The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
(Trade) War
and Peace:
How to Impose
International Trade
Sanctions
Gustavo de Souza, Naiyuan Hu, Haishi Li,
and Yuan Mei
October 5, 2022
WP 2022-49
https://doi.org/10.21033/wp-2022-49
*Working papers are not edited, and all opinions are the responsibility
of the author(s). The views expressed do not necessarily reflect
the views of the Federal Reserve Bank of Chicago or the Federal
Reserve System.
(Trade) War and Peace:
How to Impose International Trade Sanctions
Gustavo de Souza*
Naiyuan Hu
Haishi Li
Yuan Mei
October 5, 2022
Abstract
Trade sanctions are a common instrument of diplomatic retaliation. To guide current and future policy, we ask: What is the most cost-efficient way to impose trade
sanctions against Russia? To answer this question, we build a quantitative model of
international trade with input-output connections. Sanctioning countries simultaneously choose import tariffs to maximize their income and to minimize Russia’s income,
with different weights placed on these objectives. We find, first, that for countries with
a small willingness to pay for sanctions against Russia, the most cost-efficient sanction
is a uniform, about 20% tariff against all Russian products. Second, if countries are
willing to pay at least US$0.7 for each US$1 drop in Russian welfare, an embargo on
Russia’s mining and energy products - with tariffs above 50% on other products - is
the most cost-efficient policy. Finally, if countries target politically relevant sectors,
an embargo against Russia’s mining and energy sector is the cost-efficient policy even
when there is a small willingness to pay for sanctions.
Keywords: trade sanctions, tariff, tariff competition
JEL Codes: F13, O24
* Federal
Reserve Bank of Chicago
of Economics, Singapore Management University
School of Global Policy and Strategy, UC San Diego and University of Hong Kong
School
1
1
Introduction
Trade sanctions are a common instrument of diplomacy. In 2016, more than 100 trade
sanctions were active.1 Following the sanctions imposed against Russia, trade sanctions
have once again gained the spotlight in public debate.2 Ultimately, these sanctions are
meant to reduce Russia’s ability to wage war. However, if they severely restrict trade, they
also hurt the sanctioning country. This trade-off raises the question: What is the most costefficient way to apply trade sanctions? How can a government reduce economic activity in
the sanctioned country while minimizing local economic costs?
To answer these questions, we build a model of tariff competition with international trade
and input-output connections.3 In the model, firms produce using labor, locally produced
inputs, and inputs from other countries. To import inputs, firms have to pay an import
tariff. Tariffs are chosen by governments trading off two objectives with different weights.
On the one hand, they want to maximize domestic real income, which is also a measure
of households’ welfare. On the other hand, they want to minimize Russian welfare. If the
government has a high willingness to pay for sanctions, it puts higher weights on hurting
the Russian economy. As the government is trading off the cost of sanctioning Russia and
its own welfare, we refer to these sanctions as cost-efficient sanctions.
To make reliable counterfactuals, we estimate the model to reproduce the effect of tariffs
on the international trade of Russia. Using the difference-in-differences estimation strategy
introduced in de Souza and Li (2021), we found that a 10% ad-valorem tariff against Russia
decreases imports of Russian products by 43% and total imports of the taxed good by 19%,
showing that both Russia and the sanctioning country are negatively affected by tariffs.
Using the method of de Souza and Li (2021) and tariff variation from all countries, we
estimate sectoral trade elasticities to be equal to 6 on average, and they correlate with
estimates by Caliendo and Parro (2015) and those that we recover with the Feenstra (1994)
1
See Kirilakha et al. (2021).
On March 11 2022, US President Biden announced that the United States (US), the European Union
(EU), Group of Seven (G7), and North Atlantic Treaty Organization (NATO) countries were to jointly
remove Russia’s “Most Favored Nation” (MFN) status. The removal of MFN status allowed the allies to
arbitrarily raise tariffs, impose more sanctions, or even bans on Russian imports without explicitly violating
the WTO rules. Since then, a number of countries have announced trade sanctions against Russia.
3
The model builds on Ossa (2014) and Caliendo and Parro (2015).
2
2
method.
We highlight five main findings. First, we show that for countries with a small willingness
to pay for sanctions, the best policy is to impose a small tariff on all products. For instance,
if the sanctioning countries are willing to pay US$0.1 for every dollar of economic damage
in Russia, import tariffs should average 20%.4
Second, if the European Union (EU) is willing to pay above 0.7 dollars for each 1 dollar of
real income loss in Russia, an EU embargo on the mining and energy sectors with 50% tariff
on other sectors is cost-efficient.5 For countries with a large willingness to pay for sanctions,
the main driver of cost-efficient tariffs is the import share – to cause more harm in Russia,
sanctioning countries should target what Russia exports the most, i.e., mining and energy
products.
Third, we show that the EU is the group of countries that can hurt the Russian economy
the most, not the US or other sanctioning allies. Russia exports more to the EU than to
the US or other sanctioning allies.6 Because of that, tariffs imposed by the US or other
sanctioning countries can, at most, reduce Russian real income by only 0.07% or 0.22%,
respectively. In contrast, the EU can reduce real income in Russia by as much as 0.8%.
Therefore, the burden of trade sanctions against Russia has to be carried by the EU.
Fourth, if Russia retaliates, i.e., if it also chooses tariffs to punish the sanctioning countries, the economic consequence of tariff sanctions on Russia would more than double. Because the EU is an important importing origin for Russia but Russia is not an important
exporting destination for the EU, Russia imposing high tariffs on EU cannot decrease EU
welfare much but it causes a large decline in its own welfare.
Finally, we show that if sanctions target sectors with larger political relevance in Russia,
an embargo on Russian mining and energy sectors is cost-efficient even for countries with a
4
For nations with a small willingness to pay for sanctions, there are two forces affecting the choice of
tariffs. On the one hand, the government wants to impose high tariffs on sectors with low trade elasticity,
i.e., on products in which trade flows are less affected by tariffs. Yet, on the other hand, the government
also wants to impose high tariffs on products whose lower import shares from Russia so it does not affect
the local economy very much.
5
The mining and energy sectors include the extraction of crude oil, natural gas, and other energy products
(D05 and D06 in International Standard Industrial Classification (ISIC) Rev. 4) and the coke and refined
petroleum sector (D19 in ISIC Rev. 4).
6
More specifically, Russia’s exports to the EU are 4.85% of country’s total production; its exports to the
US, 0.55%; and its exports to other sanctioning allies, 1.63%.
3
low willingness to pay for sanctions. To calculate political relevance, we link each Russian
individual sanctioned by the US, UK, or EU, whom we call an oligarch, to the companies
associated with that individual in Russia. Using the company’s revenue, we calculate the
revenue share of oligarch-owned companies by sector. If the sanctioning countries are willing
to pay $0.1 for each $1 of consumption drop in Russia’s oligarch income, tariffs on Russian
mining and energy sectors should be above 80%, which would cause imports of mining and
energy products to drop by almost 100%. Therefore, if the goal of an embargo is to target
politically influential sectors, it is the optimal policy even for nations with a low willingness
to pay for sanctions.
This paper contributes to the literature on tariff competition. One strand of this literature investigates the welfare consequences of trade policies that countries implement in
cooperative and non-cooperative games. These settings include tariff cooperation (Ossa
2014), competition on non-tariff trade barriers (Mei 2021), “Most Favored Nation” (MFN)
rules (Bagwell et al. 2021), export subsidies (Beshkar and Lashkaripour 2020), market access concessions (Beshkar et al. 2022), deep trade agreements (Lashkaripour and Lugovskyy
2021), and industrial policies (Bartelme et al. 2021). They find that tariffs should be larger
on sectors with larger trade elasticity and that there are welfare gains from cooperation.
Another strand of this literature develops theories for punitive tariffs. Punitive tariffs can
sustain a cooperative equilibrium (Dixit and Bewley 1987) and thus lead to welfare gains
(Mei 2020), should be higher when trade volume surges (Bagwell and Staiger 1990) and
in small countries (Park 2000), can be more effective when implemented in a multilateral
framework (Maggi 1999, Klimenko et al. 2008) and are easier to enforce than monetary fines
(Limao and Saggi 2008).
We contribute to this literature by studying the problem of countries trading-off sanctions
and welfare-maximizing trade policy. In our model, countries simultaneously choose import
tariffs not only to maximize their income, as in Ossa (2014), Mei (2020), Mei (2021), Bagwell
et al. (2021), Beshkar et al. (2022) and many others, but also to minimize Russia’s income,
with different weights placed on these objectives. As countries put more weight into hurting
the Russian economy, optimal tariffs rise more in the sectors that have larger trade flows
and higher trade elasticities. If the willingness to pay is above US$0.7 for each US$1 drop
4
in Russian welfare, tariffs should target the main sectors of Russian exports regardless of its
trade elasticity.
Our work contributes to the literature on the economic impacts of sanctions. Sanctions
and sanction threats are more effective if they impose more harm on the target and if the
sender is more patient (Eaton and Engers 1992, Lacy and Niou 2004, Whang et al. 2013) and
should optimally trade off between the punishment on the target’s leader and the general
public (Baliga and Sjöström 2022). Empirical works show that the number of sanctions has
risen over time (Elliott and Hufbauer 1999, Felbermayr et al. 2020a, 2021, van Bergeijk 2022).
In the target country, sanctions exacerbate regional inequality (Lee 2018), induce firm exit
(Ahn and Ludema 2020, Crozet et al. 2021) and lower stock market valuation (Draca et al.
forthcoming). In the sender country, they also negatively impact firm business (Felbermayr
et al. 2020b, Gullstrand 2020, Besedes̆ et al. 2021). Sanctions disrupt international trade
(Crozet and Hinz 2020, Miromanova 2021a,b, Kwon et al. 2022).
A few works study the impact of sanctions on Russia in response to its recent aggressions
in Ukraine, including how they affect the Ruble exchange rate (Lorenzoni and Werning 2022,
Itskhoki and Mukhin 2022), the welfare implications of western countries increasing tariffs
and non-tariff trade barriers on Russia (Evenett and Muendler 2022a,b), the consequences
of banning Russian oil imports (Bachmann et al. 2022), and how Russian retaliation by
restricting imports affects Russia’s consumer prices (Hinz and Monastyrenko 2022).
In this paper, we estimate cost-efficient trade sanctions and their economic impacts. Ours
is the first paper to study optimal economic sanctions in a quantitative trade framework.
Different from Eaton and Engers (1992), Lacy and Niou (2004), Whang et al. (2013) and
Baliga and Sjöström (2022), we take the motivation for sanctions as given, and we compute
the set of tariffs that hurt the targeted country the most while costing the least to the
sanctioning countries. Inspired by Baliga and Sjöström (2022), we also calculate cost-efficient
sanctions when the sanctioning countries target the politically-relevant sectors.
We also contribute to the literature that develops new methods to estimate trade elasticities. Previous estimates rely on orthogonality or structural assumptions for identification of
the trade elasticity. Feenstra (1994), Broda and Weinstein (2006) and Soderbery (2015) assume that shocks to export supply and import demand are orthogonal. Head and Ries (2001),
5
Anderson and Van Wincoop (2004), Romalis (2007), Boehm et al. (2020) and Fontagné et al.
(2022) assume that, conditional on non-tariff trade barriers, shocks to trade flows are uncorrelated with tariffs. Eaton and Kortum (2002), Simonovska and Waugh (2014), Bergstrand
et al. (2013), Head et al. (2010) and Caliendo and Parro (2015) impose structural assumptions on non-tariff trade costs.
Different from the previous works, we estimate trade elasticities using difference-indifferences with anti-dumping tariffs. We show that the products that face anti-dumping
investigations that lead to tariff increases have similar trends prior to the tariffs as those
that face anti-dumping investigations that conclude with no tariffs. We extend the framework, which was used in the context of Brazil, to all countries and their trade partners,
which gives us sufficient variations to estimate the trade elasticities at the sector level. Our
estimates of trade elasticity average 6 and correlate with previous estimates in the literature.
The rest of the paper proceeds as follows. In Section 2, we present empirical evidence
for how a large increase in tariffs can disrupt international trade with Russia. In Section 3,
we present the model and the governments’ problems. In Section 4, we calibrate the model.
In particular, we take the sectoral trade elasticities to our empirical estimates. In Section 5,
we show our findings based on model simulations. In Section 6, we conclude.
2
Empirics
We take advantage of the difference-in-differences strategy introduced by de Souza and Li
(2021) to investigate how increase in tariffs disrupts trade, in particular the trade with
Russia. We learn lessons from global anti-dumping (AD) investigations. The empirical
analysis serves two purposes. First, many AD investigations were deployed against Russia
and they often resulted in large tariff increases on Russia. Therefore, using their variations,
these AD policies can help us understand how recent and upcoming import sanctions on
Russia can disrupt trade. Second, using AD investigations that all countries impose on all
trade partners, we can estimate the trade elasticity by sector. We defer the second point to
Section 4, where we calibrate the model to these sectoral estimates.
6
2.1
Institutions
Dumping refers to an act of price discrimination in which an exporter charges a lower price
in the destination market than in its home market.7 The World Trade Organization (WTO)
allows the destination government to impose anti-dumping tariffs to correct for such price
differences, but requires that they must follow certain procedures.8 First, a sufficient number
of firms in a domestic industry should submit a written request to the government. The
request should provide evidence that import competition imposed harm on the domestic
industry. It should also show that the foreign exporters engaged in dumping. Second, upon
receiving the request, the government should establish a committee that investigates the
case. Third, using the evidence collected in the investigation and following WTO rules, the
committee calculates the normal value of the foreign product and the export price. If the
committee finds that the foreign exporter charges a lower price in the export destination
that in its home market, the government will conclude that the foreign exporter is dumping,
and it will impose an AD tariff that equals the price difference.9 If the committee finds
otherwise, no AD tariff will be introduced. Anti-dumping duties should terminate no later
than five years after first being imposed.
2.1.1
Data
Our data source for AD investigation is the Global Anti Dumping Database (Bown 2005).
It contains all AD investigations that 31 major economies conducted on all trade partners.
For each investigation, the database covers the investigated product and its Harmonized
System (HS) code, the exporter and importer, the beginning and conclusion dates of each
investigation, and the measures taken. Our data source for international trade is the United
Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and
Zignago 2010). We merge the two data sets on the country-bilateral and HS 6-digit level.
7
Adjusted for allowances, trade costs, and currencies in different markets. See Section “Fair comparison of
normal value and export price” of WTO’s technical note on anti-dumping (https://www.wto.org/english/
tratop_e/adp_e/adp_info_e.htm).
8
See WTO’s anti-dumping rule ( https://www.wto.org/english/docs_e/legal_e/19-adp_01_e.htm).
9
Some investigations ended with the foreign exporter raising their price to avoid an AD tariff. See WTO
Agreements on Anti-dumping, subsidies, safeguards: contingencies, etc. (https://www.wto.org/english/
thewto_e/whatis_e/tif_e/agrm8_e.htm) These observations are dropped from both the treatment and
control groups.
