A man has free choice to the extent that he is rational.

Thomas Aquinas

Introduction

In the past 20 years, there has been a great deal of effort expended within criminology to explain desistance from crime [27, 29, 31, 49, 54, 55, 61, 66]. Arguably,Footnote 1 the two most prominent theoretical explanations of criminal desistance currently in the field are the age-graded theory of informal social control by Sampson and Laub [66]; [49]) and the theory of cognitive and emotional transformations by Giordano and her colleagues [27–29].Footnote 2 Although there are some important differences between these two theories of desistance—Sampson and Laub’s theory rests on a social control foundation that emphasizes the role of conventional social bonds such as a stable job and good marriage, while Giordano’s symbolic interactionist/neo-Meadian theory relies much more on emotional growth and changes in the preferences of offenders—there is important common ground between them. Both theories seem to give a great deal of weight to the importance of human agency in their theories. Unfortunately, although both discuss the importance of human agency in desistance, there is an unreconciled tension and discrepancy between the main theoretical account of desistance and the role played by human agency. Recent empirical work that relies on the voices of offenders themselves, however, is illuminating the explanatory gaps in these theories [12, 59, 70, 73]. As such, it would appear that the time has come for theoretical reflection within the discipline.

In this paper, we suggest that there are untenable theoretical linkages in both the age-graded theory of informal social control and the theory of cognitive and emotional transformations. Second, we contend that both theories show tendencies to portray those who commit criminal offenses as less than purposive, intentional actors, and more at the mercy of social forces upon which they have little control. In short, while actors in both theories are fully social, they seem to lack the basic components of strong human agency—intentionality, power, reflexivity, and the capacity for self-examination or monitoring. They are, in the words written long ago by Wrong [82], “over-socialized” (homo sociologicus). While we acknowledge that the genuine human agency is differentially provided in each theory, with the offenders in the age-graded theory possessing far less human agency compared to the cognitive and emotional transformations theory, actors in both theories lack what we think are full properties of human agents.

While Paternoster and Bushway’s [61] identity theory of desistance (ITD) can resolve some of the theoretical difficulties, it too can be accused of treating actors as behaving without full human agency because of its reliance on rational choice theory. Instead of being creatures whose skills, motivations, and competencies are entirely bestowed by society, as many social theories portray persons, critics have argued that the traditional rational choice theory of neo-classical economics paints an equally non-agentic picture of humans as socially atomized bargain hunters simply looking for the best price [6, 37, 38]. However, important work in behavioral economics has left this caricature behind. Nobel prize winner George Akerlof and Rachel Kranton have taken the lead in creating a more human model of rational choice by incorporating the idea of identity into economics in their book Identity Economics [3]. In this paper, we agree with Akerlof and Kranton that productive rational choice models will be as far from homo economicus as it is from homo sociologicus.

In the next section, we briefly discuss prevailing theories of criminal desistance including a detailed discussion of what we see to be the critical differences among them for the purpose of illustrating the contribution of Paternoster and Bushway’s ITD to the field. This is followed by a summary of the extant literature predicting desistance from crime and how the findings from this empirical work question the conclusion that jobs, marriages, and cognitive/emotional transformations are all that matter for desistance to occur. The paper concludes with some discussion of the role of identity change in desistance and directions for further theoretical development and empirical research.

Theories of Desistance from Crime

While theoretical speculation about the factors related to desistance from crime go as far back in criminological history as the original work of the Gluecks [30], desistance as a source for criminological theorizing was sadly dormant until Gottfredson and Hirschi [31] and Sampson and Laub [66] breathed some badly needed life back into this area of inquiry. Although perhaps best known for their self-control theory, Gottfredson and Hirschi also presented what would turn out to be an equally controversial theory of desistance. To them, desistance was a process of moving from offending to non-offending that was rather uniform; offenders begin to cease offending from a common starting point and at roughly the same rate over time. Unlike any other past or present desistance theorists, Gottfredson and Hirschi argued that given this uniformity, desistance from crime need not, and indeed could not, be explained by any set of traditional criminological or sociological variables. Instead, this uniformity urged that desistance from crime was simply a biological consequence of age—the “inexorable aging of the organism” ([31]: 141).

While this theory was certainly parsimonious, Gottfredson and Hirschi’s denial of all things sociological (or psychological or economic) was not likely to endear them to many criminologists who saw a great deal of heterogeneity rather than uniformity in the desistance process. For critics of the theory, some offenders continued to offend throughout their lives, and those who quit did so at different times and different rates, clearly indicating different causal mechanisms. Into this explanatory vacuum stepped Sampson and Laub [66], who developed their age-graded theory of informal social control. Although this first work briefly mentioned desistance from crime, their desistance theory was not fully elaborated until their later book, Shared, Beginnings, Divergent Lives [49].Footnote 3 Here, they used their revised age-graded informal social control theory to explain desistance by extending the Glueck offending data to age 70 and conducting extensive life history narratives with a small group of the original sample. The Glueck data were comprised of white adolescent boys who, though delinquents during adolescence, were transitioning to adulthood during the 1950s. This historical context included many opportunities to obtain well-paying manufacturing jobs along with educational benefits via the GI Bill. The backbone of this life course theory of desistance relied on the assertion that offenders quit crime when they established strong conventional bonds through marriage, military service, and stable employment (the legacy from their original 1993 theory). The slightly revised theory developed in 2003 added to this “the interplay of human agency and choice, situational influences, routine activities, local culture, and historical context” (2003: 9). Essentially, this life course theory of desistance hypothesized that exogenously generated turning points, such as finding the right intimate partner, landing a stable and satisfying job, or having a successful stint in the military, could serve to generate a downward deflection in a previous offender’s criminal offense trajectory because it strengthened a weak social bond and gave offenders what Toby [74] originally called a greater “stake in conformity.”

For Laub and Sampson [49]: 148–149) then, entering pro-social roles initiates and is both a necessary and sufficient part of the desistance process. Although turning points can have diverse effects such as “identity transformation,” it is clear from their work over the years that Sampson and Laub’s position is that any internal cognitive or psychological change that offenders undergo occurs later in the desistance process and only as a result of participation in conventional social roles. Although there are some ambiguities in the theory,Footnote 4 it is clear that identity change or other individual-level changes are not needed as initial steps, or not even needed at all, for desistance to occur. In fact, they clearly state, “Our stance on the desistance process contrasts with emerging theories of desistance that emphasize cognitive transformations or identity shifts as necessary for desistance to occur…” (2003: 278).

