Asymmetric effects of foreign direct investment and trade openness on economic growth in Somalia: Evidence from a non-linear ARDL approach

Abstract

Globally, the spread of economic integration and the lowering of trade protectionism through multiple trade pacts offered the products of many developing nations the chance to access global markets. The international economics literature deeply discusses the influence of trade liberalization and foreign investment on economic expansion; however, most assume a symmetric association. It is crucial to evaluate the nonlinear connection, as reliance on linear models might produce biased results. Accordingly, this study explored the asymmetric effects of trade openness and FDI inflows on the economic growth of Somalia using yearly data from 1990 to 2020. The outcomes of the non-linear autoregressive distributed lag (NARDL) suggest that both an increase and decrease in FDI strengthen the economic expansion of Somalia in the short-run and long-run. The domestic investment role in stimulating national output was also considerable. Moreover, the study outlined that a decline in trade openness shrinks growth in the long-run. Besides, we observed from the vector error correction modeling (VECM) a one-way causal linkage from labor force, capital, FDI, and trade openness to GDP in the short-run but not in the long-run. Moreover, short-run unidirectional causation from GDP, trade openness, labor force, and capital to FDI is observed. According to the outcomes, the study proposes that policymakers intensify trade liberalization, encourage local investment, and channel foreign investment toward export-oriented industries.

Keywords:

• Trade openness
• non-linear ARDL
• foreign direct investment
• domestic investment
• Somalia

Reviewing Editor:

• Aye Goodness
• University of Agriculture
• Makurdi Benue State
• NIGERIA

Subject:

• International Trade (incl. trade agreements & tariffs)

1. Introduction

Due to the expansion of globalization, the role of free trade and foreign investments in promoting economic development has dominated the literature in recent decades. Theoretically, open economies thrive quicker than closed economies, which allows for a feasible degree of growth. On the basis of endogenous growth theory, trade liberalization could enhance capital formation, domestic producers’ access to international markets, and resource efficiency (Keho, 2017; Rivera-Batiz & Romer, 1991; Rodrik, 1988). Moreover, Romer (1986) and Belloumi (2014) argue that developing countries can only attain advanced technology and upgrade the skills of their workforce via trade liberalization. Through the importation of highly sophisticated manufacturing technologies and knowledge dissemination, many countries became competitive in the global market (Abdi, Zaidi, et al., 2023). For instance, East Asian countries’ extraordinary performance can be attributed partly to their trade openness promoting policies in the 1970s (Stiglitz, 1996). During this period, the export promotion strategies of the developing countries led to the efficient use of domestic resources and productivity gains, which attracted further international investments for industries with a comparative advantage (Burange et al., 2019). Apart from the truth that trade liberalization is beneficial to the economic development of emerging economies, the expansion of the countries that engage in the production and export of low-value products is slow (Abdi, 2023; Hausmann et al., 2007). Besides that, trade openness has a crucial role in attracting foreign capital to emerging economies (Liargovas & Skandalis, 2012). Increased market spillovers from trade openness are likely to raise foreign investments, which boost output growth through improved local competition and the adoption of modern technology.

The significant role of foreign direct investment (FDI) inflows in the growth of developing economies is well documented in the literature (Cambazoglu & Simay Karaalp, 2014; Popescu, 2014; Sirag et al., 2018). Omri and Sassi-Tmar (2015) demonstrated that the accumulation of international investments enhances the growth patterns of the receiving nations. FDI impacts the growth of the host nation through various channels, including augmenting the supply of funds for local investment, enhancing technology transfer and know-how, and creating new jobs (Belloumi, 2014; Klasra, 2011). Moreover, it has a potential spillover effect on domestic firms through the production chain, where local enterprises buy intermediate inputs and sell their domestically made inputs to foreign companies. By achieving economies of scale, lowering costs, and motivating greater production via comparative cost advantages among countries, foreign investments strengthen the export capacity of the host country and raise foreign currency revenues (Banday et al., 2021; Belloumi, 2014).

Even though foreign investments accelerate the growth process of poor economies, its distribution has been uneven for the least developed countries (LDCs) within the OIC that attracted the slightest share of FDI inflows (Dabour, 2000). In Somalia, FDI inflows soared slightly from USD 447 million in 2019 to USD 464 million in 2020 (United Nations Conference on Trade & Development, 2021). The country adopted a new investment promotion strategy in 2020, which included ten areas of focus for foreign investment, including agriculture, fisheries, banking and financial sectors, energy, information and communications technology (ICT), and manufacturing. According to Nor and Masron (2018), FDI inflows not only improve the economic conditions of Somalia but also contribute to the nation’s conflict reduction and peacebuilding processes. On the other hand, it’s notable that Somalia is a small open economy dependent on international trade, which constitutes a considerable portion of its economic activities. However, the country’s export volume has been lower for the past three decades. This can be attributed to that some of Somalia’s export partners imposed non-tariff measures on its livestock and agricultural products, the highest comparative advantage sectors of the economy (Ali Warsame & Hassan Abdi, 2023). The agriculture sector has also seen yields drop due to unfavorable climatic circumstances that decreased the nation’s exports (Abdi et al., 2023; Warsame et al., 2023). According to the Observatory of Economic Complexity (2020), Somalia exported USD 276 million worth of commodities, while the value of imported goods was USD 4.2 billion in 2020. The main goods exported by Somalia include gold, insect resins, oily seeds, sheep and goats, and other animals to its partners, including the United Arab Emirates (UAE), Saudi Arabia, Japan, Bulgaria, and China. Additionally, the top imports were raw sugar, broadcasting equipment, rice, other vegetables, and rolled tobacco sourced from the UAE, Turkey, China, India, and Ethiopia.

The GDP growth rate and trade openness index trends from 1990 to 2020 are depicted in Figure 1. The trade openness of Somalia reached as low as 11% of the country’s GDP in 1990. However, the country’s index has shown an upward trend for the subsequent decade, surpassing 100% in 2007. Although trade openness decreased in 2009 and 2010, it increased dramatically in the succeeding years, peaking in 2013. The trade openness index has remained over 100 in successive years, except for the latest period. Empirically, a nation with a scale of trade openness of 100 percent is a completely open economy, whereas a country with a scale of trade openness index of 0 percent is a fully closed economy. However, the GDP growth rate of Somalia indicated a massive decline from 1990 to 1994, although it exhibited a positive growth rate in 1996. In the subsequent decades, the GDP growth rate was steady, even though it showed a negative growth rate in 2020 due to the COVID-19 pandemic.

Figure 1. Trade openness and GDP growth rate of Somalia from 1990 to 2020. Sources: UNTCAD (2020) and SESRIC (2020).

