The futility of economic sanctions in a globalized and interdependent world: a data-driven game theoretical study (2025)

Economic sanctions, defined as the imposition or the threat imposition of economic measures on one country (the target country) by another (the sender country) to achieve political goals, have been an important weapon in response to geopolitical conflicts (Bornstein, 1968; Hufbauer and Jung, 2020; Hufbauer et al., 1990; Kaempfer and Lowenberg, 1988; Pintor et al., 2023; Syropoulos et al., 2024). Due to the growing economic interdependence of countries worldwide, economic sanctions have been increasingly used in recent decades (Biersteker et al., 2016, Felbermayr et al., 2020; Morgan et al., 2014, 2023). Various forms of economic sanctions, such as trade restrictions, economic boycotts, and asset freezes, have been deployed to serve a wide range of policy objectives, including defending geopolitical interests, promoting democracy, and fighting terrorism (Carter, 1987; Doxey, 1980; Martin, 1994; Morgan et al., 2023; Peksen, 2009).

Economic sanctions inflict severe economic damage to both the sender and target countries due to reduced trade flows between countries and disruptions of international economic order and result in negative political consequences, especially on the target country, by threatening political stability and violating human rights (Allen and Lektzian, 2013; Hufbauer and Jung, 2021; Neuenkirch and Neumeier, 2015; Özdamar and Shahin, 2021; Peksen and Drury, 2010). Countries vulnerable to economic sanctions seek effective countermeasures to prevent and combat sanctions (Hackenbroich et al., 2020). Some researchers and policymakers suggested reducing the reliance on trade with potential sender countries through supply chain decoupling to lower the risk of being subjected to sanctions (Eppinger et al., 2021; Hufbauer and Jung, 2020). Others argue that complete decoupling is impossible under the deep integration of economies worldwide (Eppinger et al., 2021; Farrell and Newman, 2020). Temporary retaliation, in the form of retaliatory economic measures imposed by the target country to raise the sender’s cost of sanctions and make it lift sanctions, is regarded as an affordable solution (Cranmer et al., 2014; Goldstein and Pevehouse, 1997). However, there are concerns that individual retaliation against powerful countries is unlikely to be effective. To address these concerns, a new strategy of “collective resilience”, where vulnerable countries team up against the common adversary, has recently been proposed as a promising alternative (Cha, 2023). Although the literature on the effectiveness of economic sanctions is extensive (Eaton and Engers, 1992; Egger et al., 2024; Hufbauer et al., 1990; Kaempfer and Lowenberg, 1988; Lacy and Niou, 2004; Pape, 1997; Whang, 2010), the effectiveness of counter-sanction measures has rarely been discussed (Cranmer et al., 2014; Joshi et al., 2024; Maggi, 1999; Peksen and Jeong, 2022). Recent studies provide theoretical analysis and empirical evidence to explore the impact of countermeasures in specific economic sanction cases (Dong and Li, 2018; Hedberg, 2018). However, a comprehensive data-driven evaluation of the effectiveness of counter-sanction measures under different geopolitical scenarios is still lacking.

To fill the gap, we propose a data-driven game theoretical framework to study economic sanctions considering the impact of counter-sanction measures adopted by the target country. Details of the framework are shown in Fig. 1. In summary, the interactions between the sender and target countries are modelled as a three-stage sequential game, which includes the threat, sanction, and counter-sanction (or retaliation) stages. We focus, in particular, on trade sanctions, where economic sanctions and counter-sanction measures in the game take the form of import and (or) export restrictions on certain products. Leveraging the latest global input–output data, we measure the economic loss of sanctions and counter-sanction measures for different countries based on quantitative trade models. We analyse the impact of three counter-sanction measures: decoupling, individual short-term retaliation, and collective short-term retaliation. Our game-theoretical framework assesses each country’s capability of and vulnerability to economic sanctions and enables quantification of the feasibility and effectiveness of different counter-sanction measures.

