Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 10 CR 907-1--Samuel Der-Yeghiayan, Judge.
The opinion of the court was delivered by: Posner, Circuit Judge.
Before POSNER, MANION, and TINDER, Circuit Judges.
The defendant pleaded guilty to conspiring to make and sell false identification docu- ments, such as documents identifying the bearer as a permanent resident of the United States, in violation of 18 U.S.C. §§ 1028(a)(1), (2), (f). His guidelines sentencing range was 37 to 46 months, but the judge sentenced him to 60 months, as urged by the government, a sentence that although above the guidelines range was well within the statutory maximum of 15 years. §§ 1028(b)(1)(A), (B). He has appealed, challenging his sentence, but his lawyer has filed an Anders brief in which he argues that there is no valid ground for challenging the sentence and asks to be allowed to withdraw from representing the defendant.
We write to clarify an ambiguity concerning the scope of appellate review of an above-guidelines sen- tence. We have said that "the farther the judge's sentence departs from the guidelines . . . the more compelling the justification based on factors in section 3553(a) that the judge must offer in order to enable the court of appeals to assess the reasonableness of the sentence imposed." United States v. Courtland, 642 F.3d 545, 550 (7th Cir. 2011), quoting United States v. Dean, 414 F.3d 725, 729 (7th Cir. 2005); see also Gall v. United States, 552 U.S. 38, 50 (2007);
United States v. Bradley, 675 F.3d 1021, 1025 (7th Cir. 2012) (per curiam). The ambiguity is in the word "farther." It can be conceived of in either relative or absolute terms. A sentence of 60 months is 30 percent longer than a sentence of 46 months (the top of the applicable guide- lines range in this case); and a 30 percent increase is large in relative terms. But in absolute terms, given the severity of federal criminal punishments, it is a smallish 14 months; the average federal prison sentence in 2009 was 57 months. Mark Motivans, "Federal Justice Statistics 2009--Statistical Tables" 27 (Bureau of Justice Statistics, U.S. Dep't of Justice, Dec. 2011) (table 5.4), http://bjs.ojp.usdoj.gov/content/pub/pdf/fjs09st.pdf (visited Aug. 9, 2012).
It seems to us that the relative is generally more impor- tant than the absolute, as is implicit in a number of our previous decisions. See United States v. Snyder, 635 F.3d 956, 961 (7th Cir. 2011) ("the sentence imposed by the court was over two-and-a-half times greater than [the guidelines range]"); United States v. Munoz, 610 F.3d 989, 995 (7th Cir. 2010) (sentence "roughly 50% more than his advisory guideline range"); United States v. Miller, 601 F.3d 734, 739-40 (7th Cir. 2010) ("sentence that was fifty percent above the high end of the advisory Guidelines range"); United States v. Kirkpatrick, 589 F.3d 414, 415 (7th Cir. 2009) ("more than double" the guidelines range); United States v. Jackson, 576 F.3d 465, 470 (7th Cir. 2009) (same); United States v. Higdon, 531 F.3d 561, 563 (7th Cir. 2008). The guidelines range is the Sentencing Com- mission's estimate of the reasonable range of punish- ments for the defendant's offense. Usually (an important qualification, as we're about to see), a judge who imposes a sentence far above the top or far below the bottom of that range is challenging the Commission's penal judgment, and given that the Commission's knowl- edge of penology exceeds that of most judges, the judge needs to provide more in the way of justification than if he were departing incrementally.
Guidelines ranges are inherently arbitrary, so had the judge in this case sentenced the defendant to 47 months instead of the guideline maximum of 46 it would not have been a significant challenge to the Commission's penal judgment and so would not have required much in the way of justification. A 30 percent departure requires more; "substantial variances from the Sentencing Commission's recommendations require careful thought." United States v. Kirkpatrick, supra, 589 F.3d at 415. Yet less thought is necessary when the applicable guide- line is "not the product of the Commission acting in 'its characteristic institutional role,' in which it typically implements guidelines only after taking into account 'empirical data and national experience.'" United States v. Reyes-Hernandez, 624 F.3d 405, 418 (7th Cir. 2010), quoting Kimbrough v. United States, 552 U.S. 85, 109 (2007). And that is the case here. The Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. 104-208, 110 Stat. 3009, required the Commission to increase the base-offense level for the false-document offense by two levels and the sentencing enhancement for number of documents by "at least 50 percent." Id., §§ 211(b)(2)(A), (B). The Act also directed the Commis- sion to "consider whether any other aggravating or miti- gating circumstances warrant upward or downward sentencing adjustments," § 211 (b)(2)(E), and shortly afterward the Commission added to the guideline, as we'll see, an open-ended upward-departure provision. In these circumstances, the judge, not having to confront an exercise of considered penal expertise (so far as ap- pears) by the Commission, was under less compulsion to provide a comprehensive explanation for giving a sentence substantially above the top of the guidelines range.
