1 Extralegal governance is not extraordinary

A vast amount of economic and social interactions occur outside of the control and oversight of state-based legal institutions. For instance, there are more than two billion informal workers in the global economy and they constitute roughly half of the world’s workforce (Hummel 2021, 1). These workers are often unlicensed and untaxed. In politics, international relations between countries operate in a setting whose main ordering principle is anarchy (Waltz, 2010). There is no single, centralized authority that can establish international order. According to one measure, in 2009, about half of countries were on, or near, the cusp of state failure (Leeson & Williamson, 2009). More generally, in 2004, then-Chief Economist and Director of Research at the International Monetary Fund, Raghuram Rajan, argued that we are better off to “assume anarchy” in our efforts to solve the problem of poverty around the globe because most economically underdeveloped places lack effective state-based legal institutions (Rajan, 2004). Even in developed countries, the illicit economy and cross-border smuggling often thrive (Andreas, 2013). More generally, from an historical perspective, most societies have likewise lacked a strong, effective state; instead, an overwhelming majority of states in human history have been predatory and extractive (North et al. 2009, xii).

The inability or unwillingness to rely on state-based legal institutions arises for numerous reasons. In many cases, the state simply lacks the capacity to govern. Non-governmental organizations have a poor record of building high-quality political and economic institutions (Easterly, 2002). In other cases, religious, ethnic, and other minority groups fear reliance on a strong state that can prey upon, rather than protect, their interests (Scott, 2010). Racism and violent lynchings in the Jim Crow South were a strong signal that Black people could not, and should not, turn to the state and state-adjacent groups for safety and property protection (Dray, 2003; Obert & Mattiacci, 2018). Likewise, people with minority sexual and gender identities have historically been discriminated against by state institutions. Even when not explicitly legally criminalized, people in marginalized groups often avoid legal institutions out of fear, shame, or uncertainty. War and conflict also often prevent reliance on high-quality state-based institutions, both because war is destructive and also because rebuilding efforts are often unsuccessful (Coyne, 2008).

In the context of organized criminal activity, precisely because of its illicit nature, people engaged in criminal activity cannot rely on legitimate, state-based governance institutions in their formal capacity. Such people lack court-enforced contracts, police protection of body and property, employment regulations, and access to legal and conventional banking, insurance, and licensing services. People might gain access to these services illegally or through corruption, but they cannot use them with the legitimacy of a law-abiding citizen. Nevertheless, for economic life to occur in such spaces, people need to find alternative sources of governance.

There are several well-known ways in which self-enforcing exchange can be successful in the absence of strong, effective states: mechanisms based on reputation, exclusion, and inclusion (Boettke, 2005). First, even in the context of an indefinitely repeated prisoner’s dilemma game, the promise of future interactions for people with good reputations can elicit cooperative behavior (Fudenberg & Maskin, 1990). Gordon Tullock described this as the “discipline of continuous dealing” (Tullock, 1985). In a pioneering article, Macaulay (1963) finds that most business people view court-based contractual enforcement as costly, disruptive, and unprofitable, preferring instead to resolve conflict privately, with the aim of maintaining a good business relationship.Footnote 1 The presence of transaction costs likewise means that all contracts must be incomplete, so there is no reason to think that courts can deliver acceptable rulings in all, or even many, instances (Coase, 1937; Hart & Grossman, 1986; Holmström, 1982). Lengthy and costly legal proceedings are often more trouble than they are worth (Benson 1989, 179-196). Instead, people rely on a variety of alternative dispute resolution services to solve problems “in the shadow of the law.”

However, reputation mechanisms do not always work well. They tend to be more successful in environments where information costs are relatively low, which often arises in small, tight-knit communities. Learning someone’s reputation is easier in such settings, and tarnishing someone’s reputation with other people is as well. Robert Ellickson’s famous study of rural communities in Shasta County demonstrates how local reputation pressures can be quite powerful (Ellickson, 1991). Neighbors knew each other (and each other’s cattle), so they knew who to blame for trampled crops. Local residents could use the power of gossip and ostracism to incentivize good behavior. Locals also followed shared norms about liability for damages, instead of using the actual liability laws, when interacting with neighbors (but not outsiders). As a result, much of their interactions emerged as a type of “order without law.” Ellickson (Ellickson (1991), 167-183) notes, however, that this would not scale up in larger, more anonymous communities.

A second common mechanism to facilitate self-enforcing exchange is by exclusion mechanisms through clubs (Buchanan, 1965) and other organizations that generate homogeneous and relatively trustworthy communities. Exclusion mechanisms work by finding a way to prevent uncooperative people from joining a trading community in the first place (Powell and Stringham 2009, 513-514). If there are credible ways to identify who to trust, then within the community, people can be confident in trading with other group members. For example, the earliest stock exchanges in London and Amsterdam were based on self-enforcing exchanges (Stringham 2015, 39-78). Many of the trades they made were illegal, so such deals could not be enforced in state courts. To overcome problems of opportunism, trading houses had strict rules for who could join. Membership generated substantial rents, so people wanted to join and feared being kicked out. As such, when problems arose, the traders and the house would adjudicate such problems privately. Lisa Bernstein likewise shows that the diamond trade in New York City did not rely on written contracts (Bernstein, 1992). Instead, a combination of extensive social capital based on religious ties and access to the supply of diamonds itself were so valuable that people feared ostracism. More recently, Richman ((2017), 148-176) finds many of the same people in this community have turned to written contracts because of the increasing complexity of global trade, the diversification of the diamond supply chain, and the advent of new technologies.