The economic sanction game

We develop a three-stage sequential game of perfect information and common rationality, as depicted in Fig. 1. Two players, the sender and target countries, are involved in the sanction game. The sender country threatens to impose economic sanctions on the target country to persuade the target to make political concessions. At the threat stage, the target country has the first move in the game and decides whether or not to make political concessions (i.e., “violate” or “comply” with the political goals sought by the sender country). Then, at the sanctions stage, the sender country decides whether or not to impose economic sanctions to punish the target (i.e., “sanction” or “not sanction”). Finally, at the counter-sanction stage, the target country decides whether or not to resort to retaliatory measures (or counter-sanction measures; we use both terms interchangeably in this paper) against the sender’s sanctions (i.e., “retaliate” or “not retaliate”). We assume that the sender will not resort to punishment if the target complies, and the target will not retaliate if the sender does not impose sanctions.

The sender imposes sanctions by temporarily cutting off the imports of certain products from the target, i.e., imposing temporary import restrictions. The target retaliates through one of three counter-sanction measures: (a) decoupling, where the target will permanently break all trade linkages with the sender on the global supply chain network; (b) individual short-term retaliation, where the target will temporarily cut off the imports of certain products from the sender; and (c) collective short-term retaliation, where the target and its bloc members will temporarily cut off the imports of certain products from the sender. The intensity of sanctions/counter-sanction measures for a country is measured by its aggregate gross domestic product (GDP) loss based on the hypothetical extraction method (Los et al., 2016). Economic sanctions are credited with success if the target complies; otherwise, sanctions fail. We define the effectiveness of sanctions as the probability that sanctions succeed in supporting an arbitrary political goal and the effectiveness of counter-sanction measures as the reduction in the effectiveness of sanctions under the threat of retaliation.

Formulation of the game

We assume a political conflict exists between the sender and the target countries that provides a benefit \({b}_{{\rm {P}}}^{{\rm {S}}}\) to the sender and inflicts a cost of \({c}_{{\rm {P}}}^{{\rm {T}}}\) on the target. If the target compiles, the sender will not impose sanctions, and the game ends. The sender’s and target’s payoffs are \({{\mathcal{P}}}_{{\rm {C}}}^{{\rm {S}}}={b}_{{\rm {P}}}^{{\rm {S}}}+{b}_{{\rm {E}}}^{{\rm {S}}}\) and \({{\mathcal{P}}}_{{\rm {C}}}^{{\rm {T}}}={b}_{{\rm {E}}}^{{\rm {T}}}-{c}_{{\rm {P}}}^{{\rm {T}}}\), respectively. Here, \({b}_{{\rm {E}}}^{{\rm {S}}}\) and \({b}_{{\rm {E}}}^{{\rm {T}}}\) are the economic benefits of pre-sanction trade linkages for the sender and the target, respectively. If the target violates, the payoffs are \({b}_{{\rm {E}}}^{{\rm {S}}}\) and \({b}_{{\rm {E}}}^{{\rm {T}}}\) for the sender and the target, respectively, at the threat stage. Then, at the sanction stage, the sender decides whether or not to sanction by cutting off mutually beneficial trade linkages with the target. Sanctions inflict economic costs to both the sender and the target, denoted by \({c}_{{\rm {E}}}^{{\rm {S}}}\) and \({c}_{{\rm {E}}}^{{\rm {T}}}\) for the sender and the target, respectively, but benefit the sender politically in the long term (measured as \({b}_{{\rm {PI}}}^{{\rm {S}}}\)) by building a reputation for toughness (Eaton and Engers, 1999). If the sender decides not to sanction, the game ends. The sender’s and target’s payoffs are \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {NS}}}^{{\rm {S}}}={b}_{{\rm {E}}}^{{\rm {S}}}\) and \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {NS}}}^{{\rm {T}}}={b}_{{\rm {E}}}^{{\rm {T}}}\), respectively. If the sender imposes sanctions, the payoffs at the sanction stage are \({b}_{{\rm {E}}}^{{\rm {S}}}-{c}_{{\rm {E}}}^{{\rm {S}}}+{b}_{{\rm {PI}}}^{{\rm {S}}}\) and \({b}_{{\rm {E}}}^{{\rm {T}}}-{c}_{{\rm {E}}}^{\rm {{T}}}\) for the sender and the target, respectively. The target country might resort to retaliatory measures against the sender’s sanctions. Retaliation inflicts economic harm to both the sender and the target (denoted as \({c}_{{\rm {ER}}}^{{\rm {S}}}\) and \({c}_{{\rm {ER}}}^{{\rm {T}}}\) for the sender and the target, respectively) but benefits the target politically by functioning as deterrence against the sender’s sanctions in the future (denoted as \({b}_{{\rm {PR}}}^{{\rm {T}}}\)). If the target decides not to retaliate, the payoffs for the sender and the target at the end of the game are \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {NR}}}^{{\rm {S}}}={b}_{{\rm {E}}}^{{\rm {S}}}-{c}_{{\rm {E}}}^{{\rm {S}}}+{b}_{{\rm {PI}}}^{{\rm {S}}}\) and \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {NR}}}^{{\rm {T}}}={b}_{{\rm {E}}}^{{\rm {T}}}-{c}_{{\rm {E}}}^{{\rm {T}}}\), respectively. Otherwise, the payoffs for the sender and the target at the end of the game are \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {R}}}^{{\rm {S}}}={b}_{{\rm {E}}}^{{\rm {S}}}-{c}_{{\rm {E}}}^{{\rm {S}}}+{b}_{{\rm {PI}}}^{{\rm {S}}}-{c}_{{\rm {ER}}}^{{\rm {S}}}\) and \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {R}}}^{{\rm {T}}}={b}_{{\rm {E}}}^{{\rm {T}}}-{c}_{{\rm {E}}}^{{\rm {T}}}-{c}_{{\rm {ER}}}^{{\rm {T}}}+{b}_{{\rm {PR}}}^{{\rm {T}}}\), respectively. We solve the game by backward induction (Fudenberg and Tirole, 1991). See Supplementary Material for details.

