Said Simon

Inchoate thoughts on my stuff

Why Max Abrahms’ argument doesn’t work

I want to review an article called ‘Why Terrorism Does Not Work’ by Max Abrahms (International Security, Volume 31, Number 2, Fall 2006, pp. 42-78). I discuss this article because it has showed up in several syllabi in courses related to terrorism, and appears to have flaws which so fundamentally undermine its central thesis that I do not think it deserves the high regard in which many seem to hold it.

To briefly summarise Abrahms’ argument and methodology: Abrahms compiled a case set of 28 terrorist organisations listed by the U.S. Department of State as such since 2001, and examined their demands and political impact to see whether they were able to achieve their objectives. He categorised their particular objectives and the extent to which each group was successful in reaching them. He then used statistical methods to examine how often terrorist groups achieved their those objectives in either partial or full form. His finding was that they very rarely achieve them, which he explains using psychological theories and illustrates with three in-depth case studies.

Abrahms begins by defining terrorism as attacks on civilians intended to coerce their governments into making policy concessions. (42) I don’t like this definition. It does not well capture the nuances of terrorist strategy, where attacks – be they on civilian or combatant targets – can be used not only to directly pressure a population to demand that its government make concessions, through psychological intimidation, but also among other things to prompt disproportionately harsh counterterrorism that alienates a government from its constituents, to bolster a terrorist group’s status amongst its own constituents, or to scuttle a peace process and prevent concessions. In fact, terrorism under Abrahms’ definition describes a lot of what militaries do and have done in both ‘conventional’ and ‘irregular’ warfare. However, let’s be charitable and assume that Abrahms is identifying one particular type of terrorism.

I also don’t like Abrahms’ assertion that the time period from 2001-2006 is adequate to assess the success or failure of terrorist groups. This assertion, which defends the 5-year period of his analysis, would be difficult to sustain in a good assessment of groups using terrorism as part of a ‘long strategy’ such as the IRA. But if one doesn’t accept it then one would not accept Abrahms findings no matter how well argued they otherwise are.

Abrahms critiques what he identifies as a trend in the current literature to presume that terrorism by this definition is effective. He reviews much of the literature discussing the strategic nature of terrorism – as  a rational means to an end – and notes that this literature claims terrorism’s efficacy on poor empirical foundations. It is in this review that Abrahms deals a terrible blow to his own credibility, while discussing the Israeli case:

[Robert Pape] counts as a victory the Israeli decision to release Hamas leader Sheik Ahmed Yassin [sic] from prison in October 1997, ignoring the hundreds of imprisonments and targeted assassinations of Palestinian terrorists throughout the Oslo “peace process.” (46)

Though it is certainly true that the release of Sheikh Ahmed Yassin was a product of Israel’s failed attempt to kill Khaled Meshaal and King Hussein‘s subsequent rage over Petra having been the site for the attempted assassination, and not any particular campaign of terrorism, the phrase ‘hundreds of imprisonments and targeted assassinations’ needs attention.

During the Oslo process and the subsequent years of the 1990s, Israel carried out two counterterrorist assassinations: the leader of PIJ Fathi Shiqaqi, in 1995, and Hamas/PIJ bombmaker Yahya Ayyash, in 1996. Two. Not hundreds. I suppose if one makes a category called ‘imprisonments and assassinations’ then one can perhaps talk about such numbers.

But isn’t this basically analogous to saying that Glenn Beck and Ted Bundy collectively hold dozens of parking tickets and murder convictions?

That’s a bit facetious. What I mean is, creating a category of ‘arrests and assassinations’ is a pretty bizarre thing to do. There is a huge difference between an arrest and an assassination, legally, operationally, and strategically. To be honest, it is such an absurd grouping that the most charitable interpretation by far is to ascribe it to an ignorance of how many assassinations actually took place.

In which case, one has to wonder just how much research Max Abrahms has put into his cases.

Abrahms explains his methodology for determining the objectives of terrorist groups. He writes ‘Instead of arbitrarily defining the objectives of the terrorist groups in this study, I define them as the terrorists do. In general, the stated objectives of terrorist groups are a stable and reliable indicator of their actual intentions.’ (47) [Italics mine] This is a  problematic statement. Even in advanced democracies the stated objectives of leaders or parties are not always a reliable indicator of their actual intentions. In situations of armed conflict or terrorism, it is downright negligent to assume this without further enquiry.

For example, most scholars on Hamas, inside Israel and out, agree that the organisation is likely to, or at least potentially could, agree to some sort of stable and long-term peace within the framework of a two state solution, despite the hardline language of its charter. What was once a group with maximalist nationalist goals is now a pragmatic and canny political actor. This highlights another important consideration: objectives change, can be hierarchically arranged, and may be compromised upon. If one views conflicts as political bargaining situations, then it would be unreasonable to typically expect that any party’s demands will remain absolute. Yes, there is plenty of psychological study on radicalisation and groupthink (incidentally showing that one major goal of terrorist organisations is survival), but there is also plenty of evidence to show that terrorists compromise as much as any other ‘rational’ actor in many situations, under the right circumstances (IRA, ETA, PLO, etc).

