Do personal remittance outflows impede economic growth in Saudi Arabia? The role of trade, labor force, human, and physical capital

The study addresses whether personal remittance outflows (RMO) impede economic growth (EG) in Saudi Arabia. Prior research has largely focused on remittance inflows due to their sizable role in financing growth in recipient countries, while evidence on outflows is limited, primarily for GCC economies where outflows are relatively small shares of GDP. Saudi Arabia hosts millions of expatriates, generating substantial outbound remittances (e.g., about 4.9% of GDP in 2020). Descriptive trends show nominal RMO rising but RMO/GDP falling over time, while per capita income rises slowly, suggesting a complex, potentially nonlinear relation. The study formulates and tests: Null hypothesis: Remittance outflows impede economic growth in the Kingdom of Saudi Arabia. Alternative hypothesis: Remittance outflows don't impede economic growth in the Kingdom of Saudi Arabia. The aim is to estimate the effect of RMO on per capita GDP controlling for trade openness, physical capital, human capital, and labor force, using asymmetric (nonlinear) ARDL, cointegrating regressions, and VEC Granger causality.

Evidence on RMO’s effect on growth is sparse and mixed, mainly for GCC countries. Key studies: Alkhathlan (2013) used ARDL for Saudi Arabia (1970–2010), finding negative short-run but insignificant long-run effects of RMO on EG. Kaabi (2016) using panel least squares (2004–2014) found RMO negatively affected EG only in Saudi Arabia among GCC. Hathroubi and Aloui (2016) via wavelet analysis (1980–2013) found a positive short-term association between RMO and real GDP growth in Saudi Arabia. Rahmouni and Debbiche (2017) with ARDL (1970–2014) found insignificant short- and long-run effects of RMO on Saudi EG. Khan et al. (2019) using bootstrap panel causality (1996–2017) found heterogeneous bidirectional/unidirectional relations across GCC countries. Alsamara (2022) for Qatar using NARDL (2000Q1–2019Q4) documented negative effects of RMO on EG; simple ARDL found none. Other determinants: Trade often supports EG (e.g., Yaya 2017; Islam 2021a,b; Islam et al. 2022), though some find insignificance (Rahmouni and Debbiche 2017). Physical capital’s effect varies by context. Human capital (school enrollment ratios) shows mixed effects across settings. Labor force-growth links are also mixed; for Saudi Arabia, several studies report positive contributions of LF. Gaps: asymmetric (nonlinear) effects for Saudi Arabia and causality analyses are largely unexamined; robustness checks via CCR and FMOLS are uncommon. This study fills these gaps with NARDL asymmetry, CCR/FMOLS robustness, and VEC Granger causality.

Data: Annual time series for Saudi Arabia, 1985–2019, sourced from World Bank World Development Indicators (2022) and Migration and Remittances Data (2021). Variables: Economic growth measured by per capita GDP (GDPC, 2015 USD, log). Key regressor: remittance outflows (RMO, 2015 USD, log). Controls: total trade (TR, log), gross capital formation (GCF, proxy for physical capital, log), tertiary enrollment ratio (TEN, proxy for human capital, log), labor force (LF, log). Model specification: LnGDPC = f(LnRMO, LnTR, LnGCF, LnTEN, LnLF). To capture nonlinear/asymmetric effects of RMO, positive and negative partial sum decompositions of ΔLnRMO are constructed: LnRMO_P (positive shocks) and LnRMO_N (negative shocks), following Shin et al. (2013). Econometric approach: - Unit root tests (ADF, ERS, PP) confirm variables are I(0)/I(1) mixtures; dependent variable is I(1), supporting ARDL bounds framework. - Nonlinear ARDL (NARDL) model estimated to capture asymmetric long-run and short-run effects. Selected lag structure includes NARDL (1,0,2,0,1,1,0) for ECM representation. - Bounds test for cointegration uses Narayan (2005) finite-sample critical values due to small sample. - Long-run level relationship and short-run error correction model (ECM) reported. - Long-run asymmetry tested via joint equality restrictions on RMO shock coefficients. - Diagnostics: normality (Jarque-Bera), heteroskedasticity, serial correlation, Ramsey RESET, and stability via CUSUM and CUSUMSQ. - Robustness: Cointegrating regressions using Fully Modified OLS (FMOLS) and Canonical Cointegrating Regression (CCR). - Causality: Vector Error Correction (VEC) Granger causality with up to 2 lags to assess directions among variables. Descriptive statistics (Table 1) provided for all variables in logs. Key model selections/results: - Bounds F-statistic = 10.5402, exceeding Narayan (2005) critical values, confirming cointegration. - ECM shows significant and negative error-correction term, indicating rapid adjustment to long-run equilibrium.

- Cointegration: Bounds F-statistic = 10.5402 (> finite-sample 1% upper bound), confirming a long-run relationship among LnGDPC, RMO shocks, and controls. - Long-run level effects (NARDL): • LnRMO_P (positive shocks): coefficient 0.0558, p=0.3254 (insignificant) → increases in RMO do not affect EG. • LnRMO_N (negative shocks): coefficient 0.3637, p<0.001 → declines in RMO significantly increase per capita GDP. • LnTR: coefficient 0.2293, p<0.001 → trade openness strongly and positively affects EG. • LnGCF: coefficient -0.0160, p=0.8300 (insignificant). • LnTEN: coefficient 0.0164, p=0.8084 (insignificant). • LnLF: coefficient 0.3592, p=0.0059 → labor force significantly and positively affects EG. - Long-run with lags (Table 5): • LnRMO_N: contemporaneous 0.1536 (p=0.0627), LnRMO_N(-2) 0.2789 (p=0.0024), LnRMO_N(-1) -0.1379 (p=0.2407). • LnTR: 0.1857 (p=0.0002). • LnGCF(-1): -0.0953 (p=0.0338); LnGCF contemporaneous insignificant. • LnLF: 0.2909 (p=0.0012). • Long-run asymmetry check: F=2.4140, p=0.0967 (10% significance), supporting asymmetric effects of RMO. - Short-run ECM (Table 7): • D(LnRMO_N): 0.1536 (p=0.0210); D(LnRMO_N(-1)): -0.2790 (p=0.0002) → mixed short-run effects with likely negative net impact. • D(LnGCF): 0.0824 (p=0.0129) → physical capital boosts short-run growth. • D(LnTEN): 0.1348 (p=0.0457) → human capital boosts short-run growth. • ECT(-1): -0.80995 (p<0.001) → about 80.99% annual adjustment to long-run equilibrium. Model fit: R²≈0.823 (short run); R²≈0.934 (long run). - Diagnostics (Table 6): Residuals normal (JB p=0.3333), no serial correlation (p=0.3656), homoscedasticity (p=0.1300), no misspecification (RESET p=0.4269), CUSUM/CUSUMSQ stable. - Robustness (Table 8): CCR and FMOLS confirm NARDL long-run results: LnRMO_N significant positive (0.2632–0.2788); LnRMO_P insignificant; LnTR positive significant (0.1500–0.1739); LnLF positive significant (~0.226); LnGCF and LnTEN insignificant. - Causality (Table 9): • LnRMO_P does not cause EG; EG causes LnRMO_N (p≈0.076). • Trade → EG (unidirectional). • Labor force ↔ EG (bidirectional). • EG → TEN; TR → TEN. • LnRMO_N ↔ LF; TR ↔ LF. • LnRMO_P, GCF, TEN → LF (unidirectional). Overall: In the long run, lower remittance outflows (negative shocks) are associated with higher per capita GDP, while increases in outflows do not spur growth. Trade and labor force consistently support growth; physical and human capital are insignificant in the long run but matter in the short run.

The findings address the hypothesis by showing that remittance outflows exhibit asymmetric effects on economic growth. Long-run results indicate that decreases in outflows (negative shocks) significantly raise per capita GDP, while increases in outflows do not significantly affect growth, suggesting that maintaining lower levels of outflows is beneficial for Saudi growth. Short-run dynamics show mixed effects of negative RMO shocks on growth, with a likely negative net impact in the immediate term, but rapid adjustment to the long-run equilibrium. These outcomes underscore the importance of the remittance channel for a labor-importing economy: expatriates contribute to output domestically, yet their savings remitted abroad can dampen domestic demand and investment. The positive role of trade and labor force reinforces Saudi Arabia’s comparative advantage in global trade and the contribution of labor availability to output. The insignificance of physical and human capital in the long run suggests potential inefficiencies in capital utilization and gaps in the quality or relevance of human capital to the labor market, possibly due to skill mismatches and reliance on expatriate labor. Policy relevance: channel expatriate savings into domestic consumption and investment through expatriate-friendly policies and incentives (including residency/citizenship options), and upgrade education and vocational training to enhance local labor productivity, which may reduce reliance on expatriates and hence outbound remittances. Causality evidence supports these channels, indicating that trade and labor dynamics are key drivers of growth and that several factors influence growth indirectly via the labor force.

Using annual data for 1985–2019 and a nonlinear ARDL framework with robustness and causality checks, the study finds a cointegrating relationship among per capita GDP, remittance outflows, and key controls in Saudi Arabia. Long-run results show that negative shocks (declines) in remittance outflows significantly increase per capita GDP, while positive shocks have no significant effect. Trade openness and the labor force significantly and positively influence growth; physical and human capital are not significant in the long run, though they contribute positively in the short run. Policy recommendations include incentivizing expatriates to channel savings domestically (via pragmatic labor laws, incentives, and residency/citizenship pathways) and upgrading the quality of education and training to raise local labor skills and efficiency, thereby reducing reliance on expatriates and outbound remittances. Future research can exploit more granular labor data and replicate the analysis in other GCC economies.

- Labor force data include both expatriates and Saudi nationals; separate data for Saudi labor force were unavailable, limiting interpretation of labor force effects by nationality. - Findings pertain to Saudi Arabia and may not generalize beyond GCC contexts. - Potential data limitations due to availability constraints (1985–2019) and annual frequency. Future research: Use disaggregated labor data (Saudi vs expatriate) if available; replicate analyses in other GCC countries; explore sectoral and micro-level channels of remittance outflows’ effects on growth.