We acknowledge that focus on the sentencing judge's percentage deviation from the guidelines range can mislead, at least when the sentence is below rather than, as in this case, above the sentencing range; an example given in United States v. Gall, supra, 552 U.S. at 48, is that "a sentence of probation [the sentence the Court was reviewing] will always be a 100% departure regardless of whether the Guidelines range is 1 month or 100 years." But it's hard to see how a court can carry out the command of Gall to require a justification "sufficiently compelling to support the degree of the variance," 552 U.S. at 50 (emphasis added)--"degree" being a relative rather than absolute measure--without at least con- sidering the percentage deviation. We are not the only court to give weight to it. See United States v. Ressam, 679 F.3d 1069, 1089-90 (9th Cir. 2012) (en banc); United States v. Irey, 612 F.3d 1160, 1196 (11th Cir. 2010) (en banc); In re Sealed Case, 527 F.3d 188, 192 (D.C. Cir. 2008). Granted, there is resistance, see, e.g., United States v. Burns, 577 F.3d 887, 905 n. 8 (8th Cir. 2009) (en banc); United States v. Evans, 526 F.3d 155, 166 n. 5 (4th Cir. 2008)--and a note of ambivalence in our own opinions in United States v. Brown, 610 F.3d 395, 398 (7th Cir. 2010), and United States v. Omole, 523 F.3d 691, 698 n. 1 (7th Cir. 2008).
The judge's failure in the present case to give extended consideration to the appropriateness of a 30 percent departure, large as that departure is in relative terms, is easily excused. The defendant's lawyer, while recommending a much lower sentence--a below-guide- lines sentence of 30 months--did not challenge the gov- ernment's recommendation for 60 months on the ground that it exceeded the top of the guidelines range.
More important is a curious feature of the applicable guideline. The guideline range for the false-document offense rises in stages as the number of false identification documents increases--but only up to 100. U.S.S.G. § 2L2.1(b)(2). It is unclear why the count stops at 100, a suspiciously round number. Unclear to the Sentencing Commission itself, perhaps, because Application Note 5 to the guideline says that if the defendant's crime involves substantially more than 100 documents an upward departure "may be warranted." This is what is called an "encouraged departure," Koon v. United States, 518 U.S. 81, 94 (1996); see also United States v. Miller, 343 F.3d 888, 891-92 (7th Cir. 2003)--encouraged by Congress, as we pointed out. A departure is a departure; it is a sentence above or below the applicable guidelines range. United States v. Guyton, 636 F.3d 316, 319-20 (7th Cir. 2011); see also United States v. Ortega-Galvan, 682 F.3d 558, 562 (7th Cir. 2012). And though the encouragement to depart in cases like this one came from Congress rather than from the Commission and is not based (so far as ap- pears) on any empirical study, it makes common sense and provided a solid basis for a substantial increase in sentence above the top of the guidelines range in this case.
Under the guideline, as the number of false docu- ments increased from 99 to 100 or more, Castillo's offense level rose by three points, §§ 2L2.1(b)(2)(B), (C), which translates into an increase in the guidelines range from 27-33 months to 37-46 months--a 39 percent increase in his maximum guidelines sentence. He was estimated to have been responsible for 2800 documents. We don't see how imposing a sentence 30 percent above the guidelines range could be thought excessive punish- ment for 28 times the number of fraudulent documents that triggers the highest guideline sentence, when a 1 percent increase in documents (from 99 to 100) would have increased his maximum guideline sentence by 39 percent. It thus was icing on the cake that the judge noted additional aggravating factors besides the number of documents, such as the defendant's marketing computer software for producing still more counterfeit personal-identification documents.
United States v. Burns, supra, 577 F.3d at 905 n. 8, along with other Eighth Circuit decisions cited in it, and United States v. Morace, 594 F.3d 340, 345 (4th Cir. 2010), recommend using number of offense levels rather than percentage deviations in sentence length to gauge the magnitude of a variance between the sentence imposed and the guidelines sentence. So let's do that as a check on our analysis. The guideline in question provides no increase in offense levels as the number of documents increases from 1 to 5. When it hits 6, how- ever, there is a 3-level increase; at 25 there is a 6-level increase and at 100 a 9-level increase. So from 6 to 100 each quadrupling of the number of documents adds three offense levels. If we extrapolate, 400 documents would earn a 12-level increase and 1600 documents a 15-level increase. This would mean that just as going from 6 documents to 100 (two consecutive quadruplings) generates a 6-level increase (3 levels from 6 to 25 and 3 more from 25 to 100), going from 100 to 400 and then 1600 (again, two consecutive quadruplings) would generate a further 6-level increase. So if the defendant in our case had produced only 1600 documents rather than 2700, the ...