As Ernest Willis tells it, he woke up in a house on fire. It was around four in the morning in Iraan, an oil-field town in West Texas, on June 11, 1986. He had fallen asleep on the living room couch fully clothed except for his eel-skin boots, which lay beside him on the floor. It was the smoke that awakened him, and he ran to the rear bedroom to get the woman who had passed out there a few hours earlier, but the flames and smoke pushed him back. He ran to the front bedroom, where his cousin Billy had gone with another woman a few hours before, but flames again forced him back. He ran through the house and out the door, yelling, “Fire!” and then around the side and rear, banging on windows. As Willis stood in the back yard, Billy came diving naked through a bedroom window. Betsy Beleu and Gail Allison, whom the Willis cousins had just met the day before, never made it out.
At first, the police thought the fire, which came after a night of drinking and pill popping at the house, whose owners had been arrested earlier in the day, was drug-related. Maybe someone had been freebasing or cooking heroin. Later they thought that maybe it was set by an ex-husband of one of the women or a Mexican drug dealer named Santana who Allison had said was after her husband. They found no evidence of arson—for example, no one smelled gasoline—but they were suspicious of Willis. He just wasn’t acting right. He didn’t seem to be coughing as much as his cousin, he didn’t seem concerned about the dead women, and his clothes and hair weren’t singed. He’d said (and Billy had confirmed) that he had run through a burning house, yet his feet weren’t burned. He stood around smoking and acting distant as firefighters fought the blaze. Later, Willis failed a polygraph test, and the police developed a theory that marks on the floor were “pour patterns,” suggesting that an accelerant like gasoline had been used. But they had no evidence to support their suspicions: no fingerprints, no bodily fluids, no flammable liquids in the house or on Willis’ clothes or body, no witnesses, no motive.
Nevertheless, four months later Willis was arrested, charged with arson and murder, and taken into the ruthless grasp of the Texas death penalty process. Though the state had a weak circumstantial case, the cops and the prosecutors adamantly pressed ahead. Cliff Harris, then the chief of deputies and now the Pecos County sheriff, recalls, “When we took it to the grand jury, we didn’t feel that we had the evidence to get him indicted.” District attorney J. W. Johnson told the Odessa American after the trial that he had thought he had only a 10 percent chance of winning a conviction. Willis had no history of mental illness, but he was given high doses of anti-psychotic drugs, making him appear zombielike at trial—a look that prosecutors used to full advantage, vilifying him whenever they could. He was represented by well-meaning but inexperienced lawyers who made serious errors that doomed him to death row. Finally, he was abandoned by the appeals process that is supposed to be a safety net for questionable cases like his. Now he waits on death row while his final appeal before execution works its way through federal court.
It is the combination of unfairness and persistence that has put Texas under national and international scrutiny. We have been criticized for executing people who are mentally retarded, for executing people who were juveniles at the time of their offense, for trying to execute—before the federal courts stepped in to prevent it—people whose lawyers slumbered in court. These are the kinds of cases that get national attention, but there are many more that go unnoticed. Like Willis’. His case had it all: overzealous police officers and prosecutors, inadequate defense counsel, and an appellate court, the Texas Court of Criminal Appeals, that seemed almost desperate for him to die. The 57-year-old former roughneck is a poster child for what is wrong with the capital punishment system in Texas.
No one can know with absolute certainty that Willis is innocent. But innocence is not the issue here. Nor is capital punishment. Texas is a law-and-order society. We execute more criminals than any other state and most countries. Support for this policy is overwhelming; capital punishment is favored by 68 percent of Texans, compared with 59 percent of all Americans. Texas is going to have capital punishment as long as the United States Supreme Court allows it.
The issue is fairness. Our adversarial process of justice rests on an essential assumption: that the fight is fair. We should be tough on criminals, but when the moment comes that the last appeal is denied and the accused faces death by injection, we want to be able to look at ourselves in the mirror and believe that the State of Texas gave the condemned man a fair trial. The statistics say that this is not always the case. Since 1976, when the U.S. Supreme Court reinstated the death penalty after abolishing it four years earlier, 927 people have been sentenced to death in Texas. Of these, 285 have been executed (as of press time), and 188 have escaped the needle by having their sentences reduced, most of them for procedural violations. Some call these violations “technicalities,” but they can be fundamental, such as the withholding of exculpatory evidence by prosecutors. Twelve of the 188 went free—their convictions reversed or overturned or their cases dismissed or sent back for a new trial that resulted in an acquittal. It’s hard to know how many of them were actually innocent, as opposed to benefiting from some serious procedural violation by the state, but there are a handful who we can almost certainly say didn’t do the crime but were sentenced to die.
And there are still men on death row who were put there unfairly. In addition to Ernest Willis, there is César Fierro, who confessed to murder after police officers in El Paso threatened him with the torture of his mother and stepfather by police officers in Juárez; the El Paso police have admitted this, but Fierro remains on death row. Michael Blair was convicted of murder on discredited scientific evidence; even though recent DNA tests appear to exonerate him, he too remains on death row. Wrongs like these will always occur. Our criminal justice system is a government system, and the government—in this case, the courts, the cops, the district attorneys—will inevitably make mistakes. The issue is whether we are willing to correct them, as other states have done. The Republican pro-death penalty governor of Illinois, George Ryan, instituted a moratorium on executions in 2001 until the state could work out the bugs in its death penalty process, which Ryan called “fraught with errors.” In May 2002 the governor of Maryland, also a death penalty proponent, followed suit. The Texas Legislature voted down a proposed moratorium in 2001, though lawmakers couldn’t ignore the criticism of the death penalty system. So they made three changes: a new DNA testing program, a revised method for providing court-appointed defense lawyers, and a prohibition on executing the mentally retarded—the latter a bill that was vetoed by Governor Rick Perry.
