Evaluating the conﬂict-reducing eﬀect
of UN peacekeeping operations∗
H˚avard Hegre1
, Lisa Hultman2
, and H˚avard Mokleiv Nyg˚ard1,3
1
Centre for the Study of Civil War, PRIO
2
Uppsala University
3
University of Oslo
June 11, 2013
Abstract
During the past two decades there has been a dramatic increase in both funds spent
and troops sent on peacekeeping operations (PKOs). At the same time, systematic research
on the eﬃcacy of PKOs to guide policy making is still scarce. We approach this
question by simulating the eﬀect of various possible UN peacekeeping policies. We base
the simulation on a statistical model that estimates the eﬃcacy of UN PKOs in preventing
the onset, escalation, continuation, and recurrence of internal armed conﬂict in the world
for the period 1970–2008. Apart from new data on UN PKO budgets and mandates, the
model includes the most important predictors of armed conﬂict. We use out-of-sample
validation of prediction performance to identify the best statistical model and to evaluate
its predictive performance. Predictions of how various PKO policies aﬀect future conﬂict
levels are then obtained through simulating the behavior of the conﬂict variable as implied
by the estimates from the statistical model, using projections of demographic and
education-related variables from the UN and the IIASA. Our results show that in a scenario
where the UN is willing to issue PKOs with strong mandates and increase its PKO
budget by 50 percent, the risk of armed conﬂict in the world in 2035 would be reduced by
up to two thirds relative to a scenario without PKOs. Considering the enormous costs of
armed conﬂict, in terms of both human suﬀering and foregone economic development, our
results suggest that UN peacekeeping is a cost-eﬀective way of increasing global security.
∗
The paper has been funded by the Research Council of Norway project 217995/V10. We thank Paul
Huth, Burcu Savun, Halvor Mehlum, Nikitas Konstantinidis, ..., and participants at several workshops and
conferences for their comments on previous versions of this paper, and Andreas Forø Tollefsen for help with
the map.
1
1 Introduction
Peacekeeping has become a common tool for resolving conﬂicts and establishing conditions for
a stable peace in war-torn countries. The United Nations spends more money on peacekeeping
today than ever before. Against this background, we are interested in evaluating the eﬀect
of peacekeeping operations (henceforth PKOs) and their potential for reducing conﬂict in the
future. How eﬀective are PKOs in decreasing the risk of conﬂict? And what type of eﬀect can
we expect from PKOs in the future, depending on what peacekeeping policy the UN employs?
Answers to these questions are not so straightforward. PKOs may aﬀect future conﬂict
through several pathways. They may increase the duration of post-conﬂict peace; they may
prevent contagion to neighboring countries; they may reduce the lethality of ongoing conﬂicts;
and may even reduce the intensity of the conﬂict should it recur. The impact of a PKO is
likely to last for a long time. Internal conﬂicts that break out typically last 5–10 years, and
the risk of conﬂict recurrence is high for at least a decade after the war ends. Recurrent wars
also tend to drag out for years. A successful conﬂict prevention, then, will beneﬁt the country
and its neighborhood for 20–30 years relative to the counterfactual.
Several studies show the beneﬁcial eﬀects of PKOs along one of these pathways: peacekeeping
reduces the amount of violence during conﬂict (Gilligan and Sergenti 2008), it increases
the chances of conﬂict ending (Doyle and Sambanis 2006a) it reduces the risk of conﬂict recurrence
a few years after a war has ended (Doyle and Sambanis 2000; ?), and PKOs limit the
onset of conﬂict in neighboring countries (Beardsley 2011). However, none of these studies
assesses the total eﬀect of PKOs along multiple pathways, and they are therefore likely to
severely under-estimate the beneﬁts of PKOs.
In this paper, we make use of simulations based on a statistical model to evaluate how
PKOs aﬀect future incidence of armed conﬂict along all these pathways. Earlier studies on
peacekeeping have shown that the size of PKO budgets and the robustness of their mandates
are important for building peace. We use simulations to evaluate the substantial impact of
those variables on the risk of conﬂict in the period 2010–2035. We specify eight scenarios
reﬂecting diﬀerent potential policies on how much to spend on peacekeeping and what mandates
to provide, which countries to target, and how soon a mission is deployed after a major
conﬂict breaks out. These scenarios are informed by previous research on where peacekeepers
go, our own statistical estimations of relevant factors, and reports by UN sources about the
likely future of peacekeeping.
Our ﬁndings show that peacekeeping works. The more the UN is willing to spend on
peacekeeping, and the stronger the mandates provided, the greater is the conﬂict-reducing
eﬀect. We estimate that an ambitious UN peacekeeping policy will reduce the global incidence
of armed conﬂict by two thirds relative to a no-PKO scenario. This reduction is maintained
throughout our 25-year time frame. This is a substantial eﬀect for an intervention that often
is practical to implement if the political will is present. Even if a strong commitment scenario
means an initial sharp increase in the total UN PKO budget, our simulations show that
2
the budget would only increase for approximately ten years, and then start decreasing as a
consequence of the global reduction in the incidence of conﬂict. PKOs can thus be viewed as
a long-term investment for peace.
The paper is organized as follows. We begin by providing a review of previous research on
the conﬂict-reducing eﬀect of PKOs. Subsequently, the methodology is presented, describing
the simulation procedure as well as the data used. After that we present the results of the
eﬀect of our peacekeeping variables in the period 1970–2009, based on our statistical analysis.
We then discuss and assess the determinants of PKO deployment in order to formulate a
number of likely future PKO scenarios. Thereafter, the simulation results for the various
scenarios for the period 2010–2035 are presented. The last section oﬀers some conclusions.
2 The conﬂict-reducing eﬃcacy of PKOs
The literature has identiﬁed three pathways through which PKOs may be eﬀective. One such
pathway is by preventing conﬂict from breaking out or recurring. The task of maintaining
peace in a post-conﬂict situation was the original intention of peacekeeping, and remains the
most studied eﬀect of PKOs. Doyle and Sambanis (2000) was the ﬁrst quantitative analysis
of the eﬀect of PKOs on the duration on post-conﬂict peace. The authors ﬁnd a signiﬁcant
and substantial positive eﬀect of peacekeepers on peace building, measured two, ﬁve, or ten
years after the end of the conﬂict. This conclusion holds in several later studies. ?Fortna
(2008) ﬁnds that the risk of repeat war drops ‘by 75%–85% or more when peacekeepers are
present’ (Fortna 2008, 125).
? ﬁnds a marked diﬀerence between the eﬀectiveness of PKOs during and after the cold
war. She ﬁnds no signiﬁcant eﬀect of PKOs on peace duration for the full post-World War II
period, but a substantial and signiﬁcant eﬀect of all types of PKOs after the cold war (?, 283).
Similarly Sambanis (2008) concludes from analyzing the short and long term eﬀects of UN
PKOs that ‘the UN has actually become better at peacekeeping over time’. More generally, he
ﬁnds that the eﬀect of PKOs is strongest in the ﬁrst few years, but in the long run only local
economic recovery and institution building can ensure a lasting peace. The same conclusion
is reached by Collier, Hoeﬄer and S¨oderbom (2008). They argue that economic recovery is
the best way to achieve a stable peace, but that PKOs can make a substantial diﬀerence.
Looking more broadly at third-party enforcement of peace settlements, Hartzell, Hoddie and
Rothchild (2001, 200) ﬁnd that ﬁve years after ‘the signing of a peace agreement, the survivor
rate among settlements with an external assurance is 68 percent compared with 32 percent
for arrangements lacking such promise’.
A second pathway by which peacekeeping beneﬁts peace is by enabling the cessation of
ﬁghting or by reducing the intensity of violence in an ongoing conﬂict. Doyle and Sambanis
(2000) show that UN PKOs can in fact be eﬀective in ending ongoing violent conﬂict, at
least when provided with a strong enforcement mandate. PKOs with strong mandates are
3
also eﬀective in managing violence against civilians in an ongoing armed conﬂict (Kreps and
Wallace 2009; Hultman 2010) which may in turn have positive eﬀects on the prospects of
peace.
A third pathway through which peacekeeping works is by limiting the spatial and temporal
contagion of conﬂict. Beardsley (2011) argues that the eﬀect of peacekeeping goes beyond the
mandated scope of the mission, and shows that PKOs are eﬀective in reducing the likelihood
of conﬂict in neighboring countries. By creating stability in one country, the risk of conﬂict
contagion demonstrated by other studies (Gleditsch 2002; ?) is thus strongly reduced.
One serious methodological challenge for these studies is the issue of selection bias – if
the UN only sends missions to the easiest conﬂicts, the success rate of missions will be overestimated.
This seems not to be a major problem, however. Gilligan and Sergenti (2008)
explicitly address the non-random way in which PKOs are deployed and utilize a matching
model to guard against selection bias. They construct a new dataset where cases of countries
in which PKOs were deployed are matched to similar cases in which PKOs were not. They then
ﬁnd a clear peace-prolonging eﬀect of UN PKOs (Gilligan and Sergenti 2008, 104). This eﬀect
is stronger than in the non-matched dataset, meaning that previous research most probably
have underestimated the eﬀect of PKOs – at least on peace duration after war. Moreover,
Melander (2009) argues that since peacekeepers are sent to the most diﬃcult cases, it is
diﬃcult to observe the positive eﬀects. By accounting for this selection bias, he demonstrates
that peacekeeping can also prevent genocidal violence breaking out.
From previous research we can conclude that peacekeeping in general has a conﬂictreducing
eﬀect. However, all peacekeeping operations are not equally eﬀective. The two
characteristics that seem to be the most important are the operations’ mandate and their
size in terms of budget and troop strength. These are also the main aspects of PKOs that
are politically established by the UN Security Council. Doyle and Sambanis (2000) ﬁnd that
traditional PKOs, characterized by unarmed or lightly armed troops with very limited mandates,
do not have any eﬀect on peace duration.1 Multidimensional PKOs, on the other hand,
‘are extremely signiﬁcant and positively associated with’ peace-building success (Doyle and
Sambanis 2000, 791).2 Similarly, Doyle and Sambanis (2006a) ﬁnd that multidimensional
and enforcement missions have a signiﬁcant and substantial positive eﬀect on peace-building
success. Diﬀerentiating between a strict and a lenient deﬁnition of peace, they ﬁnd that multidimensional
PKOs ‘works well with respect to both measures, [but] UN missions in general
seem to have their greatest eﬀect in preventing lower-level violence and enabling countries
to democratize and rebuild institutions after civil war rather than prevent the resumption of
full-scale war’ (Doyle and Sambanis 2006a, 110).
1
Interestingly, ?, 238 ﬁnds that ‘traditional peacekeeping missions and observer missions have been the most
successful’ while Doyle and Sambanis (2006a, 111) ﬁnd that ‘traditional peacekeeping does not work well, and
may even have negative eﬀects’.
2
Discussing the problem of counterfactuals, King and Zeng (2007) argue that some of the Doyle and Sambanis
(2000) ﬁndings are model dependent and unsupported by empirical evidence. Sambanis and Doyle (2007)
dispute this claim.
4
Findings for the size of missions are a bit mixed. Doyle and Sambanis (2006a) argue that
the number of peacekeeping troops is a poor predictor of peace-building success – the number
of ‘boots on the ground’ must be considered in relation to the PKO’s mandate. The reason
for this, they argue, is that a ‘large troop deployment with a weak mandate is a sure sign
of lack of commitment by the Security Council (...) This suggests a mismatch between the
nature of the problem and the treatment assigned by the UN’ (Doyle and Sambanis 2006a,
113). However, most studies indicate that the size is important. Kreps (2010) argues that the
capacity of a UN mission may explain the variation in their success, suggesting that military
force is central for peacekeepers to succeed in conﬂict situations. In a study of micro-level
eﬀects of peacekeeping, Ruggeri, Gizelis and Dorussen (2013) show that the mission size
increases the level of co-operation by the conﬂict parties.3 In addition, when estimating the
determinants of post-conﬂict risk Collier, Hoeﬄer and S¨oderbom (2008) ﬁnd that ‘doubling
[PKO] expenditure reduces the risk from 40% to 31%’. While some missions receive an annual
budget of well over a billion USD, other budgets are limited to less than 50 millions. Since
the budget sets clear limits to the number of troops that can be employed, it should inﬂuence
the prospects for peace.