7
In the Appendix, Table A.1 shows the summary statistics of AD investigations that targeted Russia.10 During the sample period (1995-2020), Russia faced 393 AD investigations,
among which 298 (75%) were ruled positive. Figure A.1a shows the number of AD investigations and affirmative investigations by year. Figure A.1b shows the average tariff rate
by sector conditional on an affirmative ruling. Table A.3 shows the summary statistics, by
country, of the AD investigations that targeted Russia. The US conducted the most AD
investigations on Russia.
Table A.2 shows the summary statistics of AD investigations that all countries imposed
on their trade partners. During the sample period, there were a total of 15,131 AD investigations, among which 10370 (68%) ruled positive. Figure A.2a shows the number of global
AD investigations and affirmative investigations by year. Figure A.2b shows, at the world
level, the average tariff rate by sector conditional on an affirmative ruling. Table A.4 and
A.5 shows the summary statistics of global AD investigations by the investigating country
and the exporting country.
2.1.2
Discussion
Based on the WTO anti-dumping rules, two important lessons can be learned about how
one should identify the impact of AD tariffs. First, one should not compare the products
that receive an AD tariff to those that do not. To initiate an AD tariff, an investigation
committee has to be formed first. As Staiger and Wolak (1994), Prusa (2001), Lu et al.
(2013), and Besedeš and Prusa (2017), among others, show, these investigations can create
trade policy uncertainty and disrupt trade even if they do not conclude with tariff changes.
Furthermore, as de Souza and Li (2021) show, the investigated products have a lower price,
a higher trade volume, a decreasing price price trend, and an increasing trade volume trend,
compared with the products that are not investigated. Both the trade policy uncertainty
and different trends can be a confounding factor when one compares the tariffed products
versus the non-tariffed products without controlling for AD investigations. Therefore, we
limit our sample to only the investigated products.
Second, conditional on an AD investigation, the WTO rules require that whether a tariff
10
We discuss these summary statistics more in Appendix Section A.1.
8
should be imposed and, if so, the size of the tariff should depend on the price difference
that the foreign exporter set in the origin and destination countries in the pre-investigation
period. We can use the product-country fixed effect to control such a difference, and once we
do so, the AD tariff should be exogenous to the potential trends of the treatment and control
groups. To test that countries indeed follow the WTO rules, we confirm in this paper such
compliance with an event study design that shows parallel trends between the two groups
before the treatment. In de Souza and Li (2021), we supplement it with additional evidence:
(1) The AD tariff can be predicted by the exporting country’s price that it charges and the
AD tariff that it faces on the same product in a third country with a high R-squared. (2)
The AD policy applied for a sector is not correlated with the sector’s other benefits from the
government, including political connections, public procurement, subsidies, and tax breaks.
(3) Placebo tests shows that if we replace the real treatment group with one that has similar
trends or if we move the treatment time five years earlier, we do not identify any effect of
AD tariffs.
2.2
2.2.1
Empirical Strategy
Impact of Import Restrictions on Russian Trade
Following de Souza and Li (2021), we use study the impact of AD tariffs on Russian trade
using:
0
yp,c,t = θτp,c,t + βIp,c,t {After AD} + Xp,c,t
β + p,c,t ,
(1)
where yp,c,t is imports of product p by country c from Russia in year t;11 τp,c,t denotes the
AD tariff that country c imposes on product p from Russia in year t; Ip,c,t {After AD} is a
dummy after the beginning of the first AD investigation (it captures common trends between
0
treatment and control leading to the investigation); and Xp,c,t
is a set of controls.12
Our variable of interest is θ, which captures the average effect of AD tariffs on trade.
11
A product is a Harmonized System (HS) 6-digit code.
The controls are a product-country fixed effect, product-year fixed effect, and dummies for the number
of AD investigations.
12
9
As it is common in differences-in-difference, the identifying assumption is of parallel trends
between treatment and control groups. To show supportive evidence for this assumption, we
test for the existence of parallel trends (prior to the beginning of the investigation) using:
yp,c,t =
X
θj τp,c,first Ip,c,t {j Yrs. to AD} +
j
X
βj Ip,c,t {j Yrs. to AD} + ηp,c + ηc,t + p,c,t ,
j
(2)
where yp,c,t refers to imports of product p by country c from Russia in year t and where
τp,c,first denotes the first AD tariff that country c imposed on product p from Russia (this
variable equals zero for the control group).13 Moreover, in this equation, the dummy variable
Ip,c,t {j Yrs. to AD} takes one if year t is j years to the beginning of the first AD investigation; and ηp,c and ηc,y are the country-product fixed effect and country-year fixed effect,
respectively. We are interested in the coefficient θj , which captures the dynamic effect of
AD tariffs in the jth year. Having no pre-trend is equivalent to θj = 0, ∀j < 0. We limit the
sample to product-country pairs that have at least one AD investigation.
2.3
Results
Figure 1 shows how AD tariffs against Russia affect a country’s imports from Russia. AD
tariffs cause a significant and large drop in imports – a 10% increase in AD tariffs is associated
with a nearly 40% decline in imports of the targeted products. The figure also confirms the
nonexistence of a pre-trend; that is, before the increase in tariffs, the treatment and control
groups display similar import trends.
In the Appendix, Figure A.3 shows the impact of AD tariffs on the quantity and price
of the targeted product imported from Russia. Similar to the value of imports, AD tariffs
significantly reduce the quantity of imports from Russia. Figure A.3a shows that, five years
from the beginning of an AD investigation, a 10% increase in AD tariffs leads to about 30%
drop in the quantity of imports from Russia. Figure A.3b shows that it takes longer for the
13
In the difference-in-differences analysis, we limit our sample to the first AD investigation for a product
that a country conducts on Russia. In this way we ensure that there is no other investigation in the preperiod.
10
-6
-4
log(Imports)
-2
0
2
Figure 1: Impact of AD Tariffs on Imports
-5
-3
-1
1
3
Years to Investigation Beginning
Parameter Estimate
5
95% CI
Description: This figure shows the dynamic impact of AD tariffs on imports using Model 2. The impact on yearly imports is
plotted on the y-axis. The number of years to the beginning of the investigation is plotted on the x-axis. We study Harmonized
System (HS) 6-digit level imports. Imports are measured in free on board (FOB), current dollar value terms. The import data
are from the United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and Zignago 2010).
The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020. The sample
includes the product-origins that faced at least one AD investigation. The shaded area contains the 95% confidence interval.
Standard errors are clustered at the product-country level.
import price to respond to AD tariffs. A 10% increase in AD tariffs leads to about a 5% drop
in import prices. This suggests Russian exporters have to lower prices to remain competitive,
and there is incomplete pass-through of AD tariffs to consumers in the destination country.
AD tariffs significantly reduce total exports (to all destinations) of the targeted Russian
product, according to Figure 2. A 10% increase in tariffs leads to about a 15% decline in
the total exports five years after the AD investigation. This indicates that the decline in the
Russian exports of a product to the destination that imposes import restrictions dominates
the potential increase in Russian exports of the same product to other destinations. Indeed,
Column 1 of Table A.10 shows that Russia can only weakly divert exports to other destinations.14 These findings suggest that import sanctions by other countries on Russia will likely
reduce Russian output and income, a hypothesis we build on in Section 3.
Similarly, Figure 2b shows that AD tariffs significantly reduce total imports (from all
14
This result is consistent with de Souza and Li (2021), who also find an insignificant trade diversion effect
of the AD tariffs that the Brazilian government impose on other countries.
11
origins) of the targeted product to the country that imposes the import restriction: 10%
increase in tariffs leads to about a 20% decline in that country’s total imports in the fifth
year from the beginning of the AD investigation. This demonstrates that the decline in
imports of the targeted product from Russia dominates the potential increase in imports of
the same product from other origins. This is further confirmed by Column 2 of Table A.10,
whose results show that AD tariffs also only weakly divert the sanctioning country’s imports
to other origins. These findings suggest that import sanctions by other countries on Russia
will likely also reduce the sanctioning country’s consumption and income, a hypothesis we
will evaluate in Section 3.
Using variation from all AD tariffs imposed against Russia, we show in Table 1 that
tariffs against Russia decrease imports, prices, and total exports of Russian products. A
10% increase in tariffs causes a 43% drop in imports of the targeted product from Russia
(Column 1), with a 37% drop in quantity imported (Column 2) and a 6% drop in the price
of imports (Column 3). Columns 4 and 5 show that a 10% increase in tariffs reduces total
Russian exports of the targeted product by 16% and total imports of the targeted product
by the tariffing country by 19%, respectively.
Table A.7, A.8 and A.9 show that the impacts of tariffs on imports from Russia, total
exports, and total imports are robust to different combinations of fixed effect controls. Column 1 of these figures show the baseline estimates. Column 2 clusters standard errors at
the 4-digit product and importer level. Column 3 uses a dummy that denotes whether an
investigation committee is formed (instead of the number of committees) to control for the
impact of AD investigations. Column 4 controls product, importer, and year fixed effects
separately.
2.4
Discussion of Empirical Results
Tariffs against Russia decrease total imports of the taxed good and exports by Russia. This
result has two implications for sanctions against Russia – one on local welfare and another
on Russian welfare.
First, the empirical results suggests that trade sanctions decrease local welfare. Because
Russian products cannot be easily replaced, the local economy has to pay higher prices to
12
Figure 2: Impact of AD Tariffs on Product-level Total Exports and Total Imports
(b)
Imports from All Origins
-3
-3
log(Total Imports)
-2
-1
0
log(Total Exports)
-2
-1
0
1
Exports to All Destinations
1
(a)
-5
-3
-1
1
3
Years to Investigation Beginning
Parameter Estimate
5
-5
95% CI
-3
-1
1
3
Years to Investigation Beginning
Parameter Estimate
5
95% CI
Description: This figure shows the dynamic impact of AD tariffs on total exports and total imports using Model 2. The
impact on total imports and total exports is plotted on the y-axis. The number of years to the beginning of the investigation is
plotted on the x-axis. Total exports refer to the total exports of the Harmonized System (HS) 6-digit level product by Russia to
all destinations; these exports are of the same the same 6-digit product for which other countries initiated an AD investigation
on Russia. Total imports refer to the imports of the HS 6-digit level product from all origins by the country that initiated
an AD investigation on Russia; these imports are of the same 6-digit product on which the AD tariff has been imposed. The
import data are from the United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and
Zignago 2010). The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
The sample includes the product-origins that faced at least one AD investigation. The shaded area contains the 95% confidence
interval. Standard errors are clustered at the product-country level.
either produce it locally or import it from other countries. Therefore, due to higher prices,
local real income goes down.
Second, the empirical results also suggest that trade sanctions can decrease Russian
welfare. Because Russia decreases its total exports of the tariffed good, it must be the case
that it cannot easily supply it to other countries. The decrease in total Russian demand
leads to a drop in income, output, and prices in Russia.
Therefore, the empirical results indicate that if countries want to sanction Russia, they
have to incur economic loss. Given certain wiliness to pay for sanctions, how should countries
impose tariffs to maximize their own and punish Russia? To answer this question, we build
a model of international sanctions with input-output connection to understand the costefficient sanctions.
13
Table 1: Effect of AD Tariffs on Russian Trade
VARIABLES
Anti-dumping Tariff
(1)
(2)
(3)
Log Value Log Quantity Log Price
-4.295**
(1.890)
-3.695*
(1.951)
-0.552**
(0.217)
(4)
Log Total Exports
(5)
Log Total Imports
-1.577**
(0.726)
-1.867**
(0.743)
Observations
1,534
1,524
1,524
1,534
1,534
R-squared
0.807
0.811
0.872
0.804
0.839
Fixed Effects
Product X Importer, Importer X Year, Number of AD committee, After AD investigation
Cluster
Product X Importer
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Description: This table presents the impact of anti-dumping tariffs that other countries imposed on Russia on Russian trade, estimated with Model 1. Log
Value denotes the log of Harmonized System (HS) 6-digit level free on board (FOB) current dollar value imports from Russia. Log Quantity denotes the log
of the quantity (in metric tons) imported by another country from Russia on the HS 6-digit level. Log Price denotes the log of import price (measured with
value per metric ton) by another country from Russia on HS 6-digit level. Log Total Exports denotes the log of HS 6-digit level total exports value by Russia to
all destinations in the same HS 6-digit product that other countries initiated an AD investigation on Russia. Log Total Imports denotes the log of HS 6-digit
level total imports from all origins by the country that initiated an AD investigation on Russia in the same 6-digit product that the AD tariff is imposed. The
import data is from the United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and Zignago 2010). The AD data is from
the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
3
Model
In this section, we present a multi-sector, multi-country quantitative trade model with inputoutput linkage. The exposition of the model has two sections. In the first section, tariffs are
taken as given. In the second section, we present how governments choose tariffs.
3.1
Demographics
The global economy consists of N countries and J sectors. Each country has a mass Ln of
households. The preference of country n households is a Cobb-Douglas function of sectorlevel consumption goods, Cnj . Households supply labor inelastically.
The household’s problem is the following:
j
max Un =
J
C
{ nj }j=1
s.t.
J j αn
Y
C
j=1
J
X
n
j
αn
,
where
J
X
αnj = 1
j=1
Pnj Cnj = In ,
j=1
where Pnj denotes sector j composite goods price in country n. In denotes the country’s total
14
income. The consumer’s problem implies country n’s households face the following consumer
price index:
PnC
=
J
Y
Pnj
αjn
.
(3)
j=1
3.2
Intermediate Goods Producer
We assume that all markets competitive, just as Caliendo and Parro (2015). Labor is freely
mobile across sectors within a country but is immobile across countries. A representative
firm in country n and sector j produces with labor and intermediate inputs from all sectors
with a Cobb-Douglas technology:
Ynj
=
Ajn
Ljn
γnj
γnj Y
J
Mnj,k
k=1
γnj,k
γnj,k
,
where Ajn denotes the TFP, Ljn denotes sectoral employment, and Mnj,k denotes the quantity
of sector k composite goods that are used by sector j as an input. γnj and γnj,k are input-output
P
coefficients with γnj + Jk=1 γnj,k = 1.
Profit maximization implies that the output price equals the marginal cost:
pjn =
J
j Y
γnj,k
1
γn
[w
]
Pnk
,
n
j
An
k=1
(4)
where wn denotes the wage of country n.
3.3
Composite Goods
A country’s consumers and firms source their composite goods from other countries. Let Qjn
be the quantity of composite goods of sector j used in country n:
"
N
X
j (σ j −1)/σ j
Qjn =
(qni
)
i=1
15
#σj /(σj −1)
,
j
where qni
denotes the quantity of sector j output that country n buys from country i and
where σ j is the elasticity of substitution between countries. Because composite goods are
used as consumption and inputs, it must be the case that:
Qjn
=
Cnj
+
J
X
Mnj,k
(5)
k=1
3.4
Expenditure Share
To get a unit of sector j output from country i, consumers and firms in country n need to
pay:
j j
pn ,
pjni = tjni kni
j
where tjni = 1 + τni
is one plus the ad-valorem tariff that country n imposes on country i and
j
where kni
denotes the iceberg trade cost to ship one unit of sector j’s output from country
i to country n.