We live in an era of little sympathy for criminals, especially violent ones. Gone are the old notions that “there but for the grace of God go I” and “it is better for one hundred guilty people to go free than for one innocent person to be executed.” Today, you will find death penalty proponents who argue the opposite—that it is unfortunate if occasionally a possibly innocent person is put to death, but the public interest requires that those found guilty of capital murder be executed. The assumption is that only the bad guys get caught up in the system, and that is generally true. But every once in a while, it’s the hapless ones, the losers, who go to sleep on a strange couch and are unlucky enough to wake up in a house that’s on fire.
DEATH BY LOTTERY
ERNEST WILLIS WAS A SAD sack, a drunk, a onetime oil-field hand from New Mexico who was cursed with a bad back, caused by a 1970 accident, that often prevented him from working. By age forty he had had six wives, three DWIs, and four back surgeries, the most recent one a month before the fire, and he had developed a fondness for pain pills. In addition to the DWIs, he had been in trouble with the law on a couple of occasions—once, in his twenties, when he was arrested for cruising by the drive-through window of a fast-food restaurant naked and drunk, and another time, a couple of years ago, for making obscene phone calls. But he had no violent criminal past, not even a juvenile record. Lately he hadn’t been able to work and was living on food stamps, so he had moved to Odessa to live with Billy, a guy who sometimes made and sold bathtub speed. They had come to Iraan hoping their luck would change.
Instead, Willis’ got worse: He lost the capital punishment lottery. Only about one in a hundred killings ends up as a death penalty case. Who decides? The local district attorney. What does he base his decision on? There’s no simple answer. Prosecutors have enormous discretion and are accountable to no one, except to the voters who elect them. You might think that politics would cause all DAs to be death penalty advocates, but this is not borne out by the facts. Since 1976, only 116 of Texas’ 254 counties (fewer than half) have sentenced a person to death; more than half the counties (138) have never sent anyone to death row. In theory, the odds were with Willis in Pecos County; before his case, according to prosecutor Johnson, authorities had not sought the death penalty since the days of Judge Roy Bean, when the rope was used, not the needle. So what made Johnson decide that the Willis case should be treated as capital murder? He insists he didn’t. “I presented the evidence and witnesses to the grand jury,” Johnson says, “and they are the ones who made a determination it was capital murder.”
Most district attorneys would admit to taking a more active role. Retired Harris County DA Johnny Holmes, who won more death sentences than any DA in Texas history, always made the call on whether to seek the death penalty. “The most important issue to me,” he says, “is whether a reasonable cross section of the public in this jurisdiction, sitting as a jury, would vote to impose death. There are many factors that go into that decision.” Interviews with prosecutors and defense attorneys produced a long list of such factors: politics, the heinousness of the crime, the chance of winning, how good the defense attorney is, the willingness of a defendant to accept a plea bargain for a lengthy sentence, and how much publicity the case is getting. “I think the press has a lot to do with it,” says Robert Icenhauer-Ramirez, an Austin criminal defense attorney for 23 years. “If the case is high-profile and the DA figures he will have an easy time making the case, he’ll go for the death penalty. I’ve had horrendous cases with horrible facts that got no publicity. The DA will treat them as non-death penalty cases.”
One of the biggest factors is money. Many counties have never sent anyone to death row because they can’t afford to. It costs anywhere from $50,000 to $100,000 to plan and prosecute a capital murder case. Some counties don’t have their own medical examiners and have to hire one to do an autopsy. Some don’t have a crime lab and have to pay another county to test forensic evidence. Some counties have only one judge; since a trial can take two to three months, they have to pay a visiting judge to take care of all the other cases backed up behind the murder trial. Judges and DAs are beholden to county commissioners, who control the purse strings. Norman Lanford was a former district judge in Harris County as well as a visiting judge in various other counties. Out there, he says, “The commissioners would tell judges, ’Don’t ever do a capital murder case. We can get a road grader for that kind of money.’”
In other words, if you have to kill someone during a robbery, do it in Waller County, which has never prosecuted anyone for capital murder. Don’t, however, do it next door, in Harris County. Like most urban counties, it has a prosecution machine. The DA has a budget of $37 million and 233 attorneys (54 of whom do nothing but try the eight to fourteen capital murder cases a year and another 10 who just work on appeals), access to the Houston Police Department and Department of Public Safety crime labs, as well as secretaries, psychologists, forensics experts, investigators, and the budget to hire expert witnesses. The same is true in Dallas, San Antonio, El Paso, and Austin. Prosecutors there are specialists at trying capital murder cases.
At his trial, Willis appeared lost in a fog. His court-appointed lawyer, Steven Woolard, gave him a legal pad and a pencil. “He said to doodle, do anything—just look busy,” Willis says now. “He asked me if it was the pain medication causing me to act like this. I thought I was acting normal. I didn’t know.” In fact, while Willis sat in the Pecos County jail awaiting trial, someone—no one remembers who, but it had to be someone connected with the state’s side of the case—ordered that he be given high daily doses of Haldol and Perphenazine, two anti-psychotic medicines, along with the pain pills for his back. Haldol especially is given to people with severe mental illness, and according to a doctor who testified in a 1996 hearing to reopen the Willis case, the standard dosage for a person who is “barking at the moon, a danger to other people and himself,” is fifteen milligrams a day. Willis was given forty milligrams a day, on top of an undetermined daily dosage of Perphenazine.