To summarize, PKOs are eﬀective – and they are eﬀective in generating peace through
diﬀerent pathways. While selection bias may lead scholars to underestimate the eﬀect of
peacekeeping, so does a focus on single pathways to peace. It is thus possible that PKOs
are even more eﬀective than previously suggested. The factors that have been emphasized
as particularly important for enhancing the eﬀectiveness of PKOs are the type of mandate
provided by the Security Council, as well as the size of the mission.4 Based on the theoretical
explanations proposed by previous research, we should thus expect PKOs with stronger and
wider mandates as well as larger budgets to be more successful. But how much more successful
can we expect them to be? We formulate several diﬀerent PKO scenarios in which we vary
these crucial PKO components when making predictions about the eﬀectiveness of PKOs in
reducing armed conﬂict. Before turning to these scenarios, we introduce the methodology we
use to estimate the eﬀect of PKOs and make predictions into the future.
3 Methodology
3.1 Statistical model
Earlier studies of PKOs limit their attention to particular pathways of eﬀects, and consequently
restrict the analysis to a subset of the situations in which PKOs may aﬀect the
occurence of conﬂict. Doyle and Sambanis (2000), for instance, only analyze post-conﬂict
3
This positive eﬀect also seems to exist at the macro level. Time trends presented by Heldt and Wallensteen
(2006) suggest that an increase in the number of UN troops deployed in peace operations during the 1990s
coincided with a decrease in the number of intrastate armed conﬂicts.
4
These are often closely related, since a robust mandate requires a larger budget to be implemented, but
not necessarily so, as argued by Doyle and Sambanis (2006a).
5
countries, and restrict attention temporally to the ﬁrst ten years or until conﬂict reerupts,
whichever comes ﬁrst. ?Fortna (2008) has a similar setup, and includes post-conﬂict peace
periods also after the ﬁrst ten years (but disregards countries if conﬂict reerupts). Beardsley
(2011) has the most extensive dataset, including all state-months at risk of armed conﬂict
onset, but does not include information on conﬂict duration.
A PKO that succeeds in restraining a conﬂict to a few scores of annual deaths, may shorten
the conﬂict, increase the post-conﬂict duration and even decrease the duration and intensity
of any recurrence that occurs, as well as decreasing the risk and intensity of contagion to
other countries. In principle, potential contagion has no limits. If the Afghan internal armed
conﬂict could have been restrained in the mid-1970s, there might have been no attack on the
World Trade Center in New York in 2001. To assess the total eﬀect of PKOs along all the
pathways, we must analyze all country years within the period we are studying, not only those
where PKOs are deployed. We must also use available information on the intensity of armed
conﬂict to see whether intensity of conﬂict is aﬀected. The UCDP dataset records whether
conﬂicts are minor (25–999 deaths per year) or major (1000+ deaths).
To achieve this, a central feature in our modeling is the (annual) transition probability
matrix for the transitions between peace, minor, and major conﬂict. The observed transition
probability matrix is given in Table 1. The relative frequency of transition in a given year from
minor conﬂict to major conﬂict, for instance, have been 0.103, whereas the relative frequency
of transition from major to minor conﬂict was 0.205.
Table 1: Transition probability matrix: Conﬂict at t vs. at t − 1, 1970–2009
(Conﬂict level at t)
Conﬂict at t-1 No conﬂict Minor conﬂict Major conﬂict Total
No conﬂict 5078 (0.965) 155 (0.029) 21 (0.004) 5254 (1.000)
Minor conﬂict 145 (0.207) 481 (0.689) 72 (0.103) 698 (1.000)
Major conﬂict 24 (0.077) 70 (0.205) 247 (0.724) 299 (1.000)
Observations 5247 706 340 6239
Row proportions in parentheses.
To simultaneously determine how PKOs (and other explanatory variables) have aﬀected
the probability of onset, escalation, deescalation and termination of armed conﬂict in the
1970–2009 period, we estimate a multinomial logit model with lagged dependent variables
and interaction terms between explanatory variables and the lagged dependent variables.5
This model allows representing the transition probabilities in Table 1 as functions of the
explanatory variables we describe in the next section.6
We estimate the statistical relationship between the incidence of conﬂict and the presence
of PKOs of various types and budget sizes, controlling for other factors that have been shown
5
Such models are often referred to as ‘dynamic’ models, e.g. in Przeworski et al. (2000).
6
We also make use of information of conﬂict history before t − 1, see Section 3.3.
6
to aﬀect the risk of conﬂict.7 The models are estimated on data for all countries for the
1970–2009 period.
In what follows, we treat the deployment of peacekeeping operations as an exogenous
variable. In Appendix A.1, we discuss this issue and test formally that the assumption of
exogeneity indeed holds.
3.2 Simulation procedure
Our statistical model is able to capture the eﬀects of PKOs along all three pathways for
individual years, but further analysis is required to assess the eﬀects along all the pathways
seen over multiple years. To do so, we have developed a simulation routine that takes the
estimated annual transition probabilities described above as its point of departure, but repeats
the transitions for several consecutive years.8
This allows us to estimate the complete eﬀect of PKOs. If a minor conﬂict breaks out
in a hitherto peaceful country, this increases the estimated risk of conﬂict in that country
every year for a couple of decades afterwards, as well as the risks of conﬂict in neighboring
countries. If our statistical model ﬁnds that a PKO prevents the onset (or recurrence or
escalation) of such a conﬂict, that is reﬂected in several subsequent transitions, too. Our
simulation procedure allows us to estimate the probability of conﬂict for every country for
every year over a 25-year period under diﬀerent scenarios presented below, such as one where
the UN stops deploying PKOs whatsoever, or one where the UN expands its level of ambition
further. By comparing the global and regional incidence of conﬂict under these scenarios, we
can aggregate the short-term eﬀects identiﬁed by the statistical model up to a level which
makes more sense for decision makers.
Evaluating the eﬀect over as much as 25 years may seem excessive, but the eﬀects of
large-scale violent conﬂict do frequently last for at least as long as that (Collier et al. 2003).
Hence, the beneﬁcial eﬀects of PKOs should be seen in a long perspective.
The general setup of the simulation procedure is illustrated in Figure 1 and summarized
below. We use the methodology developed by Hegre et al. (2013). We perform the following
steps: (1) Specify and estimate the underlying statistical model; (2) Make assumptions about
the distribution of values for all exogenous predictor variables for the ﬁrst year of simulation
and about future changes to these. In this paper, we base the simulations for the predictor
variables on UN projections for demographic variables and IIASA projections for education
(see Section 3.3); (3) Formulate a set of scenarios for future values of PKO variables (see
Section 5); (4) Start simulation in ﬁrst year. We start in 2010 for the forecasts presented in
7
for a review of conﬂict risk variables, see Hegre and Sambanis (2006).
8
To illustrate using the transition probabilities in Table 1: The probability of going from no conﬂict to
minor conﬂict is 0.033. If that happens, the probability that this country sees an escalation to major conﬂict
is 0.103. If that happens, the probability of sustaining major conﬂict is 0.724. Over two years, countries can
go from no conﬂict to major conﬂict through several intermediate steps. Matrix calculation on the transition
probability matrix shows that the probability of going from no conﬂict to major conﬂict over 2 years through
all possible pathways is 0.010.
7
Figure 1: Simulation ﬂow chart
Estimate model
(multinomial logit)
Load first
simulation year
(2001/2009) Calculate
transition
probabilities for all
countries
Draw transition
outcomes and
update all
variables
For all years up to 2008/2050
Draw realizations
of coefficients
Section 6.1: (5) Draw a realization of the coeﬃcients of the multinomial logit model based on
the estimated coeﬃcients and the variance-covariance matrix for the estimates; (6) Calculate
the probabilities of transition between levels for all countries for the ﬁrst year, based on the
realized coeﬃcients and the projected values for the predictor variables; (7) Randomly draw
whether a country experiences conﬂict, based on the estimated probabilities; (8) Update the
values for the explanatory variables. A number of these variables, most notably those measuring
historical experience of conﬂict and the neighborhood conﬂict variables, are contingent
upon the outcome of step 6; (9) Repeat (4)–(7) for each year in the forecast period, e.g. for
2010–2035, and record the simulated outcome; and (10)Repeat (3)–(8) a number of times
to even out the impact of individual realizations of the multinomial logit coeﬃcients and
individual realizations of the probability distributions.
The simulation methodology is reasonably accurate. Hegre et al. (2013) show that the
model speciﬁcation used in this paper is able to predict about 63% of conﬂicts (minor or
major) 7–9 years after the last year of data, with about 4% false positives.9
3.3 Description of data
3.3.1 Dependent Variable
We are interested in evaluating the eﬃcacy of PKOs in ending armed conﬂicts as well as
preventing escalation and future recurrences. Therefore, the dependent variable in this study
is a three-category variable denoting whether there is a minor conﬂict, a major conﬂict, or no
conﬂict going on in a country in a given year.
The conﬂict data used in the estimation phase of the simulation are from the 2009 update
9
Hegre et al. (2013) estimate the relationship between predictors and risk of conﬂict based on data for
1970–2000, simulates up to 2009 and compares simulation results for 2007–2009 with the most recent conﬂict
data available for the same years (Harbom and Wallensteen 2010).
8
Figure 2: Map of conﬂicts ongoing in 2008
^
^^
^
^
^
^
^
^
^
^
^
^ ^
^
^
Legend
^ PKO
No Conflict
Minor Conflict
Major Conflict
Source: Harbom and Wallensteen (2009) and Weidmann, Dorussen and Gleditsch (2010)
of the UCDP/PRIO Armed Conﬂict Dataset (ACD; Harbom and Wallensteen 2010; Gleditsch
et al. 2002). The Armed Conﬂict Dataset records conﬂicts at two levels, measured annually.
Minor conﬂicts are those that pass the 25 battle-related deaths threshold but have less than
1000 deaths in a year. Major conﬂicts are those conﬂicts that pass the 1000 deaths threshold.
We only look at internal armed conﬂicts, and only include the countries whose governments
are included in the primary conﬂict dyad (i.e., we exclude other countries that intervene in
the internal conﬂict). Figure 2 shows the conﬂicts active in 2009 as well as the UN PKOs in
action in that year.
3.3.2 PKO variables
We use data on PKOs from three diﬀerent sources. We use Doyle and Sambanis (2006a)’s
coding of mandates (hereafter we refer to Doyle & Sambanis 2006a as ‘DS’). Based on their
categorization, we code two types of mandates:10
• Traditional PKO
10
The classiﬁcation is discussed in detail in Doyle and Sambanis (2006a, p. 11–18).
9
1. Observer missions – restricted to observing actions such as a truce, troop withdrawal,
or a buﬀer zone. Always deployed with the consent of the parties to the
conﬂict. Examples are the UNMOT and UNMOP missions in Tajikistan and Croa-
tia.
2. Traditional missions – also deployed with the consent of the parties, but with somewhat
extended mandates such as policing a buﬀer zone and assisting in negotiating
a peace agreement. Examples are the UNPRESEP mission in Macedonia 1995–99
and the UNIFIL mission in Lebanon.
• Transformational PKO
1. Multidimensional missions – referred to as ‘second-generation operations’, the mandates,
also consent-based, are extended with activities intended to go to the roots of
the conﬂict, such as economic reconstruction, institutional transformation (reform
of police, army, judicial system, elections). Examples are the ONUSAC mission in
El Salvador 1991–95 and the UNMIT mission in Timor-Leste (2006– ).