Plenty of earlier empirical research investigated the role of trade openness in enhancing output levels (Hye et al., 2016; Keho, 2017; Kong et al., 2021; Pahlavani et al., 2005). While some empirics supported the hypothesis of trade-led growth, others confirmed the growth-led exports hypothesis. Burange et al. (2019) argue that productivity increases lower production costs per unit, which improves international export competitiveness. However, another body of empirical studies concluded FDI-led growth, which asserts that FDI inflows stimulate the economic activities of the receiving nation (Güngör & Ringim, 2017; Mwakabungu & Kauangal, 2023; Pandya & Sisombat, 2017). Additionally, some studies admit that trade openness and FDI inflows together boost the economic growth of various economies (Kalai & Zghidi, 2019). In addition to the significance of FDI and trade, the literature has also acknowledged the relevance of capital formation for economic growth. Consequently, the extent of such an effect varies over time depending on the macroeconomic stability, degree of human resources, freedom to trade, capital purchases, and infrastructure of each country. Much of the literature did not clearly conclude the genuine linkage among FDI, trade openness, and economic growth due to the methodologies adopted and the differences in developmental stages. Several empirical studies concluded that remarkable countries experienced opposite repercussions from FDI inflows and trade openness (Belloumi, 2014; Hussein et al., 2023; Hye, 2012; Kumari et al., 2021). It is momentous that the path of causality was also ambiguous. Despite the enormous research in the literature, the undertakings that focus on the case of Somalia are lacking. Notwithstanding that Somalia is an extensively open economy, it is essential to identify whether the hypotheses of trade-led growth and/or FDI-led growth are actual.

In light of this, the present investigation employs time series data from 1990 to 2020 to estimate the asymmetric effects of FDI inflows and trade openness on economic growth in Somalia. For this purpose, we incorporate the non-linear autoregressive distributed lag (NARDL) method for cointegration established by Shin et al. (2014). This method has advantages in comparison to other traditional cointegration approaches. These include that it is robust to endogeneity problems, applicable to regressors irrespective of their order of integration, and performs better with smaller sample sizes. With several exceptions, previous studies on the linkage between FDI, trade openness, and economic growth assumed a linear interaction among the variables. According to the symmetric assumption, both increases and decreases in FDI and trade openness have an equal role on economic growth. Hence, Baharumshah et al. (2017) suggest that adopting a symmetric equation can dramatically mislead estimations when an asymmetric association is present. Also, the study evaluates the short-run and long-run causality among the interested parameters using the Johansen and Juselius cointegration and vector error correction modeling (VECM), respectively.

This research adds to the body of literature in the following ways. First, this investigation offers the first scientific evidence of the contribution of FDI inflows and trade openness to the economic growth of Somalia. Second, in contrast with prior studies, this study quantifies the coefficient responses of FDI and trade openness on economic growth by combining their asymmetric effects into a single model. Third, the study determines the pattern of economic growth in relation to shocks in trade openness and FDI using a cumulative dynamic multiplier. Fourthly, understanding the economic growth response to shifts in international trade and FDI is crucial for effective policymaking. For Somalia, this knowledge aids in formulating strategies to enhance trade performance and attract foreign capital. Insights into trade dynamics guide the prioritization of sectors and the creation of an attractive investment climate. Balanced economic reforms, informed by the insights from this study, can ensure sustainable development and integration into the global economy for Somalia. The following sections will be organized as follows. A review of relevant theoretical and empirical research is covered in the second section. The third section demonstrates the data sources and econometric methodology of the study. The fourth section exhibits the results and discussion, and the final section concludes as well as proposes the relevant policy implications.

2. Literature review

The essence of the connection between capital transfers, trade openness, and economic growth continues to be the subject of intense theoretical and empirical investigation. The relevant literature reveals conflicting results from countries at varying levels of development by adopting various econometric methodologies and sample periods. Many scholars studying the impact of openness and foreign investment on accelerating growth were motivated by Solow’s seminal papers (Solow, 1956, 1957). The papers argued that FDI promotes growth since the discovery of modern technology improves the production process.

In theory, the classical and contemporary growth models have different stances on the role of FDI in economic growth. The neoclassical growth model holds that FDI affects the economy’s production level in the short-run but has little ramification on the long-run growth rate. In the long-run, the repercussions of FDI inflows on the growth level are routed via technological advancement and labor force expansion, which are considered exogenous factors (Kalai & Zghidi, 2019). Exogenous FDI would cause an expansion of the investment level that leads to a rise in short-run per capita income because long-run growth is constrained by the diminishing marginal product of capital (Belloumi, 2014). Correspondingly, the majority of the theories that identify the interplay of trade openness and long-run economic growth rely on models of endogenous technological change. These models demonstrate that emerging economies can attain long-run economic growth, which is determined endogenously, as opposed to the neoclassical growth model, which considers exogenous factors under the premise of increasing returns to scale.

The significance of trade openness on the output level is well documented in the literature. Using panel data from countries of different development levels, Yanikkaya (2003) applied several openness measures to delve into the association between trade openness and growth. The results were in line with the theory that trade supports growth through various channels, including scale economies, technology transfers, and comparative advantage. In parallel, Keho (2017) demonstrated using the ARDL approach that domestic investment and trade openness have a favorable impact on economic growth. By applying a similar methodology, Kong et al. (2021) observed that trade openness promotes the short- and long-run quality of economic growth in China. Some research employed exports and imports as measures of trade openness to explore the impact of trade openness on economic performance (Klasra, 2011). In the long-run, the results of these undertakings supported the hypotheses of export-led growth and growth-led exports by confirming the two-way causal connection among trade openness and economic growth (Burange et al., 2019; Hye & Lau, 2014). Moreover, Hye et al. (2016) and Pahlavani et al. (2005) adopted individual trade indicators and composite trade openness indexes by observing that openness enhances economic performance in the short- and long-run. They outlined that human and physical capitals also enhance the economic growth of China and Iran, respectively. Furthermore, Udeagha and Ngepah (2021) examined the asymmetric impacts of trade openness on output levels in South Africa. The NARDL framework results confirmed that the short- and long-run repercussions of trade openness on economic growth were asymmetric.