A third common mechanism that facilitates self-enforcing exchange is through inclusion mechanisms that make trade possible in heterogeneous communities. Pre-colonial African trade provides an example of such mechanisms at work (Leeson, 2008). Powerful European traders desired opportunities to exchange with relatively weak indigenous groups, but there was little to no trust between these parties. Locals had little reliable information about the unobservable characteristics of potential traders, including most importantly, their trustworthiness. In order to credibly signal the desire to invest in long-term, mutually beneficial relationships — and thus their trustworthiness — European parties made relationship-specific asset investments. These included learning local languages and adopting local customs, religions, cultures, and intermarriage with local people. These investments would only pay off over time. By investing in nonrefundable relationship-specific assets, traders reduced their social distance to local people, signaled trust, and promoted cooperation.Footnote 2 This same mechanism appears to be in operation in the context of the underworld, where sociologist Diego Gambetta has identified numerous ways that illicit actors rely on credible signals to find trading partners and induce cooperation (Gambetta 2011, 6-18).

In summary, self-enforcing exchange happens in a wide range of industries and communities and over a vast scope of exchange scenarios. Some of these are successful cases of exchange in the “shadow of the law” and others are instances of “order without law.” Across both types of self-enforcing exchange, public choice economists have identified numerous mechanisms and institutions that solve (or at least improve on) the problem of costly access to high-quality state-based legal institutions (e.g., Stringham 2005).Footnote 3

2 Criminal governance

In this article, I focus on extralegal criminal governance as a type of self-enforcing exchange produced both by organizations and institutions (North, 1990). Governance institutions, in general, define and enforce property rights, facilitate economic exchange, and aid in the production of collective action. “Extralegal” refers to the fact that these institutions are produced outside of the scope of state-based legal institutions, but are aimed at similar ideal-type ends (Bateson 2021, 927). Of course, state-based actors can likewise act in extralegal ways, but in doing so, they act outside of what people perceive as the legitimate scope of state action (Bateson 2021, 928).

Organized crime groups often provide criminal governance. By organized crime, I mean something more specific than crime carried out in groups or by groups with an internal hierarchy (Schelling, 1971). While there have been numerous competing definitions and explanations (for example, Buchanan (1973); Varese (1994); Dick (1995); Leeson and Rogers (2012)), the definition that has become focal in the literature is that an organized crime group “attempts to regulate and control the production and distribution of a given commodity or service unlawfully” (Varese 2010, 14). This might include something like illegal drug distribution of heroin or the illegal trafficking of legal substances, such as cigarettes from a low-tax state to a high-tax state. A “mafia” is a type of organized crime group that specifically “attempts to control the supply of protection” (Varese 2010, 17). Mafias seek to create and govern the market itself.

Recent work conceptualizing criminal governance identifies three common areas of operation (Lessing, 2021). First, internal governance institutions regulate interactions within an organized crime group (e.g., Skarbek and Wang 2015; Piano 2017; Sløk-Madsen et al. 2021). This might entail rules for how people gain positions of authority, taxes that have to be paid on illicit activities carried out by members, or rules for approving assaults and murders. For example, Leeson shows how sailors aboard 18th-century pirate ships established constitutional democracies onboard and overcame problems of predatory captains (Leeson, 2007, 2009, 2010; Leeson & Skarbek, 2010). In Brazil, the Primeiro Comando da Capital prison gang uses a consignment model of drug distribution based on detailed records, loans, and relatively light punishments for those who default (Lessing & Denyer Willis, 2019).

Second, criminal-market governance institutions focus on how organized crime groups regulate illicit markets. It is well established that mafia groups — in Italy, Russia, China, Brazil, and many other places in the world — often provide protection to people engaged in illicit activity in places with weak and failed states (Reuter, 1983; Gambetta, 1993; Skaperdas, 2001; Varese, 2001; Bandiera, 2003; Varese, 2011; Buonanno et al., 2015; Dell, 2015; Dimico et al., 2017; Wang, 2017; Barnes, 2022). However, mafia organizations can also play an important role in illicit markets where the state is strong enough to enforce regulations, which in turn generates a demand for their services. For example, Gemma Dipoppa (2023) finds that mafias in Northern Italy had a market to provide illegal labor because the state was, in fact, fairly effective at enforcing local labor regulations that made hiring legal workers relatively expensive. It was the strength of the state, rather than its weakness, that motivated a demand for extralegal governance.

Finally, criminal-civilian governance institutions involve organized crime groups that govern civilian populations who are not actively involved in illicit activity (e.g., Frye 2002). Recent research on Colombia, for example, finds that hundreds of local neighborhood gangs called “combos” regulate daily life in the community much like we might expect civil servants and police officers to do in more developed countries (Blattman et al., 2023). Residents turn to combo members to regulate noisy neighbors, intervene in domestic violence, disrupt street fights, respond to robberies and thefts, and intervene when someone is doing drugs near children (Blattman et al. 2023, 13). Residents tend to view combos as less effective and less legitimate than the state’s role in these activities, but they often still view combos quite positively in both regards.