2. Enforcement missions – ‘third-generation operations’ that do not require the consent
of both parties, and therefore must draw on the authority of UN Charter
articles 25, 42, and 43 to apply force to protect the activities of the operation. Examples
are the UNPROFOR mission in former Yugoslavia 1992–95 and the UNMIS
mission in Sudan (2005– ).
The simpliﬁcation of creating two categories out of the original four is based on the ﬁnding
by DS that the latter two are signiﬁcantly more eﬀective than the two former types. We
have also estimated alternative models using the original four-category variable as either a
nominal or an ordinal variable, with fairly similar results. These estimations show that the
transformational missions are more distinct from the traditional missions, both when it comes
to when they are deployed and their eﬀects on subsequent conﬂict risk. The remaining withincategory
diﬀerences are not suﬃciently large to warrant splitting a small number of missions
into many categories.
Since the DS dataset is not time-varying, we have coded changes in mandate based on the
comments on adjustments to the mandate in Doyle and Sambanis (2006b). In some unclear
cases, Fortna (2008)’s version of the DS data was consulted (which is time-varying but not
annual). The DS data are coded up to 1999. For the years 2000–2009, we have coded the
mandate on the basis of the deﬁnitions provided by DS, using UNSC resolutions and mandate
information available at the DPKO website.11 Appendix A.3 gives a list of all PKOs by
mandate.
In order to capture the size of the PKO, we have coded the yearly expenditure for each
mission, based on United Nations General Assembly published appropriation resolutions from
11
http://www.un.org/en/peacekeeping
10
1946 to the present. The variable gives the yearly amount allocated by the UN for each speciﬁc
mission. UN PKOs are mostly funded outside the ordinary UN budget, and appropriation
resolutions were therefore quite straightforward to collect and code. A small number of
missions, e.g. the United Nations Truce Supervision Organization (UNTSO), are funded
directly through the UN’s operating budget, and yearly expenditure data are harder to single
out from other budget items. These missions, however, are all small and limited. For PKO
years without expenditure data we use the average for the mission type as our best guess.
We have removed international PKOs such as the UNIKOM mission monitoring the IraqKuwait
conﬂict 1991–2003 – i.e. UN PKOs that are deployed in more than one country
simultaneously under the same mandate. There are only four such missions and they are also
listed in Appendix A.3.
3.3.3 Other predictor variables
To predict the future incidence of conﬂict, we add predictor variables that are associated with
the risk of conﬂict and for which we have good projections for the 2010–2035 period.12 As our
baseline model, we use the model speciﬁcation that was shown to produce the most accurate
out-of-sample predictions in Hegre et al. (2013). For more information see this article.
Conﬂict History We model the incidence of conﬂict, i.e. whether the country is in a
minor or major conﬂict in a given year. To model this appropriately, we include information
on conﬂict status (no conﬂict, minor, or major conﬂict) at t − 1, the year before the year of
observation in the estimation phase in order to model the probability of transitions between
each conﬂict level. The log of the number of years in each of these states up to t − 2 is also
included. We refer to this set of variables jointly as ‘conﬂict history’ variables.
Neighborhood We include information on conﬂicts in the neighborhood in order to model
and simulate the spatial diﬀusion of conﬂicts. The neighborhood of a country A is deﬁned as
all n countries [B1...Bn] that share a border with A, as deﬁned by Gleditsch and Ward (2000).
More speciﬁcally, we deﬁne ‘sharing a border’ as having less than 100 km between any points
of their territories. Islands with no borders are considered as their own neighborhood when
coding the exogenous predictor variables, but have by deﬁnition no neighboring conﬂicts.
The spatial lag of conﬂict is a dummy variable measuring whether there is conﬂict in the
neighborhood or not. Hegre et al. (2013) does not ﬁnd any diﬀerence between minor and
major conﬂicts in terms of their diﬀusion potential.13
12
Plausible and authoritative forecasts are required for our simulation exercise. This precludes including
numerous interesting variables to the model, such as level of democracy, or characteristics of the termination
of a previous conﬂict such as military victories or aspects of peace agreements. Taking these factors fully into
account would require specifying a forecasting model also for these.
13
Beardsley (2011) does not analyze this particular question.
11
Socio-economic data We use two indicators of socio-economic development, given development’s
strong relationship with the risk of conﬂict (Collier and Hoeﬄer 2004; Fearon and
Laitin 2003; Hegre et al. 2001): The extent of secondary education and the infant mortality
rates. Both variables are highly correlated with GDP per capita, for which we have no
authoritative projections.
We use the education data of Lutz et al. (2007), providing historical estimates for 120
countries for the 1970–2000 period. The dataset is based on individual-level educational attainment
data from recent Demographic Health Surveys (DHS), Labour Force Surveys (LFS),
and national censuses. Historical estimates are constructed by ﬁve-year age groups and gender
using demographic multi-state methods for back projections, and taking into account gender
and education-speciﬁc diﬀerences in mortality. We employ a measure of male secondary education,
deﬁned as the proportion of males aged 20–24 years with secondary or higher education
of all males aged 20–24. For the 2001– period (including forecasts) we use the accompanying
scenario for educational attainment until 2050 (Samir and Lutz 2008). Our base scenario is
their General Trend Scenario.
Infant mortality is deﬁned as the probability of dying between birth and exact age 1 year,
expressed as the number of infant deaths per 1000 live births. We use the medium scenario
from the population projections, where total fertility rates for all countries are assumed to
converge towards 1.85 children per woman according to a path similar to historical experiences
of fertility decline.
Demographic data The demographic variables originate from the World Population Prospects
2006 (United Nations 2007), the most authoritative global population data set which
covers all states in the international system between 1950 and 2005 and provides projections
for the 2005–2050 period. Two key demographic indicators are used in this study. Total
population is deﬁned as the de facto population in a country, expressed in thousands. The
measure has been log-transformed following an expectation of a declining marginal eﬀect on
conﬂict risk of increasing population size (see Raleigh and Hegre 2009).
We also add a variable reﬂecting the country’s age structure. Cincotta, Engelman and
Anastasion (2003) and Urdal (2006) report increasing risks of minor armed conﬂict onset
associated with youth bulges. An emerging consensus is that youth bulges appear to matter
for low-intensity conﬂict, but not for high-intensity civil war. Age-speciﬁc population numbers
are provided by the United Nations (2007), and youth bulges are measured as the percentage
of the population aged 15–24 years of all adults aged 15 years and above. For the youth bulge
measure, the three scenarios yield identical estimates until 2024 since the relevant youth
cohorts were already born by 2005. Beyond 2025, the diﬀerent fertility assumptions lead to
signiﬁcant variation in the youth bulge projections for many countries.
Temporal and regional dummies We could ﬁt the model better to the data by adding
yearly ﬁxed eﬀects – there are good reasons to believe that the underlying transition proba-
12
bility matrix for a country with a given set of characteristics is ﬂuctuating over the observed
period. Hegre et al. (2013), however, are unable to ﬁnd temporal dummies that unambiguously
improve the predictive performance of the model. Consequently, we do not include such
terms in the model for this paper.
We include three regional dummies to account for residual regional diﬀerences in risk of
conﬂict after controlling for all predictor variables. Hegre et al. (2013) only ﬁnd three regions
to be at least vaguely distinct in this manner: Eastern Europe, Western Africa, and the rest
of Africa south of Sahara. The rest of the world is the reference category for the regional
variable.
Interaction terms Our control variables may not have the same eﬀect on the probability
of conﬂict onset as on conﬂict termination. To model this ‘dynamic’ model (Przeworski
et al. 2000), we include multiplicative interaction terms between the control variables and the
conﬂict history variables.14
4 Estimation results, 1970–2009
Table 2 shows the results of estimating a multinomial logistic regression model including the
log of annual PKO expenditures as well as our control and interaction variables.15 Table 3
shows the results for a model distinguishing between the diﬀerent PKO mandates.
4.1 Short-term eﬀects of PKOs
Increasing PKO expenditures does not aﬀect the probability that a country is in minor conﬂict
in a given year, but clearly reduces the probability of major conﬂict. Figure 3 shows the
estimated short-run eﬀect on the risk of major conﬂict of the budget of PKOs based on the
results in Table 2.
A conﬂict country where a peace-keeping operation with an annual budget of USD 15
million per year is in place, has a 50% lower risk of major conﬂict than a conﬂict country
without any PKO. A mission with an annual budget of 500 million has more than 80% lower
risk than the no-PKO country.16 The eﬀect is comparable to that found by Collier, Hoeﬄer
14
The sizeable number of interaction terms entails some loss of eﬃciency, but also improves the predictive
performance of the model (Hegre et al. 2013). Since we assess the total impact of our variables by means of
simulations, the high number of parameters do not give rise to interpretational or collinearity problems. The
only concern is whether the complexity of the model gives rise to ‘empty cell’ problems. As can be seen from
the frequencies in Table 1, this is not likely to be a problem. The estimates obtained below (e.g., Table 2) do
not indicate any such diﬃculties.
15
For reference, we report the results for a model ignoring peace-keeping operations entirely in Appendix
A.4 , Table A-5.
16
We have also estimated models with a squared log expenditure variable to investigate whether the relationship
between PKO expenditure and the risk of conﬂict might be curvilinear. The squared variable did not
improve the goodness-of-ﬁt of the model.
13
Table 2: Estimation results, determinants of conﬂict, PKO budget variables
1 2
Log PKO expenditures -0.00792 (-0.18) -0.259∗∗
(-3.11)
Minor conﬂict t-1 2.443 (1.33) 3.018 (0.95)
Major conﬂict t-1 0.234 (0.07) 4.383 (1.07)
Log time in status c0 -1.240∗∗∗
(-14.72) -1.586∗∗∗
(-9.87)
Log time in status c1 1.153∗∗∗
(10.00) 0 (.)
Log time in status c2 0 (.) 1.217∗∗∗
(7.13)
Conﬂict in neighborhood 0.651∗∗
(2.89) 0.792 (1.63)
NC * minor conﬂict at t-1 -0.612∗
(-2.32) -0.611 (-1.17)
NC * major conﬂict at t-1 -1.323∗∗∗
(-3.41) -1.304∗
(-2.25)
NC * time in status c0 -0.134 (-1.40) -0.203 (-0.97)
Log population 0.345∗∗∗
(3.48) 0.188 (1.09)
Population * minor conﬂict at t-1 -0.00258 (-0.02) 0.187 (0.93)
Population * major conﬂict at t-1 -0.0394 (-0.21) 0.138 (0.58)
Population * time in status c0 -0.0483 (-1.17) 0.0709 (0.89)
Log infant mortality rate 0.0196 (0.06) 1.976∗∗
(3.00)
IMR * minor conﬂict at t-1 -0.246 (-0.61) -1.783∗
(-2.41)
IMR * major conﬂict at t-1 -0.214 (-0.40) -1.996∗
(-2.51)
IMR * time in status c0 0.264∗
(1.99) -0.304 (-1.11)
Youth bulge 0.00887 (0.23) -0.141 (-1.89)
Youth * minor conﬂict at t-1 0.00103 (0.02) 0.176∗
(2.09)
Youth * major conﬂict at t-1 0.105 (1.45) 0.244∗
(2.49)
Youth * time in status c0 -0.00773 (-0.47) 0.0457 (1.34)
Education -1.662∗
(-1.99) 1.284 (1.03)
Education * minor conﬂict at t-1 -0.00717 (-0.01) -1.520 (-1.01)
Education * major conﬂict at t-1 2.405 (1.62) -1.960 (-1.12)
Education * time in status c0 0.442 (1.35) -0.332 (-0.56)
Log IMR in neighborhood -0.307 (-1.43) -0.0915 (-0.30)
Education in neighborhood -0.505 (-0.83) -0.759 (-0.87)
Eastern Europe -0.427 (-1.15) 0.423 (0.75)
Western Africa -0.140 (-0.58) -1.844∗∗∗
(-3.40)
Rest of SS Africa 0.0731 (0.44) -0.0641 (-0.28)
Constant -3.269∗
(-2.49) -8.542∗∗
(-3.26)
N 5942
ll -1518.4
t statistics in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
14
Table 3: Estimation results, determinants of conﬂict, PKO mandate variables
1 2
Traditional PKO -0.0757 (-0.28) -0.462 (-1.14)
Transformational PKO -0.0934 (-0.29) -2.816∗∗
(-2.68)
Minor conﬂict t-1 2.434 (1.32) 3.020 (0.94)
Major conﬂict t-1 0.337 (0.10) 4.560 (1.11)
Log time in status c0 -1.241∗∗∗
(-14.72) -1.585∗∗∗
(-9.85)
Log time in status c1 1.151∗∗∗
(9.98) 0 (.)
