How Do Non-UN Peacekeepers Affect Civilian Violence? An Instrument Investigation

ABSTRACT

This research note extends the Bara and Hultman (2020) study on the effectiveness of non-UN peacekeeping missions in terms of curbing one-sided violence (OSV) against civilians. In particular, we employ two novel instruments to address the two-way causality between the number of non-UN peacekeepers and OSV measures. For each panel year, our instruments involve the interaction between the sum of various designated peacekeepers contributed and the inverse distance between the capitals of contributor and conflict countries. As required, the instrument satisfies the necessary inclusion and exclusion (exogeneity) requirements. The instrument-based results establish a robust reduction in government OSV stemming from the number of non-UN peacekeepers deployed. That reduction also holds for propensity-score matching and the inclusion of UN peacekeepers in the same regression. Non-UN peacekeepers did not have a robust influence on rebel OSV.

KEYWORDS:

• Instrument
• one-sided violence
• non-UN peacekeeping operations
• UN peacekeeping operations
• propensity-score matching

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Data Availability Statement

Replication material available at: https://personal.utdallas.edu/~tms063000/website/downloads.html.

Notes

1 Sandler, “International Peacekeeping Operations.”

2 E.g. Doyle and Sambanis, “International Peacebuilding”; Hultman, Kathman, and Shannon, “United Nations Peacekeeping”; Hultman, Kathman, and Shannon, “Beyond Keeping Peace”; Kim, Sandler, and Shimizu, “A Multi-transitional Approach”; Walter, Howard, and Fortna, “The Extraordinary Relationship.”

3 E.g. Fortna, “Does Peacekeeping Keep Peace?”; Sambanis and Schulhofer-Wohl, “Evaluating Multilateral Interventions in Civil Wars.”

4 Walter, Howard, and Fortna, “The Extraordinary Relationship.”

5 Bara and Hultman, “Just Different Hats?”; Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

6 Sandler, “International Peacekeeping Operations,” 1882.

7 Bara and Hultman, “Just Different Hats?”; Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

8 Bara and Hultman, “Just Different Hats?,” 341.

9 Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

10 Bara and Hultman, “Just Different Hats?”; Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

11 On reverse causality bias, see Basu, Asymptotic Bias of OLS in the Presence of Reverse Causality.

12 On propensity-score matching, see Angrist and Pischke, Mostly Harmless Econometrics; Wooldridge, Econometric Analysis, 920. On differences between propensity score matching and the use of instrumental variables, see Ichimura and Taber, “Propensity-Score Matching with Instrumental Variable.”

13 Wooldridge, Econometric Analysis.

14 See surveys by Sandler, “International Peacekeeping Operations”; Walter, Howard, and Fortna, “The Extraordinary Relationship.”

15 E.g. De Rouen and Chowdhury, “Mediation, Peacekeeping and Civil War”; Kim, Sandler, and Shimizu, “A Multi-Transitional Approach.”

16 Doyle and Sambanis, “International Peacebuilding”; Doyle and Sambanis, Making War and Building Peace.

17 Fortna, Does Peacekeeping Work?; Hegre, Hultman, and Nygård, “Evaluating the Conflict-Reducing Effect.”

18 E.g. Haass and Ansorg, “Better Peacekeepers”; Hultman, “Keeping Peace or Spurring Violence?”; Kathman and Wood, “Stopping the Killing.”

19 Fortna, “Does Peacekeeping Keep Peace?”; Gilligan and Sergenti, “Does UN Intervention Cause Peace?”

20 Fjelde, Hultman, and Nilsson, “Protection Through Presence.”

21 Ibid., Ruggeri, Dorussen, and Gizelis, “Winning the Peace Locally”; Ijaz, “The Selection Problem.”

22 Fortna, “Does Peacekeeping Keep Peace?”

23 Sambanis and Schulhofer-Wohl, “Evaluating Multilateral Interventions in Civil Wars,” 254.

24 Bara and Hultman, “Just Different Hats?”

25 Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

26 Heldt, “UN-Led or Non-UN-Led Peacekeeping.”

27 Kathman and Benson, “Cut Short?”

28 Ibid., 1609.

29 For similar arguments see Fjelde, Hultman, and Nilsson, “Protection Through Presence”; Hultman, Kathman, and Shannon, “United Nations Peacekeeping”; Kim, Sandler, and Shimizu, “A Multi-Transitional Approach.”

30 Regional missions refer to a regional organization (e.g. the EU) dispatching peacekeepers within their immediate region, while international missions refer to a regional organization deploying peacekeepers outside its immediate region (e.g. NATO peacekeeping in Afghanistan).