Nor can anyone remember why the medication was ordered. Back in June, shortly after the fire, Willis had told deputy sheriff Larry Jackson about sometimes feeling tense and nervous, but he had no history of mental illness or psychosis, and the jailers all said he had been a model prisoner. “Ernest was never any problem,” says then-deputy Cliff Harris. “He was always quiet.” Willis did what he was told and took the pills.
Prosecutors Johnson and Albert Valadez both say they were never aware of the doping and that Willis didn’t appear to be acting strangely. Yet, the trial transcript reveals that Johnson repeatedly used the defendant’s doped-up demeanor against him, calling him an “animal” and a “satanic demon” and referring to “this deadpan, insensitive, expressionless face” and “cold fish eyes”—symptoms that, according to psychologists testifying at a later hearing, are typical side effects of anti-psychotic drugs.
The jury didn’t believe Willis’ story that he had woken up in a burning house or his attorneys’ theory that the fire was accidental. The prosecution’s theory of cold-blooded arson was much easier to believe. All the jurors had to do was look at the remorseless monster sitting there blank-faced, with “these weird eyes,” as Johnson said, that would “pop open like in some science-fiction horror film.” The verdict was guilty; the jury took only an hour to give him death. Later, juror Roy Urias said he was convinced of Willis’ guilt “by his failure to deny the charges against him. Specifically, when the prosecutor referred to Mr. Willis as ’vicious,’ with his ’fish eyes,’ I expected Mr. Willis to deny the accusations. I also expected a denial when the prosecutor presented the photographs of the charred bodies of the victims. Instead, Ernest Willis remained seated, completely expressionless.” Of course, he was in no condition to do much else.
The prosecution also failed to turn over a psychological report about Willis that might have saved him from death row during the punishment phase of the trial. To give the death penalty, the jury must find that the defendant is a future danger to society and that there are no mitigating reasons to spare him from capital punishment. Court records indicate that Johnson had hired a San Angelo psychologist named Jarvis Wright to test Willis, but Wright wrote that he had found nothing in the defendant’s personality to indicate such danger. The prosecution didn’t reveal the report to the defense, as the U.S. Supreme Court requires.
Prosecutors, like all lawyers, are officers of the court, which means that their first duty is not to win but to see that justice is done. Yet this responsibility is all too often overlooked in the heat of battle. It’s a war out there, and the state wants to win. In fact, prosecutors have to win. They are under far more pressure than defense lawyers, who, most of the time, are trying to get the least possible sentence for clients who are almost certainly guilty. The DA is a politician, an elected servant of the people, and he constantly needs to prove that he is winning the war against crime. And in war, anything goes. Prosecutors and police officers sometimes lie, evade the truth, and suppress evidence. They don’t do it because they are evil; rather, they do it because they are certain the defendant is evil. So in their relentless pursuit of a conviction, they sometimes fail to disclose information that would help him, as they are required to do. They don’t disclose the names of other confessors or witnesses who saw something that would help the defendant. They don’t tell the whole truth. It’s not in their interest. The attitude of defense lawyers toward prosecutors is summed up by veteran Houston defender Randy Schaffer: “You will always have prosecutors and police cutting corners, whether it’s a death penalty case or a traffic stop. It’s indigenous to the beast—what they do. And the more severe the case, the more likely they’ll do it.”
Two aforementioned cases, those of César Fierro and Michael Blair, illustrate the lengths to which the state and its agents will go to get a conviction. In 1980 Fierro was convicted of killing a cab driver the year before in El Paso. The evidence against him was the testimony of a sixteen-year-old boy who said he was with Fierro at the time of the killing, and Fierro’s confession. At his trial, Fierro, a Mexican citizen who lived in both El Paso and Juárez, said that detectives had coerced his confession by threatening to have Mexican police officers torture his mother and stepfather, who lived in Juárez, with the dreaded chicharra, an electric generator that the Juárez police were infamous for using, applying it to an interviewee’s genitals, occasionally after wetting him or her down. At trial the lead detective, Al Medrano, denied colluding with the Mexican police, and the jury convicted Fierro and sent him to death row. Fifteen years later appellate attorneys for Fierro found in his file a report written by Medrano, in which he told how he had indeed contacted the Juárez police. Armed with rifles, they raided Fierro’s parents’ home in the middle of the night and took them to the city’s police station. Later that day Fierro, in El Paso police custody, was told where his parents were. Medrano handed Fierro the phone, and he spoke briefly with Jorge Palacios, the Mexican police chief. He hung up and immediately signed a confession.
The Fierro case involves conduct the police obviously are not supposed to engage in, but equally troublesome is something the state is allowed to do: rely on forensic evidence that—TV shows such as CSI notwithstanding—often sounds more convincing than it really is, from bite marks to blood spatters. Improved scientific methods have cast doubts on the reliability of the traditional tests used to support this kind of evidence. One of the most unreliable techniques is hair-comparison analysis. In 1996 the Justice Department did a study of 240 crime labs and found hair-comparison error rates ranging from 28 percent to 68 percent. The testimony is outlawed in Michigan and Illinois, but unfortunately for Michael Blair, it is admissible in Texas.