Besides the fact that trade liberalization promotes economic growth, it also attracts foreign capital to the host country. By implementing the Johansen Cointegration approach in Nigeria and South Africa, Güngör and Ringim (2017) and Tshepo (2014) highlighted that FDI inflows increase the economic performance of the host nation in the long-run. Similarly, Pandya and Sisombat (2017) argued that FDI inflows not only augment the GDP of the receiving country but also enhance its export performance and employment. In addition, Cambazoglu and Simay Karaalp (2014) confirmed from the vector auto-regression (VAR) model the long-run association among economic growth, inward FDI, and exports in Turkey. Although increasing FDI levels encourage economic expansion, some researchers suggest that expanding macroeconomic indicators are encouraging to potential foreign investors. For example, Omri and Sassi-Tmar (2015) confirmed from the GMM estimator the interrelationship among foreign investment and economic growth in Tunisia, Morocco, and Egypt. Using yearly data from 1970 to 2014, Sirag et al. (2018) observed from different cointegration methods that FDI is a major driver of Sudan’s economic growth. Likewise, Popescu (2014) exhibited that FDI is a mechanism in the market transition, while there are differences in FDI-assisted development strategies among the Central and Eastern Europe (CEE) countries. Moreover, Amin et al. (2022) studied the short- and long-run asymmetric effects of FDI on Romanian economic growth from 1990 to 2019. The findings show that both an increase in FDI and a decrease in FDI have a favorable and substantial influence on economic growth.

A number of investigations have studied the extent of a country’s dependence on trade and FDI inflows, asserting that they enhance economic upsurge. By applying the ARDL approach, Kalai and Zghidi (2019) disclosed the existence of a one-way long-run causation from FDI to economic growth in Middle Eastern and North African (MENA) countries. Likewise, Sakyi et al. (2015) found that the pairing of FDI and exports has been pivotal in encouraging economic growth and confirmed the critical role of domestic investment in FDI-driven growth in Ghana. Moreover, the empirical findings of Banday et al. (2021) revealed that FDI and trade openness have a favorable influence on long-run GDP growth in BRICS nations. The study has also discovered that gross capital formation has a long-run association with economic growth. According to Klasra (2011), the long-run openness-growth interaction is still valid for Pakistan, while the growth-driven exports theory is empirically supported in Turkey. The study also established a bidirectional causal association between Turkey’s FDI and exports and Pakistan’s exports and trade openness. Despite FDI and trade openness being linked to long-run economic development in the literature, other researchers found the opposite that FDI inflows hamper economic growth (Belloumi, 2014; Hye & Lau, 2014; Kumari et al., 2021). The authors conducted various cointegration techniques and confirmed the presence of an association between FDI inflows and economic growth in the long-run. Comparably, Bilas (2020) and Kumari et al. (2021) conducted various cointegration techniques and concluded no evidence of a long-run connection between FDI inflows and economic growth across countries.

Even though the prior studies concluded with mixed results, they mainly focused on the linear linkage between trade openness, FDI, and economic growth. This indicates that the bulk of the literature disregarded potential asymmetries in trade and FDI's impact on economic growth. Additionally, the methodologies implemented for existing studies failed to convey the causal inference of the short- and long-run asymmetric ties between FDI and economic performance. However, Amin et al. (2022) and Udeagha and Ngepah (2021) are a pair of studies that examined the asymmetric impacts of FDI and trade openness on economic performance individually. The first study investigated the FDI-growth nexus, while the latter focused on trade openness-growth ties. Notwithstanding that, the prior studies extensively investigated the significance of openness to international markets and the attraction of foreign investment for economic growth in many countries and regions. Hence, the studies that examined the relationship in Somalia are missing in the literature. Given this backdrop, his paper examines the asymmetric correspondence between FDI, trade openness, and economic growth in Somalia. Unlike previous studies that relied on linear models, this study implements a non-linear ARDL approach to observe the long-run and short-run magnitudes of the effect of increases and decreases in FDI and trade openness on economic growth.

3. Materials and methods

3.1. Sampling and data sources

This study aims to estimate the asymmetric effects of FDI and trade openness on the economic growth of Somalia during the period 1990–2020. Since the collapse of the Somali Republic in early 1990, the country has mainly been reliant on international trade, which has become a large portion of the nation’s economy. Despite insufficient FDI inflows before the civil war, FDI inflows are now increasing and influencing the country’s output levels, employment, and stability (Nor & Masron, 2018). To achieve the objectives of the study, gross domestic product (GDP) serves as the dependent variable, whereas FDI inflows, trade openness, capital, and labor are the explanatory variables. GDP, expressed in millions of US dollars at constant 2015 prices, is used to estimate the country’s economic growth (Klasra, 2011). FDI is the amount of foreign investment net inflows as a percent of GDP (Amin et al., 2022). Dissimilar to most of the previous studies, which emphasized exports as a gauge of trade openness while ignoring the role of imports, the study utilized the total of exports and imports as a percent of GDP (Hye & Lau, 2014; Keho, 2017). The country’s total labor force was also considered to estimate the role of labor on the output level (Keho, 2017). Additionally, gross fixed capital formation, expressed in millions of US dollars at constant 2015 prices, is used as a proxy for capital. It shows the aggregate of gross additions to fixed assets and changes in stocks during a period of account and the net acquisition of valuables (Pahlavani et al., 2005). The data for these variables were gathered from the World Development Indicators (WDI), United Nations Conference on Trade and Development (UNTCAD), and SESRIC databases.

3.2. The econometric model

To empirically assess the asymmetric effects of trade openness and FDI on the economic growth of Somalia, the econometric model of this undertaking follows the previous framework of Keho (2017), Banday et al. (2021), Udeagha and Ngepah (2021), and Amin et al. (2022). These studies augmented their models with FDI inflows, trade openness, capital, labor, and economic growth. To reduce the problem of heteroskedasticity and to make our interpretation straightforward in percentage form, the study transformed all examined variables into a natural logarithm. Therefore, the following specification is utilized for this study: (1) $$\mathit{\text{ln}}{Y}_t={\beta }_0+{\beta }_1\mathit{\text{ln}}{K}_t+{{\beta }_{2}l\mathit{\text{nLF}}}_t+{\beta }_3{\mathit{\text{lnFDI}}}_t+{\beta }_4\mathit{\text{ln}}{\mathit{\text{TO}}}_t+{\mathrm{Ɛ}}_t$$(1) where Y, K, LF, FDI, and TO stand for gross domestic product, capital, labor force, foreign direct investment, and trade openness, respectively. β₀ is the intercept term, β₁, β₂, β₃, and β₄ are coefficients to be approximated, ln is the logarithm function, and ε_t is an error term.