After country n chooses the quantity to source from each origin country i to minimize
the cost of producing Qjn , country n’s expenditure share on sector j’s output from country
i equals:
j
πni
j j 1−σ
pi )
(tjni kni
j
= PN
(6)
j
j
j 1−σ j
h=1 (tnh knh ph )
The composite goods price is thus given by:
"
Pnj =
N
X
#1/(1−σj )
j j 1−σ
(tjni kni
pi )
j
.
i=1
From now on, we use θj = σ j − 1 to denote the trade elasticity.
16
(7)
3.5
Market Clearing
Let Xnj = Pnj Qjn denote country n’s total expenditure on sector j’s composite goods. The
market clearing condition for the composite goods implies that:
Xnj
=
J
X
γnk,j
k=1
N
X
X k πk
i
in
+ αnj In ,
tkin
i=1
(8)
where the first term is country n’s demand for inputs and the second term is the consumer’s
demand.
Household income, In , must be equal to labor income, tax revenue, and the trade deficit:
In = wn Ln + Rn + Dn
(9)
where wn Ln is labor income, Rn is tariff revenue, and Dn is the trade deficit. Tariff revenue
can be written as
Rn =
N
X
X k πk
j
τni
i
in
tkin
i=1
.
(10)
Using Equation 8 and the definition of the trade deficit, we can write the labor market
clearing condition:
wn Ln =
J
X
γnj
j=1
N
X
X j πj
i
i=1
tjin
in
.
(11)
With that, we are ready to define an equilibrium given tariffs.
j
Equilibrium given Tariffs Given tariffs {τni
}j,n,i , an equilibrium is defined as a set of
sectoral prices ,{Pnj }n,j , and wages, {wn }n , such that
1. firms maximize profit (Equation 4);
2. the price index satisfies Equations 6 and 7;
3. the goods markets clear, satisfying Equations 8 and 9;
17
4. the labor market clears, satisfying Equation 11;
5. the government budget constraint (Equation 10) holds.
3.6
Tariff Competition
Import tariffs are chosen by governments. Countries are in three groups according to how
they choose tariffs. There are sanctioning countries, the sanctioned country (Russia), and
neutral countries (the rest of the world, ROW). The sanctioning countries choose tariffs
trading off between two objectives. On the one hand, they want to maximize domestic
households’ welfare. On the other, they want to minimize Russian welfare. Russia also
chooses tariffs to maximize its own welfare and to reduce the sanctioning countries’ welfare.
We assume that the neutral countries do not change tariffs.15
Before we define formally the problem of a sanctioning country, let τnR be the vector of
sectoral tariffs that country n imposes on Russia. Let τ−nR be all global tariffs except what
n imposes on Russia. Use Gn (τnR , τ−nR ) to denote the equilibrium welfare in country n
under tariff policy (τnR , τ−nR ):
Gn (τnR , τ−nR ) =
In (τnR , τ−nR )
,
PnC (τnR , τ−nR )
(12)
where In (τnR , τ−nR ) denotes household income (Equation 9) and PnC (τnR , τ−nR ) denotes
the consumer price index (Equation 3).
Conditional on τ−nR , the objective of sanctioning country n is:
gn (τ−nR ) ∈ argmax{τnR } ρGn (τnR , τ−nR ) − (1 − ρ)GR (τnR , τ−nR ),
(13)
s.t. Equilibrium Conditions 4-11,
where ρ is the willingness to pay for sanctions against Russia. In other words, the local
government is willing to pay $ 1−ρ
for every $1 of consumption forgone in Russia. This
ρ
specification nests two special cases. When ρ = 1, country n maximizes its own real income,
and when ρ = 0, country n minimizes Russia’s real income without consideration of its own
15
We also consider the case in which Russia keep its tariffs constant.
18
welfare.16
Russia, the sanctioned country, trades off maximizing its own welfare and retaliating
against the countries that impose the sanctions. Use S to denote the set of sanctioning
countries. Russia’s problem is the following:
gR (τ−RS ) ∈ argmax{τRS } ρGR (τ−RS , τ−RS ) − (1 − ρ)
X Gn (τ−RS , τ−RS )
NS
n∈S
(14)
s.t. Equilibrium Conditions 4-11,
where GR (τ−RS , τ−RS ) is the equilibrium welfare in Russia and
P
n∈S
Gn (τ−RS ,τ−RS )
NS
is the
average real income of the sanctioning countries. As with sanctioning countries, ρ captures
the willingness to pay for tariff retaliation against sanctioning countries.17
j
j
j
Equilibrium with Sanctions Given {{τni
}j,n∈S,i6=R , {τRi
}j,i∈S
/ , {τni }j,n∈S,i6
/ =R }, an equilib-
rium with optimal sanctions is given by tariffs imposed against Russia by sanctioning countries, {τnR }n∈S , tariffs imposed against sanctioning countries by Russia, {τRn }n∈S , a set of
sectoral prices, {Pnj }n,j , and wages, {wn }n , such that
j
1. given tariffs {τni
}j,n,i , {{Pnj }n,j , {wn }n } is an equilibrium;
2. sanctioning countries and Russia optimally choose their tariffs:
τnR = gn (τ−nR ), ∀n ∈ S
τRS = gR (τ−RS ).
To solve a counterfactual equilibrium, we rewrite the model in changes. In this way we
eliminate the fundamentals that are invariant to tariff changes and are difficult to calibrate
16
In Section C.4, we consider an alternative cost-efficient sanction problem where the sanctioning countries
minimize Russia’s welfare but require that their own welfare does not decrease.
17
In the baseline scenario in Section 5, we assume that Russia has the same ρ as the sanctioning countries.
In Section 5.4.2, we also consider that Russia retaliates by always maximizing its own welfare ρRU S ≡ 1
and by always minimizing the sanctioning countries’ welfare ρRU S ≡ 0. We show that sanctioning countries’
strategies and their real income changes are not significantly affected by Russia’s retaliation strategy.
19
j
(for example, non-tariff trade barriers {kni
}j,n,i ). We present the sanction equilibrium in
changes in Appendix Section B.1.
4
Calibration
To calibrate our model, we rely on two main data sources: 1) the OECD Inter-Country
Input-Output (OECD ICIO) Database and 2) estimates of the trade elasticity. We calibrate
the baseline global economy to their levels in 2018, the latest year for which a world inputoutput table is available. We let each sector j ∈ {1, 2, ..., 22} denote the 22 goods sectors
considered in OECD ICIO and j = 23 denotes a merged service sector.18 Countries i, n ∈
{EUN,OSA,ROW,RUS,USA} denote the European Union, other sanctioning countries, rest
of the world, Russia, and United States.19 European Union and United States are the
two largest economies that sanction Russia. Other sanctioning countries comprise Australia,
Canada, Israel, Japan, South Korea, New Zealand, Norway, Singapore, Switzerland, Taiwan,
and United Kingdom, which are the economies that have joined sanctions on Russia by March
31, 2022.20 We combine these other sanctioning economies because of the collaborative
nature of the sanctions, and we reduce the number of countries for which we have to show
the optimal sanctioning tariffs. Rest of the world includes all other economies that are
covered by OECD ICIO. These countries have not imposed sanctions on Russia and will
thus not change their tariffs throughout our analysis. Therefore we combine them into one
economy.21
j
We calibrate country-bilateral and sector level expenditure shares, πni
, country-level
input-output coefficients, γnk,j , country-level value added, wn Ln , and country-level trade
deficit, Dn , directly to their data counterparts in OECD ICIO.
18
See Appendix Table A.11 for the list of OECD ICIO sectors and their correspondence with the International Standard Industrial Classification (ISIC) Rev. 4 sectors. As there is no import tariff variation on the
service sectors, we merge all service sectors into one single sector.
19
The European Union countries that are covered by OECD ICIO are as follows: Austria, Belgium,
Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary,
Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Morocco, Netherlands, Poland, Portugal, Romania,
Slovakia, Spain, and Sweden.
20
See https://graphics.reuters.com/UKRAINE-CRISIS/SANCTIONS/byvrjenzmve/ the evolving list of
countries that have sanctioned Russia.
21
Section C.1 in the appendix discusses statistics of Russian trade.
20
4.1
Estimation of Sector-level Trade Elasticity
We estimate the sector-level trade elasticity, θj , which is arguably the most important parameter for tariff analysis (Hillberry and Hummels 2013).22 We use a similar empirical strategy
as in Section 2. We take advantage of the AD investigations and tariffs that all countries
impose on all their trade partners to identify the elasticity of imports to tariffs, i.e., the
trade elasticity, on the sector level. We estimate it for each goods sector listed in the 2018
OECD Inter-country Input-output Database (OECD ICIO 2018).23 To this end we use a
specification similar to Equation 1:
yp,d,o,t = θj(p) τp,d,o,t + βIp,d,o,t {After AD} + γNp,d,o,t {Committee}
+ ηp,t + ηp,d + ηp,o + ηd,t + ηo,t + p,d,o,t ,
(15)
(16)
where p denotes the product, d denotes the destination, o denotes the origin of the trade
flow, t denotes the year, and j denotes the sector that product p belongs to. Moreover,
yp,d,o,t denotes the trade values of product p from country o to country d in year t. τp,d,o,t
denotes the AD tariff that country d imposes on country o in year t on the same product p.
Ip,d,o,t {After AD} takes one if year t is after the first AD investigation that country d conducts
on product p from country o. Np,d,o,t {Committee} controls the number of investigation
committees formed on the same product-country-bilateral-year level. ηp,t , ηp,d , ηp,o , ηd,t , and
ηo,t denote product-year, product-destination, product-origin, destination-year, and originyear fixed effects, respectively. Table A.6 shows the summary statistics of the variables
included in this regression.
Table 2 shows our estimated trade elasticity, θj , by sector. These elasticities range from
1.36 (other non-metallic mineral products) to 8.98 (mining and energy products). After all
sectors are pooled together, the estimated average trade elasticity equals 6.09. Consistent
22
Remember that θj = σ j − 1 where σ j denotes the elasticity of substitution across countries (see Section
3.3).
23
To ensure that there is sufficient cross-product variation that helps us identify the trade elasticities by
sector, we estimate the elasticity by pooling together the agriculture sector (D01-D02 of ISIC Rev. 4) and
food sector (D10-D12 of ISIC Rev. 4), and pooling together all mining and energy sectors (D05-D09 and
D19 of ISIC Rev. 4). We control for product-year, product-destination, product-origin, destination-year,
and origin-year fixed effects separately to allow for sufficient variation to identify the elasticities by sector.
21
with our intuition, sectors that are perceived less substitutable across countries, for example, minerals and manufactured products, have lower trade elasticities than those that are
perceived more substitutable across countries, for example, energy and chemical products.
Figure A.5 shows that our estimated elasticities are positively correlated with the estimates acquired by Caliendo and Parro (2015) and the values that we estimate with using the
Feenstra (1994) method.24 On average, our estimates are lower than those found in Caliendo
and Parro (2015) (in Figure A.5a more than half of the sectors are below the 45-degree line,
and their average estimate across all sectors equals 9.1).25 Our estimates are higher than
those that we estimate with the Feenstra (1994) method (in Figure A.5b, most sectors are
above the 45-degree line).26
Table 2: Estimated Sectoral Trade Elasticity, θj
Sector
Agriculture
Fishing
Mining energy
Mining non-energy
Mining support
Food
Textiles
Wood
Paper
Petroleum
Chemical
Pharmaceuticals
All
Estimate Standard Err
5.18
1.17
6.96
1.34
8.98
1.47
8.98
1.47
8.98
1.47
5.18
1.17
6.96
1.34
6.01
1.48
4.44
1.71
8.98
1.47
7.45
1.26
5.80
1.27
6.09
0.86
p-value
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.000
0.000
0.000
0.000
Sector
Plastic
Mineral
Basic metals
Fabricated metals
Computer
Electrical
Machinery n.e.c.
Auto
Other transport
Manufacturing n.e.c.
Service
Estimate
5.56
1.36
6.59
5.19
4.97
5.44
5.22
5.98
5.33
4.55
4.17
Standard Err
1.06
1.69
1.20
1.11
1.11
1.29
1.05
1.46
1.17
1.09
1.27
p-value
0.000
0.423
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.001
Description: This table presents the sector-level trade elasticities that we estimate with the difference-in-differences method. The import data are from the
United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and Zignago 2010). The AD data are from the Global Anti-dumping
Database (Bown 2005). The sample runs from 1995 to 2020.
5
Results
24
The method relies on time series variation in prices and market shares of imported varieties of goods.
The identifying assumption is that shocks to import demand and export supply are uncorrelated, which
serves as the moment condition. The trade value and quantity data are from the BACI Database of CEPII
(Gaulier and Zignago 2010), covering the time period 1995–2019.
25
The reason why our estimates are lower than those recovered in Caliendo and Parro (2015) is likely the
role of trade policy. de Souza and Li (2021) shows that our treatment and control groups face similar trade
policy uncertainty before a tariff was imposed. However, Caliendo and Parro (2015) do not control trade
policy uncertainty in their regressions. As trade policy uncertainty is likely positively correlated with tariffs
(Handley and Limão 2017), this can exaggerate the estimates.
26
Using the Feenstra (1994) method, the average elasticity across all sectors equals 2.6.
22
5.1
Cost-Efficient Sanctions
In this section we discuss the cost-efficient sanctions imposed by the EU. Figure 3 shows
statistics of the optimal sanctions imposed by the EU according to the different levels of
willingness to pay for sanctions, ρ. The first panel plots the cost-efficient sanctions for
selected sectors. The second panel plots change in imports in the EU implied by different
sanctioning schemes.
Cost-efficient sanctions are small and uniform across sectors for a small willingness to pay
for sanctions, according to Figure 3. If the EU is willing to pay $0.1 for each $1 of income
drop in Russia, i.e., ρ = 0.9, tariffs should average about 20% for all sectors. They increase
with a higher willingness to pay but the dispersion across sectors is small.27
Even for a small willingness to pay for sanctions, imports from Russia should drop by
more than 80%. If the EU chooses tariffs to maximize its own welfare, i.e., ρ = 1, trade
with Russia would drop by 60%. Higher tariffs against Russia increase the EU’s real income
by reducing the price of imported goods relative to exported ones.28 If the EU has positive
willingness to pay for sanctions, it wants to decreases trade with Russia even further. For a
willingness to pay of only $0.4 per dollar for $1 income dropped in Russia, i.e., ρ = 0.7, the
EU impose tariffs that decrease Russian imports by 95%.
If the EU is willing to pay above $0.7 dollars for each $1 dollar of income loss in Russia,
i.e., ρ is below 0.6, high taxes on mining and energy products are optimal. Figure 4 extends
the horizontal axis of Figure 3 to ρ = 0.4, under which the EU is willing to pay $1.5 for each
$1 of income drop in Russia, and shows optimal tariffs and trade flows changes accordingly.
If the EU would like to pay $0.7 dollars to reduce Russian real income by $1, tariffs on mining
and energy products are specially high, according to Figure 5. A tariff on energy extraction
sector products, which includes crude oil and natural gas, is above 300%, and a tariff on
petroleum is above 200%. In this case, an embargo on Russian oil and gas combined with
high tariffs on other sectors is the most cost-efficient policy. When the willingness to pay
rises to $1.5, tariffs on all sectors are above 80% and an embargo on all sectors is optimal.
