Do Humanitarian Military Interventions Foster Economic Growth?

Abstract

Humanitarian military interventions (HMIs) are undertaken to end violent conflicts and establish political and economic stability in targeted countries. Their frequency has risen since the end of the Cold War and the United Nations’ adoption of the Responsibility to Protect (R2P) doctrine. Previous research shows neutral HMIs - targeting all perpetrators - reduce conflict intensity, while biased HMIs-supporting one side-escalate violence. This study empirically tests the effects of these interventions on long-run economic growth in targeted countries from 1960 to 2019. The findings suggest that countries receiving neutral interventions experience approximately 14 percentage points higher cumulative GDP per capita growth over 5 years compared to conflict-affected countries that do not. No statistically significant evidence is found that biased HMIs contribute positively to economic growth. Furthermore, the positive effect of neutral HMIs on growth diminishes as prior conflict duration increases. These results highlight the importance of impartiality in HMIs for sustainable economic recovery.

Keywords

humanitarian interventions military interventions economic growth violent conflict

Introduction

Humanitarian military interventions (HMIs) are launched with the intent of terminating violent conflicts and establishing political and economic stability in the targeted countries. While these interventions fall within the broader category of third-party foreign military interventions (MIs), they are uniquely characterised by the simultaneous presence of two specific factors, setting them apart from other non-HMIs. First, HMIs are launched in the context of ongoing violent emergencies, and the intervener explicitly articulates the intention of “saving the strangers”. Second, these interventions involve threats or the deployment of military force in combat operations to stop violent atrocities (Gromes and Dembinski 2019). Examples include the Arab League’s intervention in Lebanon (1976–79), the UN’s intervention in Bosnia and Herzegovina (1993–95), the French intervention in Rwanda (1994), and the UN intervention in Burundi (2001–08). In each of these instances, the targeted countries were already experiencing violent conflict before the intervention began. Also, on each occasion, the objective of stopping the atrocities and establishing peace using military means was explicitly stated by the intervener. In contrast, interventions may be launched in countries without any active conflict. In some cases, the intention to stop atrocities may be absent altogether, or it may not form the primary objective of the intervening party. For instance, Kuwait suffered no violent conflict before the Iraqi intervention in 1990, which was motivated by economic and strategic rather than humanitarian interests (Pickering and Kisangani 2009). This and similar interventions can be referred to as non-HMIs. Moreover, not all interventions entail the use or threat of deploying force in combat operations to influence the security environment, such as evacuation missions, which are categorised as military interventions (Pearson and Baumann 1993; Pickering and Kisangani 2009).¹ In short, HMIs only include interventions in which violent force is either actively deployed or a threat is made for its use to alter security dynamics and establish peace in a country suffering from an ongoing violent conflict.

Foreign military interventions can directly shape political and economic institutions in target countries (Vishwasrao et al. 2018). Economic transformation is sometimes an explicit objective; even when not, using military force changes the security environment and may have economic consequences. In HMIs, improving security is central, so if violence falls, economic benefits may follow. Conversely, worsening security after intervention may contract economic output.

This study examines the impact of HMIs on economic growth in targeted countries from 1960 to 2019. The empirical analyses utilise Gromes and Dembinski’s (2019) database, which encompasses all cases of HMIs in the post-Second World War period. HMIs are defined as the use or threat of force across state borders by a state or group of states with the declared intention of saving strangers threatened by a violent emergency that has resulted in at least 25 fatalities in one calendar year. For robustness, the consolidated databases of Pearson and Baumann (1993) and Pickering and Kisangani (2009), covering the 1946–2006 period, are also employed in supplementary analyses. This research builds upon the author’s prior work demonstrating that biased HMIs escalate conflict, while neutral HMIs have a pacifying effect (Saeed 2022). This analysis examines how different types of HMIs influence short- and long-run economic growth in targeted countries.

The findings indicate that neutral HMIs lead to approximately a 14 percentage point increase in 5-year cumulative GDP per capita growth following intervention. Interventions implemented earlier tend to produce greater benefits than those initiated after extended periods of conflict. The positive impact of neutral HMIs peaks at 3.71 years of prior conflict duration, after which the rate of gains declines. No statistically significant evidence is found that biased HMIs contribute positively to long-run economic growth. Thus, the evidence suggests that neutral HMIs not only pacify conflicts (Saeed 2022) but also generate positive long-run economic outcomes.

The remainder of the paper is organised as follows: The next section reviews the historical background of humanitarian military interventions (HMIs). The following section develops the theoretical framework linking HMIs to economic growth. The subsequent section presents the empirical analyses. The final section concludes with a summary of the main findings and their implications.

Humanitarian Military Interventions

While the term “humanitarian military interventions” has a modern origin, the concept and practice have a long history. The great Roman poet Virgil exhorted the Romans to crush the proud and impose peace in foreign lands (Osborne 2007, 4). In The Rights of Magistrates, Theodore Beza lends support to using force to depose tyrants (Trim 2011). Edmund Burke campaigned to depose the most infernal tyranny of the Jacobins (Simms 2011, 103). In his A Few Words on Non-Intervention, John Stuart Mill questioned the idea of non-intervention.

The contemporaneous support for HMIs is grounded in the normative perspective in international law that assigns primacy to the rights of humanity over those of nation-states. According to this pro-interventionist view, the protection of civilian lives must be prioritised over the respect for the territorial integrity of nation-states (Smith 1994; Walzer 2006). Suppose a government is unable to stop or, worse, is complicit in atrocities against civilians. In that case, the international community must intervene to protect the lives of the “strangers” and hold the perpetrators accountable.

The anti-interventionists disagree and propose that the idea of non-intervention, inspired by the Treaty of Westphalia (1,648), is fundamental to the stability of the international political order. The political will of nation-states should be the principal determinant of international law (Seybolt 2007). While the idea of intervention has normative appeal, the international political order would collapse if intervention were to become a norm. This argument rests on an implicit assumption that the instability caused by intervention entails greater costs than inaction.

