Determinants of International Trade: Nepalese Perspectives

The paper examines how key macroeconomic variables shape Nepal’s international trade, emphasizing the importance of trade for growth, competitiveness, and access to goods and technologies. It highlights Nepal’s heavy reliance on imports—largely influenced by remittances that raise purchasing power—alongside the roles of GDP, international prices, and exchange rates. Given Nepal’s strong trade links with India, international price movements in India and exchange rate dynamics are especially consequential. The study notes a large and persistent trade imbalance with imports far exceeding exports and identifies a gap in the literature regarding a comprehensive empirical assessment of the specific determinants of Nepal’s imports and exports. The research aims to quantify how GDP, remittances, international prices, exchange rates, and capital stock affect Nepal’s trade, to inform policies that improve trade performance and economic resilience.

Prior studies using gravity and panel approaches indicate that trade is influenced by economic size and distance. Acharya (2012/2013) finds Nepal’s exports and imports rise with partner GDPs, with distance dampening trade; Nepal’s trade deficit persists despite higher export growth rates. Bista and Adhikari (2021) report that Nepal’s and partners’ GDP, real effective exchange rate, and SAFTA membership support exports, while distance reduces trade; partners’ GDP and per-capita GDP differentials raise imports. Sharma (2020) documents Nepal’s growing trade deficit and concentration of trade with India, recommending diversification. Contextually for Nepal, slower GDP growth, rising international prices, and exchange rate volatility constrain competitiveness, raise import costs, and introduce uncertainty for traders. Nepal’s fixed NPR-INR peg and central bank interventions under a fixed regime are noted. Collectively, the literature points to GDP, prices, exchange rates, distance, regional integration, and economic freedom as key determinants, and underscores Nepal’s dependence on India and need for broader market engagement.

The study uses secondary annual time series data from July 1975 to July 2023 sourced from Nepal Rastra Bank, Ministry of Finance, Central Bureau of Statistics, World Bank, IMF, and Reserve Bank of India. Two econometric models are estimated using OLS with logs to stabilize variance and interpret elasticities; stationarity concerns are addressed via log-differences in export models where needed. Diagnostics include checks for multicollinearity (notably between Nepal and India CPI), autocorrelation (Durbin-Watson), and model fit. Import Model (Model A): Imports are regressed on real GDP, remittances, CPI India, and CPI Nepal. Severe multicollinearity between CPI India and CPI Nepal (correlation ≈ 0.998) leads to dropping CPI Nepal. A COVID-19 dummy is tested but found insignificant and removed. Export Model (Model B): Exports are regressed on the international price ratio (CPI India/CPI Nepal), exchange rate (NPR/USD), and capital stock (real gross fixed capital formation). Subsequent specifications use differenced logs (DLOG) to address potential spurious regressions and achieve stationarity. An alternative export specification replaces the price ratio with separate CPI terms and later introduces real GDP in differenced log form. Coefficients are interpreted as elasticities due to log transformations.