27
In Section C.4, we solve the alternative cost-efficient sanction problem where the sanctioning countries
minimize Russia’s welfare with their own welfare non-decreasing. We find that those cost efficient sanction
tariffs resemble the optimal tariffs under low willingness to pay for sanctions.
28
This is the traditional terms-of-trade effect discussed in, for example, Bagwell and Staiger (1990).
23
Figure 3: Cost-Efficient Tariffs in the EU for Different ρ’s, ρ ∈ [0.7, 1.0]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions against Russia that vary by the level of
willingness to pay for them. ρ ranges from 0.7 to 1.0. Figure 3a plots the cost-efficient tariff on mining energy, mining nonenergy, petroleum, and the average unweighted tariff across sectors by different levels of willingness to pay for sanctions, ρ.
Figure 3b plots the percentage change in imports in the EU for different sectors by different levels of willingness to pay for
sanctions, ρ.
Figure 4: Cost-Efficient Tariffs in the EU for Different ρ’s, ρ ∈ [0.4, 1.0]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions against Russia that vary by the level of
willingness to pay for them. ρ ranges from 0.4 to 1.0. Figure 3a plots the cost-efficient tariff on mining energy, mining nonenergy, petroleum, and the average unweighted tariff across sectors by different levels of willingness to pay for sanctions, ρ.
Figure 3b plots the percentage change in imports in the EU for different sectors by different levels of willingness to pay for
sanctions, ρ.
Conditional on the willingness to pay for sanctions, trade elasticities and initial import
share are important determinants of tariffs. For a low willingness to pay, tariffs target
products with a low trade elasticity and a low import share from Russia. In this case,
countries use tariffs to manipulate the terms of trade, i.e., to raise the export price relative
to the import price (see, for example, Bagwell and Staiger 1990). The products in which
24
Figure 5: Cost-Efficient Tariffs in the EU for ρ = 0.6
Description: This figure shows the cost-efficient sectoral tariffs that the EU imposes on Russia at ρ = 0.6 – the EU is willing
to pay 0.7 dollars for one dollar real income loss in Russia.
the terms of trade are more affected by tariffs are the ones with lower demand elasticity and
lower import share.29
When there is a large enough willingness to pay for sanctions, tariffs are targeted at
sectors with a large import share from Russia and with a high trade elasticity. Sanctioning
these sectors can divert more Russian exports to other countries, reduce Russian output
more, and cause more harm in Russia. Appendix C.2 discusses these intuitions in detail.
The USA and Other Sanctioning Allied Countries (OSA) follow the same pattern of costefficient tariffs, according to Figures 6 and 7. In both cases, cost-efficient sanctions are small
and uniform across sectors for small willingness to pay for sanctions but they still cause a
large drop in trade with Russia. Figures C.3 shows that as the willingness to pay increases,
US optimal tariffs increase uniformly across sectors. The reason is that, as Figure A.4a
shows, US expenditure share on Russia is small and similar across sectors. Figure C.4 shows
that for other sanctioning countries, embargo on mining and energy sectors is optimal if they
29
For a formal proof, see Gros (1987), Broda et al. (2008), Opp (2010), Costinot et al. (2015), and
Lashkaripour and Lugovskyy (2021), who derive theories that link optimal tariffs to market shares and trade
elasticities.
25
would like to pay $1 to reduce Russia’s income by $1. If the willingness to pay rises to $1.5,
an embargo by all sanctioning countries on all Russian products is optimal.
Figure 6: Cost-Efficient Tariffs in the USA for Different ρ’s, ρ ∈ [0.7, 1.0]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure 6a
plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across sectors
for different willingness to pay for sanctions, ρ. Figure 6b plots the percentage change in imports in the USA at different sectors.
Figure 7: Cost-Efficient Tariffs in the OSA for Different ρ’s, ρ ∈ [0.7, 1.0]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure 7a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure 7b plots the percentage change in imports
in the other sanctioning allies (OSA) at different sectors.
26
5.2
5.2.1
The Welfare Cost of Sanctions
The Welfare Cost of Sanctions on Russia
How much welfare loss can sanctions cause in Russia? To answer this question, Figure 8
shows the welfare changes of Russia and the sanctioning countries under two scenarios –
with and without Russian retaliation. In the case with retaliation, similar to the sanctioning
countries, Russia chooses tariffs based on Problem 14. In the case without Russian retaliation
(8b), Russian tariffs are constant at the calibrated value.
According to Figure 8, sanctions can decrease Russian welfare between 0.5% and 3%.30
Without Russian retaliation, i.e., if Russian tariffs are constant at the calibrated values,
the welfare loss in Russia ranges from 0.5% to 1.2%, depending on the willingness to pay
of the sanctioning countries. If Russia retaliates, the welfare cost of sanctions can be as
large as 3%. The reason is the economic size difference between Russia and the sanctioning
countries. The sanctioning countries are an important sourcing origin for Russia, whereas
Russia is not an important exporting destination for the sanctioning countries.31 Because
of this, restricting imports from the sanctioning countries cannot reduce the sanctioning
countries’ income much, but it induces large price increase and real income loss in Russia.32
5.2.2
The Welfare Cost of Sanctions on Sanctioning Countries
How much do sanctions cost the sanctioning countries? According to Figure 8, the welfare
cost of sanctions is small. The sanctioning countries face a welfare loss of between 0.1% and
0.2%, depending on if Russia retaliates or not. Despite the fact that Russian imports are a
large share in some sectors of the sanctioning countries, on average the share corresponds to
only 6%, 1%, and 2%, in the EU, USA, and OSA, respectively. Therefore, the losses that
the allies can incur are limited.
30
Under ρ = 1 where the sanctioning countries optimally choose tariffs to maximize their own welfare,
these tariffs also reduce Russia’s welfare (see Figure 8). This is the classical “beggar thy neighbor” effect
considered in, for example, Maggi and Rodriguez-Clare (1998).
31
Russia spends 5.2% of their total expenditures on the sanctioning countries. The sanctioning countries
sell 0.2% of their output to Russia. See Figures A.4c and A.4d.
32
This finding also indicates that, as long as Russia cares about domestic welfare (ρRU S ≥ 0.1), Russia
should not impose high retaliatory tariffs on the sanctioning countries, and the consequences of sanctions
are similar with and without Russian retaliation. In Section 5.4.2 we elaborate this point further. We show
that the sanctioning countries’ optimal tariffs are not significantly affected by Russia’s retaliation strategies.
27
Figure 8: Welfare Changes with and without Russian Retaliation
(a) With Russian Retaliation
(b) Without Russian Retaliation
Description: This figure shows the welfare change by the different willingness to pay for sanctions, ρ, under two variations
of the model discussed in Section 3. Figure 8a displays the welfare changes in Russia, the European Union nations (EUN),
the other sanctioning countries (OSA), and the USA when Russia also changes its tariff to affect the welfare of the sanctioning
countries. Figure 8b displays the welfare change in Russia, the European Union nations (EUN), the other sanctioning countries
(OSA), and the USA under the assumption that Russia keeps its tariffs constant. Welfare refers to the equilibrium real income.
Changes are calculated by comparing the resulting equilibrium from the new tariffs to the equilibrium with initial, pre-sanction
tariffs.
To understand which of the sanctioning countries can impose a larger cost on Russia, we
consider the case where the EU, the USA, and other sanctioning countries individually set
tariffs to target Russia’s real income. For simplicity, we assume that Russia keeps its tariffs
constant. Figure 9 shows the welfare changes of Russia and the sanctioning countries under
unilateral sanctions. Each of the three sub-figures plots the counterfactual equilibrium in
which one country chooses tariffs on Russian imports based on Problem 13 while Russia and
all the other countries keep their tariff constant.
According to Figure 9, the EU is the sanctioning group most affected by sanctions against
Russia. If ρ = 0, i.e., tariffs against Russia are chosen to minimize Russian welfare, the EU
has a welfare loss of 0.1%. The US and OSA would have a welfare loss of only 0.01% and
0.02%. The reason is that the EU is the country that trades the most with Russia.
The EU is also the trade partner that can cause the largest welfare damage in Russia.
The EU alone can reduce welfare in Russia from 0.26% to 0.78%, whereas US sanctions can
only reduce Russian welfare by no more than 0.1%.
28
Figure 9: Welfare Changes with Individual Sanctions
(a)
(b)
EU
(c)
USA
OSA
Description: This figure shows the welfare change for different willingness to pay for sanctions, ρ, under the equilibrium tariffs
of the model with individual sanctions. Figure 9a shows the welfare change if the EU chooses tariffs against Russia to maximize
13 while all the other countries have tariffs constant. Figure 9b shows the welfare change if the US chooses tariffs against Russia
to maximize 13 while all the other countries have tariffs constant. Figure 9c shows the welfare change if OSA chooses tariffs
against Russia to maximize 3 while all the other countries have tariffs constant. Welfare refers to the equilibrium real income.
Changes are calculated comparing to the equilibrium with current tariffs.
5.3
Political Weights
In this section, we calculate cost-efficient sanctions if the sanctioning countries target politically relevant sectors instead of the whole Russian economy. We show that an embargo on
the Russian mining and energy sectors are optimal even when there is a small willingness to
pay for sanctions.
Government’s Problem We assume now that the sanctioning countries want to target
particular sectors in Russia according to their political relevance. Let Gpol
R (τ ) be the politi29
cally weighted welfare in Russia and τ the vector of tariffs imposed by all countries. Formally,
the politically weighted welfare is given by
Gpol
R (τnR , τ−nR )
=
J
X
λj IRj (τnR , τ−nR ),
(17)
j=1
where λj is the political weight of sector j and IRj (τnR , τ−nR ) is real income in sector j.33
34
35
The best response of sanctioning country n can now be formulated as the following:
gnpol (τ−nR ) ∈ argmax{τnR } ρGn (τnR , τ−nR ) − (1 − ρ)Gpol
R (τnR , τ−nR ),
(18)
s.t. Equilibrium Conditions 4-11,
where ρ is the willingness to pay for sanctions against politically relevant sectors.
Calibration We calibrate political weights to reflect the revenue share of companies owned
by individuals sanctioned by the EU, UK, or the USA. First, we collect the names of the
Russian individuals who have been sanctioned by European Union, United Kingdom, and
United States by Mar 10, 2022.36 . Those are part of the Russian political elite, called
oligarchs, which are believed to support the current regime. We also acquire the names of
the companies that they own. Second, we collect the names, sales, and industries of the
top 100 Russian companies by revenue from RBC 500, a website that publishes ratings for
Russian companies, and match them to the list of sanctioned people.37 Third, we connect the
industry names used in RBC 500 to OECD ICIO sectors. In the last step, we calculate, for
33
LjR (τnR ,τ−nR ) IR (τnR ,τ−nR )
,
C (τ
LR
PR
nR ,τ−nR )
IR (τnR ,τ−nR )
is real income.
C (τ
PR
nR ,τ−nR )
Real income in sector j is given by
where
LjR (τnR ,τ−nR )
LR
is Russia’s employ-
ment share in sector j, and
34
In this model, sector j’s employment share is also the sector’s share in Russian GDP.
35
In Appendix Section B.2, we rewrite this optimal sanction problem in changes, where we further elaborate
on this point.
36
The source is an article by the Guardian: https://www.theguardian.com/world/2022/mar/04/
russia-oligarchs-business-figures-west-sanction-lists-us-eu-uk-ukraine
37
The RBC 500 rating has been published since 2015. The rating is to identify the largest Russian
companies in terms of net revenue. The rating involves companies owned by Russian individuals and legal
entities, regardless of their registration - in Russia or abroad. The main source of the financial indicator
comes from consolidated financial statements. In the case of no available consolidated financial statements,
indicators would be estimated.
30
each OECD ICIO sector, the share of sales by major Russian companies owned by oligarchs
in the sector’s total sales (output). We use these shares as our political weights, λRj .
Table 3: Summary Statistics of Political Weights
Sector
# Firms owned by Oligarchs
Agriculture
2
Mining energy
11
Mining non-energy
6
Petroleum
7
Chemical
5
Basic metals
7
Machinery n.e.c.
1
Manufacturing n.e.c.
1
Service
32
Oligarch Share (λRj )
4.89%
47.56%
57.57%
57.84%
41.47%
12.79%
14.31%
9.75%
8.82%
Description: This table presents summary statistics of the political weights. The table shows, in each sector, the
number of top 500 Russian firms owned by Russian oligarchs and the revenue share generated by these firms. We
omit the sectors without major oligarch-owned firms. The data is compiled for this research.
Table 3 shows the summary statistics of political weights computed using Russian oligarchs’ share of revenue in each sector. Nine out of 23 sectors have oligarch-owned firms,
among which the petroleum sector has the highest political weight, 57.84%, indicating that
over half of the revenue in this sector is generated by firms owned by oligarchs.
Figure 10: Cost-Efficient Tariffs with Political Weights in the EU for Different ρ’s,
ρ ∈ [0.7, 1.0]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay when the EU
uses political weights described in 3. Figure 4a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum,
and the average unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure 4b plots the percentage
change in imports in the EU at different sectors.
31
Results Even for countries with a small willingness to pay for sanctions, an embargo
against Russian mining and energy sectors are the most cost-efficient sanction, according to
result in Figure 10. If the EU is willing to pay $0.1 for each $1 of income drop in Russia,
i.e., ρ = 0.9, tariffs should be concentrated in the mining and energy sector. That happens
because those are the sectors with the highest concentration of firms owned by Russian
oligarchs. Moreover, tariffs should be high enough to decrease imports of mining and energy
products from Russia by almost 100%.
Figures C.5 and C.6 show the optimal sanction tariffs and resulting import changes by
the US and other sanctioning countries. Similar to the EU, for small willingness to pay for
sanctions, i.e. $0.1 to reduce Russian real income by $1, an embargo on mining and energy
sector imports from Russia is optimal.38
5.4
Robustness
In this section, we show that the cost-efficient sanctions that we derive are robust to two alternative model specifications. First, we replace the trade elasticities that we estimated with
those that Caliendo and Parro (2015) acquired. Second, we consider alternative retaliation
strategies by Russia.
5.4.1
Caliendo and Parro (2015) Trade Elasticities
We show that the cost-efficient sanctions are robust, if we replace the sectoral trade elasticities that we estimate using the difference-in-differences strategy with the estimates that
Caliendo and Parro (2015) acquired. The reason is that, as we show in Figure A.5a, these
two sets of elasticities are positively correlated.39
Figures C.7 to C.10 show the sanctioning countries’ optimal strategies under these elasticities. If they would like to pay $0.1 for $1 decline in Russia’s welfare, the optimal tariffs
38
Figures C.5 and C.6 also show that if the US and other sanctioning countries only target politically
relevant sectors in Russia, total imports will increase. The reason is that they have incentives to lower tariffs
on the products from Russia that are not politically relevant, such that Russian employment and output
can be reallocated from the politically relevant sectors to these other sectors. As mining and energy sectors
do not account for a major share of these countries’ imports from Russia, a combination of high tariffs on
mining and energy sectors and low tariffs on other sectors can lead to an increase in total imports.