Since the end of the Cold War, the number of HMIs has increased significantly. Gromes and Dembinski (2019) classify 5 interventions as HMIs during 1960–85.² The number of interventions increased to 35 in the latter period (see Figure 1). The post-Cold War period was marked by several significant developments, including the collapse of the Soviet Union, the transition to a unipolar world order, and, more importantly, the adoption of the doctrine of the Responsibility to Protect (R2P) by the United Nations.³

Figure 1.

Humanitarian military interventions 1960–2019.

Countries subjected to these interventions experienced varying degrees of violent political conflict. To analyse conflict dynamics in relation to interventions, we exploit variation in average conflict intensity over the 3-year pre-intervention period, the intervention years, and the 5-year post-intervention period. Conflict intensity is an ordinal variable measured on a scale from 0 to 2, where 0 = less than 25 battle-related deaths, 1 = between 25 and 999 battle-related deaths, and 2 = 1,000 or above battle-related deaths (Kreutz 2010). In 20 of the 39 HMIs, the intervener acted neutrally and targeted all perpetrators of violence. The remaining 19 interventions can be categorised as biased HMIs, in which the intervener acted discriminatorily, targeting either the government or the rebel forces. Neutral and biased HMIs lasted for 73 and 70 years, respectively (Gromes and Dembinski 2019) (Online appendix Table AP.1). In both neutral and biased HMIs, the conflict situation worsened during the intervention years, albeit more so in the biased intervention. However, trends in the post-intervention period clearly show a reduction in conflict intensity in countries experiencing neutral interventions. On the other hand, countries experiencing biased interventions continue to suffer from high-intensity conflict, with average conflict intensity exceeding 1 for most of the period following the intervention years. The key takeaway from these trends is that the conflict environment improves following neutral interventions, whereas the opposite trend is observed with biased interventions.

Additionally, Saeed (2022) provides evidence from multivariate analyses that neutral HMIs pacify conflict, whereas biased interventions escalate it. The long-run estimates show that country-years experiencing neutral HMIs are 2.8–3.8 percentage points less likely to experience minor conflict (25–999 battle deaths) and 2.0–2.7 percentage points less likely to experience war (1,000+ battle deaths) relative to conflict-affected countries without such interventions. These results confirm that the post-intervention improvements observed in the trends are not merely descriptive but persist after accounting for confounding factors. Since the security environment directly impacts economic outcomes, we separately account for neutral and biased HMIs in the model for economic growth.

The next section provides a theoretical framework for understanding the impact of neutral and biased HMIs on economic growth.

Theoretical Perspective

There are two distinct channels through which a third-party military intervention can impact economic development in the targeted country. The first channel involves direct spending, including local expenditures generated by the intervention, such as logistics, services, and employment. This spending can foster growth by increasing demand for locally produced services and manufacturing products (Bove et al. 2022). Additionally, the development of new infrastructure and supporting services, like security, may be necessary to sustain a foreign military presence. Resources may also be acquired locally. Consumption spending by military personnel and their supporting staff, including some locally hired workers, helps stimulate growth. This direct channel may also include funding for reconstruction, humanitarian aid, and governance support. For instance, when the UK intervened militarily in Sierra Leone in 2000, it pledged £5.5 million to support humanitarian aid, the UN mission, and the wider peace process.⁴ However, the direct effects are likely limited or absent due to several factors. Not all HMIs involve ground troop deployment; interventions in Iraq (1992, 2014), Libya (2011), and Kosovo (1999) didn’t deploy ground forces. Additionally, foreign militaries often depend on external firms for logistical support (Rolandsen 2015).⁵ Most missions are of limited duration, with a median length of approximately 1 year, and only rarely do they extend for prolonged periods. Some operations, such as the French intervention in Côte d’Ivoire (2011) and NATO’s intervention in Bosnia and Herzegovina (1995), lasted about 3 months, while a small number of exceptional cases-most notably NATO’s intervention in Afghanistan (2003) and the UN intervention in the Democratic Republic of Congo (2000) -spanned 11 and 13 years. Thus, while a direct effect cannot be fully ruled out, they are unlikely to be a key determinant of long-run economic growth, as their influence typically does not extend beyond the life of the mission.

The second channel, perhaps more enduring in its impact, is the intervention’s influence on the local security environment (Bove et al. 2022). Economic prosperity is directly linked to peace and security, as violence threatens life, property, and contracts. If a third-party intervention reduces the intensity of violence, it can stimulate economic development by boosting consumer and investor confidence. A reduction in violence may also enhance expected profit margins by lowering transaction costs of doing business. Improved security conditions can facilitate greater domestic and foreign investment (Bussmann 2010), enhance labour mobility, and restore disrupted supply chains. Reduced conflict also allows governments to reallocate resources away from military expenditures toward productive public goods such as infrastructure, health, and education, further supporting long-term growth (Besley and Persson, 2011; Collier, 1999). Moreover, sustained reductions in violence strengthen institutions and contract enforcement, which are central to economic development and private-sector expansion (North 1990; Rodrik et al., 2004).

Our empirical strategy is to distinguish between the effects of neutral and biased HMIs on economic growth. Using the same Gromes and Dembinski’s (2019) database employed in the current analyses, Saeed (2022) has also shown that neutral HMIs lower conflict intensity, whereas biased HMIs result in conflict escalation. Suppose that the decision to continue fighting is a function of the cost of violence and the likelihood of victory. Neutral HMIs, which involve indiscriminate action against all perpetrators of violence (Gromes and Dembinski 2019), can lead to an increase in the cost of violence and lower odds of victory for all parties. These changes in turn are likely to lower expected payoffs from fighting and can encourage all belligerents to opt for a negotiated solution (Kydd and Straus 2013). Hence, we can expect neutral HMIs to lower violence levels, thereby positively impacting economic growth. As already noted in Figure 2, conflict intensity declines and remains low following neutral interventions, while it stays high after biased interventions. In short, neutral interventions improve the security environment, as also supported by empirical findings in Saeed (2022). We argue that their neutral mandates are central to reducing conflict intensity and facilitating institutional improvement by increasing the costs of violence and encouraging negotiated settlements. Over the longer run, improved security can also support growth through indirect channels such as infrastructure restoration, expanded public goods provision, and institutional stabilisation. Importantly, the long-term positive effects are likely to accumulate. Improvements in the security environment create conditions in which institutional reforms, infrastructure restoration, and public goods provision can gradually build, reinforcing one another and generating sustained economic benefits. By contrast, short-run effects may be weaker or less consistent, as adjustment costs, residual insecurity, and uncertainty in the immediate post-intervention period can dampen investment and economic activity. As these transitional constraints fade, the compounding effects of improved security and institutional stabilisation become increasingly visible, explaining why the growth-enhancing impact of neutral interventions is more pronounced in the long run.