Log time in status c2 0 (.) 1.203∗∗∗
(7.02)
Conﬂict in neighborhood 0.648∗∗
(2.87) 0.778 (1.60)
NC * minor conﬂict at t-1 -0.610∗
(-2.31) -0.614 (-1.17)
NC * major conﬂict at t-1 -1.329∗∗∗
(-3.41) -1.293∗
(-2.23)
NC * time in status c0 -0.132 (-1.39) -0.206 (-0.97)
Log population 0.343∗∗∗
(3.47) 0.186 (1.07)
Population * minor conﬂict at t-1 -0.00438 (-0.03) 0.190 (0.94)
Population * major conﬂict at t-1 -0.0446 (-0.24) 0.152 (0.64)
Population * time in status c0 -0.0477 (-1.16) 0.0751 (0.93)
Log infant mortality rate 0.0204 (0.06) 1.999∗∗
(3.05)
IMR * minor conﬂict at t-1 -0.238 (-0.59) -1.752∗
(-2.37)
IMR * major conﬂict at t-1 -0.210 (-0.39) -2.040∗
(-2.57)
IMR * time in status c0 0.263∗
(1.98) -0.307 (-1.12)
Youth bulge 0.00932 (0.24) -0.135 (-1.82)
Youth * minor conﬂict at t-1 0.000446 (0.01) 0.170∗
(2.03)
Youth * major conﬂict at t-1 0.104 (1.43) 0.241∗
(2.46)
Youth * time in status c0 -0.00781 (-0.47) 0.0450 (1.32)
Education -1.659∗
(-1.98) 1.419 (1.13)
Education * minor conﬂict at t-1 0.0136 (0.01) -1.455 (-0.95)
Education * major conﬂict at t-1 2.379 (1.61) -2.061 (-1.17)
Education * time in status c0 0.437 (1.33) -0.347 (-0.58)
Log IMR in neighborhood -0.307 (-1.43) -0.0850 (-0.28)
Education in neighborhood -0.485 (-0.80) -0.830 (-0.95)
Eastern Europe -0.423 (-1.13) 0.480 (0.85)
Western Africa -0.138 (-0.57) -1.838∗∗∗
(-3.39)
Rest of SS Africa 0.0756 (0.45) -0.0443 (-0.19)
Constant -3.271∗
(-2.48) -8.847∗∗∗
(-3.36)
N 5942
ll -1516.2
t statistics in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
15
and S¨oderbom (2008), although somewhat weaker – their estimate for log expenditures is
larger (in absolute terms) than –0.4, compared to our estimate of –0.259.17
Figure 3: Estimated eﬀect of budget
Also when represented in terms of their mandates, transformational PKOs directly aﬀect
the risk of major conﬂict only. The estimate for the transformational PKO is large and clearly
signiﬁcant. It implies that a transformational PKO reduces the risk of major conﬂict relative
to no conﬂict by more than 90%. The estimate for traditional PKOS is negative but not
statistically signiﬁcant. The parameter estimate implies that the risk of major conﬂict is 35%
lower in the presence of a traditional PKO.18
4.2 Conﬂict history
That we do not ﬁnd any direct or short-term eﬀects of peace-keeping operations on minor
conﬂicts does not mean that PKOs only reduce the intensity of conﬂicts. The transition
probability matrix in Table 1 shows that the probability of no conﬂict in a year is 0.182 after
a minor conﬂict, but only 0.077 after a major conﬂict. The probability of minor conﬂict in
a year after major conﬂict is 0.264. Given that PKOs increase the probability of transitions
from major to minor conﬂict in year t, they will also increase indirectly the probability of no
conﬂict at t + 1.
The estimates for the conﬂict history variables in Tables 2 and 3 show that this holds
more generally. The probability of minor conﬂict is much higher if there was a minor or
17
Collier, Hoeﬄer and S¨oderbom (2008) include a dummy for ‘no PKOs’. This may explain much of the
diﬀerence in estimates.
18
Although the categorizations are diﬀerent, these ﬁndings are slightly diﬀerent from Fortna (2008) who
ﬁnds that consent-based missions are in general more successful than enforcement missions. However, this
discrepancy is likely to be a result of diﬀerent designs, since she only measures the duration of peace given
a cease-ﬁre agreement, which leads to a particular selection of cases. Our results show the general ability of
peacekeeping to reduce the likelihood of conﬂict along all pathways, and in that context it is not surprising
that more extensive mandates are more successful; see e.g. Doyle and Sambanis (2000).
16
major conﬂict the year before.19 Moreover, the estimates for the ‘log time in status c0’ terms
show that the probability of conﬂict is much lower if the country has been at peace for several
years.
Eﬀective prevention of major conﬂict, then, may reduce the incidence also of minor conﬂicts
since minor conﬂicts in general more easily come to an end. The best way to assess
the combined eﬀects of these estimates is by looking into the simulated results we present in
Section 6.1. First, though, we present the PKO scenarios that we evaluate in the simulation.
5 Description and motivation of scenarios
Given that the UN has gone through a qualitative and quantitative change during the last
two decades, it is diﬃcult to predict exactly what the future of UN peacekeeping will look
like. According to a recent report by the UN which reﬂects on the future of peacekeeping,
resources are already stretched to its limits (United Nations 2009). With the global economic
crisis, potential resources are also shrinking. At the same time, the demand for peacekeeping
might become more intense (United Nations 2009).
5.1 PKO deployment rules for simulations
In our simulations, future conﬂicts occur randomly albeit with probability distributions according
to the results in Tables 2 and 3. Since we do not know where conﬂicts will occur, we
cannot know where PKOs will be needed. We therefore have to specify rules for where our
simulations will ‘send’ PKOs. These rules are based on studies identifying factors inﬂuencing
the likelihood of intervention in internal conﬂicts by the UN or other third parties.
Gilligan and Stedman (2003, 38) argue that ‘the UN acts in ways that corroborate its
humanitarian and security missions (...) one of the best predictors of UN intervention is the
number of deaths in a conﬂict’. Similarly, ?Fortna (2008) ﬁnds that UN peacekeepers tend
to ‘deploy to more diﬃcult cases rather than to easier ones’ (Fortna 2008, 44), where diﬃcult
cases are mainly deﬁned as conﬂicts with strong rebels.20
In an analysis of where PKOs are deployed (reported and discussed in Appendix A.2), we
conﬁrm that PKOs indeed are more frequently deployed to major conﬂict than to minor ones
in our dataset. Given limited resources, the UN prioritizes the most intense conﬂict areas
which constitute the greatest threats to regional stability. Our ﬁrst rule is accordingly:
Rule 1 Peace-keeping operations are initiated if the conﬂict is major (more than 1,000 battle
deaths in the previous year).
19
This inference is based on the multiple interaction terms involving conﬂict at t − 1.
20
In contrast to Gilligan and Stedman (2003), ? does not ﬁnd that the number of fatalities or the duration
of the conﬂict is a signiﬁcant predictor of UN intervention. Still, the authors at least tacitly agree that
peacekeepers are sent to the more intractable conﬂicts, although they diﬀer on what exactly intractability
implies.
17
The UN is unlikely to deploy a mission in the ﬁrst year of armed conﬂict. Other diplomatic
tools are considered ﬁrst, and the Security Council needs to come to an agreement before
a PKO can be established. Gilligan and Stedman (2003) also ﬁnd that the duration of
conﬂict matters signiﬁcantly. The longer a conﬂict lasts, the higher the probability of a UN
intervention. To give two recent examples, the mission in Sierra Leone was initiated in the
second year of major conﬂict, and the mission in the Democratic Republic of Congo was
initiated in the fourth year of major conﬂict. This very general rule roughly captures the
reaction time of the UN. We consequently specify a second rule:
Rule 2 A PKO is established in the third consecutive year of major armed conﬂict.
The third rule speciﬁes the duration of PKOs. The exact number of years chosen is
somewhat arbitrary, but is supported by the estimates in the incidence model in our analysis
of where PKOs go (Table A-3).
Rule 3 Peacekeepers remain for ﬁve years after last year with conﬂict activity (more than
25 battle-related deaths within a calendar year). This rule also applies to all PKOs active in
2009.
The fourth and ﬁfth rules restrict PKOs from being deployed in certain countries. Mullenbach
(2005) argues that international-level factors are more important than state-level factors
in determining where third parties intervene. Controlling for state- and conﬂict-level factors,
he ﬁnds that third-party interventions are less likely when the government of the target state
has a military alliance with a major power and signiﬁcantly less likely when the target state
is a major power (Mullenbach 2005, 549–52). Major powers are reluctant to welcome international
involvement in their internal aﬀairs, and have as permanent members of the Security
Council (P5) authority to veto such decisions.
Rule 4 PKOs are never deployed in permanent UNSC members.
Moreover, the UN is also highly unlikely to establish a PKO in states with very large
populations. (Gilligan and Stedman 2003, and our analysis in Table A-3). The largest country
ever to attract a PKO is Sudan, with a population of 37 millions in 2005. Therefore, in all
scenarios except S4, S7, and S8, our simulations adher to a ﬁnal rule:21
Rule 5 For most scenarios, PKOs are deployed only in countries that have smaller populations
than 100 millions in 2009.
21
This precludes PKOs in Bangladesh, Brazil, India, Indonesia, Japan, Mexico, Nigeria, and Pakistan in
addition to the permanent UNSC members.
18
5.2 Specifying PKO scenarios
We specify eight diﬀerent scenarios to explore the eﬀect on the global incidence of conﬂict of
various UN policies. The ﬁrst scenario (S1) is a comparison scenario where the UN terminates
all PKO activity in 2010. Here, the only policy rule is no deployment of PKOs.
Provided that the UN decides to establish a PKO, there are diﬀerent potential scenarios in
terms of mandate and budget – two factors that have been emphasized by previous research
to have substantial consequences for the eﬀectiveness of the mission. When it comes to
mandates, this is an area in which UN PKOs have recently undergone a major change. While
observer missions and traditional peacekeeping mandates used to dominate the actions of the
UN, recent operations have seen more multidimensional and enforcement mandates.
Figure 4: Number and total budget of UN PKO missions by mandate type, 1970–2009
Figure 4 depicts the number of and total budgets of UN PKO missions in our dataset
by mandate type. Multi-dimensional and enforcement missions were inventions of the early
1990s. Complex situations in for example the Balkans, Somalia, and Rwanda led to a surge
of PKOs with more robust mandates, but the perceived failures of several such missions led
to a slight decrease in UN peacekeeping initiatives (Durch and Berkman 2006). At the turn
of the century, the Brahimi Report (United Nations 2000) set the agenda for the future of
UN peacekeeping, and the UN again initiated a number of enforcement missions in conﬂict
situations.
Several facts are readily apparent from Figure 4: First, both the frequency and types
of PKOs changed after the end of the Cold War – in terms of frequency (left panel), the
traditional and observer missions were supplemented by multidimensional and enforcement
missions. The right panel clearly shows that enforcement missions account for an increasing
share of the total UN PKO budget. Because of the shift in both composition and scale of
PKOs after the end of the Cold War, we will mainly focus on the 1990s and 2000s in the
remainder of this section.
