SENTENCING MEMORANDUM

GERTNER, District Judge.

Anthony Oliveira ("Oliveira") pled guilty to the crime of felon-in-possession of a firearm in violation of 18 U.S.C. § 922(g)(1). On July 18, 2006, Oliveira, who had a lengthy felony record, possessed an Imez .38-caliber pistol and ammunition. His sentencing raises a number of substantial issues.

First, his sentence will be determined not merely by the offense and his criminal record. It will be determined by one of the federal recidivist enhancement statutes, the Armed Career Criminal Act ("ACCA"), 18 U.S.C. § 924(e)(1), which increases the ten year maximum sentence associated with felon-in-possession of a firearm to a fifteen year mandatory minimum sentence. Oliveira had previously been convicted of a series of felonies, including two larceny from a person convictions, one resisting arrest, and one assault and battery upon a guard. His counsel challenges whether the two larcenies and the assault and battery should qualify as crimes of violence under the ACCA. Three predicate offenses are required for ACCA sentencing.

Second, Oliveira's sentence will be substantially impacted by the three other gun and drug transactions which the government chose not to charge, merely adding them as "relevant conduct" under the Sentencing Guidelines. As a result of that prosecutorial decision, I am obliged to evaluate these allegations using the sentencing burden of proof, preponderance of the evidence, rather than beyond a reasonable doubt, the trial burden. Should I find that the uncharged conduct occurred as alleged, Oliveira's Guideline score would be increased by nearly six points. Guideline enhancements are triggered because the transactions in total involved four guns (the charged transaction and three others), and the guns were possessed "in connection with" an uncharged drug offense. See United States Sentencing Guidelines Manual ("U.S.S.G.") § 2K2.1(b)(1)(A), (b)(4)(A), (b)(6).

Finally, Oliveira's sentence will be determined by what has been called "imperfect entrapment" under U.S.S.G. § 5K2.12, which is "aggressive encouragement of wrongdoing" not amounting to a complete defense. United States v. Garza-Juarez, 992 F.2d 896, 912 (9th Cir.1993). See also United States v. Bala, 236 F.3d 87, 90 (2d Cir.2000). A man who was at best an ordinary thief—with no gun sales in his record, surely nothing on the scale of the conduct the government's witness engineered—was transformed into a gun dealer because of the government's unsavory and not entirely credible cooperating witness.

Sentencing took place over three days, with video and audio exhibits, testimony from an agent and the government's cooperating witness. I concluded that the ACCA does not apply, that larceny in Massachusetts is a stealth crime, not a crime of violence, and should not be considered an ACCA predicate offense. The paradigmatic larceny in Massachusetts involves a pickpocket—with no weapon, no one placed in fear, no one even aware that a crime has been committed. I further concluded that I was obliged to consider the "uncharged conduct" under the Sentencing Guidelines, as the government urged. The "relevant conduct" approach, however criticized,1 is the law. But I will evaluate that uncharged conduct in the context of the very real entrapment that the government's cooperator engendered.

For the reasons I describe below, I sentenced Oliveira to 100 months' incarceration and 60 months of supervised release— not the fifteen years the ACCA required, if Oliveira qualified under its provisions, nor the ten year maximum sentence that the drug statute would have required. (The Guideline score of 168 to 210 months would have been trumped by the ten year maximum penalty of the offense of conviction, 18 U.S.C. § 924(a)(2).) Make no mistake: This is a substantial sentence, particularly for a man who has never served a term of imprisonment remotely at this level. It is proportionate, meets the purposes of sentencing under 18 U.S.C. § 3553(a) and, more important, it is fair.

I. FACTS

A. Offense Conduct

This case arose from a gun sale by Oliveira to a confidential witness, Tom Heroux ("Heroux"), working for the Bureau of Alcohol, Tobacco, and Firearms ("ATF"). Heroux had been an informant for a considerable period of time. He had a long criminal history, including four charges for possession of a dangerous weapon or assault and battery with a dangerous weapon. See Heroux CORI ¶¶ 63, 65, 93, 98 (document # 59-1). There was no question that he, unlike Oliveira, had substantial experience with firearms and their sales. Oliveira claims, and the record supports, that Heroux instigated the gun transactions at issue in this case, encouraging Oliveira to deal more and more because he was willing to purchase more and more—all with government money. Although Oliveira had certainly offended before, he was a thief, not a gun dealer.

Heroux befriended Oliveira. Heroux's relationship with Oliveira derived from the relationship of the women in their lives. Heroux's wife, Rayna Heroux, was the personal caregiver to Oliveira's long-term girlfriend, Susan Facchetti, who suffered from multiple sclerosis.2 It was an odd friendship. Oliveira was African American. Heroux sported tattoos of a swastika on his arm and the words "White Pride" on his torso. His truck had a Confederate flag on it.

On November 27, 2005, Heroux informed Special Agent Robert White ("Agent White") that Oliveira had offered him numerous handguns for sale. I do not credit Heroux's information but Agent White did. Heroux has an extensive criminal record, with over 100 charges listed on his CORI form, including a charge for Possession of a Firearm (CORI ¶ 99), four charges for Possession of a Dangerous Weapon or Assault and Battery Dangerous Weapon (CORI ¶¶ 63, 65, 93, 98), Escape (CORI ¶¶ 57, 59), and numerous drug possession and theft offenses. His main source of income seemed to derive from being a government informant, participating in sting after sting.

Agent White subsequently set up an undercover operation in which, between January 26, 2006, and April 12, 2007, Heroux arranged four separate undercover purchases involving firearms, ammunition, and drugs at Oliveira's home. To the extent that they could (some of the equipment malfunctioned), ATF monitored these transactions via covert audio and visual recording equipment.

The government only chose to indict Oliveira on one offense, the one that took place on July 18, 2006. The day before, Oliveira had contacted Heroux and offered to sell him a pistol for $600. Heroux said that he would pay $100 more if Oliveira would hold the pistol an extra day. ATF agents followed Heroux to Oliveira's residence, then monitored the interaction through audiovisual monitoring devices. Video footage shows Heroux entering Oliveira's bedroom and asking for the gun, which Oliveira removed, along with ammunition, from a red backpack near his bed. Heroux subsequently paid Oliveira $700 in cash. The gun was later examined by an expert, who identified it as a Russian-manufactured Imez .38mm caliber pistol. The firearm was manufactured outside of Massachusetts and had been reported stolen four years before. There was no indication that Oliveira knew its status.

In addition to the sale on July 18, 2006, however, Heroux and Oliveira engaged in the following transactions, which the government chose not to charge: Six months before, on January 26, 2006, Heroux purchased a shotgun from Oliveira for $400 as well as three bags of cocaine for $100. Then, on February 25, 2006, Heroux purchased a 30-30 rifle from Oliveira for $500 and two $50 rocks of cocaine. Almost a year later, on April 12, 2007, Heroux purchased a loaded handgun from Oliveira for $800.3 There is no explanation for the hiatus. It took another year before the government chose to arrest Oliveira. Apart from Heroux's deals, there was no indication that Oliveira ever sold guns to anyone else, before. after or during this period.

It is not at all clear that without Heroux's exhortations and government money, Oliveira would have been dealing in guns at all. While he apparently knew where to get them, and showed some sophistication (putting on gloves, for example, when touching the guns), the transactions show him to be a broker—pairing Heroux with other dealers whose identity he knew. Heroux was very comfortable and familiar with guns and gun sales—not so Oliveira. Heroux often appeared to be the person in charge, explaining pricing and logistics to a relatively passive Oliveira. The government offered no information as to whether the dealers—the sources of the guns, the "big fish," to whom Oliveira was arguably leading—were ever indicted.

The failure to charge the three "relevant conduct" transactions is telling. If charges had been brought with respect to these transactions, and particularly, if the sources of the guns had been included, Oliveira would have appeared to be even less of a player. Indeed, under those circumstances, I would have considered a downward role adjustment. Moreover, the government's failure to formally charge these offenses—in the face of audiotapes and videotapes—strongly suggests that they were less than confident in their cooperating witness before a jury.

B. Defendant's Background

Oliveira is forty-four years old. His father left when he was three years old; he was raised by his mother in a stable home environment. She worked as a certified nurse's aide and owned their family home in Wareham, MA. Oliveira was very close to his mother. He was devastated when she was diagnosed with brain cancer in 1982 and died a year later at the age of forty-five. Oliveira was seventeen. He was so overwhelmed by her death that he could not attend her funeral and collapsed outside the service. He subsequently dropped out of school and moved around between family and friends. Almost immediately, he got into trouble; over the next few years he was arrested many times. From age eighteen on he was, for the most part, a petty thief—breaking and entering, larceny with a credit card, forgery, shoplifting, receiving stolen property—no guns, and surely no gun dealing. He developed an alcohol and drug habit; he was eventually hospitalized with heart problems in 1997. He has been drug-free since 2005.

Oliveira has had four serious relationships with women. Significantly, three of them have sought and obtained protection orders against him. He is currently in a long-term relationship with Shelley Couto, whom he plans to marry during his incarceration. He is the father of four children.

II. GUIDELINE CALCULATION

The Guidelines for Oliveira's offense are calculated as follows:

I accept the Guidelines calculation prepared by Probation in the Pre-Sentence Report ("PSR"), although as I explain below, I depart downward from the Guideline range for what has been called "imperfect entrapment" under U.S.S.G. § 5K2.12. See Garza-Juarez, 992 F.2d at 912 (affirming downward departure where there is "aggressive encouragement of wrongdoing [by the informant]"). But the Guidelines frame is a small part of this sentencing picture. Oliveira's sentence will be driven by two statutes. If I found him to be an armed career criminal under the ACCA, he would be subject to the mandatory minimum of fifteen years under 18 U.S.C. § 924(e). If I found that he was not an armed career criminal under the ACCA, the guideline range of 168-210 months would be trumped by a statutory maximum for the offense of ten years under 18 U.S.C. § 924(a)(2).

Below, I address first the ACCA; I then turn to the Guidelines calculations and the 18 U.S.C. § 3553(a) factors.

III. ARMED CAREER CRIMINAL ACT

The threshold question in this case is whether Oliveira is an armed career criminal. In light of a series of recent Supreme Court cases on the so-called residual clause of the ACCA, this is not a question with an obvious answer.

The ACCA was enacted in 1984 with the aim of targeting repeat armed offenders.4 After a few amendments in the interim, the ACCA underwent substantial changes as part of the Anti-Drug Abuse Act of 1986, Pub.L. No. 99-570, §§ 1401-1402, 100 Stat. 3207, 3207-39 to 3207-40, which brought it to its present form in 18 U.S.C. § 924(e). Notably, in the subsequent amendments, the range of predicate convictions were extended far beyond the original offenses of robbery and burglary. At the same time, that expansion dramatically complicated the statute's interpretation.

Since most criminal law is defined by state statutes, the predicate offenses on which the ACCA is based necessarily require translating state crimes into the federal categories, frequently with anomalous results. For example, an individual could be considered an armed career criminal no matter how long ago the predicate offense was committed. In the instant case, the alleged predicate offenses were committed when Oliveira was 22, 25, 26, and 28. He is now 44. Nor is there any opportunity for collateral challenges to the predicate convictions under the ACCA, as there is under 21 U.S.C. § 851, the enhancement to drug sentences based on prior convictions.

Second, the ACCA categories are less than clear. It is surely troubling that substantial punishment enhancement should turn on as ambiguous a category as the "residual" clause of the ACCA, sweeping within it offenses that "otherwise involve[] conduct that presents a serious potential risk of physical injury to another." § 924(e)(2)(B). A bedrock principle of criminal law is the rule of lenity, that criminal laws are to be narrowly construed in favor of the defendant. See, e.g., United States v. Villanueva-Sotelo, 515 F.3d 1234, 1246 (D.C.Cir.2008) ("The rule of lenity would forbid us from `interpreting] a federal criminal statute so as to increase the penalty that it places on an individual when such interpretation can be based on no more than a guess as to what Congress intended.'" (quoting Ladner v. United States, 358 U.S. 169, 79 S.Ct. 209, 3 L.Ed.2d 199 (1958))).

In this concern, I join the ranks of several judges and justices. See Chambers v. United States, 555 U.S. 122, 129 S.Ct. 687, 694, 172 L.Ed.2d 484 (2009) (Alito, J., concurring) ("After almost two decades with Taylor's `categorical approach,' only one thing is clear: ACCA's residual clause is nearly impossible to apply consistently. . . . What is worse is that each new application of the residual clause seems to lead us further and further away from the statutory text."); Sykes v. United States, ___ U.S. ___, 131 S.Ct. 2267, 2287, 180 L.Ed.2d 60 (2011) (Scalia, J., dissenting) ("The residual-clause series will be endless, and we will be doing ad hoc application of ACCA to the vast variety of state criminal offenses until the cows come home. That does not violate the Constitution. What does violate the Constitution is approving the enforcement of a sentencing statute that does not give a person of ordinary intelligence fair notice of its reach, and that permits, indeed invites, arbitrary enforcement. The Court's ever-evolving interpretation of the residual clause will keep defendants and judges guessing for years to come." (internal quotation marks and citations omitted)). See also Phillip M. Spector, The Sentencing Rule of Lenity, 33 U. Tol. L. Rev. 511 (2002).

The only arguable justification for permitting this ambiguity is that the ACCA enhancement involves sentencing factors to be decided by a judge, rather than substantive elements to be decided by a jury. As such they can be vague—as vague as other aggravating or mitigating sentencing criteria may be. The categorization of predicate offenses as sentencing factors rather than as substantive elements relies on the Supreme Court's decisions in Harris v. United States, 536 U.S. 545, 122 S.Ct. 2406, 153 L.Ed.2d 524 (2002), and Almendarez-Torres v. United States, 523 U.S. 224, 118 S.Ct. 1219, 140 L.Ed.2d 350 (1998). The Court held that factors determining the mandatory minimum are sentencing factors, Harris, 536 U.S. at 550, 568-69, 122 S.Ct. 2406, and especially so, when they are prior convictions, Almendarez-Torres, 523 U.S. at 243-44, 118 S.Ct. 1219.

But whatever the continuing validity of these cases,5 the application of the ACCA requires careful consideration. The ACCA factors do more than enhance a sentence to some degree. They trigger a substantial mandatory minimum sentence and worse, rest not on the fact of a conviction, but a post hoc characterization of what that conviction necessarily entailed.

To that post hoc exercise, I will now turn:

A. ACCA Standards

The ACCA provides that where a felon found to be in possession of a firearm has had three previous convictions for a "violent felony" or serious drug offense, he must be sentenced to a mandatory minimum prison term of fifteen years. 18 U.S.C. § 924(e)(1).

In the case at bar, Oliveira had previously pled guilty to four felonies that arguably qualify: assault and battery by a prisoner on a guard (which occurred on July 26, 1988, when he was 22 years old), two instances of larceny from a person (on Nov. 16, 1991 and Sept. 27, 1992, when he was 25 and 26), and resisting arrest (on September 23, 1994, when he was 28).6

At issue in this case is whether larceny from a person and assault and battery of a guard are violent offenses under 18 U.S.C. § 924(e)(2)(B).7 Because I find that the former offense does not qualify under the ACCA, I need not address the latter.

§ 924(e)(2)(B) (emphasis added). The first provision (i) is generally referred to as the "force clause," and the italicized portion of (ii) is generally referred to as the "residual clause." In other words, an offense that is not enumerated (burglary, arson, extortion, or an offense that involves explosives) will qualify as a "violent felony" under the ACCA only if it falls within either the force clause or the residual clause.

1. Categorical Approach

To determine whether a particular offense—not enumerated in the statute itself—fits into one of these two provisions, the court applies a "categorical approach." The court considers the elements of the offense, and their interpretation by the highest state court, and asks whether those elements constitute a violent felony under the ACCA, "without inquiring into the specific conduct of this particular offender." James v. United States, 550 U.S. 192, 202, 127 S.Ct. 1586, 167 L.Ed.2d 532 (2007). See also Taylor v. United States, 495 U.S. 575, 600, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990). The categorical approach is critical. It reflects Sixth Amendment concerns, not merely a policy preference. Just as a judge cannot rely on facts not found by a jury to increase the sentence under United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), he or she cannot do so to determine whether a prior conviction is a violent offense. The court must be bound by the specific elements found by a jury or admitted by the defendant.

Only where the crime includes a range of conduct, some of which falls outside the ACCA's definition, may the court "peek under the coverlet" to determine what offense was at issue in the underlying conviction. Gautier, 590 F.Supp.2d at 222 (noting that we may not peek very far— and may look only to the formal language of jury instructions, charging document, plea agreement, transcript of the plea colloquy, or any facts confirmed by the defendant at sentencing and any comparable judicial record (citing Shepard v. United States, 544 U.S. 13, 16, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005); United States v. Winter, 22 F.3d 15, 18 (1st Cir.1994); United States v. Holloway, 630 F.3d 252, 259 (1st Cir.2011))), overruled on other grounds by Almenas, 553 F.3d 27, and Weekes, 611 F.3d 68.

Where the defendant is convicted under a statute that encompasses both violent and non-violent offenses (as in assault and battery), and the government has failed to show under Shepard that the defendant was convicted of the more violent offense, the court applies the categorical approach to the non-violent offense. See generally Holloway, 630 F.3d at 262 (applying the categorical approach to reckless battery since the government had not established harmful battery). Since, by definition, the non-violent version of the offense does not include an element of force and thus does not fall within the force clause, § 924(e)(2)(B)(i), it will be deemed a "violent felony" only if it falls within the residual clause of 924(e)(2)(B)(ii): "burglary, arson, or extortion, involves use of explosives, or otherwise involves conduct that presents a serious potential risk of physical injury to another."

Critically, force is not an element of the offense of larceny of a person. As such I must determine whether that offense falls within the residual clause of the ACCA.8 Before I turn to that question, however, I must decipher what the residual clause means today in light of changing case law. When I do, I am obliged to conclude that this offense does not fit within the residual clause.

2. Interpreting the Residual Clause

The residual clause of the ACCA has become something of a moving target. The Supreme Court has issued four opinions interpreting the clause in as many years and provided at least four tests to determine whether a crime meets its definition.

In James, the Supreme Court applied the "closest analog" test. 550 U.S. at 203, 127 S.Ct. 1586. The court asked whether the risk posed by the felony at issue "is comparable to that posed by its closest analog among the enumerated offenses." Id. (concluding that attempted burglary is as risky as burglary). The question is whether the crime, as it is ordinarily committed, entails a serious risk of physical injury. We are to ignore the outlier conduct, where a crime may be committed in such a way as to pose no risk at all or where the crime may be committed in such a way as to pose an extreme risk of physical injury. Id. at 208 ("One can always hypothesize unusual cases in which even a prototypically violent crime might not present a genuine risk of injury—for example, an attempted murder where the gun, unbeknownst to the shooter, had no bullets . . . As long as an offense is of a type that, by its nature, presents a serious potential risk of injury to another, it satisfies the requirements of § 924(e)(2)(B)(ii)'s residual provision." (emphasis added) (internal quotation marks and citations omitted)).

In Begay, in turn, the Supreme Court applied what I describe as the "purposeful, violent, and aggressive" test. Begay v. United States, 553 U.S. 137, 128 S.Ct. 1581, 170 L.Ed.2d 490 (2008). Begay involved a two-part analysis: (1) whether the offense "involves conduct that presents a serious risk of physical injury to another;" and (2) whether the crime is "roughly similar, in kind as well as in degree of risk posed" to the enumerated offenses. Id. at 143, 128 S.Ct. 1581 (internal quotation marks and citations omitted). An offense is similar in kind to the enumerated offenses if it involves "purposeful, violent, and aggressive conduct." Id. at 145, 128 S.Ct. 1581 (concluding that driving under the influence is not a violent felony under the residual clause of the ACCA). The Court explained that Congress did not intend to cover every offense that presents a serious risk of physical injury to another, but rather only those crimes that are similar to the enumerated offenses: burglary, arson, extortion, or offenses involving explosives. Id. at 144, 128 S.Ct. 1581.

In the following term, the Supreme Court added a statistical analysis to the inquiry. In Chambers, the Court determined that the offense of "failure to report" for incarceration "is a far cry from the `purposeful, violent and aggressive' conduct" at issue in the enumerated offenses. Chambers, 129 S.Ct. at 692 (quoting Begay, 553 U.S. at 144-45, 128 S.Ct. 1581). The Court seemed convinced by statistical information provided by the United States Sentencing Commission that failure to report offenses are associated with violence in only 3.1% of cases. Chambers, 129 S.Ct. at 692-93. After a short discussion with "little more than a statistical analysis of a research report prepared by the United States Sentencing Commission," the Court concluded that failure to report falls outside the scope of the ACCA definition of "violent felony" under the residual clause. Id. at 695 (Alito, J., concurring).

Finally, last month, in Sykes, the Supreme Court combined all of these tests to determine that the offense of intentional flight from a enforcement officer by vehicle is a violent felony under the residual clause. 131 S.Ct. 2267. Rather than clarifying a definitive test, the Court referred to the approaches above as a series of considerations to be factored into the analysis. See id. at 2273 ("For instance, a crime involves the requisite risk when `the risk posed by [the crime in question] is comparable to that posed by its closest analog among the enumerated offenses.'"); id. ("Another consideration is a comparison to the crime of burglary."); id. at 2274 ("Although statistics are not dispositive, here they confirm. . . ."). See also id. at 2285-86 (Scalia, J., dissenting) (criticizing these approaches as applied to this case). Importantly, Sykes did not overrule any of the previous tests, and James, Begay, and Chambers remain good law.

In effect, then, district courts are left with all of these tests. We can glean from the above that to qualify as a "violent felony" under the residual clause of the ACCA, the crime at issue must: (i) in the ordinary case, (ii) pose as great a risk of physical injury as its closest analog among the enumerated offenses, (iii) as shown by statistical evidence where available, and (iv) be roughly similar in kind (as in active, purposeful, violent and aggressive) as the enumerated offenses. James, 550 U.S. at 208, 127 S.Ct. 1586; id. at 203, 127 S.Ct. 1586; Chambers, 129 S.Ct. at 692; Begay, 553 U.S. at 143, 128 S.Ct. 1581. See also Sykes, 131 S.Ct. 2267. This is by definition an offense by offense analysis.

This year (albeit before Sykes), the First Circuit has applied these principles to conclude that harmful battery is a violent felony but that reckless battery is not, and reserved on the question of offensive battery. Holloway, 630 F.3d at 262 (Jan. 21, 2011). An ordinary conviction of "assault and battery," therefore, without additional information that qualifies under Shepard, does not amount to a predicate offense. Id.

Assault and battery of a police officer, by contrast, is a violent felony under the residual clause. The First Circuit compared the risk of injury of assault and battery of a police officer to the enumerated offenses and concluded that the likelihood of physical injury was greater. United States v. Dancy, 640 F.3d 455, 470 (1st Cir.2011). Law enforcement officers carry weapons, and an altercation with an officer poses a serious risk of confrontation that may lead to injury to the officer, the offender, as well as to bystanders. Id.

B. Discussion

With this guidance in mind, I will turn to the question at hand: whether larceny of a person and reckless assault and battery of a guard constitute "violent felonies" under the residual clause of the ACCA.

1. Larceny of a Person

The First Circuit has not addressed the question of whether larceny of a person is a predicate offense under the ACCA since 1993—well before the recent flurry of Supreme Court precedent. In United States v. De Jesus, the First Circuit relied on the following comment of the United States Sentencing Guidelines Manual: "[o]ther offenses are included where . . . the conduct set forth (i.e., expressly charged) in the count of which the defendant was convicted. . . by its nature, presented a serious potential risk of physical injury to another." § 4B1.2, comment n. 2 (1987); United States v. De Jesus, 984 F.2d 21, 24 (1st Cir.1993). The court found that larceny of the person poses a risk of physical injury and therefore falls within the residual clause of the ACCA. De Jesus, 984 F.2d at 25. In light of the most recent Supreme Court and First Circuit cases interpreting the residual clause, however, the First Circuit is considering the matter afresh. See United States v. Rodriguez, No. 10-1891 (1st Cir. June 14, 2011), (requesting parties and amicus curiae public defender to submit supplemental briefs addressing Sykes).

In Begay, the Supreme Court explicitly rejected the notion—as articulated in the old comment of the Guidelines Manual— that any crime that poses a serious risk of physical injury is a predicate offense under the ACCA:

Begay, 553 U.S. at 142, 128 S.Ct. 1581 (emphasis in original).

It is not enough, therefore, that larceny of a person poses a serious risk of physical injury. In the ordinary case it must pose as great a risk of physical injury as burglary (its closest analog among the enumerated offenses) and be roughly similar in kind (as in active, purposeful, violent and aggressive) as the enumerated offenses. James, 550 U.S. at 203, 127 S.Ct. 1586; Begay, 553 U.S. at 143, 128 S.Ct. 1581. Because De Jesus did not address these questions, I must consider the issue anew.

In this analysis, I defer to the Massachusetts courts' interpretation of the larceny statute. Holloway, 630 F.3d at 259 ("[I]n considering whether an offense should be considered a violent felony under the ACCA, federal courts must utilize state court constructions of state law.").

a. Larceny of a Person under Massachusetts Law

The elements of larceny of a person in Massachusetts are: (1) that the defendant took and carried away property; (2) that belonged to another; (3) that the defendant took the property from the person or someone who owned or possessed it or from such a person's area of control in his or her presence; (4) with the intent to deprive that person of the property permanently. Mass. Dist. Ct. Criminal Model Jury Instructions No. 8.560 (1988).

Unlike larceny of a person in other states, in Massachusetts, the property need not be physically connected to the victim; it is enough that the property be around the victim. Commonwealth v. Subilosky, 352 Mass. 153, 166, 224 N.E.2d 197 (1967). Cf. United States v. Thrower, 584 F.3d 70, 73-74 (2d Cir.2009) (examining New York larceny from the person that required physical contact between the victim and the property and concluding that "surely larceny that requires a physical nexus between the victim and the property, as opposed to larceny generally, creates a risk of violent confrontation.").

Larceny of a person in Massachusetts is primarily a stealth crime. In Commonwealth v. Jones, the Supreme Judicial Court ("SJC") of Massachusetts held that if there is enough force—however slight— to make the victim aware of the taking, then the appropriate crime is not larceny of a person but rather robbery. 362 Mass. 83, 89, 283 N.E.2d 840 (1972). In Jones, the defendant was convicted of robbery for snatching the purses of two women as they came home from a shopping trip. Neither victim was touched, threatened, or placed in fear. The SJC affirmed the convictions, holding that purse-snatching constitutes robbery, rather than larceny of the person, where "the actual force used is sufficient to produce awareness, although the action may be so swift as to leave the victim momentarily in a dazed condition, the requisite degree of force is present to make the crime robbery." Id. In other words, where the victim is made aware of the taking, there is enough "force" to constitute robbery.

To be sure, in Commonwealth v. Davis, the Massachusetts Appeals Court declined to extend the Jones rule to pickpocketing. 7 Mass.App.Ct. 9, 385 N.E.2d 278 (1979). In Davis, the victims were approached by two groups of youths, who formed a circle around them while they were walking in a park. Id. at 9-10, 385 N.E.2d 278. The defendant leaned in and pulled the victim's wallet out of his back pocket. The victim felt the tug, turned around, and said that the defendant had taken his wallet. The defendant turned around and gave it back. The victim testified that he felt "kind of afraid." Id. at 10, 385 N.E.2d 278. On appeal, the defendant challenged his conviction for unarmed robbery on the grounds that this taking was merely larceny of a person. The government argued in turn that according to Jones, because the defendant felt a "tug" and was thus aware of the taking, there was sufficient force to constitute unarmed robbery. The Davis court disagreed and, declining to extend the Jones rule to pickpocketing, held that a "tug" that makes a victim aware of the pickpocketing is not sufficient force to constitute unarmed robbery. Id. at 11, 385 N.E.2d 278.9 Nevertheless, the Davis court upheld the conviction for unarmed robbery because the defendant had placed the victim in "fear." Id. at 12, 385 N.E.2d 278.

In any event, the Davis court recognized that pickpocketing is generally a stealth crime. Id. at 11, 385 N.E.2d 278 ("Pickpocketing characteristically involves stealth and a lack of awareness of the taking by the victim."). And where there is fear—even so much as such that the victim is "kind of afraid"—the crime is not larceny of a person.

According to Jones and Davis, then, larceny of a person in Massachusetts is a stealth crime, where the victim is generally unaware of the taking and not placed in any fear.

b. Larceny of a Person in its Ordinary Form Does Not Pose a Serious Risk of Injury

With Massachusetts' interpretation of larceny of a person in mind, I turn to the first step of the ACCA residual clause analysis: whether the ordinary conduct involved in larceny of a person poses a serious risk of physical injury. James, 550 U.S. at 208, 127 S.Ct. 1586. Under the case law, the typical offense of larceny of a person is generally pickpocketing or purse-snatching.

I find it helpful to consider the crime in context. One could imagine that the typical offense might occur on public transportation. Consider Sue, sitting on a subway at rush hour with her purse by her feet; when she gets off, she discovers that her purse is missing. Sue's offender has committed the stealth crime of larceny of a person. (If Sue had felt a tug on her bag when it was taken, her offender would have committed the crime of robbery under Massachusetts law. Jones, 362 Mass. at 89, 283 N.E.2d 840.)

Since, by definition, the typical larceny does not include even the slightest degree of force or fear, we can conclude that the risk of confrontation is rather slight. Indeed, the Massachusetts version of larceny of a person is uniquely non-confrontational. Here, the victim of larceny of a person is unaware of the taking. It is hard to imagine, then, that the crime as it is ordinarily committed poses any risk of injury at all, much less a serious risk of injury to qualify as a "violent felony" under the residual clause of the ACCA. If the victim is unafraid and unaware of the offender's presence—or even mildly aware of a pickpocketer's "tug"—there will not be a confrontation between them in the heat of the moment.

c. Larceny of a Person Does Not Pose as Great a Risk of Physical Injury as Burglary

The offense of larceny of a person therefore does not pose a serious risk of physical injury as to fall within the residual provision of the ACCA. For completeness sake, however, I will continue the analysis and assume for the sake of argument that larceny of a person entails some risk of physical injury. I must then compare that risk of physical injury to its closest analog among the enumerated offenses: burglary.

Burglary entails a high risk of violence because there is a possibility of a "face-to-face confrontation between the burglar and a third party—whether an occupant, a police officer, or a bystander—who comes to investigate." James, 550 U.S. at 203, 127 S.Ct. 1586. Of course many crimes, including simple assault and battery, entail a risk of confrontation. But the risk of burglary is heightened because it occurs within a dwelling. A victim of a burglary may be—or may feel—trapped in his home with no means of escape. Indeed, in Massachusetts, a person has the right to use deadly force to respond to an assault threatening bodily harm by an unlawful intruder—even where the homeowner had a reasonable means of retreat or escape. See Commonwealth v. Peloquin, 437 Mass. 204, 207-08, 770 N.E.2d 440 (2002). The risk of a confrontation between a burglar and a victim may even involve weapons used either by the victim, the burglar, or of course an investigating third party.

In contrast to burglary, it is highly unlikely that the victim of a larceny of a person would accost his offender rather than safely retreat.10 Likewise, it is unlikely that a pickpocketer would assault his victim. Unlike a burglar who could hurt a homeowner behind closed doors, the pickpocketer is unlikely to resort to violence in a populated area. And in any event, by definition the pickpocketer is at least attempting to operate in stealth.

d. Larceny of a Person Is Not a Purposeful, Aggressive, Violent Offense as the Enumerated Offenses

I could end the analysis here, but again, to err on the side of completeness, I will consider the final prong: whether larceny of a person—as it is understood under Massachusetts law—is the type of purposeful, aggressive, and violent behavior as the enumerated offenses, burglary, arson, extortion, or crimes involving explosives. Obviously, it is not. As a stealth crime in Massachusetts, it is not aggressive or violent. A pickpocketer is a thief, to be sure, but he is not a robber. Nor is he committing the kind of aggressive violent act as an arsonist. Indeed, since he is silent, he does not threaten as an extortionist. There is no grand act of setting off explosives that risk injury to others.

Simply put, the stealth crime of larceny of a person does not pose a serious risk of physical injury, much less a risk of injury that is as great as the offense of burglary, and it is not the same type of aggressive and violent crime as the enumerated offenses. Larceny of a person, under Massachusetts law, fails at every prong of the analysis and does not constitute a "violent felony" under the ACCA.

2. Assault and Battery on a Guard

Without larceny of a person, Oliveira would have only two violent felony predicates, even assuming that assault and battery of a guard constitutes a violent felony. I therefore need not reach that question here.11

IV. GUIDELINES CONSIDERATIONS

Having determined that Oliveira is not an armed career criminal, I now turn to the Guidelines calculations. Here, there are two issues that affect Oliveira's criminal offense level: (1) enhancements for relevant conduct; and (2) imperfect entrapment—whether Oliveira was unduly encouraged in his illegal activity by the informant.

A. Relevant Conduct

The government argues that Oliveira is subject to a total of six points in enhancements for his conduct in connection with the one count of "felon-in-possession of a firearm" that was actually charged. Oliveira in turn argues that it is unsound for the Court to consider evidence of related conduct at sentencing.

I am required by statute to consider relevant conduct. 18 U.S.C. § 3553(a)(4)(A) (providing that the Court "shall" consider the category of offense committed as calculated by the Guidelines, which includes relevant conduct). Indeed, "[n]o limitation shall be placed on the information concerning the background, character, and conduct of a person convicted of an offense which a court of the United States may receive and consider for the purpose of imposing an appropriate sentence." 18 U.S.C. § 3661. Relevant conduct is incorporated into the Guidelines calculations, with requisite enhancements for various related offenses. Where relevant conduct is in dispute, the parties have an opportunity to present information to the Court, which the Court may consider so long as there is an "indicia of reliability to support its probable accuracy." U.S.S.G. § 6A1.3(a). Ultimately, of course, the Guidelines calculation is advisory, and the Court may weigh other factors at sentencing. Booker, 543 U.S. 220, 125 S.Ct. 738.

It is true that this process has considerable drawbacks and has provided an effective backdoor for the prosecution. As in this case, the prosecution may decide to bring only one count against a defendant and then seek to introduce all of the other evidence of other offenses—for which perhaps they did not have as much evidence or worried about the credibility of a witness—as part of sentencing.

Nevertheless, I am satisfied that Oliveira did in fact commit the related offenses. Over the course of three sentencing hearings, the government produced surveillance to prove that Oliveira: sold a shotgun (that had been reported stolen) and three bags of cocaine to Heroux on January 26, 2006; sold a 30-30 rifle and two $50 rocks of cocaine to Heroux on February 25, 2006; and sold a loaded handgun to Heroux on April 12, 2007.

But while I agree that the government met its burden with respect to this related conduct, I am troubled by the Guidelines' severe treatment of that conduct. This related conduct results in a two-level enhancement for conduct that involved three to seven firearms under § 2K2.1(b)(1)(A)12 and a four-level enhancement for using or possessing a firearm "in connection with another felony"—in this case the distribution of cocaine—under § 2K2.1(b)(6).13

B. Imperfect Entrapment

Finally, I turn to the defendant's argument that he is entitled to a downward departure under U.S. S.G. § 5K2.12 for imperfect entrapment. Oliveira concedes that he was not entrapped in the legal sense. He argues, however, that Heroux's unique involvement with his family—at a particularly sensitive time given his girlfriend's illness—and his exhortations to buy more and more guns (with government money, to be sure) encouraged him to become involved in an area of illegal activity with which he was unfamiliar and at a level which he would not have participated but for Heroux.

I may depart downward from the Sentencing Guideline under § 5K2.12 where there is "aggressive encouragement of wrongdoing" by an informant. Garza-Juarez, 992 F.2d at 912. Given the evidence that I heard at sentencing, I agree that Oliveira is entitled to a downward departure in this regard. While Heroux had a serious criminal history involving firearms, Oliveira did not. Up until this point, Oliveira had been a thief rather than a firearm dealer. Prior to the first recorded transaction, the government had no information that Oliveira dealt guns—apart from Heroux's tip, which I do not credit. There is nothing to suggest that Oliveira would have been involved in this conduct had it not been for Heroux's encouragement and solicitation. The video and audiotapes show Heroux as the one suggesting other purchases, determining what kind of guns he was interested in, manipulating the fee to create a greater incentive to sell, indicating that he was willing to pay more and more for the guns.

In effect, § 5K2.12 reflects the sense that a defendant who is encouraged by the government to commit other crimes is less blameworthy than a defendant not so encouraged, and indeed, less likely to commit other crimes. As such it is not merely a Guidelines departure ground; it is keyed to the statutory purposes—just deserts (because he is less blameworthy), and the likelihood of recidivism (because he is not likely to reoffend with guns). 18 U.S.C. § 3553(a).

V. SECTION 3553(a) FACTORS

The Guideline computation here would be 168 to 210 months. Since the statutory maximum is ten years, 120 months defines the outer limit of Oliveira's sentence. I will vary from that sentence, taking into account imperfect entrapment and Oliveira's role. He was essentially a broker, bringing together a willing buyer and a willing seller. He did not purchase the guns for resale. There was no gun stash at his house. The guns he had on hand were the ones that Heroux wanted. In effect, he merely took a commission for matching a buyer to a seller.

In addition, as I indicated to the parties, I requested statistics from probation concerning the average sentences for felons in possession of firearms, similarly-situated with respect to offense conduct and criminal record. Probation reported that the range of sentences for a person with Oliveira's criminal history category of VI for this offense is 84-110 months.

One hundred months, the sentence I selected, is the high end of the range. It reflects Oliveira's substantial record and the very serious nature of the crime. "Imperfect entrapment" or not, Oliveira succumbed to Heroux's inducement to sell guns and ammunition. Since Oliveira has never before been sentenced to prison for more than two and a half years, one hundred months is also a sentence that is "sufficient, but not greater than necessary" to achieve the purposes of sentencing. 18 U.S.C. § 3553(a). It is a long sentence—far longer than any he has ever experienced, but it is proportional to the offense, will deter other would-be dealers, and promotes public safety.

SO ORDERED.