Other much less structural theories of desistance were emerging at the same time as Sampson and Laub’s. One of these was developed and elaborated by Maruna and colleagues [23, 54-56]. Maruna’s explanation of desistance relied on the notion of an offender’s identity, but not on the idea that an identity change either preceded or was an important causal mechanism of desistance. For Maruna, offenders do not change their identity from an offending antisocial person embedded in a life of crime to one who now sees themselves in a different, conventional light. Rather, offenders who already have pro-social views of themselves in the present deliberately reinterpret their offender pasts to make previous criminal actions both explicable and consistent with their current favorable views of who they are and what they are “really like.” In Maruna’s own words ([54]: 154): “[D]esisting is framed as just another adventure consistent with their lifelong personality, not as a change of heart. Again, this allows the individual to frame his or her desistance as a case of personality continuity rather than change” (emphasis added). As such, Maruna’s desisting offenders do not change who they are as much as they change the interpretation and understanding of their criminal past, so that it is reconciled with their current view of themselves as “good” people. This does not involve change as much as it does a “willful cognitive distortion” of the past to align it with the present; it is the cognitive work described as “making good” (p. 9).

A theory that did put great weight on the mental work that must accompany desistance from crime was developed by Giordano and her colleagues [27, 29]. In its original formulation, Giordano et al. [27] argued that while turning points such as jobs and marriages, or what they called “hooks for change,” were important for desistance to take place, there must first occur two types of cognitive transformations in the offender: (1) There must be a greater openness to change, and (2) the individual must perceive the hook as being salient or important for them. Other types of cognitive changes, such as changes in how one views crime or deviant behavior and the fashioning of a new conventional “replacement self,” occurs only after and as a result of involvement in conventional roles or “hooks.”

In their subsequent enhancement of the theory, Giordano et al. [29] posit an important additional type of transformation not discussed in their earlier paper. In addition to the cognitive transformations that make up the bulk of the 2002 paper, this revision included emotional transformations as a critical component of desistance from crime.Footnote 5 They argued that in many if not most offenders’ lives, there had been conflict with parents and intimate others and that recurring instances of these conflicts eventually molds an angry or a depressive self that was instrumental in offenders’ continued involvement in crime. Left unaddressed, this angry/depressive self leads to persistence in offending.Footnote 6 However, some (many? most?) offenders eventually find pro-social romantic partners who provide them with both a role model and social support for their initial steps toward breaking from crime. Also, in a process of social learning, these intimate pro-social intimates provide new pro-social definitions. In the 2007 version of the theory, then, participation in social roles and role taking with a pro-social partner helps offenders break from crime through both emotional (changing the way anger is understood and managed) and cognitive (new definitions of a criminal lifestyle) transformations.

Paternoster and Bushway [61]; [14, 15])Footnote 7 offer the most recent theoretical formulation explaining desistance, which they call an ITD. It starts with a clear definition of human agency. Following Dietz and Burns [19] and Bandura [7], they state that human agency consists of the following properties: intentionality, forethought, reflexivity, and power. Persons possess intentionality when they have a deliberate purpose for their action; in short, they make things happen according to a strategy rather than solely reacting to things happening to them. Forethought is the capacity to create future goals and plans, and through these “cognitive representations, visualized futures are brought into the present as current guides and motivators of behavior” ([7]: 164). Reflexivity is the ability to self-monitor and self-regulate so that decisions made and actions taken can be revisited and revised. Finally, human agency requires some amount of power or self-efficacy which the real or perceived ability a person has that their actions can produce the desired outcome that is attended.

Agentic offenders, Paternoster and Bushway [61] contend, will retain an “offender” working identity as long as they perceive that they are getting more benefits than costs from crime. Every criminal offender confronts failure (they get apprehended, the person being robbed fights back, they go to prison and their loved one takes up with another, the stolen goods net very little money), however, as long as they attribute these failures to something beyond them, such as “bad luck” (that is, they possess a self-serving bias), [58]) and as isolated events, they are able to continue offending. The process of changing an offender’s identity occurs when, after repeated failures, the illusion of the self-serving bias begins to be questioned. Specifically, “when perceived failures and dissatisfactions within different domains of life become connected and when current failures become linked with anticipated future failures” ([61]:1105). Once offenders come to the realization that their criminal offending is more costly than beneficial, and the failures they are experiencing are due to their own insufficiencies and are therefore likely to continue into the foreseeable future, they make initial moves to change their identity (and ultimately their life) to one that is more pro-social. A key feature of the ITD, then, is that initial moves toward desistance come about as the result of a “feared self”—an image of what the person does not want to be or fears becoming. The feared self provides the first step toward desistance, but to maintain these initial steps, the offender must craft a new, more positive image of what they want to become, “the possible self.”Footnote 8 This newly emerging pro-social identity or possible self then triggers a change in the person’s preferences for things like quick and easy money (via theft or drug dealing), or the “party life” [71], for more pro-social things like conventional employment and social relationships. It is this cluster of internal changes in identity and preferences and the crafting of the kind of person that one wants to be that both motivates behavior consistent with a pro-social identity (a life free of crime and drugs) and sends a signal to others (like potential pro-social intimates and employers) that the person is making a change in their life, which, in turn, makes pro-social opportunities more likely.

The important time order articulated by the ITD, then, is a change in a former offender’s identity that both explains the movement into conventional roles or “hooks” and explains why those who had previously been involved in crime would ever be receptive to these pro-social influences.Footnote 9 Further, while conventional turning points are extremely useful in desisting, they are not essential since one with a changed identity can, though not without difficulty, cobble together a life that does not involve criminal behavior, even if their life does not include a good job or good partnership. In sum, emotionally satisfying intimate relationships and stable employment are not essential for desistance, but a change in one’s identity is. It is this change in identity that is the willful purposive act of self-improvement that leads, in turn, to other pro-social changes.

There is of course some common ground between the ITD and other theories of desistance. For example, the identity theory is similar to the age-graded theory in that conventional social roles such as having a good job or being in a satisfying intimate relationship are implicated in desistance. In contrast, however, involvement in these roles does not initiate desistance. The kinds of stable jobs and emotionally satisfying marriages that are at the core of Sampson/Laub’s theory are obviously helpful but neither necessary nor sufficient for desistance to occur. Comparable to Giordano et al.’s theory, the identity theory posits that there must be substantial “up front” work before things like stable jobs and pro-social partners can either arrive or be successfully taken advantage of. However, ITD contends that identity change initiates rather than follows changes in preferences and social roles. In the ITD, the up front work that is required implicates individual mental processes rather than purely social mechanisms as in the theory of cognitive and emotional transformations. We highlight the differences between ITD and other desistance theories in the following section.

Differences Between Identity Theory and Other Theories of Desistance

We begin with Sampson and Laub’s age-graded theory. As just noted above, in Paternoster and Bushway’s schema, involvement in pro-social turning points or hooks like full-time employment and marriage are obviously helpful, but not essential for desistance. Serious criminal offenders, many with drug and alcohol addiction, are unlikely to have the skills for what are today virtually non-existent well-paying jobs in the largely urban areas where they are likely to be released [60, 81]. For these same reasons, they are unlikely to make attractive marriage partners for pro-social mates [21]. Second, available turning points can only be initiated and can only be effective and sustained after an offender has first determined that their previous criminal self is no longer acceptable and they want to change. That is, good jobs and good partnerships can only be found and kept once offenders have first made an initial identity change toward a pro-social self.