Quantitative trade models

We quantify the economic impact of sanctions and counter-sanction measures based on two quantitative trade models, in which the sender and the target are connected by trade linkages in the global supply chain network: an adaptive multi-regional input-output model (Guan et al., 2020; Otto et al., 2017; Wenz and Levermann, 2016) to characterize the short-term effects of trade restrictions and a general equilibrium model with input–output linkages (Caliendo and Parro, 2015; Eppinger et al., 2021) to characterize the long-term effects of decoupling. The adaptive multi-regional input-output model is demand-driven, wherein firms adjust production based on client demand, and no new trade linkages can be established in the short term. Under these assumptions, import restrictions are more likely to be effective than export restrictions: if export restrictions are implemented, the imposing country will lose the client (i.e., the affected country) and cannot build new trade linkages to boost sales in the short term, resulting in significant economic loss; while if import restrictions are adopted, the imposing country can still get supplies from alternative sources based on existing trade linkages, thus mitigating the potential economic loss (see Supplementary Material for details). Hence, we take import restrictions as measures of sanctions/retaliation.

We construct the global supply chain network using data from the latest version of the Global Trade Analysis Project database (GTAP 10) (Aguiar et al., 2019). GTAP 10 provides an MRIO table representing the annual monetary transactions between 65 sectors of 141 countries. For each country, we treat each sector as a representative firm and the entire final demand system as a representative household. A firm is represented by the index pair ai, where a denotes the products produced by the firm and i denotes the region in which the firm is located. Following (Joshi and Mahmud, 2016, 2020), we denote the global supply chain network as \({\mathcal{G}}=({g}_{ai\to bj}| ai,bj\in N)\), where N is the set of firms in the global supply chain network. g_ai→bj = 1 if there are trade flows from firm ai to firm bj, otherwise g_ai→bj = 0. If the sender (represented by the index S) cuts off imports of products in sector r from the target (represented by the index T), the trade linkages, denoted by g_rT→S = (g_rT→aS∣g_rT→aS = 1), will be deleted from the global supply chain network. The incurred economic cost of trade sanctions and counter-sanction measures for a country is measured using the hypothetical extraction method (Los et al., 2016), which compares the country’s actual gross domestic product (GDP) with its GDP under a hypothetical scenario where some trade flows are set to zero.

