Davila Won’t Open the Flood Gates: Answering Justice Breyer’s Request for Empirical Data

April 26, 2017
Guest Post

by Brian W. Stull, Senior Staff Attorney, Capital Punishment Project at the American Civil Liberties Union

Monday’s argument in Davila v. Davis involved a technical question of habeas corpus procedure: can a prisoner’s failure to raise the ineffectiveness of his direct appellate counsel in a collateral state post-conviction proceeding be excused by the ineffectiveness of the post-conviction lawyer? But the answer may turn on the more familiar judicial concern over opening floodgates. Some of the justices appeared concerned that the extension of Martinez v. Ryan and Trevino v. Thaler Davila proposed would do just that. To ground that discussion in facts, Justice Breyer asked for empirical data. While the advocates did not have ready answers at the lectern, answers were relatively easy to find at the computer.

Justice Breyer asked petitioner Davila’s attorney to “what extent has the Martinez claim proved a burden on Federal court? Is there any empirical information?” Counsel did not have a ready answer. Justice Breyer later asked the same of the Solicitor General of Texas, who repeated the figure of 3,800 cases, which Texas had previously cited in its Respondent’s Brief.

It is easy to reproduce this figure by searching on Westlaw for those district court decisions citing either Martinez or Trevino. Sure enough, the number is nearly exact (3,837). But spot checking of the 3,800 leads to at least three pertinent observations. First, not all cases citing one of the two decisions involved a claim under Martinez/Trevino. A court may cite Martinez or Trevino for a proposition stated within, without deciding a claim based on the decisions.

Second, filtering out those decisions in cases brought by pro se litigants leaves less than 700 decisions citing Martinez or Trevino. The Court will reach its own conclusions about the weight of the burden of pro se filings, but it should be uncontroversial to say that many pro-se petitions (even assuming they raise Martinez/Trevino claims) will be quickly dismissed based on any number of problems inherent in petitions drawn up by litigants unschooled in the law.

Third, and overlapping with the pro-se observation, where litigants did bring claims under the decisions, they were often creative to the point of being easily dismissed by the courts, such as arguments to toll the statute of limitations for federal habeas petitions. The true burden of Martinez and Trevino, then, may be less than Texas’s figure suggests.

Considering the 3,800 figure, Justice Breyer asked Texas’s Solicitor General “out of, roughly, how many habeas petitions?” Texas did not have those statistics, but they are readily available. The Administrative Office of the United States Courts collects this data (and much more). It annually records the number of habeas corpus cases decided by federal district courts reviewing state-court judgments. In the four full years elapsed since Martinez v. Ryan was decided in 2012, the federal district courts disposed of 16,218 non-capital and 182 death-penalty habeas petitions in 2013, 16,899 (non-death) and 166 (death) in 2014, 15,927 and 211 in 2015, and 15,770 and 201 in 2016. The sum total of these cases disposed (not including the partial years 2012 and 2017) is nearly 65,000. Reviewing instead the number of habeas petitions filed in these same years, the numbers are very similar – roughly 16,000 annually, from 2013 to 2016, a figure declining slightly each year after Martinez. Whether evaluated as a fraction of the total number of petitions filed or decided Texas’s figure of 3,800 decisions (even if not discounted) represents a tiny fraction of habeas cases, and if anything the number has declined since Martinez.

Chief Justice Roberts and Justice Alito looked at the issue from a different angle – questioning how many additional states would give rise to Martinez and Trevino claims if Davila prevailed. In Trevino, the Court decided that because in Texas the first effective opportunity to raise a claim of ineffective trial counsel based on needed information outside of the record came in state post-conviction review (as in most states), that post-conviction counsel’s incompetent failure to raise ineffective assistance of trial counsel could serve as cause for excusing procedural default. As amici for the state of Utah and 24 other states in support of Texas argued in that case – with a detailed documentation of the appellate and post-conviction systems of every state – the same could be said “virtually nationwide,” i.e., for nearly all of the states not already covered under Martinez. Indeed, review of the 3,800 decisions discussed above also reveals that discussions of Martinez and Trevino (or in some cases, claims) have come from district courts reviewing state judgments in every single state but Vermont. So the answer, as a matter of counting states, is that Davila does not open the gates any wider.

The ultimate question is the effect of Davila should the petitioner prevail. Will litigants, as Texas has argued “sandbag on state habeas and save their appellate claims for federal habeas[?]” This would not be a data-driven strategy. As a Bureau of Justice Statistics study shows, 15 percent of appellants convicted of felonies and 30 percent of those sentenced to death win their direct appeals to the intermediate state appellate courts generally charged with addressing trial error. By contrast, empirical data shows that less than one percent will prevail on habeas review, even if it is de novo. Saving a claim of trial error for federal habeas review under the rubric of ineffective assistance of appellate counsel thus appears a risky gambit.

As our amicus brief (with the National Association of Criminal Defense Lawyers) in Davila was intended to show, “[t]he right to effective appellate counsel is particularly important in light of the critical role the direct appeal plays in criminal cases.” Fears of expending too much justice to protect that right in Davila may be overblown.