New Analysis: Innocent Death-Sentenced Prisoners Wait Longer than Ever for Exoneration

On July 1, after wait­ing 41 years for his name to be cleared, Larry Roberts became the 200^th per­son exon­er­at­ed from death row. A new Death Penalty Information Center analy­sis finds that Mr. Roberts’ expe­ri­ence illus­trates a trou­bling trend: for inno­cent death-sen­tenced pris­on­ers, the length of time between wrong­ful con­vic­tion and exon­er­a­tion is increas­ing. In the past twen­ty years, the aver­age length of time before exon­er­a­tion has rough­ly tripled, and 2024 has the high­est-ever aver­age wait before exon­er­a­tion, at 38.7 years. Our research sug­gests that two of the fac­tors con­tribut­ing to this phe­nom­e­non are pro­ce­dur­al rules restrict­ing pris­on­er appeals and resis­tance by state offi­cials to cred­i­ble claims of innocence.

Our Innocence Database tracks peo­ple who were exon­er­at­ed since 1973, after the Supreme Court’s deci­sion in Furman v. Georgia inval­i­dat­ed the nation’s death penal­ty statutes and com­mut­ed the sen­tences of every­one on death row. The ear­li­est death sen­tence in our data­base was in 1963. As a result, it makes sense that the aver­age length of time before exon­er­a­tion would increase to some degree since the begin­ning of our dataset; a per­son exon­er­at­ed in 1973 would have spent a max­i­mum of 10 years on death row, while a per­son exon­er­at­ed today could have spent over 50 years. However, one would expect that with a func­tion­ing legal sys­tem that prop­er­ly iden­ti­fies wrong­ful con­vic­tions, the num­ber would plateau regard­less of the max­i­mum time spent on death row. For exam­ple, if the appeals process con­sis­tent­ly iden­ti­fied wrong­ful con­vic­tions with­in 15 years, the aver­age time on death row would increase until the late 1970s, at which point the data would show a rel­a­tive­ly sta­ble hor­i­zon­tal line around the 15-year mark through today. Instead, the aver­age length of time has sky­rock­et­ed, espe­cial­ly since the ear­ly 2000s. As the years pass, peo­ple con­tin­ue to be exon­er­at­ed who were sen­tenced to death at the very begin­ning of the mod­ern death penal­ty era. Last year, Oklahoma released Glynn Simmons, who was con­vict­ed and sen­tenced to death in 1975. He endured more than 48 years of appeals in which courts ignored and turned away his pleas of innocence. 

Though the devel­op­ment of DNA test­ing has great­ly assist­ed in the iden­ti­fi­ca­tion of decades-old wrong­ful con­vic­tions, the avail­abil­i­ty of DNA evi­dence does not appear to sig­nif­i­cant­ly affect the trend. Only one-sixth of death row exon­er­a­tions involved DNA. Many cap­i­tal con­vic­tions did not rely on any phys­i­cal foren­sic evi­dence, and when evi­dence was col­lect­ed that could be test­ed for DNA, it may have been improp­er­ly stored, con­t­a­m­i­nat­ed, or lost under the state’s care. Further, many pris­on­ers have dif­fi­cul­ty access­ing crime scene evi­dence over the objec­tions of pros­e­cu­tors so they can test it for DNA in post-con­vic­tion pro­ceed­ings. The dis­tri­b­u­tion of DNA exon­er­a­tions in our analy­sis tracks the over­all dis­tri­b­u­tion of exon­er­a­tions, with trend­lines show­ing only a mar­gin­al dif­fer­ence — and impor­tant­ly, the vast major­i­ty of the longest-serv­ing exon­er­at­ed cas­es did not involve DNA. ​“DNA is not the answer,” said Ray Krone, a DNA death row exoneree and co-founder of Witness to Innocence, upon Mr. Roberts becom­ing the 200^th exoneree. Mr. Krone said that DNA exon­er­a­tions ​“are like a canary in a coal mine, alert­ing us to the hard­er to prove fac­tors that cause most wrong­ful capital convictions.” 

