Walkabout

Forecasting 9/11:  A Scenario-Based Approach

 

By Ralph M. Hitchens

 

Disclaimer:  Although I was an intelligence analyst for more than 20 years (retiring in February 2004), during the last several years of my career I was involved in information technology program management and maintained only a cursory awareness of current intelligence issues.  Earlier in my analytical career I dealt mainly with military capability issues and nuclear proliferation.  I was only peripherally involved with terrorism and related matters.  Everything I know about the events of September 11th, 2001 – what the Government knew and when it knew it – I learned from published government and media reports and other open sources.

 

Until now, the conventional wisdom held that the Sept. 11 assault was not a foreseeable event and it would be wrong to hold the administration accountable for failing to stop it. […]  But one Democratic commissioner after another insisted that the United States was not blind at all.  As early as July, said former senator Robert Kerrey, information was available suggesting that potential terrorists were taking lessons at flight schools.

E. J. Dionne, Jr.

Op-Ed Columnist

(Washington Post, April 9, 2004)

 

“Now, in retrospect to go back and find the report six years earlier that said perhaps they were going to use aircraft as weapons is easy to do now.  But I think the Intelligence community analysts can be forgiven for not thinking about it, given the fact that they hadn’t seen a lot in the five or six years intervening…”

Richard A. Clarke

Former NSC Counterterrorism Coordinator

(Washington Post, April 9, 2004)

 

“[T]his kind of analysis about the use of airplanes as weapons actually was never briefed to us.”

Condoleezza Rice

National Security Advisor to the President

(Washington Post, April 9, 2004)

 

Nevertheless, FBI information since that time indicates patterns of suspicious activity in this country consistent with preparations for hijackings or other types of attacks, including recent surveillance of federal buildings in New York.

President’s Daily Brief, August 6, 2001

(Declassified on April 10, 2004)

 

Investigators from [the National 9/11 Commission and the joint House-Senate inquiry into intelligence failures] have cited the FBI’s failure to act aggressively on a July 2001 report from a Phoenix FBI agent that al-Qaeda operatives might be taking flight training in the United States[.]

The Washington Post, April 13, 2004

 

This is an exercise in shameless second guessing.

 

From official testimony before Congress, press reports, and media leaks, we know there were several warnings of a highly general nature published by the Intelligence Community before September 11th, 2001 concerning al-Qaeda’s possible intention to use commercial aircraft as a terror weapon.  The earliest of these warnings may have come in 1996 or 1997.  Right down to what was probably the last official warning from the intelligence community before 9/11, the August 6, 2001 President’s Daily Brief (a declassified excerpt from which is shown above) these are maddeningly vague.  None of them represented what we like to call “actionable intelligence” – i.e., there was nothing specific (an individual, location, or event) on which to base an immediate response, nothing to justify taking prompt action. 

 

We should acknowledge at the outset that there actually was some actionable intelligence lurking below the radar, although it almost certainly was not known by the authors of the August 6th PDB.  In early 2000 CIA’s Directorate of Operations (the Clandestine Service) had, with the assistance of Saudi Arabian Intelligence and the security services in Malaysia, learned about two individuals known to be affiliated with al-Qaeda.  A successful “bag job” allowed the Agency to surreptitiously photocopy one individual’s passport, which was found to have a multiple-entry US visa.  Days later this individual and another al-Qaeda member attended the now-famous “secret meeting” of terrorists in Malaysia.  They subsequently flew into the United States, where they were routinely granted six-month residence.  Both traveled about, took some flying lessons, and communicated by cell phone with several other al-Qaeda terrorists residing in the US.  They joined the other 17 hijackers on 9/11.  The CIA’s early knowledge of these two men and their ability to legally enter the United States was certainly actionable intelligence, but their failure to share this information with the FBI seems to have been “soft-pedaled” in the aftermath of 9/11, lost in the thickets of “mea culpa” that we have seen since that tragic day and not given the critical attention it deserves.[1] 