Across these three areas, criminal governance often performs one or several of five functions (Lessing 2021, 860). First, criminal governance often provides policing and enforcement. Organized crime groups might prohibit theft and enforce property rights, regulate interpersonal violence, enforce social control in public places, and control entry and exit from certain territories. Second, criminal governance sometimes provides judicial functions. This includes adjudication of disputes, trials and punishments, debt collection, and contract enforcement. The Primeiro Comando da Capital prison gang in São Paulo, Brazil, for example, has performed these functions effectively throughout many parts of the urban landscape, despite its leaders being incarcerated (Lessing & Denyer Willis, 2019). A third important function performed by criminal governance institutions is of a fiscal nature. Gangs often extort both criminal and civilian actors (e.g., Shortland 2019; Magaloni et al.2020; Sviatschi 2022; Gilbert 2022). In El Salvador, for example, one study reports that perhaps as many as 70 percent of businesses pay extortion to local gangs (Brown et al., 2023). As I discuss below, the Mexican Mafia prison gang extorts “gang taxes” from Hispanic drug-dealing gangs throughout much of Los Angeles County and Orange County (Skarbek, 2011). Fourth, some organized crime groups provide regulation in licit and illicit markets. In Rio, for instance, drug gangs prohibit residents from bringing stolen cars into their neighborhood for disassembly (Lessing 2021, 862).

Finally, an important area of recent scholarship focuses on the role of organized crime in politics. Mafias have often used violence against elites, bribed political actors, sold candidates access to voters, and coerced voters and politicians into supporting preferred policies, candidates, and parties, among many other relationships (Arias, 2006; Dal Bó et al., 2006; Durán-Martínez, 2015; Cruz & Durán-Martínez, 2016; Arias, 2017; Daniele & Dipoppa, 2017; De Feo & De Luca, 2017; Durán-Martínez, 2017; Albarracín, 2018; Alesina et al., 2019; Moncada, 2019; Phillips & Ríos, 2020; Trejo & Ley, 2021; Blume, 2022; Moncada, 2022; Siddiqui, 2022).Footnote 4 These relationships take different forms. Sometimes, political actors solve coordination and commitment problems in the illicit economy. Snyder and Durán-Martínez (2009), for example, focus on the role of state-sponsored protection rackets in Mexico. They show that when the PRI party dominated politics, it was relatively easy for drug trafficking organizations to establish reliable relationships with political actors. This led to relative peace for many years. However, as PRI lost its dominant control over politics and the country became both more democratic and less corrupt, these protection rackets broke down, giving rise to staggering levels of violent crime and unprecedented homicide levels (see, also, (Trejo & Ley, 2018, 2020)). In other cases, organized crime groups solve problems facing politicians. In Brazil, Jessie Trudeau (2022) uses novel data and a unique natural experiment to show that local gangs in favelas undermine elections by corralling residents to the polls to support favored candidates and prevent alternative candidates from campaigning in the voters’ neighborhoods. Finally, members of organized crime groups can also simply serve as “hired guns” for politicians. For example, one explanation for the emergence of the Sicilian mafia at the end of the 19th century is that they were born in response to the rise of the socialist Peasant Fasci movement (Acemoglu et al., 2020). Local political elites hired mafias to intimidate and suppress the budding socialist movement. In short, organized crime groups are often important political players.

While data limitations make it extremely challenging to estimate the global prominence of criminal governance, we know that it is neither rare nor limited to past historical periods or particular regions in the world. For example, one survey finds that 14 percent of respondents across 18 Latin American countries report that criminal groups are involved in order-provision and crime-reduction, affecting somewhere between 77 and 101 million people (Uribe et al., 2023). This estimate would obviously expand substantially with a wider global focus and by giving greater attention to places of punishment in both developed and developing countries.

More generally, the study of criminal governance is emerging as an important concept in political science and economics. Public choice—and political economy more broadly—provides many foundational concepts, theories, and empirical approaches that have been central to this work.Footnote 5 In particular, several classic questions in public choice research have motivated and framed the analysis of criminal governance. First, why do people need extralegal governance? Public choice scholars have long argued that we cannot simply assume that effective state-based institutions exist; instead, we need to theorize and model social interactions to see if, and how, such institutions emerge and work (Tullock, 1972, 1974; Buchanan, 1975).Footnote 6 Second, what internal governance problems do organizations face, and how do they respond? Gordon Tullock wrote extensively about the information and incentive problems associated with “non-market decision making” in bureaucratic organizations and in autocratic regimes (Tullock, 1965, 2005). Mafias face many of the same challenges. Likewise, in The Calculus of Consent, Buchanan and Tullock take up the Madisonian challenge about how people can create organizations that are powerful enough to govern but constrain the abuse of such power ((Buchanan & Tullock, 1962); also, see (Weingast, 1995)). Third, under what conditions do governance institutions emerge? Mancur Olson’s theory of the stationary bandit provides substantial explanatory power in understanding where, why, and how states and criminal actors actually govern (Olson, 1993). Finally, under what conditions are self-governing institutions more likely to be successful? Elinor Ostrom developed theory and evidence from a wide-ranging sample of settings to explain the tremendous diversity of institutions and institutional outcomes (Ostrom, 1990, 2009). In the following sections, I draw on some of my past work on the California prison system and on the history of prison gangs within their walls to discuss how these public choice insights contribute to our understanding of criminal governance.