Figure 5 shows the budgets of all PKOs active in 2000 (left ﬁgure) and 2009 (left ﬁgure).
These display a tendency to provide larger budgets for missions with more robust mandates.
19
Figure 5: Budget of UN PKO missions by mandate type, 2000 (left) and 2009 (right)
These operations are more complex and are consequently likely to have larger budgets. In
2000, the Brahimi report emphasized the need for more robust mandates and an increase in
resources (United Nations 2000). This marked a shift in both the nature of and the resources
spent on peacekeeping. As shown by Figure 4, the number of peace enforcement missions
have increased substantially since 2000, and as a consequence the total budget has increased
dramatically in the same period.
We outline four scenarios (S1–S4) in which the UN chooses to spend diﬀerent amounts on
each mission, ignoring the mandates. The ﬁnal four scenarios (S5–S8) vary the mandates of
the PKOs, ignoring the budget of the mission.
There are of course economic constraints which sets certain limits to the number of peacekeeping
operations that the UN can manage at the same time, as well as to the resources that
can be allocated to these missions. In a scenario with many enforcement missions, the total
amount spent on PKOs would be substantially larger than today’s levels.22 However, it seems
robust mandates are here to stay. In 2006, the Secretary-General noted that ‘United Nations
peacekeeping succeeds or fails depending on the provision of suﬃcient capacity to implement
a mandate’ (United Nations 2006). One of the main points made in United Nations (2009) is
that the UN needs to strengthen partnership with e.g. the African Union and the European
Union. Parts of the budget could thus be borne by these partners in joint operations as the
one in Darfur.
The eight scenarios are summarized in Table 4.
6 The simulated eﬀect of PKOs, 2010–2035
6.1 Prediction results
Figure 6 shows the estimated proportion of countries in conﬂict – major conﬂicts only, for
the baseline scenario without any future peace-keeping operations (S1). The simulations are
22
Below, we estimate the most expansive policy to double UN PKO expenditures.
20
Table 4: Overview of PKO scenarios
Scenario Description
1 No PKO
2 PKO, unknown mandate, budget 100 million USD per year, no large countries
3 PKO, unknown mandate, budget 800 million USD per year, no large countries
4 PKO, unknown mandate, budget 800 million USD per year, also in large countries
5 PKO, traditional mandates, unknown budget, no large countries
6 PKO, transformational mandates, unknown budget, no large countries
7 PKO, transformational mandates, unknown budget, also in large countries
8 PKO, transformational mandates, unknown budget, also in large countries, deploy in ﬁrst year
based on the estimates reported in Table 2. The left panel shows the mean proportion of
countries in both types of conﬂict and the 10th and 90th percentile over 1,000 simulations.
The right panel shows the same for major conﬂicts only.
Figure 6: Simulation 2010–2035, No PKO scenario
Left: All conﬂicts. Right: Major conﬂicts only.
Socio-economic development variables are important predictors of conﬂict, and our UN/
IIASA forecasts expect positive changes for most countries over the next 25 years. Hence, we
predict a moderate decline in the proportion of countries in conﬂict even without any PKOs,
and that the incidence of major conﬂict remains at the same level as the last ﬁve years.
Figure 7 compares the simulated proportion of countries in conﬂict in the baseline scenario
with the proportion obtained in the other scenarios. In the left panel, we compare the baseline
scenario with scenarios S2, S3, S4 varying the budget of missions. These simulations are
based on the estimates reported in Table 2. The set of black lines represent the incidence of
all armed conﬂicts, the gray set major conﬂicts only. All three scenarios imply a reduction in
the incidence of conﬂict, although the estimated eﬀect is not very large – the diﬀerence in the
incidence of conﬂict is about one percentage point. We ﬁnd a diﬀerence between scenarios
without PKOs and those with PKOs, but beyond that there is little diﬀerence between the
predictions for the various budget levels. Note that the reduction in the incidence is somewhat
larger for major conﬂicts than for minor conﬂicts. This is particularly true if we look at the
21
proportional reduction in the incidence of major conﬂict. In 2035, the predicted incidence of
conﬂict for the most extensive scenario is less than half that of the baseline scenario.
Figure 7: Simulation 2010–2035, The eﬀect of various PKO scenarios
Left: Budget scenarios. Right: Mandate scenarios.
The right panel shows the predicted incidence of conﬂict varying the mandates of the
PKOs. These simulations are based on the estimates reported in Table 3. As expected from
the estimates in the two results tables, the eﬀect of varying mandates is stronger than of
varying budget. The predicted reduction in the incidence of both levels of conﬂict is about
2%, and the reduction in the incidence of major conﬂict is about the same. This means that
the most extensive scenario reduces the risk of major conﬂict in 2035 with about two thirds.
Note that the incidence of minor conﬂict – the diﬀerence between the black and gray lines
in Figure 7 – is not reduced by any large extent. Our simulations indicate that the indirect
eﬀect of PKOs on minor conﬂict is weak. However, it is not so that PKOs merely reduce the
intensity of conﬂict without increasing the chance of peace. If that was the case, the incidence
of both levels of conﬂict would not be reduced in any of the scenarios. Our simulations imply
that for every successful transition from major to minor conﬂict due to the presence of a PKO,
there is one transition from minor conﬂict to no conﬂict. In sum, these results imply that UN
PKO policy matters signiﬁcantly, and that the UNSC has the power to substantively enhance
global security.
Our model allows for capturing long-term and spatial eﬀects of conﬂict. The estimates for
the ‘log time in status’ variables indicate that the probability of no conﬂict increase strongly
with several consecutive years of peace, and decrease with several consecutive years of conﬂict.
Likewise, conﬂicts in neighboring countries increase the risk that conﬂicts erupt. Given that
we ﬁnd that PKOs have a clear short-term eﬀect, we might expect the diﬀerence between
scenarios to increase over time. There are indications that this is the case. The diﬀerence
for the overall incidence of conﬂict between the the no-PKO and PKO scenarios in Figure 7
clearly widen from the ﬁrst year of simulation up to 2020. After the initial 10 years, the lines
are roughly parallell. The lines become parallell considerably earlier for the incidence of major
22
Table 5: List of regions
Number Region Name
1 South America, Central America, and the Caribbean
2 Western and Southern Europe, North America, and Oceania
3 Eastern Europe
4 Western Asia and North Africa
5 Western Africa
6 East, Central, and Southern Africa
7 South and Central Asia
8 Eastern and South-East Asia
conﬂict. The indirect eﬀect over time and space is probably the reason for this diﬀerence in
response over time to the various scenarios.23
6.2 Regional eﬀects
We deﬁne 8 regions as listed in Table 5. The list is a condensed version of the UN region
deﬁnition.24 In Figure 8 we show simulated incidence of conﬂict in six of these regions, varying
the mandates of the scenarios as in the right panel of Figure 7.
We also plot the observed proportion of countries in conﬂict for the 1995–2009 period for
each region. Since there are only a handful of conﬂicts within each region, individual conﬂicts
are discernible in the plots. In West Africa, for instance, there were no major conﬂicts from
1995 to 1997, and one conﬂict in 1998–99. In the 1995–2004 period the total number of
conﬂicts ﬂuctuated between 2 and 4. In the prediction part of the ﬁgure for the West Africa
region, 10% of the countries are predicted to be in conﬂict. This, then, corresponds to about 2
conﬂicts every year. The expected number of major conﬂicts (more than 1,000 battle-related
deaths) is less than 0.5 for this region.
PKOs seem to have the strongest eﬀects in ‘West Asia and North Africa’, ‘East, Central,
and Southern Africa’ and ‘South and Central Asia’ (regions 4, 6, and 7). Since the other
three regions have had few major conﬂicts in the post-Cold war period, the model predicts
a continued low incidence of these conﬂicts. Since PKOs in our scenarios are initiated only
in major conﬂicts, we consequently predict fewer deployments in these regions, and they
therefore only marginally aﬀect the regional incidence of conﬂict.
In the other three regions, however, PKOs substantially reduce the number of conﬂicts. In
the ‘West Asia and North Africa’ region, we predict a clear decline in the incidence of conﬂict
because of the relatively high levels of socio-economic development in the region. Particularly
in the ﬁrst 15 years of the simulation, PKOs with strong mandates would according to our
model reinforce this declining trend. Since there are few large countries in the region, there
23
The gradual widening of the diﬀerence between PKO and no-PKO scenarios is also due to a gradual
readjustment to a new steady-state equilibrium for the incidence of conﬂict in a probabilistic model, so it is
not straightforward to quantify the long-term eﬀect of PKOs in this manner.
24
The UN list is found at http://www.un.org/depts/dhl/maplib/worldregions.htm.
23
Figure 8: Simulation 2010–2035, both conﬂict levels. Varying mandates, six regions
is little diﬀerence between S6 and S8 for this region.
The ‘East, Central, and Southern Africa’ and ‘South and Central Asia’ regions are the
ones with the highest incidence of conﬂict in the post-Cold War period, and also in our predictions.
In ‘East, Central, and Southern Africa’, the strong mandate scenario (S6) reduces
the predicted incidence of major conﬂict from about 8% of the countries to about 3%, corresponding
to more than two conﬂicts every year to less than one. In ‘South and Central Asia’,
the predicted incidence of conﬂict is about 11–12% in 2015 and slowly decreasing under the
baseline no-PKO scenario. Scenario S6, with enforcement mandates for all conﬂicts in smaller
countries, reduces the predicted incidence to about 7%. Scenario S8, which allows for PKOs
in large countries (but not in the permanent members of the UNSC) and deployment in the
24
ﬁrst year of a major conﬂict reduces the incidence by another couple of percents. If Gilligan
and Stedman (2003) are correct that the UN is less inclined to intervene in Asian conﬂicts,
the UNSC has strong reasons to reconsider this policy. The potential eﬀect of PKOs is strong
in this area, and a policy shift would substantially decrease the incidence of armed conﬂict.
6.3 Budget eﬀects
As noted earlier, the high-commitment scenarios are costly. We are therefore interested in
assessing how much the UN budget would increase in our various scenarios. Figure 9 shows the
average simulated total budgets for UN PKOs under scenarios S2, S3, and S4. The low-budget
scenario (S2) would imply a strong reduction in UN peace-keeping expenditures, whereas the
other two scenarios represent an increase in total annual expenditures of 50–70% compared
to what the UN spent in 2009. These are signiﬁcant increases in expenditures, especially
since the UN is already struggling to get the funds and the troop commitments required to
carry out the current missions. Nevertheless, would the member states of the UN be willing
to increase their support for PKOs, there is much to be gained – not only in terms of global
security, but also in terms of development and economic growth. Gates et al. (2012) show the
detrimental eﬀects of armed conﬂict on development. According to their analysis, a major
armed conﬂict with 2500 battle deaths – i.e. those conﬂicts that PKOs are particularly good
at reducing – increases undernourishment by 3.3 percent and infant mortality by 10 percent.
deGroot and Br¨uck (2012) ﬁnd that the world would have seen a 16 percent larger global GDP
in the absence of war in the last ﬁve decades. They particularly emphasize the economic gains
to be made by ending wars earlier, which is what PKOs are well designed to do according to
our analysis. Considering the enormous negative externalities of armed conﬂict, an increase
in the UN PKO budget by 50 percent may be a relatively cheap way of investing in future
global security and development.
Figure 9: Simulated total UN PKO budgets, 2010–2035
7 Conclusion
In this paper we have evaluated the prospects of PKOs in reducing conﬂict in the future. By
simulating diﬀerent scenarios, we have estimated the eﬀect on the future incidence of conﬂict
of diﬀerent types of missions and of varying the money spent on PKOs. The results show that
25
PKOs have a clear conﬂict-reducing eﬀect. The eﬀect of PKOs is largely limited to preventing
major armed conﬂicts. However, there is a discernible indirect eﬀect since the reduction of
conﬂict intensity also tends to increase the chances of peace in following years. There are also
some interesting regional diﬀerences. PKOs have the strongest eﬀect in three regions: West
Asia and North Africa; East, Central, and Southern Africa; South and Central Asia. This
reﬂects that these regions have had more major conﬂicts compared to other regions.