Further, in our view and others,Footnote 10 one of the great weaknesses of Sampson and Laub’s theorizing is that they do not fully explain two critical components of the turning point theory of desistance: (1) how opportunities such as good jobs and pro-social intimates arrive in the lives of criminal offenders, and (2) should they find them, why former offenders would even be responsive to the direct or indirect social control efforts of conventional partners or employers. Sampson and Laub repeatedly note that turning points arrive randomly, “selection is surely operating at some level, but most marriages originate in fortuitous contacts rooted in everyday routine activities” ([49]: 45; emphasis added), and that desistance primarily occurs because conventional jobs and marriages provide direct social control or monitoring of the activities of former offenders. They state, “what has not received enough attention is the role that marriage plays in restructuring routine activities and the direct social control that spouses provide” ([49]: 135). … “even more than marriage, work, especially full-time work, leads to a meaningful change in routine activities. Work restricts many criminal opportunities…” ([49]: 47). This is problematic given what we know about assortative mating with respect to both attitudes and behavior [48] as well as the role of conventional social networks in providing information about job opportunities and how conventional opportunities are difficult to come by for those embedded in crime [32–34, 39]. Absent some internal change in the offender and a consequent “signal” that they have changed,Footnote 11 conventional opportunities either for marriages or full-time jobs are unlikely to present themselves. For ITD, an identity change comes first in the causal sequence, and only then can it be followed by changes in one’s social networks, preferences or tastes (say for excitement, “easy money,” or the “party life”), conventional employment, and pro-social relationships like legal or common law marriage. This view was also expressed by Bushway and Reuter's [16] earlier observation that employment was unlikely to lead to desistance in the absence of a personal commitment and deliberate intention to quit crime on the part of the offender, and Giordano et al. [28] point that unless a former offender had already committed to change, they would be unmoved by a spouse’s effort to supervise and/or restrict their social interactions.

It is important to note that there are significant contextual differences between the sample upon which Sampson and Laub's Theory [66]; [49]) is based and contemporary offenders leaving prison today. The Glueck boys, which formed the empirical foundation of Sampson and Laub’s theory, transitioned to adulthood during the post-WWII period when there were opportunities such as the GI Bill, and well-paying union manufacturing jobs that also enabled them to be good marriage prospects for women who were comparatively more pro-social. Further, while many of the Glueck men reported having problems with alcohol, none had experience with drugs that are a common feature in the landscape today—crack and powdered cocaine, heroin, and others. Many offenders now exist in an environment of concentrated disadvantage, with high rates of substance abuse, where there are low rates both of marriage and living-wage job opportunities for ex-convicts, particularly African-Americans [60].Footnote 12 Communities into which offenders are returned now also have higher rates of crime. Uniform Crime Report statistics indicate that the violent crime rate in 2013 was more than twice as high as the rate in 1960, and the property crime rate was more than one and one half times higher [76]. The ITD recognizes that if there is a change in identity and a motivation to break from crime, ex-offenders can still carve out a life generally free of crime without a full-time job and supportive spouse. An offender who has decided to quit crime can live with family or friends; work hourly jobs at a labor pool or temp agency; sell their blood; receive financial help from family, relatives, and acquaintances; and develop a support structure that would not likely help or help for long without a recognized general commitment to change. In fact, in today’s bleak economic environment with many collateral consequences, such as employment and housing restrictions that accompany a criminal record [60], informal work and assistance may be the only way many offenders can desist.

Another important difference between the ITD and Sampson and Laub’s revised age-graded theory of desistance is that the latter takes a rather ambiguous position with respect to human agency. In different iterations of the theory, Sampson and Laub have at times argued that their theory of desistance gives great weight to human agency. In fact, the provision of human agency was posited to be one of the important revisions between the 1993 and the 2003 and later versions of the theory (2003: 9; 2005: 38). For example, they contend in places that human beings are willful, intentional, agentic persons (2003:56): “Fortunately, as developed in more detail below, what is most striking in the narratives we collected is the role of human agency in processes of desistance from crime and deviance. The Glueck men are seen to be active players in their destiny, especially when their actions project a new sense of a redeemed self.” Elsewhere they appear to double down: “What is most striking in the narratives we collected is the role of human agency, or choice, in desistance from crime and deviance. The men who desisted are ‘active’ players in the desistance process … “the men we studied were active participants in the decision to give up crime” (2003: 141, 146). We are, however, not at all convinced that human agency, at least in terms of purposive action,Footnote 13 is an important feature of their theory.Footnote 14 For example, they conclude that desistance by default “best fits the desistance process we found in our data” (2003: 278; emphasis added), which occurs without any conscious awareness: “Desistance for our subjects was not necessarily a conscious or deliberate process”… many men made a commitment to go straight without even realizing it” (2003: 278–279), and “our main point is that many of the desisters did not seek to make good—they simply desisted with little if any cognitive reflection on the matter” (2003: 279). It is difficult to rectify these statements with their other statements regarding agency.

More recent summary statements regarding their theory do not facilitate a remedy for this quandary. For example, in 2005, they assert that theirs is “a life course view that takes human agency seriously” (p.14), but also state that, “many men made a commitment to go straight without even realizing it” (p. 37, emphasis added) and that desistance actions are “willed by the offender” though they may be “below the surface of active consciousness” (p. 38). In the articulation of their “revised” theory in 2007 (p. 326), they note that it consists of (1) social control, (2) routine activities, and (3) human agency but that their conception of desistance “rejects the notion that cognitive transformation is necessary for desistance to occur” (p. 326). To avoid ambiguity, they defined human agency as “intentional action that may or may not be accompanied by an identity change” (p. 326). Despite this homage to agency, we contend that offenders in this theory do not act, they react—they simply respond to the demands placed on them by spouses and employers.Footnote 15 The sense that Sampson and Laub’s offenders are merely reacting to the demands of the roles, they find themselves in can be vividly seen in a passage from Howard Becker that they cite approvingly:

“It is enough to create situations which will coerce people into behaving as we want them to and then to create the conditions under which other rewards will become linked to continuing this behavior” ([49]: 149).

Reactors who passively respond to the role demands made on them are simply not agentic according to our definition of agency, which stresses intentionality and forethought. Footnote 16

In sum, we contend that the identity theory of desistance can help address many questions left unanswered in Sampson and Laub’s age-graded informal theory. Perhaps the most important issue is the identification of the force or forces that motivate “intentional action.” In other words, “why would those who have had active criminal lives want to have good jobs and conventional spouses in the first place, especially when they are virtually impossible to obtain with a criminal record?” The ITD also explains the receptivity of former offenders to opportunities to turn their life around, thereby providing an answer to the question, “once former offenders have access to conventional opportunities, what prevents them from mishandling them, as many offenders have done on more than one occasion?” Pro-social opportunities only arrive and are only made successful, when there has first been an internal change in criminal offenders—a realization that their previous involvement in crime is no longer appealing and that they want to change who they are. Finally, the ITD makes a clear statement about the importance of willful, intentional conduct, or the capacity to act on one’s own behalf—one important dimension of human agency [22, 36] in desistance from crime.

