To recap: Since 1945 nearly three-fifths of all civil wars have experienced a third-party intervention, making self-contained civil wars the exception, not the norm. The bulk of them have entailed the supply of arms, aid, and bases, not the deployment of boots on the ground. Of the conflicts with no third-party intervention, the average length of conflict was 1.5 years. By contrast, those with outside intervention saw an average length of seven years. The longest of these wars were typically framed as part of a larger (and often nonexistent) Cold War narrative (not all were – Lebanon, Ethiopia, to name just a few). There are obvious selection effects present: Much of the conflict data is drawn from the Correlates of War (COW) dataset, which may be heavily skewed because of a few “never-ending” conflicts — namely civil wars in places like Angola that saw no shortage of external involvement. These analyses also suffer from selection issues. For instance, it is possible that outside powers select into civil wars because they are protracted, not that the act of arming factions makes war last longer.
Scholarship generally finds that third-party intervention on the side of rebel forces makes conflicts longer, bloodier, and more difficult to resolve through peaceful means. There is some dispute over the mechanisms: Some point to the proliferation of “veto players” which gives rise to more spoilers of any eventual settlement. Others suggest that outside interventions, especially for humanitarian purposes, create moral hazards, which actually encourage rebellion, not vice versa. Another line of argument argues that it is the spread of refugees, diaspora networks and other transnational communities that sustain conflicts.
There is some new and enterprising research that paints a more nuanced and mixed picture of the effects and unintended consequences of external interventions. Fotini Christia, for example, examines the impact of shifting alliance formations in Afghanistan and other war zones (e.g. Bosnia). She argues in her 2013 book (p.330-331):
I have suggested that in absence of a warring actor that can win the war on its own, the vicious cycle of alliance shifts and fractionalization is likely to go on until the intervention of a powerful and determined external arbiter who can enforce peace. Though this book is by no means a work on external intervention or civil war termination—subjects that span rich literatures in their own right—it does put forth the claim that for a civil war deadlock to come to an end, it may often require a credible external intervener willing to commit massive resources. This should not be interpreted as a case for imperialism or encouragement of third party actors to meddle in the internal affairs of sovereign states. It just recognizes that external meddling is almost ubiquitous in civil wars, and that the resultant deadlocks and quagmires are unlikely to come to an end without the involvement of a credible external guarantor.
Recent unpublished research by Jonah Schulhofer-Wohl finds that it is not outside intervention that prolongs fratricidal infighting among rebel actors. Rather, it is weak and inconsistent outside support which can prompt these groups turn against one another by ensuring their survival but not improving their odds of winning – a case in point being the Lebanese civil war. Paul Staniland in his new book provides a very useful typology of insurgent networks. He finds that outside support does not change their organizational structures, but adds, “It may be better to sponsor particular sub-factions that look fairly effective and integrated rather than relying on ineffectual central leaders.” That advice sounds remarkably similar to the 2012 plan backed by Hillary Clinton and General David Petraeus to arm and train select groups of rebels (but ultimately nixed by Obama).
Moreover, there is other research on interventions that often goes unmentioned. Nile Metternich, for example, finds that interventions by international organizations (e.g. NATO), especially those with democratization mandates, are associated with shorter conflicts, provided rebel leaders come from ethnic groups representing more than 10 percent of a country’s population (which would fit Syria’s largely Sunni opposition). Likewise, Clayton Thyne looks at unobserved variables – such as high levels of resolve among the combatants – that contribute to the resolution of conflicts even with third-party interventions. Finally, as Ann Hironaka finds, Cold War civil wars were protracted not because of the presence of third-party interventions per se, but because outsiders – namely the United States and Soviet Union – were intervening in weak states with relative parity on both sides (Interestingly, neither is the case in Syria; the Russians and Iranians are far outspending us, and until 2011 Syria was a relatively strong and secure state, ranked just above Israel in the 2010 Failed States Index). By contrast, she argues, 19th century interventions resulted in shorter civil wars, not longer ones, because they were typically one-sided affairs and more forceful. Civil wars back then also did not flair up again post-intervention, unlike today where we see repeated ceasefires and renewals of violence (see Chechnya, Sri Lanka, etc.).
Perhaps, then, the question we should be asking is not whether third-party interventions are, on average, helpful or harmful to civil war termination. The answer invariably is: Well, it depends. Rather, we should be asking: Is the world back in a 19th century multipolar paradigm, whereby civil wars were primarily fought between pro-democracy versus conservative/monarchist forces, and the latter typically won because their interventions were more robust and one-sided? As Hironaka and some historians (Sperber 2000) find, during 19th century interventions, the side most willing to use greater force was the anti-democratic side (in this case, typically the Concert of Europe).
In the current context, the anti-democratic axis as it were – that is, the Russia’s and Iran’s of this world – appear more willing to go “all in” to support their “proxies” than their pro-democracy counterparts in the West. That means we may be getting into bidding wars not that we can’t win – we have the bank and arms to outspend and out-supply just about anyone – but which we lack the will to win, whether due to flagging public support, setting too high a bar of excellence for our rebel or regime proxies (or fear of Mujahidin-like blowback), or – and this is where the 19th century comparison may be apt – because the stakes for us are perceived to be lower than they are for the Putins of this world.
In this sense, those like Marc Lynch who opposed intervention in Syria may be right: Maybe we do lack the political will to compete and should get out of the proxy-war business. I also agree with his point on the folly of trying to screen these rebels, who will say anything to get more arms, as we have seen, and, having interviewed several ex-rebels, the myth of a unified Free Syrian Army. Then the takeaway becomes effectively: If we want such wars – and I would put Ukraine in this basket – to burn out quickly, then our policy of doing nothing is in fact a smart one. There is no point in getting into a larger proxy war with Putin, who treats every sliver of land as an existential claim. But if our goal is for our side to win, then we are deluding ourselves that a slow trickle of non-lethal support will have any impact on the balance of power on the ground.
To be sure, we should not dismiss the findings of existing studies, which provide a very useful starting point for how to think about the unintended consequences of interventions. Nor should we hold it up as Exhibit A to stay out of Syria (or by extension, Ukraine). It is incorrect to claim that there is a scholarly consensus on third-party interventions, when in fact the findings are far more nuanced and in dispute. To make the point from a scholarly perspective would require findings that outline under what conditions external support for rebels, or lack thereof, can provide space for transnational groups like ISIS to thrive. To my knowledge, no such research exists (Idean Salehyan’s research on refugee flows probably comes the closest; also see Jeffrey Checkel’s book of essays from a more qualitative perspective).
It is impossible to know whether a more robust U.S. intervention in Syria would have ended the war or prevented the rise of the Islamic State. Former ambassador Robert Ford still believes there is enough of a “moderate” wing of the opposition that can still depose Assad. Yet, we should not be drawing the wrong conclusions from the civil war literature, which draws heavily from the Cold War era at the expense of 19th century interventions, which may be more applicable of this new multipolar era. Nor should we treat it as scripture, given its thinness on mechanisms, its data shortcomings (i.e. selection effects), and so forth. Indeed, the literature on interventions in civil wars, as well as their unintended consequences, is a work in progress and anything but settled.