3.3. Asymmetric autoregressive distributed lag (NARDL) model

A critical weakness of the linear ARDL technique is that it does not acknowledge the non-linear connection between variables. Under the symmetric ARDL, both positive and negative changes in FDI and trade openness have an equal impact on economic growth. However, Shin et al. (2014) proposed non-linear ARDL, which is a relatively recent technique for finding both long-run and short-run asymmetries among variables. Hence, to consider the two periods of increases and decreases simultaneously, this study employs a NARDL approach. The NARDL has the credibility to grasp asymmetries that are due to unpredictable incidents in macroeconomic variables such as financial and economic turmoil as well as political instability. This technique is suited to capture these influences and explain the asymmetric impacts of FDI and trade openness on economic performance in Somalia. This approach is superior to other conventional cointegration methods such as Engle and Granger as well as Johansen cointegration in estimating variables regardless of whether they are stationary at the level I(0), the first difference I(1) or both. Moreover, it is more robust and appropriate for a small sample size and allows for predicting the asymmetric cointegration model in a single equation. The following is the basic asymmetric cointegration regression equation as proposed by Shin et al. (2014). (2) $$Y_t={\phi }^{+}{\mathit{\text{xt}}}^{+}+{\phi }^{-}{\mathit{\text{xt}}}^{-}+{\epsilon }_t$$(2) where ${\phi }^{+}$ and ${\phi }^{-}$ are the long-run partial sum of positive and negative shocks that arise in $Y_t$ while xt is a k × 1 vector of regressors. In order to account for the asymmetries in the association between FDI, trade openness, and economic growth, the initial step in the asymmetric cointegrating relationship under the NARDL specification is to decompose FDI and trade openness into partial sum processes. As a result, our non-linear specification of equation (1)(1) $$\mathit{\text{ln}}{Y}_t={\beta }_0+{\beta }_1\mathit{\text{ln}}{K}_t+{{\beta }_{2}l\mathit{\text{nLF}}}_t+{\beta }_3{\mathit{\text{lnFDI}}}_t+{\beta }_4\mathit{\text{ln}}{\mathit{\text{TO}}}_t+{\mathrm{Ɛ}}_t$$(1) is as follows: (3) $$\mathit{\text{ln}}{Y}_t={\beta }_0+{\beta }_1\mathit{\text{ln}}{K}_t+{{\beta }_{2}l\mathit{\text{nLF}}}_t+{\beta }_3{\mathit{\text{lnFDI}}}_t^{+}+{\beta }_4{\mathit{\text{lnFDI}}}_t^{-}+{\beta }_5{\mathit{\text{lnTO}}}_t^{+}+{\beta }_6{\mathit{\text{lnTO}}}_t^{-}+{\mathrm{Ɛ}}_t$$(3)

where ${\mathit{\text{lnFDI}}}_t^{+},{\mathit{\text{lnFDI}}}_t^{-},{\mathit{\text{lnTO}}}_t^{+} \mathit{\text{and}} {\mathit{\text{lnTO}}}_t^{-}$ are the long-run partial sums of positive and negative changes in FDI and trade openness which are defined as follows. (4) $${\mathit{\text{lnFDI}}}_t^{+}=\sum _{i=1}^t{\Delta \mathit{\text{lnFDI}}}_i^{+}=\sum _{i=1}^t\mathrm{\text{max}}\left({\Delta \mathit{\text{lnFDI}}}_i,0\right)$$(4) (5) $${\mathit{\text{lnFDI}}}_t^{-}=\sum _{i=1}^t{\Delta \mathit{\text{lnFDI}}}_i^{-}=\sum _{i=1}^t\mathrm{\text{min}}\left({\Delta \mathit{\text{lnFDI}}}_i,0\right)$$(5) (6) $${\mathit{\text{lnTO}}}_t^{+}=\sum _{i=1}^t{\Delta \mathit{\text{lnTO}}}_i^{+}=\sum _{i=1}^t\mathrm{\text{max}}\left({\Delta \mathit{\text{lnTO}}}_i,0\right)$$(6) (7) $${\mathit{\text{lnTO}}}_t^{-}=\sum _{i=1}^t{\Delta \mathit{\text{lnTO}}}_i^{-}=\sum _{i=1}^t\mathrm{\text{min}}\left({\Delta \mathit{\text{lnTO}}}_i,0\right)$$(7)

The fundamental step of the NARDL technique is to determine the long-run asymmetric linkage between the variables using the bounds test F-statistic. The null hypothesis of no asymmetric cointegration is set as H₀: ${\phi }^{+}$= ${\phi }^{-}$ against the alternative hypothesis of asymmetric cointegration H₁: ${\phi }^{+}$≠ ${\phi }^{-}.$ The calculated F-statistic is compared to the asymptotic critical values proposed by Narayan (2005). Based on the estimated F-statistics comparison of the lower bound I(0) and upper bound I(1) critical values, it is decided whether to refuse the null hypothesis or not. Typically, there are three possible outcomes. Firstly, when the value of the F-statistics is beyond the upper bound, the null hypothesis of no cointegration is rejected. Secondly, if the value of the F-statistics is less than the lower bound value, one fails to reject the null hypothesis. Thirdly, if the F-statistics value falls between the upper bound and the lower bound value, it can be inconclusive to decide the existence of cointegration.

According to Shin et al. (2014), the asymmetrical cointegration equation is specified by including the positive and negative sums on the equation as follows. (8) $${\Delta \mathit{\text{lnY}}}_t={\alpha }_0+{\phi }_1{\mathit{\text{lnY}}}_{t-1}+{\phi }_2{\mathit{\text{lnK}}}_{t-1}+{\phi }_3{\mathit{\text{lnLF}}}_{t-1}+{\phi }_4^{+}{\mathit{\text{lnFDI}}}_{t-1}^{+}+{\phi }_5^{-}{\mathit{\text{lnFDI}}}_{t-1}^{-}+{\phi }_6^{+}{\mathit{\text{lnTO}}}_{t-1}^{+}+{\phi }_7^{-}{\mathit{\text{lnTO}}}_{t-1}^{-}+\sum _{i=1}^p{\beta }_k{\Delta \mathit{\text{lnY}}}_{t-k}+\sum _{i=1}^q{\delta }_k{\Delta \mathit{\text{lnK}}}_{t-k}+\sum _{i=1}^q{\rho }_k{\Delta \mathit{\text{lnLF}}}_{t-k}+\sum _{i=1}^q{{\gamma }_k^{+}\Delta \mathit{\text{lnFDI}}}_{t-k}^{+}+\sum _{i=1}^q{\gamma }_k^{-}{\Delta \mathit{\text{lnFDI}}}_{t-k}^{-}+\sum _{i=1}^q{\lambda }_k^{+}{\Delta \mathit{\text{lnTO}}}_{t-k}^{+}+\sum _{i=1}^q{\lambda }_k^{-}{\Delta \mathit{\text{lnTO}}}_{t-k}^{-}+{\epsilon }_t$$(8) where α₀ is the intercept, $\phi$ represents the coefficient of long-run variables, $\beta ,\delta ,\rho ,\gamma ,$ and $\lambda$ are the coefficients of the short-run, k represents the lagged values, Δ is the operator of the first difference, p and q characterize the number of lags. We will assess the short-run and long-run non-linear effects by applying the Wald test. The short-run asymmetric effect is captured if ${\sum }_{t=1}^q{\gamma }_k^{+}={\sum }_{t=1}^q{\gamma }_k^{-}$ and ${\sum }_{t=1}^q{\lambda }_k^{+}={\sum }_{t=1}^q{\lambda }_k^{-}.$ However, the long-run consequences of positive and negative changes in FDI and trade openness on economic growth are obtained as ${\phi }_4^{+}={\phi }_5^{-}$ for foreign direct investment whereas ${\phi }_6^{+}={\phi }_7^{-}$ for trade openness.