The convergence of Paternoster and Bushway’s identity theory with Giordano’s cognitive/emotional theory is closer, but there nevertheless are fundamental differences between them. First, it is clear that Giordano et al. [27] posit that a change in an offender’s identity is included within the domain of “cognitive transformations.” What is not clear, however, is whether the offender must first change his or her identity in order for “hooks for change” (like jobs and marriage) to arrive or successfully acted upon, if the hooks trigger identity change, or both (if both, which is more salient than the other). The fact that cognitive transformations are referred to as the “up front work” (2002: 992) of the desistance process would imply that identity change comes before anything else; however, it is also stated that “[h]ooks for change can provide an important opening in the direction of a new identity” and “hooks influence the shift in identity” (p. 1002).Footnote 17 The latter view about temporal order is also consistent with the causal diagram of their theory (Figure 1, p. 1029) wherein identity transformation is a consequence of involvement in conventional roles like parent, spouse, and worker. Further, since emotional transformations involve role taking experiences in pro-social romantic relationships, we would presume that they occur only after exposure to conventional hooks. Footnote 18

A second point of departure between the ITD and Giordano’s is the context in which human agency is relevant. In the theory of cognitive/emotional transformations, human agency has a more limited role compared to the ITD. Giordano et al. [27]: 1026) argued that a general proposition of their theory is that “on a continuum of advantage and disadvantage, the real play of agency is in the middle.” Their position appears to be, then, that human agency is not relevant when social and personal capital is high, and not likely to be enough to overcome daunting obstacles for those under great economic and social adversity. In the ITD, however, the agentic action of deciding to quit crime and moving in that direction by changing one’s identity is requisite at all levels of the advantage/disadvantage continuum. Because it makes this clear hypothesis, the ITD is easily falsifiable and would constitute another way to empirically distinguish it from Giordano’s desistance theory. At the risk of being too redundant, ITD contends that dissatisfaction with one’s old identity and the decision to transform one’s self is required for any desistance from crime to take place. While resources for desistance may be available, even if they are considerable resources, they will not be utilized in the long term unless someone has first decided that they no longer want to be committed to crime. As we will show in the next section, empirical qualitative evidence is beginning to illuminate this. Further, agency may be even more important among those with very few economic advantages, because self-determination and intention may be the most powerful factors driving change. Those who are resolute about changing their self and life, but who have neither a full-time job, nor a supportive pro-social spouse, and are further hobbled by drug addiction, an arrest history and their marginalized gender, race or ethnicity, may, nevertheless, with considerable effort and assistance from family, welfare agencies, and other sources, patch together a life free from crime and substance abuse.

A third point of departure of Paternoster and Bushway’s ITD and Giordano et al.’s theory is that the former places much greater emphasis on a negative conception of self as initiating desistance from crime, which only later leads to a more positive or possible self. Paternoster and Bushway [61] argued that at least the initial strides in breaking from crime come about because of the feared self. After the linking of failures, projecting of these failures into the future, and the attribution of failures to one’s own shortfalls, offenders begin to contemplate their futures. The feared self is literally what offender’s do not want to become and fear becoming—a homeless drug addict, imprisoned for long periods without seeing children or other family—and it is this fear that provides motivation for crafting what Giordano et al. [27]: 1001) have called a “replacement self,” a more positive identity based on what the person wants to be.

This leads to the fourth difference between Paternoster and Bushway’s theory and Giordano et al.’s theory—it does not clearly specify where the cognitive and emotional transformations originate. For example, in the 2002 paper where cognitive transformations such as an openness to change are important, Giordano et al. do not explain what brings about this new openness or recognition. These cognitive transformations are brought about (2002: 1003) by unexplained “agentic moves”—“… the actor creatively and selectively draws on elements of the environment in order to affect significant life changes.” In the 2007 revision, role taking and social reinforcement of conformity in pro-social relationships are added as emotional transformations to the theory, but it is not explained how former offenders are able to find these pro-social romantic partners. While Giordano et al. are fully aware of this difficulty and do not imply as Laub and Sampson [49] did that conventional partners arrive randomly for offenders, they are no more specific in explaining how romantic partners arrive. They simply note that the offender makes “agentic moves” (2007: 1607) toward pro-social others that in some way and for some reason gets reciprocated: “…we highlighted that, particularly in adulthood, the individual has an important role in making agentic moves in the direction of others who subsequently provide and reinforce the new definitions.” In this regard, and similar to Sampson and Laub, Giordano et al. do not explain in any detail either what these agentic moves are or what changes in attitude offenders must undergo in order for them to secure pro-social romantic partners.

Finally, the fifth and perhaps most important divergence between the ITD and Giordano’s cognitive/emotional theory is that the former is anchored in rational choice theory and the latter in Meadian and neo-Meadian symbolic interactionism, a distinction which has pronounced implications for the role and importance of human agency. Giordano et al.’s symbolic interactionist theory of desistance places great emphasis on the social causes of human action, including desistance from crime [13]. In spite of the talk about agentic moves, the desire for desistance is not something that emerges out of the heads of offenders, but through individuals’ participation in social role taking and social learning processes of imitation and social reinforcement. As social constructionists, therefore, Giordano et al. highlight the consequences of role taking experiences and social interactions and eschew more individualistic processes. In fact, they ([29]: 1607) are openly hostile to the possibility that desistance is a result of individual mental processes, arguing that even one’s personal thoughts are not the construction or possession of individuals, but are social products: “According to Mead and other symbolic interactionists, then, thoughts, while located within the individual, are nevertheless deeply social in origin. This is an important point, because it steers us away from a view of cognitive transformations as deriving from individualistic mental processes.” In fact, one of the reasons for the revised theory in 2007 (p. 1614) was that Giordano et al. wanted to further distance themselves from any scent of an individualist explanation and toward a more social constructionist interpretation: “The theory of cognitive transformations, in turn, likely over-theorizes actor-based changes in perspective and the primacy of associated agentic moves. A focus on role taking and the character of emotions elicited through these positive social interactions, however, serves to highlight the fully social aspects of the catalyst-actor relationship.” In other words, the 2002 theory of cognitive transformations, which was more centered on individual agency, was corrected in the 2007 version and replaced with role taking in pro-social romantic relationships.

In contrast, Paternoster and Bushway’s ITD builds on developments in both realist social theory [6] and rational choice theory [3, 10, 11, 35, 46] and precisely emphasizes “individualistic mental processes” which exist within a given social context.Footnote 19 While a comprehensive discussion of the implications of a rational choice theory of identity is beyond the scope of this paper, some ideas do need to be articulated to highlight this difference between the more social view of Giordano et al.’s cognitive/emotional theory and Paternoster and Bushway’s ITD. To do this, we appeal to one older theoretical system (Max Weber) and two more recent ones (Margaret Archer and Raymond Boudon).

