The International Trade Law Review: Must an Effective Activist State Harm Trading Partners?

I Background

A long-standing presumption in trade policy circles is that governments should be afforded significant discretion in responding to crises. Even so, it is recognised in the declarations of many international summits and elsewhere that there should be limits on the extent to which external stability is jettisoned in the pursuit of internal stability. In this chapter I present evidence that the frequency of cross-border harm inflicted by G20 members varied considerably in response to the covid-19 pandemic. G20 policy choice differed along a number of other critical dimensions too. Such evidence, I argue, begs the question posed in the title of this chapter; the tension between internal and external stability may be more apparent than real. Technocratic work to identify trade-friendly crisis response packages should commence, providing a robust factual basis for discussions between WTO Members once the covid-19 pandemic is behind us.

II Introduction

The economic fallout from the covid-19 pandemic is the second systemic crisis to confront the multilateral trading system in 15 years. In light of the duration of such crises, including the years taken to recover from them, these events can no longer be dismissed as once-in-a-lifetime events or black swans.

Given the aftermath of the global financial crisis, the potential for crisis responses to permanently scar cross-border commercial flows of goods, services, investments, ideas, data and people can no longer ruled out either.2 Plenty is at stake here. For one, the conclusion of certain leading economic historians of international trade – namely, that the world trading system is less affected by 'shocks' (unanticipated crises) than 'shifts' (evident trends) – may need to be revisited too.3

A rethink should also call into question certain sacred cows that underpin the post-war economic settlement, perhaps the most prominent of which is John Ruggie's notion of embedded liberalism.4 One tenet of this theory is that international arrangements governing trade between nations do not impede governments seeking to restore internal stability (or threats to such stability) even if that means discriminating against foreign commercial interests (thereby threatening external stability). In short, when responding to crises, is it acceptable for governments to sacrifice external stability on the altar of internal stability? This proposition is tantamount to arguing that crisis-induced protectionism is permissible.

There are, of course, stronger and weaker versions of this proposition. Discrimination against foreign commercial interests could be time-limited. It could be targeted and proportionate. Governments engaging in discrimination could inform trading partners and make transparent the nature, extent and duration of measures taken that disadvantage foreign commercial interests.

Arguably, senior officials have recognised the policy dilemmas that arise during systemic crises. That G20 heads of government repeated the mantra in 2009 that they had learned the lessons of the 1930s suggests they recognised the perils of an unconstrained free-for-all in state responses – in particular trade policy responses – to the global financial crisis. Likewise, as the covid-19 pandemic unfolded, on 30 March 2020 the trade ministers of the G20 group declared: 'We agree that emergency measures designed to tackle COVID-19, if deemed necessary, must be targeted, proportionate, transparent, and temporary, and that they do not create unnecessary barriers to trade or disruption to global supply chains, and are consistent with WTO rules.'5

Still, this non-binding G20 statement implies no formal constraint on the crisis-era choices of governments during the covid-19 pandemic. On other occasions officials were at pains to emphasise that multilateral trade rules posed no constraint when formulating crisis responses.6

But what if the tension between internal and external stability is more apparent than real? What if it is possible to devise effective, activist policy intervention during crises that creates limited or no harm to trading partners? That a state must take action to protect its residents during a crisis, such as a pandemic, is not in question. The presumption that international rules should not influence crisis-era response is.

There would be little point asking the question 'must an effective activist state harm trading partners?' if every government response was broadly similar and significantly discriminated against foreign commercial interests. Yet, this is an empirical question that, remarkably, few have addressed either in the context of the global financial crisis or the covid-19 pandemic.7

The bulk of this chapter is devoted to assessing the similarities and differences in the responses of G20 members to the covid-19 pandemic during the calendar year 2020. The purpose is to highlight that G20 members have used the discretion afforded to them in markedly different ways since the pandemic arose.

Some findings presented here also challenge the often-heard contention that large trading nations comply less with long-standing norms of the world trading system. While some readers may draw country-specific lessons from the results presented here, my goal is to encourage further analysis and, ultimately, deliberation on the balance of state discretion and obligations appropriate during system-wide crises.

