Political and productive capacity characteristics as outward foreign direct investment push factors from BRICS countries

Pull factors are host-country attributes motivating inward FDI, while this study focuses on push factors—home-country characteristics that drive outward FDI (OFDI). OFDI can benefit home economies via access to markets, inputs and technology, enhancing productivity and competitiveness. BRICS countries have accounted for a large share of the surge in global OFDI by firms from developing countries (62% of the increase from 1995 to 2015). Prior work recognizes home-country measures and institutions as important, but gaps remain regarding the roles of political stability and productive capacity as home-country push factors in emerging markets. This study addresses that gap by empirically introducing two innovative variable sets: a productive capacity index (PCI) and a newly derived political stability index (PSI) built from ICRG components. The research examines BRICS home-country push factors affecting OFDI and their dynamic short- and long-term effects using a panel autoregressive distributed lag (PARDL) model and impulse response functions (IRF) for 2000–2018. The contributions include employing a PARDL framework specific to BRICS and constructing a comprehensive PSI. Results confirm significant short- and long-run relationships, particularly strong long-run effects of PCI and PSI on OFDI.

The study is grounded in Dunning’s eclectic (OLI) paradigm and subsequent extensions emphasizing institutions (Dunning and Lundan, 2008, 2010) and the investment development path (Dunning and Narula, 1996), which identify market size, resources, and policies as key determinants varying by development stage (BRICS largely in stages 3–4). Prior evidence suggests OFDI can enhance home-country employment and productivity (Lipsey, 1995; Desai et al., 2005; Navaretti and Castellani, 2004). Country-specific studies (e.g., Russia: Kalotay & Sulstarova, 2010; China: Hong & Kim, 2003; Morck et al., 2008; Yu et al., 2019; India: Nayyar, 2008) emphasize liberalization, institutional reforms, access to finance, and exchange rate expectations. Roy and Narayanan (2020) use quantile panels distinguishing developing vs. developed markets and find income, technology, and integration positively associated with OFDI and interest rates negatively so, but do not isolate BRICS or include political stability and PCI as in this study. Political risk has been shown to affect FDI (Lucas, 1990; Jeutang & Kesse, 2021), and components of productive capacity (human capital, natural/energy resources, technology) are linked to OFDI (Lall, 1980; Clegg, 1987; Al-Sadig, 2013). However, direct treatment of PCI and a comprehensive PSI as home-country OFDI push factors is limited. The review supports classifying push factors into: (1) economic openness and stability (trade openness, inward FDI, GDP, exchange rate, interest rate), (2) productive capacity and technology, and (3) political stability.

Sample and data: Balanced panel of BRICS (Brazil, Russia, India, China, South Africa) for 2000–2018, constrained by PCI availability. Data sources: UNCTAD (OFDI stock, trade openness, PCI), World Bank WDI (GDP, real exchange rate, real interest rate, patents/technology), and ICRG (PRS Group) for political risk components. Variables (logs used): LOFDIS (OFDI stock, dependent), LIFDI (inward FDI), LEIGDP (trade openness), LGDP (income level), LEXCH (real exchange rate), LRIR (real interest rate), LPCI (productive capacity index), LTechno (patents), Lindex (constructed political stability index). Model specification: OFDIS = f(IFDI, EIGDP, RIR, EXCH, PGDP/GDP, Techno, PCI, Index). A PARDL approach estimates short- and long-run effects via MG, PMG, and DFE estimators; a VAR-based IRF assesses dynamic responses to shocks. PSI construction: Following Jeutang & Kesse (2021), principal component analysis (PCA) on 12 ICRG political risk subcomponents; although only the first four PCs had eigenvalues ≥1 (cumulative 0.752), all PCs were retained to capture all aspects. Weights for each PC are proportional to eigenvalues, and the PSI is the weighted sum of PCs. Descriptive PSI stats: mean 2.1803, min 1.99, max 2.55, SD 0.1313. Diagnostics and tests: Cross-sectional dependence (Breusch–Pagan LM, Pesaran scaled LM, Pesaran CD) indicate no/weak CSD (all p-values > 0.05). Unit root tests (IPS, ADF-Fisher, PP-Fisher, LLC) with intercept and trend show some nonstationarity in levels but stationarity at first differences. Variance inflation factor mean 5.8 (<10), suggesting no critical multicollinearity. Panel cointegration (Pedroni, Kao, Westerlund) rejects no-cointegration (all statistics significant at 1%), supporting a long-run relationship. Optimal lag determined as 1 using MMSC, Hansen J, MBIC/MAIC/MQIC. Estimation: MG, PMG, and DFE PARDL models estimated; Hausman test favored MG as the preferred estimator (reported p-value 0.7755). Error-correction terms are negative and significant across estimators, confirming convergence to long-run equilibrium. IRF: VAR-based impulse responses trace LOFDIS reactions to one-standard-deviation shocks in Lindex, LPCI, LTechno, LGDP, and own shocks over a 10-period horizon. Robustness: Replaced Lindex with World Bank Governance Political Stability (GPS) indicator; re-ran all steps (CSD, unit roots, cointegration, lag selection). Signs and significances were consistent; Lindex exhibited a relatively larger long-run coefficient than GPS.

- Economic openness and stability: Positive and significant associations of trade openness (LEIGDP) and inward FDI (LIFDI) with OFDI in both short and long run across estimators. Example long-run MG coefficients: LIFDI ≈ 0.955 (significant), LEIGDP ≈ 0.569 (significant). GDP (LGDP) and real exchange rate (LEXCH) also positive and significant long run (MG LGDP ≈ 0.786; LEXCH ≈ 0.834). Real interest rate (LRIR) negative and significant (MG ≈ −0.207 long run; short run ≈ −0.584), indicating higher borrowing costs dampen OFDI. - Productive capacity and technology: PCI (LPCI) and technology (LTechno) are positive and significant drivers in both horizons. Long-run MG LPCI ≈ 0.953; LTechno ≈ 0.397; short-run MG LPCI ≈ 0.145; LTechno ≈ 0.307. - Political stability: The constructed PSI (Lindex) is positive and significant in short and long run; long-run effects are comparatively stronger than short-run effects. Long-run MG Lindex ≈ 0.610; short-run MG ≈ 0.236. - Adjustment dynamics: Error-correction terms are negative and significant (e.g., MG EC(−1) ≈ −1.825, p<0.01), confirming long-run equilibrium adjustment. Panel R-squared around 0.60 (MG ≈ 0.604). - IRF evidence: Shocks to Lindex produce a positive, significant response in LOFDIS from about period 4 onward; shocks to LPCI diverge positively from around period 2; LTechno and LGDP also show positive long-term responses. A shock to LOFDIS causes a self-correcting downward drift from period 2. - Robustness: Substituting GPS for Lindex preserves signs and significance; Lindex’s long-run coefficient magnitude exceeds GPS, supporting the comprehensiveness of the PCA-based PSI. - Overall: Both PCI and PSI exhibit relatively strong long-run impacts on OFDI compared with other controls, validating them as key push factors in BRICS.

The findings directly address the research question by demonstrating that home-country political stability and productive capacity are significant and durable push factors for OFDI from BRICS, beyond standard macroeconomic openness and stability determinants. The positive long-run coefficients for PCI and PSI, corroborated by impulse responses, indicate that improvements in governance stability and the breadth of productive capabilities (human capital, infrastructure, institutions, private sector dynamism, ICT, etc.) create conditions that encourage firms to expand abroad. The negative effect of real interest rates reinforces the role of financing conditions. The results align with theory (OLI, investment development path) and prior empirical evidence on the roles of institutions and capabilities, but extend the literature by quantifying the joint and long-run effects in a BRICS panel with PARDL and a composite PSI. Policy-wise, the results underscore the importance of strengthening domestic productive systems and institutional stability to sustain and scale OFDI, which can, in turn, feed back into home-country development via productivity, learning, and market diversification.

Using PARDL estimations and VAR-based IRFs on BRICS (2000–2018), the study shows that OFDI is jointly and significantly driven in the short and long run by macroeconomic openness and stability (trade openness, inward FDI, GDP, exchange rate, interest rates), productive capacity, technology, and political stability. The strongest long-run push factors are the productive capacity index and the political stability index, whose shocks produce persistent positive effects on OFDI. Policy implications emphasize enhancing human capital, energy and natural resource security, structural transformation, transport and ICT, institutional quality, and private-sector development, alongside improving governance and political stability, to foster outward investment. Future research should unpack the individual components of political stability and productive capacity with methods that can handle multicollinearity and explore heterogeneous sectoral or country-specific effects within BRICS.

- Temporal and coverage constraints: Analysis restricted to 2000–2018 due to PCI availability, limiting longer-horizon inference. - Composite indices and multicollinearity: High multicollinearity among political risk components necessitated a PCA-based PSI, precluding analysis of individual component effects; similar aggregation for PCI may mask heterogeneous sub-dimension impacts. - Estimator assumptions and endogeneity: Dynamic panel estimators (MG/PMG/DFE) carry assumptions about homogeneity/heterogeneity; potential endogeneity between lagged dependent variables and errors is acknowledged (especially for DFE), though cointegration and error-correction mitigate concerns. - Generalizability: Results pertain to BRICS and may not generalize to other developing or developed country groups without further study. - Data access: ICRG data are proprietary; replication requires licensed access, though UNCTAD and WDI data are publicly available.