Effect of COVID-19 on the mutual trade between Germany and the Visegrad Four

The study investigates how COVID-19 affected bilateral trade between Germany and the Visegrad Four (Czechia, Hungary, Poland, Slovakia). Germany is the V4’s closest and dominant trade partner and an industrial anchor in the EU, linked through extensive cross-border value chains. The research question is whether, and to what extent, the pandemic disrupted these trade flows. Contextually, the V4 economies are highly open, manufacturing-intensive, and integrally connected to German-led value chains, making them susceptible to pandemic-induced supply chain shocks. The purpose is to quantify COVID-19’s impact on exports to and imports from Germany and to discern short-run versus long-run dynamics. The importance lies in informing policymakers and industries in these closely integrated economies about resilience and recovery patterns following systemic shocks.

Recent research documents broad COVID-19 impacts on trade, supply chains, and policy responses. Studies highlight: global value chain disruptions and partial deglobalization (Antràs, 2020); initial supply-side shocks cascading to demand reductions (Zhang et al., 2022); heterogeneous trade resilience with lower-income countries less robust (Mena et al., 2022); negative effects across exporters and importers, with labor-intensive sectors more vulnerable (Hayakawa & Mukunoki, 2021); and generally justified but varied trade policy restrictions (Curran et al., 2021). For the V4, analyses report idiosyncratic regulatory effects (Urbanovics et al., 2021), health-system management weaknesses (Sagan et al., 2022), and uneven labor market impacts, particularly on women (Zieliński, 2022). Broader determinants of international trade—demographics, investment, technology diffusion, energy and transport costs, and institutions—also frame V4–Germany trade relations (Kocsis & Karácsonyi, 2022; Robertson, 1967; Rogers et al., 2008; Alt et al., 1996). Historically, the V4’s transition and EU accession increased openness, FDI, and trade intensity (Molendowski, 2015; Ivanová & Masárová, 2018; Melikhova et al., 2015). Prior V4-focused work found statistically significant but country-specific COVID-19 trade effects (Jindřichovská & Uğurlu, 2021; Ugurlu & Jindřichovská, 2022). This paper adds by isolating bilateral trade with Germany, the V4’s most significant partner, and explicitly quantifying COVID-19’s impact within that relationship.

- Data: Monthly series from 2010M1 to 2021M4. Trade (exports/imports with Germany) and GDP from Eurostat (EU trade since 1988 by SITC [DS-018995], values in euros; GDP in current prices, million euros). Because GDP is quarterly, it is converted to monthly via EViews 10 linear interpolation. Real broad effective exchange rates (index 2010=100) from FRED; Slovakia uses the euro but a common USD-based approach is used to keep currency treatment consistent. All trade and GDP variables are seasonally adjusted (SA) and then log-transformed; RB (exchange rate index) remains in levels (not logged). - Variables/Notation: L denotes logarithm; EX (exports), IM (imports); CZ (Czechia), HU (Hungary), PL (Poland), SK (Slovakia); RB real broad exchange rate; GDP gross domestic product; SA seasonally adjusted. Example: LEXCZ = log of seasonally adjusted exports of Czechia. - COVID-19 shock: An impulse dummy D equals 1 in 2020M03 (onset of the shock) and 0 otherwise (Ugurlu & Jindřichovská, 2022). - Econometric approach: Unit root tests (PP and ADF) indicate mixed orders of integration, I(0) and I(1), across variables, motivating ARDL bounds testing for cointegration (Pesaran et al., 2001). Country-specific ARDL models are estimated for exports and imports, including both host (V4) and partner (Germany) incomes, exchange rates, and the COVID dummy. Where cointegration is found, long-run relationships and short-run error-correction models (ECM) are estimated. Lag selection is AIC-based (maximum lag 12). Estimation and plotting use EViews 12. - Model specification (illustrative export for Czechia): A distributed lag of the dependent variable and regressors in differences, plus lagged levels for bounds testing; analogous specifications for Hungary, Poland, and Slovakia. Import models similarly include lagged dependent variables, host and partner GDPs, and exchange rates. - Diagnostics: Breusch–Godfrey LM tests (serial correlation), Ramsey RESET (functional form), and CUSUM/CUSUMSQ (stability). VIFs indicate multicollinearity primarily among GDP variables (VIF > 10); exchange rate VIFs < 5. The authors proceed despite multicollinearity, citing that OLS remains unbiased/consistent though precision may be affected (Gujarati & Porter, 2008). Notably, Slovakia’s export model fails stability diagnostics and is not interpreted further; import models largely pass with some instability in Hungary’s CUSUMSQ.