Blair, a convicted child molester, was arrested for one of the highest-profile crimes in Texas: the 1993 murder of seven-year-old Ashley Estell, who was kidnapped from a crowded Plano soccer tournament. The police had no fingerprints, body fluids, or eyewitnesses who could place Blair and the girl together that morning. After several days, however, Charles Linch, the trace-evidence analyst from the Southwestern Institute of Forensic Sciences, concluded that hairs found in Blair’s car “appeared similar” to Ashley’s, and hairs in a clump found at another park two miles from the abduction site looked like they belonged to the suspect and the victim. This evidence gave the police probable cause to arrest Blair.
At his trial, the police produced three witnesses who had come forward after Blair’s arrest, when his photo was blanketing the local news, and a fourth who said she’d seen a car that bore a tenuous resemblance to Blair’s Ford EXP near the area where the body was recovered. The only substantive evidence came from Linch, who said that three hairs found in Blair’s car had the same “microscopic characteristics” as Ashley’s. Two tiny black hairs, found on and near the body, were too small for comparison, but Linch said they had Mongolian characteristics, which could apply to Blair, who is half Thai. And, Linch said, a fiber found on Ashley’s body was similar to fibers from a stuffed rabbit found in Blair’s car. In his closing arguments, Collin County prosecutor J. Bryan Clayton said of the hairs, “You can call it a link, you can call it association, you can call it a match, or any other darned thing they want to call it.” The jurors did, and Blair was convicted and sent to death row.
In 1998, however, the case against Blair began to unravel when a series of newer mitochondrial DNA tests revealed that none of the hairs belonged to either him or Ashley. The latest of the four test results, on the clump of hair, came only two months ago. And, defense lawyers say, the fiber was from a stuffed rabbit bought at Target that was indistinguishable from any one of half a million stuffed animals. It’s clear now: Blair was convicted and sentenced to death on junk science.
Dubious forensic evidence also played a central role in the Ernest Willis case. Arson investigation is an inchoate “science”; in 1987, when Willis was convicted because prosecutors said pour patterns indicated he had dumped an accelerant throughout the house, it was even more so. The first national standards for fire investigation weren’t even published until 1992. “For many years fire experts looked at things like spalled concrete or crazed glass and speculated, dreamed up theories,” says Arizona State University law professor and noted authority on forensic evidence Michael Saks. “Finally, after sending umpteen people to prison, they did empirical testing. They set buildings on fire and went in and looked for spalled concrete and crazed glass. It turns out those things are unrelated to whether a fire was arson or not. It was all guesswork and imagination.”
Perhaps the most unreliable experts are those who, during the punishment phase, predict that a defendant will be a continuing danger to society. Such evidence is necessary before a jury can impose the death penalty. Though the American Psychiatric Association has said such predictions are wrong two thirds of the time, Texas prosecutors have relied on a handful of psychiatrists who can be counted upon to answer, emphatically and almost every time, yes, this person is a future danger, thus dooming him or her to die. The most notorious was James Grigson, of Dallas, a psychiatrist who was known in legal circles as Dr. Death because of the scores of Texas capital murder cases in which he testified, using phrases like “absolutely certain” or “one hundred percent sure.” Many times he never even interviewed the person he testified about.
Defense attorneys must object to such flimsy evidence to preserve the right to object to it on appeal, but they know they will be overruled. In the view of many defense attorneys, judges are not neutral referees, assuring that a trial is fair, but adversaries, especially in capital punishment cases. Former judge Lanford says, “Generally, the judge is the second or third prosecutor in the courtroom. The state is going to win on most things.” Many district judges are former prosecutors; some even worked for the prosecutor’s office that is trying the case before them. In Harris County 20 of the 22 judges in local felony courts previously worked in the DA’s office. Like prosecutors, judges are elected; they will be tough on crime, especially in capital punishment cases. Most defendants have only one chance for a fair trial: a court-appointed lawyer who knows how to defend a death penalty case. In Texas, that chance isn’t very good.
THE DEFENSE RESTS—LITERALLY
I PUT MY HEART AND soul into the defense of Ernest Willis,” says Steven Woolard, the lead counsel in the case. “But if I were a judge today, there’s no way I would have appointed me then.” Not many attorneys were available in Pecos County in 1986, Woolard recalls, and he and another lawyer were named by Judge Brock Jones to defend Willis. Woolard was zealous but inexperienced: At the time, he had been practicing law for less than four years. He had never tried a death penalty case.
Willis’ attorneys never tried to poke holes in the prosecution’s unlikely theory that a pillhead with “surgical failed back syndrome” (Willis, vomiting from pain, had seen two doctors the day of the fire), who had drunk a six-pack of Coors, could have gone out—would have wanted to go out—at three-thirty in the morning to siphon ten to fifteen gallons of gas into a can, douse the house from one end to the other, set the place on fire, and then get rid of the can without waking anyone up or getting even a drop on his hands, feet, or clothes. They rarely objected when Johnson called Willis an “animal.” Worst of all, during the punishment phase of the trial, they asked only two perfunctory questions of the state’s two witnesses, who claimed Willis had a “bad” reputation but gave no details, and they rested their case without calling character witnesses who might have persuaded the jury to spare Willis’ life. Many years later Pecos County deputy sheriff Larry Jackson (now deceased) told the Dallas Morning News, “If he’d had sufficient counsel, he wouldn’t be on death row. . . . They messed this old boy around for years.”
