Crime, Consumption, and Choice: On the Interchangeability of Licit and Illicit Income

Objectives
 
 We examine the rational assumption of the interchangeability of legal and illegal monies. Drawing from economics, behavioral economics, and sociology we answer two main research questions: (1) Do offenders perceive money earned across various income-generating activities (legal vs. illegal) in the same way? (2) How do consumption patterns (spending and saving) differ across various forms of income-generating activities? 
 
 Methods
 
 We use an a priori mixed methods approach with two interrelated studies; a quantitative survey of incarcerated offenders (N  =  58) and a qualitative study of semi-structured interviews from four separate previous research projects (N  =  107). 
 
 Results
 
 We find evidence for the existence of differential consumption patterns based on quantitative and qualitative data from both incarcerated and active offenders regarding their patterns of spending legal and illegal money. 
 
 Conclusions
 
 Our findings have implications for choice theories of crime, for public policy approaches to poverty, and crime prevention interventions.


Introduction
The intrinsic meaning of money, beyond simply its transactional value, has long been a key topic of scholarship in fields as diverse as sociology, economic sociology, and behavioral economics. Scholars have questioned its social importance and how it structures human relations, often drawing distinctions between money and the value assigned to it (Zelizer 2017). Conversely, criminological inquiry into money has been largely focused on establishing its role as a primary driver of income-generating crimes (Baumer and Gustafson 2007;Coleman 1992;Sullivan 1989). For instance, in their study of approximately 300 convicted burglars from England, Bennet and Wright (1984) noted that most burglars tended to cite the need for money as a primary motivation in the decision to engage in the crime (see also, Cromwell et al. 1991;Hagan and McCarthy 1998;Shover 1990). Recent quantitative research suggests that, across multiple samples, the frequency of engaging in income-generating crimes is highly responsive to illicit monetary incentives (Nguyen et al. 2021). Qualitative accounts from active offenders reinforce this point (e.g., Baron 2004;Fader 2016;Tunnell 2006). Despite these findings, the monetary gain has mostly been relegated to a motivational assumption in criminology (see Brezina and Topalli 2012), but not as a factor worth specifically theorizing about, as witnessed in sociology (e.g., Zelizer 2017Zelizer , 1993Zelizer , 1989 and economics (e.g., Ingham 2001;Thaler 1987). This is unfortunate because the study of money has important implications for criminological discourse. First, understanding how money is used for future consumption, (i.e., spending on goods, services, and necessities) is a key omission in choice theories of crime (Nagin 2007). Traditional economic theory dating back over half a century assumes individuals are rational and view money gained from any source as equivalent for the purposes of future consumption (Friedman 1957;Modigliani and Brumberg 1954). Yet, the traditional notions of rational consumption have been challenged by multiple intellectual traditions. In fact, there is considerable evidence suggesting individuals might not treat income from different sources as fungible, or completely interchangeable (Shefrin and Thaler 1988). Specifically, evidence from behavioral economics and sociology 1 challenges the rational assumption of the fungibility of legal and illegal monies, notably in ways specifically related both to the nature (objective characteristics) and source (criminal origins) of illegal funds in ways that differentiate them from legal funds. If the assumption of fungibility is unwarranted, this has clear implications for theoretical studies of offender decision-making. Primarily, this would mean that consumption choices are differently impacted by the nature and source of income, which would, in turn, differentially influence the decision to engage in additional income-generating activity. Such a reciprocal relationship would thereby assume consumption as one of the key decision-making points in offender decision-making.
Second, the putative (non)interchangeability of money portends a substantial impact on policy, specifically those where the program strategy is designed to offset criminal gains with legal income. For instance, the logic of supported work and income maintenance programs is predicated on the notion that improving financial well-being through providing or enhancing legal opportunities will help to reduce crime by discouraging an individual's decision to offend (Berk, Lenihan, and Rossi 1980). Yet the empirical support for the success of these programs remains mixed at best (Visher et al. 2005), especially in the longer term (Cook et al. 2015). One possible explanation for the limited success of such programs in reducing recidivism is a potentially flawed implicit assumption that legal sources of money, gained through "legitimate" employment and income payments, will offset the need to earn money through illegal activity. In other words, a lack of attention to the intrinsic meanings of different monies could constitute a "theory failure" resulting in the overall ineffectiveness of such programs (Berman and Fox 2016;Kraemer et al. 2001).
Finally, prior research suggests that offenders' consumption of nonstandard or non-necessity commodities such as luxury items, or illicit expenditures such as drugs, alcohol, or gambling drive illegal income-generating activities over legal income-generating activities (Felson et al. 2019). Offenders instigate a great deal of economic activity in impoverished neighborhoods, where their participation in drug selling, theft, trading in stolen items, and offering of illegal services undergird black and gray market economies (see Goel and Nelson 2016;Restrepo-Echavarria 2014). In this way, understanding consumption and the fungibility of different sources of income has implications for addressing conspicuous consumption, which contributes to the cycle of poverty and crime (Jacobs and Wright 1999).
This paper provides evidence for the existence of differential consumption patterns among offenders based on quantitative and qualitative data from incarcerated and active offenders regarding their patterns of spending legal and illegal money. We pose two main research questions: (1) Do offenders perceive money earned across various income-generating activities (legal vs. illegal) in the same way? (2) How do consumption patterns (spending and saving) differ across various forms of income-generating activities? We begin by discussing theoretical expectations for the fungibility of different monies and rational consumption, and the theoretical critiques of the rational model put forth by economics, behavioral economics, and sociology. We then present our findings and implications for theoretical studies of choice theories of crime (Nagin 2007;Pogarsky et al. 2018) and for public policy and poverty and crime prevention interventions.