- Cointegration evidence: - Exports (bounds F-statistics): Czechia 4.7398 (significant), Hungary 26.9125 (significant), Poland 5.1189 (significant), Slovakia 5.1189 (significant). Cointegration holds in all; however, Slovakia’s export model fails stability diagnostics and is not used for interpretation. - Imports (bounds F-statistics): Czechia 4.6891***, Hungary 25.6632***, Poland 4.9442*, Slovakia 9.8885***; cointegration present in all. - Long-run export determinants (Table 3): - Germany’s GDP (LGDPDE) positively drives V4 exports except Slovakia: Czechia 1.3562**; Hungary 1.4860***; Poland 2.3803*; Slovakia −0.0826 (ns). Host-country GDPs are generally not significant for exports. - Real broad exchange rate: Hungary’s RBHU_SA is negative and weakly significant (−0.0041*); others not significant in long-run export equations. - Short-run export dynamics and COVID effect (Table 4): - Error-correction terms (ECT): significant and negative for Czechia (−0.7573***), Hungary (−1.0384***), Slovakia (−0.3882***); Poland’s ECT is positive and not significant (0.2467), implying no short-run adjustment for Poland’s exports. - COVID-19 dummy D is negative and significant across countries, indicating a contemporaneous trade shock: Czechia −1.4100**; Hungary −0.2349***; Poland −0.1379***; Slovakia −0.2639***. - Long-run import determinants (Table 6): - Partner GDP (Germany): positive and significant in Czechia (1.4614***) and Hungary (1.8571***); not significant in Poland and Slovakia. - Host GDP: significant and positive for Poland (1.0453**); significant and negative for Slovakia (−0.1802***); not significant in Czechia and Hungary. - Exchange rate: Slovakia’s RBSK_SA is large and positive (7.9615***); others not significant in the long run. - Short-run import dynamics and COVID effect (Table 7): - ECTs are negative and highly significant for all (Czechia −0.9669***; Hungary −1.0504***; Poland −0.5112***; Slovakia −1.0725***), indicating effective short-run adjustment toward long-run equilibria. - COVID-19 dummy D is negative and significant in all: Czechia −0.1039*; Hungary −0.1999**; Poland −0.2021***; Slovakia −0.3495***. - Overall interpretation: - COVID-19 exerted a significant, negative short-run shock to both exports and imports between Germany and all V4 countries. - Depending on dummy specification timing, the effect is concentrated during the pandemic onset with trade reverting toward prior trends, indicating transience of the initial shock and subsequent recovery. - Structural interdependence is strong (cointegration), with Germany’s income a key driver of V4 exports (except Slovakia). Exchange rate impacts are heterogeneous across countries and between exports and imports.

The findings directly answer the research question: COVID-19 significantly depressed bilateral trade flows between Germany and the V4, primarily as a sharp, short-run shock. The presence of cointegration in both export and import systems indicates deep structural linkages and value-chain integration with Germany. Germany’s income is the principal driver of V4 exports in the long run (except for Slovakia), consistent with gravity-type expectations that partner demand conditions shape export volumes. Short-run negative and significant COVID dummies show that initial containment and supply-chain disruptions curtailed trade, while significant and appropriately signed ECTs (except for Poland’s export model) show rapid adjustment back toward long-run equilibria. The heterogeneity in exchange rate effects and significance levels reflects differences in currency regimes, industrial structures, and trade composition across V4 countries. Collectively, the results imply that while the pandemic shock was sizable, bilateral trade relationships proved resilient and recovered toward trend, with industries and value chains adapting to operate more flexibly post-shock.

This paper contributes by isolating Germany–V4 bilateral trade and quantifying COVID-19’s impact using country-specific ARDL models on monthly data. Key contributions are: (1) establishing cointegration and strong long-run interdependence between Germany and each V4 economy; (2) showing that Germany’s GDP predominantly drives V4 exports in the long run; (3) documenting a negative and statistically significant short-run COVID-19 shock to both exports and imports across all V4 countries; and (4) highlighting heterogeneous exchange rate effects across countries. The evidence suggests the pandemic’s trade disruption was largely transitory, with trade reverting toward pre-shock paths, and that production networks adapted to increase flexibility. Future research should extend the sample to include more post-pandemic observations, incorporate sectoral/firm-level data to capture value-chain specifics, test alternative break specifications or regime-switching structures, and account for concurrent shocks (e.g., energy prices and inflation linked to the Ukraine conflict).

- Temporal scope: The analysis covers 2010M1–2021M4, providing limited post-pandemic observations; COVID-19 effects may evolve with a longer horizon. - Concurrent shocks: Rising inflation and energy price spikes associated with the Ukraine conflict may confound estimates of COVID-19’s impact in the late sample. - Model stability and specification: Slovakia’s export model failed stability diagnostics (CUSUM/CUSUMSQ), so its export results are not interpreted; some instability appears in Hungary’s import CUSUMSQ. - Multicollinearity: High VIFs for GDP variables indicate multicollinearity, potentially inflating standard errors, though the authors proceed given OLS properties. - Data construction: Quarterly GDP interpolated to monthly via linear methods may introduce measurement error relative to true monthly activity; exchange rate indices are broad-effective measures and not bilateral rates.