Additionally, we also calculate the short-run cumulative dynamic multiplier effect of a unit change in ${\mathit{\text{lnFDI}}}_{t-1}^{+},$ ${\mathit{\text{lnFDI}}}_{t-1}^{-},$ ${\mathit{\text{lnTO}}}_{t-1}^{+}$ and ${\mathit{\text{lnTO}}}_{t-1}^{-}$ using the following formula: (9) $$m_t^{+}=\sum _{i=0}^h\frac{{\partial \mathit{\text{lnY}}}_{t+k}}{\partial {\mathit{\text{lnFDI}}}_{t-1}^{+}},m_t^{-}=\sum _{i=0}^h\frac{{\partial \mathit{\text{lnY}}}_{t+k}}{\partial {\mathit{\text{lnFDI}}}_{t-1}^{-}},h=0,1,2,\dots$$(9)

Note that h → ∞, $m_t^{+}$ → ${\phi }_4^{+}$ and $m_t^{-}$ → ${\phi }_5^{-}$ (10) $$m_t^{+}=\sum _{i=0}^h\frac{{\partial \mathit{\text{lnY}}}_{t+k}}{\partial {\mathit{\text{lnTO}}}_{t-1}^{+}},m_t^{-}=\sum _{i=0}^h\frac{{\partial \mathit{\text{lnY}}}_{t+k}}{\partial {\mathit{\text{lnTO}}}_{t-1}^{-}},h=0,1,2,\dots$$(10)

Note that h → ∞, $m_t^{+}$ → ${\phi }_6^{+}$ and $m_t^{-}$ → ${\phi }_7^{-}$

4. Empirical results and discussion

4.1. Summary statistics

The descriptive summary of the series and the correlation matrix of the parameters are demonstrated in Table 1. Panel A of the table illustrates the average, median, maximum, minimum, and standard deviation, among others. We can observe that the lnGDP indicator has the highest mean of 3.067 as well as a maximum of 3.206, while lnLF has the lowest average of 0.274, and lnFDI exhibited the smallest minimum value of -0.732. In addition, the lnFDI indicator has shown the highest standard deviation of 0.84. As shown in Panel A, all the variables have normal distributions with constant variance and zero covariance, as indicated by the Jarque–Bera statistic. Nevertheless, Panel B highlights the correlation between the series. Trade openness, FDI, labor force, and gross fixed capital formation were positively associated with the real economic output of Somalia during the period of the study. The correlation test further suggests the absence of multicollinearity among the study’s variables.

Table 1. Descriptive information and correlation matrix.

Panel A: Summary of descriptive statistics
	lnGDP	lnK	lnLF	lnTO	lnFDI
Mean	3.067	2.399	0.274	1.843	0.294
Median	3.069	2.397	0.273	1.895	0.106
Maximum	3.206	2.519	0.460	2.125	1.471
Minimum	2.937	2.287	0.130	1.070	−0.732
Standard deviation	0.085	0.073	0.110	0.264	0.842
Skewness	−0.002	0.125	0.138	−1.428	0.116
Kurtosis	1.858	1.722	1.690	4.534	1.315
Jarque-Bera	1.684	2.190	2.315	13.568	3.737
Probability	0.431	0.335	0.314	0.001	0.154
Observations	31	31	31	31	31
Panel B: Correlation matrix
lnGDP	1
lnK	0.906	1
lnLF	0.851	0.574	1
lnTO	0.413	0.032	0.797	1
lnFDI	0.885	0.661	0.930	0.701	1

4.2. Unit root test

To conduct the NARDL cointegration technique, it is imperative to ascertain the stationarity of the variables to avoid biased results. The initial analysis of the time-series data is to establish the integration order of the variables. The NARDL approach can be implemented irrespective of whether the variables are stationary in the order of I(0), I(1), or a combination of both. We utilized various unit root tests, such as the Augmented Dickey-Fuller (ADF) and Philips–Peron (PP). The null hypothesis (H₀) of the ADF and PP tests is that the series has a unit root, whereas the alternative hypothesis (H₁) specifies that the series is stationary. The results shown in Table 2 indicate that most of the series have a unit root at level, but they are all purely stationary at I(1). This signifies the appropriateness of the NARDL technique in this study, where different orders of integration were observed.

Table 2. Unit root tests.

ADF Test
	Level		At first difference
Variables	Constant	Constant with trend	Constant	Constant with trend
lnGDP	−0.0013	−18.589***	−4.469***	−5.243***
lnK	1.571	−1.961	−4.887***	−3.914**
lnLF	−0.876	−8.563***	−4.288***	−3.012
lnTO	−2.180	−1.520	−4.947***	−4.124**
lnFDI	−0.139	−3.056	−6.608***	−6.442***
PP Test
	Level		At first difference
Variables	Constant	Constant with trend	Constant	Constant with trend
lnGDP	−0.507	−4.390***	−4.610***	−5.385***
lnK	−1.593	−6.110***	−4.976***	−6.043***
lnLF	1.655	−5.358***	−2.277	−1.173
lnTO	−4.313***	−0.479	−4.936***	−6.018***
lnFDI	−0.067	−3.059	−6.589***	−6.423***

Note: ** and *** symbolize significance at 5% and 1%, respectively.

4.3. Nonlinearity of BDS test

To determine the nonlinearity of the interested parameters, we employed the BDS test displayed in Table 3. This method was postulated by Broock et al. (1996) to observe and examine the projected residuals of the time series model, which are identically distributed errors. The H₀ of this method is that the variables are evenly distributed, which means that the data series are dependent (linear). In contrast, the alternative hypothesis present that the data series violate the normality distribution, which implies that the data series are independent. Hence, the z-statistics of the entire sampled variables, except capital and FDI, are significant, leading to the denial of the null hypothesis. This verifies the nonlinearity of the data parameters, which further emphasizes the suitability of nonlinear ARDL in the study.