The remainder of this chapter is organised as follows. Section III addresses foundational questions (such as which crisis-era policy interventions are of interest to trading partners), discusses the evidence base used in this study and presents some summary statistics relating to the quantum of G20 policy responses during 2020. Section IV demonstrates that last year there was considerable variation across the G20 members in their policy responses, which created cross-border spillovers and thereby implicated trading partners. Section V focuses on the general economic support measures taken by the G20 members during 2020 and Section VI presents concluding remarks.

III Foundational matters, evidence deployed and summary statistics on relevant G20 policy responses during 2020

In assessing the significance for the world trading system of G20 policy intervention during the covid-19 pandemic, which policy intervention is in scope? From the perspective of trading partners, not every government policy response during a crisis is of interest to commercial policymakers.8 Economists have long argued that only those policy interventions that induce cross-border spillovers are of interest. Such spillovers may harm or benefit trading partners and policy monitoring exercises should collect information on both.

A slew of crisis-era policy interventions does not generate direct cross-border spillovers by affecting conditions of competition in relevant goods and services markets or by affecting the incentive to engage in foreign direct investment, for example. Unemployment insurance payments do not directly alter the relative treatment of domestic and foreign firms and are out of scope. State contributions to firm labour costs during the covid-19 pandemic need not be in scope either, in particular if a condition for receipt of those payments is that the employees in question remain at home and do not engage in commercial activity. Under these circumstances, the recipient firm is de facto temporarily part of the national welfare state, disbursing monies to idled workers.

Another important design choice relates to the stated motive and purpose of a policy intervention. Some may be tempted to take these into account. However, it is difficult to verify any statement concerning motive or purpose. Furthermore, in the case of subventions to firms, money is fungible. Consequently, if the stated purpose of a subsidy is to (without strings attached) contribute to the labour costs of a company still in commercial operation, then there is no guarantee that the money is not used for some other purpose that affects conditions of market competition. For this reason, a blanket exclusion for subsidies for employee compensation is inappropriate.

Moreover, once certain motives or purposes are deemed benign and therefore out of scope, this risks encouraging inappropriate relabelling of policy intervention. For these reasons, the Global Trade Alert team regards as in scope any policy intervention likely to alter the relative treatment of domestic commercial interests vis-à-vis any group of foreign rivals – irrespective of stated motive or purpose.

A further design choice is whether to include only those policies that are crisis-related. In the context of the covid-19 pandemic, this would amount to including only those policies that are said to have arisen solely because of the pandemic. This is a difficult assessment to make as even routine trade policy decisions (or decisions scheduled before the emergency arose) can be coloured by an unfolding crisis. For this reason, the source used here collected information on all policy intervention likely to induce cross-border spillovers of some kind that were implemented between 1 January 2020 and 31 December 2020. Perhaps it would be better to refer to crisis-era policy intervention than crisis-related policy measures.

In mid-May 2021 information from the Global Trade Alert database was extracted on different facets of in-scope G20 policy response and this provided the factual basis for the findings presented in this chapter.9 This database is a point of reference for policy developments in the world trading system and is increasingly used in official circles, by scholars and analysts and the business community. At the time of writing, over 2,500 entries in the Google Scholar database refer to the Global Trade Alert.

By mid-May 2021, a total10 of 2,204 distinct policy interventions undertaken by G20 members between 1 January and 31 December 2020 had been documented by the Global Trade Alert team. To put this total in context, consider the following. In November 2020, the 26th report of Global Trade Alert was published and it contained information on 1,371 policy interventions undertaken by the G20 members during the first 10 months of 2020.11 Much of the 50 per cent difference in the quantum of G20 policy intervention documented refers to earlier policy intervention that came to light after information collected ceased for the 26th report.12

With product-level international trade data of the finest grain on underlying good trade flows13 and information on the goods and sectors implicated by each recorded G20 policy intervention, it is possible to identify at the product level cross-border spillovers associated with G20 commercial policy intervention.

On the basis of available data in mid-May 2021, I estimate that G20 policy intervention during the calendar year 2020 generated a total of 54,844 beneficial cross-borders spillovers and 129,313 harmful spillovers at the product level. Intra-G20 spillovers totalled 37,901. Least-developed countries saw their commercial interests affected just under 40,000 times by G20 policy intervention implemented last year, providing one indication of the cross-border fallout from G20 commercial policy response during the covid-19 era.

