The replication data, code, and appendix for this article are available in the electronic supplementary materials on the Review of International Organizations’ webpage.

1. For example, many UN treaty-bodies receive individual petitions from states that have ratified optional protocols. Schoner (2023) shows that repressive regimes have incentives to commit to these optional protocols when faced with international incentives and few domestic institutional constraints and Schoner (2024) shows that individual petitions can improve human rights behavior when publicized by civil society organizations. This important research suggests that even repressive regimes receive reparations orders from international bodies (including orders to pay monetary damages) and repressive states have incentives to change human rights behavior following violation decisions by international bodies.

2. See Salvador Chiriboga v. Ecuador, Partially Dissenting Vote of Judge Leonardo A. Franco. Judgment on Reparations and Costs, Inter-Am.Ct.H.R. (March 3, 2011), para 9.

3. Altwicker-Hámori et al. (2016) make a similar argument with respect to application of the equity principle in the context of the European Court of Human Rights (ECtHR).

4. Although our focus in this paper is on monetary damages, we also know little about judicial decision-making when it comes to the package of remedies that are ordered in adverse judgments and whether some remedies are considered substitutes for one another. For example, other remedial orders (e.g. creating a scholarship or memorial) can also be costly, and the extent to which Inter-American Court judges balance the costs of these orders with monetary damages awarded to victims is unknown. This question is beyond the scope of this paper, but represents an avenue for future research.

5. Indeed, international law does not recognize punishment as a valid objective of reparations. See Johns (2022), pgs. 120–123. Punitive damages are also not part of the case law of the Inter-American Court of Human Rights.

6. Because this form of damage is rather objective in its determination, usually involving sums easily quantified by receipts, our focus will be on the awarding of pecuniary and non-pecuniary damages.

7. Of note, the value of damages awarded in the ECtHR are generally smaller than those awarded by the IACtHR and the AfCtHR has examined fewer cases than the other regional bodies and has awarded damages in a relatively small number of cases.

8. The ICJ ordered compensation for mass atrocities for the first time in a 2002 case, Armed Activities on the Territory of the Congo.

9. Even so, this principle itself is not always well understood (Keller et al., 2022).

10. Lost earnings refer to loss of economic income, both past and future, and are calculated based on income the victim would have received up to the time of their possible natural death. The Court is careful to take into account the specifics of the victim’s life, including age, occupation, and education, in calculating the presumed lifetime income. Consequential damages which typically refer to conventional costs associated with human rights violations, such as funeral expenses, medical bills, or the costs of searching for victims who have disappeared.

11. See Gelman v Uruguay, Merits and Reparations. Judgment, Inter-Am.Ct.H.R. (Feb 24, 2011), para 98.

12. See Al-Skeini and Others v. UK. Judgment, ECt.H.R. (July 7, 2011), para 182.

13. María Claudia García’s potential earnings are much higher than average, as she was only 19 years old when she disappeared and was studying at the university.

14. See Rio Negro Massacres v Guatemala and Ituango Massacres v Colombia, for example.

15. Quoted in Schwöbel-Patel (2018), pg. 710.

16. See Rosendo Cantú v. Mexico, Preliminary Objections, Merits, Reparations and Costs. Judgment, Inter-Am.Ct.H.R. (Aug 31, 2010), para 115.

17. See Rosendo Cantú v. Mexico, Preliminary Objections, Merits, Reparations and Costs. Judgment, Inter-Am.Ct.H.R. (Aug 31, 2010), para 18.

18. Not all marginalized groups represent “ideal” victims. For example, although indigenous victims come from marginalized groups, Vandermaas et al. (2024) note that women and children are more likely to be identified as victims (weak, vulnerable, and passive) than men because “the constructions of masculinity do not allow space for vulnerability and passivity” (175). Vandermaas et al. (2024) also note that “claims to intergenerational victimhood – that underpin demands for restitution of indigenous rights or reparations to descendants of past repression – tend to be more socially controversial and resisted than do claims to ‘direct victimhood’” (175). As a result, it is unclear the amount of public sympathy that exists in society for indigenous victims. Future research should assess the impact of indigenous group identity on the value of monetary awards.

19. This includes awards from states that withdrew from the Court during this time period, like Trinidad and Tobago and Venezuela, as the cases occurred prior to their withdrawal.

20. The selection process into reports at the Commission is beyond the scope of this paper; see Parente (2022).

21. The cases excluded by a cutoff of 25 include massacres and cases involving indigenous populations in which an entire indigenous population is considered the victim of a human rights violation. We also choose this threshold because many datasets on political violence consider 25 deaths in a single conflict to be indicative of a mass killing, which is fundamentally related to, yet different from, human rights violations.

22. The IACtHR usually orders judgments in USD. For the few cases that are done in a currency other than USD, we use the average exchange rate of the year of the judgment to convert the award into USD. Inflation adjustment is done by the authors.

23. If the Court does not clearly distinguish between pecuniary and non-pecuniary damage, we divide the total damages equally between the two categories. Results are also robust to excluding these 15 cases altogether.

24. As discussed below, we model awards to individual victims to test hypotheses 3 and 4, in which “total victims” is a control variable.

25. As noted above, the total award in any single case has increased by about $24,000 each year, so if the amount per victim is staying largely constant, the number of victims being recognized in a single case has increased over time. In part, this is driven by the increase in cases relating to massacres and indigenous groups (where an entire indigenous community is found to be the victim), but it could also be the Court expanding its definition of what counts as an indirect victim.

26. For each case, we calculated an average award per victim based on the total number of victims in the case and the total damages awarded. The entries in this table correspond to the average of this value, across all cases from a particular state.

27. Both cases involved petitioners on the death row; the Court awarded court costs only.

28. The Inter-American Court is clear that it designates someone as a minor if they are below the age of 18, following the Convention on the Rights of the Child.

29. We model direct and indirect victim awards separately in the appendix.

30. In cases where the merits judgment is separate from the reparations and costs judgment, this reflects the difference between year of the violation and the year of the merits judgment.

31. This is the “NHRI type” variable from Conrad et al. (2013), which is nominal. The two types of NHRIs coded are ombudsmen and human rights commission.

32. See Myrna Mack Chang v. Guatemala, Reasoned and Partially Dissenting Opinion of Judge Arturo Martínez Gálvez. Judgment on Merits, Reparations, and Costs, Inter-Am.Ct.H.R. (Nov 25, 2003), para IV.

33. Nevertheless, while we do not think this is the primary consideration for state capacity, we do include models with this measure as robustness checks in the appendix.

34. States frequently labeled dissidents as terrorists, which is why we include them in this category.

35. Because the value of damages is logged, we interpreted this finding by exponentiating the coefficient, subtracting one and multiplying by 100.

36. See Aloeboetoe v. Suriname, Reparations and Costs. Judgment, Inter-Am.Ct.H.R. (Sept 10, 1993), para 91. The Commission ruled that all seven victims suffered Article 4 and 5 violations, even though the seventh survived. Suriname accepted responsibility and the Court did not issue a separate ruling on the merits.

37. Both the dependent and independent variables are logged in the models. Values are calculated as (1.5^−0.25−1) ∗ 100.

38. In Ecuador’s case, this is driven by two cases in particular in which a single victim was awarded one million and nearly 29 million USD, respectively. In Barbados’ case, no damages were ever awarded.

39. See Plan de Sanchez Massacre v. Guatemala, Separate Opinion of Judge Sergio García Ramírez. Judgment on Reparations, Inter-Am.Ct.H.R. (Nov 19, 2004), para 27.

40. Ibid., para. 28.

41. Interestingly, the coefficient for pecuniary damages is statistically significant and negative, which is the opposite of what we expected. However, we believe this is driven by the fact that indirect minor victims are highly unlikely to receive anything in pecuniary damages, as they are not the ones responsible for paying consequential damages like funeral or search costs.

42. Myrna Mack Chang v. Guatemala, Merits, Reparations and Costs. Judgment, Inter-Am.Ct.H.R (Nov 25, 2003), para 264(b).

43. See Mohamed v. Argentina, Preliminary Objection, Merits, Reparations and Costs. Judgment, Inter-Am.Ct.H.R. (Nov 23, 2012), para 164.

44. Ibid., at para 171.

45. The exception is the coefficient on Total victims, which just misses the cutoff for statistical significance at .

For their helpful comments, we thank Gino Pauselli, Wayne Sandholtz, Beth Simmons, Øyvind Stiansen, and Justin Wedeking. We also thank participants in the 2022 Annual Meeting of the International Studies Association, the 2022 Annual Meeting of the Peace Science Society (International), and the 2023 Annual Meeting of the American Political Science Association. We are also grateful for the helpful feedback received at the Niehaus Center for Globalization and Governance in March 2023 and various renditions of the Law and Courts Junior Women’s Writing Group. Katelyn Bennett, Mara Bravin, Alexandria Brooks, Louise Byrne, Sam Ellyson, Alexandra Fiul, Maggie Flanagan, Hunter Lang, and Ann Silva all provided excellent research assistance.

Research assistants were paid by Christopher Newport University’s Reiff Center for Human Rights and Conflict Resolution. We also received research assistant support from University of Kentucky’s Political Science Department and we gratefully acknowledge their support.

Authors and Affiliations

1. Department of Political Science, University of Kentucky, Lexington, KY, USA

   Jillienne Haglund

2. Department of Political Science, Christopher Newport University, Newport News, VA, USA

   Francesca Parente

Contributions

Author contributions to research design and conceptualization: J.H. (50%), F.P. (50%); statistical analysis: J.H. (40%), F.P. (60%); writing: J.H. (60%), F.P. (40%). The order of the authors is chosen alphabetically.

Corresponding author

Correspondence to Jillienne Haglund.

Conflicts of interest

The authors declare they have no conflict of interest or competing interests in relation to this article.

Responsible editor: Axel Dreher