Denote \({\pi }^{{\rm {S}}}({\mathcal{G}})\) and \({\pi }^{{\rm {T}}}({\mathcal{G}})\) as the pre-sanction GDP for the sender and the target, and \({\pi }^{{\rm {S}}}({\mathcal{G}}-{g}_{r{\rm {T}}\to {\rm {S}}})\) and \({\pi }^{{\rm {T}}}({\mathcal{G}}-{g}_{r{\rm {T}}\to {\rm {S}}})\) as the GDP for the sender and the target after sanctions, then, we define

where

are the reduction in GDP relative to pre-sanction levels after the imposition of sanctions for the sender and the target, respectively. Denote the trade linkages deleted under retaliation imposed by the target T against the sender S as \({g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R}}}\in \{{g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,D}}},{g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,I}}},{g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,C}}}\}\). \({g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,D}}}\), \({g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,I}}}\), and \({g}_{{\rm {S}}\to {\rm {T}}}^{{\rm {R,C}}}\) are selected under the decoupling, individual short-term retaliation, and collective short-term retaliation strategies, respectively. Then, we denote

where

are the reduction in GDP relative to pre-sanction levels after the imposition of sanctions and retaliation for the sender and the target, respectively. See model details in the Supplementary Material.

Capability of and vulnerability to economic sanctions

We define country i’s pre-sanction trade-to-GDP (PT) ratio with country j in sector r as

where \({Z}_{ri\to uj}^{* }\) is the trade flow from sector r in country i to sector u in country j before sanctions. Country i’s largest pre-sanction trade-to-GDP (LPT) ratio with country j is defined as \({\rm {LP{T}}}_{i\to j}={\max }_{r}{\rm {P{T}}}_{ri\to j}\). Then, country i’s capability of and vulnerability to economic sanctions are defined as

and

respectively. Here, \({\mathcal{N}}\) is the set of countries/regions on the global supply chain network. Trade partners with which country i’s LPT ratios are larger than \(\hat{{\rm {LPT}}}\) are defined as important trade partners. We set \(\hat{{\rm {LPT}}}\) as 5% in the main text. Results based on other values of \(\hat{{\rm {LPT}}}\) can be found in the Supplementary Material.

Effectiveness of sanctions and counter-sanction measures

We define the effectiveness of trade sanctions imposed by the sender S on the product r from the target T as the probability that sanctions succeed in supporting an arbitrary political goal. Economic sanctions are credited with success if the target complies, i.e., the “comply-not sanction” equilibrium is achieved. To quantify political benefits and costs, we define

Here, ρ^S, ρ^T, ξ^S, and ξ^T are the ratios between political benefits/costs and the pre-sanction GDP. We assume that “comply-not sanction” is the optimal outcome for the sender, i.e., the sender can obtain the optimal payoff if she can achieve her political goals without resorting to punishment. Thus, \({{\mathcal{P}}}_{{\rm {C}}}^{{\rm {S}}}=\max \{{{\mathcal{P}}}_{{\rm {C}}}^{{\rm {S}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {NS}}}^{{\rm {S}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {NR}}}^{{\rm {S}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {R}}}^{{\rm {S}}}\}\). Then, we define

where q^S > 1 is a constant value. The larger q^S is, the less political benefit the sender can obtain by maintaining a reputation for toughness. Similarly, we assume that “violate-not sanction” is the optimal outcome for the target. Thus, \({{\mathcal{P}}}_{{\rm {V}}\to {\rm {NS}}}^{{\rm {S}}}=\max \{{{\mathcal{P}}}_{{\rm {C}}}^{{\rm {T}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {NS}}}^{{\rm {T}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {NR}}}^{{\rm {T}}},{{\mathcal{P}}}_{{\rm {V}}\to {\rm {S}}\to {\rm {R}}}^{{\rm {T}}}\}\). Then, we define

where q^T > 1 is a constant value. The larger q^T is, the less political benefit the target can obtain through retaliation.