Rational Choice and Consumption
Despite the influence of multidisciplinary perspectives, rational choice theories are the predominate theoretical framework guiding both the quantitative and qualitative literatures on offender decision-making (Bernasco, Elffers, and van Gelder 2017). Quantitative analyses of offending decisions, largely focused on how individuals perceive and weigh risks and costs (e.g., Loughran et al. 2016) and to a lesser extent rewards to offending (Nguyen 2020), have drawn heavily on conventional economic choice models in line with Becker's (1968) original conceptualization of rational choice. His version of the theory explicitly proposes that offending decisions are the product of the marginal benefits of a potential action exceeding its marginal costs. In comparison, qualitative work on offender decision-making typically draws heavily on Clarke's (1986, 2002) more catholic conceptualization of rational choice theory, which allows for the influence of situational factors and the adaptive nature of would-be offenders as they contemplate potential offending choices (e.g., Jacobs et al. 2003;Rossmo and Summers 2019;Wright, Logie, and Decker 1995). 2 A commonality among quantitative and qualitative work on offender decisionmaking is the prominence of illegal monetary incentives as motivators of illegal income-generating activities. Importantly, while both variants of the theory agree on the importance of illicit financial incentives as inducements for illegal activities, they are effectively silent in terms of differentiating illicit proceeds from legal monies for purposes of the eventual consumption of future goods.
That said, formal theories of consumption are embedded in extensions of traditional economic rational choice theory congruent with Becker's perspective (e.g., Friedman 1957;Modigliani and Brumberg 1954). As in the criminological version of neoclassic rational choice, one goal of an individual facing a consumption choice is to maximize utility or benefits. The traditional approach economists use to describe consumption is the life cycle hypothesis originally proposed by Modigliani and Brumberg (1954). This theory states that to maximize utility, present consumption should depend on current wealth and expected future earnings, such that lifetime consumption should equal lifetime income. Though the life cycle model is still widely used (Deaton 2005), scholarship in economics, as well as behavioral economics and sociology have challenged key assumptions of the rational notions of consumption, reframing it to consider both behavioral and structural factors which may also affect how individuals choose to spend their money (e.g., Duesenberry 1949;Shefrin and Thaler 1988;Thaler 2015).
Most importantly, rational models of consumption like the life cycle hypothesis are reliant on assumptions about the interchangeability, or fungibility, of money, which for the purpose of future spending, specifies there should be no difference between a dollar in your pocket obtained illegally versus one obtained legally. 3 Importantly, if illegal funds tend to be dedicated to specific categories of spending (e.g., partying, drugs), then the penchant to continue these activities can be driver of criminal behavior. This suggests that extant versions of rational models of consumption, which equate to the meaning of different forms of income, are incomplete. This has potential implications not only for revising and developing choice theories of crime but also for the effectiveness of policies-such as income maintenance and employment programs-derived from the validity of extant theories of choices and incentives.

Predicted Departures from Rational Consumption
Departures from the strict assumptions of rational consumption have long been the subject of theorizing and empirical inquiry. For instance, early on Duesenberry (1949), an economist proposed the relative income hypothesis arguing consumption is governed by an individual's income relative to both others (rather than to an absolute ideal) and one's own previous level of consumption. This suggests that once levels of spending on certain types of commodities are established by an individual, departures from these levels will become more difficult in future periods. 4 However, the most persuasive challenge to the rational assumption of the fungibility of legal and illegal monies comes from research in behavioral economics and sociology that has focused specifically on both the nature and source of illegal funds that may differentiate them from legal monies. According to Coulborn's (2013), money can be "concrete" or "abstract." First, what we term the nature of money is related to Coulborn's notion of concrete money in that the tangible characteristics of illegal income are different from those of legal income because it is volatile and inconsistent, payoffs are frequent but in smaller sums, and largely cash-based (Naylor 2004;Nguyen et al. 2021). Second, the source of money drives its meaning and abstract value. Illegal money is gained through the use of force or fraud. It often carries negative connotations and is referred to as "dirty money," "wages of crime," and "wages of sin" (see MacCoun and Reuter 1992;Naylor 2004), which can be related to the value that offenders assign to it compared to legal money. Nature and source have analogs to relevant challenges to the assumptions of rational consumption from both behavioral economics and sociology, respectively. Here, we present nature and source as characteristics of money in separate subsections below; however we underscore that these attributes of money are often interdependent and overlap in how individuals value and consume money.

The Nature of Illegal Money
One of the strongest sets of challenges to the neoclassic economic framework of consumption is offered by behavioral economics, developed primarily to reconcile observed deviations from rational predictions in empirical studies of the life cycle hypothesis (Thaler 1987). 5 Broadly, Simon (1955) introduced the notion of "bounded rationality" which relaxes one or more of the assumptions of rational expectations. Bounded theories assume there are fixed sets of alternatives and uncertainty in the true probability of outcomes. Rather than striving for utility maximization, decision-makers are limited in cognitive capacity and undergo a satisficing strategy. More specifically, in response to the life cycle hypothesis, Shefrin and Thaler (1988) proposed the behavioral life-cycle hypothesis, which offers several modifications to the standard theory. First, challenging the rational assumption that money acquired by different means is fungible, and thus holds no intrinsic meaning beyond purchasing power, Thaler (1987Thaler ( , 2015 instead contends that individuals directly violate the principle of fungibility by cognitively establishing different pots or mental accounts. For example, people often carry large credit card debt but still keep a "money jar" to set aside money for a vacation. Second, as in earlier challenges to the standard rational model proposed by Duesenberry (1949), Shefrin and Thaler (1988) highlight the salience of the framing of choices in how resources ultimately might be consumed. 6 They point out that consumption is extremely sensitive to income earned in a given period, arguing that, "income paid in the form of a lump sum will be treated differently from regular income" (p. 610). Research also indicates that consumption behaviors are sensitive to the amount of income, where larger sums are more likely to be saved than smaller sums (Shefrin and Thaler 1988;Zagorsky 2013). There is also literature which suggests that foreknowledge of the income increase (Arkes et al. 1994) and the amount of effort invested in the income increase (Kivetz and Simonson 2002) are important considerations when individuals decide to allocate funds to certain mental accounts.
Much of what we know about the nature of illegal income comports with each of these proposed behavioral departures from standard rational choice and suggests that a conceptualization of consumption related to illegal income is in order. For instance, illegal income is often paid as a lump sum or "fast cash" (Sampson and Laub 2003). It's accrual is episodic and less predictable than money gained from legal activities, which is often scheduled and delayed in the form of a regular paycheck, not to mention traceable, taxable, and typically deposited into a bank account. In terms of decision-making, the intermittency and variable amounts of illegal money likely result in different realized incomes from period to period, which violates rational expectations of future income. This suggests that illegal income could serve as a source of variable reinforcement for offenders (see Gerhart and Rynes 2003;Yukl et al. 1972), which has implications for how they value such funds and their motivation to acquire it. From a decision-making standpoint, this likely differentiates its value from income obtained legally, which is more likely to be distributed on a regular basis and would therefore conform to patterns of constant and consistent reinforcement. Consequently, it is easy to see how the more transitory nature of fast cash fails to comport with rational beliefs associated with the nature of legal income (Loughran et al. 2013).