 

What has been cited as theoretically actionable and long available in the public domain is the famous “Phoenix Memo,” outlining the suspicions of the Phoenix Office of the FBI concerning “individuals of investigative interest” – known or suspected Usama bin Laden (UBL) supporters – receiving aviation-related education and training in Arizona.  To be sure, this memo fell well short of being what the Press likes to call a “smoking gun.”  None of the individuals named in the memo were subsequently proven to have had any involvement in the 9/11 terrorist attacks, and the thrust of the author’s suspicion was misplaced – he suspected al-Qaeda’s intent to be the infiltration of trained personnel into the international civil aviation community in order to strike targets within that community.  Still, the Phoenix Memo did recommend sharing information about these Islamic aviation students with other Intelligence Community agencies and examining State Department (i.e., INS) records to track other potential bin Laden supporters receiving aviation training in the US.  The memo was dispatched to FBI Headquarters on July 10, 2001, and surely had not moved very far up the ladder before the tragic events of September 11th occurred.   

 

Under the best of circumstances it was impossible for this memo (sent with “Routine” precedence) to have been of any use in acquiring early warning of the 9/11 hijackings.  Direct dissemination of the memo to other intelligence agencies would not have been contemplated by anyone at the Hoover Building.  The specific information contained in the memo would need to be “repackaged” as an FBI intelligence report, which would take time.  Following up on the recommendations in the memo would also have taken a great deal of time.  Prompt liaison with the INS and the Department of State might have yielded, after an uncertain interval (weeks or months at best), a list of people from Islamic countries narrowed, presumably, by gender and age range, who had entered the U.S. during the past few years.  The various FBI field offices would, ideally, have freed up the necessary manpower to contact aviation schools and obtain student records going back for a number of years.  Cross-referencing all this data would eventually have yielded a list of potential suspects that would have included some of the al-Qaeda 9/11 hijackers.  But anyone familiar with the government can easily see that this process could not have been accomplished within two months.  A year or more is a realistic timeline.

 

A different approach would have taken as a point of departure the original, vague warnings in the late 1990s about al-Qaeda’s plan to use commercial aircraft as a terror weapon and employed a methodology commonly referred to as scenario-based forecasting.  Intelligence analysts would begin by assessing whether the threat was plausible in the broadest sense, and then identify the necessary events – preconditions – to make it come to pass.  For example, hijacking an airliner is within the capabilities of a trained terrorist, but even with a gun to their heads the pilots of a hijacked airliner are not likely to willingly crash into buildings.  It follows logically that al-Qaeda would need its own pilots, able to handle large commercial aircraft and also willing to become martyrs to the cause.  Could such individuals be recruited from the pool of qualified airline pilots who were also practicing Muslims?  Possible, but not likely.  Therefore, individuals with proven motivation needed to be trained to fly airliners.  Where might they obtain such training?  Not in remote camps in Afghanistan or using Microsoft Flight Simulator, to be sure.  It’s not too hard to spin this out, and establish real-world benchmarks to track al-Qaeda’s progress toward its goal.  Had the intelligence community endorsed this methodology several years ago, and followed up by establishing contacts within the commercial jet training industry and monitoring the foreign student presence, the events of 9/11 might have been forestalled.

 

Scenario-based forecasting is not about predicting the future.  Rather, it is way to anticipate future events that may lie outside our frame of reference.  In practical terms it involves deconstructing future outcomes of varying plausibility into a series of dependencies or enabling events, which can then be monitored to provide decision-makers with advance warning that a particular outcome is increasingly likely to come to pass.  This methodology, which bears some resemblance to Bayesian analysis, was popularized by the futurist Peter Schwartz more than two decades ago in his book The Art of the Long View.  Schwartz’s involvement with the strategic planning unit of the Royal Dutch Shell Corporation included what has become a landmark case study of applied scenario-based forecasting:  “The Greening of Russia.” 