Table 3. Nonlinearity of BDS test.

	lnGDP		lnGFCF		lnLF		lnTO		lnFDI
Dimension	BDS Stat	z-Stat	BDS Stat	z-Stat	BDS Stat	z-Stat	BDS Stat	z-Stat	BDS Stat	z-Stat
2	−0.013	−0.868	−0.003	−0.273	0.197	9.111	0.127	5.672	0.019	0.846
3	0.07	2.363	−0.003	−0.137	0.342	9.369	0.202	5.458	0.053	1.42
4	0.151	3.344	−0.005	−0.159	0.445	9.654	0.236	5.201	0.076	1.643
5	0.227	3.795	0.052	1.126	0.515	10.132	0.243	4.945	0.085	1.712
6	0.295	3.999	0.106	1.798	0.561	10.766	0.228	4.649	0.075	1.503

4.4. Bounds test

After noticing the integration order of the variables, we further our analysis by checking the prevalence of long-run cointegration among the explained variable and the regressors. Nevertheless, the bounds testing approach is implemented to inspect the symmetric and asymmetric cointegration of the series by testing the linear and non-linear models. The test procedure is based on the critical values arranged by Narayan (2005). Table 4 summarizes the results of the F-bounds cointegration test. The outcomes from the linear model exhibit that the calculated F-statistics is 3.303, which falls between the lower and upper bound critical values. Thus, the long-run cointegration of the linear model is inconclusive. However, the findings of the non-linear model demonstrate that the F-statistics is 7.306, which is higher than the upper bound critical values. Therefore, we refuse the null hypothesis of no long-run non-linear cointegration among the variables. This ascertains that the long-run equilibrium cointegration relationship between lnK, lnLF, lnTO, lnFDI, and lnGDP is asymmetric. Baharumshah et al. (2017) suggest that using symmetric equations can drastically misinterpret estimations when asymmetric cointegration is present. Since the variables do not exhibit any violation of the diagnostic tests, we can advance by estimating the long-run parameters.

Table 4. F-bounds test.

Model specification	Lower bound	Upper bound	F-statistics	Conclusion
Linear	2.62	3.9	3.303	Inconclusive
Non-linear	2.04	3.24	7.306	Cointegration

Note: the critical values, at the 5% significance level, are based on Narayan (2005).

4.5. Long-run and short-run results

The long-run coefficients were estimated subsequent to ascertaining the long-run asymmetric ties among the variables. The long-run estimates of the coefficients are demonstrated in Table 5 using the NARDL technique. The study found a strong empirical association among the explanatory and dependent variables since they are all significant at the 1% and 5% threshold levels except for positive shocks in trade openness. The outcomes indicate that in the long-run, the coefficient of capital has a favorable effect on GDP in Somalia. This implies that a 1% rise in average capital formation enhances economic growth by 1.309%. Our findings suggest that domestic investment activities, including equipment purchases, plants, and machinery improve the economic output of Somalia. These results of the positive contribution of domestic investments to economic growth are equivalent to the prior results of Pahlavani et al. (2005), Hye (2012), and Hye et al. (2016). On the contrary, the outcomes indicate a negative association between the labor force and GDP. The results reveal that a percentage expansion in the labor force will lead to, on average, a 1.168% decline in the economic growth of Somalia. This outcome is similar to Keho (2017), who observed that the labor force hampers the economic growth of Cote d‘Ivoire.

Table 5. Long-run coefficient estimates.

Variable	Coefficient
lnK	1.309***
	(47.289)
lnLF	−1.168**
	(-3.008)
lnFDI^+	0.103***
	(3.713)
lnFDI^—	0.172***
	(2.726)
lnTO^+	0.084
	(1.072)
lnTO^—	−0.380*
	(-1.828)

Note: *, **, *** signifies significance at 10%, 5% and 1%, respectively. Numbers in parenthesis indicate the t-statistic.

Moreover, the long-run asymmetric findings demonstrated that the coefficients of lnFDI + and lnFDI—are favorable for economic growth. Interpretively, a 1% positive shock in FDI (positive shocks in the partial sum of FDI) improves the economic growth of Somalia by 0.103%. Similarly, a 1% negative shock in FDI (negative shocks in the partial sum of FDI) significantly improves economic growth by 0.172%. This can be interpreted that the increases and decreases of FDI tend to enhance the economic growth of Somalia in the long-run, even though a negative shock has a stronger effect on growth than a positive shock. Another striking result from the study is that the decline in trade openness (negative shocks in the partial sum of TO) adversely affects the national output of Somalia. This can be interpreted as a 1% negative shock in trade openness leads to a 0.380% fall in economic growth. This result implies that trade activities are crucial for the economy where any kind of restriction on the export and import of Somalia will diminish its growth level. However, the findings disclose that the asymmetric linkage between increases in trade openness and GDP was found to be statistically insignificant. This suggests that long-run economic performance in Somalia is not influenced by an increase in trade openness (positive shocks in the partial sum of TO). These findings contradict the outcome of Sriyana and Afandi (2020), who ascertained that in the long run, negative shocks in trade openness do not influence growth for the Philippines and Singapore, while positive shocks in trade openness enhance economic performance for these countries.

Besides the long-run parameters is estimating the short-run effects and the error correction term (ECT). The short-run estimates of the regressors and the ECT are exhibited in Table 6. The results propose that capital and labor have a favorable and statistically significant effect on GDP. A 1% increase in domestic investment and labor force in Somalia leads, on average, to a 0.63% and a 2.71% increase in output growth, respectively. It indicates that capital formation and labor force are crucial engines for the economic performance of Somalia in the short-run. This finding is equivalent to Hye and Lau (2014), who found similar results in India. The most remarkable finding is that the coefficient of increase in FDI has a constructive and statistically significant impact on GDP. Interpretively, a 1% upturn in the positive shocks of FDI enhances economic growth by 0.036%, on average, in the short-run. Likewise, the previous lag in the decrease in FDI indicated a favorable impact on GDP in the short-run. Interpretively, economic growth rises by 0.074%, on average, for a percentage increase in the negative shocks of FDI at lag one. This means that FDI changes have a significant contribution on economic growth in Somalia.

Table 6. Short-run dynamic effect and error correction model.