IV No set blueprint: Differential G20 commercial policy responses during 2020

Having summarised statistics on the collective G20 policy response and associated spillovers, attention turns to the contributions of individual G20 members. Of particular interest is whether G20 governments responded to the onset of the covid-19 pandemic in ways that have broadly similar implications for trading partners.

If so, one might conclude there is a widely used blueprint for crisis response. If that blueprint involved frequently harming trading partners, then it would reveal the degree to which external stability is sacrificed to restore internal stability. For some, such a finding might justify affording governments considerable discretion in responding to a crisis. On this view, the collateral damage done to trading partners is an unfortunate consequence of effective crisis response.

However, if G20 members differ in the degree to which their responses to covid-19 inflicted harm on trading partners, then the question of whether effective pandemic responses can be devised at little or no cost to other nations arises. Such a finding might call into question the wisdom of giving governments a free pass during crises.

Taking this discussion forward requires some assessment of how aggregate commercial policy stances can be compared across governments in anything like real time. After a decade of official and independent monitoring of national trade policy stances, a number of practices have emerged and they have been outlined in this chapter.

For example, reports by the World Trade Organization (WTO) Secretariat present counts of policy intervention, typically distinguishing between trade restrictive and trade liberalising measures. For policy intervention implicating trade in goods, estimates of the amounts of trade covered have been presented as well. Similar coverage estimates for services trade or for foreign direct investment flows have not been presented to date, largely on account of limited availability of detailed information on relevant cross-border flows.

The drawbacks of frequency and trade coverage measures are well known to those who use and present them. However, given the lags in the publication of international trade data and of measures of national business environments, macroeconomic policy stances and other relevant variables, little more can be done to inform policymaking in the shorter term. Indeed, at this point it is appropriate to note that the tendency of many analysts to make the perfect (simulations/counterfactuals, advanced econometric estimation, etc.) the enemy of the good (frequency, trade coverage and other simple measures) needs to be kept in check.14

Table 1 presents the simplest way to compare the commercial policy stances of G20 members last year. In this table, for each G20 member, the total number of implemented policy interventions that created negative cross-border spillovers found in the Global Trade Alert database are reported as harmful interventions in the second column. Likewise, all implemented policy interventions that benefited trading partners ('liberalising intervention') are shown in the third column. As there is interest in the total amount of interventions, the number of initiated contingent protection investigations that have yet to result in the imposition of import duties is reported too.

The G20 members were then ranked from the member with the least number of recorded interventions in total to the member with the most. Significant lengths were taken to ensure that the amount of time reporting on each G20 member was comparable. Since the principal source of information used by the Global Trade Alert are official documents, any delays in such publication or any denial of access to official websites could result, in principle, in different recorded totals of information.

Table 1: Measured in terms of frequency of policy intervention, there are three groups of G20 member15
G20 memberTotal number of interventions
HarmfulLiberalisingOngoing contingent protection investigations
United States2824636
Germany219911
Brazil104996
United Kingdom1701815
India974439
France1221111
Canada11974
Italy1071111
Turkey99214
Indonesia78286
Argentina56495
Russia67254
China47376
Australia66715
Japan5511
Republic of Korea4941
South Africa37112
Saudi Arabia3551
Mexico2094

Examining the data presented in Table 1, there are three groups of G20 members. As far as a resort to policy intervention is likely to cause cross-border spillovers, there is a group of five relatively inactive nations. These five nations implemented fewer than 60 relevant policy interventions last year. Then there is a group of five very active G20 members, the governments of which each undertook more than 175 policy interventions last year. In between are nine G20 members that undertook between 75 and 175 steps last year likely to implicate trading partners.

Such differences across G20 members are associated with variations in the number of cross-border spillovers that implicate the commercial interests of trading partners. The five G20 members that implemented the fewest measures last year together created 1,873 product-level spillovers benefiting trading partners and 5,350 spillovers harming them. In contrast the five most active G20 members reported in Table 1 generated just under 20,000 positive product-level spillovers and over 47,000 harmful spillovers.