39
The correlation is 0.57.
32
should equal, on average, about 15%. Compared to the our estimated trade elasticities,
lower willingness to pay for sanction can justify mining and energy sector embargo under
Caliendo and Parro (2015) trade elasticities. Petroleum sector embargo by the EU is optimal
if the willingness to pay is as low as $0.1 to reduce Russian real income by $1. If the EU’s
willingness to pay rises to $0.7, an embargo on all mining and energy sectors is optimal.
As Figure A.5a shows, Caliendo and Parro (2015) trade elasticities are higher than ours on
average, and especially so for the energy sectors. Given the willingness to pay for sanction,
higher trade elasticities provide the sanctioning countries incentives to impose higher tariffs
because they can divert more exports away from the opponent and harm their income more.
Similar to Figure 4, if the sanctioning countries are willing to pay $1.5 for $1 reduction
in Russia’s welfare, an embargo on Russian imports in all sectors is optimal.
5.4.2
Russian Retaliation Strategies
The cost-efficient sanctions are also robust to Russian retaliation strategies. The reason is
that, as Russia is a relatively small export destination for the sanctioning countries (see
Figure A.4c), Russia’s retaliation on the sanctioning countries’ exports should not strongly
affect the latter group’s output, income, nor their incentives to impose sanctions.40 In this
section we consider two alternative Russian retaliations: in solving Problem 14, Russia always
sets their retaliation tariff to maximize their own real income (ρRU S ≡ 1) and to minimize
the sanctioning countries’ real income (ρRU S ≡ 0).
Figures C.11 to C.16 show the sanctioning countries’ optimal tariffs and the associated
import changes, with ρRU S ≡ 1. Figure C.17 to C.22 show the same set of variables for
ρRU S ≡ 0. Similar to Section 5 where ρRU S equals that of the sanctioning countries, for
small willingness to pay to sanction Russia ($0.1 for $1 real income drop in Russia), the
sanctioning countries should optimally impose around 20% tariffs on all sectors. If the
sanctioning countries would like to pay $0.7 of their real income to reduce Russian real
income by $1, an embargo on mining and energy sectors is optimal for the EU. For willingness
to pay higher than $1.5, a embargo on all Russian products by all sanctioning countries is
40
This is also corroborated by Figure 8 which shows that the sanctioning countries’ real income is not
significantly affected by ρ – the willingness to pay to minimize the opponent’s real income in both the the
sanctioning countries and Russia.
33
optimal.
6
Conclusion
In this paper we study how countries should optimally impose import sanctions. We investigate how these sanctions depend on countries’ willingness to pay for sanctions, trade
shares, and trade elasticities. We develop a model of tariff competition which features multiple countries, multiple sectors and input-output linkages. Countries weigh the objectives
of maximizing their own income and diminishing their opponent’s income, and they respond
optimally to other countries’ tariff strategies.
Russia’s invasion of Ukraine has caused significant causalities, economic damage, and
threatened global stability and economic prosperity. We apply the model to study the
cost-efficient sanctions on Russia. Basing on the difference-in-differences empirical strategy
developed in de Souza and Li (2021) and global anti-dumping investigations and tariffs
from the Global Anti-dumping Database (Bown 2005), we first document that tariffs on
imports from Russia strongly decrease Russian total exports and the sanctioning country’s
total imports in the targeted products. Using the same empirical strategy, we estimate the
model’s trade elasticity for each sector.
We find that if the sanctioning countries would like to pay $0.1 of real income to diminish
Russian real income by $1, the sanctioning tariffs should be about 20% and similar across
sanctioning countries and across sectors. If the sanctioning countries’ willingness to pay
rises to $0.7, the EU should impose an embargo on the energy and mining sectors. If the
willingness to pay increases to $1.5, an embargo on all sectors is close to optimal.
We also find that sanctions by the EU can lead to larger real income loss in Russia than
the USA and other sanctioning allies. Russian retaliation slightly increases the welfare loss
in the sanctioning countries. However, it leads to substantially larger welfare loss in Russia.
Furthermore, if sanctions target the sectors that are politically relevant, a global embargo on Russia’s mining and energy sectors is optimal even with low willingness to pay for
sanctions.
Many countries have implemented trade sanctions on Russia. With these analyses, we
34
propose a rationale why the observed sanctions differ across countries and sectors. We
provide a toolbox that helps policy makers optimally impose sanctions, as they trade off
between undermining Russia’s capacity to continue its war and the cost on domestic welfare.
35
References
Ahn, D. P. and R. D. Ludema (2020): “The sword and the shield: The economics of
targeted sanctions,” European Economic Review, 130, 103587.
Anderson, J. E. and E. Van Wincoop (2004): “Trade costs,” Journal of Economic
literature, 42, 691–751.
Antràs, P., D. Chor, T. Fally, and R. Hillberry (2012): “Measuring the upstreamness of production and trade flows,” American Economic Review, 102, 412–16.
Bachmann, R., D. Baqaee, C. Bayer, M. Kuhn, A. Löschel, B. Moll, A. Peichl,
K. Pittel, and M. Schularick (2022): “What if? The economic effects for Germany
of a stop of energy imports from Russia,” EconPol Policy Report 36, ifo Institute.
Bagwell, K. and R. W. Staiger (1990): “A Theory of Managed Trade,” American
Economic Review, 80, 779–795.
Bagwell, K., R. W. Staiger, and A. Yurukoglu (2021): “Quantitative Analysis of
Multiparty Tariff Negotiations,” Econometrica, 89, 1595–1631.
Baliga, S. and T. Sjöström (2022): “Optimal Sanctions,” Working Paper.
Bartelme, D., A. Costinot, D. Donaldson, and A. Rodrı́guez-Clare (2021):
“The Textbook Case for Industrial Policy: Theory Meets Data,” Working Paper.
Bergstrand, J. H., P. Egger, and M. Larch (2013): “Gravity Redux: Estimation of
gravity-equation coefficients, elasticities of substitution, and general equilibrium comparative statics under asymmetric bilateral trade costs,” Journal of International Economics,
89, 110–121.
Besedeš, T. and T. J. Prusa (2017): “The hazardous effects of antidumping,” Economic
Inquiry, 55, 9–30.
Besedes̆, T., S. Goldbach, and V. Nitsch (2021): “Cheap talk? Financial sanctions
and non-financial firms,” European Economic Review, 134, 103688.
36
Beshkar, M., P.-L. Chang, and S. Song (2022): “Balance of Concessions in the World
Trade Organization,” .
Beshkar, M. and A. Lashkaripour (2020): “Interdependence of Trade Policies in General Equilibrium,” Working Paper.
Boehm, C. E., A. A. Levchenko, and N. Pandalai-Nayar (2020): “The long and
short (run) of trade elasticities,” Tech. rep., National Bureau of Economic Research.
Bown, C. P. (2005): “Global antidumping database version 1.0,” World Bank Policy
Research Working Paper.
Broda, C., N. Limao, and D. E. Weinstein (2008): “Optimal tariffs and market power:
the evidence,” American Economic Review, 98, 2032–65.
Broda, C. and D. E. Weinstein (2006): “Globalization and the Gains from Variety,”
The Quarterly journal of economics, 121, 541–585.
Caliendo, L. and F. Parro (2015): “Estimates of the Trade and Welfare Effects of
NAFTA,” The Review of Economic Studies, 82, 1–44.
Costinot, A., D. Donaldson, J. Vogel, and I. Werning (2015): “Comparative
advantage and optimal trade policy,” The Quarterly Journal of Economics, 130, 659–702.
Crozet, M. and J. Hinz (2020): “Friendly fire: the trade impact of the Russia sanctions
and counter-sanctions,” Economic Policy, 35, 97–146.
Crozet, M., J. Hinz, A. Stammann, and J. Wanner (2021): “Worth the pain?
Firms’ exporting behaviour to countries under sanctions,” European Economic Review,
134, 103683.
de Souza, G. and H. Li (2021): “The Employment Consequences of Anti-Dumping Tariffs:
Lessons from Brazil,” .
Dekle, R., J. Eaton, and S. Kortum (2007): “Unbalanced Trade,” American Economic
Review, 97, 351–355.
37
Dixit, A. and T. F. Bewley (1987): “Strategic aspects of trade policy,” in Advances
in Economic Theory: Fifth World Congress, New York, NY: Cambridge University Press,
329–362.
Draca, M., J. Garred, L. Stickland, and N. Warrinnier (forthcoming): “On
Target? Sanctions and the EconomicInterests of Elite Policymakers in Iran,” Economic
Journal.
Eaton, J. and M. Engers (1992): “Sanctions,” Journal of Political Economy, 100, 899–
928.
Eaton, J. and S. Kortum (2002): “Technology, geography, and trade,” Econometrica,
70, 1741–1779.
Elliott, K. A. and G. C. Hufbauer (1999): “Same Song, Same Refrain? Economic
Sanctions in the 1990’s,” American Economic Review, 89, 403–408.
Evenett, S. J. and M.-A. Muendler (2022a): “ISOLATING THE RUSSIAN ECONOMY: THE OVERLOOKED ROLE OF INTERNATIONAL SHIPPING COSTS,”
https://www.globaltradealert.org/reports/download/91, accessed: 2022-07-10.
——— (2022b):
“MAKING MOSCOW PAY: HOW MUCH EXTRA BITE WILL
G7 & EU TRADE SANCTIONS HAVE?” https://www.wita.org/atp-research/
making-moscow-pay/, accessed: 2022-07-10.
Feenstra, R. C. (1994): “New Product Varieties and the Measurement of International
Prices,” The American Economic Review, 84, 157–177.
Felbermayr, G., A. Kirilakha, C. Syropoulos, E. Yalcin, and Y. V. Yotov
(2020a): “The global sanctions data base,” European Economic Review, 129, 103561.
Felbermayr, G., T. C. Morgan, C. Syropoulos, and Y. V. Yotov (2021): “Understanding economic sanctions: Interdisciplinary perspectives on theory and evidence,”
European Economic Review, 135, 103720.
38
Felbermayr, G., C. Syropoulos, E. Yalcin, and Y. V. Yotov (2020b): “On the
Heterogeneous Effects of Sanctions on Trade and Welfare: Evidence from the Sanctions on
Iran and a New Database,” School of Economics Working Paper Series 2020-04, Drexel
University.
Fontagné, L., H. Guimbard, and G. Orefice (2022): “Tariff-based product-level trade
elasticities,” Journal of International Economics, 137, 103593.
Gaulier, G. and S. Zignago (2010): “Baci: international trade database at the productlevel (the 1994-2007 version),” .
Gros, D. (1987): “A note on the optimal tariff, retaliation and the welfare loss from tariff
wars in a framework with intra-industry trade,” Journal of international Economics, 23,
357–367.
Gullstrand, J. (2020): “What goes around comes around: The effects of sanctions on
Swedish firms in the wake of the Ukraine crisis,” World Economy, 43, 2315–2342.
Handley, K. and N. Limão (2017): “Policy uncertainty, trade, and welfare: Theory and
evidence for China and the United States,” American Economic Review, 107, 2731–83.
Head, K., T. Mayer, and J. Ries (2010): “The erosion of colonial trade linkages after
independence,” Journal of international Economics, 81, 1–14.
Head, K. and J. Ries (2001): “Increasing returns versus national product differentiation
as an explanation for the pattern of US-Canada trade,” American Economic Review, 91,
858–876.
Hillberry, R. and D. Hummels (2013): “Trade elasticity parameters for a computable
general equilibrium model,” in Handbook of computable general equilibrium modeling, Elsevier, vol. 1, 1213–1269.
Hinz, J. and E. Monastyrenko (2022): “Bearing the cost of politics: Consumer prices
and welfare in Russia,” Journal of International Economics, 137, 103581.
39
Itskhoki, O. and D. Mukhin (2022): “Sanctions and the Exchange Rate,” Tech. rep.,
National Bureau of Economic Research.
Kirilakha, A., G. Felbermayr, C. Syropoulos, E. Yalcin, and Y. Yotov (2021):
“The Global Sanctions Data Base: An Update that Includes the Years of the Trump Presidency,” School of Economics Working Paper Series 2021-10, LeBow College of Business,
Drexel University.
Klimenko, M., G. Ramey, and J. Watson (2008): “Recurrent trade agreements and
the value of external enforcement,” Journal of International Economics, 74, 475–499.
Kwon, O., C. Syropoulos, and Y. V. Yotov (2022): “The Extraterritorial Effects of
Sanctions,” Working Paper.
Lacy, D. and E. M. S. Niou (2004): “A Theory of Economic Sanctions and Issue Linkage:
The Roles of Preferences, Information, and Threats,” The Journal of Politics, 66, 25–42.
Lashkaripour, A. and V. Lugovskyy (2021): “Profits, Scale Economies, andthe Gains
from Trade and Industrial Policy,” Working Paper.
Lee, Y. S. (2018): “International isolation and regional inequality: Evidence from sanctions
on North Korea,” Journal of Urban Economics, 103, 34–51.
Limao, N. and K. Saggi (2008): “Tariff retaliation versus financial compensation in the
enforcement of international trade agreements,” Journal of International Economics, 76,
48–60.
Lorenzoni, G. and I. Werning (2022): “A Minimalist Model for the Ruble During
the Russian Invasion of Ukraine,” National Bureau of Economic Research Working Paper
Series, No. 29929.
Lu, Y., Z. Tao, and Y. Zhang (2013): “How do exporters respond to antidumping
investigations?” Journal of International Economics, 91, 290–300.
Maggi, G. (1999): “The Role of Multilateral Institutions in International Trade Cooperation,” American Economic Review, 89, 190–214.
40
Maggi, G. and A. Rodriguez-Clare (1998): “The value of trade agreements in the
presence of political pressures,” Journal of Political Economy, 106, 574–601.
Mei, Y. (2020): “Sustainable cooperation in international trade: A quantitative analysis,”
Journal of International Economics, 123, 103305.
——— (2021): “Regulatory Protection and the Role of International Cooperation,” .
Miromanova, A. (2021a): “The Effectiveness of Embargoes: Evidence From Russia,”
Working Paper.
——— (2021b): “Quantifying the Trade Reducing Effect of Embargoes: Firm Level Evidence
from Russia,” Working Paper.
Opp, M. M. (2010): “Tariff wars in the Ricardian Model with a continuum of goods,”
Journal of International Economics, 80, 212–225.
Ossa, R. (2014): “Trade Wars and Trade Talks with Data,” American Economic Review,
104, 4104–46.
Park, J.-H. (2000): “International trade agreements between countries of asymmetric size,”
Journal of International Economics, 50, 473–495.
Prusa, T. J. (2001): “On the Spread and Impact of Anti-Dumping,” Canadian Journal of
Economics, 591–611.
Romalis, J. (2007): “NAFTA’s and CUSFTA’s Impact on International Trade,” The review
of Economics and Statistics, 89, 416–435.
Simonovska, I. and M. E. Waugh (2014): “The elasticity of trade: Estimates and
evidence,” Journal of international Economics, 92, 34–50.