Kevin McNally, a Kentucky attorney who analyzes cases for evidence of bad lawyering, testified in a hearing to determine whether Willis should get a new trial that the punishment-phase lawyering was “in the bottom one half of one percent” of the two hundred cases he had looked at. That would make it the very worst. He referred to witnesses Woolard knew about but didn’t call: family members and friends, some of whom came forth at this hearing and testified that Willis was a loving father, a good boss, and a decent man. Willis’ brother Alton related a story about a family gathering at Lake Stamford, when Willis had seen a boy accidentally back a truck into the lake. The man J. W. Johnson had called a “satanic demon” had pulled off his boots, dived into the water, broken a window, pulled the kid out, and then refused money for saving his life. “Most capital defense lawyers would trade their right arm” for this kind of mitigating evidence, McNally said. Today Woolard says he didn’t call the character witnesses because of concerns about their credibility: “Their presentation, manner of dress, cultural affectations.” In other words, they were rednecks. In Pecos County, of all places.
Once again, the Willis case shows the extent to which the death penalty system is like a lottery. A few counties, such as Dallas, have public-defender systems with experienced attorneys. In most counties, however, the trial judge appoints attorneys for indigent defendants from a list of available volunteers. Some are experienced lawyers, but many more are inexperienced (sometimes only a couple of years out of law school); they are easily confused by the arcane rules of capital punishment cases and cowed by the prosecutorial juggernauts. Court-appointed attorneys are frequently solo practitioners with little support staff to investigate, find witnesses, and keep track of motions to file. They object when they shouldn’t and don’t object when they should. As in the Willis case, they don’t question the obvious or do the basic work to save their clients’ lives; as in Michael Blair’s case, they don’t hammer away at flimsy evidence. They cut corners. Sometimes they just give up. A Dallas Morning News investigation in 2000 found that one quarter of all death row inmates had been defended by attorneys who had been or were later disciplined by the State Bar of Texas for everything from lying to neglecting their cases.
One of the reasons for the bad lawyering is bad pay, which chases away good people and makes a good defense impossible. Court-appointed defense attorneys often lack the budgets to hire their own experts to attack those of the prosecution. In Willis’ case, Woolard hired a fledgling arson investigator, whose credentials Johnson mocked mercilessly. “I felt inhibited somewhat,” says Woolard now about his choice. “I had to justify expenses to [trial judge Brock] Jones.” Unlike prosecutors, judges, police officers, or jailers, court-appointed attorneys are the only ones in the criminal justice system who work for less than the going rate for their profession. County officials can think of a lot of ways they would rather spend tax dollars than defending accused murderers. So appointed lawyers have to battle to get paid. “You take a voucher to the judge after the case,” says an Austin defense attorney. “Let’s say you worked three hundred and fifty hours. The judge would cut it in half.” Such penny-pinching stifles the lawyers’ incentive to investigate and put on a vigorous defense. In the case of Federico Macias, the federal court that overturned a guilty verdict because of ineffective assistance of counsel noted in its opinion that the trial attorney had been paid roughly $11.84 an hour. “Unfortunately,” the court said of defense counsel, who had failed to interview the witnesses who would one day exonerate the defendant, “the justice system got only what it paid for.”
Proof of the low quality of court-appointed lawyers in Texas came in a state bar committee study of 2,983 Texas defense attorneys, prosecutors, and judges that appeared in 2000. Called “Muting Gideon’s Trumpet” (the reference is to a book about the case of Gideon v. Wainwright, in which the U.S. Supreme Court said that every indigent criminal defendant had to be provided with a lawyer), the study revealed a system in which judges appointed attorneys who were friends or campaign contributors, especially if they were good at speeding the case through the court. The study confirmed that many state trial judges operated a patronage system: Attorneys who were beholden to judges for work turned around and made campaign contributions to those same judges. Former judge Lanford remembers a colleague, George Walker, who gave death penalty cases to a friend, the late Joe Cannon. “Joe was a nice man, but he was incompetent to handle capital cases,” Lanford recalls. “He was George’s buddy. He got the cases because he moved them. There was pressure—keep costs down, keep things moving.”
Such a system inevitably wound up embracing incompetence. Cannon was one of the infamous sleeping lawyers; he bragged about hurrying through trials. In the murder trial of Calvin Burdine, Cannon slept during the questioning of witnesses, and though he knew about mitigating character witnesses, he failed to bring them into court to testify. Burdine was convicted and got the death penalty. Then there was Ronald Mock, who kept getting appointments (and a steady paycheck) despite sloppy lawyering that caused him to be disciplined five times by the state bar. Mock defended more than a dozen men who wound up on death row.
To the Legislature’s credit, the patronage system that bumbled so many men to death row has been improved by the passage, in 2001, of the Fair Defense Act. Though it is still up to individual counties how they appoint attorneys to defend the poor, the act says that judges have to adopt stricter procedures for appointing attorneys, specify their qualifications, and pay a “reasonable fee.” Counties also have to set standards (at least five years of criminal-law experience) and require continuing-education seminars in defending criminals. For the first time, the state has provided money to supplement what counties pay for indigent defense—a total of $19.7 million for 2002 and 2003. That is approximately 10 percent of the total cost; most states pay half. The law has been in force only since January 2002, so it’s difficult to gauge its effectiveness. Jim Bethke, the director of the Task Force on Indigent Defense, says, “Anecdotally, things have improved.” But some defense lawyers remain unimpressed. “Texas has developed a culture of bad legal representation,” says veteran Austin attorney Rob Owen, who has defended more than fifty death penalty cases. “Just paying more money per hour provides more money for poor representation.”