One like­ly expla­na­tion for the trend involves long­stand­ing efforts to restrict the num­ber and scope of appeals, begin­ning with the Anti-Terrorism and Effective Death Penalty Act (AEDPA) in 1996. Death-sen­tenced pris­on­ers are guar­an­teed direct appeals — one review by the high­est state court and then a dis­cre­tionary appeal to the United States Supreme Court — but these appeals are lim­it­ed to the facts and evi­dence that were intro­duced at tri­al. Only after this direct review, in state ​“post-con­vic­tion” appeals, can pris­on­ers present new evi­dence and claims such as inef­fec­tive assis­tance of tri­al coun­sel, sup­pres­sion of evi­dence by pros­e­cu­tors, or new DNA results estab­lish­ing inno­cence. However, AEDPA imposed a strict one-year dead­line for fed­er­al post-con­vic­tion appeals, and nar­rowed the grounds upon which fed­er­al courts could grant post-con­vic­tion relief. Absent extra­or­di­nary cir­cum­stances, AEDPA also bars death-sen­tenced pris­on­ers from bring­ing more than one fed­er­al post-con­vic­tion appeal. In recent years, Supreme Court deci­sions inter­pret­ing the pro­vi­sions of AEDPA have fur­ther nar­rowed the legal avenues avail­able to prisoners. 

One prob­lem is that many of the core caus­es of wrong­ful con­vic­tions, such as offi­cial mis­con­duct, false wit­ness tes­ti­mo­ny, improp­er foren­sic evi­dence or ​“junk” sci­ence, and inad­e­quate legal rep­re­sen­ta­tion, can often only be brought as claims in state post-con­vic­tion appeals, and evi­dence to sup­port those claims is some­times slow to emerge by no fault of the pris­on­er. First, there is no fed­er­al con­sti­tu­tion­al right to coun­sel in state post-con­vic­tion pro­ceed­ings, forc­ing pris­on­ers to rely on pro bono lawyers or what­ev­er coun­sel the state may choose to pro­vide (or not) — and they gen­er­al­ly have no legal rem­e­dy for the inef­fec­tive per­for­mance of coun­sel dur­ing these state pro­ceed­ings. Strict state pro­ce­dur­al laws also require that all mer­i­to­ri­ous claims must be first pre­sent­ed to state courts or they are waived forever. 

Second, cap­i­tal cas­es often have many thou­sands of pages of records and dozens of wit­ness­es with poten­tial­ly valu­able insights in the case — but state statutes of lim­i­ta­tions and the one-year dead­line imposed by AEDPA, along with chron­ic under­fund­ing of defense coun­sel, mean that claims that must be made at this stage are too often missed. If claims are belat­ed­ly dis­cov­ered, pris­on­ers must over­come a moun­tain of pro­ce­dur­al hur­dles just to con­vince a court to con­sid­er the mer­its of the evi­dence. As a con­se­quence, many pris­on­ers with cred­i­ble inno­cence claims are effec­tive­ly boxed out of court due to dra­con­ian pro­ce­dur­al rules. In Barred: Why the Innocent Can’t Get Out of Prison, law pro­fes­sor Daniel Medwed writes that ​“the rule regime is stacked against the inno­cent, con­trary to the pop­u­lar belief that the post­con­vic­tion process is full of escape hatch­es from the prison cell, those imag­i­nary ​‘tech­ni­cal­i­ties’ that let peo­ple loose.” Prisoners ​“can have evi­dence of inno­cence — and no one will­ing to hear it.” 

The declin­ing num­ber of death sen­tences over­all may also play a role in the trend. In the 1990s, near­ly 300 peo­ple were sen­tenced to death each year, com­pared with few­er than 50 per year since 2015. Since 2017, every death row exon­er­a­tion has involved a sen­tence that is at least ten years old. 

Additionally, the caus­es of wrong­ful con­vic­tion offer insight into the increas­ing length of time before exon­er­a­tion. Most wrong­ful con­vic­tions are caused by more than one rea­son — a seri­ous indict­ment of the legal sys­tem. The chart below illus­trates rela­tion­ships between dif­fer­ent rea­sons for exon­er­a­tion and the preva­lence of each rea­son in the overall data. 