The Paternoster/Bushway identity theory of desistance is consistent with Weberian rationality. In Economy and Society, Weber argued for the importance of studying action. He defined action as conduct that has a subjective meaning to the person acting and contrasts this with behavior, which is conduct that is an involuntary reaction to some external stimulus. Unlike behavior, then, action is meaningful, purposive, and voluntarily conducted. The key to explaining action, then, is to understand the meaning that the action has for the person. Weber’s methodology, his tool to understanding meaningful human action, was verstehen, or getting inside the actor’s head and taking her/his point of view, what we might easily call methodological individualism [10]. As we have documented, Giordano et al. [27, 29] on the other hand are decidedly antimentalistic [17] and reject the importance of understanding human conduct by examining individualistic mental processes. In their view, humans do not take action that is motivated by deeply held beliefs, or sentiments, rather meaning (recall their view of the social nature of the thoughts of an individual) is something that emerges out of social interaction. Giordano et al. imply that action is not preceded or initiated by any individual mental state or concern or that these cognitive processes are even important to understand because they do not motivate or guide conduct. In discussing the overly social emphasis of what he called the situationalist approach, Campbell [17]: 36) expressed it best: “…the focus is less on the meanings informing the actions of individuals than on their capacity to understand and interpret what others mean by their actions” (emphasis added). By distancing themselves from “individualist mental processes,” Giordano et al.’s work blunts a great deal of human agency. We agree with Campbell's argument that “… because meaning creation and manipulation is no longer regarded as located in the individual, but is always represented as a ‘social’ possession, the vision of a free, meaning-creating and hence action-creating, individual has vanished. By transferring such processes from an intra-subjective to an inter-subjective or social setting, the individual human being is effectively deprived of the ability to engage in willed responsible actions” ([17]: 148).

Others have also challenged the social constructionist or situationalist position. In Being Human: The Problem of Agency; Archer [6] is sharply critical of two trends in social thought that, in her words, have both impoverished human agency. The first, which has been occurring since the Enlightenment, is traditional economic thought that has made human beings little more than price-attentive bargain hunters.Footnote 20 Traditional rational choice economists have impoverished human beings, she argues, by characterizing them as instrumental strivers for the best prices of the objects they pursue, ignoring both the possibility that humans can pursue non-instrumental ends and that one’s identity might be a source of preferences and motivators for action. This is nowhere better revealed than in the neo-classical economic notion that preferences are simply revealed by the choices one makes [68] and that the source of one’s preferences are epiphenomenal—degustibus non est disputandum—in matters of taste, there can be no dispute. Economists have, therefore, typically dealt with preferences as they have been revealed and have shown little interest in the possibility that human desires are the source of those preferences. Only the external conduct of homo economicus need be taken into account in understanding their preferences or goals of action [8], so there is no need to delve into cognitions to understand beliefs or desires.

Sociologists are no less guilty in Archer’s eyes of impoverishing humanity. The sociologist’s way of impoverishment is by presenting all human capabilities and properties as things that are given by society through adopted roles and conversations with others through language. The homo sociologicus that seems to be described by social constructionists is both made and driven by social forces over which they appear to have little conscious control or understanding. Archer’s critique is reminiscent of Wrong’s [82] complaint about the “over-socialized man” in sociology—a man that has “too much society: too little of the self” ([6]: 78). Taking a position that would be highly critical of the social constructionism of Giordano et al., Archer argues that rather than viewing meaning or even one’s self as something that is passively acquired from social interaction, “Self-consciousness derives from our embodied practices in the world [praxis]…. [o]ne of the most important social properties that we have, the power to know ourselves to be the same being over time, depends upon practice in the environment rather than conversation in society” ([6]: 7).

In her realist social theory, Archer is highly critical of both overly economical and overly social depictions of human beings and human agency. Given Archer’s account of the limitations of rational choice theory, how can it be used by Paternoster/Bushway to harbor a real sense of human agency? We would point out that both the strict rational choice theorist’s ignorance of human beliefs and desires and the overly socialized social constructionist’s perception that human power is due solely to social interaction can be overcome by expanding rational choice theory.

Aguiar and de Francisco [1] have distinguished between internalist and externalist rational choice models.Footnote 21 In the externalist model, exemplified by Paul Samuelson [68] and Gary Becker [8], preferences are revealed by action and there is no need to understand the beliefs, desires, or other mental states of people. In the internalist model, however, preferences are mental states that constitute a reason or motivation for a person to take action. In this model, persons’ desires and beliefs are causally linked to their decisions and actions. This internalist position was adopted by the sociologist Boudon [10, 11], who made two important arguments. First, he suggested that a complete explanation of social phenomenon required the removal of all “black boxes,” which requires explaining why individuals “behaved the way they did” ([11]: 5). Second, he argued that the traditional, externalist conception of rational choice was too narrow and could not explain even common phenomenon like player’s behavior in ultimatum games or voting decisions.Footnote 22 Boudon’s argument in his cognitive theory of action is that people’s actions are in harmony with their beliefs or “strong reasons” ([11]: 17), which do not have to be instrumental: “any action is caused by reasons in the mind of individuals.” One source of a person’s strong reasons for action is that such action is an expression of who they are or want to be—their identity—and that both one’s identity and the strong reasons that are associated with it provide a complete explanation.

In this view, one’s identity motivates action because it is the focus of preferences, tastes, desires, and beliefs and is what Giordano et al. [29]: 1614) would call an individualistic mental process, what Boudon would call strong reasons, and what Archer [6]: 79–80) refers to as the locus of our “ultimate concerns.” While a detailed discussion of these issues is beyond the scope of this paper, one’s identity as more individualistic than social and one which is the nucleus of strong reasons held by persons is consistent with the realist social theory view of identity [5]. To Archer [6], one’s self-identity is very much a personal product achieved by intentional acting on the world, that is, engaging in practical activities with an emphasis on practical activity, praxis, rather than language (for a similar position, see [17]). With respect to the social constructionist position, she asserts (2000: 4), “Bodies have properties and powers of their own and are active in their environment, which is much broader than ‘society’s conversation.’ The resultants of our embodied relations with the world cannot be construed as the gift of society. Constructionism thus impoverishes humanity by subtracting from our human powers and accrediting all of them—selfhood, reflexivity, thought memory, and emotionality—to society’s discourse.” The personal self emerges, then, though the “primacy of practice” ([6]: 121–153), the effect of which is “to make the embodied practices of human beings in the world more important than their social relations for the emergence of selfhood…” (p. 121).Footnote 23

With respect to the externalist, rational choice model of traditional economics, Archer [6]: 4) strongly asserts that human beings are motivated by more than just pricing, that their actions are expressions of their ultimate concerns, concerns that are “not a means to anything beyond them, but are commitments which are constitutive of who we are, and an expression of our identities.”Footnote 24 The notion that one’s identity is an important source of one’s preferences and desires and that preferences motivate action has also gained adherents in economics. Akerlof and Kranton [3]: 10), for example, have argued that one’s identity is an important birthplace for preferences and motivations for action and that “what people care about, and how much they care about it, depends in part on their identity.” Importantly for the ITD, just as preferences are bundled with our identity, should we change our identity, our preferences can change: “… people’s motives, or tastes, are partly of their own making. Choice of identity, then, may be the most important ‘economic’ decision a person ever makes… identity points us to a new reason why preferences can change…” ([3]: 10).