3 Internal governance with criminal constitutions

La Nuestra Familia (NF), a prison gang originating in Northern California in the 1960s, is one of the most powerful and best organized gangs in the California prison system (Skarbek, 2010). The gang formed initially in response to high levels of violence and racial animosity from both incarcerated people and prison staff. Although it would become a major player in the underground economy, the gang initially formed as a self-protection group, mostly for Hispanic people from rural areas of Northern California. The gang today provides a variety of benefits to its members, including protection over one’s body and property, access to illicit markets, and adjudication of disputes with members and nonmembers. The NF gains revenue from numerous illicit endeavors. Inside prison, their members benefit from the drug trade, extortion, theft, gambling, racketeering, and smuggling contraband. Beyond prison walls, the primary income for NF stems from drug distribution, armed thefts, burglaries, and home invasion robberies. As articulated by a veteran NF member, joining the organization offers “increased power, enhanced knowledge, insight, and protection” (Fuentes 2006, 262).

Prisons are difficult places to carry out collective action successfully. Prisoners face the classic free-rider problem of producing local public goods (Olson, 1971). Some of what the gang does also generates benefits that are not easily excludable. For instance, gangs regulate the use of large-scale or serious violence. However, some incarcerated people prefer to free ride on the activities and behaviors needed to achieve that. Selective incentives might be difficult to provide since prisoners typically lack resources and live in relatively impoverished environments. They likewise lack autonomy, mobility, and privacy, all of which could be used to promote cooperation and reduce free-riding. Nevertheless, gang membership provides enough private benefits that many people wish to join a gang.

One Grand Jury indictment explains the process of becoming a member:

Usually membership was approved only after a period of indoctrination by Nuestra Familia members in which prospects were schooled and then tested ... about Nuestra Familia criminal activities, including the manufacture of prison weapons, or “shanks,” and the commission of murder, robbery and drug dealing inside and outside of prison. Only those who evinced both a knowledge of, and willingness to participate in, these activities were eligible for membership. This perpetration of violence and other crimes was ordinarily a prerequisite to membership in the organization, and membership, once achieved, was for life, as symbolized by the organization’s membership phrase, “blood in, blood out” (Fuentes 2006, 296).

“Blood in” means that a potential member has to assault or kill an enemy of the gang to join. “Blood out” means that the only exit from the gang is death. Thus, prisoners face a stark constraint: no exit options. Prison walls prevent physical exit. Incarcerated people have little say over where to live and whom to live with. There is no self-sorting into exclusive communities of trustworthy people. Moreover, a lifetime commitment to a gang raises the threat of ex-post opportunistic behavior. How can the gang credibly signal to high-quality potential members that they will not regret joining? Even if collective action to create a group is successful, empowering a gang raises a fundamental problem in political economy (Weingast, 1995). A group strong enough to protect one’s rights is usually strong enough to violate those rights. In one of its most elegant articulations, James Madison writes in Federalist Paper #51:

If men were angels, no government would be necessary. If angels were to govern men, neither external nor internal controls on government would be necessary. In framing a government which is to be administered by men over men, the great difficulty lies in this: you must first enable the government to control the governed; and in the next place oblige it to control itself.

Like governments, a prison gang’s coercive capacity needs constraints because its members are no angels.Footnote 7 In Californian prisons, people want to empower a gang to be strong enough to protect their rights, but also to constrain the gang from using that power to abuse their rights. With few exit options and a lifetime commitment, the problem of a predatory gang leadership loomed as a serious threat for the Nuestra Familia. In response, the founding members wrote a detailed constitution that outlined the group’s purpose and internal governance. One NF member explains “It was like the United States Constitution, but it was our constitution” (Center for Investigative Reporting, 2007).

The hierarchy of the gang, as described in the original constitution, is structured into four tiers: a General, followed by Captains, Lieutenants, and Soldiers. A single General oversaw the imprisoned members. The General directed as many as ten Captains, typically located in separate prisons. Each Captain, in turn, led Lieutenants and Soldiers within their designated correctional institutions. The Constitution states that the Nuestro General is “the supreme power in the organization ... his powers shall have no limit” (Art. 1, §1). This raises the threat that the General will use this supreme power to prey on gang members. However, the constitution actually limits the General’s power in several important ways. First, during peaceful times, the General has less authority. While he has the freedom to both appoint and dismiss Captains during times of war, he can only dismiss them during peaceful times. If he does so, the former Captain’s regiment members will choose a new Captain to safeguard their interests and oversee the General’s actions. Another mechanism that limits the General’s power is through impeachment. If all of the Captains sign an impeachment statement, the General can be removed from his position (Art. 1A, §1). His authority is stripped from him as soon as the document is presented. If he disputes the legitimacy of the signatures, a Soldier will be tasked to confirm the Captains’ votes (Art. 1A, §2). Upon the General’s impeachment, the highest-ranking Captain will take over his role. As a result, this top Captain is also motivated to keep a close watch on the General’s actions.

How can the General avoid impeachment or coups? First, if one of the Captains does not follow the constitution or poses a threat, the General has the power to dismiss him. As noted, removing a Captain does not guarantee that a more agreeable replacement will take the helm, as Captains are chosen by the collective regimental members. Nevertheless, it provides a way for the General to check bad actors and provides a deterrent. Second, any harm or assassination attempt on the General triggers an immediate declaration of war. This elevates the risks of attempting to dethrone the General. In the event of an unsuccessful coup, the General, during times of war, has the liberty to appoint Captains, bypassing the typical electoral process of the members. Third, the General decides who his successor will be. The successor, who must be a “well-versed and battle-hardened warrior” (Art. 1, §3), has an incentive to keep a vigilant eye on the other Captains. Being next in the succession line, he has a vested interest in ensuring a smooth transition. If a coup takes place, his anticipated position might be jeopardized.