The Source of Illegal Money
The origin of money likely engenders different intrinsic meanings, leading to further differences in how legal and illegal income are assessed by offenders. This contrasts with the rational consumption model that suggests money has an objective value and thus, changes in an individual's wealth-regard less of the source-are treated the same. To wit, persuasive sociological literature has arisen questioning the objective value of money (Zelizer 1993). For over a century, social theorists have studied and debated it's social meaning, dating back to Simmel (1900) who differentiated between value and money and questioned the abstract constructivist nature of value and the role of money as a vehicle for making values commensurable. This thinking was formalized by Dodd (1994) who argued that structural and societal factors might shape individuals' beliefs about the very meaning of money. More contemporary scholarship (Delaney 2012;Zelizer 2017) offers insights into how we may begin to systematically categorize the consumption of offenders working in various contexts, such as those operating in impoverished neighborhoods where a great deal of instrumental predatory crime takes place. Zelizer (2017) outlines a framework as to how and why individuals-by discriminating between different categories of money based on how it is accumulated-do not view capital as a monolithic concept. She observes for example that money earned through work is viewed differently than that which is gifted or won (Zelizer 1989(Zelizer , 2017. In doing so she rejects the notion that money is a, "… single, interchangeable, absolutely impersonal instrument …" (p. 1). Zelizer also points out that, "Unlike an 'honest dollar', 'dirty money' is stained by its ethically dubious origins." Similarly, Delaney (2012) suggests that decisions about consumption are driven by the existence of different money cultures, arguing that different kinds of work produce profound differences in how people evaluate and use money in their professional and personal lives.
Moreover, notions regarding the sociology of work align with research on emotional accounting (Levav and McGraw 2009), which posits that the affective response to a given source of income can have an impact on its consumption. For decision-making, money acquired through actions associated with negative affect can be spent on hedonistic products to engender a counteractive positive experience. Likewise, individuals may spend a portion of their ill-gotten money on strategies to reduce negative emotions such as guilt (e.g., giving to a charity; see also research on the psychology of emotional self-regulation, e.g., Bonanno 2001;Connolly et al. 2019). These consumption habits can fuel subsequent decisions regarding the frequency and type of criminal behavior.
In summary, theorizing and evidence from multiple intellectual traditions including economics, behavioral economics, and sociology have challenged the traditional rational consumption assumption of the fungibility of illegal and legal money. Instead, different themes have emerged across disciplines, namely that the nature of income (e.g., intermittent, cash, lump sums) and the source of the income (legal activities vs. illegal activities) are important determinants of consumption. These attributes of illegal funds provide important impetus to study the prospect that money acquired through different sources -in particular, legally versus illegally-are capable of driving different patterns of consumption and subsequent income-generating activities. If these challenges are valid to differentiate legal and illegal income as unique constructs that exert independent (if related) influences on an individual's utility, choice theories of crime themselves may be incomplete and should therefore be refined to accommodate such complexities (Nagin 2007).

Research Strategy
To address the questions raised above we conduct a mixed methods research project with two interrelated studies; a quantitative survey of incarcerated offenders and a qualitative, interview-based study of active offenders. First, we analyze data from a survey-based study of N = 58 incarcerated males housed at a medium security facility in Pennsylvania serving short minimum sentences for income-generating crimes (i.e., drug selling, robbery, burglary), who were soon to be released. The survey was principally designed to understand legal and illegal income-generating activities and consumption and spending patterns in the two months prior to incarceration. In addition to quantitative items about income-generating activities, the survey instrument elicited responses to one open-ended qualitative question about consumption matched to its quantitative items. Second, we analyzed interview data from active offenders (i.e., non-institutionalized and currently offending) who are demographically similar to the incarcerated sample but from St. Louis and Atlanta.
Using two data sources provides a number of advantages. The results of Study 1 provided evidence that offenders differentiate between different sources of income with congruent differences in the expenditures that follow. The data were limited, however, in their ability to illustrate the origins of these differential choices. Study 2 provided further context for the results of the first study and a clearer picture of offender consumption patterns. This enabled us to further illuminate the meanings that offenders attach to their decisions about money and income-generating activities in the fullest sense. This strategy has been recommended by Ragin, Nagel, and White (2004) as potentially "… helpful in assessing the credibility" of causal mechanisms posited by researchers, especially as a follow-up to quantitative studies (p. 15).
The criminal decision-making literature suggests that structural determinants, characteristics of decision-makers, and situational aspects of the criminal event are all factors that coalesce in the study of choice (Lindegaard and Copes 2017). Consumption patterns could also differentially be impacted in the same way, as structural and individual factors might shape individuals' beliefs about the very meaning of money. Incarcerated offenders' responses about income-generating activities and consumption are likely to be more sensitive to these structural issues, like restricted labor market opportunities, than active offenders. Active offenders who operate in various contexts, such as impoverished neighborhoods where a great deal of instrumental predatory crime takes place are likely to be more attuned to situational circumstances that are related to making and spending money. Taken together, using the two complementary data sources and assessing how they comport with each other can provide important variation in studying offender decisions and consumption patterns.

