Revisited the role of foreign aid in capital formation: experience of South Asian countries

The paper investigates whether and how foreign aid influences domestic investment in South Asian countries and through which country-specific conditioning channels aid becomes effective. Despite sizable aid inflows and the centrality of investment as a transmission channel to growth, the aid–investment relationship has received limited attention compared to aid–growth studies. Prior findings are mixed, with evidence of both crowding-in and crowding-out effects and suggestions that domestic conditions such as trade openness, institutions, infrastructure, human capital, and financial development shape aid effectiveness. Focusing on South Asia—a major aid recipient region with relatively limited investment-focused evidence—the study evaluates six conditioning factors: financial development, trade, FDI, human development, government expenditure, and external debt, over 1990–2019. It aims to address methodological gaps by accounting for cross-sectional dependence and structural breaks and to inform policy by identifying channels that enhance or hinder aid’s investment impact.

Theoretical arguments suggest aid can complement domestic savings, relax foreign exchange constraints, finance critical capital goods, and enhance public investment, infrastructure, and human capital, thereby boosting private investment via positive externalities (Papanek 1973; UNCTAD 1999; Sachs 2005; Chatterjee et al. 2003; Calderon and Serven 2004). Aid can also mitigate debt overhang and adverse shocks, improving investment prospects (Guillaumont and Chauvet 2001; Collier and Dehn 2013; Herzer and Grimm 2012). However, other theories warn aid may encourage corruption, rent-seeking, consumption, Dutch disease effects, and volatility, potentially depressing investment and exports (Easterly 2003; Economides et al. 2008; Rajan and Subramanian 2011). Empirically, early and subsequent cross-country and country-specific studies report mixed evidence: negative or crowding-out effects on investment (e.g., Mosley et al. 1987; Snyder 1996; Herzer and Grimm 2012), weak or consumption-oriented impacts (Mahdavi 1990; Werker et al. 2009), as well as some complementary/crowding-in effects (Dollar and Easterly 1999; Orji et al. 2019). Several works emphasize that aid’s effectiveness depends on conditioning factors such as trade openness, governance, institutional quality, and infrastructure, but consensus is lacking and most evidence pertains to growth rather than investment. In South Asia, literature predominantly covers growth, poverty, FDI, human development, and trade competitiveness, with limited focus on investment. This motivates an examination of channel-specific effects for the region.

Data: Annual panel for six South Asian countries (India, Pakistan, Bangladesh, Sri Lanka, Nepal, Maldives), 1990–2019. Main variables include domestic investment (gross fixed capital formation, % GDP), foreign aid (ODA, % GDP), trade (% GDP), human development (secondary enrollment, gross %), external debt (% of GNP), FDI inflows (% GDP), government expenditure (% GDP), domestic credit to private sector (% GDP), lending interest rate (%), lagged growth rate, and real exchange rate. Sources: World Development Indicators (World Bank) and IMF Outlook. Empirical strategy: The study applies second-generation panel techniques to address endogeneity, cross-sectional dependence, heterogeneity, and structural breaks. - Panel stationarity: Pesaran’s CIPS (CADF-based) unit root test and Karavias–Tzavalis (KT, 2014) unit root test allowing structural breaks (intercept and trend). - Panel cointegration: Westerlund (2007) error-correction based tests accounting for cross-sectional dependence and heterogeneity; robustness via Westerlund and Edgerton (2008) cointegration tests with structural breaks (level and slope). - Long- and short-run estimation: Panel ARDL framework using Pooled Mean Group (PMG) estimators (Pesaran et al. 1999) for long-run homogeneity and short-run heterogeneity. Robustness via Common Correlated Effects PMG (CCEPMG) to control for unobserved common factors, cross-sectional dependence, and breaks (Chudik and Pesaran 2015). Interaction terms between aid and conditioning factors (financial development, trade, FDI, human development, government expenditure, external debt) identify channel (net) effects. - Panel causality: Juodis, Karavias, and Sarafidis (2021) Granger non-causality test employing Half Panel Jackknife (HPJ) correction to mitigate Nickell bias, with lag selection by SBC. Modeling follows investment theories (Keynesian, accelerator, flexible accelerator), incorporating interest rates, output growth, and other controls. Error-correction terms (ECT) assess long-run adjustment speeds. Hausman tests check long-run slope homogeneity; CD tests assess remaining cross-sectional dependence.