Captains monitor the General, but they also have the power to coerce Lieutenants and Soldiers. How are Captains constrained in using this power to the detriment of the organization? Several mechanisms accomplish this. First, the General ranks the Captains “according to their leadership ability and overall foresight” (Art. 1, §7), providing some incentive for good leadership. In times of peace, Lieutenants who are unfairly demoted can appeal to the General (Art. 3, §2). The General will then assign at least three soldiers to review the situation. If the evidence suggests that the demotion was unwarranted, these soldiers can either discipline the Captain or reinstate the Lieutenant. If a Lieutenant feels unjustly punished by a Captain during times of peace, they can also appeal to the General (Art. 6, §1). Finally, Captains are always answerable to the General (Art. 3, §10). If serious complaints arise against a Captain, the General can remove him, allowing the regiment to elect a new Captain.

The Lieutenant acts as the organization’s representative, working closely with Soldiers. Their ranking is determined by their Captain’s assessment of their experience and leadership skills (Art. 4, §4). Lieutenants are responsible for maintaining communication with the entire regiment and consistently setting a positive example for Soldiers (Art. 4, §1). Lieutenants are also in charge of maintaining a record called the Bad News List, which lists all adversaries of the NF (Skarbek 2014, 90-95). They review daily the prisoner transfer list to identify any listed enemies entering the facility (Art. 4, §5). If such an enemy is identified, arrangements are made to assault or kill the person. Court records state, “Every lieutenant in Nuestra Familia was tasked with addressing individuals on the list. Senior members informed their juniors about the gang’s foes, and a lieutenant could independently direct his soldiers to take out a recognized enemy of the Nuestra Familia” (People, 1981). Additionally, Lieutenants converse with new NF members to gather information on any unidentified enemies of Nuestra Familia (Art. 4, §5a).

In summary, early members of the NF recognized that they faced a problem fundamental to political institutions: the need to empower a group to protect rights, but also to find ways to constrain that power from being misused. As with the US constitution, the NF constitution relied on a number of checks and balances, elections, and impeachment options. Of course, the constitutional system does not work perfectly, and over the years, it has changed in important ways. Nevertheless, the constitution works well enough that the group has operated for more than sixty years. It is currently one of the most successful and feared prison gangs in California, and they are able to recruit members year after year. Their internal criminal governance institutions have been crucial to this success.

4 Criminal-market governance in the streets of LA

Like the Nuestra Familia, the Mexican Mafia prison gang has operated in the California prison system for more than fifty years. Its members have historically been Hispanic people from Southern California. Some recent estimates suggest they have about three hundred “made” members and roughly one thousand associated or affiliated members (People, 2008; United States, 2003a). According to many people in law enforcement, the Mexican Mafia, often called La Eme, is the most powerful prison gang in California (People, 2008; United States, 2003a).

Many Mexican Mafia members are incarcerated in county jails throughout Southern California.Footnote 8 However, many members are also housed in state prisons spread across California. For example, Pelican Bay State Prison holds some of the most senior Mexican Mafia members, and that facility is more than 700 miles from the heart of Los Angeles. The fact that members are separated makes it difficult to organize collective action. They are not allowed to write or call each other, so communication is relatively costly. Even within a prison, many members are held in highly segregated units, with some members being housed in solitary confinement. The gang almost appears fully contained and hidden from the minds of people in the free world. However, that’s not the case.

Despite all of these constraints, this small group of incarcerated people serves as a quasi-government to thousands of gang and gang-adjacent criminal actors in the free world. The Mexican Mafia defines and enforces property rights over drugs and gang territory. Gangs like MS-13 that are affiliated with La Eme use the number 13 in their gang name to signal to potential predators that the Mexican Mafia protects their property. Territorial expansion can generate substantial increases in revenue, so gangs often have an incentive to dispute territorial lines and grab new turf (Levitt & Venkatesh, 2000). In such cases, the Mexican Mafia often resolves disputes between rival street gangs. There are also the typical run-of-the-mill business conflicts and disputes that arise in any industry interacting with the public and a longer supply chain. Finally, the Mexican Mafia creates rules to limit costly externalities. One of the most interesting ones is that, in the 1990s, they made a rule that street gangs could not conduct drive-by shootings (see Blatchford 2009, 121-123 and Rafael 2007, 34-37). This disruptive practice put innocent lives in danger and drew substantial negative attention from law enforcement and the media. While no individual street gangs had an incentive to stop doing drive-by shootings, the Mexican Mafia had the information and incentive to implement regulations to prevent it.

Given that the Mexican Mafia prison gang is incarcerated and physically separated from the free world, how and why does it provide criminal-market governance on the streets of Southern California? The answer comes from a classic theory in public choice: Mancur Olson’s theory of the stationary bandit (Olson, 1993; McGuire & Olson, 1996). For this hypothesis to offer a compelling explanation, several related mechanisms must be observed. First, the Mexican Mafia must be able to extract resources from the street gang members. Second, the prison gang must be “stationary” or else they would have little reason to reduce their tax demands. Finally, the prison gang must have an encompassing interest, such that it can govern across a large enough space that it will internalize local externalities.