Model A (Imports): Initial specification with LOG(RGDP), LOG(REMITTANCES), LOG(CPII), LOG(CPIN) shows LOG(RGDP) = 1.103 (p=0.009), LOG(REMITTANCES) = -0.071 (p=0.071), LOG(CPII) not significant, LOG(CPIN) = 1.476 (p=0.0067); R^2 ≈ 0.998. Due to multicollinearity between CPII and CPIN (corr ≈ 0.998), CPIN is dropped. With DUM_COVID included: LOG(RGDP) = 1.644 (p=0.0003), LOG(REMITTANCES) = -0.133 (p=0.0005), LOG(CPII) = 1.451 (p=0.0000), DUM_COVID not significant; R^2 ≈ 0.998; DW ≈ 1.02. Final import model (dropping DUM_COVID): LOG(RGDP) = 1.595 (p=0.0002), LOG(REMITTANCES) = -0.131 (p=0.0005), LOG(CPII) = 1.471 (p=0.0000); R^2 ≈ 0.9978; DW ≈ 1.05. Interpretation: 1% increases in real GDP and India’s CPI are associated with approximately 1.59% and 1.47% higher imports, respectively; a 1% increase in remittances associates with a 0.13% decrease in imports. Model B (Exports): Levels model indicates only exchange rate is significant: LOG(EXRATE) = 2.160 (p=0.0000); LOG(CPII/CPIN) and LOG(CAPITALSTOCK) are not significant; R^2 ≈ 0.960; DW ≈ 1.05. DLOG specification with price ratio: DLOG(EXRATE) = 1.813 (p=0.046); DLOG(CPII/CPIN) and DLOG(CAPITALSTOCK) not significant; R^2 ≈ 0.151; DW ≈ 1.36. Alternative DLOG model: DLOG(CPII) = -5.197 (p=0.0047), DLOG(EXRATE) = 3.001 (p=0.0002), DLOG(CPIN) and DLOG(CAPITALSTOCK) not significant; R^2 ≈ 0.467; DW ≈ 1.62. Final DLOG export model (with DLOG(RGDP)): DLOG(CPII) = -6.496 (p=0.0000), DLOG(EXRATE) = 2.490 (p=0.0005), DLOG(RGDP) = 3.128 (p=0.096); R^2 ≈ 0.471; DW ≈ 1.37. Interpretation: Higher Indian CPI significantly reduces Nepal’s exports (elasticity around -6.5), while a depreciation of NPR (higher exchange rate) raises exports (elasticity around 2.49). GDP has a positive but marginally significant association with exports.

The empirical results address the research question by quantifying how macroeconomic determinants shape Nepal’s trade. Imports rise strongly with domestic economic activity and Indian price levels, but fall with higher remittance inflows. The positive import response to India’s CPI, contrary to the initial hypothesis, may reflect inelastic demand for essential Indian goods and rising purchasing power allowing import volumes to be sustained despite higher prices. The negative linkage between remittances and imports, also contrary to the hypothesis, may indicate uses such as debt repayment or investment rather than consumption of tradables, and that imports respond to broader consumption patterns beyond remittance-receiving households. For exports, exchange rate depreciation robustly boosts competitiveness and volumes, while higher Indian prices significantly suppress Nepal’s exports—likely via input cost pass-throughs, competitiveness pressures in the India-centric market, or demand effects. GDP’s positive association with exports suggests capacity-driven gains from growth, though its statistical strength is weaker in short-run differenced models. Overall, the findings underscore Nepal’s exposure to Indian price dynamics and the exchange rate, highlighting the need for competitiveness, market diversification, and macroeconomic stability to improve trade outcomes.

The study concludes that Nepal’s import dynamics are primarily driven by real GDP growth and Indian price levels, with remittances associated with lower import volumes, while export performance is highly sensitive to exchange rate depreciation (which boosts exports) and Indian CPI increases (which dampen exports). After addressing multicollinearity between Nepal and India CPIs, the import model shows robust elasticities for GDP and CPI India and a significant negative elasticity for remittances. Export models consistently reveal the dominant role of the exchange rate and a strong negative effect of India’s CPI; GDP contributes positively but less robustly in short-run specifications. Policy implications include sustaining economic growth, enhancing competitiveness and quality of Nepalese products, diversifying export markets to reduce overreliance on India, maintaining a favorable and stable exchange rate through prudent macro policies, and channeling remittances into productive investments to expand domestic supply and reduce import dependence. These steps can strengthen Nepal’s resilience and help rebalance its trade structure.

The study relies on secondary annual data (1975–2023), which may mask intra-year dynamics and be subject to measurement revisions. Severe multicollinearity between Nepal and India CPIs required dropping CPIN from the import model, potentially omitting relevant domestic price information. Several specifications exhibit signs of positive autocorrelation (Durbin-Watson near or below 1.4), which may affect inference if not fully addressed. The export models using differenced logs explain a moderate share of variation (R^2 around 0.47) and suggest possible omitted variables. Model results capture correlations within the specified frameworks and may not fully identify causal mechanisms.