Soderbery, A. (2015): “Estimating import supply and demand elasticities: Analysis and
implications,” Journal of International Economics, 96, 1–17.
Staiger, R. W. and F. A. Wolak (1994): “Measuring Industry Specific Protection:
Antidumping in the United States,” NBER Working Papers, 4696.
41
van Bergeijk, P. A. (2022): “Sanctions and the Russian-Ukraine Conflict: a Critical
Appraisal,” Working Paper.
Whang, T., E. V. McLean, and D. W. Kuberski (2013): “Coercion, Information, and
the Success of Sanction Threats,” American Journal of Political Science, 57, 65–81.
42
A
Empirics
A.1
Summary Statistics
Table A.1 shows the summary statistics of AD investigations that targeted Russia. During
the sample period (1995-2020), Russia faced 393 AD investigations, among which 298 (75%)
ruled positive. 150 products that Russia exported faced AD investigations, among which
112 had tariff increases.41 20 countries imposed AD tariffs on Russia. Conditional on an
investigation that leads to an AD tariff increase, the average tariff was 123% and the median
tariff was 43%. Figure A.1a shows the number of AD investigations and affirmative investigations by year. Figure A.1b shows, conditional on an investigation that was ruled positive,
the average AD tariff was the highest on metal and machinery sectors.
Table A.3 shows the summary statistics, by country, of the AD investigations that targeted Russia. The United States conducted the most AD investigations, followed by the
European Union, Canada and Ukraine. Conditional on an affirmative investigation, the AD
tariff rate imposed by the United States was the highest (52.63%).
Table A.2 shows the summary statistics of AD investigations that all countries imposed
on their trade partners. There were a total of 15131 AD investigations, among which 10370
(68%) ruled positive. 1585 products faced AD investigations, among which 1298 had tariff
increases. Conditional on an investigation that lead to an AD tariff increase, the average
tariff was 128% and the median tariff was 55%. Figure A.2a shows the number of global
AD investigations and affirmative investigations by year. Figure A.2b shows, on the world
level, conditional on an investigation that was ruled positive, the average AD tariff was the
highest on mining (non-energy), mining support and automobile sectors. Table A.4 and A.5
shows the summary statistics of global AD investigations by the investigating country and
the exporting country.
41
A product refers to a Harmonized System (HS) 6-digit code.
43
Table A.1: Statistics of AD Investigations that Targeted Russia
Tariff Increase
# Investigations
298
# Products
112
# Countries
20
Avg. Tariff
1.23
Med. Tariff
0.43
No Tariff Chg
105
74
11
0
0
All
393
150
20
0.90
0.33
Notes: This table presents the statistics of the anti-dumping investigations that targeted Russia during 1995 and 2020. Each investigation is conducted on a product from
Russia. The average and median tariff is the simple average and the median across
investigations. The investigation-level tariff is computed in ad-valorem terms.
Table A.2: Statistics of Global AD Investigations
# Investigations
# Products
# Countries inv
# Countries exp
Avg. Tariff
Med. Tariff
Tariff Increase
10370
1298
31
95
1.28
0.55
No Tariff Chg
4761
808
31
89
0
0
All
15131
1585
31
106
0.88
0.29
Notes: This table presents the statistics of the anti-dumping investigations that targeted
Russia during 1995 and 2020. Each investigation is conducted on a product from Russia.
The average and median tariff is the simple average and the median across investigations.
The investigation-level tariff is computed in ad-valorem terms.
A.2
Other Figures and Tables
44
Figure A.1: Summary Statistics of AD Investigations on Russia
(a)
(b)
Number of AD Investigations by Year
Ave. AD Tariff Rate by Sector
Description: This figure shows the summary statistics of the AD investigations and AD tariffs that target Russia. The left
panel shows the number of AD investigations that are ruled affirmative and negative by year. The right panel shows, by sector,
the average tariff rate across AD investigations conditional on an affirmative ruling. The same sector classification is used as
the 2018 OECD Inter-country Input-output Database (OECD ICIO 2018). The AD data is from the Global Anti-dumping
Database (Bown 2005). The sample runs from 1995 to 2020.
Table A.3: AD Investigations that Targeted Russia by Country
Country
United States
European Union
Canada
Ukraine
India
Argentina
China
Philippines
Venezuela
Mexico
Indonesia
Pakistan
Turkey
Colombia
Brazil
South Korea
South Africa
Taiwan
Australia
Peru
Investigations
# Investigations % Treated
76
52.63%
68
82.35%
43
79.07%
32
90.63%
25
48.00%
22
63.64%
22
100.00%
21
42.86%
15
100.00%
12
83.33%
11
100.00%
9
100.00%
9
44.44%
7
100.00%
6
66.67%
5
100.00%
5
100.00%
3
33.33%
1
100.00%
1
100.00%
Treated
Avg. Tariff Med. Tariff
4.97
3.13
0.54
0.35
1.44
0.87
0.39
0.34
0.71
0.68
0.82
0.60
0.26
0.18
0.15
0.15
0.64
0.64
0.49
0.34
0.35
0.29
0.29
0.28
0.17
0.10
1.35
0.50
0.24
0.29
0.25
0.30
0.78
0.78
0.39
0.39
0.29
0.29
0.07
0.07
All
Avg. Tariff Med. Tariff
2.62
1.57
0.45
0.33
1.14
0.86
0.36
0.32
0.34
0.00
0.52
0.60
0.26
0.18
0.06
0.00
0.64
0.64
0.41
0.30
0.35
0.29
0.29
0.28
0.08
0.00
1.35
0.50
0.16
0.16
0.25
0.30
0.78
0.78
0.13
0.00
0.29
0.29
0.07
0.07
Description: This table presents summary statistics of the AD investigations that targeted Russia by the country that initiated the investigation. The table
shows the number of investigations, the fraction of the investigations that lead to a tariff increase, the average tariff rate conditional on an affirmative investigation,
and the average tariff rate of all investigations. The AD data is from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
45
Figure A.2: Summary Statistics of Global AD Investigations
(a)
(b)
Number of AD Investigations by Year
Ave. AD Tariff Rate by Sector
Description: This figure shows the summary statistics of the AD investigations and AD tariffs that all countries imposed on
their trade partners. The left panel shows the number of AD investigations that are ruled affirmative and negative by year. The
right panel shows, by sector, the average tariff rate across AD investigations conditional on an affirmative ruling. The same
sector classification is used as the 2018 OECD Inter-country Input-output Database (OECD ICIO 2018). The AD data is from
the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
Figure A.3: Impact of AD Tariffs on Quantity and Price
(b)
Price
-6
-1
-.5
log(Quantity)
-4
-2
0
log(Price)
0
.5
1
Quantity
2
(a)
-5
-3
-1
1
3
Years to Investigation Beginning
Parameter Estimate
5
-5
95% CI
-3
-1
1
3
Years to Investigation Beginning
Parameter Estimate
5
95% CI
Description: This figure shows the dynamic impact of AD tariffs on the quantity and price of imports using Model 2. The
impact on yearly import quantity and price is plotted on the y-axis. The number of years to the beginning of the investigation
is plotted on the x-axis. The import quantity is measured with Harmonized System (HS) 6-digit level metric tons imported
from Russia by the country that initiated the AD investigation. HS 6-digit level price is measured with the value per metric
ton. The import data are from the United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier
and Zignago 2010). The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to
2020. The sample includes the product-origins that faced at least one AD investigation. The shaded area contains the 95%
confidence interval. Standard errors are clustered at the product-country level.
46
Table A.4: Global AD Investigations by Investigating Country
Country
United States
India
European Union
Canada
Argentina
Brazil
Turkey
China
Australia
Sourh Africa
South Korea
Indonesia
Peru
Mexico
Pakistan
Russia
Malaysia
Colombia
Venezuela
New Zealand
Taiwan
Ukraine
Israel
Philippines
Trinidad and Tobago
Chile
Japan
Jamaica
Uruguay
Costa Rica
Ecuador
Investigations
# Investigations % Treated
3611
66.46%
1997
81.47%
1875
60.27%
1250
63.76%
885
71.75%
681
48.31%
652
75.31%
624
98.40%
544
64.34%
383
63.71%
366
68.03%
320
65.63%
310
64.19%
275
83.64%
259
81.08%
164
59.76%
150
55.33%
132
52.27%
120
84.17%
102
33.33%
96
51.04%
91
91.21%
83
46.99%
44
40.91%
28
82.14%
26
57.69%
19
89.47%
16
93.75%
10
30.00%
9
77.78%
9
11.11%
Treated
Avg. Tariff Med. Tariff
2.17
0.90
0.80
0.51
0.92
0.56
1.91
1.09
1.24
0.62
1.42
0.76
1.59
0.45
0.44
0.29
0.28
0.15
0.54
0.45
0.32
0.28
0.27
0.21
2.35
0.44
1.41
0.81
0.39
0.28
0.34
0.23
0.19
0.13
1.93
0.77
1.42
2.04
1.05
0.58
0.44
0.22
0.89
0.45
2.24
1.18
0.31
0.15
1.76
1.92
0.31
0.23
0.38
0.40
0.51
0.22
0.63
0.55
0.98
0.13
0.30
0.30
All
Avg. Tariff Med. Tariff
1.44
0.37
0.65
0.50
0.56
0.23
1.22
0.49
0.89
0.50
0.69
0.00
1.20
0.29
0.43
0.29
0.18
0.07
0.34
0.28
0.22
0.15
0.18
0.11
1.51
0.29
1.18
0.75
0.32
0.22
0.20
0.15
0.11
0.06
1.01
0.10
1.19
0.95
0.35
0.00
0.22
0.14
0.81
0.41
1.05
0.00
0.13
0.00
1.44
1.92
0.18
0.10
0.34
0.29
0.48
0.22
0.19
0.00
0.76
0.13
0.03
0.00
Description: This table presents summary statistics of the global AD investigations by the country that initiated the investigation. The table shows the number of
investigations, the fraction of the investigations that lead to a tariff increase, the average tariff rate conditional on an affirmative investigation, and the average tariff rate
of all investigations. The AD data is from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
47
Table A.5: Global AD Investigations by Exporting Country
Country
China
South Korea
Taiwan
Japan
India
United States
Indonesia
Thailand
Brazil
Russia
Viet Nam
Malaysia
Germany
Ukraine
France
South Africa
Turkey
European Union
Italy
Spain
United Kingdom
Mexico
Romania
Hong Kong
Canada
Pakistan
Singapore
Argentina
Kazakhstan
Belgium
Slovakia
Australia
Netherlands
Egypt
New Zealand
Saudi Arabia
Chile
United Arab Emirates
Peru
Austria
Bulgaria
Poland
Iran
Macao
Sweden
Venezuela
Belarus
Israel
Macedonia
Hungary
Philippines
Finland
Czechia
Oman
Greece
Uruguay
Switzerland
Luxembourg
Lithuania
Colombia
Moldova
Sri Lanka
Portugal
Denmark
Malawi
Croatia
Trinidad and Tobago
Bahrain
Libya
Norway
Dominican Republic
Ireland
Paraguay
Bangladesh
Estonia
Faroe Islands
Liechtenstein
Guatemala
Latvia
Bosnia and Herzegovina
Nepal
North Korea
Qatar
Cuba
Slovenia
Uzbekistan
Algeria
Georgia
Kyrgyzstan
Kuwait
Laos
Nigeria
Serbia
Armenia
Costa Rica
Ecuador
Jordan
Investigations
# Investigations % Treated
3791
79.87%
1258
62.80%
837
71.92%
689
64.01%
628
63.85%
522
72.22%
498
69.48%
442
72.85%
407
69.78%
393
73.54%
349
65.04%
327
64.83%
307
65.80%
295
83.39%
266
45.49%
255
73.33%
239
45.61%
224
78.57%
180
68.89%
171
51.46%
165
73.33%
162
75.31%
149
79.87%
134
48.51%
130
49.23%
127
66.93%
119
66.39%
116
44.83%
104
84.62%
101
63.37%
97
61.86%
96
51.04%
94
62.77%
83
20.48%
74
8.11%
73
28.77%
71
73.24%
70
77.14%
60
48.33%
57
66.67%
57
94.74%
57
75.44%
54
55.56%
53
5.66%
43
39.53%
40
20.00%
39
66.67%
37
54.05%
36
69.44%
34
50.00%
33
69.70%
30
53.33%
28
82.14%
22
59.09%
21
80.95%
21
42.86%
20
35.00%
20
0.00%
18
33.33%
15
66.67%
14
50.00%
13
76.92%
13
76.92%
12
83.33%
12
100.00%
11
90.91%
11
45.45%
10
0.00%
10
20.00%
10
80.00%
9
88.89%
9
88.89%
9
100.00%
6
100.00%
6
66.67%
6
50.00%
6
0.00%
5
20.00%
5
100.00%
4
75.00%
4
100.00%
4
0.00%
4
100.00%
3
66.67%
3
33.33%
3
33.33%
2
100.00%
2
100.00%
2
100.00%
2
100.00%
2
0.00%
2
100.00%
2
100.00%
1
100.00%
1
0.00%
1
100.00%
1
100.00%
Treated
Avg. Tariff Med. Tariff
2.01
0.99
0.50
0.24
0.97
0.37
0.98
0.60
1.80
0.40
0.84
0.47
0.67
0.50
0.57
0.33
1.04
0.73
1.23
0.43
2.82
0.76
0.49
0.29
0.65
0.39
1.17
0.68
0.97
0.60
1.07
0.79
1.04
0.46
0.65
0.40
0.89
0.45
0.81
0.46
1.14
0.77
0.93
0.56
1.13
0.66
0.88
0.63
0.40
0.28
0.31
0.33
0.59
0.36
0.92
0.85
1.90
0.81
0.54
0.41
1.03
0.62
1.43
0.70
0.47
0.12
0.20
0.20
0.14
0.11
0.53
0.30
0.56
0.28
0.85
0.56
0.49
0.50
1.01
0.54
1.65
0.63
0.69
0.51
0.41
0.29
0.23
0.23
0.44
0.27
2.07
1.44
2.44
1.09
0.47
0.53
2.38
1.26
1.14
1.57
1.05
0.45
1.10
0.41
1.71
0.91
0.58
0.46
0.89
0.55
0.44
0.34
1.37
0.72
0.13
0.38
9.48
0.30
0.98
2.27
1.10
0.69
0.43
0.11
0.28
11.07
0.25
1.03
1.67
1.10
0.53
0.56
1.03
0.33
0.22
0.13
0.28
0.27
0.40
0.55
1.03
0.38
0.22
0.08
0.28
0.31
0.40
0.55
0.52
0.14
0.34
0.18
0.52
0.17
0.28
0.18
0.45
0.21
0.46
0.13
0.13
0.38
0.26
0.20
0.45
0.21
0.46
0.13
0.13
0.38
0.26
0.20
0.40
0.29
0.37
0.40
0.29
0.37
0.04
0.34
0.04
0.34
Avg. Tariff
1.60
0.32
0.70
0.63
1.15
0.61
0.46
0.41
0.73
0.90
1.84
0.32
0.43
0.98
0.44
0.79
0.47
0.51
0.61
0.42
0.84
0.70
0.91
0.42
0.20
0.21
0.39
0.41
1.61
0.34
0.64
0.73
0.30
0.04
0.01
0.15
0.41
0.66
0.23
0.67
1.57
0.52
0.23
0.01
0.17
0.41
1.63
0.26
1.65
0.57
0.73
0.59
1.40
0.34
0.72
0.19
0.48
0.00
0.04
0.25
4.74
0.23
0.76
1.89
1.10
0.63
0.20
0.00
0.21
0.27
0.20
0.11
0.28
0.27
0.27
0.27
0.00
0.10
0.14
0.25
0.18
0.00
0.45
0.14
0.15
0.04
0.13
0.38
0.26
0.20
0.00
0.40
0.29
0.37
0.00
0.04
0.34
All
Med. Tariff
0.61
0.08
0.24
0.29
0.15
0.36
0.20
0.20
0.37
0.33
0.26
0.15
0.21
0.47
0.00
0.38
0.00
0.35
0.18
0.07
0.49
0.37
0.43
0.00
0.00
0.12
0.17
0.00
0.77
0.24
0.24
0.29
0.05
0.00
0.00
0.00
0.14
0.37
0.00
0.54
0.63
0.43
0.15
0.00
0.00
0.00
0.49
0.34
0.63
0.07
0.45
0.12
0.78
0.09
0.45
0.00
0.00
0.00
0.00
0.28
0.20
0.25
0.59
0.76
1.10
0.46
0.00
0.00
0.00
0.32
0.22
0.08
0.28
0.31
0.06
0.27
0.00
0.00
0.17
0.28
0.18
0.00
0.45
0.21
0.00
0.00
0.13
0.38
0.26
0.20
0.00
0.40
0.29
0.37
0.00
0.04
0.34
Description: This table presents summary statistics of the global AD investigations by the exporting country. The table shows the number of investigations, the fraction
of the investigations that lead to a tariff increase, the average tariff rate conditional on an affirmative investigation, and the average tariff rate of all investigations. The AD
data is from the Global Anti-dumping Database. The sample runs from 1995 to 2020.