Assuming that ρ^S ∈ (0, l^S] and ρ^T ∈ (0, l^T] are independent and follow uniform distribution. Then, the effectiveness of sanctions without and with the threat of retaliation are

and

respectively. We define the effectiveness of retaliation (\({\vartheta }_{{\rm {S,T}},r}^{{\rm {R}}}\)) as the reduction in the effectiveness of sanctions under the threat of retaliation, i.e.,

Detailed analysis is shown in the Supplementary Material.

Feasibility of collective short-term retaliation

In the collective short-term retaliation strategy, the target can come to an agreement with other countries to build a bloc against the sender’s sanctions. If the sender imposes economic sanctions on any country in the bloc, all countries in the bloc will retaliate by imposing economic sanctions on the sender. Assuming all countries in the potential bloc will adopt the same strategy, then, for each country c in the bloc B, the expected payoff of joining the bloc against the sender S is

where \({{\mathcal{E}}}_{{\rm {S,c}},r}^{{\rm {c,R}}}({g}_{{\rm {S}}\to {\rm {c}}}^{{\rm {R}}}={g}_{{\rm {S}}\to {\rm {c}}}^{{\rm {R,C}}})\) is the expected payoff for the country c as the target under the collective short-term retaliation strategy and \({\pi }^{c}({\mathcal{G}}-{g}_{{r}^{{\prime} }{c}^{{\prime} }\to S}-{g}_{S\to {c}^{{\prime} }}^{R,C})\) is country c’s GDP when country \({c}^{{\prime} }\) is targeted and adopts the collective retaliation strategy. Similarly, the payoff of not joining the bloc (denoted by \({E}_{{\rm {c}}}^{{\rm {NJ}}}\)) is

and

if country c decides not to retaliate, to adopt the individual short-term retaliation strategy, and to adopt the decoupling strategy, respectively. Here, \({{\mathcal{E}}}_{{\rm {S,c}},r}^{{\rm {c}}}\) is the expected payoff for country c as the target if it decides not to retaliate. The payoff gain is

Country c is willing to join the bloc if and only if ΔE_c > 0. The collective short-term retaliation strategy is feasible only if all countries in the bloc are willing to join the bloc. Detailed analysis is shown in the Supplementary Material.

Asymmetry between the capability of imposing and vulnerability to economic sanctions

For simplicity, we assume that, except for decoupling, only one sector will be affected if one country (the imposing country) imposes import restrictions on another (the affected country), and the imports of all products from this sector will be cut off. At the sanction stage, the sender is the imposing country, and the target is the affected country. At the counter-sanction stage, the target is the imposing country, and the sender is the affected country. The economic impact of import restrictions varies among sectors. Figure 2 presents the GDP loss for the imposing and affected countries during and after the restriction period (Fig. 2a–h) and the effectiveness of import restrictions (Fig. 2i–l) when five sectors, in which the affected country’s PT ratio with the imposing country are highest, are affected by import restrictions. Here, an affected country’s PT ratio with the imposing country in a sector represents the annual trade amount between these two countries in this sector as a share of the affected country’s annual GDP before sanctions (“Data and methods”). We define the effectiveness of import restrictions as the effectiveness of sanctions taking the form of import restrictions issued by the imposing country without the threat of retaliation by the affected country (see “Data and methods” for details). Generally, regardless of the size of the imposing and affected countries, import restrictions are more effective when hitting sectors in which the affected country’s PT ratios with the imposing country are higher (Fig. 2i–l). Therefore, we assume that the imposing country will always adopt import restrictions on the sector in which the affected country’s PT ratio with the imposing country is highest to maximize the effectiveness of sanctions/retaliation.