Ernest Willis would be dead today if not for his appellate lawyers. Back in 1989, Willis had lost his direct appeal before the Court of Criminal Appeals (CCA), the one that automatically follows a guilty verdict. After that came his writ of habeas corpus, an appeal that concerns new evidence and violations of constitutional rights (also filed before the CCA). Attorneys with the Texas Resource Center, a now-defunct federally funded organization that represented poor death row inmates, and then Latham and Watkins, a large international firm with offices in New York, put on the kind of vigorous defense for Willis that court-appointed lawyers in Texas could not afford. For instance, Latham and Watkins, working pro bono for Willis since 1995, has used five lawyers, a private investigator, a professor of psychiatry, a forensic psychologist, a neuropharmacologist, and an arson investigator. Willis’ appellate attorneys looked at the county jail logs and discovered the daily dopings. They tracked down the psychological evaluation that said Willis was not a future danger. And they looked into the strange story of David Long, a convicted ax-murderer who used to make and sell bathtub speed with Billy Willis. The born-again Long had met Ernest Willis in the dayroom at the Ellis prison unit in 1990. Eventually he confessed to the prison psychiatrist that he had set fire to the Iraan house. The psychiatrist believed Long and set up a videotaped confession in 1990, during which Long confessed in detail, saying he had driven from Round Rock to Iraan that night, drinking and shooting speed. When he got to the house, he started the fire with a mix of Wild Turkey and Everclear, his favorite drink. Long had motive (“I hated the dude,” he said about Billy, toward whom he had various druggy grudges) and a history of violence: In 1983, after being fired by his boss, Long had used whiskey to set the man’s trailer on fire. “I killed him because I hated the son of a bitch,” he said in a 1986 confession.
In 1995 lead attorney Jim Blank, of Latham and Watkins, went to the CCA with the new evidence. The court ruled in 1996 that trial judge Jones should hold hearings to determine whether Willis was entitled to a new trial. The wheels of justice ground slowly; hearings were held intermittently during the next three years. Blank brought forward the previously ignored witnesses, who testified to Willis’ good character. He found an arson expert who said the state’s pour-pattern theory was all wrong—the patterns on the floor could have been caused by any number of things. The expert had also done an experiment to see if Long’s Wild Turkey and Everclear cocktail was capable of setting fire to carpet and wood; it was. Blank got Woolard to admit on the stand to several serious trial errors, including failing to offer any character witnesses. (“I loaded my guns for the guilt-innocence question and felt so very strongly about that,” Woolard offered.)
Jones was convinced: Willis had not gotten a fair trial. In June 2000, in a 33-page opinion, he ordered a new trial based on the withheld psychological profile, the mind-numbing drugs, and the ineffective assistance of counsel. All Ernest Willis needed was for the CCA to uphold Jones’s order, and he would get the shot at freedom he deserved.
DISORDER IN THE COURT
THE COURT OF CRIMINAL APPEALS is no ordinary court. The idea of a separate court of last resort for criminal cases is one that has been embraced by only one other state, Oklahoma. The court has always had its critics. Its isolation in a single area of the law caused it long ago to develop a fondness for legal hypertechnicalities at the expense of justice. It used to have a reputation for being pro-defendant, overturning cases for minor procedural defects. In the forties the CCA famously reversed the conviction of a murderer who had stomped an old woman to death because the indictment didn’t say he stomped her with his feet. Through the eighties, the court kept its reputation for overturning convictions and ordering new trials, reversing up to a third of its cases. The CCA was all Democratic until 1992, when the first Republican judge was elected. Outrage in 1993 over a brutal Houston murder—in which the CCA ordered a new trial because the cards containing the names of potential jurors were shuffled an extra time—led to the elections in 1994 of Republicans Sharon Keller and Steve Mansfield, and by 1999 all nine judges were Republicans. As with elections for DAs, elections for the CCA have increasingly emphasized how tough the candidate would be on criminals. Keller, now the presiding judge, has campaigned on the idea that failure to give the death penalty is a human rights violation. (She declined to be interviewed for this story.) In 2001 Judge Tom Price, the closest thing to a voice of moderation on the court, received an official reprimand from the Commission on Judicial Conduct for his 2000 campaign literature, which included the statement: “I have no feelings for the criminal. All my feelings lie with the victim.” Indeed. Since 1994, the CCA has reversed only thirteen death penalty convictions on direct appeal, about 3 percent of the total cases—the lowest death penalty reversal rate of any state court of last resort in the country. The court is even tougher on habeas corpus appeals; since 1995, the CCA has granted new trials on death penalty writs only twice—out of more than five hundred writs coming its way. “From the seventies through the nineties, I got reversals on sixty percent of my habeas writs,” remembers Houston defense attorney Randy Schaffer. “Since the mid-nineties, I doubt if I get ten percent reversed. Did I get real stupid? I don’t think so. The judges stopped looking at the damn cases.”
When Judge Jones made his recommendation for a new trial for Willis, he was going out on a limb. Trial judges, who must face the local electorate, don’t arbitrarily recommend new trials for death row inmates. For this reason, appellate courts usually defer to trial judges, who are closest to the action. Not the CCA. “If the trial judge recommends that relief be refused, the court will follow the trial judge,” says Charlie Baird, a Democrat who served on the CCA until the end of 1998. “But if the trial judge recommends relief be granted, the court will figure out some way to get around that recommendation.”