 

In the early 1980s Royal Dutch Shell corporate planners were contemplating the construction of an expensive offshore natural gas facility in the North Sea to serve the European market.  The key question was whether such a facility would be economically viable in the long run.  It certainly would be if the price of natural gas – fairly high, at the time – could be maintained.  Were there any factors that might bring down the price of natural gas and jeopardize this offshore project?  Only one, Shell’s strategic planners concluded:  unrestricted access to Russia’s huge natural gas reserves. 

 

At the time, in response to Cold War imperatives, the European Community was voluntarily limiting natural gas imports from the Soviet Union to no more than 50% of its consumption requirements.  If the Cold War ended there would be no reason to keep such a restriction in place.  So, what was the prospect of a fundamental change within the USSR that would end the Cold War?  Schwartz and his colleagues surveyed the Russian political landscape and identified a number of key individuals who would most likely be associated with radical economic and political reform.  Mikhail Gorbachev – then a junior member of the Soviet Politburo – was one such individual, along with some lesser communist party figures and relatively unknown economists and scholars (“institutniki”) such as Abel Aganbegyan and Tatiana Zaslavskaya.  The scenario, therefore, had Gorbachev advancing to the leadership of the Politburo and the writings of these reform-minded institutniki receiving wide circulation and high-level acceptance, culminating in major reform initiatives and détente with the West.  

 

Even as late as 1983, when Schwartz and his colleagues presented this scenario to Shell’s corporate leaders, Mikhail Gorbachev’s ascendancy was by no means assured.  Few Western Kremlinologists were ready to bet on him, and even fewer were willing to predict significant reform if and when he came to power.  But within a few years everything was falling into place.  Thanks to the scenario-based forecasting of its strategic planning group, Royal Dutch Shell saw the light in time to scale back its investment in an expensive offshore natural gas facility.

 

The lesson for intelligence analysts, confronted with a vague warning about a future threat, is to set aside abstract theorizing about the probability of this threat becoming a reality, or comparison with other potential threats.  Obviously we should first determine whether a threat meets a certain standard of inherent plausibility.  In this case, it’s not hard to visualize how hijacking an airliner and crashing it into a large building would constitute a devastating blow against the United States:  military aircraft crashing into warships inflicted serious damage on the U.S. Navy in the final year of World War II, and novelist Tom Clancy updated this scenario twenty years ago in Debt of Honor, which saw a Japanese fanatic who happened to be an airline pilot crashing a Boeing jet transport into the Capitol Building.  Having made a plausibility determination we then need to ask the question, “What enabling events are we likely to see happen for this scenario to become a reality?”  Working through the specifics with subject matter experts, we will then have some idea of where to focus our collection efforts. 

 

Naturally this process will be easier in some cases than in others.  Terrorist threats of a deadly but commonplace nature – truck bombs, individual suicide bombers, that sort of thing – will seldom lend themselves to this methodology.  In such cases there is no substitute for good police work, supplemented by intelligence collaboration and information sharing in both the vertical and horizontal dimensions (which is a wholly separate problem).  But something as technically complex as using commercial aircraft to destroy buildings is another matter entirely.  In this instance there was at least one obvious dependency that would bring al-Qaeda out from the shadows.  While basic flight training on light aircraft is available almost anywhere, relatively inexpensive, and pursued by many people from diverse backgrounds, qualification training on large transport-category jet aircraft is available only at a relatively limited number of major flying schools and is prohibitively expensive for a private individual.  Nearly all students attending these schools are funded by an airline or corporate employer.  Surely the fact that students from the Middle East who lacked an airline affiliation or other corporate backing were attending commercial airliner training schools in the United States would qualify as a highly significant indicator.  It was also something that could be tracked without the complications and expense of redirecting “national technical means” of intelligence collection.    