These ﬁndings have some clear policy implications, since they illustrate the eﬀect of different
PKO policies. In one of the most extensive scenarios, in which major armed conﬂicts
are met with a PKO with an annual budget of 800 million USD, the total UN peacekeeping
budget is estimated to increase by 50–70 percent. However, in this scenario, the risk of major
armed conﬂict is reduced by half relative to a scenario without any PKO. This indicates that
a large UN peacekeeping budget is money well spent. Moreover, the total PKO budget would
increase for about ten years, and then start decreasing again as a result of a reduced number
of conﬂicts in the world. In another scenario, which speciﬁes that major conﬂicts get a PKO
with a transformational mandate in the ﬁrst year, the risk of conﬂict is reduced by two-thirds
in 2035 compared to a scenario without any PKO. If the UN is serious about maintaining
international peace and security, it is important to consider the impact of diﬀerent policies
regarding mandates and budgets, as well as the reaction-time from a conﬂict outbreak to the
deployment of a mission.
The methodology used here opens up for new interesting questions and possible extensions
to the research presented. One pertinent question is whether the quality of PKOs may not
be equally important for its eﬃciency as the mandate and the budget. Troop-contributing
countries have varying levels of military training and a large number of countries contributing
troops to a single mission may introduce coordination problems. Another relevant issue is
the impact of regional security actors. In this paper we have evaluated the eﬀect of UN
PKOs, but the UN is not the only actor doing peacekeeping. For example, the African Union
and NATO have been involved in several conﬂict and post-conﬂict situations. Therefore it
would be interesting to assess whether these actors diﬀer in their peacekeeping eﬃcacy, and
subsequently simulate a future scenario that takes into account the increasing involvement
of regional actors in peacekeeping. The simulation procedure used here oﬀers a useful tool
for evaluating the practical relevance of theoretical insights as well as assessing the impact of
diﬀerent policies.
26
References
Beardsley, Kyle. 2011. “Peacekeeping and the Contagion of Conﬂict.” Journal of Politics 73:1051–1064.
Cincotta, Richard P., Robert Engelman and Daniele Anastasion. 2003. The Security Demographic:
Population and Civil Conﬂict After the Cold War. Washington, DC: Population Action Interna-
tional.
Collier, Paul and Anke Hoeﬄer. 2004. “Greed and Grievance in Civil War.” Oxford Economic Papers
56(4):563–595.
Collier, Paul, Anke Hoeﬄer and M˚ans S¨oderbom. 2008. “Post-conﬂict risks.” Journal of Peace Research
45(4):461–478.
Collier, Paul, Lani Elliot, H˚avard Hegre, Anke Hoeﬄer, Marta Reynal-Querol and Nicholas Sambanis.
2003. Breaking the Conﬂict Trap. Civil War and Development Policy. Oxford: Oxford University
Press.
deGroot, Olaf J., Carlos Bozzoli and Tilman Br¨uck. 2012. “The Global Economic Burden of Violent
Conﬂict.” DIW Berlin.
Doyle, Michael W. and Nicholas Sambanis. 2000. “International Peacebuilding: A Theoretical and
Quantitative Analysis.” American Political Science Review 94(4):779–801.
Doyle, Michael W and Nicholas Sambanis. 2006a. Making War & Building Peace: United Nations
Peace Operations. Princeton: Princeton University Press.
Doyle, Michael W. and Nicholas Sambanis. 2006b. “Notes on the Coding of Mandates of UN Peace
Operations for Michael Doyle and Nicholas Sambanis, 2006, Making War and Building Peace:
United Nations Peace Operations (Princeton, NJ: Princeton University Press).” Available online
at http://pantheon.yale.edu/ ns237/index/research/DS2006replication.zip.
Durch, William J. and Tobias C. Berkman. 2006. Restoring and Maintaining Peace: What We Know
So Far. In Twenty-First-Century Peace Operations, ed. William J. Durch. United States Institute
of Peace and The Henry L. Stimson Center pp. 1–48.
Fearon, James D. and David D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American
Political Science Review 97(1):75–90.
Fortna, Virginia Page. 2008. Does Peacekeeping Work?: Shaping Belligerents’ Choices After Civil
War’. Princeton: Princeton University Press.
Gates, Scott, H˚avard Hegre, H˚avard Mokleiv Nyg˚ard and H˚avard Strand. 2012. “Development Consequences
of Armed Conﬂict.” World Development 40(9):1713–1722.
Gibler, Douglas M. and Meredith Sarkees. 2004. “Measuring Alliances: The Correlates of War Formal
Interstate Alliance Data set, 1816–2000.” Journal of Peace Research 41(2):211–222.
Gilligan, Michael J and Ernest J Sergenti. 2008. “Do UN Interventions Cause Peace? Using Matching
to Improve Causal Inference.” Quarterly Journal of Political Science 3:89–122.
Gilligan, Michael and Stephen John Stedman. 2003. “Where Do the Peacekeepers Go?” International
Studies Review 5(4):37–57.
Gleditsch, Kristian S. 2002. All International Politics is Local: The Diﬀusion of Conﬂict, Integration,
and Democratization. Ann Arbor, MI: University of Michigan Press.
Gleditsch, Kristian S. and Michael D. Ward. 2000. “War and Peace in Space and Time: The Role of
Democratization.” International Studies Quarterly 44(1):1–29.
27
Gleditsch, Nils Petter, Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg and H˚avard Strand.
2002. “Armed Conﬂict 1946–2001: A New Dataset.” Journal of Peace Research 39(5):615–637.
Greene, William H. 2003. Econometric Analysis. 5th Edition. Upper Saddle River: Prentice-Hall.
Harbom, Lotta and Peter Wallensteen. 2009. “Armed Conﬂicts, 1946–2008.” Journal of Peace Research
46(4):577 – 587.
Harbom, Lotta and Peter Wallensteen. 2010. “Armed Conﬂicts, 1946–2009.” Journal of Peace Research
47(4):501–509.
Hartzell, Caroline, Matthew Hoddie and Donald Rothchild. 2001. “Stabilizing the Peace After Civil
War: An Investigation of Some Key Variables.” International Organization 55(1):183–208.
Hegre, H˚avard, Joakim Karlsen, H˚avard Mokleiv Nyg˚ard, H˚avard Strand and Henrik Urdal. 2013.
“Predicting Armed Conﬂict 2010–2050.” International Studies Quarterly 55(2).
URL: http://onlinelibrary.wiley.com/doi/10.1111/isqu.12007/pdf
Hegre, H˚avard and Nicholas Sambanis. 2006. “Sensitivity Analysis of Empirical Results on Civil War
Onset.” Journal of Conﬂict Resolution 50(4):508–535.
Hegre, H˚avard, Tanja Ellingsen, Scott Gates and Nils Petter Gleditsch. 2001. “Toward a Democratic
Civil Peace? Democracy, Political Change, and Civil War, 1816–1992.” American Political Science
Review 95(1):33–48.
Heldt, Birger and Peter Wallensteen. 2006. Peacekeeping Operations: Global Patterns of Intervention
and Success, 1948-2004. 2nd ed. Folke Bernadotte Academy Publications.
Hultman, Lisa. 2010. “Keeping Peace or Spurring Violence? Unintended Eﬀects of Peace Operations
on Violence against Civilians.” Civil Wars 12:29–46.
Kennedy, Peter. 2008. A Guide to Modern Econometrics. Oxford: Blackwell Publishing.
King, Gary and Langche Zeng. 2007. “When Can History Be Our Guid? The Pitfalls of Counterfactual
Inference.” International Studies Quarterly 51:183–210.
Kreps, Sarah. 2010. Why Does Peacekeeping Succeed or Fail? Peacekeeping in the Democratic Republic
of Congo and Sierra Leone. In Modern War and the Utility of Force: Challenges, Methods, and
Strategy, ed. Jan Angstrom and Isabelle Duyvesteyn. Routledge.
Kreps, Sarah and Geoﬀrey Wallace. 2009. Just How Humanitarian Are Interventions? Peacekeeping
and the Prevention of Civilian Killings during and after Civil Wars. In American Political Science
Association Annual Meeting, Toronto, 3-6 September.
Lutz, Wolfgang, Anne Goujon, K.C. Samir and Warren Sanderson. 2007. “Reconstruction of Population
by Age, Sex and Level of Educational Attainment for 120 Countries 1970–2000.” Vienna Yearbook
of Population Research 5:193–235.
Melander, Erik. 2009. “Selected to Go Where Murderers Lurk? The Preventive Eﬀect of Peacekeeping
on Mass Killings of Civilians.” Conﬂict Management and Peace Science 26:389–406.
Mullenbach, Mark J. 2005. “Deciding to Keep Peace: An Analysis of International Inﬂuences on the
Establishment of Third-Party Peacekeeping Missions.” International Studies Quarterly 49:529–555.
Przeworski, Adam, Michael E. Alvarez, Jos´e Antonio Cheibub and Fernando Limongi. 2000. Democracy
and Development. Political Institutions and Well-Being in the World, 1950–1990. Cambridge:
Cambridge University Press.
28
Raleigh, Clionadh and H˚avard Hegre. 2009. “Population, Size, and Civil War. A Geographically
Disaggregated Analysis.” Political Geography 28(4):224–238.
Ruggeri, Andrea, Theodora-Ismene Gizelis and Han Dorussen. 2013. “Managing Mistrust: An Analysis
of Cooperation with UN Peacekeeping in Africa.” Journal of Conﬂict Resolution .
URL: http://jcr.sagepub.com/content/early/2012/07/16/0022002712448906.abstract
Sambanis, Nicholas. 2008. “Short-Term and Long-Term Eﬀects of United Nations Peace Operations.”
The World Bank Economic Review 22:9–32.
Sambanis, Nicholas and Michael W Doyle. 2007. “No Easy Choices: Estimating the Eﬀects of United
Nations Peacekeeping (Response to King and Zeng).” International Studies Quarterly 51:217–226.
Samir, KC; Bilal Barakat, Vegard Skirbekk and Wolfgang Lutz. 2008. Projection of Populations by
Age, Sex and Level of Educational Attainment for 120 Countries for 2005–2050. IIASA IR-08-xx.
Laxenburg, Austria: International Institute for Applied Systems Analysis.
UCDP. 2012. “UCDP Battle-related deaths dataset v.5-2012.” Uppsala Conﬂict Data Program.
URL: www.ucdp.uu.se
United Nations. 2000. Report of the Panel on United Nations Peace Operations. New York: United
Nations. A/55/305, S/2000/809.
United Nations. 2006. Overview of the ﬁnancing of the United Nations peacekeeping operations: budget
performance for the period from 1 July 2004 to 30 June 2005 and budget for the period from 1 July
2006 to 30 June 2007. New York: United Nations. Report of the Secretary-General 24 Feb,
A/60/696.
United Nations. 2007. World Population Prospects. The 2006 Revision. New York: UN.
United Nations. 2009. A New Partnership Agenda: Charting a New Horizon for UN Peacekeeping.
New York: UN Department of Peacekeeping Operations and the Department of Field Support.
Urdal, Henrik. 2006. “A Clash of Generations? Youth Bulges and Political Violence.” International
Studies Quarterly 50(3):607–630.
Weidmann, Nils B, Hans Dorussen and Kristian Skrede Gleditsch. 2010. “The Geography of the
International System: The CShapes Dataset.” International Interactions 36:86–106.
Wooldridge, Jeﬀrey M. 2010. Econometric Analysis of Cross Section and Panel Data. Cambridge:
MIT Press.