The affected country’s LPT ratio with the imposing country can be regarded as an indicator of the importance of the imposing country as a trade partner of the affected country. Countries occupying important positions in global trade are more able to sanction others, while countries heavily relying on foreign markets are more vulnerable to sanctions. We thus define each country’s capability of and vulnerability to economic sanctions based on the LPT ratios between countries (see “Data and methods”) to identify the primary potential senders and targets of economic sanctions globally. Figure 3 shows each country’s pre-sanction GDP, capability of, and vulnerability to economic sanctions. We observe a marked asymmetry between each country’s capability of and vulnerability to economic sanctions (Fig. 3d). The capability of sanctions among countries follows a Pareto distribution (Fig. 3b), where a small group of large economies are much more capable of sanctioning others, while the vulnerability to sanctions follows a more uniform distribution (Fig. 3c), where small economies are slightly more vulnerable. These results indicate that, although all countries face similar degrees of vulnerability to economic sanctions due to the high dependency on global trade, giant economies dominating global trade are the primary potential senders of economic sanctions.

Effectiveness of counter-sanction measures

We evaluate the effectiveness of different counter-sanction measures under varying geopolitical scenarios, with a particular focus on two representative scenarios where sanctions are initiated by highly capable countries. We name these two representative scenarios as high bilateral interdependence (such as the China–US) and high unilateral interdependence scenarios (such as China-Indonesia; see examples in Fig. 4a, b). Under the high bilateral interdependence scenarios, the sender’s LPT ratio with the target and the target’s LPT ratio with the sender are both high, indicating that both countries heavily rely on each other’s trade markets. Under the high unilateral interdependence scenarios, the target’s LPT ratio with the sender is high, while the sender’s LPT ratio with the target is low, indicating an imbalanced trade relationship between the sender and the target, with the sender taking a significantly more important position on the target’s trade market than vice versa.

We provide two examples for each representative scenario. For the high bilateral interdependence scenario, the sender–target pairs are China–US (Fig. 4c, e) and Germany–France (Fig. 4d and f). For the high unilateral interdependence scenario, the pairs are China–Indonesia (Fig. 4g, i) and the US–Singapore (Fig. 4h, j). To maximize the impact of collective retaliation, countries with which the sender’s LPT ratios are highest among all trading partners are chosen as the target’s bloc members. We set the number of countries in a bloc as three in the main text; see other settings in the Supplementary Material. Figure 4c–j shows the sender’s and target’s GDP change during and after the restriction period and the effectiveness of retaliatory measures under the base (i.e., no), individual short-term, and collective short-term retaliation strategies. We define the effectiveness of a retaliatory measure as the reduction in the effectiveness of sanctions under the threat of retaliation (see “Data and methods” for details). The duration of short-term import restrictions is set as one year. We also present the sender’s and target’s annual GDP loss caused by decoupling for comparison.

Unlike short-term retaliation, where involved countries’ GDP levels will return to pre-sanction levels shortly after the restriction period, decoupling will come at long-term GDP loss for both the sender and the target. The GDP loss caused by decoupling can exceed that by short-term retaliatory measures in less than a year. Such heavy cost makes complete decoupling impractical. Compared to the high unilateral interdependence scenario, individual short-term retaliation is far more effective under the high bilateral interdependence scenario, indicating that deeper economic integration among countries makes it more difficult to weaponize global trade. Owing to the joint efforts of bloc members, collective short-term retaliation results in an additional reduction in the effectiveness of sanctions. Such reduction is less significant under the high bilateral interdependence scenario than the high unilateral interdependence scenario since the target itself is the major contributor to the collective efforts, with limited additional support from other bloc members.

Formally, the following propositions summarized the effectiveness of counter-sanction measures under varying scenarios.