Despite their obvious physical separation from society, the Mexican Mafia is able to communicate, command, govern, and tax thousands of people on the streets of Southern California. Hispanic street gang members in Southern California, known as Sureños, pay thousands of dollars in taxes on drug sales to the Mexican Mafia every year (United States, 2007, 2008a, d, 2009, 2010). This is not a voluntary choice on the part of most street gang members. Refusal to pay or an attempt to opt-out or avoid this system is met with violence, including murder (United States, 2003b). When gangs attempt to avoid paying taxes or receiving any of the associated benefits, they are systematically targeted for assault (Blatchford 2009, 124). For example, the Maravilla gangs of East Los Angeles resisted paying gang taxes from 1993 to 2006, but at the cost of constant in-jail assaults and stabbings (Contreras, 2023). The taxation scheme is all the more striking when we consider the fact that the Mexican Mafia is outnumbered by the gang members whom it taxes. There are an estimated 21,000 Hispanic street gang members in Los Angeles, implying that there may be as many as seventy times as many street gang members than Mexican Mafia members. Given that the street gangs are armed and dangerous enough to control gang territory on the street amid rival gangs, why do they concede to the demands of distant, incarcerated rulers?

First, the Mexican Mafia has a credible threat of violence in Los Angeles jails. The gang members themselves can use violence, and many non-prison gang members will carry out their violent demands out of fear, access to resources, or to gain higher status. When a new person arrives at jail, they are often unfamiliar with the physical layout, ignorant of the social dynamics, have few allies, and are unarmed. By contrast, the incumbent prison gang members are often armed, knowledgeable, and organized. As a Special Agent with California’s correctional department explains:

A street gang member knows, “If I don’t do what I’m told to do on the streets, that when I hit the jail, or when I hit the prisons, there are those who are so loyal to the Mexican Mafia that they’re going to assault me.” So you’re going to do exactly what you’re told, when you’re told to do it and how you’re told to do it, and not ask any questions about it. The prison gang has a credible threat to do serious violence against newly arrived gang members who have not paid gang taxes. (American Public Media, 2008).

As one Maravilla gang member explains, they faced grave danger while incarcerated for not paying gang taxes:

Going to [Los Angeles] County jail was like a death sentence. You had all these Sureños who bowed down to those guys [La Eme] just coming off on [attacking] us. As soon as you were put in a cell with other guys, you knew what was going down. One of them would ask you, “Where you from, ese?” You go, “Soy Maravilla [I’m Maravilla],” and that was it. They would rat pack you ... But when you’re from Maravilla, you can’t back down. I just prepared myself for what was going to happen ... And that was it. About four or five guys started swinging on me, and I started swinging on them back. They got the best of me. They broke my jaw, broke my ribs (Contreras 2023 ,61).

For a threat of future retaliation to be credible, it has to be shared knowledge that drug dealers and gang members will end up in jail in the future. In fact, people involved in gangs and the drug trade have high rates of incarceration and recidivism. In Los Angeles in 2004, for example, nearly half (49.51%) of parolees were reincarcerated within three years. Specifically, for those imprisoned due to various drug-related crimes, the recidivism rate ranges from 45% to 66% within the same period. Focusing on a shorter time frame, within the first year, 28.82% of individuals convicted for selling controlled substances and 45.43% for possessing them are sent back to prison. Looking at a broader scope, across California, the three-year recidivism rate for male offenders stands at 58.73%. Gang membership also increases the likelihood of recidivism (Dooley et al., 2014). Many street gang members consider incarceration a normal and expected part of life; some consider it akin to going to college (Blatchford 2009, 32). Moreover, even if some drug dealer does not anticipate incarceration, it is highly likely that a friend, family member, or fellow gang member will be incarcerated or is already locked up. These associates can then be used as a hostage, forcing someone on the street to pay to stop the Mexican Mafia from hurting or killing the hostage. Between the fear of violence against oneself or against a potential hostage, the Mexican Mafia’s threats are credible.

Having an encompassing interest increases a stationary bandit’s benefits from limiting local externalities and producing public goods. The organization of the Los Angeles jails helps to create such an encompassing interest for the Mexican Mafia. Although Los Angeles is home to nearly ten million people across more than 4,000 square miles, there are only a few jails. In particular, the Men’s Jail and Twin Towers facility hold the largest portion of people. The county jail system acts like a siphon drawing drug-dealing gang members from across numerous neighborhoods across hundreds of miles to a few, localized areas where they are deeply vulnerable. Since the Mexican Mafia can tax Hispanic drug dealers from these areas, they then have an incentive to produce the criminal-market governance needed in the free world.

This is not merely an extortion racket. Olson’s stationary bandit model predicts that if the above conditions (and some other assumptions) are met, then the gang will not just tax, but govern as well.Footnote 9 The evidence supports several of the empirical implications. First, the gang does not tax all, or even most, of the gangs’ revenue. They typically only demand 25 to 30 percent. They also protect tax-paying gangs from thieves and other predators. The “13” designation in a gang name and gang graffiti deters people who know they can be harmed in jail. The gang likewise sends gang associates to kill people who have stolen from their affiliates (People, 2009). As noted, they also provide local public goods, like adjudicating disputes between gang rivals and regulating certain types of violence (United States, 2008a). Second, Olson’s theory also points to reasons why they do not tax and will not provide this type of criminal-market governance to other people and groups. For example, they do not govern Black or White street gang members because these people are protected by their own respective ethnic-gang groups while incarcerated. With no credible threat, the Mexican Mafia cannot tax them, and in turn, has little to no incentive to provide criminal governance.

Not all people are as susceptible to taxation from the Mexican Mafia (see Schelling (1971)). The Mexican Mafia cannot systematically tax people who are not involved in criminal activity. Law-abiding citizens do not anticipate incarceration, and they can turn to the police if the Mexican Mafia attempts to tax them. Even still, not all people involved in illicit activity are similarly subject to taxation. People involved in the sale of voluntarily purchased goods, like crack or meth, need to attract customers. In doing so, they have to make themselves visible to both customers and the Mexican Mafia. By contrast, the best jewel thieves and embezzlers are never visible enough to get the gang’s attention. Related to this, because the Mexican Mafia’s control is based on power in the county jail system, they also do not have a credible threat against drug dealers in counties where they do not have control. For instance, they cannot tax Hispanic drug dealers in Santa Clara County. Instead, the Nuestra Familia uses its control of jails in Santa Clara County to tax and govern local Hispanic gang members (United States Attorney’s Office, 2021). In short, Olson’s theory of the stationary bandit provides substantial explanatory power for the emergence and operation of criminal-market governance on the streets of LA and other major metropolitan areas.

5 Criminal-civilian governance: gang rule in prison

While the previous two sections discuss governance over criminal actors, organized crime groups can also govern regular citizens and people who are not engaged in illicit activity. In this section, I describe how gangs in men’s prisons in California govern the incarcerated population.Footnote 10 For more than 100 years, the California prison system did not have highly-powerful and organized prison gangs. When the prison population was relatively small and ethnically homogeneous, decentralized mechanisms like ostracism and gossip were highly effective at enforcing social control. There was a shared understanding about what constituted acceptable behavior: the Convict Code (Irwin, 1980; Sykes, 2007).Footnote 11 The Convict Code included a variety of sensible norms, including don’t lie, don’t steal, don’t complain, and don’t snitch. To the degree that a person adhered to this code, he would be in good standing. That would mean he was respected by his peers and their support would mean he would be less likely to be victimized. To the degree that someone regularly violated the code, he would have few friends and would be more likely to be victimized. This was a decentralized process that focused on individual behavior. People could adhere to the code or not. People could punish deviations from the code or not. No person or group explicitly crafted the code. An individual’s reputation is what mattered the most.

This system worked relatively effectively until the late 1950s and 1960s. It failed, however, because of striking changes to the demographics of the prison population (Skarbek, 2012; Seagren & Skarbek, 2021).Footnote 12 The prison population grew substantially in size and diversity. More first-time prisoners arrived as well. Since gossip and ostracism work best in small, tight-knit communities, these reputation mechanisms became much less effective. The 1960s saw an explosion of prisoner-on-prisoner violence and large-scale rioting. In the face of this violence, the first prison gangs came into existence as self-protection groups. These groups quickly evolved into prison gangs that engaged in a variety of criminal activities. Now, most incarcerated people in California live under gang rule.

In California, prison social order operates as a community responsibility system (Roth & Skarbek, 2014), which organizationally, looks much like clan-based societies throughout history and in many parts of the world today (Weiner, 2013). Nearly all people in prison have to affiliate with a racial group, which in turn is governed by one of several prison gangs. Within this framework, each gang member’s actions reflect on the gang and race as a whole, and the collective is held accountable for the individual’s behavior. For example, if an inmate affiliated with a gang fails to repay a drug debt to a drug dealer from another gang, it is expected that the offender’s gang will address the issue. The gang will enforce its own form of justice, whether through punishing their delinquent member or paying restitution to the aggrieved party.

This system has become an informal regulatory mechanism within the prison economy. Gangs create and enforce rules that hold their members to account, knowing that the collective reputation of the gang hinges on the trustworthiness and reliability of its individual members. By doing so, they instill a certain level of predictability and security in illicit transactions that would otherwise be rife with uncertainty and risk. The ability of prison gangs to monitor and maintain the behavior of their members enables a form of trade that, while outside the official channels and laws, operates on a set of understood rules within the prison walls. The underground economy for things like drugs, alcohol, and tobacco is obviously extremely important in this context. Even so, this system of gang-based governance rules across a much broader set of quotidian activities and behaviors that are not criminal. To see why, it is important to understand the relationship between gang membership and race and ethnicity.

While nearly all people have to affiliate with a racial group in prison, not all people are actively involved in criminal activity. There are three broad categories to distinguish. First, the California Department of Corrections and Rehabilitation recognizes the most dedicated and entrenched gang affiliates under the title of “Security Threat Group I.” This highest category has encompassed gangs such as the Aryan Brotherhood, Black Guerrilla Family, Mexican Mafia, Nazi Low Riders, Northern Structure, Nuestra Familia, and Texas Syndicate. These are the oldest and most notorious gangs. Their active “made” members are undoubtedly engaged in criminal activity in numerous ways. However, compared to the prison population as a whole, there are relatively few people at this tier. For example, law enforcement officials at one point estimated there were only about 300 official members of the Mexican Mafia at a time when the men’s prisoner population was nearly 160,000 (People, 2008; United States, 2003b).