Mixed-Methods Approach
Our methodological design follows established procedures for conceptualizing and designing mixed methods research designs (see Creswell 2015;Yvonne Feilzer 2010). Such approaches can be devised a priori to establish lines of evidence for pre-established hypotheses (i.e., a deductive approach) or they may be identified as necessary when the results of a specific study (quantitative or qualitative) produce important yet incomplete results. In addition, the design may be such that quantitative results identifying causal mechanisms are followed up by a qualitative study designed to contextualize the results from the quantitative study. Alternatively, a qualitative study may identify new relationships inductively, which may then be operationalized and tested deductively in a follow-up quantitative study. In the current project, the first study, which was predominantly quantitative in nature, documented important relationships regarding the nature of consumption in a sample of incarcerated individuals. That study benefited from an important qualitative add-on, a single open-ended item at the end of the survey. The results of the survey and suggestive results of its openended item (outlined below) led us to follow-up with our second study, a qualitative analysis of preexisting and new interview-based data from active offenders. This approach has been applied previously in criminological studies of how future time discounting affects offender decision-making (Brezina et al. 2009) and the extent to which criminal self-efficacy impacts offending . Variation in data sources across different studies in a mixed methods design can produce important benefits. Including data from not only different methods but different kinds of respondents (active offenders and incarcerated offenders) allowed for the disentangling of concepts from the identity and backgrounds of respondents themselves, thereby removing the confounding effects of respondent backgrounds and allowing the overall study focus on the illumination of concepts.
In the current article, quantitative results from Study 1 revealed that offenders differentiated between different types of work with consequences for how they engaged in mental accounting to produce different consumption patterns. This was further reified by the results of that study's open-ended question, which provided some indication of the potential importance of mental accounting in differentiating spending and the extent to which it was related to different kinds of income generation. Study 2 represents a methodological extension of these efforts, relying on interviews with active offenders. We identified two forms of data for this study. First, a combination of data from three previous studies of drug robbery, carjacking, and burglary. Data from these studies were included in the current study if the constituent interviews involved discussions of income generation, consumption, and/or mental accounting (70 percent of these datasets included such discussions). Data from these studies represent a posteriori identification of themes identified in the first study which had been produced through discussions of offenders not designed to focus on mental accounting or consumption. Their value lies in the fact that they were not subject to researcher demand effects. We also collected new data with the same kinds of populations where discussions were more consciously focused on such topics, and as such, provided more rich contextualization of the results of Study 1 and greater detail than the data previously mentioned. Importantly, this study also identified that 69 percent of respondents distinguished between different kinds of consumption and mental accounting. These considerations brought Study 1 and Study 2 together in the mixed methods study design illustrated in Figure 1.

Study 1: Survey of Incarcerated Persons
The purpose of Study 1 was to provide a prospective study specifically focused on offenders' income-generating activities and consumption decisions. We recruited participants through several different strategies, including group recruitment sessions whereby the researcher explained the nature of the study and asked for volunteers, having the unit manager post a sign-up sheet in the unit with a brief description of the survey, and through word of mouth among the residents. Each survey was administered by a member of the research team in a one-on-one setting to ensure confidentiality and took approximately 1-1.5 h. Interviewers read the consent form to each participant, assured confidentiality, and stated that the respondent could stop the interview at any time. The surveys were coded and entered into and analyzed with the statistical software package STATA (StataCorp 2017).

Measures
Measures tap into three factors: participation in legal, informal, and illegal employment, attitudes towards legal employment, and consumption patterns. To contextualize the respondent's life in the two months prior to incarceration, each incarcerated person was asked about a number of life circumstances during those two months, the same time frame as they were asked to report about their money-generating activities.
Income Generating Activities. Legal Work: was defined for the respondent as "paid work that was reported to the government." First, the respondent was asked if they ever had a legal job and if the answer was affirmative, they were asked if they had a legal job in the two months prior to incarceration. In addition to reporting participation, respondents were asked to report the number of weeks worked and weekly earnings from legal work.
Informal Work: was defined as "activities where you were paid 'off the books' or 'under-the-table', excluding activities that are against the law. These are jobs where you are paid cash and do not report the income on tax forms." Similar to the legal work questions, respondents were asked if they ever had an informal job and if they answered in the affirmative, a number of questions followed regarding their participation in informal work in the two months prior to incarceration.
Illegal Work: was defined as "making any money from activities that are against the law." The questions for illegal work followed the same pattern as legal work and informal work.
Spending Money. Expenses: Respondents were asked to think about the two months when they were last on the street, "What were the five biggest things that you spent your money on (i.e., rent, food, drugs, cell phone)?
The answers were open-ended so the respondents were free to list their top five expenses. After listing their five top expenses, they was asked to estimate the cost of each expense." Disbursement: To understand how respondents paid for their expenses, they were asked to indicate the percentage of each of the five top expenses paid by legal, informal, and/or illegal income. Among all the open-ended responses, we collapsed these into five major categories: essential housing, other essential living, family gifts, drugs and alcohol, and extras (e.g., going to a bar, women, designer shoes, etc.).
Demographic variables: Respondents were asked for a number of demographic information including age, race/ethnicity, marital status, number of children, and highest level of education (see Table 1).
In addition to the quantitative questions above, respondents were asked an open-ended question: "In general, do you think the money that you made legally is used for different things than the money you make illegally? Tell us a little about how you think about your expenses and spending."

Study 1-Results
In relation to the question, "In general, do you think the money that you made legally is used for different things than the money you make illegally? Eighty-five percent of respondents reported that they viewed money from legal sources differently than money from criminal sources. When probed with an open-ended question," "Tell us a little about how you think about your expenses and spending." Answers to this question provide some context for the respondents who answered affirmatively to the question that asked whether the respondent thinks legal money is different than illegal money.
First, multiple responses suggested that the source of the money matters. One respondent who worked in the construction industry for under the table cash and engaged in burglary talked about the positive nature of legal earnings: I spent money on positive things when I was doing good, used it on stuff I wanted and needed. If I was doing both, then no. But when just had a legal job, I had more respect for what I spent money on. (James, burglar) A 40-year-old man, incarcerated on drug manufacturing charges and who illegally sold heroin, also worked informally as a broker to sell cars, Legal money-you're gonna try to save it. You work hard for it. The illegal money -you can blow it fast because you can get it right back. (Kayo, heroin dealer) Importantly, participants provided statements that legal and illegal sources of money were consumed differently. For example, there were strong sentiments about the party lifestyle that illegal sources of income for which were deemed more appropriate. For example, a 24-year-old man who was incarcerated for carrying a firearm without a license, reported that they worked as a contractor and remodeled homes for under-the-table cash and also sold various types of illicit drugs noted, When I work hard for my money, I don't spend it on the drugs and the women. But when I spend the money from selling drugs, I know I can get it back fast so it goes for things I don't really need. (Lenny, drug dealer) Similarly, a 44-year-old respondent who was also incarcerated for selling illicit drugs and worked at the Home Depot the two months prior to incarceration recalled, The illegal cash I spend quicker and on things I didn't need like going out and eating good food. With the legal money, have to budget that more and spend it on things you need like paying bills. (Jack, drug dealer) Finally, a 39-year-old respondent who worked as a land surveyor informally and illegally made money from burglary and drug sales indicated that, The money you make legally usually goes for things that are legal like rent and bills. The illegal money that I got went to drugs, there'd be no other reason to do it-just to feed addiction. (Sneaks, burglar)

