Passive Embezzlement Schemes as Continuing Offenses

Introduction

David Brunell was indicted on October 26, 2017 on one count of embezzlement of government funds under the federal embezzlement statute, 18 USC § 641.1 Brunell’s father, prior to his death in 1993, received monthly retirement benefits from the Social Security Administration (SSA).2 After his father passed away, the SSA continued to automatically deposit retirement benefits into the joint bank account Brunell held with his father.3 The indictment alleged that from January 2004 to May 2017, Brunell allowed these benefits to accrue by concealing his father’s death from the SSA and withdrawing the funds for personal expenses.4 In response to these charges, Brunell filed a motion to dismiss the portion of the indictment alleging conduct before October 26, 2012, arguing it was barred by the five-year statute of limitations.5

The government contended that because Brunell engaged in a passive embezzlement scheme, his offense was a “continuing offense” for statute of limitations purposes.6 Section 641 criminalizes two types of embezzlement schemes—active and passive. Active embezzlement schemes require an affirmative act by the embezzler during the course of the scheme, such as making a false statement to the government in order to continue fraudulently receiving disability benefits.7 Passive embezzlement schemes, however, continue automatically once set in motion and require no action by the embezzler for the scheme to continue.8

For most offenses, the statute of limitations begins to run when the elements of an offense are satisfied.9 These are known as “discrete” offenses.10 For continuing offenses, however, the statute of limitations begins to run when the crime affirmatively ends, such as when law enforcement apprehends an escapee11 or uncovers the concealment of a fact affecting an individual’s right to government benefits.12 Continuing offenses involve a harm to society that necessarily occurs until the crime ends, as opposed to a harm stemming from repeated, but discrete, criminal acts.13 Accordingly, the question presented to the district court was whether Brunell’s passive embezzlement scheme was a continuing offense.14

Most courts agree that active embezzlement schemes are not continuing offenses. But courts of appeals are divided as to whether passive embezzlement schemes are continuing offenses.15 This disagreement stems from diverging interpretations of the Supreme Court’s test for when an offense should be deemed “continuing.” To decide whether an offense is continuing, and therefore when the statute of limitations begins to run, courts look to Toussie v United States.16 Under Toussie, an offense should be deemed continuing when either “the explicit language of the substantive criminal statute compels such a conclusion,” or “the nature of the crime involved is such that Congress must assuredly have intended that it be treated as a continuing one.”17

Courts of appeals are divided on how to apply Toussie’s second prong, which requires an analysis of the “nature of the crime involved,” to embezzlement offenses charged under § 641. Some courts interpret Toussie to require an analysis of the conduct charged in the indictment, evaluating whether an offense is continuing on a case-by-case basis.18 Courts using this type of analysis, known as the “charged conduct approach,” distinguish between active and passive schemes of embezzlement, finding only the latter continuing offenses.19 By contrast, other courts interpret Toussie to require a “categorical approach,” rejecting the distinction between active and passive embezzlement schemes. These courts evaluate the elements of § 641 to determine, on a statute-by-statute basis, whether all embezzlement schemes are continuing offenses.20 This Comment focuses on two questions raised by this split in authority: First, is the charged conduct approach correct, and second, if it is, are passive embezzlement schemes continuing offenses?

The resolution of these questions has important consequences for defendants like Brunell who raise statute of limitations defenses. Most obviously, it affects the length of the conduct for which a defendant can be indicted. In Brunell’s case, this meant the difference between facing charges for an embezzlement scheme lasting 161 months and one lasting only 55.21 It also affects sentencing and restitution. Applying the continuing offense doctrine means defendants could face higher sentences under the United States Sentencing Guidelines.22 Similarly, the shorter the defendant’s embezzlement scheme, the less the recovery will be for restitution purposes. For example, in the Second Circuit, “the statute of limitations applies to the calculation of the amount of loss for purposes of restitution,” so that the amount the defendant received from a time-barred embezzlement scheme cannot be included in the calculation of the recovery.23

More broadly, the point at which the statute of limitations begins to run for embezzlement offenses affects the delicate balance between protecting defendants from prosecutions for events in the far-distant past and ensuring that criminals are brought to justice. It is entirely possible that if only a minimal portion of the embezzlement scheme occurred within the statute of limitations and the judge finds that the embezzlement scheme is not a continuing offense, the defendant will walk free because the amount charged is less than the statutory minimum under § 641.24 Embezzlement schemes involving the fraudulent receipt of Social Security benefits like Brunell’s are both pervasive and, due to their difficulty to detect, costly to the government.25 In 2015, for example, the Office of the Inspector General for the Social Security Administration closed 529 cases of individuals fraudulently receiving their deceased relatives’ benefits, just like Brunell did.26

Despite these high stakes, the continuing offense doctrine in the statute of limitations context has gone “virtually unexplored by legal scholarship.”27 This Comment offers a novel justification for the charged conduct approach to embezzlement offenses. Part I covers the legal background on the federal statute of limitations and defines the continuing offense doctrine. Part II describes the current circuit split between the categorical and charged conduct approaches, observing that an adoption of the charged conduct approach would not upset settled precedent. Part III proposes a novel justification for classifying passive embezzlement schemes charged under § 641 as continuing offenses. This Comment argues that Toussie is an inadequate guide for choosing between the categorical and charged conduct approaches. Instead, courts should look to the Supreme Court’s Armed Career Criminal Act and Immigration and Nationality Act jurisprudence, which dictate that courts should adopt the charged conduct approach for statutes criminalizing diverse classes of offenses, like § 641. Applying this approach allows courts to properly classify passive embezzlement schemes as continuing offenses, which accords with the goals of the statute of limitations.

I. The Continuing Offense Doctrine

Criminal statutes of limitations bar the government from prosecuting certain offenses after a set period of time. The current default federal criminal statute of limitations provides that “[e]xcept as otherwise expressly provided by law, no person shall be prosecuted, tried, or punished for any offense, not capital, unless the indictment is found or the information is instituted within five years next after such offense shall have been committed.”28 As both statutory and doctrinal exceptions illustrate, the statute of limitations is not applied uniformly to all crimes. In order to understand the differing time periods for which the statute of limitations runs, Part I.A discusses the underlying purposes of the statute of limitations. Part I.B unpacks one of the ways that the statute of limitations is extended—the continuing offense doctrine—and discusses the Supreme Court’s test for determining whether an offense should be deemed “continuing.”

A. The Statute of Limitations

The federal criminal statute of limitations serves two primary purposes: fairness to defendants and efficient investigation and prosecution of crimes by law enforcement officials.29 In service of fairness, the statute of limitations offers a legislative guarantee that after a certain amount of time has passed, defendants will not have to stand trial and “defend themselves against charges when the basic facts may have become obscured by the passage of time.”30 The “passage of time” can “impair memories, cause evidence to be lost, deprive the defendant of witnesses, and otherwise interfere with [the defendant’s] ability to defend himself.”31 In an adversarial justice system, the result of the trial process is only as good as the evidence presented. In turn, the statute of limitations represents a legislative judgement that after a certain point in time, there is an “irrebuttable presumption” that a defendant will not receive a fair trial.32 After this point, “no quantum of evidence is sufficient to convict,” even if the defendant actually committed the offense.33

The time pressure that the statute of limitations places on law enforcement promotes efficiency. A hard cutoff on when prosecutors can bring charges “encourag[es] law enforcement officials promptly to investigate suspected criminal activity.”34 Moreover, it puts a premium on law enforcement’s time, forcing them to allocate their enforcement resources on the most salient and harmful offenses instead of low-level offenses in the distant past.35

But fairness and efficiency do not come without costs. Most obviously, “[e]very statute of limitations, of course, may permit a rogue to escape.”36 As a result, victims may not be made whole through post-conviction restitution,37 and society suffers every time a guilty person goes free.38 For some crimes, Congress has decided these costs outweigh the benefits of a statute of limitations. This is why “for almost as long as such rules have existed, Congress has created exceptions to them.”39

Originally enacted in 1790, the default federal criminal statute of limitations has shifted from two, to three, to now five years.40 Over the past few decades, Congress has also passed an increasing number of crime-specific exemptions, often in response to crises or changing circumstances.41 Statutes that extend the statute of limitations are often motivated by the seriousness of the crime. For this reason, murder has never had a statute of limitations.42 Elsewhere, Congress recognizes that some crimes are harder to catch than others. For example, in response to the 1980s Savings and Loan scandal and “the extraordinary complexity of procurement fraud cases,”43 Congress lengthened the statute of limitations on procurement fraud to seven years.44 In addition to statutory modifications, both courts and Congress can extend the statute of limitations by modifying the definition of the crime, thereby changing when the five-year period begins to run. This is known as the continuing offense doctrine.

B. Classifying Continuing Offenses

For most offenses, the statute of limitations begins to run when the elements of the offense are satisfied.45 These are known as “discrete” offenses.46 Take Illinois’s criminal damage statute, which prohibits, among other things, “knowingly shoot[ing] a firearm at any portion of a railroad train.”47 The moment an individual, with knowledge that he or she is shooting at a train, pulls the trigger of a firearm, the elements of Illinois’s criminal damage statute are satisfied. At that moment, the statute of limitations begins to run.

Continuing offenses involve a different sort of crime. Continuing offenses persist until they are affirmatively ended—for example, when the criminal is apprehended by law enforcement.48 Additionally, they involve a harm that necessarily persists so long as the crime does.49 Classic examples of continuing offenses include conspiracy, escape, kidnapping, and crimes of possession.50 For these crimes, the statute of limitations will not run until the conspiracy is foiled, the escapee is captured, or the victim is freed.51 In other words, for continuing offenses, the statute of limitations does not run until the crime is finished,52 even if the technical elements of the offense are satisfied before that moment. To be sure, the term “‘[c]ontinuing offense’ is a term of art that does not depend on [its] everyday notion[s] or ordinary meaning.”53 As a result, a continuing offense is not just a series of discrete acts tied together by a single overarching plan.54 So if an individual had an overarching scheme to shoot at multiple railroad cars, the fact that the individual did so over a longer period of time does not render the criminal damage charge a continuing offense. Rather, “[t]he hallmark of a continuing offense is that it perdures beyond the initial illegal act, and that ‘each day brings a renewed threat of the evil Congress sought to prevent.’”55 On the other hand, if there were a conspiracy among a group of individuals to shoot at railroad cars, the statute of limitations would begin to run not at the moment the agreement is formed (thereby satisfying the elements of conspiracy), but when the conspiracy is affirmatively ended.

An offense can be deemed continuing either legislatively or judicially. While Congress from time to time includes statutory language deeming an offense continuing,56 courts are often the arbiters of whether an offense should be treated as continuing.57 In 1970, the Supreme Court laid out the test to determine whether an offense is continuing for statute of limitations purposes in Toussie v United States.

Defendant Robert Toussie was charged with violating the Universal Military Training and Service Act58 for failing to register for the draft within five days of his eighteenth birthday.59 In 1959, Toussie did not register during the required window between June 23 and 28. As a result, he was indicted on May 3, 1967.60 Toussie moved to dismiss the indictment, arguing that prosecution was barred by the five-year statute of limitations. Because failure to register is a discrete offense, Toussie argued, the elements of the offense were satisfied when the five-day window closed in June 1959 and the statute of limitations ran out in June 1964.61 The government responded that failure to register for the draft was a continuing offense and that Toussie could be prosecuted at any time while he was unregistered.62 That is, as long as Toussie remained unregistered, the statute of limitations would not run. Accordingly, the question presented to the Court was whether failure to register for the draft was a continuing offense.

The Court laid out a two-pronged test for determining when courts should deem an offense “continuing.” An offense is continuing when (1) “the explicit language of the substantive criminal statute compels such a conclusion” or (2) “the nature of the crime involved is such that Congress must assuredly have intended that it be treated as a continuing one.”63 Applying this test, the Court made quick work of the first prong, concluding that no explicit language in the draft registration statute contemplated a prolonged course of conduct.64 Applying the second prong, the Court looked to whether there was anything “inherent in the act of registration itself which makes failure to do so a continuing crime.”65 Finding that there was not, the Court contrasted failure to register, which involved a discrete five-day window to register, with conspiracy, a well-known continuing offense.66 While “the nature of a conspiracy” is such that “each day’s acts bring a renewed threat of the substantive evil Congress sought to prevent,” the “first draft registrations clearly were viewed as instantaneous events and not a continuing process.”67 In support, the Court surveyed draft laws dating back to 1917, concluding that Congress had always intended to treat failure to register as a discrete offense.68 In other words, legislative history suggested that avoiding registration was an offense that happened at one point in time, rather than an ongoing course of conduct. As a result, the Court concluded that failure to register for the draft was not a continuing offense and reversed Toussie’s conviction.69

Since Toussie, the Court has discussed the continuing offense doctrine only once. In United States v Bailey,70 a prison escapee sought to raise a duress defense.71 The prisoner’s ability to raise this defense depended, in part, on whether escape was a continuing offense.72 With little analysis, the Court concluded that “escape from federal custody as defined in [18 USC] § 751(a) is a continuing offense.”73 Due to the “continuing threat to society posed by an escaped prisoner,” the Court noted, “the nature of the crime involved is such that Congress must assuredly have intended that it be treated as a continuing one.”74

Today, lower courts struggle to apply Toussie consistently. They seldom struggle with the first prong—few statutes contain explicit language indicating that an offense should be deemed continuing. When such language is included, interpretation is relatively straightforward.75 Instead, courts struggle to apply the second prong: whether the “nature of the crime involved is such that Congress must assuredly have intended it to be treated as a continuing one.”76 Difficulty arises because of the vagueness of the phrase “nature of the crime.” As one practitioner remarks, “Reference to the ‘nature’ of a crime is so vague as almost to defy analysis.”77

II. The Circuit Split

Circuit courts are split between two interpretations of Toussie’s second prong when determining whether embezzlement schemes charged under the federal embezzlement statute, 18 USC § 641, are continuing offenses. Some courts interpret Toussie’s second prong to require an analysis of the charged conduct in the indictment in each case.78 These courts inquire into the type of embezzlement scheme perpetrated. This is called the charged conduct approach. By contrast, other courts interpret Toussie’s second prong to require only an analysis of the elements of the statute, not the defendant’s course of conduct.79 These courts analyze whether § 641 in its entirety is a continuing offense without regard to the defendant’s conduct in the case. This is called the categorical approach.

This Part analyzes the circuit split. Part II.A examines the charged conduct approach, while Part II.B examines the categorical approach. Part II.C clarifies the nature of the circuit split, concluding that the charged conduct approach does not open the floodgates to overbroad prosecutorial discretion.

A. The Charged Conduct Approach

The Fourth Circuit80 and, in an unpublished opinion, the Ninth Circuit81 interpret Toussie’s instruction to evaluate “the nature of the crime involved” to require an analysis of the defendant’s charged conduct in each case to determine whether the defendant’s embezzlement scheme is a continuing offense. In doing so, these courts distinguish between two types of embezzlement schemes: active and passive. As described below, these courts classify only passive embezzlement schemes as continuing offenses.82

The Fourth Circuit’s decision in United States v Smith83 was the first to lay out the charged conduct approach. In that case, defendant Alfred Smith engaged in an embezzlement scheme commonly charged under § 641, which provides that “[w]hoever embezzles, steals, purloins, or knowingly converts to his use . . . any record, voucher, money, or thing of value of the United States or of any department or agency thereof . . . [is guilty of a crime].”84 Instead of notifying the SSA after his mother passed away, Smith collected his mother’s Social Security benefits through the joint bank account they shared.85 During a four-year period, Smith embezzled approximately $26,336 in benefits.86 Smith was subsequently indicted under § 641 on one count of embezzlement.87 He moved to dismiss the indictment, noting that almost all of the benefits payments deposited into the account occurred outside of the statute of limitations period.88 If Smith’s motion were granted, all but one or two checks that Smith received would have been outside the scope of prosecution.89

The Fourth Circuit, adopting the charged conduct approach, rejected this argument. Applying Toussie’s second prong, the court reasoned that “the nature of embezzlement is such that Congress must have intended that, in some circumstances, it be treated in § 641 as a continuing offense.”90 Notably, the court did not detail its reasoning for interpreting Toussie to require the charged conduct approach.91 Instead, the court focused on parsing out the types of embezzlement schemes that are continuing offenses. The court concluded that in “cases where the defendant created a recurring, automatic scheme of embezzlement” and “maintained that scheme without need for affirmative acts,” the conduct is a continuing offense.92 Although the court did not use the term, these characteristics established the definition of passive embezzlement schemes.93 The court was careful, however, to underscore the boundaries of the charged conduct approach, noting that “[not] all conduct constituting embezzlement may necessarily be treated as a continuing offense as opposed to merely ‘a series of acts that occur over a period of time.’”94

In an unpublished opinion, the Ninth Circuit also concluded that an embezzlement scheme involving Social Security benefits was a continuing offense. But in reaching this conclusion in United States v Neusom,95 the Ninth Circuit relied upon just one in-circuit case, offering no analysis regarding the choice between the categorical and charged conduct approaches.96 As a result, Neusom should not be read as a clear endorsement of the charged conduct approach by the Ninth Circuit. To the contrary, in at least one other case, the Ninth Circuit utilized a categorical approach when determining whether mail fraud was a continuing offense.97

Despite its lukewarm reception at the appellate level, the charged conduct approach has been adopted by district courts across the United States.98 In particular, district courts encountering embezzlement schemes similar to the one seen in Smith are more likely to adopt the charged conduct approach.99 But other cases involving theft of Social Security disability benefits without the use of a joint bank account also follow Smith’s charged conduct approach. In United States v Phan,100 the leading district court case applying the charged conduct approach, the court found that a case manager at a nonprofit tasked with helping immigrants obtain government benefits engaged in a continuing offense by diverting one client’s disability benefits to her personal account and letting those benefits automatically accrue.101 To justify its adoption of the charged conduct approach, the district court interpreted Toussie’s second prong to require an analysis of the “nature of the offense charged.”102 Applying this approach, the court highlighted two aspects of continuing offenses that render passive schemes of embezzlement continuing offenses. First, continuing offenses “take[ ] place not in a moment, but over a span of time,”103 continuing “until affirmatively ended.”104 Phan’s embezzlement scheme, the court held, is “analogous” because the SSA continued to rely on Phan’s initial non-disclosure and issue checks.105 Second, the court emphasized the fact that the “harm done to society” continues “for as long as the crime is ongoing.”106 The harm in Phan’s scheme fit the definition of a continuing offense because the “deleterious effect on society . . . continued at least until the payments stopped.”107

B. The Categorical Approach

The Second,108 Seventh,109 and Eighth110 Circuits embrace the categorical approach, narrowly interpreting “the nature of the crime involved” to require an abstract analysis of the embezzlement statute’s elements. Courts using the categorical approach are motivated by the Court’s instruction that statutes of limitations should be interpreted liberally in favor of repose.111

The Second Circuit was the first appellate court to adopt the categorical approach to evaluate whether embezzlement under § 641 is a continuing offense. In United States v Silkowski,112 defendant Ralph Silkowski pled guilty to theft of public funds under § 641 for fraudulently receiving auxiliary disability benefits for his wife and his daughter, despite the fact that they did not live with him.113 Before sentencing, Silkowski disputed the district court’s restitution order, which included losses outside the statute of limitations.114 To determine whether the losses were actually outside the statute of limitations, the court had to determine whether the defendant’s scheme of embezzlement was a continuing offense. The court concluded that “[i]n light of the government’s claim that it did not charge a continuing offense,” Silkowski’s one-count violation of § 641 is “not a continuing offense in this case.”115 Instead of detailing its analytical steps, however, the court just noted that under a contrary holding, “a serious issue would exist as to whether Silkowski entered a valid guilty plea.”116

Unlike the Second Circuit in Silkowski, the Seventh Circuit offered an in-depth analysis of the continuing offense doctrine in United States v Yashar.117 In that case, the defendant was a “ghost payroller,” meaning that he received compensation for his employment with the City of Chicago despite doing “little or no work.”118 Unlike the defendants in Smith, Silkowski, and Phan, the defendant was indicted for embezzlement under 18 USC § 666(a)(1)(A), which criminalizes embezzlement by a local government employee when the locality receives federal benefits.119 This statute, however, is analogous in all relevant aspects to § 641.120

In a motion to dismiss the indictment, Yashar argued that almost all of the alleged embezzlement scheme occurred more than five years before the waiver he signed, thereby tolling the statute of limitations.121 Additionally, because the government conceded that the benefits Yashar received within the five-year statute of limitations period did not meet the statutory minimum, Yashar argued that the indictment should be dismissed entirely (just as the district court held).122

The Seventh Circuit interpreted Toussie’s second prong to mean that a crime should be treated as continuing only when “the crime by its nature” is such that Congress intended it to be continuing, adopting a categorical approach focused on the “statutory language” of the offense.123 Rejecting the distinction between active and passive schemes that the Fourth Circuit would later make in Smith, the court noted that because conspiracy, a continuing offense, “may or may not involve affirmative actions,” such a distinction “has never been the benchmark of a continuing offense.”124 As a result, the court determined that embezzlement under § 666(a)(1)(A) was not a continuing offense.125 Additionally, the specter of unchecked prosecutorial discretion drove the court’s conclusion. The court noted that under a contrary holding, “the statute of limitations, designed as a control on government action, would instead be defined by it.”126

Most recently, the Eighth Circuit adopted the categorical approach in United States v Askia.127 In that case, the defendant was convicted under § 666(a)(1)(A) for embezzling federal grant money that he pledged to use to build a children’s center for after-school programs.128 While the defendant’s scheme occurred from August 2007 to April 2008, he was not indicted until March 6, 2013. The defendant subsequently moved to dismiss the indictment, arguing the five-year statute of limitations barred prosecution for his conduct before March 6, 2008.129

Applying the categorical approach, the Eighth Circuit relied on multiple factors to conclude that § 666(a)(1)(A) is not a continuing offense.130 First, the court analogized the elements of embezzlement to the elements of larceny, which is “well established” as a discrete offense.131 The court reasoned that because “embezzlement is merely a larceny from a position of trust,” then “[e]mbezzlement, like larceny, is completed with the unlawful taking.”132 Second, the court noted that unlike conspiracy, § 666(a)(1)(A) offenses do not continue over time after the elements are satisfied.133 Finally, the court found that an underlying purpose of the statute of limitations, to “encourage[ ] timely prosecutions when the facts are fresh,” would not be accomplished if § 666(a)(1)(A) was a continuing offense.134

C. Criticisms of the Charged Conduct Approach

Critics of the charged conduct approach argue it will effectively “swallow the second factor of Toussie,” which asks whether the nature of the crime involved is such that Congress intended it to be treated as a continuing one.135 An inquiry beyond the text of the statute and into the charged conduct, they argue, would no longer focus on Congress’s intent, “but on the conduct charged by the prosecutor and the language of the indictment.”136 This leaves the continuing offense doctrine “virtually unbounded,”137 effectively “transfer[ing] an excessive amount of power to the prosecutor” by allowing her to “ultimately decide when the limitations period would begin to run” based on the language she uses in the indictment.138

But adopting the charged conduct approach toward offenses charged under § 641 would not have these dramatic effects, as passive embezzlement schemes are a narrowly defined subset of all the possible schemes that could be charged under the federal embezzlement statute. As a result, prosecutors are structurally constrained by the definition of a passive embezzlement scheme. The extent of the disagreement between circuits underscores this point. Generally, courts using the categorical approach classify § 641 offenses as “discrete” in cases in which they are presented with active embezzlement schemes. Viewing these opinions narrowly on their facts means that courts using the categorical approach have not definitively foreclosed the possibility that passive embezzlement schemes are continuing offenses. Moreover, courts using the charged conduct approach also deem active embezzlement schemes discrete offenses. As such, the circuit split only concerns a small subset of embezzlement offenses—passive embezzlement schemes—and is rooted in the dicta of courts applying the categorical approach instead of settled precedent.

Most courts adopting the categorical approach are presented with active schemes of embezzlement.139 Recall that active embezzlement schemes are distinguished from passive schemes in that they require some affirmative act to keep the embezzlement going.140 In other words, active embezzlement schemes do not automatically continue. In Yashar, the defendant did a small amount of work for the City of Chicago in order to keep up appearances.141 In Askia, the defendant applied for a grant and then affirmatively withdrew that money from the community learning center’s account on multiple occasions, as opposed to effortlessly receiving checks from the government.142 Even the defendant in Silkowski “falsely certified that both his wife and daughter resided with him” in order to reinstate Social Security benefits and continue his embezzlement scheme.143 In each of these cases, the defendants’ affirmative acts—whether that be going to work to keep up appearances, transferring money between accounts, or filing false statements—rendered the embezzlement scheme active because the scheme did not continue automatically.

This trend holds true for district court cases as well. First, the trend occurs in cases that seemingly resemble Smith’s passive acceptance of benefit payments for his deceased father but, on closer examination, involve an affirmative action by the defendant. For example, in United States v Powell,144 the defendant failed to report the death of her grandmother and continued to receive her Social Security benefits.145 But Powell also took steps to make it appear that her grandmother was still living, such as establishing credit and utilities in her grandmother’s name.146 At this point, Powell’s embezzlement scheme was no longer perpetrated “automatically” or “without need for affirmative acts.”147 Nonetheless, the court relied on Yashar to categorically classify all embezzlement schemes as discrete offenses, not just active schemes.148

Second, this trend occurs in cases involving the collection of disability benefits, like Phan.149 In United States v Pease,150 the defendant falsely indicated to the SSA that he did not receive income because he was disabled.151 Unlike Phan, the benefits that the defendant received as a result of this misrepresentation were not automatic—he actively submitted false statements to the SSA in order to keep the stream of benefits flowing.152 Again, the court classified all embezzlement schemes as discrete offenses using the categorical approach.153

Admittedly, the line between active and passive embezzlement offenses is not always so clear. For example, even the defendant in Smith “wrote checks and withdrew funds” from his mother’s account, acts that could be considered active.154 Nonetheless, Smith still draws a principled line between passive and active schemes. A passive scheme is one that is “maintained . . . without need for affirmative acts” and is “automatic.”155 So if a scheme requires the defendant to keep up appearances, like in Yashar, or periodically transfer money from one account to another, like in Askia, it is active because those actions relate to the maintenance of the scheme. On the other hand, withdrawing money from a joint bank account is not an action that maintains the embezzlement mechanism (concealment), as it does not affect the stream of payments.

Of course, there are a few cases in which judges do not closely adhere to Smith’s definition of a passive embezzlement scheme. In United States v Ellis156 and United States v Thompkins,157 although both defendants failed to notify the SSA of their relatives’ deaths and fraudulently received benefits as a result, like in Smith, the defendants needed to forge their deceased relatives’ signatures in order to deposit the money.158 Because these schemes were not automatic, and an affirmative action (forgery) was required to maintain the illegal stream of benefits, under Smith these embezzlement schemes were active, not passive. Despite this, both courts found that the active schemes were continuing offenses.159 These cases, however, are outliers. Most courts applying Smith find that active embezzlement schemes are not continuing offenses.

Courts that correctly apply the charged conduct approach to active embezzlement schemes conclude that these schemes are discrete offenses. In United States v Tackett,160 the defendant engaged in an active embezzlement scheme that involved embezzling stimulus checks meant for residents of a retirement home.161 Despite endorsing the charged conduct approach,162 Judge Amul Thapar held that the active embezzlement scheme was not a continuing offense because “in order to embezzle, steal, and convert thirty-four different checks, [the defendant] would have needed to perform more than one affirmative act.”163 Similarly, in United States v Williams,164 the court distinguished the defendant’s active embezzlement scheme from an employee benefit plan with the scheme in Smith, highlighting the defendant’s affirmative acts: “[H]ad [the defendant] not taken that additional step each time, the violation would not have continued.”165

Although some courts that adopt the charged conduct approach are presented with active embezzlement schemes,166 the majority of charged conduct cases involve passive embezzlement schemes.167 In both Neusom and Smith, once the defendants began concealing information affecting their right to benefits, the defendants took no other affirmative action.168 District court cases follow this pattern as well. For example, in United States v Gibson,169 there were “no other affirmative acts by defendant to effect the embezzlement” once the initial omission occurred.170

In sum, the charged conduct approach has not opened the floodgates for prosecutors to dictate when the statute of limitations begins to run. Active schemes of embezzlement, which involve discrete events continued repeatedly over time, are not continuing offenses under both the charged conduct and categorical approaches. Passive schemes, on the other hand, simply have not been considered by courts applying the categorical approach. When read in light of their facts, the opinions in which courts adopt a categorical approach do not explicitly rule on passive embezzlement schemes. This means that the disagreement among circuits regarding whether passive embezzlement schemes are continuing offenses is not rooted in settled precedent. In other words, a narrow reading of cases in which courts use the categorical approach does not explicitly foreclose the possibility that passive embezzlement schemes are continuing offenses. Even courts that adopt the categorical approach recognize that they are only dealing with active, not passive, embezzlement schemes. In United States v Reese,171 although the district court adopted a categorical approach, it observed that “even if the court followed the reasoning [in] Smith,” it would not find a § 641 offense continuing because the defendant’s scheme was “not the type of . . . passive activity considered by Smith.”172 As a result, a court’s adoption of the charged conduct approach would not be a radical departure from precedent in a circuit that uses the categorical approach.

III. Applying the Charged Conduct Approach to Passive Embezzlement Schemes

Courts that categorically deem all schemes of embezzlement discrete offenses are too quick to do so. Instead, courts should carefully consider, on a case-by-case basis, whether the defendant’s embezzlement scheme is a continuing offense using the charged conduct approach.

To get to this conclusion, this Part proceeds in two sections, answering two questions: First, is the charged conduct approach correct? And second, if it is, are passive schemes of embezzlement continuing offenses? Part III.A surveys courts’ interpretations of Toussie, arguing that their reliance on the vague language of Toussie’s second prong is insufficient by itself to justify either the categorical or charged conduct approach. In response, this Comment observes that the decision between a categorical and charged conduct approach is one that the Supreme Court has made before in the Armed Career Criminal Act of 1988173 (ACCA) and the Immigration and Nationality Act174 (INA) contexts. In light of the Court’s decision to adopt an approach similar to the charged conduct approach for statutes that criminalize diverse classes of crimes, like § 641, this Comment argues courts should adopt the charged conduct approach. Part III.B concludes that under the charged conduct approach, passive, but not active, embezzlement schemes are properly classified as continuing offenses. Finally, Part III.C concludes that classifying passive embezzlement schemes as continuing offenses under the charged conduct approach fits within the aims of the statute of limitations.

A. Courts Should Adopt the Charged Conduct Approach

The Supreme Court provides little guidance to lower courts applying the second prong of Toussie. Read in context, the phrase “nature of the crime involved” requires neither the charged conduct nor the categorical approach. Part III.A.1 discusses this ambiguity, arguing that the simplicity of Toussie’s facts limits its applicability to the federal embezzlement statute. As a result, mere reliance on the text of Toussie fails to resolve the dispute. In response, Part III.A.2 argues that courts should adopt the charged conduct approach because the Supreme Court, when faced with a similar choice between categorical and charged conduct approaches in the ACCA and INA contexts, adopted a conduct-based approach similar to the charged conduct approach. Following the Court’s lead would result in a more accurate application of the continuing offense doctrine.

1. Toussie is an insufficient guide.

Before deciding whether passive embezzlement schemes should be treated as continuing offenses, courts need to decide which approach to take—the charged conduct or categorical. Currently, courts make this decision by interpreting Toussie’s second prong, which directs courts to look at the “nature of the crime involved.” Courts adopting the charged conduct approach interpret this phrase to require an inquiry into the “nature of the offense charged,”175 while courts adopting the categorical approach interpret it to prompt an inquiry into “whether the crime by its nature is such that Congress intended that it be a continuing one.”176 As discussed below, reliance on this language alone justifies neither the charged conduct nor categorical approach.

In Smith, the leading circuit court opinion adopting the charged conduct approach, the Fourth Circuit’s only analysis of Toussie is in passing, with the court concluding that Smith’s embezzlement scheme “is correctly considered, under Toussie, to be a continuing offense.”177 This shallow treatment is surprising; indeed, the dissent in Smith cites to cases adopting the categorical approach, like Yashar.178 Phan, the leading district court case adopting the charged conduct approach, does not fare much better. While Phan spells out its interpretation of Toussie’s second prong by paraphrasing the prong in a way that shows its interpretative step,179 the court primarily focuses on why passive embezzlement schemes are continuing offenses, omitting discussion of the choice that opens up that analysis in the first place—whether to use the charged conduct or categorical approach.

This lack of discussion of Toussie’s second prong may be driven by the plausibly straightforward nature of the Supreme Court’s language. Read in context, the test in Toussie provides that an offense should not be deemed continuing “unless the explicit language of the substantive criminal statute compels such a conclusion or the nature of the crime involved is such that Congress must assuredly have intended that it be treated as a continuing one.”180 Within this test lies a distinction between the text of the statute and the conduct charged: the Court refers to the “explicit language” of the statute in one prong and the “nature of the crime involved” in the other.181 The Court’s use of “text” in one prong and “crime involved” in the other can be read to imply an analysis of more than just the text of the statute’s elements. But there is no reason to think that “crime involved” necessarily means the crime as charged in the indictment as opposed to the crime as defined by the statute. As a result, simply interpreting this phrase from Toussie—especially when it can be read both ways—is inadequate to justify the charged conduct approach.

While courts adopting the categorical approach do a better job at justifying their interpretation of Toussie’s second prong, these justifications also fall short. First, these courts point to Toussie’s caution that “the doctrine of continuing offenses should be applied in only limited circumstances.”182 Because the charged conduct approach marginally extends the statute of limitations, they argue, courts should give the benefit of the doubt to defendants. But what does “limited circumstances” mean, and why should it apply here? To be a workable principle, “limited” must mean something beyond “less extensive.” Comparatively, far from leaving the continuing offense doctrine “virtually unbounded,”183 the charged conduct approach only extends the statute of limitations for passive schemes of embezzlement, a small subset of embezzlement offenses.184

Second, some courts describe Toussie itself as an example of the categorical approach. In Askia, the court reasoned that because Toussie examined the text and legislative history of the failure to register offense, Toussie’s second prong necessitates a categorical approach.185 Other courts and commentators point to the Court’s comment that “[t]here is [ ] nothing inherent in the act of registration itself which makes failure to do so a continuing crime”186 to argue Toussie requires the categorical approach.187 These conclusions, however, rest upon the assumption that the phrase “act of registration” refers to the elements of the statute, not Toussie’s conduct. But this is not necessarily the case. Toussie’s failure to register for the draft and the narrow statutory prohibition on failing to register are one and the same. There are no active or passive schemes to avoid registering; failure to register means failing to complete a form, which was exactly Toussie’s conduct. As a result, it is unlikely that the Court was looking to the text of the statute when the facts in the case were so simple.

It is this simplicity of Toussie, not its complexity, that produces inadequate justifications by circuits on both sides of the split. Justice Oliver Wendell Holmes Jr famously remarked that “hard cases[ ] make bad law.”188 But Toussie suffers from the opposite problem: because the statute in Toussie was simple, the Court’s guidance has limited applicability.189 In Toussie, the statute at issue was the Universal Military Training and Service Act. It requires a specific course of conduct (register for the draft) within a well-defined time frame (five days after a man’s eighteenth birthday).190 As a result of the simplicity of the underlying statute, the Court was never presented with the choice between the charged conduct and categorical approaches. So, commentators concluding that “Toussie never suggests that a particular offense may be continuing under some circumstances but not others, all depending on the conduct charged”191 are right but do not go far enough. Toussie never favored either approach because the issue was never before the Court. As a result, courts must look to other areas of law to find a workable principle that justifies their approach.

2. Broad statutes require flexible approaches.

In contexts beyond statutes of limitations, courts face the choice between the categorical and charged conduct approaches. In the context of sentencing enhancements and removal decisions, courts are frequently required to classify state statutes. For example, the ACCA imposes heightened penalties for offenders who have three prior convictions for a “violent felony.”192 In the immigration context, any alien, including a lawful permanent resident, may be deportable if convicted of an “aggravated felony.”193 As a result, courts are required to determine whether convictions under state statutes fall into either of these categories. These determinations raise a similar question: Should courts look only to the elements of the state statute or can they look to the defendant’s underlying conduct?

For most statutes, the Supreme Court employs a categorical approach. Under the ACCA, a “violent felony” includes any offense that is equivalent to the generic elements of “burglary, arson, or extortion.”194 In Descamps v United States,195 the Court considered whether a comparison of a state burglary statute to the generic elements of burglary should entail a categorical approach, or whether courts can consider “information about a case’s underlying facts,” such as the indictment.196 The Court concluded that the legislative history’s apparent focus on the statutory elements, as well as a desire for consistency and fairness, supported the categorical approach.197 Later, the Court applied this reasoning to the immigration context.198

But for statutes that prohibit alternative courses of conduct, called “divisible statutes,” the Court uses a modified categorical approach that allows courts to look to the charged conduct.199 Divisible statutes allow a defendant to be convicted of a crime without the government proving one or more elements.200 For example, in Taylor v United States,201 the Court evaluated whether a divisible state statute criminalizing burglary of a building, motor vehicle, or boat fit within the generic crime of burglary, which the ACCA considers a “violent felony.”202 Because most statutes are not divisible, courts usually know which elements of the offense the jury relied on to convict the defendant and can make a straightforward comparison using the categorical approach.203 For divisible statutes, however, courts do not know which elements the jury convicted upon without looking into the charged conduct. As a result, an application of the categorical approach to divisible statutes raises a Sixth Amendment issue, as all conduct that increases a defendant’s sentence must be put to the jury.204 In Taylor, the Court responded to these concerns by holding that when the statute is divisible, courts should use a “modified categorical approach,” looking to the charged conduct to determine the elements upon which the defendant was convicted.205 After holding that only burglary of a house was a “violent felony,” the Court remanded the case with instructions to apply this modified categorical approach to determine whether the defendant was convicted of burglarizing a building, motor vehicle, or a boat.206

Section 641 is analogous to divisible statutes. Broadly speaking, divisible statutes involve “diverse classes of criminal acts.”207 Similarly, § 641 criminalizes a diverse range of offenses, from classic theft208 to modern-day embezzlement schemes209 to theft of classified documents or government property.210 This is due to a conscious drafting choice: Congress intentionally drafted § 641 with broad language so it might cover situations not envisioned under the common law.211

While no court has weighed in on whether § 641 is a divisible statute, in the immigration context, the Second212 and Ninth213 Circuits deem a similar embezzlement statute divisible.214 18 USC § 656 provides that whoever, being a bank employee, “embezzles, abstracts, purloins or willfully misapplies any of the moneys, funds or credits of such bank” insured by the federal government is guilty of a crime. Similarly, § 641 provides that “[w]hoever embezzles, steals, purloins, or knowingly converts” property of the United States is guilty of a crime. In Carlos-Blaza v Holder,215 the Ninth Circuit recognized that § 656 prohibits at least two different types of crimes: “pure theft offenses,” which do not require proof of intent to defraud, and misapplication of bank funds, a type of embezzlement which requires proof of intent to defraud.216 Similarly, § 641 separately criminalizes theft and embezzlement, which can have different intent requirements.217 Because a statute of conviction is divisible when some crimes “involve an aggravated felony and some of which do not,”218 and only crimes involving deceit or fraud are aggravated felonies under the INA,219 the court found that § 656 is a divisible statute and utilized the modified categorical approach.220

For the same reasons that courts use the modified categorical approach to divisible statutes like § 656, courts should adopt the charged conduct approach for offenses charged under § 641. When faced with a statute that criminalizes diverse courses of conduct—some less serious than others—the Supreme Court held that courts should gather more information about the defendant’s conduct to determine what the jury intended to convict upon rather than make categorical rulings. Similarly, when faced with § 641, a statute that criminalizes offenses ranging from basic theft to complicated embezzlement, courts should gather information about the defendant’s conduct to more accurately predict what “Congress must assuredly have intended”221 regarding the statute of limitations. The categorical approach to divisible statutes carries the risk of applying sentencing enhancements or removal consequences in a manner inconsistent with Congress’s intent; the categorical approach to § 641 also carries the risk of applying the continuing offense doctrine to embezzlement offenses in a manner inconsistent with Congress’s intent.

To mitigate this risk, courts should strive for a more accurate approach that avoids a sweeping classification for two different types of conduct. Some cases involving discrete instances of theft will be easy to classify. But passive embezzlement schemes are different than active schemes: they are different both in kind (they require no action by the defendant) and in their harm to society (they are more likely to continue longer).222 As such, a more thorough approach is necessary. Just as Congress intended different consequences for different courses of conduct under the ACCA and INA, so too is it likely that Congress intended different statute of limitations consequences for passive and active embezzlement schemes. By carefully parsing out these differences, courts can more accurately apply the continuing offense doctrine.

B. Passive Schemes of Embezzlement Are Continuing Offenses

Perhaps the most important reason why courts should adopt the charged conduct approach is that passive embezzlement schemes have all the characteristics of continuing offenses, while active schemes do not. Accordingly, only under the charged conduct approach can courts accurately classify passive embezzlement schemes as continuing offenses. Judge Diarmuid O’Scannlain summarizes the two key aspects of continuing offenses: “(1) an ongoing course of conduct that causes (2) a harm that lasts as long as that course of conduct persists.”223 This Section analyzes passive embezzlement schemes in light of these two characteristics.

1. Passive schemes continue until affirmatively ended.

A central characteristic of continuing offenses is that “the commission of such crimes takes place not in a moment, but over a span of time.”224 This means that continuing offenses persist “until affirmatively ended.”225 For example, “a defendant who takes possession of contraband continues to possess the contraband day after day until the possession ceases.”226 Courts widely hold that classic continuing offenses such as conspiracy, escape, kidnapping, bigamy, and crimes of possession all share this characteristic.227 Even Yashar noted in response to the government’s argument that Toussie is “applicable only to offenses in which the defendant’s affirmative actions cease and the offense continues” that “[i]t is true that many continuing offenses, such as escape or bigamy, meet that definition.”228

Passive schemes of embezzlement, like classic continuing offenses, continue until affirmatively ended.229 In United States v Brunell,230 some intervening action was necessary in order to stop the stream of benefits paid to Brunell.231 For Brunell, the affirmative event that ended his scheme was an investigation by the SSA.232 Conversely, active schemes of embezzlement, which are not continuing offenses, do not simply continue until affirmatively ended. By their very nature, active schemes of embezzlement require affirmative conduct by the defendant. Every time that a defendant has to forge a signature,233 take steps to make it appear their relative is still living,234 or periodically submit false statements to the SSA,235 in order to “maintain”236 the embezzlement scheme, it is clear that the scheme would not have automatically continued until someone, be it the defendant or the authorities, affirmatively stepped in to end it. In each of these cases, the defendant needed to act in order to continue the scheme. Put differently, the defendant who pulls the trigger to a gun pointed at a train commits a discrete offense of criminal damage,237 like an active embezzlement scheme. But if that trigger were to cause the gun to indefinitely fire at a train until the police arrived and affirmatively ended the gun’s firing, that act would be a continuing offense, like a passive embezzlement scheme.

Yashar rejects the distinction between active and passive schemes of embezzlement as the test for whether an embezzlement offense is continuing because some conspiracies involve active steps to continue the conspiracy. The court argues that “conspiracy is a continuing offense which may or may not involve affirmative actions by the defendant after the elements of the offense are met” and thus “the active or passive nature of the defendant’s actions has never been the benchmark of a continuing offense under Toussie.”238 The Seventh Circuit correctly points out that the distinction between active and passive is not present in Toussie. As discussed in Part III.A, though, this is because the Court in Toussie was never presented with the question of whether different courses of conduct charged under the same statute may be discrete while others are continuing. In this light, it should not be surprising that the Court did not set the active or passive nature of a crime as the benchmark of a continuing offense.

Rather, the more important inquiry is whether classic continuing offenses like conspiracy continue until affirmatively ended. If an affirmative ending is not the benchmark for a continuing offense, then passive embezzlement schemes may not be continuing offenses. A conspiracy ends if the objective is achieved239 or an individual withdraws, which requires affirmative actions by the defendant to distance himself from the conspiracy or disclose the plan to law enforcement.240 Even if a defendant forms a conspiracy and does nothing else, “a defendant’s membership in the conspiracy, and his responsibility for its acts, endures even if he is entirely inactive after joining it.”241 In other words, an individual’s want of active participation is not a defense. As a result, a conspiracy “is presumed to continue until there is affirmative evidence of abandonment, withdrawal, disavowal, or defeat of the object of the conspiracy.”242 Accordingly, the characteristics of the classic continuing offense of conspiracy establish that an affirmative ending is a key characteristic of a continuing offense—a characteristic that is common to passive embezzlement schemes.

2. Passive schemes involve harm that necessarily persists so long as the offense continues.

In Toussie, the Court recognized that for continuing offenses, “each day’s acts bring a renewed threat of the substantive evil Congress sought to prevent.”243 Similarly, Bailey recognized that escape is a continuing offense due to “the continuing threat to society posed by an escaped prisoner.”244 In other words, “the harm done to society through their commission necessarily continues on for as long as the crime is ongoing.”245

Passive schemes of embezzlement result in harm that “necessarily continues” so long as the offense persists. In Brunell’s case, each day he concealed information from the SSA brought a renewed harm that § 641 was enacted to prevent. Conversely, active schemes of embezzlement do not bring a “renewed threat” each day. Rather, the renewed threat occurs each time the defendant takes an affirmative action to continue the offense. For example, in Thompkins and Ellis, which both involved active embezzlement of SSA benefits for deceased relatives,246 there was not a renewed threat unless the defendant took the additional affirmative action of forging their deceased relatives’ signatures. And although the harm to society continued over a period of time, it did not automatically continue.

There is good reason to require that the harm Congress sought to prevent necessarily occurs. Justice Byron White, dissenting in Toussie, argued:

Our own cases distinguish the “instantaneous” from the “continuing” offense on the theory that in the former case, the illegal aim is attained as soon as every element of the crime has occurred, whereas in the latter case, the unlawful course of conduct is “set on foot by a single impulse and operated by an unintermittent force,” until the ultimate illegal objective is finally attained.247

Passive embezzlement schemes fit within this framework, which the DC Circuit considers persuasive.248 Further, requiring an “unintermittent force” appreciates that the term “continuing offense” is a term of art and ensures a strong distinction between active and passive embezzlement schemes. A series of discrete offenses could never be considered a continuing offense because the harm of each subsequent discrete offense does not necessarily occur without the defendant’s intermittent intervention. As a result, the distinction between passive and active schemes appreciates the deeper meaning of the term “continuing offense.”

Unlike active schemes, the harm from passive embezzlement schemes could theoretically continue indefinitely. When a defendant makes the choice to conceal a fact affecting the SSA’s payment of benefits, the defendant exercises control not just over one payment, but over all future payments.249 Until the defendant makes an affirmative disclosure or is apprehended, the future stream of benefits is essentially in the defendant’s possession. On the other hand, when periodic verification of a condition like a disability is required, the defendant only exercises control over benefits that will be automatically disbursed until the defendant is again required to verify the disability. So even if Defendant A steals the same amount as Defendant B but uses a passive—rather than active—scheme of embezzlement, courts should worry more about Defendant A because the potential amount of loss when the scheme was set in motion was much greater than that of Defendant B.

Moreover, in practice, the harm from passive embezzlement schemes is multiplied by the fact that passive schemes are harder to detect. Most commonly, embezzlement schemes involving Social Security benefits are detected through an investigation by the SSA.250 When a defendant is able to sit back without filing paperwork to verify a disability,251 investigators are not going to have a paper trail to tip them off. 252 In one extreme example, a Tennessee woman fraudulently received at least 347 checks from the SSA over the course of twenty-nine years because the SSA had no record of her aunt’s death.253 Indeed, in some circumstances, passive schemes of embezzlement can be paperless crimes. The effortlessness of passive embezzlement schemes may make them more enticing. As a result, it is more important to create strong incentives for reporting the death of a relative, like in Brunell, than to deter individuals from filing false statements with the SSA.

3. Passive embezzlement schemes involve concealment, which itself is a continuing offense.

Treating passive embezzlement schemes as continuing offenses makes sense because the core conduct of passive embezzlement schemes—concealment—is independently considered a continuing offense. Courts deem concealment offenses continuing because they have all the characteristics of continuing offenses.254 Under 42 USC § 1383a(a)(3)(A), a person can be convicted if he or she has knowledge of an event affecting a right to any Social Security benefit and fails to disclose such event with intent to fraudulently secure the benefit. Similarly, 42 USC § 408(a)(4) prohibits concealment of information affecting a right to insurance benefits.255 First, in finding that offenses charged under both § 1383a(a)(3)(A) and § 408(a)(4) are continuing offenses, courts note “[t]he crime of concealment of events affecting benefits, like the traditional continuing offenses of conspiracy and escape, continues until it is affirmatively ended.”256 Judge Thapar notes that concealment is “quite like a conspiracy” because each day the defendant furthers their offense by concealing material facts affecting continued rights to benefits.257 Second, concealment meets the harm principle: “[R]epeated receipt of undeserved benefits through concealment was precisely the ‘substantive evil’ that Congress sought to prevent with § 1383a(a)(3)(A).”258 To be sure, courts consider an active scheme to conceal by making a false statement259 a discrete offense because of the affirmative false disclosure involved.260

Passive schemes of embezzlement from the SSA operate in the same manner as concealment offenses and should therefore be deemed continuing offenses. Often, defendants are charged under § 641 and either § 408(a)(4) (concealing facts related to Social Security benefits) or § 1383a(a)(3)(A) (concealing facts related to insurance benefits) for their embezzlement scheme.261 In a paradigmatic passive embezzlement scheme—in which the embezzlement scheme involves concealment and is automatically executed, like in Smith—the conduct that meets the elements of § 641, as well as § 408(a)(4) or § 1383a(a)(3), is identical. As a result, there is no reason to treat the same underlying conduct as both continuous and discrete at the same time. On the other hand, in a situation in which concealment is part of a larger active scheme of embezzlement, meaning the defendant engages in additional affirmative actions on top of the concealment, it makes sense to allow courts to separate the continuous element (concealment) from the discrete elements (active embezzlement). This is exactly what the district court did in United States v Bundy.262 While the court found that the defendant’s active embezzlement scheme involving false statements to the SSA regarding her past employment and disability was discrete for statute of limitations purposes, it nevertheless held separately that concealment of her ongoing employment under § 408(a)(4) was a continuing offense.263 Accordingly, when the defendant’s passive embezzlement scheme closely mirrors the type of conduct charged under § 408(a)(4) or § 1383a(a)(3)(A), courts should find that scheme a continuing offense.

C. Purposes of the Statute of Limitations

Finally, the charged conduct approach fits neatly within Congress’s purposes behind the criminal statute of limitations. First, the goal of fairness through evidentiary protections264 is not present for embezzlement offenses charged under § 641, as the evidence the government will likely rely on is a death certificate and bank records. It is unlikely that there will be eyewitnesses whose memories will fade over time or physical evidence that can deteriorate. Without a series of affirmative actions, a long paper trail of fraud is unlikely.265 Further, to the extent there are evidentiary concerns, applying the continuing offense doctrine to passive embezzlement schemes is unlikely to have any effect on wrongful convictions. As Professors Paul Robinson and Michael Cahill point out, despite wide variations in state statutes of limitations, there is no known discrepancy in wrongful convictions based on stale evidence.266 Second, the statute of limitations seeks to give law enforcement an incentive to investigate crimes promptly.267 But given the massive scale on which Social Security benefits are distributed, it is hard to see how a few limited cases are going to change the investigators’ conduct. The effect, therefore, is negligible.

Despite the lack of evidentiary and law enforcement incentivization concerns, however, courts might not be the right institution to make this conclusion. Toussie warns that the doctrine of continuing offenses should be applied in limited circumstances in light of Congress’s policy that the statute of limitations should not be extended “except as otherwise expressly provided by law.”268

Even so, while Congress did not expressly extend the statute of limitations to passive embezzlement schemes, doing so through the continuing offense doctrine is supported by Congress’s practice of extending the statute of limitations in other circumstances. As Smith recognized, passive embezzlement schemes often occur in small amounts over a long period of time, making them difficult to prosecute.269 When Congress extended the statute of limitations for procurement fraud cases, it did so because these cases were extremely difficult to prosecute and discover due to their complexity.270 While passive embezzlement schemes are not necessarily complex, the frequency of this issue in courts shows the pervasive lag time in investigation and detection. Additionally, Congress prescribed a “general rule that the limitations period does not run during a period of fraudulent concealment”271 when it passed the Major Fraud Act of 1988.272 Because passive schemes of embezzlement mirror concealment offenses, it is consistent with Congress’s intent to apply the continuing offense doctrine to these offenses.

Conclusion

Embezzlement of public benefits is prevalent in the United States. In 2015, the SSA reported over 6.5 million Social Security recipients age 112 or older, yet worldwide there are only 35 people known to reach that age.273 Given that the Social Security program in the United States is woefully underfunded, it is more important than ever to ensure those who perpetrate embezzlement schemes are held fully accountable. Of particular concern are passive embezzlement schemes, which continue automatically once set in motion. Because these schemes do not require any effort on the embezzler’s part, they often run well beyond the five-year statute of limitations period and could bar the government from prosecuting the full embezzlement scheme. The continuing offense doctrine is one legal mechanism to ensure that those who commit passive embezzlement schemes are held fully accountable. Courts, however, are divided as to whether passive embezzlement schemes are continuing offenses. Some courts use a categorical approach and hold that no embezzlement schemes are continuing offenses; other courts use a charged conduct approach and hold that only passive embezzlement schemes are continuing.

This Comment argues that courts do not adequately justify their choices between the categorical and charged conduct approaches. To decide between the charged conduct and categorical approaches, courts rely solely on Toussie v United States to guide their decision. But Toussie, in its simplicity, never implicated either approach and as a result cannot resolve the circuit split. Fortunately, the Supreme Court has made the choice between a conduct-based and categorical approach before in the sentencing and removal contexts. The Court recognized that for divisible statutes, which criminalize diverse classes of offenses, a more careful approach is required in order to accurately discern the correct sentencing and removal consequences. Recognizing that 18 USC § 641 also criminalizes a diverse range of embezzlement schemes, courts should adopt the charged conduct approach to more accurately apply the continuing offense doctrine.

Applying the charged conduct approach allows courts to correctly apply the continuing offense doctrine to passive embezzlement schemes. Not only do passive embezzlement schemes have the two key characteristics of continuing offenses, as they continue until affirmatively ended and involve a harm that necessarily persists so long as the offense continues, but courts treat passive schemes’ core conduct—concealment—as a continuing offense.

• 1United States v Brunell, 320 F Supp 3d 246, 247 (D Mass 2018).
• 2Id.
• 3Id.
• 4Id.
• 5Brunell, 320 F Supp 3d at 247. See also 18 USC § 3282(a) (establishing a generally applicable five-year limit on prosecution of offenses other than murder).
• 6Brunell, 320 F Supp 3d at 248. It is important to separate the meaning of the term “continuing offense” in the statute of limitations context from its meaning in the venue context. Whereas in a venue context the term means that an offense was started in one district and completed in another, the term means something different in the statute of limitations context. See United States v Dunne, 324 F3d 1158, 1165–66 (10th Cir 2003).
• 7See, for example, United States v Reese, 254 F Supp 3d 1045, 1050 (D Neb 2017) (finding that an embezzlement scheme involving multiple false affirmations “does not describe a continuous and passive scheme”).
• 8See United States v Smith, 373 F3d 561, 567–68 (4th Cir 2004) (“At least in those cases where the defendant created a recurring, automatic scheme of embezzlement under section 641 . . . we think that Congress must have intended that such [conduct] be considered a continuing offense.”).
• 9Pendergast v United States, 317 US 412, 418 (1943).
• 10United States v Yashar, 166 F3d 873, 878–79 (7th Cir 1999) (differentiating continuing offenses from discrete offenses).
• 11United States v Bailey, 444 US 394, 413–14 (1980) (holding that escape is a continuing offense for which the statute of limitations begins to run when the criminal is apprehended).
• 12United States v Payne, 978 F2d 1177, 1180 (10th Cir 1992) (suggesting that concealment of an event affecting the continued right to payment of social security benefits, prohibited by 42 USC § 408(a)(4), is a continuing offense). See also United States v Henrikson, 191 F Supp 3d 999, 1005 (D SD 2016) (holding that concealment of an event affecting the right to benefits under 42 USC § 408(a)(4) is a continuing offense). Another example of a continuing offense is the possession of contraband. See United States v Pease, 2008 WL 808683, *2 (D Ariz) (holding that a defendant who possesses contraband commits a continuing offense for which the statute of limitations does not begin to run until possession ceases).
• 13For example, the harm from the continuing offense of kidnapping necessarily occurs until the victim is released, whereas the harm from a string of robberies does not necessarily occur until the next robbery. See United States v Morales, 11 F3d 915, 921 (9th Cir 1993) (O’Scannlain concurring in part and dissenting in part) (explaining that kidnapping is a continuing offense).
• 14Brunell, 320 F Supp 3d at 248.
• 15Compare Smith, 373 F3d 561 (Fourth Circuit holding that a passive embezzlement scheme is a continuing offense), with Yashar, 166 F3d 873 (Seventh Circuit holding that neither active nor passive embezzlement schemes are continuing offenses).
• 16397 US 112 (1970).
• 17Id at 115.
• 18See, for example, Smith, 373 F3d at 568 (holding that only some embezzlement schemes are continuing offenses).
• 19Id. See also Jeffrey R. Boles, Easing the Tension between Statutes of Limitation and the Continuing Offense Doctrine, 7 Nw J L & Soc Pol 219, 238 (2012) (remarking that this approach can “be best described as following a ‘charged conduct’ approach”).
• 20See, for example, Yashar, 166 F3d at 877 (holding that no schemes of embezzlement are continuing offenses).
• 21Brunell was indicted on October 26, 2017 for an embezzlement scheme that ran at least from January 2004 to May 2017 (161 months). Brunell, 320 F Supp 3d at 247. If Brunell’s embezzlement scheme were not a continuing offense, then all conduct before October, 26, 2012 would be barred by the five-year statute of limitations. October 2012 to May 2017 is 55 months.
• 22Even if part of an embezzlement scheme is barred by the statute of limitations, judges have discretion over whether to consider this time-barred conduct when calculating the Guidelines range. See United States v Booker, 543 US 220, 267 (2005) (holding that judges can consider conduct not submitted to the jury if it does not increase the sentencing range above the statutory maximum). Judges disagree over whether time-barred conduct can be considered for sentencing. Compare United States v Williams, 217 F3d 751, 754 (9th Cir 2000) (holding that actions occurring outside limitations period could be considered in determining base offense level), with United States v Bell, 808 F3d 926, 928 (DC Cir 2015) (Kavanaugh concurring in denial of rehearing en banc) (“Allowing judges to rely on acquitted or uncharged conduct to impose higher sentences than they otherwise would impose seems a dubious infringement of the rights to due process and to a jury trial.”).
• 23United States v Silkowski, 32 F3d 682, 690 (2d Cir 1994). See also United States v Green, 897 F3d 443, 448 (2d Cir 2018) (holding that the district court may not order the defendant “to pay restitution for losses attributable to acts of conversion that [the defendant] committed outside the five-year limitation period”). Admittedly, other circuits find that restitution can be imposed for acts beyond the statute of limitations. See United States v Dickerson, 370 F3d 1330, 1342 (11th Cir 2004) (holding that “the defendant [must] pay restitution to all victims for the losses . . . even where such losses were caused by conduct outside of the statute of limitations”); United States v Bach, 172 F3d 520, 523 (7th Cir 1999) (upholding restitution based on entire Ponzi scheme, even though only two mailings were within statute of limitations period). In other circuits, however, the issue remains unresolved. See United States v Pelletier, 2017 WL 5162800, *5 (D Me) (noting that the First Circuit has yet to decide the issue).
• 2418 USC § 641 (establishing a $1,000 threshold). See also 18 USC § 666(a)(1)(A) (establishing a $5,000 threshold for embezzlement from an organization receiving federal funds). In United States v Askia, 893 F3d 1110, 1114 (8th Cir 2018), the court found that the defendant’s embezzlement was not a continuing offense and thus prosecution of most of the embezzlement scheme was barred. In order to save the entire indictment from dismissal, the government offered proof of embezzlement within the statutory period, meeting the $5,000 threshold by just $503.36. Id at 1115.
• 25For examples of recent convictions, see Broward Resident Found Guilty of Stealing Deceased Grandparents’ Social Security Funds (US Attorney’s Office for the Southern District of Florida, Jan 24, 2018), archived at http://perma.cc/9NTM-5K46 (woman convicted under § 641 for embezzling $130,000 in social security payments); Stafford Man Convicted of Social Security Disability Benefits Fraud (US Attorney’s Office for the Eastern District of Virginia, Feb 24, 2017), archived at http://perma.cc/T2MW-EDXD (man convicted under § 641 for embezzling over $75,000 in social security disability benefits); Ed White, Probation for Man Who Cashed $122K of Dead Grandma’s Social Security (Detroit Free Press, June 20, 2016), archived at http://perma.cc/68J4-HTQ7.
• 26Office of the Inspector General, Social Security Administration, How the OIG and the SSA Investigate Deceased Payee Fraud (Jan 28, 2016), archived at http://perma.cc/Q5GK-GMYD. While this statistic does not take into account whether these cases involved active or passive embezzlement schemes, these cases involve the same core embezzlement mechanism as many passive embezzlement schemes discussed throughout this Comment.
• 27Boles, 7 Nw J L & Soc Pol at 222 (cited in note 19).
• 2818 USC § 3282(a).
• 29Toussie, 397 US at 114–15.
• 30Id at 114.
• 31United States v Marion, 404 US 307, 321 (1971).
• 32Id at 322.
• 33Stogner v California, 539 US 607, 615 (2003).
• 34Toussie, 397 US at 115.
• 35See Lindsey Powell, Unraveling Criminal Statutes of Limitations, 45 Am Crim L Rev 115, 130–31 (2008) (arguing that a limitations period forces law enforcement to emphasize economy and security).
• 36Pendergast v United States, 317 US 412, 418 (1943).
• 37See United States v Silkowski, 32 F3d 682, 690 (2d Cir 1994) (limiting the loss calculation to losses stemming from conduct within the statute of limitations period).
• 38See Paul H. Robinson and Michael T. Cahill, Law without Justice: Why Criminal Law Doesn’t Give People What They Deserve 60 (Oxford 2005) (arguing that “[b]y overtly shielding guilty offenders from the punishment they deserve, statutes of limitation advertise the criminal-justice system as condoning a failure of justice”).
• 39Powell, 45 Am Crim L Rev at 122–24 (cited in note 35) (noting, for example, that Congress extended the statute of limitations to ten years for violations of naturalization laws and to eight years for terrorism offenses after the World Trade Center bombings).
• 40Act of Apr 30, 1790, ch 9, § 32, 1 Stat 112, 119 (establishing a three-year limitations period for capital offenses other than forgery and willful murder and a two-year period for most other offenses); Act of Apr 13, 1876, ch 56, 19 Stat 32, 32–33 (establishing a three-year limitations period for all offenses not capital or otherwise provided for in existing law); Act of Sept 1, 1954, ch 1214, § 10(a), 68 Stat 1142, 1145 (changing the general statute of limitations from three to five years), codified as amended by Act of Sept 26, 1961, Pub L No 87-299, 75 Stat 640, 648.
• 41See Powell, 45 Am Crim L Rev at 124–26 (cited in note 35) (noting that Congress extended the statute of limitations for insurance fraud in response to corporate scandals in 1994 and for offenses against children due to increasing concern for child welfare).
• 42See 18 USC § 3282 (clarifying that offenses that are “not capital” are subject to the statute of limitations).
• 43Major Fraud Act of 1988, S Rep No 100-503, 100th Cong, 2d Sess 3 (1988), reprinted in 1988 USCCAN 5969, 5970–71.
• 44Major Fraud Amendments Act of 1989, Pub L No 101-123, 101 Stat 248, codified as amended at 18 USC § 1031(f).
• 45Pendergast, 317 US at 418.
• 46United States v Yashar, 166 F3d 873, 879 (7th Cir 1999). See also Toussie, 397 US at 136 (White dissenting).
• 47720 ILCS 5/21-1(a)(7).
• 48United States v Morales, 11 F3d 915, 921 (9th Cir 1993) (O’Scannlain concurring in part and dissenting in part).
• 49Id. See also note 13.
• 50See Morales, 11 F3d at 921 (O’Scannlain concurring in part and dissenting in part) (“The classic example of a continuing offense is a conspiracy.”), quoting United States v McGoff, 831 F2d 1071, 1078 (DC Cir 1987); United States v Rivlin, 2007 WL 4276712, *2 (SDNY) (“Conspiracy, escape, kidnapping, and crimes of possession are traditional examples of continuing offenses.”).
• 51Morales, 11 F3d at 919, 921 (O’Scannlain concurring in part and dissenting in part).
• 52Toussie, 397 US at 115.
• 53United States v Jaynes, 75 F3d 1493, 1506 (10th Cir 1996) (citation omitted).
• 54Id (noting that “a continuing offense is not the same as a scheme or pattern of illegal conduct”).
• 55Yashar, 166 F3d at 875, quoting Toussie, 397 US at 122.
• 56See, for example, 22 USC § 618(e) (providing that the failure of an agent of a foreign nation to register as such “shall be considered a continuing offense for as long as such failure exists”); 18 USC § 3284 (“The concealment of assets of a debtor in a case under title 11 shall be deemed to be a continuing offense.”).
• 57See United States v Kissel, 218 US 601, 610 (1910) (finding conspiracy a continuing offense despite no explicit statutory language).
• 5865 Stat 75 (1951), codified as amended at 50 USC § 453.
• 59Toussie, 397 US at 112–13.
• 60Id.
• 61Id at 114.
• 62Id.
• 63Toussie, 397 US at 115.
• 64Id at 115–20.
• 65Id at 122.
• 66Id.
• 67Toussie, 397 US at 122.
• 68Id at 116–19.
• 69Id at 123–24.
• 70444 US 394 (1980).
• 71Id at 398.
• 72Id at 412.
• 73Id at 413.
• 74Bailey, 444 US at 413, quoting Toussie, 397 US at 115.
• 75See, for example, 18 USC § 3284; 22 USC § 618(e).
• 76Toussie, 397 US at 115.
• 77Benjamin E. Rosenberg, “Continuing Crimes” and Statutes of Limitations *4 (Westlaw Journal of White-Collar Crime, Nov 2010).
• 78See United States v Smith, 373 F3d 561, 568 (4th Cir 2004).
• 79See United States v Yashar, 166 F3d 873, 876–77 (7th Cir 1999).
• 80See generally Smith, 373 F3d 561.
• 81See generally United States v Neusom, 159 Fed Appx 796 (9th Cir 2005).
• 82See Smith, 373 F3d at 568. See also notes 6–8 and accompanying text.
• 83373 F3d 561 (4th Cir 2004).
• 8418 USC § 641. The SSA refers to this type of scheme as “deceased payee fraud.” See SSA, How the OIG and the SSA Investigate Deceased Payee Fraud (cited in note 26).
• 85Smith, 373 F3d at 563.
• 86Id.
• 87Id.
• 88Id at 563. Smith was indicted on January 24, 2003; his embezzlement scheme ran from March 1994 to February 3, 1998. Id.
• 89See Smith, 373 F3d at 568.
• 90Id at 564.
• 91The court’s only reference to Toussie in the continuing offense context is in conclusion: “We are satisfied, however, that in addition to being properly aggregated into a single count, the particular kind of embezzlement that occurred in this case is correctly considered, under Toussie, to be a continuing offense.” Id at 568.
• 92Id at 567–68.
• 93For cases using these characteristics as the definition of a passive embezzlement scheme, see United States v Reese, 254 F Supp 3d 1045, 1050 (D Neb 2017) (“Smith held that a passive scheme of fraudulent conduct which requires no action by the defendant is a continuing offense.”); United States v Phan, 754 F Supp 2d 186, 190–91 (D Mass 2010) (describing Smith’s criteria as the type of embezzlement scheme that is a continuing offense).
• 94Smith, 373 F3d at 568.
• 95159 Fed Appx 796 (9th Cir 2005).
• 96Id at 798–799, citing United States v Morales, 11 F3d 915, 918 (9th Cir 1993). In Morales, the court explicitly rejected the government’s argument that the continuing offense doctrine applied. See generally Morales, 11 F3d 918.
• 97See United States v Niven, 952 F2d 289, 293 (9th Cir 1991) (holding that “the analysis turns on the nature of the substantive offense, not on the specific characteristics of the conduct”), overruled on other grounds by Nichols v United States, 511 US 738 (1994).
• 98See, for example, United States v Orlando, 358 F Supp 3d 160, 161 (D Mass 2019); United States v Brunell, 320 F Supp 3d 246, 248–49 (D Mass 2018); United States v Turner, 2014 WL 641768, *1 (ED Ark); United States v Morris, 2014 WL 2560617, *5 (ED Ark); United States v Wharton, 2014 WL 1430387, *8–10 (D Md); United States v Tackett, 2011 WL 4005347, *3–6 (ED Ky); United States v Street, 2008 WL 4372737, *3–4 (SD Miss); United States v Gibson, 2008 WL 4838226, *2–3 (WD Mo).
• 99See, for example, Brunell, 320 F Supp 3d at 248–49; Street, 2008 WL 4372737, *3–4.
• 100754 F Supp 2d 186 (D Mass 2010).
• 101Id at 188–91.
• 102Id (emphasis added), citing Toussie, 397 US at 115.
• 103Phan, 754 F Supp 2d at 190, quoting Morales, 11 F3d at 921 (O’Scannlain concurring in part and dissenting in part).
• 104Phan, 754 F Supp 2d at 190, quoting United States v Pease, 2008 WL 808683, *2 (D Ariz).
• 105Phan, 754 F Supp 2d at 190.
• 106Id at 190, quoting Morales, 11 F3d at 921 (O’Scannlain concurring in part and dissenting in part).
• 107Phan, 754 F Supp 2d at 190.
• 108See generally United States v Silkowski, 32 F3d 682 (2d Cir 1994).
• 109See generally Yashar, 166 F3d 873.
• 110See generally United States v Askia, 893 F3d 1110 (8th Cir 2018).
• 111See id at 1119. This Comment responds to this policy argument in Part III.C.
• 11232 F3d 682 (2d Cir 1994).
• 113Id at 684–86.
• 114Id.
• 115Id at 690.
• 116Silkowski, 32 F3d at 690.
• 117166 F3d 873 (7th Cir 1999).
• 118Id at 875.
• 119Id, citing 18 USC § 666(a)(1)(A).
• 120See Askia, 893 F3d at 1118 n 3 (“Section 641 is analogous in all relevant respects here to § 666(a)(1)(A). Each offense prohibits unlawfully taking property; the offenses differ merely based on the lawful owner or possessor of the property and the property’s value.”) (citation omitted); United States v Sunia, 643 F Supp 2d 51, 73 (DDC 2009) (same).
• 121Yashar, 166 F3d at 875. Yashar signed a waiver that effectively tolled the statute of limitations on August 13, 1997. Therefore, if his embezzlement were a discrete offense, the government could only indict him for conduct after August 13, 1992. This would leave only one month of wages and benefits. Id.
• 122Id at 875. See also 18 USC § 666(a)(1)(A)(i).
• 123Yashar, 166 F3d at 877.
• 124Id. This Comment addresses this argument in Part III.B.1.
• 125Id at 879–80.
• 126Yashar, 166 F3d at 878. This Comment addresses this argument in Part III.C.
• 127893 F3d 1110 (8th Cir 2018).
• 128Id at 1115.
• 129Id. See also note 120 and accompanying text.
• 130Askia, 893 F3d at 1117, 1119 (focusing on “the language and elements of the offense, rather than the facts alleged or the charge itself”).
• 131Id at 1118, citing United States v McGoff, 831 F2d 1071, 1078 (DC Cir 1987).
• 132Askia, 893 F3d at 1118–19.
• 133Id at 1119.
• 134Id.
• 135Yashar, 166 F3d at 877. See also Andrew P. O’Shea, Note, Criminal Law—An Embezzlement Intermezzo: Scheming to Side-Step Toussie v. United States’s Continuing Offense Test, 35 W New Eng L Rev 249, 276 (2013) (concluding that “[t]o prevent such arbitrary application of criminal law, courts should limit prosecutorial discretion in deciding whether to write an indictment as a continuing offense”).
• 136Yashar, 166 F3d at 877.
• 137Id at 879.
• 138Boles, 7 Nw J L & Soc Pol at 245 (cited in note 19).
• 139A single Westlaw search of “‘embezzlement’ AND ‘statute of limitations’ AND ‘18 USC § 641’” conducted on April 9, 2019 yielded eighteen cases in which courts adopted the categorical approach, all of which involved active embezzlement schemes.
• 140See notes 92–94 and accompanying text.
• 141See Yashar, 166 F3d at 875.
• 142Askia, 893 F3d at 1114–15.
• 143Brief for Appellee, United States v Silkowski, No 93-1520, *3 (2d Cir filed Oct 15, 1993).
• 14499 F Supp 3d 262 (D RI 2015).
• 145Id at 263.
• 146United States’ Memorandum of Law in Opposition to the Defendant’s Motion to Dismiss, United States v Powell, No 14-808-M, *2 (D RI filed Feb 24, 2015) (Government’s Memorandum in Powell).
• 147Smith, 373 F3d at 568.
• 148Powell, 99 F Supp 3d at 264, 267. For another example of affirmative actions, see Government’s Sentencing Memorandum, United States v Young, No CR-140-B-W, *7 (D Me filed Aug 25, 2010) (defendant “took a number of actions to hide the receipt of this money from others, including changing the address on the joint back [sic] account to insure [sic] that [others] never saw the bank statements”).
• 149See Phan, 754 F Supp 2d at 187–88.
• 1502008 WL 808683 (D Ariz).
• 151Id at *3.
• 152Government’s Response to Defendant’s Motion to Dismiss the Indictment, United States v Pease, No CR-07-757-PHX-DCG, *3 (D Ariz filed Mar 13, 2008) (Government’s Motion in Pease).
• 153Pease, 2008 WL 808683 at *1–3. See also United States v Bundy, 2009 WL 902064, *1–2 (D Me) (defendant made multiple false representations to the SSA that she was not working to gain disability benefits).
• 154Smith, 373 F3d at 563.
• 155Id at 567–68 (emphasis added).
• 1562015 WL 248362 (WD La).
• 1572008 WL 3200629 (WD NC).
• 158Ellis, 2015 WL 248362 at *3; Thompkins, 2008 WL 3200629 at *1.
• 159Ellis, 2015 WL 248362 at *3; Thompkins, 2008 WL 3200629 at *2. See also United States v Morris, 2014 WL 2560617, *5 (ED Ark) (finding that defendant’s “regular, ongoing and active theft and concealment of SSA and VA funds” was a continuing offense) (emphasis added); United States v George, 2011 WL 4559122, *1–2 (WD Wash) (deeming active embezzlement a continuing offense under the charged conduct approach); United States v Easley, 2011 WL 2265116, *1–2 (ED Ark) (same).
• 1602011 WL 4005347 (ED Ky).
• 161Id at *1.
• 162Id at *4 (“Under some circumstances, however, theft or embezzlement can be a continuing offense.”).
• 163Id at *5.
• 1642017 WL 5957734 (D Md).
• 165Id at *4.
• 166See notes 160–65 and accompanying text.
• 167A single Westlaw search of “‘embezzlement’ AND ‘statute of limitations’ AND ‘18 USC § 641’” conducted on April 9, 2019 yielded thirty-nine cases involving the choice between the categorical and charged conduct approaches. Of the twenty-one courts that adopted the charged conduct approach, all but four did so when presented with a passive embezzlement scheme.
• 168See Neusom, 159 Fed Appx at 798; Smith, 373 F3d at 563–64.
• 1692008 WL 4838226 (WD Mo).
• 170Id at *2.
• 171254 F Supp 3d 1045 (D Neb 2017).
• 172Id at 1050. See also, for example, United States v Johnson, 145 F Supp 3d 862, 871 (D SD 2015) (finding that “[e]ven if one were to agree with the Fourth Circuit’s reasoning in Smith, [the defendant’s] conduct differs from the scheme in that case. Contrary to the embezzlement in Smith, [the defendant’s] embezzlement was not automatically recurring”); United States v Duhamel, 770 F Supp 2d 414, 416 (D Me 2011) (recognizing that “the Indictment here does not allege a ‘recurring, automatic scheme’” in holding that § 641 is categorically not a continuing offense), quoting Smith, 166 F3d at 567–68.
• 173Pub L No 98-473, 98 Stat 2185, codified at 18 USC § 924(e).
• 174Pub L No 89-236, 79 Stat 911 (1965), codified at 8 USC § 1101 et seq.
• 175Phan, 754 F Supp 2d at 188 (emphasis added).
• 176Yashar, 166 F3d at 877 (emphasis added).
• 177Smith, 373 F3d at 568.
• 178Id at 569 (Michael dissenting).
• 179Phan, 754 F Supp 2d at 188 (paraphrasing “crime involved” as “offense charged”).
• 180Toussie, 397 US at 115.
• 181Id.
• 182Askia, 893 F3d at 1119, quoting Toussie, 397 US at 115.
• 183Yashar, 166 F3d at 879.
• 184See Part II.C.
• 185Askia, 893 F3d at 1117 (describing Toussie as “examining the text and legislative history of the offense”).
• 186Toussie, 397 US at 122.
• 187See United States v Powell, 99 F Supp 3d 262, 266 (D RI 2015) (arguing that the Court is clear that “the ‘inherent nature’ inquiry is to be made of the crime itself independent of any facts peculiar or unique to a particular case”); Boles, 7 NW J L & Soc Pol at 243–44 (cited in note 19) (arguing that the charged conduct approach contravenes Supreme Court precedent by improperly “transfer[ring] the focus of the continuing offense assessment from the congressional intent inherent in a criminal statute to the particular charged conduct in a given case”).
• 188Northern Security Co v United States, 193 US 197, 400 (1904) (Holmes dissenting).
• 189For a discussion of this legal theory, see O’Bannon v Town Court Nursing Center, 477 US 773, 804 (1980) (Blackmun concurring) (arguing that “easy cases make bad law”).
• 190Toussie, 397 US at 113.
• 191Boles, 7 NW J L & Soc Pol at 245 (cited in note 19).
• 19218 USC § 924(e). Previously, this statute included a residual clause that extended the definition of “violent felony” to offenses that “present[ ] a serious potential risk of physical injury,” but this clause was held to be void for vagueness by the Court in Johnson v United States, 135 S Ct 2551, 2555, 2563 (2015).
• 1938 USC § 1227(a)(2)(A)(iii). In Sessions v Dimaya, 138 S Ct 1204, 1210 (2018), the Supreme Court invalidated the part of the definition of “aggravated felony” that mirrored the residual clause invalidated in Johnson. See note 192.
• 19418 USC § 924(e)(2)(B)(ii). See also Mathis v United States, 136 S Ct 2243, 2247 (2016) (“To determine whether a past conviction is for one of those offenses [burglary, arson, or extortion], courts compare the elements of the crime of conviction with the elements of the ‘generic’ version of the listed offense”).