Further examples of what I consider to be improper identification of goals include:

  • Ascribing to the Abu Nidal organisation the maximalist goal of destroying Israel, when good evidence suggests that Abu Nidal was interested in gaining wealth and social prestige, and that his primary concern was competition with other Palestinian groups and with prolonging the struggle that had proved so profitable for him. He may have even contracted his organisation out to Israeli intelligence services, according to some sources.
  • Ascribing to Hizbullah the maximalist goal of destroying Israel, when so much evidence suggests that Hizbullah’s attention is primarily paid to Lebanon and to consolidating its position in that country by building strategic alliances with other regional entities, such as Syria and Hamas.
  • Ascribing to the PFLP the maximalist goals of destroying Israel and establishing a Marxist state in Palestine, when the PFLP has been a participating member of the Palestinian Authority for some time, and should be considered analogous to Hamas in that it is likely to accept a compromise within the framework of a two-state solution.

To compound these mistakes, Abrahms assigns multiple goals improperly to several groups. For example, Hamas receives two goals: ‘destroy Israel’ and ‘establish Islamic state in Palestine’. However, as I mentioned earlier, there is substantial consensus amongst scholars on the group that Hamas is capable and willing to compromise on the former, and good evidence suggests that the latter is akin to any other overall ideological goal of a political party. PFLP’s goal to establish a Marxist system should be viewed similarly. No further goals are assigned to either group – not ‘gain substantial legitimacy’ and not ‘reach a position to affect policy’ – and thus these groups are considered failures. Such small flaws highlight a much larger one.

Abrahms groups all his terrorist organisations together in the same sample, making no meaningful essential differentiation between groups that use terrorism almost exclusively and consist of a small ‘vanguard’ of activists, and vast multilayered entities engaged in multiple spheres of activism. To cite but one example, Abrahm’s sample includes both Hizbullah and Aum Shinrikyo (49). This is unreasonable bordering on unforgivable. Hizbullah is a wealthy political entity engaged in widespread social welfare work, has a significant presence in the Lebanese parliament, and has a large army that receives funding, training, and direction from Iran. It operates in the fractured environment of Lebanon and grew to power during a civil war combined with an invasion by two states (Israel and Syria). It even owns its own tv channel. Aum Shinrikyo was a millenarian apocalyptic cult that attacked the Tokyo metro in an attempt to bring about the End of Days. There is little ground for comparison between these two types of groups, and to include them both in the same sample is to deprive it of much of its analytic potential.

Abrahms concludes based on his analysis of the sample that ‘In the aggregate…the terrorist groups achieved their main policy objectives only three out of forty-two times—a 7 percent success rate. Within the coercion literature, this rate of success is considered extremely low…terrorism is thus a decidedly unprofitable coercive instrument’ (51-52) Yet as I’ve already explained, many of the 42 cases in his sample are constructed from shoddy research and bad attributions.

Abrahms does recognise that large political entities that use terrorism as a component within a broad strategy of insurgency enjoy successes, though, but doesn’t attribute those successes to organisational scope or overall political strategy. Rather, he attributes them to another variable: target selection

Abrahms writes: ‘Target selection is a superior explanatory variable for predicting the outcomes of terrorist campaigns.’ (55) When terrorists attack civilians, Abrahms argues, this leads to those targeted populations assuming maximalist objectives and an unwillingness to negotiate on the part of the terrorists. He supports this with the statistics he derived from his sample, a review of psychological theory to show that people are likely to assume that terrorist groups that attack civilians are more radical than those which focus their attentions on military targets, and three case studies. Abrahms’ theory about public perceptions of target perception is incredibly simplistic. It makes no use of studies of risk assessment heuristics, nor does it try to explain how public perceptions of terrorist goals translate into government policy.  Since changes in government policy are the sort of goal that terrorism by Abrahms’ definition is used to pursue, this renders his explanation entirely insufficient.

Abrahms case studies on how terrorists that target civilians are likely to fail are the attacks by Chechen separatists which prompted the First  and Second Chechen Wars, 9/11 and US response to it, and the First Intifada.

