A closer look at reversal rates — habeas is the big news

The suddenly famous Ellen Brotman commented here yesterday that CA3’s 5% reversal rate was lower than previous years. Very true. Reversals are dropping, led by a freefall in CA3 habeas reversals. I’ve whipped up some quick graphs to illustrate.

First, CA3’s overall reversal rate over the last 13 years:

At first glance that looks like a steep, steady drop. But notice that two years — 2006 with its 28% reversal rate, and 2013 with 5.4% — are outliers. Eight of the 13 years fall between 9% and 12%.

Here are CA3’s reversal rates for the two largest categories of cases, criminal and private civil:

So civil reversals are dropping, while criminal reversals spiked in 2006* but have otherwise have held relatively stable. Bottom line: private civil and criminal appeals don’t explain the 2013 reversal-rate plunge.

To get the plunge, we need to look at prisoner post-conviction appeals.** Here there are two groups: 2255 petitions for federal convictions, and habeas corpus petitions for state convictions. The numbers are startling:

Look at that habeas relief rate, the red line. For 10 years 2002 to 2011, CA3 reversed over 12% of cases every year, peaking in 2011 at 16%. Then, the elevator shaft: 6% in 2012, 3% in 2013. That’s a historic shift happening before our eyes.

Now look at the blue line, which is post-conviction challenges to federal convictions. Except for 2009,* that’s been a fairly steady downward trend, but now “downward” has become “dream on, appellant.” CA3’s 0.8% reversal rate was the lowest of any circuit in 2013. In fact, it was the lowest one-year reversal rate of any circuit since 2004. Hard to believe.

Aside: what explains this abrupt drop in post-conviction reversals? Did district court judges suddenly get more perfect? Unlikely. Did a landmark case require appeals courts to defer more to district courts? Don’t think so. On the habeas side, there were two big 2011 decisions (Richter and Pinholster) that make it harder for courts to grant habeas relief (and another one last week); if district courts started denying everything in 2011 and CA3 started affirming them all when they arrived on appeal a year later, the drop in habeas relief would make sense. I haven’t found habeas relief-rate stats, so that’s only hypothesis, and it shouldn’t explain the 2255-reversal-rate drop. But I strongly suspect the difference is that CA3 is now affirming a lot of denials that, in past years, it would have reversed.

Anyway. So how does CA3 compare to its sister circuits? Here’s a comparison of CA3’s overall reversal rate vs. the rate for all circuits:

Until the last couple years, CA3 was reversing more often; not any more. Here are the criminal and private civil reversals, CA3 and for all circuits:

Nothing too startling. Other than 2006, criminal and civil reversals nationally have been steady. Here’s a comparison of CA3 habeas and 2255 reversal rates vs. all circuits**:

So habeas & 2255 reversals are dropping sharply nationwide. And CA3’s big habeas-reversal drop (red) brings it in line with the national average (green) after a decade of above-average reversing.

Update: I have more analysis of habeas reversal rates in other circuits here.

The source for all this data is here and here — the table I used is B-5, the lines I used are criminal, U.S. prisoner petitions (which I’ve called 2255), private prisoner petitions (which I’ve called habeas), and other private civil.

* I assume the first big CA3 reversal spike — criminal direct appeals in 2006 — is the result of Booker‘s holding in 2005 that Apprendi applies to the Sentencing Guidelines. I don’t know offhand what caused the second CA3 spike, for 2255 appeals in 2009; nationally, the 2255 reversal rate went down that year.

** Update:  The available numbers defy precise answers, and the stats I’ve posted obscure some of the complexity. The stats I gave above for 2255 cases actually are the stats for all US prisoner petitions. That includes 2255s, but, it also includes prisoner civil rights and conditions cases; roughly, 2255s were about 84% of the category nationally. The same caveat applies to the habeas stats I gave; in 2013 habeas cases were about 55% of their category. (Source: 2013 table B-7). Bottom line, it’s no easy thing to pin down causes of the drop in reversals.