48
Table A.6: Summary Statistics of Diff-in-Diff Regression by Sector
Sector
Agriculture
Fishing
Mining non-energy
Mining support
Food
Textiles
Wood
Paper
Petroleum
Chemical
Pharmaceuticals
Plastic
Mineral
Basic metals
Fabricated metals
Computer
Electrical
Machinery n.e.c.
Auto
Other transport
Manufacturing n.e.c.
Service
No. Obs
1959
304
124
72
6048
22568
3629
8738
10041
25126
784
8502
4415
100308
5039
11945
7533
13712
1148
1986
35834
1559
No. Prods
10
5
3
3
87
212
64
97
140
259
21
100
69
203
90
111
95
229
29
40
353
29
No. Importer No. Exporter Mean Ave Tariff
6
7
0.03
2
3
0.01
2
2
0.03
2
2
0.03
21
37
0.03
18
30
0.03
15
25
0.03
20
38
0.03
19
45
0.04
23
61
0.04
11
12
0.05
22
32
0.03
22
29
0.04
21
58
0.03
22
31
0.04
21
38
0.03
22
33
0.03
24
44
0.04
13
14
0.04
13
13
0.04
27
52
0.03
8
19
0.04
Mean Log Value
8.68
4.66
5.98
6.35
4.95
-2.96
2.08
2.60
2.86
3.04
6.38
4.42
3.18
-0.48
5.24
4.30
4.25
5.37
6.18
5.85
4.88
1.62
Sd. Ave Tariff
0.07
0.05
0.07
0.08
0.07
0.07
0.07
0.07
0.08
0.07
0.09
0.07
0.08
0.07
0.08
0.07
0.07
0.08
0.08
0.08
0.07
0.08
Sd. Mean Log Value
6.70
9.12
3.52
3.43
7.46
10.24
9.53
9.01
9.00
8.79
5.81
8.19
8.27
10.15
7.56
8.15
8.61
7.60
7.73
7.70
7.90
9.96
Table A.7: Effect of AD Tariffs on Imports from Russia
Dependent Variable
Anti-dumping Tariff
Log Value
(1)
(2)
(3)
(4)
-4.295**
(1.890)
-4.295**
(1.917)
-4.468**
(1.992)
-3.966*
(2.004)
1,534
0.804
X
X
1,638
0.688
Observations
1,534
1,534
R-squared
0.807
0.807
Product X Importer
X
X
Importer X Year
X
X
Number of AD committee
X
X
After AD investigation
X
X
Product
Importer
Year
Dummy for AD committee
Cluster
Product X Importer 4-digit X Importer
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
X
X
Product X Importer
X
X
X
X
X
Product X Importer
Description: This table presents the impact of anti-dumping tariffs imposed on Russia by other countries on the log of imports from Russia by the country that
initiated the AD investigation. We study Harmonized System (HS) 6-digit level imports. Imports are measured in free on board (FOB), current dollar value terms.
The coefficients are estimated with Model 1. Different columns include different combinations of fixed effects and controls. The import data are from the United
Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier and Zignago 2010). The AD data are from the Global Anti-dumping Database
(Bown 2005). The sample runs from 1995 to 2020.
49
Table A.8: Effect of AD Tariffs on Total Exports
Dependent variable
Anti-dumping Tariff
Log Total Exports
(3)
(1)
(2)
-1.577**
(0.726)
-1.577*
(0.796)
Observations
1,534
1,534
R-squared
0.804
0.804
Product X Importer
X
X
Importer X Year
X
X
Number of AD committee
X
X
After AD investigation
X
X
Product
Importer
Year
Dummy for AD committee
Cluster
Product X Importer 4-digit X Importer
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
(4)
-1.445*
(0.741)
-1.013
(0.989)
1,534
0.803
X
X
1,638
0.657
X
X
Product X Importer
X
X
X
X
X
Product X Importer
Description: This table presents the impact of anti-dumping tariffs imposed on Russia by other countries on the log of Harmonized System (HS) 6-digit level
total exports by Russia to all destinations. The coefficients are estimated with Model 1. Total exports refer to total exports of the HS 6-digit level product by
Russia to all destinations; these exports of the same 6-digit product for which other countries initiated an AD investigation on Russia. Different columns include
different combinations of fixed effects and controls. The import data are from the United Nations Comtrade Database acquired through the BACI Database of
CEPII (Gaulier and Zignago 2010). The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
Table A.9: Effect of AD Tariffs on Total Imports
Dependent Variable
Anti-dumping Tariff
Log Total Imports
(3)
(1)
(2)
-1.867**
(0.743)
-1.867**
(0.764)
Observations
1,534
1,534
R-squared
0.839
0.839
Product X Importer
X
X
Importer X Year
X
X
Number of AD committee
X
X
After AD investigation
X
X
Product
Importer
Year
Dummy for AD committee
Cluster
Product X Importer 4-digit X Importer
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
(4)
-1.890**
(0.826)
-1.499*
(0.871)
1,534
0.837
X
X
1,638
0.738
X
X
Product X Importer
X
X
X
X
X
Product X Importer
Description: This table presents the impact of anti-dumping tariffs imposed on Russia Russia by other countries on the log of Harmonized System (HS) 6-digit
level total imports by another country from all origins. The coefficients are estimated with Model 1. Total imports refer to the total imports of the HS 6-digit level
product from all origins by the country that initiated an AD investigation on Russia; these imports are of the same 6-digit product on which the AD tariff has been
imposed. Different columns include different combinations of fixed effects and controls. The import data are from the United Nations Comtrade Database acquired
through the BACI Database of CEPII (Gaulier and Zignago 2010). The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from
1995 to 2020.
50
Table A.10: Effect of AD Tariffs on Trade Diversion
VARIABLES
Anti-dumping Tariff
(1)
(2)
(3)
Log Exports to Other Destinations Log Imports from Other Origins Log Imports of Other Products
0.202
(0.525)
0.112
(0.372)
-0.214
(0.344)
Observations
1,063
1,062
1,064
R-squared
0.888
0.908
0.901
Fixed Effects
Product X Importer, Importer X Year, Number of AD committee, After AD investigation
Cluster
Product X Importer
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Description: This table presents the impact of anti-dumping tariffs imposed on Russia by other countries on the trade diversion to other destinations, origins, and products,
estimated with Model 1. Log Exports to Other Destinations denotes the log of exports of Harmonized System (HS) 6-digit level product by Russia to all destinations except the
country that imposed the anti-dumping tariff on the same 6-digit product from Russia (for which an AD investigation was initiated). Log Imports from Other Origins denotes the
log of Harmonized System (HS) 6-digit level imports by the country that imposed the anti-dumping tariff on the same 6-digit product from Russia (for which an AD investigation
was initiated). Log Imports of Other Products denotes the imports from Russia of all HS 6-digit level products within the same HS 2-digit category except the HS 6-digit product
that faces an AD investigation by another country. The import data are from the United Nations Comtrade Database acquired through the BACI Database of CEPII (Gaulier
and Zignago 2010). The AD data are from the Global Anti-dumping Database (Bown 2005). The sample runs from 1995 to 2020.
51
Table A.11: OECD ICIO Sectors and International Standard Industrial Classification (ISIC) Revision 4 Sectors
Sector
Agriculture
Fishing
Mining energy
Mining non-energy
Mining support
Food
Textiles
Wood
Paper
Petroleum
Chemical
Pharmaceuticals
Plastic
Mineral
Basic metals
Fabricated metals
Computer
Electrical
Machinery n.e.c.
Auto
Other transport
Manufacturing n.e.c.
Service
OECD ICIO
Agriculture, hunting, forestry
Fishing and aquaculture
Mining and quarrying, energy producing products
Mining and quarrying, non-energy producing products
Mining support service activities
Food products, beverages and tobacco
Textiles, textile products, leather and footwear
Wood and products of wood and cork
Paper products and printing
Coke and refined petroleum products
Chemical and chemical products
Pharmaceuticals, medicinal chemical and botanical products
Rubber and plastics products
Other non-metallic mineral products
Basic metals
Fabricated metal products
Computer, electronic and optical equipment
Electrical equipment
Machinery and equipment, nec
Motor vehicles, trailers and semi-trailers
Other transport equipment
Manufacturing nec; repair and installation of machinery and equipment
Electricity, gas, steam and air conditioning supply
Water supply, sewerage, waste management and remediation activities
Construction
Wholesale and retail trade; repair of motor vehicles
Land transport and transport via pipelines
Water transport
Air transport
Warehousing and support activities for transportation
Postal and courier activities
Accommodation and food service activities
Publishing, audiovisual and broadcasting activities
Telecommunications
IT and other information services
Financial and insurance activities
Real estate activities
Professional, scientific and technical activities
Administrative and support services
Public administration and defence; compulsory social security
Education
Human health and social work activities
Arts, entertainment and recreation
Other service activities
Activities of households as employers;
undifferentiated goods- and services-producing activities of households for own use
ISIC Rev. 4
1-2
3
5-6
7-8
9
10-12
13-15
16
17-18
19
20
21
22
23
24
25
26
27
28
29
30
31-33
35
36-39
41-43
45-47
49
50
51
52
53
55-56
58-60
61
62-63
64-66
68
69-75
77-82
84
85
86-88
90-93
94-96
97-98
Description: This table shows the relationship between the OECD ICIO sectors that we consider and the International Standard Industrial Classification (ISIC) Rev. 4
sectors.
52
Figure A.4: Trade Statistics with Russia by Sector
(a) Other Countries’ Expenditure Share
(b) Russia’s Share of Output
(c) Other Countries’ Share of Output
(d) Russia’s Expenditure Share
Description: This figure shows, by country and sector, the trade statistics with Russia. Figure A.4a shows the share from
Russia in total expenditure on the sector’s products. Figure A.4b shows the share of output sold to each country in Russia’s
sectoral total output. Figure A.4c shows the share of output sold to Russia in other countries’ sectoral total output. Figure
A.4d shows Russia’s expenditure share from other countries in Russia’s total expenditure on the sector’s product.
53
Figure A.5: Correlation between Trade Elasticities Estimated with Diff-in-Diff and
the Estimated Values (1) in Caliendo and Parro (2015) and (2) that we estimate
with the Feenstra (1994) method
(a)
(b)
Correlation with the elasticities that we estimate with
the Feenstra (1994)
Correlation with Caliendo and Parro (2015) elasticities
Description: This figure shows the sector-level trade elasticities that we estimate with the difference-in-differences method
and those estimated in the literature. The left panel shows the correlation between our elasticities with those that Caliendo
and Parro (2015) find. The right panel shows the correlation between our elasticities with what we estimate with the Feenstra
(1994) on the same level of sectors.
54
B
Model
B.1
Equilibrium in Changes
Using the “exact hat algebra” technique popularized by Dekle et al. (2007), we can express
the equilibrium conditions in terms of changes from the baseline equilibrium, given changes
j0
j
in tariffs t̂jni j,n,i (t̂jni = (1 + τni
)/(1 + τni
)):
ĉjn
= (ŵn )
j
γn
J
Y
k,j
(P̂nk )γn
(B.1)
k=1
"
P̂nj =
N
X
#−1/θj
−θj
j
πni
(t̂jni ĉji )
(B.2)
i=1
j
π̂ni
=
0
Xnj
=
ĉji t̂jni
(B.3)
P̂nj
J
X
γnj,k
k=1
In0
!−θj
0
N
X
πk
0
in
Xik
k0
t
i=1 in
= ŵn wn Ln +
ŵn =
J
X
j=1
γnj
Dn0
+
+ αnj In0
0
j0 j
j 0 Xn πni
τni j 0
tni
j=1 i=1
J X
N
X
j0
πin
j0
,
Xi j 0
tin wn Ln
i=1
N
X
(B.4)
(B.5)
(B.6)
j
To compute the counterfactuals, we only need data on bilateral trade shares πni
, the
j
share of value added in production γnj , value added wn Ln , incumbent tariffs τni
, the share of
intermediate consumption γnk,j , and sectoral trade elasticity θj .
We follow Ossa (2014) and Caliendo and Parro (2015) to construct a trade flow matrix
without trade imbalance using the approach introduced in Dekle et al. (2007). All later
calculations of welfare changes given counter-factual tariffs will treat this purged trade flow
data as the factual equilibrium.
Equilibrium in Changes Given Tariff Changes Given changes in tariffs t̂jni
j,n,i
, an
equilibrium is defined as changes in sectoral prices, {P̂nj }n,j , and wages, {ŵn }n , such that
1. Firms maximize profit (Equation B.1);
55
2. The price index satisfies Equations B.2 and B.3;
3. The goods markets clear, satisfying Equations B.4 and B.5;
4. The labor market clears, satisfying Equation B.6;
Tariff Competition with Equilibrium in Changes Following the notations in Section
3.6, we denote changes in the sectoral tariffs that country n imposes on Russia with t̂nR ,
and all other tariffs–changes in tariffs imposed on Russia by all countries except country n
and those that country n imposes on all other countries except Russia–with t̂−nR .
The change in country n’s welfare equals the following:
Ĝn (t̂nR , t̂−nR ) =
Iˆn (t̂nR , t̂−nR )
P̂nC (t̂nR , t̂−nR )
,
(B.7)
where Iˆn (t̂nR , t̂−nR ) denotes the change in country n income and P̂nC (t̂nR , t̂−nR ) denotes the
change in country n’s consumer price index.
Conditional on t̂−nR , the objective of sanctioning country n is to both maximize their own
welfare (real income, or GNI) and minimize Russian welfare in the counterfactual equilibrium:
gn (t̂−nR ) ∈ argmax{t̂nR } ρGn Ĝn (t̂nR , t̂−nR ) − (1 − ρ)GR ĜR (t̂nR , t̂−nR ),
(B.8)
s.t. Equilibrium Conditions B.1-B.6,
where Gn and GR denotes country n and Russia’s real income in the baseline equilibrium.
We calibrate them to the country’s Purchasing Power Parity (PPP) adjusted GNI in 2018,
the same year as the OECD ICIO data.42
Russia, when it retaliates, it maximizes a weighted average of maximizing its own welfare
42
The data source is the World Bank. See https://data.worldbank.org/indicator/NY.GNP.MKTP.PP.
CD?locations=1W-EU-RU-US.
56
and minimizing the allies’ welfare, both in the counterfactual equilibrium:
gR (t̂−RS ) ∈ argmax{t̂RS } ρGR ĜR (t̂RS , t̂−RS ) − (1 − ρ)
X Gn Ĝn (t̂RS , t̂−RS )
n∈S
NS
(B.9)
s.t. Equilibrium Conditions B.1-B.6
Equilibrium in Changes with Sanctions Given changes in all tariffs except what
the sanctioning countries impose on Russia and Russia imposes on sanctioning countries,
{t̂jni }j,n,i \ t̂nR , t̂Rn n∈S , an equilibrium with optimal sanctions is given by tariffs imposed
against Russia by sanctioning countries, {t̂nR }n∈S , tariffs imposed against the sanctioning
countries by Russia, {t̂Rn }n∈S , a set of sectoral prices, {P̂nj }n,j , and wages, {ŵn }n , such that
1. Given tariffs {t̂jni }j,n,i , {{P̂nj }n,j , {ŵn }n } is an equilibrium;
2. Sanctioning countries and Russia optimally choose changes in their tariffs:
t̂nR = gn (t̂−nR ), ∀n ∈ S
t̂RS = gR (t̂−RS )
B.2
Equilibrium in Changes with Political Weights
Here we rewrite the sanctioning countries problem in changes, if they target the politically
relevant sectors in Russia. The change in Russia’s politically weighted welfare equals the
following:
Ĝpol
R (t̂nR , t̂−nR ) =
J
X
PS
j=1
where
λj LjR
PS
s s
s=1 λ LR
λj LjR
s s
s=1 λ LR
L̂sR (t̂nR , t̂−nR )
IˆR (t̂nR , t̂−nR )
P̂RC (t̂nR , t̂−nR )
,
(B.10)
denotes the politically weighted initial employment share of sector j in
Russia. The politically weighted welfare increases if the whole economy’s real income,
IˆR (t̂nR ,t̂−nR )
,
C (t̂
P̂R
nR ,t̂−nR )
increases, and if employment/value added increases more in the sectors that
have higher political weights. On the other hand, if other countries would like to target the
57
politically related sectors, on top of reducing the whole economy’s real income, they can also
set tariffs to reduce employment/value added in the politically related sectors.
Similar to Problem B.8, the objective of sanctioning country n is to both maximize their
own welfare and minimize Russia’s real income in the politically connected sectors:
pol
gnpol (t̂−nR ) ∈ argmax{t̂nR } ρGn Ĝn (t̂nR , t̂−nR ) − (1 − ρ)Gpol
R ĜR (t̂nR , t̂−nR ),
(B.11)
s.t. Equilibrium Conditions B.1-B.6,
where Gpol
R =
PJ
j=1
Lj
λj LRR GR equals Russia’s politically weighted real income in the initial
equilibrium. We calibrate λj to Russia’s political weights,
LjR
LR
to Russia’s sectoral employment
shares, and GR to Russia’s GNI.
Same as Problem B.9, we assume that Russia retaliates by maximizing a weighted average
of its own welfare and punishment on the sanctioning countries’ welfare.
C
C.1
Quantitative
Sectoral Trade Statistics with Russia
In Figure A.4 we highlight two findings from analyzing the sanctioning countries’ trade
statistics with Russia. First, Russia’s exports of mining and energy products to the EU
is important for both the EU’s consumption and Russia’s production. Mining and energy
sector products from Russia accounts for about 20% of the EU’s total expenditure on these
products (Figure A.4a), and exports of these products to the EU accounts for more than
a quarter of Russia’s total output (Figure A.4b). This suggests that among all sanctioning
countries, the EU should carry a heavy load.43 If the EU sanctions mining and mining
products from Russia, this can cause significantly economic losses in Russia, but it may also
hurt the EU’s welfare.
43
The EU accounts for 36% of Russia’s total exports (5% of Russia’s total output) and 39% of Russia’s
total imports (4% of Russia’s total expenditure). The US accounts for 4% of Russia’s total exports (1%
of Russia’s total output) and 5% of Russia’s total imports (0.5% of Russia’s total expenditure). Other
sanctioning countries account for 12% of Russia’s total exports (2% of Russia’s total output) and 11% of
Russia’s total imports (1% of Russia’s total expenditure).
58
Second, the sanctioning countries are a major importing origin for Russia. However,
Russia is not a major exporting destination for the sanctioning countries. Figure A.4d shows
that Russia spends about 50% of its total expenditure on machinery and pharmaceutical
sector products, and more than a quarter of its total expenditure on electrical equipment
and chemicals, on imports from the sanctioning countries. However, the share of exports to
Russia in the sanctioning countries’ sectoral output never exceeds 5% (Figure A.4c). This
suggests that tariff retaliation by Russia may not cause substantial harm on the income of
sanctioning countries. Rather, it can significantly reduce Russia’s welfare. If Russia cares
about their own welfare, it should be optimal for them not to impose high retaliatory tariffs.‘
C.2
Cost-Efficient Tariffs and Fundamentals
To study the relationship between sectoral optimal tariffs imposed on Russia and the fundamental differences across sectors, we use the following regression:
τs,ρ = βσ,ρ σs + βups,ρ upstreamnesss + βimp,ρ ImpShrs + s,ρ
(C.1)
where τs,ρ is the optimal tariff imposed by the EU on sector s imports from Russia when
EU’s willingness to pay is governed by ρ. σs is the trade elasticity of sector s. upstreamnesss
measures sector s’ upstreamness Antràs et al. (2012) – the average number of sectors one
dollar of sector s output goes through until it reaches the final consumer. ImpShrs is the
share of imports from Russia in total sector s’ imports by the EU in the baseline economy.
To make each variable comparable, they are normalized to have mean 0 and variance 1.
In Figure C.1 we plot the raw correlations between sectoral optimal tariffs by the EU
and individual sector characteristics for ρ = 1 and ρ = 0.6. In Figure C.2, we plot how
the partial correlations change with ρ. For low willingness to pay for sanctions, i.e., when
ρ is large, higher tariffs should be set on sectors that have small import shares, lower trade
elasticities and are more downstream. For example, when ρ = 1, a sector whose import
share is one standard deviation smaller should be targeted by a tariff that is 0.5 standard
deviation higher. Similarly, a sector whose trade elasticity is 1 standard deviation larger
should face a tariff that is 0.4 standard deviations lower. These relationships change when
59
Figure C.1: Correlations of Sectoral Cost-Efficient Tariffs and Fundamentals
(a) Tariffs and Initial Import Share, ρ = 1
Tariff
0
.1
2
.11
4
.12
Tariff
6
.13
8
.14
10
.15
(b) Tariffs and Initial Import Share, ρ = 0.6
0
.1
.2
.3
.4
Init. Imp. Shr.
.5
0
.2
.3
.4
Init. Imp. Shr.
.5
(d) Tariffs and Trade Elasticity, ρ = 0.6
.1
0
.11
2
.12
4
Tariff
Tariff
.13
6
.14
8
.15
10
(c) Tariffs and Trade Elasticity, ρ = 1
.1
2
4
6
Trade Elasticity
8
10
2
6
Trade Elasticity
8
10
(f) Tariffs and Upstreamness, ρ = 0.6
Tariff
0
.1
2
.11
4
.12
Tariff
6
.13
8
.14
10
.15
(e) Tariffs and Upstreamness, ρ = 1
4
1.5
2
2.5
Sector Upstreamness
3
1.5
2
2.5
Sector Upstreamness
3
Description: This figure shows the raw correlations between the sectoral optimal tariffs imposed on Russia by the EU and
different sector-level fundamentals under ρ = 0 and ρ = 0.6.
ρ falls below 0.7, i.e. when the EU places greater weight in punishing Russia. For these
higher willingness to pay, sectors with larger import shares and higher trade elasticities are
60
targeted with higher tariffs.
-.5
0
β
.5
1
Figure C.2: Correlations of Tariffs with Different Fundamentals for Different ρ
.5
.6
.7
ρ
Trade Elasticity
Init. Imp. Shr.
.8
.9
1
Upstreamness
Description: This figure shows the partial correlations between the sectoral optimal tariffs imposed on Russia by the EU and
different sector-level fundamentals under different ρ’s (estimated with Equation C.1).
61
C.3
Other Figures and Tables
Figure C.3: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρ ∈ [0.4, 1]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure
C.3a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.3b plots the percentage change in imports in the USA at
different sectors.
Figure C.4: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρ ∈ [0.4, 1]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.4a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.4b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
62
Figure C.5: Cost-Efficient Tariffs with Political Weights for Different Levels of
Willingness to Pay in the USA, ρ ∈ [0.7, 1]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay when the
USA uses political weights described in Table 3. Figure C.5a plots the cost-efficient tariff on mining energy, mining non-energy,
petroleum, and the average unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.5b plots
the percentage change in imports in the USA at different sectors.
Figure C.6: Cost-Efficient Tariffs with Political Weights for Different Levels of
Willingness to Pay in the OSA, ρ ∈ [0.7, 1]
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the OSA under cost-efficient sanctions with different willingness to pay when the
OSA uses political weights described in Table 3. Figure C.6a plots the cost-efficient tariff on mining energy, mining non-energy,
petroleum, and the average unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.6b plots
the percentage change in imports in the OSA at different sectors.
63
Figure C.7: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρ ∈ [0.7, 1] and Caliendo and Parro (2015) trade elasticities
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure C.7a
plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across sectors
for different willingness to pay for sanctions, ρ. Figure C.7b plots the percentage change in imports in the EU at different
sectors.
Figure C.8: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρ ∈ [0.4, 1] and Caliendo and Parro (2015) trade elasticities
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure C.8a
plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across sectors
for different willingness to pay for sanctions, ρ. Figure C.8b plots the percentage change in imports in the EU at different
sectors.
64
Figure C.9: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρ ∈ [0.7, 1] and Caliendo and Parro (2015) trade elasticities
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.9a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.9b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
Figure C.10: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρ ∈ [0.7, 1] and Caliendo and Parro (2015) trade elasticities
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (USA) under cost-efficient sanctions with different
willingness to pay. Figure C.10a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.10b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
65
Figure C.11: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρSanction ∈ [0.7, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure
C.11a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.11b plots the percentage change in imports in the EU at
different sectors.
Figure C.12: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρSanction ∈ [0.4, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure
C.12a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.12b plots the percentage change in imports in the EU at
different sectors.
66
Figure C.13: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρSanction ∈ [0.7, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure
C.13a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.13b plots the percentage change in imports in the USA at
different sectors.
Figure C.14: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρSanction ∈ [0.4, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure
C.14a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.14b plots the percentage change in imports in the USA at
different sectors.
67
Figure C.15: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρSanction ∈ [0.7, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.15a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.15b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
Figure C.16: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρSanction ∈ [0.4, 1] and ρRU S ≡ 1
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.16a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.16b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
68
Figure C.17: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρSanction ∈ [0.7, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure 3a
plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across sectors
for different willingness to pay for sanctions, ρ. Figure 3b plots the percentage change in imports in the EU at different sectors.
Figure C.18: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
EU, ρSanction ∈ [0.4, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the EU under cost-efficient sanctions with different willingness to pay. Figure 4a
plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across sectors
for different willingness to pay for sanctions, ρ. Figure 4b plots the percentage change in imports in the EU at different sectors.
69
Figure C.19: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρSanction ∈ [0.7, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure
C.19a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.19b plots the percentage change in imports in the USA at
different sectors.
Figure C.20: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
USA, ρSanction ∈ [0.4, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the USA under cost-efficient sanctions with different willingness to pay. Figure
C.20a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average unweighted tariff across
sectors for different willingness to pay for sanctions, ρ. Figure C.20b plots the percentage change in imports in the USA at
different sectors.
70
Figure C.21: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρSanction ∈ [0.7, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.21a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.21b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
Figure C.22: Cost-Efficient Tariffs for Different Levels of Willingness to Pay in the
OSA, ρSanction ∈ [0.4, 1] and ρRU S ≡ 0
(a)
(b)
Tariff
Change in Imports
Description: This figure shows statistics of the other sanctioning allies (OSA) under cost-efficient sanctions with different
willingness to pay. Figure C.22a plots the cost-efficient tariff on mining energy, mining non-energy, petroleum, and the average
unweighted tariff across sectors for different willingness to pay for sanctions, ρ. Figure C.22b plots the percentage change in
imports in the other sanctioning allies (OSA) at different sectors.
71
C.4
Sanctioning Countries’ Real Income Does not Decrease
In this section we study the cost-efficient sanctions with which the sanctioning countries’
real income does not decrease and they minimize the real income in Russia. The sanctioning
country, n’s problem (in changes) is the following. Conditional on all other tariffs except
those country n imposes on Russia, t̂−nR , the objective of country n is:44
gn (t̂−nR ) ∈ argmin{τn,R } GR ĜR (t̂nR , t̂−nR ),
(C.2)
s.t. Equilibrium Conditions B.1-B.6,
Ĝn (τnR , τ−nR ) ≥ 1
Figure C.23 shows cost-efficient sanctions by the EU that satisfy Problem C.2. These
tariffs are similar to those if the sanctioning countries have low willingness to pay to sanction
Russia. They are low (about 10%) on average and similar across sectors. As the EU does not
want to decrease their own real income, lower tariffs should be imposed on energy extraction
and petroleum sectors.
44
We assume Russia’s retaliation strategy is to set tariffs on the sanctioning countries to maximize Russia’s
welfare – it follows Problem 14 where we set ρRU S = 1.
72
Figure C.23: Cost-Efficient Tariffs with EU Welfare Non-decreasing
Description: This
figure shows the cost-efficient sectoral tariffs that the EU imposes on Russia when the EU solves Problem C.2 – the EU
minimizes Russian welfare but requires that its own welfare does not decrease.
73
@FedFRASER