Shock, international trade, and diversity

Ting Lan, Davide Malacrino, Adil Mohommad, Andrea Presbitero, Galen Sher 11 May 2022

The COVID-19 epidemic led to a decline in international trade (Baldwin 2020), as security concerns prevented individuals from moving out and producing imported goods and workers for export. In the second quarter of 2020, global trade in goods decreased by 12.2% and trade in services declined sharply by 21.4% compared to the last quarter of 2019. Since then, trade in goods has recovered rapidly, returning to pre-epidemic levels by October 2021, but trade in services has remained below pre-epidemic levels. This distinction between goods and services is well documented (e.g. IMF 2021), but it is less well known that products and services have different patterns. Among products, it trades products that are typically produced in GVC, such as automobiles, electronics and clothing, which are rapidly declining and recovering. Among services, trade in travel services remains depressed while trade in telecommunications services is already strong before the epidemic.

The epidemic has broken old patterns

Prior to the epidemic, trade in goods and services closely tracked the development of overall demand. The import demand model, which relates a country’s imports to its domestic demand and relative import prices, was good for interpreting imports of goods and services before the epidemic. However, these patterns are broken during the epidemic. In 2020, commodity trade fell by less than one per cent due to changes in overall demand and relative import prices. This unexpected deviation was even greater in countries with more infections, stricter control systems, and less mobility (Figure 1).

Figure 1 Factors related to the forecast error of the demand model in 2020

Formula: World Health Security Index; Google, Community Mobility Report; Hale and others (2021); Our world in data; World Trade Organization; And the IMF staff count.
Note:: The figure reports the standard coefficient of residual regression from the demand model of the listed variable. Solid bars show coefficients that are statistically significant at the 5% level; Shows hollow bars that aren’t. The health readiness of trading partners for epidemics is measured by the Global Health Security Index. “Travel Services Imports Share” captures the share of travel services in a country’s total service imports. See IMF (2022) for more details.

In contrast, the service trade declined by less than one per cent in 2020 due to a change in overall demand, and this gap is largely explained by the disruption in imports of travel services caused by the epidemic. These patterns may reflect changes in costs from services to products under epidemic waves, or they may reflect difficulties in producing goods domestically under epidemic waves, in which case they are imported instead.

Great lockdown and unintentional international spillover

Another important reason for the decline in trade in 2020 is the international spillover from the containment policy established by trading partners (Bonadio et al. 2020, Espitia et al. 2021, Lafrogne-Joussier et al. 2022). Indeed, we find that between January and May 2020 there was a 60% decline in commodity imports, the unintended consequences of the lockdown imposed by countries’ trading partners (Figure 2). This research is based on a standard gravity model (Santos Silva and Tenreyro 2006) using detailed commodity-level data (‘HS-six-number’) of bilateral commodity trade that is responsible for the development of importing countries and industries, changing demand and Including factors such as trade agreements that may affect commodity-level trade between the two countries.

Figure 2 International spillover effect of lockdown

Formula: Hale et al (2021); IMF, Trade Statistics Directions; And the IMF staff count.

The country that dropped out of the lockdown depends on their imposition. For countries whose exporting partners are more able to rely on remote work, the spillover was less than half. Spillovers from lockdowns also change over time. They became stronger between February and April as more countries outside Asia began to impose lockdowns and then began to decline from May, suggesting that trade had adapted to them.

The elasticity of the global value chain

Like commodity trade, GVCs have adapted to the wave of transmission during epidemics. After capturing the initial wave of transition, Asia has increased its GVC-intensive product market share. Between 2019 and June 2020, Asia grew 4.6 percentage points in Europe and 2.3 percentage points in North America (Figure 3). Profits appear to be temporary, as they were returned to 3.1 percentage points and 0.6 percentage points a year later.

Figure 3 Transfer of market share

Formula: Trade Data Monitor; And the IMF staff count.
Note:: The market share of Factory Asia and Factory Europe is calculated using only GVC-intensive products as defined in IMF (2022). GVC = Global Value Chain. See IMF (2022) for more details.

Nevertheless, the epidemic greatly disrupted supply (Celasun et al. 2022) shifting demand from service to product while disrupting the production and delivery of workers, resulting in high shipping costs and congested ports. The war in Ukraine has exacerbated these disruptions by depleting energy, metals and agricultural products. Epidemics, war, climate change, and cyber attacks make GVC more resilient.