The breakdown between policy interventions that are harmful and beneficial to trading partners reported in Table 1 is also instructive. Three emerging market members of the G20 adopted policy mixes where 40 per cent or more of the implemented policy intervention benefited trading partners: Argentina (44 per cent), Brazil (48 per cent) and China (41 per cent). Meanwhile, seven G20 members adopted policy mixes where at least 90 per cent of policy intervention adopted last year was harmful to trading partners. The seven are Australia, Canada, France, Germany, Japan, the Republic of Korea and the United Kingdom. Just one out of 56 Japanese policy interventions benefited trading partners. Less than 4% of German policy intervention during 2020 benefited trading partners.

As Table 1 makes clear, G20 members differed significantly in their resort to policy intervention that implicates the interests of trading partners. On the metrics presented there, it seems some governments were able to devise pandemic responses that did not sacrifice their trading partners interests that often.

Another dimension upon which the G20 policy stance can be compared is over the length of time policy interventions that are harmful to trading partners are in force. Some may take the view that crisis-era policy intervention that harms trading partners is inevitable and so what matters is that the intervention be short-lived. Readers may recall that G20 trade ministers called on policy intervention to be temporary, which is not necessarily the same as a measure being short-lived. Still, it does beg the question of whether the share of time-limited policy intervention varies across G20 members.

Table 2 reveals considerable variation across G20 members in resorting to harmful policy intervention in terms of its duration. Over 90 per cent of US, Canadian and Saudi Arabian policy interventions that harmed the interests of trading partners do not have a phase-out date. In contrast, 35 per cent of harmful Mexican policy interventions last year have no set termination date. More than half of Chinese harmful policy interventions have a termination date.

Another difference across the G20 members revealed by Table 2 is the degree to which harmful policy intervention had been phased out by the end of last year (31 December 2020). Forty-six per cent of Chinese measures that were introduced last year and that harmed trading partners were unwound before 2021 began. Italy, the Republic of Korea, Russia and Turkey had removed between 20 per cent and 26 per cent of harmful policy interventions by the end of last year. Some G20 members may have taken their trade ministers' plea for temporary policy intervention more seriously than others.

In terms of resorting to more transparent forms of trade policy intervention, one way to compare across G20 members is to examine the percentage of national policy interventions implemented last year that take the form of the border policies the WTO Secretariat reports in its G20 trade monitoring reports. These border policies are typically salient policy interventions (such as import tariff increases). Governments can choose to implement these higher-profile policy interventions or murkier, less transparent policy interventions, such as firm-specific subsidies.

For the entire G20, a third of the policy interventions implemented last year (714 of 2,204 measures) were transparent border measures of the kind the WTO Secretariat regularly reports statistics on. On this metric, none of the measures Japan implemented last year were transparent. Less than 10 per cent of Australian, Canadian and German policy interventions last year employed policy instruments the WTO reports regularly on. Over 40 per cent of Indonesia's and South Africa's policy interventions were transparent in this sense. Over 50 per cent of Argentina's, Brazil's and Turkey's policy interventions during 2020 involved resorting to a more transparent policy intervention. In sum, the actions of G20 members vary significantly in terms of resorting to transparent and to murkier policy intervention.

Table 2: G20 members varied significantly in their resort to temporary, time-bound policy intervention during 202016
G20 memberPercentage of harmful policy interventions
Lapsed by
31 December 2020
Lapsed during
Q1 2021
Lapsed during
Q2 2021
To lapse during Q3 and Q4 2021Without
phase-out date
Argentina10.7%3.6%0.0%10.7%75.0%
Australia16.7%3.0%1.5%6.1%72.7%
Brazil10.6%1.0%4.8%1.9%81.7%
Canada0.8%3.4%0.0%0.0%95.8%
China46.8%2.1%2.1%4.3%44.7%
France16.4%0.8%5.7%13.1%63.9%
Germany9.6%0.5%2.3%1.4%86.3%
India18.6%4.1%2.1%10.3%64.9%
Indonesia14.1%1.3%0.0%1.3%83.3%
Italy20.6%2.8%8.4%5.6%62.6%
Japan12.7%5.5%1.8%0.0%80.0%
Mexico20.0%20.0%5.0%20.0%35.0%
South Korea26.5%0.0%0.0%2.0%71.4%
Russia23.9%1.5%4.5%7.5%62.7%
Saudi Arabia0.0%0.0%2.9%2.9%94.3%
South Africa13.5%5.4%0.0%0.0%81.1%
Turkey23.2%0.0%0.0%1.0%75.8%
United Kingdom7.1%4.1%1.8%2.4%84.7%
United States4.3%0.4%0.7%0.4%94.3%

As should be fairly evident from the statistics reported in this section, certain narratives that have arisen in recent years concerning which nations fail to comply with multilateral trade norms need revisiting. With regard to some of the metrics presented here, the policy stance of certain emerging markets is 'better' than those of their higher per-capita income counterparts in the G20.