Our analy­sis finds that the role of offi­cial mis­con­duct in wrong­ful con­vic­tions can­not be over­stat­ed. Prosecutors and police con­duct the ini­tial inves­ti­ga­tion and main­tain con­trol over the evi­dence, includ­ing DNA, used to con­vict and sen­tence some­one to death. When they with­hold excul­pa­to­ry evi­dence, there is vir­tu­al­ly no way for the defense to prove their client’s inno­cence. Some state offi­cials defend even egre­gious police or pros­e­cu­tor mis­con­duct and try to block the defense from devel­op­ing inno­cence claims, using tac­tics like refus­ing access to the evi­dence for DNA test­ing — all of which fur­ther delays exon­er­a­tion. We found that 71% of all exon­er­a­tions involved offi­cial mis­con­duct, but the num­ber was 95% (18 of 19) for exon­er­a­tions that took 30 years or longer. Over half of all exon­er­a­tions involved both offi­cial mis­con­duct and perjury/​false accu­sa­tions. State admis­sions of wrong­do­ing appear to car­ry great weight for courts eval­u­at­ing inno­cence claims, but such admis­sions are the excep­tion rather the rule. 

State offi­cials have enor­mous pow­er over whether a per­son can be declared for­mal­ly exon­er­at­ed, and by exten­sion whether they can receive state com­pen­sa­tion for their wrong­ful incar­cer­a­tion. By far the most com­mon method of exon­er­a­tion in our data­base is a for­mal dis­missal of charges, a deci­sion made either by the pros­e­cu­tor or by a court, which often defers strong­ly to the prosecutor’s rec­om­men­da­tion. Nearly all the longest-delayed exon­er­a­tion cas­es end­ed in dis­missed charges, con­firm­ing the prosecutor’s pow­er­ful influ­ence over the process and the fail­ure of the oth­er meth­ods, par­tic­u­lar­ly court-based appel­late acquit­tal, to cor­rect these wrong­ful con­vic­tions. Even in the face of com­pelling evi­dence of inno­cence, many death-sen­tenced pris­on­ers have been forced to plead guilty to less­er charges and forego com­pen­sa­tion in order to avoid the prosecutor’s threat of anoth­er tri­al and pos­si­ble death sentence. 

In a study of wrong­ful con­vic­tions, law pro­fes­sors Jon B. Gould and Richard A. Leo char­ac­ter­ize our crim­i­nal legal sys­tem as ​“high­ly prone to error because of struc­tur­al truth-seek­ing flaws.” They find that ​“var­i­ous state actors (police, pros­e­cu­tors, judges) remain entrenched in a high­ly adver­sar­i­al mind-set in the post-con­vic­tion exon­er­a­tion process” and ​“no sin­gle state actor or orga­ni­za­tion seems will­ing to take respon­si­bil­i­ty declar­ing inno­cence or seek­ing exon­er­a­tions.” In fact, based on empir­i­cal analy­sis, ​“police and pros­e­cu­tors were the largest source of oppo­si­tion to an exon­er­a­tion.” The study finds that polit­i­cal con­cerns may also cre­ate bar­ri­ers to exon­er­a­tion, for exam­ple when for­mer Virginia Governor Doug Wilder had evi­dence of Earl Washington’s inno­cence but com­mut­ed his death sen­tence to life in prison instead of par­don­ing him. Mr. Washington was lat­er ful­ly exon­er­at­ed. Professors Gould and Leo sug­gest that pub­licly elect­ed offi­cials are ​“con­cerned about the ensu­ing uproar if con­vict­ed defen­dants are released when the evi­dence is ​‘mere­ly’ sug­ges­tive of inno­cence,” but the pub­lic should ​“expect greater open­ness by our jus­tice offi­cials to a defendant’s show­ing of inno­cence, even if pre­sent­ed post-tri­al.” (See the Death Penalty Information Center’s recent report, Lethal Election: How the U.S. Electoral Process Increases the Arbitrariness of the Death Penalty.)

Mr. Simmons, believed to be the longest wrong­ful­ly incar­cer­at­ed exon­er­at­ed per­son in the United States, illus­trates both the sys­temic fail­ures of appel­late courts to iden­ti­fy inno­cence and the last­ing prej­u­di­cial effects of offi­cial mis­con­duct. Accused of a con­ve­nience store rob­bery and mur­der in Oklahoma, he and co-defen­dant Don Roberts were iden­ti­fied in a line­up by the sur­viv­ing wit­ness, who had been shot in the head. Four wit­ness­es tes­ti­fied that Mr. Simmons was play­ing pool in Louisiana dur­ing the mur­der, but he and Mr. Roberts were con­vict­ed and sen­tenced to death based on the sur­viv­ing witness’s tes­ti­mo­ny. Twenty years after the con­vic­tion, the tri­al pros­e­cu­tor admit­ted that the wit­ness had want­ed to think about the iden­ti­fi­ca­tion ​“overnight” and had described the per­pe­tra­tor as being much taller and heav­ier than Mr. Simmons. ​“Quite can­did­ly, it was one of the few cas­es I have been involved in that the ver­dict a week lat­er could eas­i­ly have been dif­fer­ent,” the pros­e­cu­tor admit­ted. Despite these admis­sions, Mr. Simmons’ post-con­vic­tion appeals were denied. Then the defense team found police records, nev­er dis­closed to the defense, show­ing that the wit­ness had iden­ti­fied four oth­er men in line­ups, and filed a post-con­vic­tion peti­tion in 2023. Only then, with the sup­port of the District Attorney, was Mr. Simmons for­mal­ly exon­er­at­ed by a court after a half-cen­tu­ry of incar­cer­a­tion. But the DA reject­ed Mr. Roberts’ request to over­turn his con­vic­tion in June 2024 and he is still seek­ing formal exoneration. 