Two strategies for resilience: diversity and replaceability

Companies can diversify their supplier relationships to source more material from different countries, which will reduce dependence on any single country (including the home country) and provide established relationships that can be tapped in times of crisis (Bas and Fernandes 2022). Indeed, our simulations using a global economic model (Bonadio et al. 2021) suggest that more internationally versatile sourcing will reduce the economic impact of supply disruptions in a large supplier country by about half on average (Figure 4). When multiple countries are repeatedly hit by supply shocks, high diversity reduces the volatility of economic growth by about 5%.

Figure 4 Benefit from diversity

Source: IMF Staff Accounts.
Note:: Figure shows GDP decline in response to a 25% labor supply contraction in a country that is a major global supplier of intermediaries. Bars and squares show a simple average of the GDP decline in the countries of each region Replacement elasticity = 0.5. See IMF (2022) for more details.

The greatest room for diversity is the reduction of internally acquired shares. We find that global firms tend to source most of their inputs in their own country (e.g. 82% in the Western Hemisphere). This will only make matters worse by leaving resourcing firms more open to disruption in their country.

In addition to establishing relationships with more suppliers, companies can adapt their manufacturing processes to make it easier to replace inputs provided by different suppliers. For example, in response to the shortage of semiconductors, Tesla rewrote its software to enable it to use semiconductors that were less scarce. Our simulations with the global economic model suggest that greater substitution could reduce the economic impact of supply disruptions in a large supplier country by about four-fifths in the average receiving country (but not in the source country).

Impact for economic policy makers

The epidemic has dramatically shaped trade development and spread through trade to the economy. Therefore, the protection of exports and imports is another argument for ending the epidemic in all parts of the world through mass vaccination.

Firms will ultimately decide where to produce and source input, but policy can also play a role in the externality and inconsistency of information (Baldwin and Freeman 2022). For example, further simulations in our analysis suggest that reducing trade costs, such as non-tariff barriers, could encourage international diversification. In addition, governments can provide information to companies to gain greater visibility in their supply chains and to facilitate better risk analysis of supply chain networks.

Before the war in Ukraine, the global economy was threatened by calls for restoration; War has added to the risk of global economic disintegration. At the same time, it has exposed the risk of a large concentration of important product imports. By showing the benefits of international diversity and replacement, our analysis underscores what can be achieved through close international integration and cooperation.

Editors note: The opinions expressed in this column are those of the authors, and they do not necessarily represent the views of the International Monetary Fund.


Baldwin, R (2020), “The Greater Trade Fall of 2020: Learning from the 2008-09 Great Trade Collapse”,, April 7.

Baldwin, R and R Freeman (2022), “Global Supply Chain Risk and Resilience”,, April 6.

Bas, M and A Fernandes (2022), “Trade’s Resilience to COVID-19”,, 28 April.

Bonadio, B, Z Huo, AA Levchenko, and N Pandalai-Nyar (2020), “The Role of the Global Supply Chain in the COVID-19 Epidemic and Beyond”,, 25 May.

Bonadio, B, Z Huo, AA Levchenko, and N Pandalai-Nyar (2021), “Global Supply Chains in Epidemic”, Journal of International Economics 133: 103534.

Celasun, O, NJ Hansen, M Spector, A Mineshima and J Zhou (2022), “Supply Disruption Added to Inflation and Reduced Recovery in 2021”,, 16 March.

Espitia, A, A Mattoo, N Rocha, M Ruta and D Winkler (2021), “Trade and Covid-19: Lessons from the first wave”,, 18 January.

IMF (2021), External Sector Report, Washington DC.

IMF (2022), “World Trade and the Price Chain during an Epidemic”, in War Returns to Global Recovery, World Economic Outlook, April 2022.

Lafrogne-Joussier, R, J Martin and I Mejean (2022) “Supply Chain Barrier and Mitigation Strategies”,, February 5.

Santos Silva, JMC, and S Tenreyro (2006), “The Log of Gravity”Economics and Statistics Review 88 (4): 641–58.

Leave a Reply

Your email address will not be published.