Figure 2.

Conflict intensity and humanitarian military interventions.

Hypothesis 1

All else equal, neutral HMIs have a positive impact on short-run and long-run economic growth.

The biased HMIs are launched in support of a particular party in conflict (Gromes and Dembinski 2019). These interventions are likely to have asymmetric effects on the cost of violence and the likelihood of victory for both parties. The supported party is likely to experience a reduction in the cost of violence and an increase in the likelihood of victory. As the prospects for victory improve and the cost of achieving it is subsidised through external support, this can create “perverse incentives” for the supported party to continue fighting (Grigoryan 2010).

On the other hand, the targeted party will experience a rise in the cost of violence and a fall in the likelihood of victory. Viewed in isolation from the strategy of the supported party, the rational strategic behaviour of the targeted party is to discontinue fighting. However, in situations of conflict, strategic choices are not made in isolation and are directly influenced by the choices of the adversary. In the case of neutral HMIs, when all parties come under assault and face rising costs of violence and diminishing prospects for total victory, a rational strategy is to cease fire and redirect resources either to strengthen their bargaining position in negotiations or to preserve their capacity to fight in the future. However, once external interventions tend to widen the asymmetries of power between the belligerents, the targeted party is more likely to escalate violence. Inaction in this case will inevitably lead to annihilation and destruction of all military resources, eliminating any prospect of future returns. In such a situation, the weaker are rationally motivated to fight harder (Hirshleifer 1991). Hultman (2007) has shown that mounting pressure that results from a major assault can motivate rebels to increase atrocities against civilians to increase the cost of violence for the government.

Hence, this leads us to our second hypothesis.

Hypothesis 2

All else equal, biased HMIs have a negative impact on short-run and long-run economic growth.

It is important to note that any direct effects, if present, are confined to the duration of the mission. The indirect effects of HMIs on growth are mediated by the changing conflict dynamics. These effects can be enduring and more relevant in determining long-run economic growth.

The empirical literature examining the impact of military interventions on economic growth in targeted countries remains very limited. Vishwasrao et al. (2018) analysed 214 countries covering the 1950–2013 period and found that transformative occupation – involving the constructive transformation of political and economic institutions in the targeted countries - has a positive impact on economic growth. As per the definition we are following, not all transformative occupations have a humanitarian character if they are not launched in the context of ongoing violent emergencies. A growing body of literature in peacekeeping studies shows that these missions have a positive impact on economic growth. Beber et al. (2019) found a positive impact of peacekeeping missions on local economies. Caruso et al. (2016) observed that a 10 percent increase in the size of the UN mission led to about a 600-tonne increase in agricultural productivity. HMIs differ from peacekeeping missions on several accounts. First, peacekeeping mission deployment requires the consent of the parties, and the force is normally used for defence purposes, which is clearly not the case in HMIs. Peacekeeping missions are also mostly deployed in post-conflict environments to maintain peace, whereas HMIs are primarily an offensive measure to stop atrocities.⁶

Empirical Analyses

Figure 3 illustrates the trends in average GDP per capita for the targeted countries, with the year preceding the intervention used as the base year. The index is developed using the following expression.

G D P p e r c a p i t a i n d e x = (\frac{{G D P p e r c a p i t a}_{i}}{{G D P p e r c a p i t a}_{b}}) * 100

where

{G D P p e r c a p i t a}_{i}

is the average GDP per capita in year i and

{G D P p e r c a p i t a}_{b}

is the average GDP per capita in the base year.

Figure 3.

Cumulative growth in GDP per capita and duration of violence prior to neutral HMIs.

The index compares average GDP per capita growth relative to the pre-intervention year across the targeted countries. The trend shows that conflict-affected countries experiencing neutral HMIs experienced a cumulative increase of almost 30 percentage points in their average GDP per capita compared to the pre-intervention year. Initially, the average GDP per capita falls by about 4 percentage points during the intervention years. Saeed (2022) shows that average conflict intensity dips during the neutral HMIs compared to the pre-intervention level. However, it takes 1 year for a significant drop in average conflict intensity in conflict-affected countries without intervention. Also, while the intervention may be successful in curbing some level of violence immediately, the initial climate of uncertainty that accompanies foreign military involvement is likely to diminish gradually. This may explain the simultaneous occurrence of neutral HMIs, a modest drop in violence, and a contraction of about 4 percentage points in average GDP per capita. However, in the period following neutral HMIs, average GDP per capita shows sustained growth, which coincides with declining levels of conflict intensity during these years, as observed by Saeed (2022).

Countries that experience biased HMIs show a significant contraction of over 13 percent in average GDP per capita during the intervention year as compared to the pre-intervention level. These countries also experienced a sharper spike in average conflict intensity during intervention years compared to the pre-intervention period (Saeed 2022). Following the intervention years, the growth trajectory is volatile, and 5 years after HMIs, the average GDP per capita is about 3 percentage points below the pre-intervention year. The years following biased interventions also show a higher average conflict intensity than those following neutral HMIs. Hence, the long-run trends show that neutral HMIs, which lower conflict intensity, are correlated with higher growth in GDP per capita, whereas biased HMIs, which escalate conflict intensity, are correlated with unstable growth trajectories.

The trends in Figure 3 are illuminating and appear to support Hypotheses 1 and 2. However, these average estimates are based on a large, heterogeneous sample, and some outliers may be driving these trends. One way to address this problem is to assign a weight of 1 to a country-year experiencing economic contraction (negative growth) and 0 otherwise. The proportion of country-years experiencing economic contraction is estimated for the 5 years preceding the intervention, all intervention years, and the 5 years following the intervention.