John Huffington and Kerry Max Cook are two oth­er recent exam­ples whose exon­er­a­tions were delayed for years due to pros­e­cu­to­r­i­al mis­con­duct. In Mr. Huffington’s case, District Attorney Joseph Cassilly with­held two fed­er­al reports from the defense for over a decade show­ing that an FBI agent had lied and mis­rep­re­sent­ed foren­sic evi­dence at tri­al. Instead of dis­clos­ing the reports, DA Cassilly said he had ​“decid­ed to wait a while to see if the defense files any post-con­vic­tion motions in the case” — essen­tial­ly let­ting Mr. Huffington use up his lim­it­ed chances for post-con­vic­tion appeals. At the same time, DA Cassilly fought Mr. Huffington’s motions to retest the evi­dence, con­tin­ued to with­hold evi­dence even after DNA test­ing exclud­ed Mr. Huffington as a source, and pur­sued retri­al even after a court vacat­ed Mr. Huffington’s sen­tence in 2013. Mr. Huffington was forced to enter an ​“Alford plea” in 2017 to avoid a retri­al. DA Cassilly was dis­barred in 2021 as a con­se­quence of his eth­i­cal vio­la­tions in the case, and Mr. Huffington was ful­ly par­doned in 2023, 41 years after his orig­i­nal sen­tence. Only then did he get com­pen­sa­tion from the state — $2.9 mil­lion — for his wrongful conviction.

In Mr. Cook’s case, pros­e­cu­tors with­held evi­dence through­out three tri­als and death sen­tences, includ­ing a secret deal with a ​“jail­house snitch,” and over­saw the destruc­tion of a poten­tial­ly excul­pa­to­ry human hair that the defense had sought to test for DNA. In June 2024, the Texas Court of Criminal Appeals ruled that Mr. Cook was ​“actu­al­ly inno­cent” and his case had been ​“marked by book­ends of decep­tion” span­ning 46 years. Like Mr. Huffington, Mr. Cook had pre­vi­ous­ly been forced to take a plea when pros­e­cu­tors threat­ened a new tri­al. The rul­ing means Mr. Cook will now be eli­gi­ble for about $3 mil­lion in compensation.

The increas­ing length of time before exon­er­a­tion deserves fur­ther research and atten­tion, but our analy­sis con­firms the sober­ing fact that inno­cent death-sen­tenced peo­ple are inad­e­quate­ly pro­tect­ed in our legal sys­tem. Marcellus Williams faces exe­cu­tion by Missouri in September despite mul­ti­ple DNA tests prov­ing that he did not pro­duce the bloody shoeprints, hair, and mur­der weapon DNA left at the crime scene. He has argued his inno­cence across 23 years of appeals. Prosecuting Attorney Wesley Bell sup­ports Mr. Williams’ inno­cence claim and filed a motion to vacate his con­vic­tion, but Attorney General Andrew Bailey tried to block an inno­cence hear­ing and force Mr. Williams’ exe­cu­tion. The Missouri Supreme Court recent­ly autho­rized the hear­ing to pro­ceed on August 21. A group of death row exonerees held a press con­fer­ence in Missouri on August 6 to call for Mr. Williams’ exon­er­a­tion. ​“The exe­cu­tion of an inno­cent per­son is an irre­versible trav­es­ty that should nev­er be accept­ed as an inevitable out­come,” said Herman Lindsey, death row exoneree and direc­tor of Witness to Innocence. Joseph Amrine, who spent 15 years on Missouri’s death row, said Missouri ​“has noth­ing to gain by killing the wrong person.”