Variable	Coefficient
ΔlnGDPt-1	−0.015
	(-0.170)
ΔlnLF	2.713*
	(2.098)
ΔlnLFt-1	−2.291
	(-1.663)
ΔlnK	0.631***
	(9.500)
ΔlnFDI^+	0.036**
	(2.656)
ΔlnFDI^+t-1	0.006
	(0.676)
ΔlnFDI^—	−0.009
	(-0.292)
ΔlnFDI^–t-1	0.074*
	(1.911)
ΔlnTO^+	−0.082**
	(-2.780)
ΔlnTO^–	0.078
	(0.872)
ECT(-1)	−0.905***
	(-3.330)
R^2	0.929

Note: *, **, *** present significance at 10%, 5% and 1%, respectively. Numbers in parenthesis signify t-statistic.

In addition, the short-run coefficient of an increase in trade openness has an unfavorable impact on GDP. Interpretively, a 1% rise in the positive shocks of trade openness hampers economic growth by 0.082% on average. This indicates that the tendency of trade liberalization is not productive for the economy in the short-run, which demands cautious trade policies. Equivalently, Sriyana and Afandi (2020) found that short-run increases in trade openness hamper the economic performance of Indonesia. Conversely, the short-run previous lags in GDP, labor force, and increase in FDI were found to be inconsequential in explaining the economic growth of Somalia. Additionally, the negative shocks in FDI and trade openness were not statistically different from zero. Furthermore, the ECT, also known as the cointegration term, displays that the divergence from long-run equilibrium is progressively addressed by a succession of partial short-run corrections. As shown in Table 6, the estimated value of ECT has a statistically significant and negative coefficient of -0.905. This discloses that the long-run equilibrium adjusts at a 90.5% yearly rate in reaction to the disequilibrium brought on by short-run shocks. On the other hand, the value of the R² is 0.93, which denotes that trade openness, FDI, capital, and labor force explain 93% of variations in economic growth of Somalia.

To scrutinize the consistency of the model assumptions, we performed several diagnostic tests, as illustrated in Table 7. The empirical outcomes demonstrate that the model passed the diagnostic test. The series is normally distributed, there does not appear serial correlation, heteroskedasticity, or misspecification of the functional form. Besides, the accuracy of the long- and short-run parameters as outlined by Pesaran (1998) was investigated through testing the stability of the model components. The study employed the cumulative sum (CUSUM) and cumulative sum of squares (CUSUMSQ) of recursive residuals, as depicted in Figure 2. The findings express that the plots of CUSUM and CUSUM of squares are between the critical boundaries at the 5% significance level for both tests of stability. This implies that the empirical findings of the model are accurate and reliable for policy-making decisions.

Figure 2. Model stability CUSUM test and CUSUM square test.

Table 7. Diagnostic tests.

Type	Statistic
Serial Correlation – LM test	1.384 [0.320]
Heteroscedasticity – BPG test	1.424 [0.329]
Normality – JB test	0.106 [0.947]
Ramsey RESET test	0.938 [0.379]
WLR – FDI	6.110 [0.0294]
WSR – FDI	1.214 [0.3349]
WLR – TO	17.491 [0.0013]
WSR – TO	2.588 [0.1260]

Note: values in the [.] are the p-values. W_LR and W_SR are the long-run and short-run asymmetries of FDI and TO, respectively.

To assess whether the changes in FDI and trade openness are definitely asymmetric, the short-run and long-run asymmetry are tested using the Wald test. The Wald (W_LR) results in Table 7 indicate that trade openness and FDI have asymmetric influence on economic performance in the long run. It implies that the output growth of Somalia is subject to positive and negative shocks in trade openness and FDI. This finding supports the conclusion of Amin et al. (2022), who observed a non-linear relationship in the FDI-growth nexus in Romania. In South Africa, Similar results were obtained by Udeagha and Ngepah (2021), who discovered an asymmetric connection between trade openness and GDP growth. However, the short-run Wald test (W_SR) results of trade openness and FDI became statistically insignificant at the 5% threshold level, which implies that asymmetric effects of trade openness and FDI is not present in the short-run.

The findings of the cumulative dynamic multiplier depicted in Figure 3 reveal the pattern of GDP growth due to alteration in trade openness and FDI in Somalia. The impacts of both positive shocks (increases) and negative shocks (decreases) in trade openness on economic growth have remained consistent over time, as seen in the left panel of Figure 3, even though both had shown adverse effects at the end of the period. Moreover, the results indicate that the decrease in trade openness has a more immense negative influence on economic growth than positive shocks in trade openness in Somalia, which supports our long-run findings. The fact that a negative shock has a higher impact than a positive shock suggests that policymakers foster the enhancement of trade openness by reducing trade barriers. As depicted on the right panel of Figure 3, while positive changes in FDI have a negative aftermath on economic performance in Somalia, adverse changes in FDI have a long-term positive impact. However, the cumulative negative effect of a positive change in FDI surpasses the cumulative positive impact of a negative change on GDP growth in Somalia. This signals that the government should promote a conducive environment for FDI since the current cumulative level of foreign investment hampers the economic growth of Somalia.

Figure 3. Dynamic multiplier effect graphs.

4.6. Multivariate and VECM results

Subsequently, we examine the multivariate cointegration method as a robust result of the ARDL and vector error correction modeling (VECM) for assessing the short-run and long-run causalities among the sampled variables. The results of the Johansen cointegration and VECM are presented in Tables 8 and 9, respectively. The outcome of multivariate cointegration indicates the presence of at least two cointegrating vectors among the interested variables, as shown by Trace and Maximum-Eigenvalues (see Table 8). Further, the results of VECM are summarized in Table 9. It is observed that there is a one-way causation linkage from labor force, capital, FDI, and trade openness to GDP in the short run, not in the long run. This suggests that labor, capital formation, FDI, and trade openness are essential for enhancing Somalia’s economic growth. A notable finding is a causation from FDI, GDP, and trade openness to capital in the short run. Moreover, unidirectional causation from GDP, trade openness, labor force, and capital to FDI is observed in the short run. It is observed that there is a bidirectional causality among GDP and capital in the short run. A striking outcome of the VECM is that there is no long-run causality among the interested variables (i.e., the ECT is statistically insignificant).

Table 8. Johansen cointegration test.

No. of CE(s)	Eigenvalue	Statistic	Critical Value	Prob.
Trace test
None *	0.8263	117.3888	69.8189	0.0000
At most 1 *	0.7371	68.3832	47.8561	0.0002
At most 2 *	0.4527	30.9748	29.7971	0.0364
At most 3	0.3312	14.0999	15.4947	0.0802
At most 4	0.0964	2.8379	3.8415	0.0921
Max-Eigen
None *	0.8263	49.0056	33.8769	0.0004
At most 1 *	0.7371	37.4084	27.5843	0.0020
At most 2	0.4527	16.8749	21.1316	0.1780
At most 3	0.3312	11.2620	14.2646	0.1416
At most 4	0.0964	2.8379	3.8415	0.0921

Note: * denotes rejection of the null hypothesis of no cointegration at the 5% significance level.