In countries that experience neutral HMIs, we observe that, relative to the pre-intervention period, the proportion of country-years experiencing economic contraction drops from 55 to 18 percent during the 5-year post-intervention period (Table 1). The difference is statistically significant at the 1 percent level. In the case of biased HMIs, no significant difference is observed in the proportion of country-years experiencing economic contraction between the post-intervention and pre-intervention periods.

Table 1.

Likelihood of economic contraction

	5 years pre-intervention	Intervention years	5 years post-intervention
Neutral	0.55	0.39*	0.18***
Biased	0.40	0.40	0.30

Economic contraction is defined as an episode of negative growth. ****indicates statistical significance at 1 percent. The Z test is applied for testing the difference in proportion. ***, ** and * indicate statistical significance at 1, 5 and 10 percent.

These trends provide further preliminary evidence in support of hypotheses 1 and 2: neutral HMIs contribute to economic growth, whereas biased HMIs lower economic growth. However, these trends only indicate statistical correlations and may be driven by factors independent of the HMIs. To isolate the effects of HMIs on both short-run and long-run economic growth, we develop a multivariate model that controls for other important determinants of growth.

The econometric model is specified as follows.

y_{i t} - y_{i t - 1} = β_{0} + β_{1} {B i a s e d}_{i t} + β_{2} {N e u t r a l}_{i t} + {X β}_{i} + \emptyset_{i} + γ_{t} + ε_{i t}

where

y_{i t}

= logged GDP per capita in country i year t.

X

is a matrix of control variables,

\emptyset

are country fixed effects and

γ_{t}

time dummies.

The coefficients of interest in the model are $β_{1}$ and $β_{2}$ , and our model must include all the important determinants of growth, particularly those likely to be correlated with HMIs. The decision to intervene is likely to be driven by a range of factors, some of which, such as local political, bureaucratic, and military considerations in the intervening nation, are unrelated to economic growth in the targeted country. However, other important factors may drive both intervention and economic growth in the targeted country and hence must be included in the model. For instance, conflict prevalence in the targeted country may prompt foreign intervention and is likely to affect domestic economic growth. In fact, in the case of HMIs, the intention to stop atrocities is explicitly stated by the intervener (Gromes and Dembinski 2019). Our conflict measure is the conflict intensity index, which measures conflict intensity on an ordinal scale from 0 to 2, corresponding to less than 25, between 25 and 1,000, and 1,000 or more battle-related deaths, respectively (Kreutz 2010). Violence in neighbouring states can spill over across the border (Forsberg 2014) and may also harm domestic economic growth (Murdoch and Sandler 2002). We control for the magnitude of violence in neighbouring states using a measure developed by Marshall (2019), which aggregates ordinal scores for interstate and social violence in neighbouring countries. Interventions may follow punitive actions such as sanctions, which have been documented to hurt economic growth Gutmann et al. 2023). We include a dummy variable indicating whether a sanction regime is in place, using the data from Felbermayr et al. (2020). To control for the targeted State’s material capability-which is likely to affect domestic economic growth and foreign powers’ decision to intervene-we include a measure developed by Singer et al. (1972) and Singer (1987).⁷

We also control for military spending, inflation, and total population, all of which are likely to determine economic growth. While the literature reports mixed findings, recent empirical studies using robust instrumental variable strategies suggest a strong negative effect of military spending on economic growth (d'Agostino et al. 2019; Saeed 2023). As an additional measure of militarisation, the value of arms imports is also included in the model (Stockholm International Peace Research Institute 2021). Empirical evidence indicates that higher inflation rates are generally associated with slower economic growth, as sustained inflation erodes investment and economic stability (Barro 1995). The data source for all these variables is The World Bank (2021). Ethnic fractionalization is measured as the probability that two randomly selected individuals are not from the same ethnic group (Drazanova 2019). High levels of ethnic fractionalization are often associated with lower economic growth, as heterogeneous societies may experience reduced public goods provision, political instability, and higher transaction costs (Easterly and Levine 1997).

Finally, we also include time dummies for important events that are likely to have affected economic growth worldwide, such as recession that followed the oil embargo of early 1970s (1974–75) the Iran-Iraq war (1980–88), the Gulf War (1990–91), the post-Cold-War period (1991-to date), and recessions during 1974–75 and 2008–09. We include fixed effects to account for systematic differences in baseline growth levels across countries and to control for time-invariant factors. The descriptive statistics on all these variables are reported in the online appendix Table AP.2. The analyses cover the period from 1960 to 2019.

Since the empirical analyses are based on a large panel data set, we expect heteroscedasticity and cross-sectional dependence, both of which can affect the standard errors. We address these issues by estimating Driscoll and Kraay (1998) standard errors, which are robust to heteroscedasticity, autocorrelation, and cross-sectional dependence.

Table 2 reports the results for contemporaneous effect models. We focus on the effects of intervention variables in the main text, while the full set of results are reported in the online appendix Table AP.3. We use the full sample as well as sub-samples of non-high-income⁸ conflict-affected countries⁹ and active-conflict countries¹⁰ to estimate the coefficients. This approach allows us to assess the stability and robustness of the results, particularly given that interventions are relatively infrequent events. We have a total of 73 country-years for neutral interventions and 70 country-years for biased HMIs.

Table 2.

Contemporaneous impact

Variables	2.1	2.2	2.3	2.4
Variables	Full sample	Non-high income	Conflict-affected countries	Active conflict countries
Biased interventions	−0.028* (0.08)	−0.027 (0.10)	−0.029* (0.09)	−0.002 (0.78)
Neutral interventions	0.022 (0.11)	0.021 (0.11)	0.027*** (0.0)	0.024** (0.01)
Observations	4,462	3,183	2,758	821
Number of countries	139	108	86	61
Other controls	Yes	Yes	Yes	Yes
Within R2	0.06	0.05	0.05	0.12
F-stat	13.73	12.28	12.11	8.33

Parentheses contain p value where ***p < 0.01, **p < 0.05, *p < 0.1. Driscoll and Kraay (1998) standard errors in all models.

