vol. 7, no. 2, pp. 45-53, 2021. https://doi.org/10.26417/646vvc21d [24] H. E. Chebbi and M. Olarreaga, "Investigating exchange rate shocks on agricultural trade balance: The case of Tunisia," The Journal of International Trade & Economic Development, vol. 28, no. 5, pp. 628-647, 2019. https://doi.org/10.1080/09638199.2019.1572774 [25] P. A. Olomola and J. T. Dada, "Real exchange rate, exchange rate volatility and trade balance in Sub-Saharan African countries: A Generalized method of moment (GMM) approach," International Journal of Scientific Research and Management, vol. 5, no. 11, pp. 7395-7403, 2017. https://doi.org/10.18535/ijsrm/v5i11.08 [26] D. O. Olayungbo, "Effects of global oil price on exchange rate, trade balance, and reserves in Nigeria: A frequency domain causality approach," Journal of Risk Financial Management, vol. 12, no. 1, pp. 1-14, 2019. https://doi.org/10.3390/jrfm12010043 [27] D. Sakyi and M. Immurana, "Seaport efficiency and the trade balance in Africa," Maritime Transport Research, pp. 1-10, 2021. https://doi.org/10.1016/j.martra.2021.100026 [28] O. Nyahokwe, "The impact of exchange rate on trade balances in South Africa: A sectorial analysis," International Journal of Humanities and Social Sciences, vol. 9, no. 1, pp. 43–56, 2020. [29] I. Kallianiotis, "Trade balance and exchange rate: the J-Curve," Journal of Applied Finance & Banking, vol. 12, no. 2, pp. 41-64, 2022. https://doi.org/10.47260/jafb/1223 [30] F. B. Hunegnaw and S. Kim, "Exchange rate and sectoral trade balance dynamics: Empirical evidence from eastern Africa panel data," The International Trade Journal, vol. 34, no. 6, pp. 535-551, 2020. https://doi.org/10.1080/08853908.2020.1749736 [31] B. Banerjee, H. Shi, J. Radovan, Y. Sheng, and X. Li, "The impact of the exchange rate and trade composition on China’s Trade Balance Vis-a`-Vis selected partner countries," Comparative Economic Studies, vol. 59, pp. 311-344, 2017. https://doi.org/10.1057/s41294-017-0025-5 [32] F. Eshetu, "Birr devaluation and its effect on trade balance of Ethiopia: An empirical analysis," Journal of Economics and International Finance, vol. 9, no. 11, pp. 103-119, 2017. https://doi.org/10.5897/JEIF2017.0864 [33] M. Goldstein and M. S. Khan, "Income and price effects in foreign trade," Handbook of International Economics, vol. 2, pp. 1041-1105, 1985. https://doi.org/10.1016/S1573-4404(85)01004-9 [34] A. K. Rose and J. L. Yellen, "Is there a J-curve?," Journal of Monetary Economics, vol. 24, no. 1, pp. 53-68, 1989. https://doi.org/10.1016/0304-3932(89)90006-6 [35] M. Bahmani-Oskooee, "Exchange rate uncertainty and trade flows of developing countries," The Journal of Developing Areas, vol. 25, no. 4, pp. 497-508, 1991. [36] M. Bahmani-Oskooee and H. Fariditavana, "Nonlinear ARDL approach, asymmetric effects and the J-curve," Journal of Economic Studies, vol. 42, no. 3, pp. 519-530, 2015. [37] S. Johansen, "Statistical analysis of cointegration vectors," Journal of Economic Dynamics and Control, vol. 12, no. 2-3, pp. 231-254, 1988. [38] R. Holden and A. Perlman, The foundations of microfinance. USA: Microfinance Press, 1994. https://doi.org/10.1016/j.jmacro.2008.12.005 https://doi.org/10.47738/ijaim.v2i1.23 https://doi.org/10.37284/eajbe.4.1.485 https://doi.org/10.1016/j.econmod.2015.10.040 https://doi.org/10.2478/sbe-2020-0019 https://doi.org/10.2478/sbe-2020-0019 http://dx.doi.org/10.2139/ssrn.3439302 https://doi.org/10.1080/09638199.2017.1327611 https://doi.org/10.1016/S2212-5671(16)30221-0 https://doi.org/10.1016/S2212-5671(16)30221-0 https://doi.org/10.1080/23322039.2015.1045213 https://doi.org/10.26417/646vvc21d https://doi.org/10.1080/09638199.2019.1572774 https://doi.org/10.18535/ijsrm/v5i11.08 https://doi.org/10.3390/jrfm12010043 https://doi.org/10.1016/j.martra.2021.100026 https://doi.org/10.47260/jafb/1223 https://doi.org/10.1080/08853908.2020.1749736 https://doi.org/10.1057/s41294-017-0025-5 https://doi.org/10.5897/JEIF2017.0864 https://doi.org/10.1016/S1573-4404(85)01004-9 https://doi.org/10.1016/0304-3932(89)90006-6 198 Edelweiss Applied Science and Technology ISSN: 2576-8484 Vol. 9, No. 4: 186-198, 2025 DOI: 10.55214/25768484.v9i4.5954 © 2025 by the authors; licensee Learning Gate [39] C. A. Sims, "Bayesian skepticism on unit root econometrics," Journal of Economic Dynamics and Control, vol. 12, no. 2- 3, pp. 463–474, 1988.