These are not good case studies. The choice of two cases which led to highly uncommon and extreme state responses are unlikely to produce many generalisable conclusions. The Intifada, which is broadly considered to have led to domestic and international pressure so great that Israel felt forced to accept the return of the PLO, to sign the Oslo Accords, and establish the PA, is generally considered not only to be an example of Palestinian success, but is not considered to be a terrorist campaign. Indeed, the Intifada does not actually fall under Abrahms’ definition of terrorism: ‘Throughout the intifada, only 15 percent of Palestinian demonstrations were violent. ‘ (73) Of  that violent 15% only a small fraction comprised violence against Israeli civilians. Most violence was Palestinian-on-Palestinian. Yet Abrahms argues that this is irrelevant: ‘an absolute majority of Israelis (80 percent) believed that the means employed by the Palestinians to protest Israeli rule were “mainly violent.”’ (73) However, regardless of Israeli perception, the actual behaviour of Palestinian nationalists during the Intifada does not fit Abrams definition, which does not include target perceptions as a condition.

Abrahms still believes that the Intifada can be a good case study of his argument: ‘Because the majority of Israelis regarded the intifada as a protracted terrorist campaign, and Israelis inferred from Palestinian terrorism their intentions of wanting to destroy Israel, the intifada undermined Israeli confdence in the Palestinians as a credible partner for peace.’ (74)  But regardless of Israeli public confidence, the Palestinians won semi-autonomy. The PLO returned from exile in Tunisia. Hamas became a relevant political actor, in no small part due to Israeli assistance during its early activities as a competitor to the PLO. Yassir Arafat shook hands with Yitzhak Rabin before the eyes of the world; no greater legitimacy could be bestowed upon the Palestinian leader, and by no greater enemy of his than was once Rabin. Far from being an example of the failure of an organisation to exact concessions from the government of a target population, this was an unmitigated success for Palestinian nationalist groups.

So Abrahms has given us:

  • A limited definition of terrorism
  • A false or dissembling throw-away claim about Israel’s counterterrorism that immensely undermines his credibility
  • A controversial claim about the truthfulness and robustness of the political objectives expressed in statements by terrorist organisations
  • A dataset which groups together widely disparate organisational types operating in widely disparate contexts and treats them as analogous cases
  • Statistical conclusions of dubious use, given the relevance of ‘garbage-in-garbage-out in reference to his methods for constructing a data sample
  • Some psychological theory that adds very little new perspective, and which fails to in any way explain the results he finds.
  • Two highly irregular cases of terrorism prompting extreme counterterrorism: the Chechen Wars and the US response to 9/11
  • A ‘black is white’ interpretation of the First Intifada, which not only flies in the face of the starkly obvious narrative in which Israel makes enormous concessions as a result of Palestinian activism, but which is internally self-contradictory given Abrahms’ own definition of terrorism

Anyone considering assigning Max Abrahms’ article for any other reason than to set up a straw man example of bad scholarship on terrorism should thus think twice.


7 responses to “Why Max Abrahms’ argument doesn’t work

  1. Rutger May 4, 2011 at 4:49 pm

    Good review, but for one small flaw. You label Aum Shinrikyo as a small vanguardist group, but it could just as easily qualify for you opposite category as a “vast multilayered entities engaged in multiple spheres of activism”. AS had thousands of members and was involved in all sorts of social, medical and religious activities besides their militant ones. You would have been better off using something like the RAF as the quintessential vanguardist group.

    • Said Simon May 4, 2011 at 5:20 pm

      I think it would be reasonable to state that the part of Aum Shinrikyo that was engaged in violence was a small, core vanguard, and that most adherents were unaware of it. Furthermore, I chose to highlight AS in contrast to Hizbullah because unlike RAF, AS carried out only one attack of any significance and did not last long as a violent actor. So while you’re right that they’re not as quintessentially vanguardist, I still thought they were a better example to highlight the failings of Abrahms’ dataset compared to alternatives.

  2. Danny Lieberman July 1, 2011 at 8:52 am

    Interesting reasoning that ignores the facts on the ground.

    Intifida II, the death toll, including both military and civilian, is estimated to be 5500 Palestinians and over 1100 Israelis, as well as 64 foreigners.

    Intra-Palestinian violence was a prominent feature of the Intifada, with widespread executions of alleged Israeli collaborators. While Israeli forces killed an estimated 1,100 Palestinians and Palestinians killed 164 Israelis, in Intifida I, Palestinians killed an estimated 1,000 other Palestinians as alleged collaborators, although fewer than half had any proven contact with the Israeli authorities.

    Living under military occupation and losing over 6500 lives is far from a victory.

    The plight of the Palestinians in the West Bank and Gaza, while not good is still far better then their brethren in Iraq, Lebanon, Syria and Egypt and Yemen. This has nothing to do with victories of their terror leadership but everything to do with the Israelis being רחמנים בני רחמנים – and more concerned with their PR profile than with security profile.

    The Palestinians can walk away from terror at any time. The Israelis will take the shirt off their backs for them but instead they have continued to adopt a policy of destruction and violence.