Budgeting and Spending Questions
Individuals were also asked to report more concretely on how they earned money (i.e., legally, illegally, informally), as well as their most important expenditures and the source of money used to pay for the expenses. Table 2 addresses participation in each of the income-generating activities and shows that all respondents were involved in some type of income-generating activity during the two months prior to incarceration, with important trends in their income-generating activities demonstrated. First, there is considerable variation in the types and combinations of income-generating activities the respondents engaged in during the two months prior to their incarceration. The first column shows that illegal income-generating activities (66 percent) were most common, followed by informal (57 percent) and legal work (48 percent). The percentages in the bottom row show that most of the sample (60 percent) doubled up during this period. Though approximately half the sample was involved in legal work during this time, only 16 percent exclusively earned money legally, which is important because it suggests that having legal income was not sufficient to offset engaging in crime. Second, involvement in legal and illegal income-generating activities was more consistent than involvement in informal work. Finally, doubling up appears to have paid off monetarily for the respondents in this sample. Respondents that doubled up made substantially more than respondents that did not. Table 3 displays the five top categories of expenses, which include essential housing, other essential living, family gifts, drugs and alcohol, and extras. The second column shows the number of respondents who reported the expense as one of their top five expenses. For example, 68 percent of respondents stated that housing was one of their top five expenses. Table 3 also shows the percentage of respondents who report using a particular form of income for each expense. For example, among individuals who reported having a legal job in the two months prior to incarceration, 82 percent of them used money from their legal job to pay for "housing" whereas only 65 percent of respondents who earned money from illegal work reported spending their money from illegal work on "housing." "Other essential living" expenses and "drugs and alcohol" appear to be paid for by all three sources of income. For the less essential categories of "family gifts" and "extras," a higher proportion of respondents reported using money from illegal work than money from legal work to pay for the expenses. Figure 2 presents the five categories of expenses and the income source used in each category. First, across all categories of expenses, the vast majority of respondents reported using only one source (legal only, informal only, or illegal only) rather than reporting using two or more sources of income (i.e., legal and illegal). This suggests that for most respondents,   monies are not fungible and that different sources of income are maintained in separate mental accounts. Second, when we compare across the five categories of expenses, the predominant source of income used to pay for the expenses is different. For instance, 37 percent of respondents used only legal funds to pay for essential housing. Interestingly, 23 percent reported using only their illegal funds to pay for housing. Most of these respondents reported having only illegal income in the two months prior to incarceration. The "other essential" category, which included bills and food, was more mixed with approximately similar proportions using legal, informal, and illegal sources of income to pay for bills. Expenses like cell phone bills and grocery bills are less consistent in cost and due dates compared to housing costs. It is likely that grocery bills are paid in cash and are therefore more amenable to the nature of illegal sources of income as opposed to other bills like cell phone bills. The three categories of discretionary spending "family gifts," "substances," and "extras" were predominately paid for by illegal funds. A good portion of these discretionary categories were also paid for by funds from informal work.

Study 2: Interview-Based Study of Active Offenders
In the second study, we analyze semi-structured interviews from three separate previous research projects Wright 2000, 2003) 7 as well as new interviews collected in 2019 (see Table 4). Across these projects, the interviews contained data on how offenders made money and viewed money. In those cases where interviews covered concepts relevant to the current study and where respondents were over 18 years old, we pulled and re-coded them.
Within the final sub-sample, ages ranged from 18 to 46 years old, averaging 28 years. The final sample yielded 107 participants, including one White offender, and six Latino offenders. The remainder were all African American (11 of whom were Black females). 8 The preferred criminal occupations were drug dealing, robbery, car theft, burglary, larceny, carjacking, or a combination of offense types, which is consistent with the incomegenerating crimes represented in Study 1.
We applied the categories identified in the Study 1 survey to outline broad domains for coding and analyzing the data from these projects and then used these data to break main domains into subdomains and categorizations (see coding structure, Figure 3). These include forms of income generation (legal, informal, and illegal) and consumption behaviors (expenditures vs. disbursements), as well as the nature and source of income on criminality and consumption, to see how such factors influenced offenders' decision-making processes.
All three legacy datasets as well as the interviews collected in 2019 followed a semi-structured protocol with pre-established questions drawing broadly from processual models of offending that break down offenses into compartmentalized stages and anchored to a specific offending event(s) (e.g., Decker 1995, 1997). Generally, questions probed the situational aspects of these offenses, such as the way in which they were carried out and what the offender felt and did following the offense, with specific attention paid to the way they spent ill-gotten gains. This helped establish the typicality of both the participants and their offenses, which correspond with previous research (Copes 2003;Jacobs and Wright 1999;Topalli et al. 2015) and Study 1. Income-generating activities, money, and consumption were commonly discussed in the archived data we coded, and further addressed in the original interviews we collected in 2019 (36 interviews). A particular advantage of these data was their basis in specified criminal events using the processual framework (see Topalli et al. 2020). Such descriptions of offending events served as useful interrogatory anchors for subsequent discussions about money and provided a useful way to discriminate relevant interviews from those that were less conceptually informative. Their adherence to a common methodological approach provided further consistency of content across all the interviews as well, making the combination of data logical and supporting the notion that these studies featured consistent themes regarding the making and spending of money.

Transcription and Data Analysis
All projects followed established protocols commonly employed in active offender research (see Topalli et al. 2020). Interviews for Study 2 were transcribed verbatim by a third-party transcriptionist or research assistant. Verbatim transcriptions included the use of filler phrases, grammatical errors, profanity, and the phonetic spelling of commonly used words. A second researcher would then audit the completed transcripts to ensure their accuracy and to allow for collaborative reconciliation of words or phrases deemed inaudible by other transcribers or address any discrepancies between transcribers. All audios underwent a final round of auditing before transitioning to the coding phase. Audited transcripts were coded, re-coded, and analyzed by hand (Jacobs et al. 2000(Jacobs et al. , 2003Brezina and Topalli 2012) or using NVIVO 12 (for 2019 data). Each transcript underwent a manual coding process, producing a hierarchical model of interrelated concepts with themes and sub-themes (see Figure 3).

Study 2-Results
In line with the open-ended question from the survey, we analyzed the interview data to determine how offenders differentiated forms of income and consumption. Most of them identified the three categories of money identified in Study 1 (i.e., legal, informal, and illegal) and differentiated between forms of consumption (essential, discretionary, and extras). In line with Study 1, we found that there was a differentiation between different sources of income. For example, Butter who partakes in drug robberies noted, The paycheck is for the house stuff. Food, clothes. The robbin' money is for this kinda shit [shows a gold chain and a ring] and to get me a tattoo or go to the club. Or I can buy some nice dope. Those two different monies, you know what I'm saying? There's bill money and havin' fun money. (Butter, drug robber) For respondents who partied hard and were caught up in the cycles of offending and conspicuous consumption, the difference between legal and illegal money became obvious. Ghost, also a drug robber, stated, Similar to Study 1, respondents explained how they differentiated the proceeds of different income-generating activities, applying strict rules to how money from different sources was to be consumed. Nukie-who engaged in drug dealing as well as carjacking vehicles to sell their parts on the black market-stated, I got like, two jobs, OK. The first is the gig I got working at my cousin's autobody shop, OK? He give me a paycheck and that money, I just put that up. I put that up with my girlfriend's money she get from working at the hospital. I let her handle that money. She pay the electricity and the cable and shit. Rent. Diapers. Toys. Basketballs and shit. Groceries. All that shit. The other job, that's me doing all that shit I told you about, carjacking, car stealing. Selling parts, you know … hub caps, spinners. And then I flip that and buy and sell some dope. That money is for like, doing more buying and selling. And then I spend that at like, the club or to go to a hotel and party or whatever. Sometimes, we take it to the mall and just you know blow the shit.
In those cases where respondents cohabitated with others, particularly women (significant others, mothers, etc.) they clearly articulated that legitimate money was to be used on legitimate expenses, and necessities. Illegitimate money was spent faster and on partying. Quick, who sold drugs and engaged in burglary, commented, It's like this. I keep my girl in her nice clothes and jewelry. I got kids in a few houses. I buy all the toys and video games. I pay for Six Flags and shit. She handles the bills. She handles groceries. She can cook a meal. That's on her. I don't spend my money on milk and bread. I pay for restaurants and shit, the nice stuff. (Quick, drug dealer, burglar).
Cee-Low's description similarly reveals the nuanced and dynamic nature of how offenders view the various natures and sources of money. Rather than treating them as static concepts, we note that offenders adopt various distinctions and meanings based on situational internal and external pressures. These distinctions have implications for the manner in which offenders engage in mental accounting and consumption, suggesting that considering money as a unitary concept is an inefficacious conceptualization of money for understanding offender behavior.
Well, I started robbin' and slingin' when I was younger. I was skipping school and making money and my parents didn't know about any of it until I got caught, and they went nuts man. They were pissed. But that money was good, so when I got out and went back to school I just kept doing what I was doing but I was more careful. So, I got like, an after-school job and that way, if I had new shoes or something, I just tell "em it was from my job. But I had shit stacked up in my closets. Then after I left the house, I just kept on. Keep a job and nobody hassles you. Police stop you, Where'd you get this money! You can show them a paystub." (Cee-Low, drug dealer, robber, carjacker) In sum, Study 1 and Study 2 provide complementary perspectives on the number of key ideas. First, we find that engaging in multiple incomegenerating activities is commonplace. Second, the delineation between the sources of income become apparent when respondents were queried about their consumption habits. Different sources of money seemed to hold different values to the respondents. Thus thirdly, different sources of money were spent differently. Legal income was often earmarked for essentials like housing whereas illicit income would be used towards extras, like partying. Our respondents also provided support for the idea that consumption fuels subsequent decisions about the type and frequency of income-generating activity.