• 195570 US 254 (2013).
• 196Id at 258, 263.
• 197Id at 267.
• 198See Nijhawan v Holder, 557 US 29, 41 (2009); Moncrieffe v Holder, 569 US 184, 190 (2013) (“When the Government alleges that a state conviction qualifies as an ‘aggravated felony’ under the INA, we generally employ a ‘categorical approach’ to determine whether the state offense is comparable to an offense listed in the INA.”).
• 199While Johnson and Dimaya struck down the residual clauses in the definitions of “violent felony” and “aggravated felony,” respectively, courts still make comparisons to generic offenses included in the definitions of these terms using the modified categorical approach. See, for example, United States v Garrido, 736 Fed Appx 77, 78 (5th Cir 2018) (using the modified categorical approach to determine whether a conviction is a “violent felony”); United States v Almanza-Vigil, 912 F3d 1310, 1318 (10th Cir 2019) (affirming the modified categorical approach when a statute is divisible); United States v Abbott, 748 F3d 154, 158 (3d Cir 2014) (affirming a district court’s application of the modified categorical approach to a divisible Pennsylvania controlled substance statute).
• 200See Michael McGivney, Comment, A Means to an Element: The Supreme Court’s Modified Categorical Approach after Mathis v. United States, 107 J Crim L & Crimin 421, 423–26 (2017).
• 201495 US 575 (1990).
• 202Id at 592–99. For a description of a state statute involving alternative elements, see Nijhawan, 557 US at 35 ( “A single Massachusetts statute section entitled ‘Breaking and Entering at Night,’ for example, criminalizes breaking into a ‘building, ship, vessel or vehicle.’’’).
• 203For more on divisible and indivisible statutes, see McGivney, 107 J Crim L & Crimin at 423–26 (cited in note 200).
• 204Descamps, 570 US at 269–70. Although Apprendi v New Jersey, 530 US 466, 474–476 (2000), which held that all conduct that increases the limit of the possible federal sentence must be found by a jury, was decided before Taylor, the Court recognized these Sixth Amendment concerns when affirming Taylor in Shephard v United States, 544 US 13, 24 (2005).
• 205Taylor, 495 US at 602. In Descamps, the Court defined Taylor’s approach as “a modified categorical approach,” clarifying that it “permits sentencing courts to consult a limited class of documents, such as indictments and jury instructions.” Descamps, 570 US at 257.
• 206Taylor, 495 US at 602 n 1 (noting that under the ACCA, burglary is only a violent felony if committed in a building, not a boat or motor vehicle).
• 207Dickson v Ashcroft, 346 F3d 44, 48 (2d Cir 2003).
• 208See 18 USC § 641 (providing that whoever “steals” is guilty of a crime). See also, for example, United States v Scott, 784 F2d 787, 788–89, 793 (7th Cir 1986) (upholding conviction for stealing checks from an agency receiving federal funding).
• 209See, for example, United States v Brunell, 320 F Supp 3d 246, 247 (D Mass 2018).
• 210See United States v Morison, 844 F2d 1057, 1060 (4th Cir 1988) (prosecution under 18 USC § 641 for theft of classified photographs of Soviet naval preparations).
• 211Morissette v United States, 342 US 246, 266 n 28 (1952) (“The history of [§] 641 demonstrates that it was to apply to acts which constituted larceny or embezzlement at common law and also acts which shade into those crimes but which, most strictly considered, might not be found to fit their fixed definitions.”).
• 212Akinsade v Holder, 678 F3d 138 (2d Cir 2012).
• 213Carlos-Blaza v Holder, 611 F3d 583 (9th Cir 2010).
• 214See Akinsade, 678 F3d at 145; Carlos-Blaza, 611 F3d at 588–89.
• 215611 F3d 583 (9th Cir 2010).
• 216Id at 588, citing United States v Gillett, 249 F3d 1200, 1201–02 (9th Cir 2001) (affirming conviction for theft under 18 USC § 656 in which a guard stole bags of money from a federally insured bank).
• 21718 USC § 641 (providing that whoever “steals” anything of value from the United States commits a crime). Compare Morissette, 342 US at 260–61 & n 19 (1952) (noting that § 641 imports the common law meaning of theft, which does not require intent to deceive, only an “intent to deprive an owner of his property”), with United States v Dupee, 569 F2d 1061, 1064 (9th Cir 1978) (noting that “a classic case of embezzlement” involves “the fraudulent conversion of property”); Criminal Resource Manual § 1638 (Department of Justice 2019), archived at http://perma.cc/9ZEP-JL6N (noting that “fraudulent conversion or appropriation” is an element of a § 641 prosecution).
• 218Carlos-Blaza, 611 F3d at 588 (quotation marks and citation omitted).
• 219Id at 588 n 9.
• 220Id at 589 (applying the modified categorical approach by “compar[ing] to the statutory definition the following types of documents: the charging document, the written plea agreement, the transcript of plea colloquy, any explicit factual finding by the trial judge to which the defendant assented, and comparable judicial records”). See also Akinsade, 678 F3d at 145 (proceeding under the assumption that § 656 is divisible).
• 221Toussie, 397 US at 115.
• 222See Section III.B.2.
• 223United States v Morales, 11 F3d 915, 921 (9th Cir 1993) (O’Scannlain concurring in part and dissenting in part).
• 224Id.
• 225Pease, 2008 WL 808683 at *2.
• 226Id.
• 227See id (“Classic examples of continuing offenses include conspiracy, escape, kidnapping, bigamy, and crimes of possession . . . [which], once committed, continue in effect until affirmatively ended.”); United States v Bundy, 2009 WL 902064, *12 (D Me) (holding that defendant’s failure to disclose a material fact affecting the defendant’s benefits, “like the classic continuing offenses of conspiracy, escape, kidnapping, bigamy, and crimes of possession, by its nature continues in effect until affirmatively ended”).
• 228Yashar, 166 F3d at 877.
• 229See Phan, 754 F Supp 2d at 190 (holding that “[w]rongfully continuing to receive SSI benefits as a result of an initial misrepresentation . . . continues until it is ‘affirmatively ended’’’).
• 230320 F Supp 3d 46 (D Mass 2018).
• 231Id at 247.
• 232Id.
• 233See Ellis, 2015 WL 248362 at *3 (finding no continuing offense when the defendant forged a signature in order to cash benefits checks sent to a deceased relative).
• 234See Powell, 99 F Supp 3d at 263; Government’s Memorandum in Powell at *2 (cited in note 146).
• 235See Pease, 2008 WL 808683 at *3; Government’s Motion in Pease at *3 (cited in note 152).
• 236See notes 155–59 and accompanying text.
• 237See note 47 and accompanying text.
• 238Yashar, 166 F3d at 877.
• 239United States v Payne, 591 F3d 46, 69 (2d Cir 2010) (“Conspiracy is a continuing offense . . . that involves a prolonged course of conduct; its commission is not complete until the conduct has run its course.”).
• 240See United States v Bostick, 791 F3d 127, 143 (DC Cir 2015).
• 241Smith v United States, 568 US 106, 114 (2013).
• 242United States v Castro, 972 F2d 1107, 1112 (9th Cir 1992) (emphasis added), abrogated on other grounds by United States v Jimenez Recio, 537 US 270 (2003).
• 243Toussie, 397 US at 122.
• 244Bailey, 444 US at 413.
• 245Morales, 11 F3d at 921 (O’Scannlain concurring in part and dissenting in part) (emphasis added).
• 246Thompkins, 2008 WL 3200629 at *1; Ellis, 2015 WL 248362 at *3. See also notes 156–59 and accompanying text.
• 247Toussie, 397 US at 135–36 (White dissenting) (emphasis added), quoting United States v Midstate Co, 306 US 161, 166 (1939).
• 248United States v McGoff, 831 F2d 1071, 1079 n 15 (DC Cir 1987) (noting that “[Justice White’s] dissenting opinion provides a useful discussion of the general principle of continuing offenses” because “[t]he majority did not consider the offense a continuing one and therefore had no occasion to dispute Justice White’s description”).
• 249See United States’ Response to the Defendant’s Motion to Dismiss, United States v Gibson, No 08-03057-01-CR-S-FJG, *6 (WD Mo filed Aug 22, 2008) (arguing that “the continuing failure to notify [the SSA]” was an exercise of control over not just the SSA property already in the defendant’s possession, but over “future payments to the defendant”).
• 250See, for example, Bundy, 2009 WL 902064 at *1 (investigation by the SSA Office of Inspector General that uncovered defendant’s active embezzlement). Often, an investigation is prompted by a complaint from the public. See, for example, Government’s Motion in Pease at *3 (cited in note 152) (noting that an anonymous complaint started the investigation).
• 251In Bundy, the defendant completed paperwork that falsely represented her employment history in order to obtain SSA benefits. The defendant’s fraudulent receipt of benefits was uncovered by an SSA employee reviewing the defendant’s file. Bundy, 2009 WL 902064 at *4.
• 252Compare Smith, 373 F3d at 567 (concealment of a family member’s death), with Government’s Motion in Pease at *2–3 (cited in note 152) (alleging multiple false statements made by the defendant on various applications for disability benefits).
• 253United States v Dinsmore, 585 Fed Appx 898, 899 (6th Cir 2014) (upholding sentencing for plea agreement under § 641).
• 254See, for example, Payne, 978 F2d at 1180.
• 255Compare 42 USC § 408(a)(4) (fraud related to disability insurance benefits), with 42 USC § 1383a(a)(3)(A) (fraud related to supplemental Social Security income). The statutes have identical language regarding concealment.
• 256United States v Henrikson, 191 F Supp 3d 999, 1005 (D SD 2016) (holding that 42 USC § 408(a)(4) is a continuing offense). See also Bundy, 2009 WL 902064 at *12 (finding that § 408(a)(4) is a continuing offense because “by its nature [the offense] continues in effect until affirmatively ended, in this case, by revelation of the concealment”).
• 257United States v Banks, 708 F Supp 2d 622, 627 (ED Ky 2010).
• 258Id.
• 25918 USC § 1001.
• 260See United States v Dunne, 324 F3d 1158, 1164–65 (10th Cir 2003); Brogan v United States, 522 US 398, 413 (1998) (Ginsburg concurring) (noting that 18 USC § 1001 was intended “to protect the Government from the affirmative, aggressive and voluntary actions of persons who take the initiative [to make false statements]”).
• 261See United States v Treacy, 2014 WL 12698499, *1 n 1 (WD Va) (charging defendant under both 18 USC § 641 and 42 USC § 408(a)(4)); United States v Wharton, 2014 WL 1430387, *1 (D Md) (charging defendant under both § 641 and § 1383a(a)(3)).
• 2622009 WL 902064 (D Me).
• 263Id at *11–12.
• 264See Part I.A.
• 265See Note, The Statute of Limitations in Criminal Law: A Penetrable Barrier to Prosecution, 102 U Pa L Rev 630, 637–38 (1954) (suggesting embezzlement is a crime “in which memory is not so important a factor as in the usual case, because the crimes are evidenced by a written matter”).
• 266Robinson and Cahill, Law without Justice at 61 (cited in note 38) (surveying state statutes of limitations and finding that “despite these wide variations in limitation rules, [the authors] are aware of no demonstrated discrepancy in the experience of different states in terms of the central concern of limitation rules: improper outcomes based on stale, or vanished, evidence”).
• 267See Part I.A.
• 268Toussie, 397 US at 115, quoting 18 USC § 3282(a).
• 269Smith, 373 F3d at 567. See also notes 250–51 and accompanying text.
• 270Major Fraud Act of 1988, Pub L No 100-700, 102 Stat 4631, codified as amended in various sections of Title 18; S Rep No 100-503 (cited in note 43).
• 271S Rep No 100-503 at *14 (cited in note 43).
• 272Pub L No 100-700, 102 Stat 4631, codified as amended in various sections of Title 18.
• 273Office of the Inspector General, Social Security Administration, Numberholders Age 112 or Older Who Did Not Have a Death Entry on the Numident *3 (Mar 4, 2015), archived at http://perma.cc/96VX-TYWM.