Asymmetric Nexus Among Domestic Investment and Economic Growth in Somalia: Evidence from the VEC Model and Granger Causality

Over the past century, countries have exhibited diverse patterns of economic growth. There is, however, a clear distinction between developed and developing countries, indicating that they have taken two divergent routes. Developed countries have consistently increased their per capita income and achieved transparent and sustainable development by effectively utilising investments. On the other hand, developing countries like Somalia, a country in the Horn of Africa, have faced numerous challenges to economic progress. The country has been unable to benefit from investments due to various factors such as poverty, corruption, and unstable politics. The inability to address these issues has resulted in a backward trajectory for Somalia. Over the years, Somalia has struggled with economic development due to a lack of resources and infrastructure. The country has yet to modernize its traditional rural pastoralism to adapt to urban trade and service-based industries. This has resulted in minimal progress in Somalia's economy in the last decade. Additionally, corruption has been a significant impediment to Somalia's development. Corruption has allowed the few who are in power to benefit from the country's resources while the majority of the population continues to live in poverty. This has limited the country's ability to invest in social services, education, and infrastructure, vital to economic growth. The political instability in Somalia has also been a significant challenge to the country's economic progress. The country has been in a state of civil war for over two decades, resulting in a lack of security, displacement of people, and loss of lives. The unstable political environment has made it difficult for investment and economic growth. Over the years, Somalia's inability to address poverty, corruption, and unstable politics has limited its economic development. The country needs significant infrastructure, social services, and education investments to support its transition from traditional rural pastoralism to urban trade and service-based industries. According to Somali National Bureau of Statistics (SNBS) [1], the GDP of Somalia increased by 2.9% in 2021, compared with a decrease of -0.3% in 2020. That decline comes from the COVID-19 pandemic shock in 2020.

The graph in Figure 1 illustrates the trends of domestic investment in Somalia. As indicated, the highest level of domestic investment occurred in 2008 at a value of $522 million. Subsequently, there was a decline in investment in Somalia due to factors like insecurity, political instability, and droughts. However, since 2012, domestic investment has been steadily increasing. Figure 2 displays Somalia's annual Gross Domestic Product (GDP) figures from 1992 to 2021. In general, Somalia's GDP has exhibited a gradual increase since 1996. However, the International Monetary Fund [2] states that the trade imbalance decreased GDP in 2013. Nevertheless, starting in 2014, the GDP resumed its regular growth pattern.

1.jpg

Figure 1. Domestic investment in Somalia

Data Source: OICStat

2.jpg

Figure 2. Gross Domestic Product of Somalia

Data Source: OICStat

According to the findings of Barro and Sala-i-Martin [3], economic growth can be enhanced by allocating significant resources to investment, as it has demonstrated a positive correlation with economic growth. The investment is a crucial instrument for attaining a higher standard of living and overall development. Although there are two types of investment—domestic and foreign—this study primarily focuses on domestic investment (DI) as an instrument variable to promote economic growth in low-income countries like Somalia. The fact that motivates this choice is that domestic investment facilitates the creation of job opportunities and promotes technological advancements, both of which are fundamental driving forces for economic development. It should be noted that much research has focused on the impact of Foreign Direct Investment on GDP, overshadowing the importance of domestic investment in driving economic growth. This is evidenced by abundant literature on the former topic [4-8].

The importance of domestic investment (DI) in driving economic growth has been debated among researchers for some time. Scholars such as Adams [9], Ali and Mna [10], and Keshava [11] have argued that DI impacts economic growth more than foreign direct investment (FDI). This argument is based on neo-classical and endogenous growth theories, suggesting DI is critical to promoting economic development. In light of these theories, this article aims to explore the correlation between domestic investment and economic growth in the context of Somalia.

3.1 Data

This study uses quantitative analysis of secondary data from 1992 to 2021 to determine how domestic investment affects the economic prosperity in Somalia; the data was collected from the specified period 1992 to 2021 to ensure data availability and to understand the economic impact after the military regime government collapsed. The dependent variable of this study is the Gross Domestic Product, with domestic investment, foreign direct investment, exchange rate, and exports as explanatory factors. This study utilises data from the Food and Agriculture Organization (FAO), the International Monetary Fund (IMF), and the Statistical, Economic and Social Research and Training Center for Islamic Countries (SESRIC). See Table 1 for details.