Table 9. VECM causality tests.

	Short-run causality					Long-run causality
	ΔlnGDP	ΔlnLF	ΔlnK	ΔlnFDI	ΔlnTO	ECTt-1
ΔlnGDP		3.455*	0.568*	0.009	−0.085*	0.286
ΔlnLF	0.014		−0.018	−0.0004	0.006	−0.010
ΔlnK	−1.166**	6.297***		0.003	−0.120**	0.075
ΔlnFDI	−14.068***	−89.830***	11.990***		−1.265***	12.014***
ΔlnTO	0.153	−10.688	−0.085	0.045		−0.235

Note: *, **, *** signify significance at 10%, 5%, and 1%, respectively.

5. Conclusion and policy recommendations

Many countries have experienced a rise in output growth as a result of the expansion of world economic integration and the lowering of trade barriers on a global scale through various trade agreements. The progress of improved trade and its associated competition have promised increased output that has attracted domestic and foreign investments. The existing empirical literature extensively investigated the short-run and long-run linkage between openness and FDI on the economic growth of various countries by using linear models. Hence, this study investigated the asymmetric impacts of trade openness and FDI inflows on the economic growth of Somalia using annual data from 1990 to 2020. To avoid spurious regression, the stationarity of the scrutinized variables was checked using the ADF and PP tests, which showed mixed integration orders, i.e., I(0) and I(1). The empirical investigations of the asymmetric bounds test demonstrated the presence of a non-linear cointegration association between trade openness, FDI, and economic growth. Subsequently, the estimation of the short- and long-run estimates was done using the NARDL technique. Moreover, the study conducted a VECM causality test in order to locate the direction of the causal connection among the variables. The cumulative dynamic multiplier was tested to determine the pattern of adjustment of economic growth to shocks in trade openness and FDI. It is notable that the models are free from diagnostic errors and stability issues.

The long-run findings exhibited that domestic investment improves economic growth in Somalia, although the labor force was observed to have negative effects. Furthermore, the estimates of the long-run asymmetric parameters reveal that positive and negative changes in FDI have a favorable increasing effect on economic growth. In addition, the study outlined that the decline in trade openness diminishes the growth level of Somalia. However, positive changes in trade openness do not affect the economic performance of Somalia in the long-run. Besides, the short-run estimates of the study indicate that domestic capital stock and labor force enhance the output level. It is noteworthy that a positive shock in FDI has a favorable increasing effect on output growth. Similarly, the negative shock in FDI of the previous period has a constructive contribution to the output level of Somalia. Furthermore, the study discloses that a rise in trade openness has a negative impact on the economic growth of Somalia in the short-run although the negative shock in trade openness was statistically insignificant. The estimated value of ECT shows that the long-run equilibrium adjusts to short-run shock-induced imbalances in the previous period at a rate of 90.5%. Besides, there is a one-way causality from labor force, domestic investment, FDI, and trade openness to GDP in the short-run but not in the long-run. This implies that labor, capital formation, FDI, and trade openness are essential for enhancing the economic growth of Somalia. Moreover, a unidirectional causation from GDP, trade openness, labor force, and capital to FDI is found in the short-run. A striking result of the VECM is that there is no long run causality among the interested variables (i.e., the ECT is statistically insignificant).

The empirical outcomes of the study contribute to the international economics literature by providing fresh knowledge on the asymmetric impacts of trade openness and FDI on economic growth in Somalia. Hence, the findings of the study can be derived from the following policy insights. Firstly, policymakers should implement a well-balanced growth stimulus initiative that is consistent with the enhancement of trade liberalization. Such a policy creates local competition and improves the country’s export competitiveness through an improved allocation of domestic resources. It is noteworthy that the economy is currently reliant on exporting agricultural goods, whose prices are unstable due to their vulnerability to international market fluctuations. In this regard, the new trade and investment schemes should switch the country from exporting primary products to more diversified goods, which could enhance the value of the exports. Secondly, the creation of a conducive investment environment for local and international investors. Currently, domestic investments are the key factor promoting the economic growth of Somalia, even though there has been a tremendous amount of capital flight for the past two decades. Hence, local investment should be encouraged, which is accompanied by the attraction of foreign capital in more capital-intensive sectors of the country. Additionally, FDI should be directed toward the export-oriented industrial and agriculture sectors, where Somalia has a comparative advantage in accordance with its trade policy.

Thirdly, to overcome the market access challenges of the exported goods, the Somali authorities should initiate bilateral and multilateral trade agreements with its partners in order to eliminate trade barriers, i.e., non-tariff barriers, to expand the volume of exports, and achieve economic growth. A crucial pathway to fulfilling this is to accelerate the country’s integration with international markets through the improvement of the World Trade Organization accession efforts that are still in the negotiation process. The organization’s membership will promise access to huge markets with lower trade restrictions that can boost the country’s trade volume. Fourthly, the government should improve human capital by expanding the expenditure on education because trade openness and FDI are associated with the acquisition of advanced capital goods. The skilled labor force can efficiently absorb the advanced technologies coming from developed countries, which can lead to long-run economic growth. Finally, the government should develop transportation infrastructure, the energy sector, the legal and institutional policies to achieve more sustainable trade competitiveness. The reduction of administrative and legal barriers will facilitate local firms access to finance and engagement in contracts with international consumers and suppliers.

Despite the present study demonstrating new evidence in the international economics literature, there are some limitations. Aggregated trade data were used in the study. Future studies should prioritize determining which industries have the most impact on the country’s output. It is crucial to understand whether the trade of agriculture or the non-agriculture sector has more influence on economic growth. Future research should also extend to examine the asymmetric effects of domestic investment and the labor force since the study made the assumption that changes in trade openness and FDI are non-linear in explaining economic growth.

Author contributions

The authors have contributed significantly to the writing of this article. Abdikafi Hassan Abdi and Mohd Azlan Shah Zaidi was responsible for the study’s conception, design, and development, writing the first draft of the article, data collection, analysis as well as reviewed and edited the article. Dhaqane Rooble Halane contributed to the literature review. Abdimalik Ali Warsame wrote the discussion section.

Data availability

The datasets used and/or analyzed during the current study are available from the following links: https://data.worldbank.org/indicator/SL.TLF.TOTL.IN?locations=SO; https://www.sesric.org/query.php; and https://unctadstat.unctad.org/wds/TableViewer/dimView.aspx.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research is supported by SIMAD University, Somalia (Grant number: SU-PG-2023-015).