Zartman, I William. 2001. “The timing of peace initiatives: Hurting stalemates and ripe moments.”
Global Review of Ethnopolitics 1(1):8–18. doi:10.1080/14718800108405087.
URL: http://www.tandfonline.com/doi/abs/10.1080/14718800108405087
29
A Appendix
A.1 When do they go? Exploring potential endogeneity
Several studies have looked at whether PKOs are subject to a selection eﬀect such that they
are sent to the ‘easy’ conﬂicts. So far there is little or no evidence that would support such a
claim (?Gilligan and Sergenti 2008; Gilligan and Stedman 2003). Being sent to particularly
intense conﬂicts would, however, not be the only way a selection eﬀect could inﬂuence the
estimated eﬃciency of PKOs. Another possible mechanism would be one were peacekeepers
are sent to conﬂicts after the conﬂicts have passed their intensity peak. PKOs would then
be deployed only when the ‘moment is ripe’ (Zartman 2001) and conﬂicts would nevertheless
have deescalated without the intervention. If so, it is untenable to attribute any causal eﬀect
of the PKO – it would simply signal the beginning of the end.
Table A-1: Onset of PKOs across ‘conﬂict trajectory’, 1970–2009
ds onstri
conﬂicttrajectory No PKO Traditional Transformational Total
No. % No. % No. % No. %
-5 3 75.0% 1 25.0% 0 0.0% 4 100.0%
-4 18 100.0% 0 0.0% 0 0.0% 18 100.0%
-3 22 88.0% 1 4.0% 2 8.0% 25 100.0%
-2 83 97.6% 1 1.2% 1 1.2% 85 100.0%
-1 101 96.2% 1 1.0% 3 2.9% 105 100.0%
0 712 97.9% 11 1.5% 4 0.6% 727 100.0%
1 84 96.6% 2 2.3% 1 1.1% 87 100.0%
2 75 96.2% 1 1.3% 2 2.6% 78 100.0%
3 13 100.0% 0 0.0% 0 0.0% 13 100.0%
4 22 100.0% 0 0.0% 0 0.0% 22 100.0%
5 4 100.0% 0 0.0% 0 0.0% 4 100.0%
Total 1,137 97.3% 18 1.5% 13 1.1% 1,168 100.0%
We investigate this claim in two ways and ﬁnd little evidence in its favor. First we construct
a ﬁve-category conﬂict variable that distinguishes between ﬁve levels of battle deaths incurred
in a given year.25 From this, we create an 11-category ‘conﬂict trajectory’ variable. This
variable tracks the escalatory process of conﬂicts by comparing the conﬂict level at t with the
level at t − 1. A conﬂict which stays at the same level scores 0 on this variable. A conﬂict
which escalates gets a positive score, and a conﬂict which de-escalates a negative score. Table
A-1 tabulates conﬂict trajectory against onset of PKOs for all country years in conﬂict or
within three years after the end of a conﬂict. The column to the right shows the total number
of conﬂict years. These have an approximately normal distribution across the trajectory
categories. The second and third columns reports the distribution of PKO onsets across the
conﬂict trajectory categories. There is only slight evidence for the hypothesis that PKOs are
deployed as the conﬂict is winding down. Half of the 31 PKO onsets were deployed in years
where the intensity level was the same as the preceding years, and 27 of the deployments
happened in years when the conﬂict trajectory was between –2 and 2. Only four cases break
25
The ﬁve categories are: 0–99, 100–499, 500–999, 1000–9999, 10,000–max. Data on annual battle deaths
come from the UCDP Battle Deaths Dataset (UCDP 2012).
30
this symmetric patterns: The operations in Cambodia (1992), El Salvador (1993), Lebanon
(1978), and Morocco (1991) were initiated following a noticeable decrease in conﬂict intensity.
Next, we conduct an instrumental variable analysis. We utilize exogenous variation in the
number of PKOs which are deployed to estimate the peace-building eﬀects of these PKOs.
Figure 4 clearly shows that the world saw many more PKOs in the period from 1989 to 1995,
and then again after 2000. The end of the Cold War opened up for a massive expansion of UN
peacekeeping activities, but enthusiasm dropped abruptly after the failures in Bosnia in 1995
and Somalia in 1993. UN PKO engagements then gradually re-expanded from 2000 onwards.
This temporal variation is not a function of the ‘demand’ for them – the number of conﬂicts
declined gradually from 1992 until 2009.26 In fact, the number of major armed conﬂict surged
in the 1998–2001 period, implying that the proportion of major conﬂicts with a PKO was
even lower than what appears from Figure 4. We use these time periods, in combination
with information on whether a conﬂict was going on in a country the previous year, as an
instrumental variable for PKO deployment. In addition, we use variables recording whether
the country was one of the ﬁve permanent members in the UN security council or an ally of
any of these.27
Instruments must satisfy two criteria: relevance and exclusion (Kennedy 2008; Greene
2003). Our instrument variables are relevant since they are highly correlated with the deployment
of PKOs (see Table A-3). As argued above, they are exogenous, therefore also satisfy
the exclusion criteria. We run a two-stage least squares (2SLS) model with this instrument in
the ﬁrst stage of the model and the incidence of major conﬂict as the dependent variable.28
We include the same control variables as in the analysis in the next section. The results of
the estimation are shown in Table A-2. In addition to the instrument we include the set of
covariates discussed above. The results from the ﬁrst-stage estimation is reported to the right,
those from the second stage to the left. The ﬁrst-stage results show that two of the three
instrument variables have a signiﬁcant eﬀect on the onset of PKOs.
In the second stage, the instrumented PKO variable has a negative and signiﬁcant eﬀect
on the incidence of major conﬂict. The magnitude of the estimate is fairly large for a relatively
imprecise instrumented variable – the diﬀerence between country-years with high and
low probabilities of PKO onset is about 0.2. This diﬀerence translates into more than 50%
reduction in the probability of major conﬂict. The Wald test for exogenity however is not
signiﬁcant, indicating that the original variable is not really endogenous and that it is ‘safe’
to conduct classical inferences (Wooldridge 2010, 472–77). All in all, we conclude that endogenity
is a minor problem for our analysis, and proceed to analyzing the impact of diﬀerent
scenarios for PKO involvement.
A.2 Where do peacekeepers go?
Table A-3 shows the results from estimating a multinomial regression model with a simpliﬁed
version of the categorical Doyle-Sambanis mandate variable as the dependent variable. As
26
See Harbom and Wallensteen (2010, p. 502) and Figure 6.
27
The ﬁve permanent members are the UK, France, China, Russia, and the US. Alliance data were taken
from Gibler and Sarkees (2004).
28
We restrict attention to the eﬀect on major conﬂicts given what we ﬁnd in Table 3 – PKOs are eﬀective
in reducing the intensity of conﬂict, not in preventing them. We also ran a two-stage probit model given that
the dependent variable is dichotomous. This yielded substantively the same results, but in the second stage
the instrumented variable was only borderline signiﬁcant. We rely on the linear 2SLS model even though we
have a dichotomous variable, since the instrumental probit model is less eﬃcient and produces larger standard
errors (Greene 2003).
31
Table A-2: When do they come: Instrumental variable regression
(1) (2)
conﬂict2 pko cat
PKO Instrumented -0.385∗∗∗
(-3.88)
Instrument, 90-95 + 00– 0.114∗∗∗
(8.20)
Instrument P5 Country 0.0411 (0.94)
Instrument P5 Allied 0.0322 (1.96)
Minor Conﬂict, t-1 -0.235∗
(-2.22) 0.00518 (0.04)
Major Conﬂict, t-1 0.135 (0.71) 0.229 (1.06)
Time in Peace -0.0441∗∗∗
(-7.18) -0.0156∗
(-2.30)
Time in Conﬂict -0.0601∗∗∗
(-7.20) -0.0264∗∗
(-3.08)
Time in War 0.217∗∗∗
(18.00) -0.0392∗∗
(-3.07)
Infant Mortality 0.0359∗
(2.57) 0.00696 (0.44)
IMR * minor conﬂict at t-1 0.0127 (0.54) 0.0799∗∗
(3.15)
IMR * major conﬂict at t-1 -0.0138 (-0.42) 0.136∗∗∗
(4.01)
IMR * time in statys c0 -0.0144∗∗
(-3.01) -0.0139∗∗
(-2.66)
Youth Bulges -0.00649∗∗∗
(-3.83) -0.00315 (-1.66)
Youth * minor conﬂict at t-1 0.00759∗∗
(2.82) -0.00129 (-0.39)
Youth * major conﬂict at t-1 0.0104∗∗
(2.59) -0.00112 (-0.24)
Youth * time in status c0 0.00239∗∗∗
(4.01) 0.00185∗∗
(2.86)
Population 0.000857 (0.21) 0.00120 (0.25)
Population * minor conﬂict at t-1 -0.00285 (-0.35) -0.0498∗∗∗
(-6.55)
Population * major conﬂict at t-1 -0.000655 (-0.05) -0.0804∗∗∗
(-7.33)
Population * time in status c0 -0.000298 (-0.21) -0.00242 (-1.55)
Education -0.0642 (-1.89) -0.0552 (-1.44)
Education * minor conﬂict at t-1 0.192∗∗
(2.92) 0.193∗∗
(2.75)
Education * major conﬂict at t-1 -0.172∗
(-2.02) 0.239∗
(2.57)
Education * time in status c0 0.0114 (1.07) -0.0182 (-1.52)
Neihborhood IMR 0.00406 (0.53) 0.0128 (1.48)
Neihborhood Education 0.0360 (1.25) 0.130∗∗∗
(4.61)
NC * minor conﬂict at t-1 0.0279∗∗∗
(3.99) 0.0164∗
(2.11)
NC * major conﬂict at t-1 -0.0308 (-1.58) -0.101∗∗∗
(-5.40)
NC * time in status c0 -0.00517∗
(-2.10) 0.00484 (1.73)
Eastern Europe -0.00325 (-0.32) -0.00674 (-0.58)
Western Africa -0.0385∗∗∗
(-3.74) -0.0316∗∗
(-2.75)
Rest of SS Africa -0.0260∗∗
(-2.84) -0.0690∗∗∗
(-9.26)
cons 0.138∗∗
(3.08) 0.121∗
(2.43)
ll 957.0
aic . -1847.9
N 5946 5946
t statistics in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
explained in Section 3, we have merged the ‘observer’ and ‘traditional’ categories into a new
‘traditional operation’ category, and the ‘multidimensional’ and ‘enforcement’ categories into
‘transformational operations’. The model is estimated only for the post-1989 period, and only
for country years where the country is either in conﬂict or has had a conﬂict within the last
10 years. We have excluded the permanent members of the UNSC from the data set used
here, since these countries are very likely to veto PKOs in own internal conﬂicts.
Model 1 – onset – is restricted to PKO onsets, i.e. conﬂict/post-conﬂict country years
where a peace-keeping operation continued from the previous with the same mandate have
ben removed from the data set. Model 2 – incidence – includes all conﬂict/post-conﬂict
country years for the 1990–2009 period.