    With Western European funding flowing into the pockets of Abu Maazen and friends and corrupt economic connections with Israeli ex-generals, it’s about the money not about peace.

    • Said Simon July 1, 2011 at 7:01 pm

      Hi Danny,

      Thanks for your comment. I appreciate your contribution.

      I think it is a flaw in Abrahms’ reasoning, and in yours, to view the success and failure of a given political strategy in maximalist terms, or to group the two Intifadas together despite their hugely different dynamics. It is true that the Palestinians have not yet achieved the goal of a state. It is also true that the violence of the Aqsa/Second Intifada did not bring any significant Israeli concessions – though I think it’s important to view that violence as part of an internal power struggle within Palestinian politics as well as a struggle against Israel. However, Palestinian nationalists succeeded in gaining considerable political legitimacy during the 1970s and 1980s, and the signing of the Oslo Accords in 1993 should be considered a huge victory for Fatah and the other groups within the PLO. After all, it allowed them to return from exile and form the Palestinian Authority, which allowed them to govern the Palestinians from within the Territories and to negotiate officially with Israel. In other words, the PLO went from hated enemies to negotiating partners following the Intifada. That is a major strategic success, and marks a huge shift in the conditions of the occupation under which Palestinians have lived since 1967, even if it didn’t end it.

      The social and political conditions of the Palestinians in the West Bank and Gaza differ widely per location. I doubt you’d argue that the ‘plight’ of someone living in Gaza is the same as someone living in Ramallah. By the same token, the typical social conditions of someone living in Damascus or Beirut is likely to be better than someone living in Sana’a or Basra. Under the Hamas government, the people of the Gaza Strip probably enjoyed less political freedom than most Lebanese, even though the people of the West Bank probably enjoyed more political freedom than any other Arab community in the Middle East.

      But that brings me to a more important point: do not view ‘the Palestinians’ as though they are this single, monolithic group with a single political will that could be manifested any time. The role of violent resistance against Israel, which you and I would probably call ‘terrorism’, has played a significant role in giving political legitimacy to the various groups in Palestinian politics. Anyone who calls too loudly for nationalist groups to renounce violence will face accusations of betraying their duty. Meanwhile Hamas owes much of its support to its lack of corruption compared to Fatah. There are splits within Palestinian political parties, with younger generations being either more interested in governance than in violence or less interested, and older generations riding on their legacy as PLO ‘freedom fighters’.

      How can we talk about ‘the Palestinians’ as though they are one group with one government that can adopt one policy which enjoys public support and legitimacy? Even in a well-established democracy like Israel we would not say that government policy enjoys total support of the people. No, the Palestinians cannot simply ‘walk away from terror at any time’. And while it may be that the Israelis will take the shirt of their backs for peace, it does seem as though the Israeli government is less generous about ending construction in settlements.

      By ending construction in settlements, the Israeli government would have stopped one of the largest provocations behind the lack of trust most Palestinians feel towards Israeli will for peace. That mistrust more than anything is what’s behind the current lack of political will for peace negotiations with Israel amongst Palestinian politicians. And with a PM like Netanyahu and a FM like Avigdor Lieberman, you can hardly say that Israel is more concerned with PR than with security. The world has no trust for Israel anymore, and this is terrible because in the end, PR is part of security. Without the trust of the world and the Palestinians, security will only be won at the cost of oppression.

  3. Andrew Touesnard November 16, 2012 at 4:29 am

    You’ve obviously read the literature. I’m kind of shocked that people like Abrahms and Steven Pinker apparently haven’t. Shocked and disappointed. Great write-up.

  4. Andrew Touesnard November 16, 2012 at 4:45 am

    Moderator, please remove my previous comment. The problem isn’t a lack of research. The problem with their line of reasoning is that political violence is an act of desperation — no one should expect it to work most of the time. For example, Pinker cites Abrahms’ narrow study to make the case that violence doesn’t work. He also compares nonviolent movements to violent ones, in a kind of score-keeping, but doesn’t address the fact that violent movements resort to violence out of necessity. The Syrians did not resort to violence as a first choice — they did it following homicidal repression of protests (the same with Libya, and that ultimately happened to work for reasons that have nothing to do with nonviolence versus violence). Most groups that resort to violence are already likely to fail. If 1/3 of them succeed, it is a very liberal interpretation to say that “violence does not work”. If one tried to take a lesson from this logic, one might be tempted to think that the Syrian opposition, for example, would have been better off practicing nonviolence in response to being slaughtered en masse. That just doesn’t add up in my mind.

    • Said Simon November 16, 2012 at 5:15 am

      I actually think it is both a lack of research and a problem of interpreting the data that they did have. But I’m sympathetic to Pinker’s broader thesis. As for Abrahms’ article…well, it’s pretty bad but quite popular; it’s not surprising Pinker cited it.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: