The Spy Who Was Left Behind Read online

  I considered the narrative arc of the official version: It described a triumph of professionalism over anarchy. “Very impressive,” I thought. “Almost too good to be true.”

  * * *

  The trial of Anzor Sharmaidze began on December 30, 1993. By then most newspapers had lost interest in the murder. I found a fleeting reference to it in the Washington Times. Sharmaidze testified that his own car had broken down near Tbilisi and that he had tried to stop the Niva in which Woodruff was riding. He said he fired one shot in anger when the car did not stop but that the killing of the American was unintentional. Five weeks later the Russian news service Interfax reported that Sharmaidze had been convicted and sentenced to fifteen years in prison.

  About the same time, the US State Department announced the results of an FBI investigation into the murder: Freddie Woodruff had been killed by “a random act of violence.” I thought that this was the end of the Woodruff saga. But things were about to get much more interesting.

  On February 21, 1994, thirteen days after a Georgian judge rendered his verdict in the trial of Anzor Sharmaidze, the FBI arrested Aldrich Hazen Ames for espionage. They alleged that Ames, a senior CIA operations officer, had secretly been working for Moscow since 1985. During those years, Ames had been a branch chief in Soviet counterintelligence and, after the collapse of the Soviet Union, chief of an antinarcotics intelligence task force for the Black Sea basin. He was suspected of having betrayed to their deaths at least ten people who spied on the USSR for the CIA.

  And he had been in Tbilisi a week before Freddie Woodruff was murdered.

  Director Woolsey promised “a thorough damage assessment” of the extensive injury done by the traitor. This process, known in spy argot as “walking the cat backwards,” involved a microscopic review of Ames’s thirty-two years at the Agency: all the documents he’d looked at, all the people he’d talked to, all the secrets he’d known. It was an investigation that would take decades.

  But in the days immediately after the arrest, the US government mobilized to reexamine a question of more personal interest to me: What role, if any, had Ames played in the death of Freddie Woodruff? Spokesmen for the CIA acknowledged that Ames’s duties at the Agency’s Counternarcotics Center involved tracking the flow of heroin and cocaine through Georgia, Turkey, Bulgaria, and Romania—a drug pipeline that both Senator Jesse Helms and the Wall Street Journal attributed to the KGB. Ames had traveled to Georgia in July 1993 in connection with this work. It now seemed plausible that Woodruff’s murder two weeks later was connected either to Ames’s desire to protect the Russians or the Russians’ desire to protect Ames.

  In October 1995 the former chief of Georgia’s security service, Irakli Batiashvili, publicly asserted that Woodruff was killed at the behest of the SVR (the successor organization to the KGB First Directorate). Batiashvili had directed an independent investigation of the murder; however, the press linked his accusation to a government media campaign against Igor Giorgadze—a man accused of attempting to assassinate Eduard Shevardnadze. A former Soviet general and high-ranking KGB officer, Giorgadze had replaced Batiashvili as Georgia’s security chief. After the failed assassination attempt, Giorgadze had fled to Russia, where he was living under Moscow’s protection.

  Nine months later the man who had replaced Giorgadze as security chief made a more explicit accusation. Shota Kviraya declared that his predecessor had arranged the murder of Woodruff on Moscow’s orders and that Eldar Gogoladze (Shevardnadze’s chief bodyguard and the driver of the car in which Woodruff died) was involved in the plot. A spokesman for the Russian SVR called the accusation “groundless, absurd and malicious.” The service “is not involved in terrorism,” he said. “It fights against terrorism, together with other countries, including the United States.”

  In addition to accusations of SVR involvement, the press reports provided a previously undisclosed fact about the Woodruff story: At his trial, Anzor Sharmaidze had testified that he was tortured into “confessing” to the murder. He claimed to be innocent of all charges.

  The carefully constructed official version seemed to be coming apart and what was left in its place was a tantalizing mystery. As I considered the allure of this puzzle, it occurred to me that, with a modicum of effort, I might be able to offer the Woodruff family the comfort of reliable information. I already had deep reason to be suspicious about the sincerity of pronouncements by the US government. I had just completed a decade-long investigation into the circumstances of my father’s death in the Vietnam War and had discovered that the air force had deceived my grieving family. Contrary to their earnest representations, my father did not die in a quaint hamlet in South Vietnam. Instead, he was killed in Laos fighting a war that no one would officially acknowledge.

  So I decided to make a FOIA request about the death of Freddie Woodruff.

  * * *

  The Freedom of Information Act is a remarkable law. It compels agencies of the federal government to produce information in response to an individual request. Prior to being reluctantly signed by Lyndon Johnson, the FOIA was unanimously opposed by the executive branch agencies. They warned the president that the legislation granted extraordinary powers of intrusion and examination to people who were merely “idly curious.”

  And they were right, of course. I was merely curious and the federal government was going to have to tell me at least a part of what it knew about Woodruff’s murder.

  I initiated the process in November 1997 by mailing identical FOIA requests addressed to the Central Intelligence Agency; the Federal Bureau of Investigation; the National Security Agency; and the Department of State. I did so with no inkling of where my queries might lead. If I’d known, I might have been more careful.

  The first document arrived from the FBI more than a year-and-a-half later: a three-page memo authored by a Special Agent George Shukin on August 11, 1993. The Bonn-based FBI legal attaché (legat) had arrived in Tbilisi on August 9, a little more than eighteen hours after the murder of Woodruff. Shukin, who spoke fluent Russian, had traveled to Georgia to make a preliminary assessment of whether Woodruff’s death gave rise to a federal crime that could be prosecuted in the United States.

  The Bureau had responsibility to investigate the shooting under a law that made it illegal to murder a US diplomat anywhere in the world. Thus, even though Woodruff was merely posing as a State Department employee, the FBI had jurisdiction to investigate the CIA officer’s death.

  The Georgians informed Special Agent Shukin that Woodruff had been killed by a randomly fired bullet that struck him just above his right eye. However, the FBI agent considered this explanation unlikely. His examination of the Niva revealed that the metal skin and glass of the car were undamaged. There was no evidence that a bullet had penetrated the vehicle from the outside.

  In addition, a French wire service reported an alternative scenario that seemed to be more consistent with the forensic facts. Citing sources in Tbilisi, Agence France-Presse said that Woodruff had been killed by a shot fired from inside the car in which he was riding. AFP also reported that the driver of the car, Gogoladze, was in a “state of extreme drunkenness” when he arrived at the hospital and that he was “known for his excesses when drunk.”

  More troubling still was the condition of Woodruff’s body. An American embassy employee who was present when Gogoladze first arrived at the Kamo Street Hospital at 10 p.m. on August 8 reported that Woodruff’s body was already in an advanced state of rigor mortis, a stiffening that would not normally occur until several hours after death. Gogoladze explained the rigor and his apparent delay in reaching Tbilisi by claiming that after the shooting he had unsuccessfully searched for a hospital for three-and-a-half hours. The image of Gogoladze frantically hauling Woodruff’s dead body from village to village searching for medical care seemed implausible to Special Agent Shukin. The shooting had allegedly occurred twenty miles from Tbilisi and, as chief of Shevardnadze’s personal protection force, Gogoladze would
have been expected to know every trauma facility in the area.

  In light of these inconsistencies, Shukin said in his memo that he considered it possible that Woodruff had been killed somewhere other than the car and then put in the car. Shukin recommended that an FBI shooting team be dispatched to investigate each of the places visited by Woodruff on the day he died—but only if the Georgian government provided adequate security. It was clear to the special agent that Georgia was a very dangerous place. Shukin had already twice informed US ambassador Kent Brown that FBI investigators would not leave the United States until written assurances regarding their safety had been provided.

  I read the memo a second time. It made me queasy. The memo was exculpatory evidence. In the hands of a skilled trial lawyer it could be used to liberate an innocent man from prison. And I was a skilled trial lawyer.

  The absence of a bullet hole in the Niva meant that Sharmaidze could not have fired at the back of the car and struck someone inside. The presence of rigor mortis meant that Woodruff died long before Sharmaidze allegedly appeared at the scene of the crime. The official version of the murder was impossible.

  It was a fantastically inconvenient discovery.

  I spent weeks wrestling with my conscience. What duty did I owe Sharmaidze? I did not know him and he did not know me. I was not responsible for his imprisonment—but if I acted I might be able to set him free. And all the while I heard Edmund Burke, the eighteenth-century Irish statesman and philosopher, whispering in my ear, “All that is necessary for Evil to triumph is that good men do nothing.”

  It was 2001 and newspaper pundits were describing the rise of a uniquely modern phenomenon—the super-empowered angry man: agents of radical change who use wealth and technology to cause devastation on a scale that formerly only nation-states could do. I was no super-empowered angry man. But I was a moderately-empowered curious man.

  “I wonder what I can do with a law license, a passport, and a credit card?” I thought.

  I was about to find out.

  CHAPTER 2

  * * *

  “DO THEY THINK WE’RE IDIOTS?”

  Four years passed before the next tranche of documents arrived from the FBI. Three hundred fifty-six thoroughly censored pages, a kind of evidentiary Swiss cheese in which all the tastiest morsels have been removed.

  The redactions made most of the documents unintelligible. Long ribbons of bowdlerizing black punctuated by abbreviations, acronyms, and the occasional noun. Witness statements were expurgated. Sources and methods were obscured. Proper names were deleted. Anything that might reveal a special agent’s suspicion or conclusion was sanitized.

  What remained was the silhouette of an investigation. But it was enough to make me take notice, because I specialize in what is self-importantly called complex commercial litigation—basically complicated business disputes involving lots of documents and lots of money. It is an area of the law in which the peculiarities of my personality are reinvented as virtues: a passion for mysteries, a memory for trivia, an obsessive attention to detail, and an annoying refusal to quit when both logic and good sense demand it.

  Thus, by both trait and training, I was well suited to tease the truth out of the bits left by FBI censors. I took it as a personal challenge—to immerse myself in the minutiae of the Bureau documents and divine the content of the redactions. The project promised to be deliciously tedious.

  The foundation of my reconstruction effort was a time line, a chronology of any intelligible data that could be squeezed from the FBI documents. In order to make these un-redacted data fragments comprehensible, I had to teach myself the specialized language of Bureau investigators and censors. This had the happy and unexpected effect of making some of the redactions meaningful.

  For example, an “FD 302” was a special agent’s report on a witness interview and the marginal notation “b1” opposite a deletion was the censor’s judgment that the blacked-out information was properly classified in order to protect national defense or foreign policy. Thus, a b1 redaction of an entire FD 302 told a story: The FBI had interviewed a witness and the evidence provided by that witness was classified for reasons of national security. The extent of the deletion gave a hint about the amount of relevant evidence elicited in the interview.

  In addition, it was possible to derive information about both the witness and the interrogation based on where a document appeared in the chronology. For example, if the censored FD 302 was preceded by correspondence to FBI headquarters requesting “country clearance,” then the special agents had traveled to a foreign country to conduct the interview. If one of the unnamed agents requesting such travel authorization was identified as a polygrapher, then the FBI deemed the prospective witness cooperative enough to voluntarily submit to a polygraph but unreliable enough to necessitate the use of a lie detector.

  Further, an appreciation of the historical context in which a document was created could give insight into redacted content. For example, if an FD 302 indicated that special agents interrogated a witness in the days immediately following a newsworthy event—for example, the arrest of CIA traitor Aldrich Ames—then it could be deduced that the interview was informed by and perhaps related to that event. This logical inference is bolstered if the interview occurred during a flurry of classified Bureau activity that commenced immediately following the event.

  Finally, my time line was organic. As I learned new information I added it to the chronology. Thus, if I interviewed a witness who said that he or she had been previously interrogated by the FBI, I would search for the corresponding FD 302 and identify the prior interview on the time line. If the witness gave me details regarding the evidence he provided to the special agents, I would attempt to identify that evidence in the redactions and to carry the thread of that evidence forward into subsequent documents.

  My first step was to put the FBI documents in date order. When I did so, it became apparent that there had been two completely separate FBI investigations into the murder of Freddie Woodruff. The first investigation began immediately after the shooting and ended when the Georgian authorities convicted Anzor Sharmaidze six months later. The second investigation started eight days later—two days after the FBI arrested Aldrich Ames for espionage.

  According to the documents, the Bureau’s first concern was for the safety of its investigators. As recommended in the Shukin memo, the FBI demanded that the Georgian government give written assurances that it would provide adequate security. The issue of lawlessness in Georgia had been raised the day after the murder by George Shukin. In his report to FBI headquarters Shukin recommended that any other FBI personnel sent to Tbilisi “be accompanied by an appropriate number of fully-armed HRT personnel.” The need for the Bureau’s elite hostage response team was not hypothetical. By the summer of 1993, Georgia had devolved into a state of almost perfect anarchy. The only recognized authority was the gun.

  However, some of those guns were in the hands of American commandos.

  In 1979 the newly mobilized US Army Delta Force agreed to provide operators to train and lead local bodyguards at the State Department’s most threatened embassies. This arrangement gave Delta a permanent covert presence in some of the most unstable places in the world. A month before Woodruff’s death, the Washington Times reported that these Special Forces soldiers had been sent to Georgia in a secret mission to protect Eduard Shevardnadze. The Georgian-speaking Delta operators provided the chairman’s bodyguards with antiterrorism training, equipment, and weapons. In addition, they brought two squads of Georgian paramilitary to the United States for training at Fort Bragg, North Carolina.

  The CIA had proposed the program and Bill Clinton had approved it in a presidential directive. It was the first operation of its kind on the soil of the former Soviet Union and it did not go unnoticed in Moscow.

  After his murder, Woodruff was quickly identified as the CIA officer in charge of this covert operation. Journalists speculated that this US security initiative (or
Clinton’s recent offer to mediate armed conflicts between former Soviet states) might have provoked a violent response from revanchist elements in Russia. Nevertheless, on August 12—the same day the FBI team arrived in Georgia—US ambassador Kent Brown told the press that he expected the investigation to show two things: that the killing was not premeditated and that Woodruff had not been targeted because he was an American.

  Brown’s statement seemed to be a none-too-subtle message from the State Department to the FBI: It was politically unacceptable for Woodruff’s death to be anything other than a senseless tragedy. If the facts suggested otherwise, the facts would have to be adjusted.

  While in Tbilisi the special agents were billeted at the Sheraton Metechi Palace—an outpost of Western comfort and a hub of social activity for newly prosperous Georgians. “Unfortunately numerous Georgians who visit the Metechi carry handguns,” said the embassy’s understated security guidelines for Tbilisi. “There have been several instances where these firearms have been discharged in the hotel. While there are no absolutes in gun play, there seem to be two recurring characteristics when these incidents have occurred: the gun wielders have been drinking heavily and/or have been engaged in arguments with other Georgians. If you observe obviously drunken and/or angered Georgians, avoid them.”

  But in the summer of 1993, it was very hard to avoid drunken or angry Georgians. There was a lot to be angry about. And some good reasons to drink. The government had failed to obtain a monopoly on violence. As a result, Georgia was governed not by the rule of law but by each local strongman’s will to power. It was perfect anarchy: a society dominated by criminals and warlords. Crime was not the exception; it was the rule.