 

The tragedy of 9/11 is sometimes compared to the attack on Pearl Harbor as an example of a catastrophic surprise that could have been foreseen but was not.  The comparison is apt, although the use of scenario-based forecasting was much less feasible in the case of Pearl Harbor.  Certainly in both cases there was considerable awareness at the policy level that an attack was likely, and many believed it imminent.  What surprised everyone was the where and how.  In 1941, though, scenario-based forecasting offered less promise, as Japanese military and naval movements were difficult to discern and analyze, and the cabinet-level decision-making process in Tokyo was virtually opaque to Western politicians.[2]  Still, historical analysis and recent US Navy exercise experience should have pointed toward more focused warning.  Japan had begun its last war, in 1904, with a surprise naval attack on the enemy’s principal fleet base.  Also, at least one of the US Navy’s “Fleet Problem” exercises during the 1930s featured a carrier strike against Hawaii coming from the North Pacific, an area of sparse to nonexistent shipping which an attack force, maintaining radio silence, might traverse unnoticed.  Still, in both cases only an imaginative methodology would have allowed intelligence analysts and policymakers to distinguish signals from noise, and get outside their existing frame of reference to consider non-intuitive scenarios.  

 

Why the intelligence community has never strongly embraced scenario-based forecasting or other nontraditional analytical methods is a good question, one that some of us “in the business” have often asked ourselves.  It may be that the deference toward academic scholarship as a highly-valued credential within the analytical ranks (particularly in the Directorate of Intelligence at the CIA) leads us toward inductive reasoning as the default methodology, and in an era of “information overload” only those analysts with the strongest and most retentive intellects will prosper in their careers.  Even those analysts who embrace deductive reasoning – distant kin to scenario-based forecasting – may stand the inductive-deductive duality on its head by concocting hypotheses from the residue of a great deal of inductive drudgery accomplished by others.  Compelling theories are hardly likely to spring, fully formed, from the mind of Zeus. 

 

Anyone who has sat in on an interagency working group will tell you that what counts in the intelligence community is knowledge accumulation.  Analysts instinctively defer to those in possession of more facts than they have.  An individual from the Directorate of Intelligence who is chosen to write an article for the President’s Daily Brief will clearly be an analyst on top of his or her game, someone who has “sipped from the tsunami” of unstructured data pouring in from a myriad of intelligence sources.  But absent a methodology that would impose some structure on the analytical process, would any coherent picture about the terrorist threat emerge in time for intelligence consumers – policymakers, the military, law enforcement – to do anything about it?  Not likely, judging from what we read in the PDB of August 6th, 2001.  Accurate information, as far as it went, but it didn’t go very far.  Like it or not, we have to let President Bush and Condoleezza Rice off the hook.  Reduced to reading this sort of mushy, scattershot intelligence analysis, topped off by reassurance that the FBI had opened many new field investigations, nobody at the policy level can be accused of being “asleep at the switch.”

 

Spinoza, who lived and died long before there was an intelligence community, stated our conundrum in the most concise fashion:  all our knowledge is about the past and all our decisions are about the future.  Scenario-based forecasting may help us rattle the bars of the cage in which we are, all too often, imprisoned.

 

 

[1] CIA’s failure to inform FBI of their existence and activities of these two terrorists was noted in the 9/ll Commission Report.  Finding 5b in the declassified Top Secret Annex reads:  The Intelligence Community acquired additional, and highly significant, information regarding Khalid al-Mihdhar and Nawaf al-Hazmi in early 2000. Critical parts of the information concerning al-Mihdhar and al-Hazmi lay dormant within the Intelligence Community for as long as eighteen months, at the very time when plans for the September 11 attacks were proceeding. The CIA missed repeated opportunities to act based on information in its possession that these two Bin Ladin-associated terrorists were traveling to the United States, and to add their names to watchlists.

 

[2] One exception was the principal diplomatic code, which used a machine cipher broken by the Americans.  However, the famous multi-part message of December 7th to the Japanese ambassador in Washington DC, decrypted by the Americans, was not interpreted by any of its few cleared readers – up to and including President Roosevelt – as an explicit war warning, despite the tendency of some postwar historians to regard it as such.