- Stationarity and cointegration: CIPS and KT tests indicate a mix of I(0)/I(1) variables with evidence of structural breaks. Westerlund (2007) and Westerlund–Edgerton (2008) tests support cointegration among variables even accounting for breaks. - Long-run direct effect of aid: Foreign aid significantly reduces domestic investment in South Asia. In PMG estimates, a 1% increase in aid (% GDP) is associated with about a 0.78% decrease in domestic investment (% GDP) in the baseline long-run specification; AID remains negative and significant across specifications. CCEPMG results corroborate a negative and significant long-run coefficient on AID. - Channel (net) effects of aid interacting with conditioning factors: • PMG net effects: financial development (+0.27), FDI (+0.19), trade (+1.10), human development (+0.43) are positive (complementarity); external debt (−0.26) and government expenditure (−0.35) are negative (substitutability). • CCEPMG net effects: financial development (+0.46), FDI (+0.46), trade (+0.63), human development (+0.09) positive; external debt (−0.19) and government expenditure (−0.18) negative. - Short-run dynamics: Short-run coefficients indicate DAID often positive and significant (e.g., around 0.12–0.18), and domestic credit also supports investment in the short run; lagged growth and lending rate changes can reduce investment in the short run. Trade and real exchange rate short-run effects are generally insignificant. - Other controls (long run): • Lending rate: unexpectedly positive and significant (≈0.14–0.22), interpreted as reflecting financial repression/administered rates where higher real lending rates expand credit supply and savings, supporting investment. • Trade: positive and significant (≈0.05–0.20). • Domestic credit to private sector: positive and significant (≈0.14–0.24). • Real exchange rate: negative and significant (≈−0.03 to −0.06), with depreciation making capital goods imports costlier. • Lagged growth: positive and significant, consistent with the (flexible) accelerator hypothesis. - Adjustment: ECT terms are negative and significant (about −0.11 to −0.17), indicating convergence to long-run equilibrium. - Causality: Juodis et al. (2021) Granger tests show bidirectional causality between foreign aid and domestic investment (e.g., W_HPJ ≈ 23.70** for AID→DINV and 13.52** for DINV→AID), supporting a feedback mechanism consistent with the crowding-out finding.

The findings address the research question by demonstrating that foreign aid does not directly stimulate domestic capital formation in South Asia; rather, it tends to crowd out investment in the long run. However, aid becomes effective when it augments complementary domestic channels—trade openness, financial development, FDI inflows, and human capital—through which investment is fostered. Negative net effects when aid interacts with external debt and government expenditure suggest that using aid to service debt or to finance unproductive public spending undermines investment formation. These results reconcile mixed prior evidence by highlighting that context and channels determine aid’s investment impact. The robustness across PMG and CCEPMG underscores the stability of the conclusions after controlling for cross-sectional dependence and structural breaks. The bidirectional causality implies that not only can aid influence investment (often negatively), but investment outcomes also shape aid dynamics, reinforcing the need for policies that strengthen complementary channels and improve aid management to mitigate crowding-out.

The study shows that in South Asia (1990–2019), foreign aid directly depresses domestic investment but can promote it indirectly when paired with conducive conditions: deeper financial systems, greater trade integration, stronger human capital, and higher FDI. Conversely, channeling aid toward external debt service or unproductive government spending reduces its effectiveness. Policy implications include: aligning aid with national investment priorities; strengthening governance and aid management institutions; enhancing complementary channels (trade facilitation, financial sector development, human capital formation) to unlock aid’s investment potential; coordinating aid and FDI policies; and reducing external debt burdens to avoid crowding-out effects. Future research could explore country-specific heterogeneity in channels, the role of institutional quality and governance in moderating channel effects, and disaggregated aid types to identify which modalities most effectively catalyze capital formation.