Dependent Variable: GDP Per Capita Growth Rate from year t–1 to year t.

As expected, the estimated coefficient for biased interventions is negative in all models. It is statistically significant at the 10 percent level in Models 2.1 (full sample) and 2.3 (conflict-affected countries sample), and borderline significant in Model 2.2 (non-high-income countries sample). The coefficient size remains between 0.027 and 0.029 across Models 2.1-2.3. This suggests that the countries targeted with biased interventions experience around 2.8 to 3 percentage points lower economic growth as compared with the baseline samples.

The coefficient for neutral interventions is positive in all models and statistically significant in Models 2.3 and 2.4, suggesting around 2.4-2.7 percentage points higher economic growth in the targeted countries compared to the baseline sample of countries that suffer from conflict but experience no neutral intervention.¹¹

Regarding the control variables, we observe that inflation has a statistically significant negative impact on economic growth, suggesting that higher price instability tends to hinder output growth. Similarly, increased military spending is associated with lower growth, suggesting that resources diverted to defence reduce investment in productive sectors. Both domestic and regional conflicts also exert a significant negative effect on growth, reflecting the disruptive consequences of violence on economic activity, trade, and investment. These findings are consistent with prior research highlighting the adverse effects of macroeconomic instability and conflict on development.

Time dummies capturing major global shocks, such as the Iran-Iraq War and the 2008 global recession, also have statistically significant negative effects on economic growth. These dummies account for the impact of extraordinary events that affected many countries simultaneously, helping to isolate the effect of interventions and other control variables from broader, economy-wide shocks. Their significance underscores how large-scale conflicts or global recessions can disrupt trade, investment, and overall economic activity, slowing growth across affected countries.

The trends in the annual frequency of humanitarian interventions showed that most of these interventions took place after 1985 (see Figure 1)- a period that coincides with the winding down of the Cold War. The dissolution of the Soviet Union gathered pace during the mid-1980s, and the international political order began transitioning from a bipolar to a unipolar world order. It was also during the post-Cold War period that the notion of humanitarian intervention began to gain legitimacy, culminating in the adoption of the Responsibility to Protect (R2P) doctrine in 2005. We extend the analysis by focusing on the 1985–2019 period to reduce heterogeneity, with full results presented in Appendix Table AP.4. Across all models, biased interventions consistently show a negative and statistically significant effect on economic growth, with countries experiencing such interventions growing on average 4.2 percentage points less than those without. Evidence for neutral HMIs is weaker and less consistent, but the negative impact of biased interventions is strengthened in the post-1985 sample.

Overall, the results from contemporaneous effects models (Tables 2 and AP.4) seem to suggest that biased interventions harm economic growth. This aligns with the escalatory effect of these interventions on conflict intensity observed by Saeed (2022). While the effect of neutral HMIs appears to be positive, the evidence for statistical significance is weak. The weak contemporaneous significance may reflect a lagged intervention effect, whereby conflict pacification precedes and gradually translates into improvements in economic growth rather than producing immediate economic responses.

Long-Run Effects

We now estimate the long-run effects of neutral and biased HMIs on economic growth. We observe contributions to cumulative growth over 5 years following the HMIs. The econometric model is specified as follows.

y_{i t + k} - y_{i t - 1} = β_{0}^{k} + β_{1}^{k} {B i a s e d}_{i t} + β_{2}^{k} {N e u t r a l}_{i t} + X β_{i}^{h} + \emptyset_{i}^{k} + γ_{t}^{l} + ε_{i t}^{k}

where

y_{i t + k}

= logged GDP per capita in country i, year t = 0 and year k = 1 to 5. X is a matrix of control variables,

\emptyset_{i}^{k}

are country fixed effects and

γ_{t}^{l}

are time dummies.

Hence, a total of 5 models are estimated to measure the impact of biased and neutral HMIs on cumulative GDP per capita growth over 5 years following the intervention.

Table 3 presents results on the long-run impacts of biased and neutral interventions. We estimate models using samples of conflict-affected and non-high-income countries as well as those experiencing ongoing conflict. The estimated coefficient for biased intervention is statistically insignificant in all the models (Models 3.1–3.7). Saeed (2022) shows that biased HMIs are followed by escalating conflict levels, and the present results similarly indicate no positive effect on long-run growth. The results are also consistent with the trends observed in Figure 3 and Table 1. On the other hand, the coefficients for neutral interventions are positive and statistically significant at the 5 and 1 percent levels. The impact of neutral HMIs snowballs from 5 percentage points to 14 percentage points over a 5-year post-intervention period. Despite some differences in the size of the coefficients, the positive impact is consistently observed in the estimations based on samples of non-high-income countries (Model 3.6) and active conflict countries (Model 3.7).

Table 3.

Cumulative impact

Variables	3.1	3.2	3.3	3.4	3.5	3.6	3.7
	K = 1	K = 2	K = 3	K = 4	K = 5	K = 5	K = 5
	Conflict-affected					Non- high income	Active conflict
Biased interventions	−0.021 (0.29)	−0.005 (0.82)	−0.002 (0.93)	0.004 (0.97)	0.018 (0.66)	0.022 (0.51)	0.021 (0.51)
Neutral interventions	0.050*** (0.00)	0.071** (0.01)	0.097** (0.01)	0.123*** (0.00)	0.141*** (0.00)	0.148*** (0.00)	0.116*** (0.00)
Observations	2,758	2,758	2,758	2,758	2,758	3,183	821
Number of countries	86	86	86	86	86	108	61
Other controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Within R2	0.07	0.09	0.10	0.10	0.10	0.08	0.14
F-stat	16.79	20.92	48.11	60.68	40.23	68.78	53.62

Parentheses contain p value where ***p < 0.01, **p < 0.05, *p < 0.1. Driscoll and Kraay (1998) standard errors in all models.

Dependent Variable: GDP Per Capita Growth Rate from year t–1 to year t + k.