Proposition 1

Decoupling is an impractical counter-sanction measure.

Proposition 2

Short-term individual retaliation strategies are more effective under high bilateral interdependence than high unilateral interdependence scenarios.

Proposition 3

The incremental increase in the effectiveness of retaliation contributed by bloc formation is less significant under the high bilateral interdependence than high unilateral interdependence scenarios.

Proof of Propositions 2 and 3 can be found in the Supplementary Material.

Dilemmas in achieving both the effective and feasible collective retaliation

Despite the improvement in the effectiveness of retaliation due to the support from bloc members, collective retaliation is impracticable because the underlying feasibility conditions cannot be met. Bloc formation is feasible if and only if all bloc members are willing to join the bloc, i.e., the payoff gain of joining the bloc versus not joining the bloc is positive.

Figure 5 shows the expected payoff of countries in the bloc under different retaliation strategies for each example in Fig. 4. Under the high bilateral interdependence scenario (Fig. 5a–f), the payoff gain of joining the bloc relative versus retaliating alone or not retaliating is positive for all bloc members except for the target whose payoff of joining the bloc is higher than not retaliating but lower than retaliating alone. This indicates that the target, with which the sender’s LPT ratio is relatively high among all countries in the bloc, is unwilling to join the bloc. Under the high unilateral interdependence scenario (Fig. 5i–p), targets achieve much higher payoffs by joining the bloc, and the effectiveness of sanctions is significantly reduced under the threat of collective resilience, whereas the major contributors in the bloc (e.g., the US in China-Indonesia game and Canada in the US–Singapore game) get negative payoff gains by adopting the collective retaliation strategy. These results indicate that there is a conflict between the effectiveness and feasibility of bloc formation. Major contributors in the bloc, game changers in the collective sanction, are unwilling to join the bloc because their deep economic interdependence with the sender effectively resists potential sanctions, making bloc formation an unnecessary option.

The increasing deployment of economic sanctions calls for effective counter-sanction measures to safeguard economic and political interests. Despite efforts to develop and assess the impact of potential countermeasures, data-driven understanding and evidence of their effectiveness and feasibility are lacking. In this study, we propose a game-theoretical framework to model the interactions between countries involved in economic sanctions as a three-stage sequential game and to simulate the economic impact of sanctions based on real-world international trade data. We mathematically characterize the equilibrium paths of the sanction game and investigate the role of countermeasures in sanction issues. Our framework allows us to conduct both analytical and numerical investigations to assess the global distribution of capability and vulnerability to economic sanctions and to examine the effectiveness of different countermeasures under varying geopolitical scenarios.

Although the growing economic interdependence among countries increases the risk of exposure to sanctions for all countries, only a few giant economies have the practical power to weaponize economic interdependence. Analysis of the effectiveness of different counter-sanction measures against these highly capable countries shows that a highly interdependent and diversified international trade system makes it difficult to weaponize trade. The major source of sanction risk for vulnerable economies is the imbalanced trade relationships with giant economies, under which collective retaliation has the potential to be significantly effective. However, the high effectiveness of collective retaliation largely comes from the contribution of bloc members having deep economic integration with the sender and bearing losses by joining the bloc. Such difficulty in reconciling the effectiveness and feasibility of collective retaliation makes it impractical to combat economic sanctions.