As noted by previous studies, it is diﬃcult to identify circumstances in which conﬂict
32
Table A-3: Where do they go: Determinants of peace-keeping operations, 1990–2009
(1) (2)
Onset Incidence
Traditional Transformational Traditional Transformational
Traditional operation t–1 0 4.733∗∗∗
6.168∗∗∗
4.676∗∗∗
(.) (6.90) (13.77) (7.59)
Transformational operation t–1 3.028∗∗∗
0 2.726∗∗
6.878∗∗∗
(3.30) (.) (3.17) (11.79)
Major conﬂict t 1.882∗
1.600∗
1.232 1.932∗∗
(2.38) (2.00) (1.78) (2.85)
Minor conﬂict t–1 0.286 1.080 0.0936 -0.700
(0.38) (1.43) (0.14) (-1.05)
Major conﬂict t–1 -0.0883 -0.547 -0.610 -1.536
(-0.09) (-0.47) (-0.65) (-1.65)
Post-conﬂict year 1–3 0.509 0.182 0.0138 -0.739
(0.56) (0.19) (0.02) (-0.99)
Post-conﬂict year 4–6 -0.293 -1.898∗
(-0.37) (-2.36)
Post-conﬂict year 7–10 -0.326 -3.741∗∗
(-0.43) (-2.75)
Log population -0.387 -0.494 -0.295 -0.391
(-1.81) (-1.92) (-1.70) (-1.83)
Log infant mortality rate 0.0611 0.515 -0.126 0.250
(0.17) (1.38) (-0.52) (0.96)
1990s 21.56∗∗∗
-0.819 0.982∗
-0.754
(9.08) (-1.35) (2.25) (-1.69)
cons -22.49 -2.531 -1.693 -1.031
(.) (-0.89) (-0.86) (-0.45)
N 1002 1152
t statistics in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
Unit of observation: Country years at conﬂict or in post-conﬂict state (less than 10 years after end of conﬂict).
countries will receive PKOs, but Model 1 give some indications. First, both traditional and
transformational PKOs are about six times more likely to be initiated in countries with major
conﬂict (more than 1,000 battle deaths) than in conﬂicts that are less intense or just have
ended. The UN occasionally starts up PKOs in countries that have had up to three years
after conﬂict, but almost never after that.29 There is some indication that conﬂicts that have
lasted a year or more have a larger probability of attracting PKOs.
Secondly, PKOs are less frequent in large countries. This is particularly true for transformational
operations. The odds of PKO initiation in a country with 10 million inhabitants
is more than three times higher than in a country with 100 million inhabitants. This is also
evident from the list of all PKOs (Table A-4).
Thirdly, transformational PKOs are more likely in under-developed countries, but the
relationship is not very strong. A conﬂict country with an infant mortality rate at 100 (per
1,000 live births) is about twice as likely to receive PKOs as one with 20.
Finally, traditional operations often initiate after transformational ones, and vice versa.
Moreover, as evident from Figure 4, traditional PKOs were more frequent in the 1990s than
in the 2000s, whereas transformational operations became more numerous in the most recent
29
Estimates for the coeﬃcients for ‘Post-conﬂict year 4–6/7–10’ are typically smaller than –30, reﬂecting the
almost perfect absence of such cases. Given the estimation problems associated with such relationships we
opted not to present these results.
33
decade.
Model 2 – incidence – complements this picture by showing that PKOs also tend to
continue if the conﬂict remains at the major conﬂict level. The probability of continuation
decreases quickly over the post-conﬂict period.
A.3 List of peace-keeping operations
34
TableA-4:ListofUnitedNationspeace-keepingoperations,1970–
2009
AcronymMissionnameStartdateClosingdateCountries
Observermissions
UNDOFUnitedNationsDisengagementObserverForceJun-74PresentSyria
UNAVEMIUnitedNationsAngolaVeriﬁcationMissionIJan-89Jun-91Angola
ONUCAUnitedNationsObserverGroupinCentralAmericaNov-89Jan-92CostaRica,
ElSalvador,
Guatemala,Honduras
MINURSOUnitedNationsMissionfortheReferenduminWesternSaharaApr-91presentMorocco(WesternSahara)
UNAMICUnitedNationsAdvanceMissioninCambodiaOct-91Mar-92Cambodia
UNOMURUnitedNationsObserverMissionUganda-RwandaJun-93Sep-94Rwanda,Uganda
UNOMIGUnitedNationsObserverMissioninGeorgiaAug-93Jun-09Georgia
UNOMILUnitedNationsObserverMissioninLiberiaSep-93Sep-97Liberia
UNASOGUnitedNationsAouzouStripObserverGroupMay-94Jun-94Chad
UNMOTUnitedNationsMissionofObserversinTajikistanDec-94May-00Tajikistan
UNMOPUnitedNationsMissionofObserversinPrevlakaJan-96Dec-02Croatia,FederalRepublicofYugoslavia
MIPONUHUnitedNationsCivilianPoliceMissioninHaitiDec-97Mar-00Haiti
UNPSGUNCivilianPoliceSupportGroupJan-98Oct-98Croatia
Traditionalmissions
UNFICYPUnitedNationsPeacekeepingForceinCyprusMar-64PresentCyprus
UNIFILUnitedNationsInterimForceinLebanonMar-78PresentLebanon
UNGOMAP30
UnitedNationsGoodOﬃcesMissioninAfghanistanandPakistanMay-88Mar-90Afghanistan,Pakistan
UNAVEMIIUnitedNationsAngolaVeriﬁcationMissionIIJun-91Feb-95Angola
UNOSOMIUnitedNationsOperationinSomaliaIApr-92Mar-93Somalia
UNAMIRUnitedNationsAssistanceMissionforRwandaOct-93Mar-96Rwanda
UNAVEMIIIUnitedNationsAngolaVeriﬁcationMissionIIIFeb-95Jun-97Angola
UNPREDEPUnitedNationsPreventiveDeploymentForceMar-95Feb-99Macedonia
UNCROUnitedNationsConﬁdenceRestorationOperationinCroatiaMay-95Jan-96Croatia
UNSMIHUnitedNationsSupportMissioninHaitiJul-96Jul-97Haiti
MINUGUAUnitedNationsVeriﬁcationMissioninGuatemalaJan-97May-97Guatemala
MONUAUnitedNationsObserverMissioninAngolaJun-97Feb-99Angola
UNTMIHUnitedNationsTransitionMissioninHaitiAug-97Dec-97Haiti
UNOMSILUnitedNationsObserverMissioninSierraLeoneJul-98Oct-99SierraLeone
30
UNGOMAPiscodedasactiveonlyinAfghanistan
35
Multi-dimensionalmissions
UNTAGUnitedNationsTransitionAssistanceGroupApr-89Mar-90Namibia
ONUSALUnitedNationsObserverMissioninElSalvadorJul-91Apr-95ElSalvador
UNTACUnitedNationsTransitionalAuthorityinCambodiaMar-92Sep-93Cambodia
ONUMOZUnitedNationsOperationinMozambiqueDec-92Dec-94Mozambique
UNMIBHUnitedNationsMissioninBosniaandHerzegovinaDec-95Dec-02Bosnia&Herzegovina
MINURCAUnitedNationsMissionintheCentralAfricanRepublicApr-98Feb-00CentralAfricanRepublic
UNMISETUnitedNationsMissionofSupportinEastTimorMay-02May-05Timor-Leste
UNMITUnitedNationsIntegratedMissioninTimor-LesteAug-06PresentTimor-Leste
Enforcementmissions
UNPROFORUnitedNationsProtectionForceFeb-92Mar-95Croatia,
Bosnia&Herzegovina,
Macedonia
UNOSOMIIUnitedNationsOperationinSomaliaIIMar-93Mar-95Somalia
UNMIHUnitedNationsMissioninHaitiSep-93Jun-96Haiti
UNTAESUnitedNationsTransitionalAdministrationfor
EasternSlavonia,BaranjaandWesternSirmiumJan-96Jan-98Croatia
UNMIKUnitedNationsInterimAdministrationMissioninKosovoJun-99PresentKosovo
UNTAETUnitedNationsTransitionalAdministrationinEastTimorOct-99May-02Timor-Leste
UNAMSILUnitedNationsMissioninSierraLeoneOct-99Dec-05SierraLeone
MONUCUnitedNationsOrganizationMissioninthe
DemocraticRepublicoftheCongoNov-99PresentDemocraticRepublicofCongo
UNMILUnitedNationsMissioninLiberiaSep-03PresentLiberia
UNOCIUnitedNationsOperationinCˆoted’IvoireApr-04PresentCoted’Ivoire
MINUSTAHUnitedNationsStabilizationMissioninHaitiJun-04PresentHaiti
ONUBUnitedNationsOperationinBurundiJun-04Dec-06Burundi
UNMISUnitedNationsMissionintheSudanMar-05PresentSudan
UNAMIDAfricanUnion-UnitedNationsHybridOperationinDarfurJul-07PresentSudan
MINURCATUnitedNationsMissionintheCentralAfricanRepublicandChadSep-07PresentCentralAfricanRepublic,Chad
Missionsassociatedwithinternationalconﬂictsthatareexcludedfromouranalysis
UNEFIISecondUnitedNationsEmergencyForceOct-73Jul-79Egypt
UNIIMOGUnitedNationsIran-IraqMilitaryObserverGroupAug-88Feb-91Iran,Iraq
UNIKOMUnitedNationsIraq-KuwaitObservationMissionApr-91Oct-03Iraq,Kuwait
UNMEEUnitedNationsMissioninEthiopiaandEritreaJul-00Jul-08Ethiopia,Eritrea
36
A.4 Additional estimation tables
37
Table A-5: Estimation results, determinants of conﬂict, baseline model without PKO variables
1 2
Log PKO expenditures
Log expenditures squared
Traditional PKO
Transformational PKO
Minor conﬂict t-1 2.441 (1.32) 3.043 (0.98)
Major conﬂict t-1 0.100 (0.03) 4.098 (1.02)
Log time in status c0 -1.240∗∗∗
(-14.72) -1.573∗∗∗
(-9.81)
Log time in status c1 1.146∗∗∗
(9.97) 0 (.)
Log time in status c2 0 (.) 1.260∗∗∗
(7.41)
Conﬂict in neighborhood 0.649∗∗
(2.88) 0.803 (1.67)
NC * minor conﬂict at t-1 -0.613∗
(-2.32) -0.612 (-1.18)
NC * major conﬂict at t-1 -1.297∗∗∗
(-3.37) -1.219∗
(-2.13)
NC * time in status c0 -0.133 (-1.40) -0.214 (-1.02)
Log population 0.346∗∗∗
(3.50) 0.212 (1.21)
Population * minor conﬂict at t-1 -0.00127 (-0.01) 0.199 (0.98)
Population * major conﬂict at t-1 -0.0350 (-0.18) 0.160 (0.67)
Population * time in status c0 -0.0485 (-1.18) 0.0638 (0.79)
Log infant mortality rate 0.0131 (0.04) 1.979∗∗
(3.01)
IMR * minor conﬂict at t-1 -0.247 (-0.61) -1.820∗
(-2.48)
IMR * major conﬂict at t-1 -0.197 (-0.37) -2.050∗∗
(-2.60)
IMR * time in status c0 0.265∗
(2.00) -0.293 (-1.07)
Youth bulge 0.00915 (0.24) -0.146∗
(-1.96)
Youth * minor conﬂict at t-1 0.00109 (0.02) 0.176∗
(2.09)
Youth * major conﬂict at t-1 0.105 (1.44) 0.246∗
(2.52)
Education -1.661∗
(-1.99) 1.172 (0.94)
Education * minor conﬂict at t-1 -0.0122 (-0.01) -1.477 (-0.99)
Education * major conﬂict at t-1 2.391 (1.61) -1.856 (-1.06)
Education * time in status c0 0.445 (1.36) -0.247 (-0.42)
Log IMR in neighborhood -0.294 (-1.37) -0.0996 (-0.33)
Education in neighborhood -0.497 (-0.82) -0.934 (-1.08)
loi 0 (.) 0 (.)
loic1 0 (.) 0 (.)
loic2 0 (.) 0 (.)
lois0 0 (.) 0 (.)
Youth * time in status c0 -0.00783 (-0.47) 0.0461 (1.35)
Eastern Europe -0.423 (-1.14) 0.298 (0.54)
Western Africa -0.144 (-0.59) -1.884∗∗∗
(-3.49)
Rest of SS Africa 0.0742 (0.45) -0.0652 (-0.28)
Constant -3.327∗
(-2.53) -8.551∗∗∗
(-3.32)
N 5942
ll -1525.0
t statistics in parentheses
∗
p < 0.05, ∗∗
p < 0.01, ∗∗∗
p < 0.001
38