The trend in Figure 3 shows that countries with neutral HMIs experience a 30 percent points increase in average GDP per capita 5 years after the intervention. The regression results suggest that about 46 percent of that growth (14/30) can be attributed to neutral HMIs.

We also estimate the long-run impacts after restricting the time period to the post-1985 period. The results are presented in the online appendix Table AP.5. The results for neutral HMIs are quite consistent. The coefficients are statistically significant in all the models, and the contribution of neutral HMIs to the 5-year cumulative growth is 10.6 percentage points. As for the biased HMIs, the estimated coefficients are negative but mostly not statistically significant.

For an additional robustness check, we employ an entirely different dataset of Pearson and Baumann (1993) and Pickering and Kisangani (2009)- hereafter referred to as the International Military Interventions (IMI) database-that collectively covers the period of 1946–2005. The IMI database employs a more liberal definition of humanitarian intervention than Gromes and Dembinski (2019). For instance, unlike Gromes and Dembinski’s, the IMI database also codes post-conflict peacekeeping, relief efforts, and evacuation missions as humanitarian intervention. The IMI database does not code whether an active conflict preceded the launch of intervention; therefore, it is not possible to restrict IMI purely to cases where intervention followed an ongoing violent emergency. Despite these limitations, IMI represents the closest and most comprehensive alternative dataset available for assessing the robustness of our core findings.

The regression results based on the IMI database seem to reinforce the narrative drawn from the analyses using Gromes and Dembinski’s (2019) database, namely that neutral HMIs positively contribute to long-term economic growth. The coefficients in Model 4.2 suggest a 5 percentage point contribution to the 5-year cumulative growth. These coefficients are smaller than those reported in Table 3, but they still show a significant impact. Note that the IMI database is likely to include cases where there was no ongoing violent emergency, which may explain some of the reduction in the growth differential relative to the baseline sample. Again, we observe no statistically significant evidence of a positive impact of biased HMIs on long-run economic growth (Table 4).

Table 4.

Cumulative impact

Variables	Model 4.1K = 1	Model 4.2K = 5
Biased interventions (IMI Database)	−0.001 (0.88)	0.012 (0.62)
Neutral interventions (IMI Database)	0.023** (0.02)	0.050** (0.02)
Observations	2,240	2,240
Number of countries	80	80
Other controls	Yes	Yes
Within R2	0.09	0.12
F-stat	32.54	47.38

Parentheses contain p value where ***p < 0.01, **p < 0.05, *p < 0.1. Driscoll and Kraay (1998) standard errors in all models.

Dependent Variable: GDP Per Capita Growth Rate from year t–1 to year t + k.

Size Effects

The effects we have estimated so far are referred to as intervention-level effects. Our independent variables assume a value of 1 if a country-year experiences biased or neutral HMIs. However, we observe that the number of ground forces deployed in these interventions varies considerably. The distributional trends in Table 5 suggest that biased interventions typically entail a much larger commitment of ground troops than neutral interventions.

Table 5.

Distribution deployment of ground troops

Ground troops per 100,000 population	Neutral	Biased
Less than 100	52	24
Between 100 and 300	9	21
Between 300 and 500	3	12
Between 500 and 700	3	3
Between 700 and 900	0	1
Between 900 and 1,200	1	0

Author’s calculations using Gromes and Dembinski (2019) database.

To measure the size effect, we introduce the number of ground troops deployed (logged) as a proxy for intervention. Here again, we observe a statistically significant and positive impact of neutral HMIs on cumulative growth (Table 6). The estimate from Model 6.1 shows that a 1 percent increase in troop deployment contributes 4.9 percentage points to 5-year cumulative growth. We observe no statistically significant impact of biased HMIs on 5-year cumulative growth.

Table 6.

Cumulative impact (size effect)

Variables	6.1K = 5Conflict-affected	6.2K = 5Non-high income
Biased interventions (no. of troops per 100,000 logged)	−0.005 (0.62)	−0.004 (0.66)
Neutral interventions (no. of troops per 100,000 logged)	0.049** (0.02)	0.046** (0.01)
Observations	2,757	3,182
Number of countries	86	108
Other controls	Yes	Yes
Within R2	0.09	0.08
F-stat	44.19	46.10

Parentheses contain p value where ***p < 0.01, **p < 0.05, *p < 0.1. Driscoll and Kraay (1998) standard errors in all models.

Dependent Variable: GDP Per Capita Growth Rate from year t–1 to year t + k.

Prior Duration of Conflict

Next, we test whether the duration of violence before the start of intervention affects the subsequent effects of these interventions on economic growth. Figure AP.1 presented in the online appendix plots the cumulative increase in GDP per capita (where y is logged GDP per capita) on the duration of violence before the onset of neutral HMIs. The trend appears to be non-linear, and the appropriate fit is quadratic. At first, growth is positive and rising; however, after nearly 4 years of prior conflict leading up to the intervention, the trend shifts downward. In other words, the gains in cumulative growth in response to neutral HMIs increase at a decreasing rate once the prior duration of conflict exceeds 4 years.

Table 7 reports results for the model that accounts for prior conflict duration. Using the coefficients from the conflict-affected countries sample Model 7.1, we observe that the cumulative gains in GDP per capita that result from neutral HMIs maximise at about 3.86 years (neutral*prior duration of violence (in years) which is almost the same as observed in the Figure AP.1.

= 0.147 + 2 * (- 0.019) neutral * violence duration = 3.86 years

Table 7.

Cumulative impact (prior duration of violence)

Variables	7.1K = 5
Biased interventions	0.028 (0.44)
Neutral interventions * prior duration of violence	0.147*** (0.00)
Neutral interventions * prior duration of Violence^²	−0.020*** (0.00)
Observations	2,758
Number of countries	86
Other controls	Yes
Within R2	0.10
F-stat	32.08

Parentheses contain p value where ***p < 0.01, **p < 0.05, *p < 0.1. Driscoll and Kraay (1998) standard errors in all models.

Dependent Variable: GDP Per Capita Growth Rate from year t–1 to year t + k.

We observe no reliable evidence for a statistically significant effect of the interaction of biased intervention and prior duration of conflict on growth.