Our study has limitations. First, the effectiveness of sanctions and counter-sanction measures is associated with a sophisticated combination of economic, political, social, and historical factors (Dashti-Gibson et al., 1997). The proposed game theoretical framework captures the core aspects of economic sanctions but overlooks the influence of various political and cultural factors, such as historical grievances, national pride, and the stability of governments to withstand economic and diplomatic pressure. While these elements significantly shape the feasibility and effectiveness of sanctions in the real world, their quantification is challenging due to the lack of data. Addressing these challenges can enhance the model’s capacity to elucidate complex real-world cases. Second, decision-makers’ personal political and economic considerations, including the potential impact of sanctions on their popularity, electoral prospects, and personal financial interests tied to the target country, wield substantial influence over their stance on economic sanctions and their decisions regarding the imposition, continuation, or removal of sanctions. Thus, integrating decision-makers’ payoffs into the game-theoretical framework represents a crucial avenue for future research. Third, economic sanctions take various forms with different intensities and duration (Hufbauer et al., 1990; Morgan et al., 2014). Our analysis and conclusions focus primarily on trade sanctions, a common form of economic sanctions targeting countries, based on hypothetical sanction scenarios. Our findings may not be generalized to other forms of economic sanctions, especially those targeting individuals or groups. However, the proposed game-theoretical framework can be flexibly adapted to investigate diverse forms of economic sanctions, such as capital flow restrictions and arms embargoes, by quantifying the economic costs associated with sanctions and retaliatory measures. Finally, we assume a static structure of global supply chain networks under short-term sanctions and counter-sanction measures, i.e., countries involved in the sanction game cannot diversify their trade to alternative markets to mitigate the impact of sanctions or retaliation in the short study period. This simplification needs to be amended when modelling sanctions or retaliation that last years. Investigating the interplay between the interaction of countries’ policies and the evolution of trade networks is an interesting topic for future research.

In conclusion, our data-driven analysis suggests that engagement and cooperation are better strategies than offensive punishment and isolation in dealing with geopolitical conflicts, as seen in many real-world cases. For example, in 2001, the US lifted economic sanctions imposed on India in response to the 1998 nuclear tests, thereby fortifying their defense and counter-terrorism partnership through strategic dialogue and collaboration. This waiver also facilitated the resumption of trade and investment between India and the US, opening up avenues for economic engagement and cooperation. The 2019–2023 Japan–South Korea trade dispute also serves as a compelling case study illustrating the limitations of economic sanctions. This multifaceted conflict, rooted in historical disagreements, led to a significant deterioration in diplomatic relations between the two countries and disrupted global supply chains for the electronics industries. Despite the imposition of trade sanctions, the dispute persisted, highlighting the ineffectiveness of such measures in addressing international conflicts. Ultimately, both Japan and South Korea reinstated each other as preferred trading partners to rebuild their security and economic ties, underscoring the necessity of dialogue and cooperation in easing tensions and fostering relations.

Despite being deemed ineffective in our model, economic sanctions remain widely utilized in the real world, often guided by decision-making processes informed by past experiences. The pervasive use of these seemingly ineffective tools underscores a critical evidence gap in policy-making, emphasizing the need for more data-driven approaches to inform and refine international strategies. We recommend countries pursue collaboration, cooperation, dialogue, and diplomacy to address differences and prompt mutual understanding, trust, and global peace and stability. We hope that our studies can offer useful insights and guidance for policymakers and stakeholders who are involved in or affected by trade interdependence and economic sanctions. We also hope that our studies can inspire more research on the economic and political implications of these complex and dynamic phenomena.

The simulation code can be accessed at https://github.com/jianan0099/sanctions.

Author notes

1. These authors contributed equally: Yang Ye, Qingpeng Zhang.

Authors and Affiliations

1. Center for Infectious Disease Modeling and Analysis, Yale School of Public Health, Yale University, New Haven, CT, USA

   Yang Ye

2. Musketeers Foundation Institute of Data Science, The University of Hong Kong, Hong Kong SAR, China

   Qingpeng Zhang

3. Department of Pharmacology and Pharmacy, LKS Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China

   Qingpeng Zhang

Contributions

YY: conceptualization, data curation, formal analysis, methodology, visualization, writing—original draft, writing—review and editing; QZ: conceptualization, funding acquisition, investigation, methodology, project administration, supervision, writing—original draft, writing—review and editing.

Corresponding author

Correspondence to Qingpeng Zhang.

Competing interests

The authors declare no competing interests.

Ethical approval

Ethical approval was not required as the study did not involve human participants.

Informed consent

This article does not contain any studies with human participants performed by any of the authors.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions