Revisiting sanctions data: what can we learn?

Sanctions are one of the main tools used by the EU to uphold the rules-based international order and to respond to international crises. International sanctions are reshaping the global order and shifting centres of power. They have been applied in response to Russian aggression against Ukraine, rigged elections in Belarus, Iran’s programme for developing weapons of mass destruction, and the fight against international terrorism. The EU is expected to increase the use of sanctions, which is why studying this topic from multiple perspectives is important.

In the last year, I engaged in two important projects exploring how data can help clarify the nature, scope, and design of sanctions. The first research project focused on unpacking sanctions figures, and the second one addressed mapping sanctions circumvention schemes.

Why does sanctions data matter?

First, there is a large volume of academic literature on sanctions, particularly from legal, political science, and economic perspectives. However, very little existing literature examines sanctions from a data perspective.

Second, numbers and data matter because they provide another perspective on the discussed topic. Can we say whether sanctions are effective or not? Studies often rely on complex economic models, measuring trade parameters, GDP, access to technologies, and so on. But what if we evaluate effectiveness through the scope of sanctions and how broad they are – for instance, by asking whether placing ten or twenty designated individuals and companies under a cyber sanctions regime can meaningfully impact efforts to counter malware operations?

On the other hand, we have North Korea – the most heavily sanctioned country under UN Security Council resolutions. Yet, despite being under a sanctions siege, Pyongyang has orchestrated a vast network of hackers involved in stealing funds from cryptocurrency exchanges. While the scale of these cyber operations is enormous, the number of designated individuals and entities under cyber-sanctions regimes remains surprisingly low.

Another example is Russia, which has around 1,400 military factories and companies. Yet, less than half of them are subject to sanctions, despite significant efforts to raise the cost of aggression. This raises important questions about how comprehensive sanctions really are, and whether the current number of listings is adequate compared to the scope of the goals.

After Brexit, the UK appeared to adopt a more active and engaged role in applying sanctions, as outlined in its strategic policy document. However, what the numbers show is the opposite: surprisingly, the UK lags behind the EU’s sanctions mechanism, which is often considered to be more bureaucratic and slower than that of the UK.

Figure 1. Dynamics of sanctions imposition by the EU, US, and UK over a 10-year period

Source: Sanctions Finder

The UK implemented less than half as many personal sanctions in 2023, and even fewer in 2024, compared to the EU, despite having significantly more autonomous freedom to impose new listings.

Adjusting the lens through which we view sanctions becomes possible when we rely on data to measure impact, examining whether the numbers reflect the level of political commitment.

What sanctions can teach us?

In the paper ‘Un(packing) sanctions data,’ I assessed different aspects of sanctions-related numbers, specifically in the context of comparing approaches to sanctions policies in the EU and the US. For instance, I looked at the number of companies and individuals targeted by sanctions and the number of sectoral sanctions imposed, as well as comparing specific sanctions programmes (such as those targeting Russia) by examining how many sanctions were adopted by the US versus the EU. There is a common consensus that trade embargoes and comprehensive sanctions are less preferable than  ‘smart’ or ‘targeted’ sanctions.

I argue that the concept of ‘smart sanctions’ should be expanded beyond its individualised nature. Practitioners usually frame smart sanctions by applying certain restrictions to them, usually mandating that humanitarian exemptions, judicial review mechanisms, or periodic review processes be put in place to qualify sanctions as ‘smart.’

The argument of the paper is that we need to build a more reliable and extensive concept of smart (targeted) sanctions — one that is not only limited by specific parameters but also guided by certain policy considerations that, if put in place, will allow us to consider sanctions truly ‘smart.’ If so, should the frequency of sanction imposition or whether the pace of designations is accelerating or slowing down be viewed as criteria to evaluate how well sanctions are being delivered and how ‘smart’ they are? Can the critique of the disproportionate nature of sanctions against Russia be considered legitimate if, in reality, certain sanctions programmes (e.g. sanctions in relation to cyber-attacks, in response to Turkey’s unauthorised drilling activities, or the situation in Guatemala targeting individuals undermining democracy) list only a few high-level officials, well-documented perpetrators, or military factories? If such sanctions are indeed considered to be overly wide-ranging, what could then be a reasonable criterion for a sanction proportionality test? I leave these questions open in my paper. I consider such discussions to be important, and I believe that further development of these themes can shed light on whether it is necessary to address such issues before enacting new sanctions or refining existing frameworks.

Sanctions circumvention: lack of data make compliance challenging

In the second piece to which I had the privilege to contribute while at the EUI, I aimed to shed light on schemes employed to circumvent sanctions. Within the framework of the report, my co-authors and I identified nearly 100 cases connected to sanctions evasion schemes. This compilation was divided into three principal groups: violations of sectoral sanctions, violations of targeted sanctions, and compliance failures.

Figure 2. Distribution of analysed cases of sanctions violations

Source: ‘Restrictive measures: the challenges in front of the EU: report on schemes employed to circumvent sanctions,’ p.41.

As official data on sanctions enforcement across the EU is currently unavailable, this project represents the first attempt to map sanctions violations and group them into defined categories. Based on publicly available information, there are approximately 300 cases (including 100 episodes we reviewed) related to sanctions violations from 2022 to 2024 across the 27 EU member states.

Among these cases, violations of sectoral sanctions have been the most prevalent. The financial sector is particularly affected, as most investigated cases involve monetary or banking transactions, including the use of cryptocurrency transfers. Identifying the most vulnerable areas in sanctions compliance provides an opportunity to properly adjust and refine sanctions regulations.

Moreover, the year with the highest number of sanctions violations was 2022, following the imposition of what I call ‘sweeping sanctions’ against Russia and Belarus (i.e. sanctions that are more than just targeted or smart sanctions but do not reach the extent of a full embargo), with one third of the violations occurring during that time. This data may indicate two key dynamics: first, that evasion strategies evolve over time and become harder to detect; second, that businesses adapt quickly to new restrictions. Both hypotheses need to be examined, and only the establishment of a centralised, open database of all enforcement actions against sanctions violators will make definite conclusions reliable.

Concluding remarks

Open and publicly accessible sanctions data has the potential to immediately make sanctions enforcement and implementation more coherent and streamlined. Access to sanctions data could lead to the development of better policies and allow sanctions circumventions to be investigated more effectively, as gaps will be easier to map and detect.

Tags: sanctions, Data