Table 1. Data description

3.2 Methodology

The Augmented Dickey-Fuller (ADF) unit root test and Johansen cointegration test are used in this study to determine the long-run association between domestic investment and economic growth. This study also analyses the Toda-Yamamoto (T-Y) approach of the Granger causality test to determine causality among the study's variables and the Vector Error Correction (VEC) model. Toda-Yamamoto is deemed superior to the traditional Granger causality test due to its capacity to work with any level of integration for the used series, whether it is I (0), I (1), or I (2) [34]. Granger Causality also helps this study identify the direction of causality between variables, which is essential for understanding whether domestic investment drives economic growth or vice versa [35-37]. On the other hand, The VEC model can handle data with different levels of stability, which is essential since economic data is often not perfectly stationary. It also provides insights into short-term and long-term relationships between variables [38, 39]. It's like an expert detective who examines the data and reconstructs the timeline of events (the relationships between variables). Together, these tools help to determine whether domestic investment drives economic growth in Somalia.

3.3 Unit root test

Assessing the stationarity of time series data is essential. Various unit root tests are documented in the literature, including the NG-perron, ADF, and Phillips and Perron tests. However, this study uses the ADF unit root test because it includes extra lags for the dependent variable to eliminate autocorrelation. It can assume three possible questions.

$\begin{matrix}

  \text{No Constant, No Trend} \\ 

  \Delta y_{t-1}=\gamma y_{t-1}+\sum_{i=1}^p \beta_i \Delta y_{t-1}+\varepsilon_t \\ 

  \text{Constant, No Trend} \\ 

  \Delta y_{t-1}=\alpha_0+\gamma y_{t-1}+\sum_{i=1}^p \beta_i \Delta y_{t-1}+\varepsilon_t \\ 

  \text{Constant, Trend} \\ 

  \Delta y_{t-1}=\alpha_0+\alpha_0 T+\gamma y_{t-1}+\sum_{i=1}^p \beta_i \Delta y_{t-1}+\varepsilon_t \\ 

\end{matrix}$    (1)

The $\tau$ test value is computed as:

$\tau=\frac{\hat{\gamma}}{\sigma_{\hat{\gamma}}}$    (2)

where, $\hat{\gamma}$ is the estimation coefficient, and $\sigma_{\hat{\gamma}}$ is the standard error of the estimation coefficient.

The ADF‘s Null hypothesis (H₀): data series is nonstationary or unit root.

3.4 Johansen cointegration test

The framework created by Johansen enables the incorporation of five types of trends and constants in both the short-run equations and cointegration relationships, which are restricted trend, unrestricted trend, restricted constant, unrestricted constant, and no trend or constant, the specificities of trends and constants may vary depending on the data. This study adapts unrestricted constant. Therefore, after specifying the cointegration type, the VECM equation looks as follows:

$\begin{aligned} & \Delta ln G D P_t =v+\alpha \beta^{\prime}\left(G D P_{t-1}+\varepsilon\right)+\theta_1 G D P_{t-1} +\theta_2 G D P_{t-2} \ldots .+\theta_{P-1} G D P_{t-(\rho-1)}+\mu_t\end{aligned}$    (3)

where, $\alpha \beta^{\prime}$ indicates the rank of Metrix that is explained in Eq. (3). $\alpha \beta^{\prime}=\left[\begin{array}{l}\alpha_1 \\ \alpha_2\end{array}\right]$ are short-run adjustment coefficients, and $\beta_s$ are cointegrating vectors of long-run relationship. At the same time, ε is constant in the cointegration equation.

3.5 Model specification

This study adopted the following model to determine the nexus between domestic investment and economic growth in Somalia:

$\boldsymbol{G D P}=(F D I, F D I, E R, E X P,)$    (4)

where, GDP is Gross Domestic Product, which measures the economic growth of Somalia, while DI is Domestic Investment, FDI is Foreign Direct Investment, ER is Exchange Rate, and EX is Exports.

The entire variable converted into natural Logarithmic to avoid heteroscedasticity and expression as an Econometric model:

$\begin{gathered}ln G D P_t=\beta_0+\beta_1 ln D I_t+\beta_2 ln F D I_t+\beta_3 ln E R_t +\beta_4 ln E X_t+\varepsilon_t\end{gathered}$    (5)

where, $\beta_s$ are coefficients of the study and expected to be> 0, and ε is the error term, Subscript t is time, which indicates that this study uses time series data.

3.6 Vector error correction model

This practices the VEC model to estimate both short-run and long-term relationships between the variables. It can also aid in determining how errors in equilibrium are fixed over time. Finally, yet importantly, adjustment coefficients show how well the changes in disequilibrium were made up. The general VECM questions can be written as:

$\Delta y_t=\alpha+\sum_{i=1,2 \ldots}^{\rho-1} \beta_i \Delta y_{t-1}+\lambda E C T_{t-1}+\varepsilon_t$    (6)

where, ∆ indicates the first difference is vector coefficient, $\beta_i$ is short-term dynamic coefficients of the model’s long-run adjustment equilibrium, λ is a speed of adjustment parameter, ECT_t-1 is error correction term legged from cointegrating regression of the dependent variable on regressors, and ε_t is a vector of impulse. The VECM model used in this study constitutes two parts; the first part is the lagged explanatory variables coefficients, which tell the causality of short-term relationships among the variables in the estimated model. The second part is ECT_t-1, which describes the long-term causality between the variables of the estimated model. When the VECM model in Eq. (6) applies to our study and expands as matrix form, it looks like the following equation:

$\begin{aligned} {\left[\begin{array}{c}\Delta ln G D P_t \\ \Delta ln D I_t \\ \Delta ln F D I_t \\ \Delta ln E X_t \\ \Delta ln E R_t\end{array}\right]=} & {\left[\begin{array}{l}\alpha_1 \\ \alpha_2 \\ \alpha_3 \\ \alpha_4 \\ \alpha_5\end{array}\right]+\left[\begin{array}{l}\lambda_1 \\ \lambda_2 \\ \lambda_3 \\ \lambda_4 \\ \lambda_5\end{array}\right] E C T_{t-1} } +\sum_{i=1}^{\rho-1}\left[\begin{array}{lllll}\beta_{1 i} & \beta_{1 i} & \beta_{1 i} & \beta_{1 i} & \beta_{1 i} \\ \beta_{2 i} & \beta_{2 i} & \beta_{2 i} & \beta_{2 i} & \beta_{2 i} \\ \beta_{3 i} & \beta_{3 i} & \beta_{3 i} & \beta_{3 i} & \beta_{3 i} \\ \beta_{4 i} & \beta_{4 i} & \beta_{4 i} & \beta_{4 i} & \beta_{4 i} \\ \beta_{5 i} & \beta_{5 i} & \beta_{5 i} & \beta_{5 i} & \beta_{5 i}\end{array}\right] \times\left[\begin{array}{c}\Delta ln G D P_{t-1} \\ \Delta ln D I_{t-1} \\ \Delta l n F D I_{t-1} \\ \Delta l b E X_{t-1} \\ \Delta l b E R_{t-1}\end{array}\right]+\left[\begin{array}{c}\varepsilon_t \\ \varepsilon_t \\ \varepsilon_t \\ \varepsilon_t \\ \varepsilon_t\end{array}\right]\end{aligned}$    (7)

where, ρ-1 is the optimal lag length, which is reduced by 1. i indicates the number of variables used in the article. According to Toda and Yamamoto [34], using the Granger causality test is the main requirement for short-term causality relationships among the endogenous variables in the VECM. Therefore, this study will use a pairwise Granger causality test.

The purpose of this research was to explore the correlation among the level of domestic investment and the economic growth of Somalia over almost three decades, ranging from 1992 up to 2021, by using annual time series data and both the Granger causality test and the Vector Error Correction Model (VECM). The study found that domestic investment, foreign direct investment, and exports positively contributed to economic growth, while the exchange rate hurt economic growth. The results showed bidirectional causal relations among economic growth and domestic investment in Somalia and a unidirectional causality association from exports and exchange rates to GDP in Somalia. The causal analysis also confirmed a unidirectional causal relation from economic growth to FDI in Somalia.

Recommendations from this study suggest that Somali government policymakers should prioritise creating a favorable environment for domestic investments by reducing bureaucratic obstacles, simplifying regulations, and offering incentives to attract local and foreign investors. In order to make the investment process more efficient, policymakers should focus on creating a business environment that is both transparent and effective. Offering tax incentives, such as reductions or breaks, can help ease the financial burden on businesses, making for a more competitive investment landscape. Improving the domestic investment climate can increase capital infusion and drive economic growth. Other suggestion for policymakers is to invest in research and development, enhance infrastructure for production and transportation, and facilitate access to credit for businesses that export goods. Boosting the export sector can make a significant contribution to long-term economic growth. Policymakers must prioritise human capital development to drive economic growth through foreign direct investment. Investing in education and skills training can attract foreign direct investment and establish Somalia as a desirable location for skilled workers. Fluctuations in exchange rates can harm a country's economic growth. Policymakers can manage this by increasing foreign exchange reserves, stabilising exchange rates, and promoting currency stability. A stable exchange rate regime can foster economic growth.