The more significant point, which has systemic as opposed to country-specific implications, is that G20 members did not resort to the same playbook or blueprint in designing their responses to the covid-19 pandemic. To the extent that some G20 members were able to develop effective responses to the pandemic without harming the commercial interests of trading partners, there may be important lessons for devising responses to future crises. Indeed, there may be more than one effective strategy that does little harm to trading partners – there is no presumption here that there is a single blueprint that could guide responses to future crises.

V Subsidies dominate G20 resort to General Economic Support measures

Despite the frequent resort to behind-the-border policy responses, many G20 members long gave up cooperating with the WTO Secretariat on reporting their general economic support measures. This is regrettable as such measures can affect trading partners. For example, a very generous state-provided bailout may have important knock-on effects for competitive dynamics in a sector deemed important for a trading partner's interests.

More generally, concerns about subsidies and other non-market practices have been frequently raised in recent years. Yet, the evidence base provided by governments to the WTO Secretariat is thin and arguably this has hampered deliberation and finding common ground. For the purposes of this chapter, however, general economic support measures are important as they account for 1,345 of the 2,204 policy interventions undertaken by G20 members last year and recorded in the Global Trade Alert database.

For the purposes of what follows here, general economic support measures are taken to include:

  1. subsidies to firms that are likely to affect conditions of competition inside the borders of the implementing jurisdiction;
  2. state provided support to exporters shipping to third markets and to national firms undertaking investments (including mergers and acquisitions) abroad;
  3. the introduction or expansion of localisation measures; and
  4. changes to public procurement regulations or public procurement tenders that direct purchases to domestic suppliers and hiring to nationals.

Table 3 reports on each G20 member's resort to these four types of support measure during 2020. It is worth noting that two-thirds (883) of all general economic support measures involve some type of financial support for firms competing in domestic markets. Another 18 per cent of measures relate to export incentives or support for domestic firms' operations in foreign markets. Therefore, subsidies in sectors where international competition is possible accounted for a significant share of G20 pandemic-era response last year.

Of the subsidies offered by G20 governments, 277 were in the form of state loans, 261 were financial grants, 155 involved state-supported trade finance, 117 were in the form of loan guarantees, 98 involved some type of tax relief and 74 involved state-provided financial support for overseas operations, including the acquisition of corporate assets. Almost all (98 per cent) of subsidy policy changes taken last year by G20 governments disadvantaged foreign commercial interests.

The second column of Table 3 makes clear that every G20 government availed themselves of subsidies during 2020. In terms of resort to subsidies affecting conditions of competition in domestic markets, the United States, France, the United Kingdom, Italy and Germany stand out. In contrast, Mexico and Saudi Arabia both granted 15 or fewer such subsidies last year.

When it comes to state-provided incentives likely to alter conditions of competition abroad, Canada and Germany have been most active. Between them, these two governments account for more than half of such state support awarded by the G20 last year that has been recorded in the Global Trade Alert database. Localisation measures, often linked to state-provided trade finance, are common in Germany and the United Kingdom. Use of public procurement regulations to support domestic firms remains common in the United States.

The final column in Table 3 reinforces the earlier finding that there is considerable variation across the G20 in the resort to time-limited or temporary measures. Mexico and China have the largest shares of time-limited general economic support measures. In fact, those shares are an order of magnitude larger than the shares of time-limited measures implemented during 2020 by Canada and the United States.

Table 4 provides further information on the beneficiaries of the general economic support measures, which could be sector-specific or firm-specific. Alternatively, state support provided by G20 members could be horizontal – that is, in principle benefiting all sectors. Once again, comparing across G20 members, the patterns of intervention vary considerably.

Only in Mexico and Turkey did more than one-fifth of support measures favour firms in agriculture; for every other G20 member the share is less than 0.1. Over 40 per cent of US, Brazilian, French, German, Indian and Russian measures favour manufacturing industries. Two-fifths or more of US, Indonesian and Turkish measures support target commercial interests in service sectors.

Argentina, China and Italy stand out for adopting such high shares of support measures that cover all sectors of the economy (deemed 'horizontal' measures). In each case, more than two-fifths of general support measures benefiting local firms did not explicitly discriminate between those local firms. In contrast, as the last column of Table 4 shows, more than four-fifths of US, Canadian and German general economic support measures benefited specific local firms.

The significant differences across G20 members in the manner in which they implemented general economic support measures in 2020 raises important questions as to what, from the perspective of limiting distortions to the world trading system, are better practices during systemic crises.

The aftermath of the global financial crisis witnessed rising shares of world goods trade competing in markets where some exporting firms were subsidised.17 It remains to be seen whether the covid-19 pandemic will result in a similar ratcheting up of potentially subsidy-distorted trade. Surely the time has come to devise guidelines for governments to support firms and their employees during systemic crises that limit the fallout for the world trading system in both the short and medium term.

Table 3: Subsidies dominate general economic support measures taken by the G20 during 2020
Implementing G20 memberNumber of non-export-related subsidies (subsidies affecting competition in domestic markets)Number of export incentives and support in foreign marketsNumber of localisation measuresNumber of government procurement measuresShare of time-limited measures
Argentina212200.20
Australia427430.30
Brazil473900.19
Canada3171010.02
China205010.69
France930200.37
Germany70646100.12
India1885110.38
Indonesia492410.16
Italy793200.39
Japan2325110.22
Mexico120010.77
Republic of Korea307200.23
Russia3171130.33
Saudi Arabia152730.07
South Africa183000.10
Turkey386300.26
United Kingdom82302600.14
United States16410520.04
Table 4: The domestic beneficiaries of discriminatory general economic support measures vary considerably across the G20
Implementing G20 memberShare of harmful economic support measures that affect agricultural sectorsShare of harmful support measures that affect manufacturing sectorsShare of harmful support measures that affect service sectorsShare of harmful support measures that are classified as 'horizontal'Share of harmful support measures that are classified as firm-specific
Argentina0.0800.0400.2400.4800.000
Australia0.0540.2140.3390.2320.357
Brazil0.0850.4240.1360.2030.492
Canada0.0390.3300.3300.0100.942
China0.0380.1150.3850.4620.000
France0.0740.4740.1680.2210.453
Germany0.0210.4150.1130.1130.815
India0.0950.4760.1190.1670.214
Indonesia0.0180.1250.4640.2680.446
Italy0.0600.2980.1430.4290.286
Japan0.0600.3600.1800.3000.460
Mexico0.2310.2310.0000.2310.154
Republic of Korea0.0000.1540.4100.3590.154
Russia0.0580.4810.4040.0190.135
Saudi Arabia0.0000.1480.4070.3330.111
South Africa0.0000.3330.5240.0480.619
Turkey0.3190.2550.2130.1700.085
United Kingdom0.0000.2460.3550.1300.674
United States0.0180.4150.5440.0280.903

VI Concluding remarks

When the covid-19 pandemic hit, governments inherited multilateral trade rules and underlying conventions concerning appropriate state behaviour when crises threaten internal stability. Those rules essentially give governments wide discretion to tackle crises, even if that discretion involves harming the commercial interests of trading partners. The current rules have been defended by some on the grounds that, during emergencies, multilateral trade disciplines do not stand in the way of governments.

When those rules were devised in the aftermath of the Second World War, national markets were significantly less integrated into the world economy. Under those circumstances privileging internal stability over external stability may have made sense, perhaps on the grounds that there was limited external exposure to foreign government actions and, correspondingly, the risks to external stability were not that large.

With the highly integrated world economy of today, circumstances have changed markedly. So the question arises as to whether it still makes sense to jettison external stability and the benefits it confers whenever a crisis looms into view. This question is all the more compelling given the finding that the trading nations responsible for most of world trade responded in such different ways during the initial phase of the covid-19 pandemic.

Last year, five G20 members were able to devise pandemic responses that generated less than one-eighth of the cross-border spillovers of the five most active G20 members. This begs the question of whether effective state action during crises must inflict considerable harm on trading partners. If it is possible to devise crisis responses – or more importantly, guidelines for governments devising crisis responses – then the tension between internal and external stability loses much of its force and, with it, the logic of the current rules on exceptions at the WTO.