31 See Bara and Hultman, “Just Different Hats?”; Kim and Sandler, “Non-UN Peacekeeping Effectiveness.”

32 See, especially, Bellamy and Williams, “Who’s Keeping the Peace?” These authors emphasize that UN PKOs have greater legitimacy and experience than non-UN PKOs. However, Chapter VIII allowing for missions by regional organizations and experience gained by non-UN PKOs since 2000 partly counter their arguments.

33 Cameron and Trivedi, Microeconometrics Using STATA.

34 Gillian and Sergenti, “Do UN Intervention Cause Peace?”

35 E.g. Bara and Hultman, “Just Different Hats?”; Haass and Ansorg, “Better Peacekeeping, Better Protection?”

36 Adding Europe and Central Asia, East Asia and Pacific, Latin America, and South Asia results in multicollinearity problems.

37 Ijaz, “The Selection Problem.”

38 Fjelde, Hultman, and Nilsson, “Protection Through Presence”; Ruggeri, Dorussen, and Gizelis, “Winning the Peace Locally.” Unlike our exercise, both of those articles considered subnational UN conflict data.

39 Bun and Harrison, “OLS and IV Estimation”; Dreher and Langlotz, “Aid and Growth”; Dreher, Gassebner, and Schaudt, “The Effect of Migration on Terror”; Nunn and Qian, “US Food Aid and Civil Conflict.”

40 Based on an initial weight matrix, the two-step GMM initially estimates parameters, which are then employed to compute a new weight matrix, used to re-estimate the parameters. Ideally, the estimated GMM parameters with the IV makes the sample versions of population-moment conditions close to the true value – see StataCorp, Stata 16.

41 Cameron and Trivedi, Regression Analysis; Wooldridge, Econometric Analysis.

42 Sundberg and Melander, “Introducing the UCDP GED.”

43 The Nigerian conflict lasted for four months; however, its period is extended to 28 months to include 24 post-conflict months – see discussion in the text.

44 Bara and Hultman, “Just Different Hats?,” 357; Haass and Ansorg, “Better Peacekeeper, Better Protection.”

45 Högbladh, UCDP GED Codebook.

46 Kreutz, “How and When Armed Conflict End” and its updates indicate conflict end dates.

47 Arpino and Mealli, “The Specification of the Propensity Score.”

48 Ibid.

49 International Institute for Strategic Studies, The Military Balance.

50 Bara and Hultman, “Just Different Hats?”

51 International Peace Institute (IPI), “IPI Peacekeeping Data.”

52 Those inverse distances are from Gleditsch and Ward, “Measuring Space.”

53 Lagging OSV as a binary control follows Bara and Hultman, “Just Different Hats?” and Kim and Sandler, “Non-UN Peacekeeping Effectiveness.” A count OSV lagged control would raise endogeneity concerns.

54 Gleditsch et al., “Armed Conflict 1946–2001”; Pettersson and Öberg, “Organized Violence, 1989–2019.”

55 World Bank, “World Development Indicators.”

56 In the online appendix, Figure A1 displays the balancing of the covariate distributions after matching. The displayed box plots of the estimated propensity scores between the control (without non-UN deployment) and treated (with non-UN deployment) for the matched sample are similar in terms of the median, the 25th percentile, and 75th percentile. This is consistent with the covariates being balanced after matching. So, the mean bias reduces from 21.0 in the unmatched samples to 8.8% in the matched samples.

57 Bara and Hultman, “Just Different Hats?”

58 Fjelde, Hultman, and Nilsson, “Protection through Presence.”

59 Ibid.

60 Bara and Hultman, “Just Different Hats?”

61 Ijaz, “The Selection Problem.”

62 Ruggeri, Dorussen, and Gizelis, “Winning the Peace Locally.”

63 Bara and Hultman, “Just Different Hats?”

64 Ichimura and Taber, “Propensity-Score Matching with Instrumental Variable”; Wooldridge, Econometric Analysis, 802–9.

Additional information

Notes on contributors

Wukki Kim

Wukki Kim is an associate professor in the Department of Economics and Law at the Korea Military Academy, Seoul, Republic of Korea with a PhD from the University of Texas at Dallas. He holds the rank of Major in the Korean armed forces. His research interests are in defense economics, peacekeeping, and terrorism. His work has appeared in Journal of Conflict Resolution, Journal of Peace Research, European Journal of Political Economy, and Defence and Peace Economics.

Todd Sandler

Todd Sandler is Emeritus Chair in the Department of Economics at the University of Texas at Dallas. His research interests are in collective action, public economics, defense economics, peacekeeping, and terrorism. His work has appeared in American Political Science Review, American Journal of Political Science, American Economic Review, Economic Journal, International Organization, and Quarterly Journal of Economics.