Paternoster and Bushway [61] is consistent with these extensions of rational choice theory and is different in important ways from extant theories explaining desistance. In particular, both Sampson and Laub’s age-graded informal theory of social control, nor Giordano et al.’s theory of cognitive and emotional transformations are not what Boudon [11] would call complete explanations because they do not explain why some offenders have access to conventional opportunities like jobs, marriages, and pro-social romantic partners and others do not. As such, these theories are not helpful in explaining desistance for offenders who have no job skills, are substance abusers, or who have little or no prospect of being married or having a relationship with a conventional partner. The ITD provides an alternative to a social constructionist image of one’s identity, thoughts, and beliefs as being entirely created through social interactions and can provide an explanation of intentional self-change and desistance among even those offenders who are the most disadvantaged and isolated. A summary of the positions taken by desistance theories in criminology is provided in Table 1. In the next section, we briefly review the empirical evidence with respect to these various desistance theories and conclude that the empirical support is equivocal at best. We present this view not as a prelude to an empirical study but to suggest that it is not a settled factual matter that desistance among today’s offenders can be attributed to either the causal effect of conventional turning points like jobs and marriages or to the consequences of better social and intimate relationships. In fact, many of the published empirical studies are consistent with the expectations of the identity theory. Similar to Paternoster and Bushway's [61] theoretical paper, our intent here is to compel the discipline to expand its collective imagination.

Table 1 Summary of desistance theories of crime
Full size table

Recent Empirical Studies of Desistance: Marriage, Employment, and Identities

In Shared Beginnings, Divergent Lives [49], which follows the original sample of Glueck delinquents with official arrest data until age 70 supplemented by extensive interviews with a subsample of 52 of the men, the authors offered several different theoretical mechanisms through which turning points, such as stable employment and marriage, led to desistance, seeming to privilege the direct controlling effect of stringent supervision. The official arrest data did, however, unequivocally demonstrate that by age 70, virtually all of the men had stopped accumulating arrests so that desistance from crime was normative. Further, both the initiation of desistance and its rate demonstrated such heterogeneity that they argued that it could not possibly be due to a uniform process like the biological aging of the organism and must instead be caused by social factors. Although there were several possible candidates for the mechanism of this causal process, Laub and Sampson seem to contend that it was not so much that former offenders became better persons (a change in propensity or criminality) but that crime became much more difficult for them to do (a change in opportunity or crime).Footnote 25

In subsequent publications with the Glueck data, these findings have been enhanced and elaborated but not substantially altered. For example, using data from the Glueck delinquent sample to age 32 and group-based trajectory models, Laub et al. [50] identified four distinct groups of offenders: two that clearly evidenced desistance and two that did not. They found that being married was related to desistance, but not all marriages were the same—early marriages that gradually built up social control over time were the most important for crime cessation. In Divergent Beginnings, Laub and Sampson [49] conducted a quantitative analysis in addition to their interviews with offenders and found that there was about a 30 % reduction in crime over the life course for those who were married. Revisiting these data with an inverse probability of treatment weighting method, Sampson et al. [67] found that being married was associated with an average 35 % reduction in the odds of crime compared with the hypothetical odds had the man not married. Because this effect lasted only as long as the man was in the state of marriage, social control was hypothesized to be responsible: “[a]pplication of IPTW [inverse probability of treatment weighting] to account for selection into marriage thus produces a large and consistent estimate of crime reduction during the years former delinquents are married” ([67]: 490; emphasis added).Footnote 26 The beneficial marriage effect did not differ across conditions; it reduced crime regardless of whether the marital attachment was strong or weak or if the spouse was herself criminal rather than conventional.

Not all the support for the crime reducing effect of marriage or intimate relationships is based upon the Glueck data, however. For example, while not examining desistance per se, two studies speak to the role of marriage in adult offending. First, Horney et al. [40] examined the month-to-month offending patterns over a period that ranged from 24 to 36 months in a sample of nearly 600 serious male adult offenders who were released from incarceration. They found that offending was lower in those months when former offenders were living with their spouses, but that living with an unmarried partner actually increased crime. Second, King et al. [45] looked at the relationship between marriage and offending up to approximately age 24 in the National Youth Survey. Using a propensity score matching method, they found that selection into marriage was quite large, indicating that entrance into marriage was not largely a matter of chance or luck as Sampson and Laub contended. When selection was controlled, marriage did have a weak but significant inverse effect on crime into early adulthood. They also found, however, that while marriage inhibited offending for males, it had no such crime reduction effect for women.

Empirical evidence of marriage and offending also comes from other nations. Using growth curve models for a sample of nearly 5000 Dutch men and women born into three different birth cohorts, Bersani et al. [9] found that being married was associated with a 35 % reduction on the odds of conviction—a marriage effect of comparable magnitude to other studies. However, they too reported that the crime inhibiting effect of being in a marriage was nearly twice as high for males compared with females (though still significant). The Bersani et al. results solidify the existence of a marriage effect and its existence in a non-US sample during a different historical period than that covered by the Glueck data. Van Schellen et al. [78] utilized the same Dutch data as Bersani and colleagues and constructed criminal histories for nearly 5000 persons from 12 years old to a maximum age of 70. For men, they found that marriage was related to a reduction in offending over the life course, but only when the wife was a non-offender. For women, marriage was associated with desistance no matter what the criminal history of the husband. The importance of the character of the spouse was revealed in the fact that even an unstable marriage to a criminal spouse had a crime reduction effect on long-term offending.

In contrast, research using more contemporary samples of US offenders emerging into adulthood does not support the turning point hypothesis that marriage is associated with desistance from crime. In an analysis of offending over a 13-year period in a sample of male and female adolescents who had been incarcerated in 1982, Giordano et al. [27] found no relationship between attachments to a spouse or to children and subsequent adult criminal offending. Their follow-up study with the same sample reported in 2007 also failed to find a relationship between marriage and desistance from crime. Evidence that the relationship between marriage and desistance from crime may be spurious rather than causal can be found in a study of Norwegian men by Lyngstad and Skardhamar [52]. Using official data to estimate the propensity of offending both 5 years before and after marriage, they report that there was a large decrease in crime before the marriage took place (a finding more consistent with identity theory).

The extant literature regarding the relationship between employment and desistance are even more equivocal. In their 1993 analysis of the Glueck boys up to age 32, Sampson and Laub [66] consistently found that stable employment was related to reduced offending and that a longer period of incarceration as a juvenile was indirectly associated with persistence in crime in large part by reducing the chance that a youth would land a secure job. In the 1993 volume, marriage and stable employment were the twin pillars of their age-graded informal theory of social control. By 2003, however, with the follow-up arrest data extended to age 70, support for a desistance effect for employment was more tepid: “A second surprise was that although all three men we profiled displayed remarkable employment stability in light of their childhood and adolescent backgrounds, none of them pointed to work as a major turning point in his life. This suggests that stable work may not trigger a change in an antisocial trajectory in the way that marriage or serving in the military does, even though employment may play an important role in sustaining the process of desistance” (2003: 129; emphasis added).Footnote 27 Nevertheless, in their quantitative analysis, they found a lower rate of offending during months when these former offenders were employed.

Other evidence paints a more mixed picture about the effect of employment on desistance. Horney et al. [40], for example, found no consistent short-term reduction in offending during months when former prison inmates were employed—property offending was significantly higher during months when former offenders worked, but assault was reduced by about an equal measure (though not significant). In an experimental study of recidivism, Uggen [75] found that employment was related to the hazard of arrest within a sample of previously arrested offenders but only among older offenders (those at least 26 years old); employment had no effect on the arrest hazard for younger offenders. In a related study of released parolees over a 7-year period, Piquero et al. [63] found that in only one instance (violent crime for white parolees) was employment related to the risk of subsequent arrest. van der Geest et al. [77] examined the effect of employment on adult offending (ages 18 to 32) for a sample of approximately 300 serious offending Dutch youth. After identifying five groups with distinct developmental trajectories, they found that being employed for a full year resulted in a lower conviction rate compared with those who were unemployed for the same period. Moving from chronic unemployment to average employment also resulted in a reduction in conviction rates. This crime reduction effect was stronger for high-quality jobs though stability in employment was unrelated to offending. In their study with a high-risk sample of Ohio youth, Giordano et al. [27] found that job stability was unrelated to early adult criminal involvement and those with a “full respectability package,” which consisted of both being married and having stable employment, were no less likely to offend than others. In their 2007 follow-up study, occupational prestige was unrelated to desistance from crime. Other studies, that are careful with the direction of causal inference, suggest that desistance from crime precedes employment and other adult roles, findings that resonate more with Paternoster and Bushway’s identity theory than the age-graded informal social control theory of Sampson and Laub. For example, Massoglia and Uggen [57] found within a sample of conventional youth that desistance from delinquency occurred before a successful transition to adult roles like employment was made. Skardhamar and Savolainen [72] used smoothing spline regression models to identify changes in criminal offending at the time of stable employment. Contrary to the “work induces desistance” argument, but also consistent with the ITD, they found that desistance from crime came before the transition to stable work and taking on stable work did not result in a further reduction in crime. They did, however, find that for a very small group (less than 2 % of the sample), legitimate work during active crime periods led to a substantial decline in offending.

With respect to the role of emotional transformations in desisting from crime, Giordano et al. [29] provide both quantitative and qualitative evidence that emotional selves are important. Based on three waves of interviews with a sample of former Ohio delinquents who had transitioned into adulthood, they reported that having an anger identity was positively related to crime, relationship violence, and drug and alcohol use. This was so even net of demographic characteristics, adult social bonds like marital happiness, and occupational prestige. With longer follow-up offending data, they also reported that those with an anger identity were significantly less likely to be a stable desister from crime (compared with a persister) and more likely to be an intermittent offender over time rather than a desister. Much the same pattern prevailed for the emotional-self of depression. Those who were experiencing depression were more likely to report crime, relationship violence, and drug/alcohol use and less likely to be desisters from crime. It appears that both negative affect states, anger and depression, make it less likely that previous offenders will be able to use conventional turning points to their advantage and break from crime. Further, in both the angry and depressive-self models, marital status, marital happiness, and employment quality were each unrelated to desistance from crime.

While the empirical evidence regarding marriage and employment remain equivocal, recent empirical work is illuminating the importance of identity in the desistance process. Before we highlight this research, it is important to note that there is evidence in Laub and Sampson’s own follow-up of the Glueck boys that is also consistent with the ITD. For example, they refer ([49]: 142) to the case of Michael who made “a conscious decision” to enter the military because he feared what would happen if he did not: “If I’d gone back out on the corner-I’d get mixed up with the same gang that I got involved with, so I didn’t want to do that.” As Michael attests, he underwent a change of identity and preferences before entering the military. Similar processes are evident in other cases. Laub and Sampson ([49]: 142–143) related that he changed in part because of a feared self (our words, not his) and that “what motivated him in large part was the fear of losing his wife and family if he did not straighten out.” John also desisted when he “was ready and willing to take advantage of opportunities that came his way to avoid repeating what he saw as his father’s mistakes.” In each of these cases, it can be argued, desistance was propagated by a feared self that initiated an identity change.

Recent research has more clearly attributed these changes to a change in identity. LeBel et al. [51], for example, examined desistance among the 130 male property offenders from the Oxford Recidivism Study who were initially interviewed in the 1990s and were followed up some 10 years later. They found that a previous offender’s “subjective states,” which included an identity as a conventional family man, was indirectly related to long-term recidivism risk through its effect on reducing reentry problems. Opsal [59] examined the role of both employment and identity in desistance from crime within an interviewed sample of 43 female ex-offenders, with the interviews taken both immediately after the women were released from incarceration and again approximately 3 months later. To summarize the results, Opsal found that both conventional employment and identity change were important in the desistance process though it was impossible to make a clean causal inference as to whether identity change or jobs came first. There was clear evidence from the narratives, however, that the process of desistance involved the ideas of a feared and possible self that are central to the identity theory. One of Opsal’s women ([59]: 388) reflects this: “So, I got another chance and I’m gonna do it this time, because I want to change. I want to go home and be with my kids. I want to live a drug-free life. I want to be able to be an abiding citizen and do what I need to do and not always be in trouble and be bad-ass. That is not me.” The female offenders in Opsal’s sample illustrate the difficulties that serious offenders have upon reentry. All were unable to find the kinds of stable good-paying jobs that the Glueck males enjoyed during a time of economic prosperity. Like virtually all offenders released from prison today, when these women found jobs, they were generally in the food service and janitorial sector, paying at the minimum wage.

Based upon intensive in-person interviews, Stevens [73] explored the relationship between identity and offender rehabilitation in three English prison-based therapeutic communities (TCs). She reported that an important component of moving toward desistance was the creation, via participation in the therapeutic community, of a better “possible self.” Through participation in TC events, serious offenders learned that they had a stock of valuable and worthwhile skills which led them to think better of themselves and that a life without crime in the future was achievable. The connection between identity change toward a pro-social possible self and quitting crime was vividly expressed by one member of the TC: “Before [TC], I was kind of lost and really broken… I’ve changed so much. I honestly don’t believe I will ever offend again because I’m not that person now. I’ve found a better person here” ([73]: 540; see also [44]). Breen [12] conducted a mixed-methods study of 27 pregnant and parenting women with extensive criminal records and reported results consistent with the identity theory. First, many of the women initiated desistance when they became dissatisfied with their life of crime, particularly important was the feared self with respect to their children. For example, one woman (Jennifer) stated that she started to go straight: “After almost getting arrested and the cops telling me that if I get arrested I’m gonna be having this child in jail, and then, my child would have been take away”. Jennifer also indicated that she had constructed a positive “positive self” and a roadmap to get there: “… now I realized I have to finish my schooling, I have to get a job”. Sharpe [70] conducted in-depth interviews with 19 mothers with a non-trivial history of criminal offending. These women revealed how difficult it was for serious offenders to link with good, well-paying jobs and to a person they expressed in vivid terms the stigma they felt they faced in turning their lives around. For many of these women, it was the expectation of the feared self—having their children taken from them by the state—that started them down the path of desistance. Breen [12]: 66–67) also reported that these desisting women changed their preferences as well as their identities and adopted more pro-social values.