Discussion and Conclusion

Humanitarian military interventions (HMIs) are launched on the pretext of resolving ongoing violent conflicts. These interventions involve the deployment of military force against either the government, the rebels, or both, with the intention of compelling the targeted parties to commit to peace. The analyses in Saeed (2022) show that conflict-affected countries targeted with neutral HMIs-in which the intervener targets all violent actors-lower conflict intensity, whereas the biased interventions, where the intervener acts discriminately, escalate violence. This contribution examines whether the positive impact of neutral HMIs is also experienced on the economic front, particularly on long-run economic growth. While interventions can cause direct economic consequences-as likely to result from setting up supporting infrastructure, local spending by foreign troops and locally hired staff-our main focus is on the long-run impact, which is likely to last beyond the life of the mission itself. Neutral interventions that indiscriminately target all violence actors are likely to raise the cost of violence and reduce the likelihood of victory for all the parties in the conflict, and hence compel them to opt for a negotiated solution. The resulting reduction in conflict intensity can lead to a positive impact on long-run economic growth. Biased interventions are likely to encourage the supported parties to increase their atrocities and hence lead to more violence. As a result, this should harm long-run economic growth. Results from regression analyses across different samples and databases suggest that neutral HMIs contribute around 14 percentage points to the 5-year cumulative GDP per capita following interventions. These results are consistent even when we control for interventions using the number of troops deployed. We observe no evidence for biased HMIs to positively contribute to long-run economic growth. We also observe that the gains from neutral HMIs tend to diminish with an increase in prior duration of violence before the start of the intervention -hence, an earlier intervention is more likely to lead to larger economic gains for the targeted countries.

These macro findings also find support in case-study–based micro evidence. The humanitarian intervention in Burundi (2001–08), classified as a neutral HMI, occurred amid a civil war that began in 1993 (Nkurunziza 2018) and culminated in the Arusha Peace and Reconciliation Agreement (2000). Regional and UN forces were deployed with the consent of the parties to provide security guarantees, monitor ceasefires, and implement disarmament, thereby stabilising the post-agreement transition. Consistent with the pacifying logic of neutral HMIs, conflict intensity began to decline from the second year of intervention and fell below 25 battle-related deaths by the end of the mission. This security improvement coincided with a reversal in economic performance: after a decade of uniformly negative GDP per capita growth (1993–2003), Burundi recorded positive 5-year cumulative GDP per capita growth, calculated over rolling windows (e.g., 2003–2008 for 2003), amounting to 3.3, 6.7, and 7.5 percent for the periods ending in 2003, 2005, and 2007, respectively.

The intervention in Chad (2008–10) occurred amid acute insecurity in eastern Chad driven by spillovers from the Darfur conflict, mass refugee inflows, and persistent attacks on civilians and humanitarian actors. The EU and UN-led mission involved the deployment of troops with government consent to protect civilians, secure refugee camps, and facilitate humanitarian access, rather than to intervene in Chad’s internal power struggle. Coinciding with this neutral stabilisation effort, conflict intensity declined from pre-intervention levels. This economic improvement is reflected in 5-year cumulative GDP per capita growth-measured relative to the year preceding each intervention year—which reached approximately 15, 19, and 17 percent in 2008, 2009, and 2010, respectively, consistent with the pacifying and growth-supporting effects of neutral HMIs.

In contrast, the intervention in Libya (2011) was a biased HMI, directly supporting anti-Gaddafi forces rather than maintaining neutrality. This coercive approach coincided with a massive increase in violence across the country. Economic consequences were severe: 5-year cumulative GDP per capita growth—measured relative to 2010, the year preceding the intervention—fell by 53 percent, indicating that roughly half of per capita GDP was lost, consistent with the destructive impact of non-neutral military interventions.

The policy implications from the analyses are straightforward. The intervener should take a politically neutral stance and target all perpetrators of violence, as this is more likely to not only reduce atrocities but also yield positive economic gains for the targeted country. The intervening country is likely to commit significant resources to the intervention, so a neutral battlefield strategy is more likely to lead to a positive outcome for such a resource-intensive venture.

Supplemental Material

Supplemental Material - Do Humanitarian Military Interventions Foster Economic Growth?

Supplemental Material for Do Humanitarian Military Interventions Foster Economic Growth? by Luqman Saeed in Journal of Conflict Resolution.

Supplemental Material

Supplemental Material - Do Humanitarian Military Interventions Foster Economic Growth?

Supplemental Material for Do Humanitarian Military Interventions Foster Economic Growth? by Luqman Saeed in Journal of Conflict Resolution.

Footnotes

ORCID iD

Luqman Saeed

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data supporting the findings reported in this paper are openly available from Ulster University’s Research Portal at http://doi.org/10.21251/65b3fb59-abc0-4065-bbc7-e2f9c010d8e0 (Saeed, 2026).

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Barro

Robert J

. 1995. Inflation and Economic Growth, 5326. NBER Working Paper. https://doi.org/10.3386/w5326

Beber

Bernd

Gilligan

Michael J.

Guardado

Jenny

Karim

Sabrina

. 2019. “The Promise and Peril of Peacekeeping Economies.” International Studies Quarterly 63 (2): 364–79.

Besley

Persson

. 2011. Pillars of Prosperity: The Political Economics of Development Clusters. Princeton University Press.

Bove

Vincenzo

Salvatore

Jessica Di

Elia

Leandro

. 2022. “UN Peacekeeping and Households’ Well-Being in Civil Wars.” American Journal of Political Science 66 (2): 402–17.

Bussmann

Margit

. 2010. “Foreign Direct Investment and Militarized International Conflict.” Journal of Peace Research 47 (2): 143–53.

Caruso

Khadka

Petrarca

Ricciuti

. 2016. “The Economic Impact of Peacekeeping: Evidence from South Sudan.” Defense and Peace Economics 28 (2): 250–70.

Collier

1999. “On the Economic Consequences of Civil War.” Oxford Economic Papers 51 (1): 168–83.

d’Agostino

Giorgio

Paul Dunne

John

Pieroni

Luca

. 2019. “Military Expenditure, Endogneity and Economic Growth.” Defence and Peace Economics 30 (5): 509–24.