In the change from Democrat to Republican, the court changed its philosophy but not its character: It is still hypertechnical. The most notorious example of this did not involve the death penalty. Roy Criner had received a 99-year sentence for the rape of a woman who was also murdered. In 1998 a DNA test proved that the sperm in the victim wasn’t Criner’s, and the trial judge ordered a new trial. In a 5-3 opinion written by Judge Keller, the CCA denied Criner a hearing on the new evidence. “The DNA evidence . . . does not establish his innocence,” she wrote, noting that Criner could have used a condom or not ejaculated. Former judge Baird, who dissented, is still outraged: “The problem with Keller’s position was that those arguments were never made by the state. Keller left any semblance of being an impartial judge behind and became a partisan advocate for the prosecution. And it begs the question, Why would anyone want an innocent man to stay in prison?” Judge Price later wrote that the decision had made the CCA a “national laughingstock.” Keller didn’t help matters when she gave an interview in 2000 for Frontline, discounting the DNA evidence and calling the victim “promiscuous.” About Criner’s little innocence problem, she said, “He has to establish unquestionably that he is innocent, and he hasn’t done it.” When asked how a person could prove he was innocent, she replied, “I don’t know. I don’t know.” She’s right: It’s almost impossible under the court’s standard, which is “clear and convincing evidence.” If exonerating DNA isn’t “clear and convincing,” what is? (Criner was eventually freed after the Board of Pardons and Paroles recommended that he be pardoned.)
The CCA has also made it almost impossible to show that the state violated a defendant’s right to a fair trial. The court typically describes mistakes or misconduct during a trial as “harmless error.” In other words, the defendant would have been convicted anyway. Perhaps the most infamous examples of harmless error occurred in the sleeping-lawyer cases, one of which involved Calvin Burdine. Even though the trial court said he should get a new trial, the CCA overruled. (A federal judge rejected the CCA’s opinion in June, and he will get a new trial next year.)
The most troublesome use of harmless error was in 1996, when the CCA ruled on the capital murder conviction of César Fierro, the suspect who had confessed after being warned that his parents would be tortured in Mexico. In a July 1994 affidavit, the DA at the time of the trial, Gary Weiser, said, “I believe that Medrano and Palacios colluded to coerce Fierro’s confession.” Had he known, he says, he would have recommended that the judge suppress the confession and dismiss the case unless he could have corroborated other testimony. The trial judge found a “strong likelihood” that the confession had been coerced and said Fierro should get a new trial. Alas, the CCA overruled a judge once again. Yes, Keller wrote, the police had lied about coercing a confession, but the trial court would have found Fierro guilty anyway. “[W]e conclude that applicant’s due process rights were violated,” she wrote. “But, because we conclude that the error was harmless, we deny relief.” Though he believes Fierro committed the murder, Weiser thinks he deserves to go free. “I was a prosecutor for ten years, and I put a lot of people to death,” he says. “I never lost one. But to execute a man on illegally produced evidence—it’s wrong. It’s not justice. Nobody should be convicted on illegally obtained evidence.” Once again: If violating a citizen’s right to due process and threatening torture isn’t harmful, what is?
One of the more baffling things about the CCA is its failure to respect the fact that competent counsel is an essential part of the constitutional guarantee of a fair trial. The CCA’s position is that any licensed attorney meets the competency standard. For example, in 1997 a death row inmate named Ricky Kerr wrote the CCA, saying he was worried that his neophyte court-appointed appellate lawyer was overlooking his constitutional claims and that he wanted a new attorney. The court refused. The attorney subsequently botched the appeal, which the CCA dismissed. A federal judge stayed Kerr’s execution and called the CCA’s actions in the case “a cynical and reprehensible attempt to expedite petitioner’s execution at the expense of all semblance of fairness and integrity.”
The CCA’s critics say that the court is result-oriented, ruling on ideology. Asked to explain the court’s sometimes bizarre opinions, former judge Baird says, “They are beyond comprehension. They cannot be understood because they are the product of judges who are intellectually dishonest. They first determine the result they want, and then they distort the law to fit that result.” But a former colleague of Baird’s, Mike McCormick, who served on the court from 1980 to 2000, thinks critics have an agenda of their own: “Calling a court result-oriented, well, it depends whether you’re on the winning side or the losing side.”
The real question is whether the role of the court should be restricted to construing the law or broadened to include dispensing justice. McCormick believes the CCA’s job is to interpret the law: “One individual judge’s concept of justice is not what the court is all about.” And what about cases like Ernest Willis’, where it looks like an injustice is being done, where it looks like the guy really might not have committed the crime? “If you have evidence of actual innocence,” says McCormick, “the vehicle to get it in the system is the governor and clemency.”
The Board of Pardons and Paroles is often the last chance for the condemned. It isn’t much of a chance, though, and it isn’t much of a board either. The eighteen members, all appointed by the governor, have never gotten together to vote in the past quarter of a century. They’ve never even conducted a hearing. They individually consider the cases and then vote, by fax and e-mail. “We vote on our best gut feeling,” says member Paul Kiel, “with all the information we have.” A pardon can be granted by the governor only on the board’s recommendation. But the board has granted only two death penalty pardons since 1990, and both were requested by prosecutors. In short, the board has neither the desire nor the authority to deliberate issues of innocence. Board chairman Gerald Garrett says that innocence should be up to the judicial system. “I don’t think we should casually set aside rulings of the courts,” he says. It’s a catch-22 worthy of the whole Texas death penalty system: No one cares about the possibility of innocence.