Finally, using longitudinal data (the Rutgers Health and Human Development Project) and growth-curve models, Rocque et al.[65]) found that even net of a cluster of control variables, a favorable improvement in one’s identity over time was related to a decline in offending. In sum, while there has yet to be a definitive study of identity change over time and desistance, these empirical efforts support the idea that one’s identity is an important component of the desistance process.

Conclusion

Theoretical work in desistance from crime was greatly advanced by the important work of Sampson and Laub [66]; [49]), and their idea that pro-social “turning points” like jobs and marriages are critical in the desistance process. Subsequent to this, Giordano and colleagues [27, 29] made an equally important contribution by highlighting both the importance of the up front work that offenders must first engage in for turning points (hooks in their parlance) to be effective in bringing about desistance. Although both theories have provided vital insight into how desistance occurs, there are critical unresolved theoretical issues as well as equivocal empirical support. Among the nagging theoretical issues for the age-graded theory is the following question, “If pro-social turning points or conventional hooks do not come randomly, then what is the mechanism by which they arrive for criminal offenders, and how can offenders now take advantage of these conventional opportunities?” Another question relates to historical context, “Although marriage to a conventional partner and a good-paying job were available to the Glueck boys, is the age-graded theory relevant for today’s offenders, particularly the majority of offenders released today, who have few educational or employment opportunities, have substance abuse issues, and face other collateral consequences of having a criminal records. Paternoster and Bushway’s [61] ITD is an attempt to address these limitations, and in so doing, bring into prominence the role of human agency in the desistance process.

In this paper, we further develop this theory, with the explicit goal of further highlighting the unique elements of the ITD. We do this because we believe that the identity theory of desistance is a viable alternative perspective on desistance from crime. It is founded on an “internalist” model of rational choice wherein preferences and behaviors are shaped by actor’s strong reasons [11] or their ultimate concerns [6], which include one’s self-identity. This internalist model, which is also consistent with Weberian action theory, relies on the very individualistic mental processes that Giordano et al. (and Sampson and Laub) explicitly reject. The relationships among purpose or strong reasons, human agency, and one’s identity is nicely captured by the philosopher Hollis [37]: 101) who argued that a person “… acts freely, only if he has good reasons for what he does (and no better reasons for doing something else). He has good reasons, only if he acts in his ultimate interests. His ultimate interests derive from what he essentially is” (emphasis added).

An important contribution of the identity theory, then, is that it hypothesizes that human agency is critically involved in the desistance process, as well as individual cognitive processes. Specially, it insists that unlike both the age-graded and cognitive transformation theories, identity change must come first and it initiates subsequent changes in preferences, social networks, and behavior. The ITD argues that offenders must first change their self-identity (who they are and want to be) before pro-social opportunities arrive and can successfully be taken advantage of or sustained. This theory challenges Sampson and Laub’s and Giordano et al.’s notion that internal change is not necessary and that turning points such as romantic partners appear by chance. In the identity theory, desistance is viewed as intentional self-change [43].

Identity theory’s invocation of cognitive processes as integral to desistance also harmonizes well with other cognitive theories. A substantial literature in behavioral economics by Kahneman [42] and Frederick [25] have linked good decision making with pro-social outcomes, and Paternoster and Pogarsky [62] have found that those who are more cognitive and less intuitive in their decision making are at significantly lower risk of criminal offending as well as more likely to develop conventional social bonds. Clearly, then, identity change is not the only cognitive process involved in offenders desisting from crime, and the development of the identity theory will benefit in the future from a convergence with other cognitive theories of decision making.

Finally, unlike Giordano et al. who limit the workings of human agency to only those at the mid-range of disadvantage, Paternoster/Bushway’s theory views human agency as essential for all desistance, but particularly among the highly disadvantaged. Without resources, it would appear that willful, purposive actions would be even more critical. This last point is particularly important since reentry in America in the twenty-first century is coupled with all of the deficits noted above, a life that is unlikely to involve the kinds of employment and marriage opportunities that were available to the Glueck boys. Under these conditions, offenders wishing to quit crime must have strong reasons anchored in their belief that their future self will be one relatively free from crime and drugs.

This assertion, which we have highlighted more clearly in this paper, represents a clear leverage point for empirical tests of ITD, particularly tests that seek to compare ITD with other versions of identity theory. Objectively, this strong claim also makes ITD falsifiable, an important feature of good theory that should aid empirical tests and future theoretical development. However, we believe strongly that the development of ITD is more than just a job program for academic criminologists. We have entered an interesting time where policymakers are increasingly interested in topics such as reentry and alternatives to incarceration (National Research [18]). As such, policymakers are eager to find ways to not (just) punish but also help offenders exit criminal activity. But, not surprisingly, there is no clear consensus about the nature of effective programs. In this vacuum, theories of desistance play an important role because they point toward key features of programs. For example, policies built around Sampson and Laub’s theory of social control might focus primarily on providing active offenders with pro-social experiences such as employment. However, ITD predicts that these opportunities will be largely wasted on individuals who have not yet reached the point where they have decided to adopt a new identity. The idea of cognitively preparing offenders to make good use of conventional opportunities is an integral component of successful rehabilitation programs such as Andrews and Bonta’s risk-need-responsivity model [4] and Ward et al. [79] Good Lives Model. Further, MacKenzie’s [53] review of correctional treatment effectiveness has highlighted the success that cognitive-behavioral programs have had in reducing recidivism. The identity theory of desistance is perfectly compatible with treatment programs directed at other cognitive processes like anger management and developing better decision making and problem solving skills. Policies informed by ITD might require proactive action on the part of individuals prior to providing employment. Such selection would avoid wasting resources with people who are not yet ready to desist (and will not benefit from employment) and avoid creating perverse incentives for those who are still actively involved in crime.

One implication of ITD is that active criminals already have an identity. As a result, change will not come about simply as the result of a small change in the nature of the incentives. Instead, we argue that an individual needs to make a fundamental shift in their approach to life. This will not be easy, nor will it occur simply because new opportunities are presented. It is also not clear to us how to instigate the choice to try out new identities. However, ITD does suggest that opportunities offered to individuals who have not chosen to try a different way might actually confuse the matter by delaying the encounter with the feared self. At the same time, finding ways to respond positively to people who are in fact trying to change might both encourage further change and allocate scarce resources efficiently.