Now is not the time to add reform of exceptions to the WTO work programme. This would be premature, not least because much more fact collection and analysis needs to be done. Such technocratic groundwork could be undertaken after the conclusion of the Twelfth WTO Ministerial Conference, drawing upon the expertise of officials from principal international economic institutions as well as independent experts. Crisis response is typically a complex and fraught matter and any analysis of these matters would need to take account the realities of decision-making in extremis. That such analysis is difficult makes it no less valuable.

Given that in less than 15 years governments have faced two episodes of intense system-wide stress, it is advisable to think more in terms of the world trading system facing recurring emergencies than crises and black swans. The timing, form, context and threats posed by future emergencies cannot be known for certain, but it is prudent to expect that there will be emergencies and that it is our responsibility to prepare for them. Those personnel who fight fires and staff ambulances and hospitals have developed protocols for effective decision-making during emergencies. Trade policy officials need to catch up.


Footnotes

1 Simon J Evenett is professor of international trade and economic development at the University of St Gallen. The author thanks Silvan Hofer and Piotr Lukaszuk for their assistance in preparing the tables for this chapter. The evidence presented in this chapter relates to policy intervention undertaken by members of the G20 during 2020 and recorded in the Global Trade Alert database by 25 May 2021.

2 Baldwin, R.E. and S. J. Evenett (eds.) (2020). COVID-19 and Trade Policy: Why Turning Inward Won't Work. CEPR Press.

3 Irwin, D. and K. O'Rourke (2011). 'Coping with Shocks and Shifts: The Multilateral Trading System in Historical Perspective', National Bureau of Economic Research working paper number 17598. November.

4 Ruggie, J. (1982). 'International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order', International Organization 36(2): 379-415.

5 G20 (2020). G20 Trade and Investment Ministerial Statement. 30 March.

6 If anything, the ongoing debate as to whether to relax certain WTO rules on intellectual property rights relevant to bearding covid-19 is further evidence of the current direction of travel. Multilateral rules, on this view, are to play no role in shaping crisis responses – a position consistent with a hard line interpretation of Ruggie's embedded liberalism.

7 The 26th Global Trade Alert report, published in November 2020, addressed this matter at considerable length. This book chapter can be thought of as updating some of the more pertinent facts concerning G20 pandemic-era policy response. (Evenett, S.J. and J. Fritz (2020). Collateral Damage: Cross Border Fallout from Pandemic Policy Overdrive. The 26th Global Trade Alert report. CEPR Press.)

8 During a pandemic, when public health concerns are of considerable importance, policy changes by foreign governments may be a concern on non-commercial grounds.

9 For a description of the methodology used to collect information in the Global Trade Alert database and a comparison with the reporting by the World Trade Organization see Evenett, S J. (2019), 'Protectionism, state discrimination, and international business since the onset of the Global Financial Crisis', Journal of International Business Policy 2(1): 9-36. For a detailed account of how the Global Trade Alert classifies policy intervention see Evenett, S.J. and J. Fritz (2020), The GTA Handbook, 14 July.

10 This total excludes any trade defence or safeguard investigations initiated by G20 members during 2020 that did not result in duties that year.

11 Evenett, S.J. and J. Fritz (2020). Collateral Damage: Cross Border Fallout from Pandemic Policy Overdrive. The 26th Global Trade Alert report. CEPR Press.

12 That the Global Trade Alert team updates earlier published totals on policy intervention is one of the reasons why the quantum of policy intervention found in that independent initiative exceeds those reported by public sector international organisations.

13 Specifically, the six-digit HS code international trade data found in the United Nations Comtrade database.

14 Moreover, what passes for the 'perfect' often involves deeper analysis of some narrowly defined aspect of the trade policy matter in question. This became evident in the 'serious' academic papers seeking to quantify what amounted to selected aspects of the impact of the US-China trade war. There are significant trade-offs between 'depth' and 'breadth' in applied international trade analysis.

15 Source: Global Trade Alert.

16 Source: Global Trade Alert.

17 See Figure 2 on page 7 of Baldwin, R.E. and S. J. Evenett (eds.) (2020). COVID-19 and Trade Policy: Why Turning Inward Won't Work. CEPR Press.

Get unlimited access to all The Law Reviews content