Drazanova

Lenka

. 2019. Historical Index of Ethnic Fractionalization Dataset (HIEF). Harvard Dataverse. https://doi.org/10.7910/DVN/4JQRCL

10.

Driscoll

Kraay

A. C.

. 1998. “Consistent Covariance Matrix Estimation with Spatially Dependent Data.” The Review of Economics and Statistics 80: 549–60.

11.

Easterly

William

Levine

Ross

. 1997. “Africa’s Growth Tragedy: Policies and Ethnic Divisions.” Quarterly Journal of Economics 112 (4): 1203–50.

12.

Felbermayr

Gabriel

Kirilakha

Aleksandra

Syropoulos

Constantinos

Yalcin

Erdal

Yoto

Yotov

. 2020. “The Global Sanctions Database.” European Economic Review 129: 103561.

13.

Forsberg

2014. “Transnational Transmitters: Ethnic Kinship Ties and Conflict Contagion 1946-2009.” International Interactions 40 (2): 143–65.

14.

Grigoryan

2010. “Third-Party Intervention and the Escalation of State-Minority Conflicts.” International Studies Quarterly 54: 1143–74.

15.

Gromes

Thorsten

Dembinski

Matthias

. 2019. “Practices and Outcomes of Humanitarian Military Interventions: a New Data Set.” International Interactions 45 (6): 1032–48.

16.

Gutmann

Jerg

Neuenkirch

Mattias

Neumeier

Florian

. 2023. “The Economic Effects of International Sanctions: an Event Study.” Journal of Comparative Economics 51 (4): 1214–31.

17.

Hirshleifer

Jack

. 1991. “The Paradox of Power.” Economics & Politics 3 (3): 177–200.

18.

Hultman

. 2007. “Battle Losses and Rebel Violence: Raising the Costs for Fighting.” Terrorism and Political Violence 19 (2): 205–222.

19.

Kreutz

Joakim

. 2010. “How and when Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset.” Journal of Peace Research 47 (2): 243–50.

20.

Kydd

A. H.

Straus

. 2013. “The Road to Hell? Third-Party Intervention to Prevent Atrocities.” American Journal of Political Science 57 (3): 673–84.

21.

Marshall

Monty G

. 2019. Major Episodes of Political Violence (MEPV) and Conflict Regions, 1946-2018. Center for Systemic Peace.

22.

Murdoch

James. C.

Sandler

Todd

. 2002. “Economic Growth, Civil Wars, and Spatial Spillovers.” Journal of Conflict Resolution 46 (1): 91–110.

23.

Nkurunziza

Janvier D.

2018. The Origin and Persistence of State Fragility in Burundi. International Growth Centre.

24.

North

D. C.

1990. Institutions, Institutional Change and Economic Performance. Cambridge University Press.

25.

Osborne

Roger

. 2007. Civilization: A New History of the Western World. Pimlico.

26.

Pearson

Frederic S.

Baumann

Robert A.

. 1993. International Military Interventions 1946-1988. Inter-University Consortium for Political and Social Research, Data Collection 6035, University of Michigan.

27.

Pickering

Jeffrey

Kisangani

Emizet F.

. 2009. “The International Military Intervention Dataset: an Updated Resource for Conflict Scholars.” Journal of Peace Research 46 (4): 589–99.

28.

Rodrik

Subramanian

Trebbi

. 2004. “Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development.” Journal of Economic Growth 9 (2): 131–165.

29.

Rolandsen

Øystein

. 2015. “Small and far Between: Peacekeeping Economies in South Sudan.” Journal of Intervention and Staebuilding 9 (3): 353–71.

30.

Saeed

Luqman

. 2022. “Conflict Escalation During Neutral and Biased Humanitarian Military Interventions.” The Economics of Peace and Security Journal 17 (2): 19–31.

31.

Saeed

Luqman

. 2023. “The Impact of Military Expenditures on Economic Growth: a New Instrumental Variables Approach.” Defense and Peace Economics 36 (1): 86–101.

32.

Saeed

(2026) Replication data for “Do Humanitarian Military Interventions Foster Economic Growth?” Ulster University. https://doi.org/10.21251/65b3fb59-abc0-4065-bbc7-e2f9c010d8e0

33.

Seybolt

T. B.

2007. Humanitarian Military Interventions: The Conditions for Success and Failure. Oxford University Press.

34.

Simms

Brendan

. 2011. “‘A False Principle in the Law of Nations’: Burke, State Sovereignty, [German] Liberty, and Intervention in the Age of Westphalia.” In Humanitarian Intervention: A History, edited by Simms

Brendan

Trim

D. J.

, 89–110. Cambridge University Press.

35.

Singer

David J.

1987. “Reconstructing the Correlates of War Dataset on Material Capabilities of States, 1816-1985.” International Interactions 14: 115–32.

36.

Singer

David J.

Bremer

Stuart

John

Stuckey

. 1972. “Capability Distribution, Uncertainty, and Major Power War, 1820-1965.” In Peace, War, and Numbers, edited by Russett Bruce , 19–48. Sage.

37.

Smith

Tony

. 1994. “In Defense of Intervention.” Foreign Affairs 73 (6): 34–46.

38.

Stockholm International Peace Research Institute . 2021. SIPRI Arms Transfer Database. Stockholm International Peace Research Institute.

39.

The World Bank . 2021. World Development Indicators Database. The World Bank. https://databank.worldbank.org/reports.aspx?source=world-development-indicators

40.

Trim

D. J.

2011. ““If a Prince Use Tyrannie Towards his People”: Interventions on Behalf of Foreign Populations in Early Modern Europe.” In Humanitarian Intervention: A History, edited by Simms

Trim

D. J.

, 29–66. Cambridge University Press.

41.

Vishwasrao

Sharmila

Schneider

Matthew

Chiang

Eric P.

. 2018. “The Effects of Military Occupation on Growth.” Kyklos 72 (1): 183–207.

42.

Walzer

2006. Just and Unjust Wars: A Moral Argument with Historical Illustrations. Basic Books.