Discussion
Relating to the broader theme of context and choice, which unifies the articles in this special issue, the present study is concerned with the context in which monies are acquired and how this ultimately relates to choices about the consumption of these monies and the subsequent link to the criminal choice. Despite generating interest across multiple disciplines, the study of the meaning of money in criminological discourse is curiously absent, even though money is the primary motivator for acquisitive crimes and is the backbone of the illicit economy, which supports and is supported by crime. As such, our study was motivated by two primary research questions-do individuals view legal and illegal sources of income differently, and are patterns of consumption motivated by different sources of income? To answer these questions, we drew on evidence from behavioral economics and sociology that challenges the rational assumption of the fungibility of money. Our findings, gained through both quantitative and qualitative measures observed in two unique datasets, provide strong evidence to question the assumption that funds earned legally are conceptualized as fungible with those earned illegally.
Our conclusions regarding the offender perceptions of the fungibility of different sources of income are based on the underlying assumption that individuals are most often involved in multiple income-generating activities, either consecutively or concurrently. That is, it is not uncommon for individuals who have low-wage jobs to engage in multiple forms of incomegenerating activities, including informal and illegal activities (Crutchfield 2014;Nur and Nguyen 2022). Fagan and Freeman (1999) argued that legal and illegal income-generating activities exist on a continuum, with many individuals "doubling up" by participating in multiple income-generating activities simultaneously. Our findings reveal a majority (∼60 percent) of respondents in Study 1 (our incarcerated sample) "doubled up" during the two months prior to their incarceration, with considerable variation in earnings. This finding is important because it suggests that stable employment does not always serve as a turning point, or deflection, in an individual's offending trajectory (Nguyen and Loughran 2018). In fact, we find evidence that legal employment alone is not a sufficient condition for many individuals to refrain from illegal income-generating crimes. 9 While it may be that this is due to insufficient income generated legally, it also questions the idea that legal and illegal monies are perceived as pure substitutes.
We found support for the theoretical concepts which guided our study. First, we found that the episodic, frequent, and cash-based nature of illegal payouts played a large part in how our respondents disbursed their various sources of income. For many of them, the liquidity of illegal money made it easier to budget and pay for products consumed at high rates, such as illicit drugs or alcohol. Additionally, the source of income matters. Contrary to the traditional economic assumption of the fungibility of money, respondents from both studies attached different meanings to the money they made legally compared to the money they made illegally. Legal money was referred to as "hard earned money" or "regular money" and many respondents were proud of the income they earned from their legal job. Legal money was often earmarked for housing or essential bills. Sentiments of money earned through illegal activity were often thought of as "extra money," with little pride or attachment to illegal money.
Second and importantly, findings from this study suggest because of the nature and source of illegal money, it was often spent on partying, which is in line with the extant research on offender lifestyles (e.g., Felson et al. 2019;Shover and Honaker 1992;Short 1966). This consistency between the source of income and its intended use demonstrates the decision-making factors inherent in money acquisition are based on systematic modes of thinking about money and consumption. It also suggests a reciprocal relationship between sources and the nature of money. We find that criminal pursuits result in the acquisition of funds that are treated differently than those acquired legitimately. Offenders were more likely to spend ill-gotten gains on partying and conspicuous consumption, and expend legitimate resources (e.g., from a job) on mainstream expenditures (e.g., paying bills). At the same time, we find that the acquisition of "dirty" money further encouraged predatory crime to extend the pursuit of partying and other conspicuous consumption. This is in keeping with previously described etiological models of street crime (Lofland 1969;Wright and Decker 1997;see also, Felson et al. 2019). The attraction of such consumption was clearly evident in our sample. Offenders who wish to engage in partying would increasingly pursue illegal sources of income for these purposes, while increasingly abandoning legitimate sources of income and legitimate pursuits.
The upshot of this is that the extent to which offenders can exercise some level of self-control and conscious decision-making about money is likely subject to prevailing situational conditions. Throughout the interviews of offenders in Study 2 we noted that the clear separation of "dirty" versus "legit" money was more pronounced when offenders were going through periods of "plenty" and that the cross-use of funds (i.e., using legitimate money for partying or illegitimate money for necessities) was more pronounced when offenders experienced pressures (be they financial, social, or personal). A key driver of this break-down in discrete consumption patterns was the extent to which offenders were caught up in cycles of crime and partying noted by Wright and Decker (1997;see also, Lofland 1969). This type of lifestyle was more frequent in Study 2 (our active offender sample). Offenders trapped in such cycles were more likely to violate the separation of monies for legitimate or illegitimate pursuits. Importantly, even when there were such break-downs offenders acknowledged that this was undesirable, 10 confirming the importance of mental accounting in their day-to-day lives. Future studies should examine the situational factors that can impact the extent to which fungibility is violated.
Similarly, there are policies which target cashless transactions to reduce more readily available sources of monies (Wright et al. 2017). Such cashless environments could foster a different means of internalization of different types of income across illegal market participants. Offenders rely heavily on cash as a spending medium because of its transactional anonymity. For some offenders, the cycle of criminality mentioned previously involves individuals brought into patterns of criminality designed to bankroll their partying (see Felson et al. 2019;Shover 1996). While there has been working that has identified money as important (e.g., Jacobs and Wright 1999;Katz 1988;Wright and Decker 1997), they do not expressly focus on the extent to which differential perceptions of money, based on its nature and source, drive consumption (and offending more broadly). Programs that remove cash from the streets are likely to change the manner in which offenders engage in consumption (see Tekin et al. 2014). As such, it is not only the source of activity (legal or illegal) that drives differential consumption patterns but the nature of gains obtained from those activities (cash vs. checks, vs. debit cards for example).
Our findings also provide insight into some important extensions for the study of choice theories in crime (Nagin 2007). First, our results offer continued support for the integration of concepts from behavioral economics and sociology to refine and redefine traditional rational choice models of offending (Pogarsky et al. 2018). As with any application of behavioral economics, a necessary starting condition is a clearly specified rational model predicting how a rational individual "should" act (Camerer and Loewenstein 2011;Loughran 2019). In the case of consumption, the traditional life cycle theory is the classic rational consumption model (Modigliani and Brumberg 1954), which among other things assumes fungibility of monies, an assumption which has been challenged on the basis of empirical evidence (Shefrin and Thaler 1988;Thaler 1990). In particular, there is now strong evidence of what Thaler (1990) refers to as "mental accounting," by which "various forms of wealth do not appear to be as close substitutes as the theory would suggest" (p. 194). Our results imply that the legitimacy of the source of earnings plays an important role in delineating mental accounts. As such, we continue to implicate the use of predictable departures from rational behavior in shaping criminological choice theories through the systematic refinement of rational choice.
Second, our findings support the role of social structure, context, and culture in shaping the choices that are available to individuals. The extant literature on decision-making has typically conflated broader rational choice and deterrence by largely focusing primarily on individuals' responsiveness to risk and costs (Paternoster 2010). Although there have been some recent development in extending our understanding of offender decision-making in response to rewards for offending (e.g., Loughran et al. 2016;Nguyen et al. 2021;Thomas and Vogel 2019;Thomas et al. 2022), empirical evidence on the relationship between larger societal structures and social organization on decision making is woefully absent. Similarly, the cultural context in which individuals are surrounded and the values and norms of work and money play a pivotal role in decision-making regarding making and spending money. Our respondents, both incarcerated and active offending samples, are highly constrained in their set of choices. They are part of the low-wage labor markets and often from impoverished communities consisting of a largely low-wage workforce with high rates of unemployment. Low human capital and criminal social capital can push individuals in the low-wage labor market to take on additional, alternative sources of income (Freeman 1996).
Just as the social structure and culture shape decision-making on making money, consumption is also impacted. Scholars have argued that inequality can shape perceptions among the poor to consider conspicuous goods as essential (Banerjee and Duflo 2007). That is, relative deprivation may raise their minimum required spending on conspicuous goods, such as entertainment, eating out, clothing, or luxury items. Research suggests that some individuals who are socioeconomically disadvantaged and who have low human capital invest in relatively high levels of conspicuous consumption because they lack other signals or recognition of other accomplishments (Moav and Neeman 2012;Ordabayeva and Chandon 2011). Moav and Neeman (2012) argue that conspicuous consumption can suppress upward mobility and prevent the poor from escaping a poverty trap. The salience of conspicuous consumption might reduce the effectiveness of programs such as income maintenance or cash transfer in reducing criminal behaviors. Christiaensen, Pan, and Wang (2013) argue there is a direct link between spending behavior and the source of income-money that is earned more easily is also spent more easily. The authors analyze the consumption of households and found that there is a higher marginal propensity to spend unearned income on alcohol and tobacco, clothing, transportation, and communication, while earned income is more likely to be spent on education. A concerted effort to better understand the dynamics of consumption among those who are involved in criminal activity might allow us to inform policies and programs that create contexts in which people spend more responsibly.
Third, it is likely that our findings have implications to extend theories of desistance from crime. For theories that focus on choice and identity in the desistance process (e.g., Giordano 2016; Paternoster and Bushway 2009), different forms of income could be used to more directly test essential concepts of identity change and cognitive transformation in the desistance process. For instance, it could be posited that individuals actively engaged in crime have two unique and distinct utility functions, one for regular consumption and one for the "pleasures of life," both of which might be governed by illegal income though the former is only governed by legal income. Desistance would then be hypothesized to occur as a utility from the latter decreases, for instance during the transition from adolescence to adulthood (e.g., Thomas and Vogel 2019). Specifying and measuring the parameters of this type of model (e.g., the marginal rate of substitution between a legal dollar and an illegal dollar at different points of the life course) could represent concrete ways in which to define, model, and test the role of humans agency in the desistance process (Paternoster 2017;Thomas et al. 2021).
Finally, it is important to point out that policies seeking to address issues such as employment/income maintenance are predicated on the idea that money from legal avenues can offset the need for illicit income. Yet, there is accumulating evidence that such programs are largely ineffective in achieving the desired goals of reducing both recidivism specifically, and the attractiveness of criminal opportunities more generally. A key finding in both of our studies is that offenders maintain strong motivations to persist in offending that go beyond objective income-generating capacities. We suspect those less entangled in the cycle of criminality mentioned previously will be more amenable to programs that focus on replacing illicit income with legal income, but such programs will have to additionally counteract many of the behavioral notions of value, consumption, and the mental accounting that are embedded within illegal income. The chances of achieving meaningful results with more serious offenders are more equivocal, given the existing evaluations of employment programs in prisons as well as post-release programs (Visher et al. 2005). Nonetheless, given the reciprocal nature of making money and spending money, incorporating research on consumption patterns can provide interesting avenues for programs aimed at improving the employment prospects of individuals involved in income-generating criminal activity.
We believe that the study of consumption in criminological discourse can be developed through a number of strategies. First, we urge scholars to develop tools to measure consumption that not only examines how consumption varies by the source of the income but also how consumption varies in response to the amount of the income (Shefrin and Thaler 1988;Zagorsky 2013), foreknowledge of the income (Arkes et al. 1994), and amount of effort invested in acquiring income (Kivetz and Simonson 2002). Second, it would be important to study the consumption of incarcerated and non-incarcerated individuals, and a non-offending sample. To comprehensively understand how offenders think about and engage in consumption, their responses need to be measured against those who do not engage in crime, a common practice in the literature on criminal expertise (see, e.g., Nee et al. 2019;Topalli 2005a;Wright et al. 1995). Third, systematic inquiry into between individual and within individual factors that are related to variation in mental accounting could shed light on the desistance process and the success of income or employment programs. For example, it is possible that stronger prosocial ties could impact an individual's tendency to differentiate between legal and illegal money and thus alter the crime choice. Fourth, concepts and theoretical ideas related to consumption including loss aversion, incometargeting, and reference dependence in illegal earnings all yield potentially substantive implications to the study of labor market incentives, income maintenance, and the theoretical foundations of offender decision-making.
Our paper merged quantitative and qualitative information to illustrate that uniting multiple, disparate lines of scholarship focused on offender decision-making can provide substantial insight into a novel line of inquiry. Despite their similar purpose, the quantitative and qualitative literature focused on offender decision-making have tended to develop along somewhat independent trajectories. Yet, the salience of money in the lives of offenders resonates across both quantitative and qualitative literatures on offender decision-making. One benefit of our mixed methods approach in studying consumption among offenders is that we were able to capitalize on the integration of behavioral economics embraced by empirical quantitative studies (e.g., Loughran 2019; Pogarsky et al. 2018), and draw on qualitative works to provide detailed information on individual thought processes. We advocate for an interdisciplinary framework, that includes mixed methods and sociological approaches, in studying context and choice.
2. Interestingly, research which has challenged standard rational choice in terms of incorporating both principles of behavioral economics (Pogarsky et al. 2018) and emotions and affect (van Gelder 2017) has largely followed this divide in terms of points of departure. 3. To be clear, Becker's (1968) model differentiates between risk premiums for acquiring a dollar illegally (which involves considerable more risk and cost) than acquiring a dollar legally, and hence would-be offenders should demand a higher premium for engaging in illicit activities (Grogger 1998). 4. In many ways, Duesenberry's theory is a forerunner to both structural and behavioral explanations of departure from predicted rational behavior (McCormick 2018). 5. These systematic departures from the normative model were defined by Thaler (1987:167) as anomalies, "if [they are] difficult to 'rationalize', or if implausible assumptions are necessary to explain [them] within the paradigm." 6. A number of empirical studies demonstrate support for mental accounting principles and the framing of choices (e.g., Heath et al. 1995;Prelec and Loewenstein 1998;Shafir and Thaler 2006). 7. In each of these three studies, data were collected 12-18 months prior to their publication. 8. This is reflective of the offender population of the metropolitan areas of St. Louis and Atlanta as well as the network of offenders available to project recruiters (all of whom were African American males operating in African American majority neighborhoods in these cities). 9. As confirmed in Study 2 it may even be part of a larger strategy to conceal one's criminal involvement. In addition to the comment by Cee-Low in that study, there were similar comments by offenders such as, "I just work the job at the carwash so when the police stop me and go through my pockets and ask where I got this cash from I can just say it's my tips from the carwash." (Bluey, drug seller) 10. As Ghost (drug dealer, robber) stated, "I hate to use my paycheck for dope, but sometimes, you gotta do what you gotta do." Similarly, Bluey (drug dealer) commented, "I was gonna take that little money from that [dope] deal and go to the club but my girlfriend wouldn't shut up about the diapers and the formula and all that other baby shit and I just gave it to her to shut up."