And so, six months after Judge Jones ordered a new trial for Ernest Willis, the Court of Criminal Appeals, in a six-page reply, denied all relief. The state’s court of last resort found that Willis hadn’t proved that he took the anti-psychotic drugs involuntarily; that perhaps the state had an “essential state policy” in giving them that Judge Jones never asked about; that Woolard used “reasonable professional judgment” in not calling character witnesses; and that the suppressed psychological report regarding future dangerousness was “inconclusive,” an interpretation the psychologist, Jarvis Wright, disagreed with, later saying in an affidavit that he saw “no evidence that Mr. Willis would pose a future danger.” Doping, cheating, bumbling—if these don’t trouble the Court of Criminal Appeals, what will?
JIM BLANK HAS FILED A habeas petition in federal court and hopes for oral arguments in Midland soon. But the odds aren’t good—since Congress passed the Anti-Terrorism and Effective Death Penalty Act in 1996, the federal courts have been severely limited in granting habeas relief. Willis is beginning his sixteenth year on death row. In October 2000 he married for the seventh time, to Verilyn Harbin, the sister of former death row inmate Ricky McGinn, who was executed that same year. The two started writing each other a few years ago, then met and fell in love through the Plexiglas windows of the visitor’s cage. She says, “He is the most loving person I’ve ever met.” He says if not for her, he would have given up already.
Meanwhile, former Pecos County prosecutor Johnson, now a defense attorney, still thinks Willis is guilty. “Twelve grand jurors and twelve members of the jury—that’s twenty-four people who made the decision unanimously,” he says. He and other defenders of the Texas death penalty process insist that the system works. This, says Judge Michael Keasler, of the CCA, is the reason the court doesn’t overturn more cases: “They’re tried well. That’s a tribute to the jobs the trial judges, prosecutors, and defense lawyers are doing.” Try telling that to the federal judges who have castigated the system. Or to Blank. He and his firm have worked billable hours in excess of $1 million trying to get Willis off death row. Ultimately, the only death row inmates who stand a chance in Texas are those with pro bono attorneys—lawyers with the resources, experience, and desire to take on the state. And this is the final proof that the system doesn’t work. Every Texan who has walked free from death row has done so with outside help—filmmakers, TV stars, preachers, activists, and pro bono lawyers, not the attorneys appointed by the state to represent them. They got out in spite of the system, not because of it.
It’s an unfair system, and we need a statewide debate on how to straighten it out. The 2001 legislative session showed that capital punishment sits heavily on people’s minds. Lawmakers made some changes, but more are needed: For example, beef up the Fair Defense Act, ensure that claims of innocence backed by new evidence get a hearing, restructure the Board of Pardons and Paroles, change the way we select judges to the CCA—or maybe just abolish the damned thing. At the very least, the Legislature should institute a two-year moratorium on executions, the length of one legislative cycle, while it studies the problem. In the long run, this won’t prevent any justifiable executions, but it will make sure that every execution is, in the best sense of the word, justified.
SIX WHO WALKED OFF DEATH ROW.
RANDALL DALE ADAMS was convicted of the 1976 killing of a Dallas policeman who had stopped a car driven by David Harris. Adams said he wasn’t even in the car at the time, but Harris said Adams was the gunman. At the trial, the state relied on an eyewitness who had picked someone else out of a lineup, as well as Harris, a sixteen-year-old with a long juvenile record. As a juvenile, he wasn’t eligible for the death penalty. But Adams was, and he got it. In 1988 The Thin Blue Line, a documentary about the case, was released to great acclaim and Harris confessed. A year later Adams’ conviction was overturned, and he was granted a new trial. The DA dismissed charges, and Adams was freed.
CLARENCE BRANDLEY was a black janitor convicted of the 1980 rape and murder of a white teenager at Conroe High School. Police officers intimidated a witness who said one of the school’s four white janitors might have done it and refused to seek evidence that would have exonerated Brandley, such as blood samples from other potential suspects. Ten years later, with new attorneys, a New Jersey ministry, and 60 Minutes on his side, his conviction was overturned, and after the prosecution declined to retry, he was released in 1990. Whitewash, a Showtime movie about the case, was released earlier this year.
RICARDO GUERRA, an illegal immigrant, was riding in a car in Houston in 1982 with friend Roberto Flores when they were stopped by a policeman. Shots were fired from Flores’ gun, and both the officer and Flores were killed. Though all evidence pointed to Flores as the shooter, the police went after Guerra, hiding evidence pointing to Flores, bullying witnesses to lie, even threatening to take one witness’s child away if she didn’t cooperate. After Vinson and Elkins took the case pro bono, Guerra’s conviction was overturned in 1997 by a federal court, which called the police misconduct “outrageous.” The DA dropped all charges, and Guerra was freed.
KERRY MAX COOK was convicted of sexually mutilating and killing a woman in 1977. His sentence was overturned in 1997 because the prosecution had withheld evidence and used statements from a witness who had previously made conflicting statements. Days before a new trial, Cook pleaded no contest and was released. Soon afterward, DNA tests on semen found in the woman’s underwear proved it didn’t belong to Cook after all but to her married boyfriend.
MUNEER DEEB was convicted of hiring David Wayne Spence to kill one of the three victims of the 1982 Lake Waco murders. The only substantive evidence against him came from two jailhouse informants, one of whom later recanted. The Texas Court of Criminal Appeals overturned the conviction in 1991, and Deeb was acquitted at a new trial.
FEDERICO MACIAS was convicted in 1984 of killing an El Paso couple. His attorney never called alibi witnesses or questioned the witnesses who placed Macias at the crime scene. A New York firm took the case pro bono, found the alibi witnesses, and raised doubts about the prosecution’s witnesses. A federal court overturned the conviction in 1992 based on bad lawyering. A grand jury refused to reindict, and Macias was released.