The second tier of prison gangs, known by officials as “Security Threat Group II,” is composed of people associated with smaller, newer, and less dominant gangs and street-level organizations. This includes members of the Crips, Bloods, Norteños, Sureños, White Supremacists, and Northern Riders, among others. People affiliated with these groups are typically street gang members engaged in varying levels of criminal offending. There are two important points to make about these groups. The first is that many of these street gangs are akin to a local franchise from a larger entity (e.g., Levitt and Venkatesh 2000, 762). For example, there are many autonomous local Sureño cliques that all call themselves Sureños, but they are not working together. Moreover, they are also often rivals and go to war. Yet, when incarcerated, members of Sureño gangs are expected to drop their street rivalries and act in a united fashion. Second, many of these tier-two groups align with one of the tier-one prison gangs. For Sureños, they unite under the umbrella of the Mexican Mafia. Members of these tier-two groups typically act under the direction of the more dominant gangs, often playing a secondary role within the prison hierarchy. Some street gang members will participate in criminal activity at the behest and in coordination with their respective umbrella gang.Footnote 13 As discussed in the previous section, the Mexican Mafia can tax and govern Sureños street gang activity, so this hierarchical relationship is not limited to within prison walls.

Finally, there are many incarcerated people who enter prison with no street gang affiliation at all. Nonetheless, these people will still have to affiliate with their respective racial and ethnic groups. At the most aggregate level, this typically breaks down along the lines of Black, White, Southern Hispanic, Northern Hispanic, Paisa, Native American, and Asian. People also tend to group by their home area code or city (often called your “car”). Because the tier-one gangs act as an umbrella organization over their own race and ethnic group, affiliating with one’s racial group means that one is living under the rule of the gangs. Hence, these people are not formal gang members, but they must navigate the intricate ties between race, ethnicity, and gang association within the prison context. This association is essential for their survival in the prison social order. It is also mandatory.

As one person explains, “I’m a white guy, so I can only hang with white guys in prison. The COs [correctional officers] told me that I would only run with the white guys. I knew that going in, but they told me too” (Trammell (2012), 27). Each racial and ethnic group is expected to incorporate new members in this way. Each gang wants the other gangs to do so because it facilitates community responsibility. Unaffiliated people lack the standing through which gangs impose social control. One person complains about when this process does not operate quickly:

The black dudes were slow at training their new guys. When they come in, we have a talk and set them straight. My boys were on it; you cannot fuck around and let this slide. The Mexicans, they know what’s what and they were quick, but the black guys, that’s a different story. I had to remind them several times, you know, ‘You have a new boy on the yard’ and stuff (Trammell 2012, 28).

If gangs were merely rival armies, we would not expect to see one side encouraging the other to recruit more soldiers. Instead, the system works more smoothly if everyone is affiliated. As a consequence of this three-tier structure and the community responsibility system, most incarcerated people live under gang rule, regardless of whether one is in a gang or involved in criminal activity.

The system of criminal-civilian governance operates across a wide range of activities and relationships that impact essentially everyone in a facility. The two most important functions of criminal governance are related to policing and enforcement and to judicial issues.Footnote 14 On the former, gangs make rules that enforce property rights for their ethnic group, regulate interpersonal violence, control entry to a housing area, prohibit informing to officials, and regulate other public behavior. One important role is to regulate racial lines and interracial interactions. For example, one formerly incarcerated White person in a leadership role explains:

I was fine with the boys playing cards or dealing meth to the blacks and the Southerners ... Celling-up is another story ... It’s a lot about appearances; I really don’t give a shit about your friends. I had a guy that played serious poker and won lots of money from the black guys in prison; that was great, keep up the good work, my brother. But if he told me he was going to the blacks to cell up, then I check him (Trammell 2012 47).

Another person explains, “You’re not to mix races, you’re not to sit next to someone of a different race other than your own race, eat with someone other than of your own race, smoke with someone other than of your own race. Someone has to control this, and so our leader regulates these things (Trammell 2012, 48). Another person explains, “There are so many rules about who goes first in line for meals and who gets the TV first. If you follow all these rules, you end up doing easy time. I was a con which means I follow the code so you have to know the rules and you have to teach the new guys how to be a con and follow the rules” (Trammell 2009, 756). Prison gangs decide, supervise, and enforce these social rules.

The gang-based system also monitors and controls entrance to a housing area. Gang leaders inspect the prison paperwork of new arrivals to find out if they are sex offenders, enemies of the gang, or people in good standing. As one guy explains, “Everyone comes in with a jacket, it’s like your paperwork that tells what you did, and so you have to show it to everyone, and if you don’t then we all know what’s up” (Trammell 2012, 112). Gang shot-callers do this for all new arrivals, not just those with a gang affiliation or someone who is involved in the drug trade. If someone is convicted of a sex offense, for instance, then the gang leader will arrange for him to be assaulted (Trammell & Chenault, 2009). This will often happen repeatedly if the person does not immediately request officials to move him to protective custody. In this way, gangs curate the population of the housing unit.

A second major function within gang rule is judicial. Gangs play a vital role in resolving disputes and punishing bad behavior. For instance, one person explains, “That’s why we have shot-callers so when a couple of idiots get into it in the yard, instead of letting them kill themselves, the shot-caller goes out and works it out. He talks to these guys and finds out what happened, who did what to who, it’s very simple” (Trammell 2012, 47). Another person explains:

The boys inside, they follow the rules and that means you work with your own boys and do what they say. Look, there is a lot of problems caused by the gangs, no doubt. The thing is, they solve problems too ... you don’t start a fight in the yard and stuff. Gangs are a problem but we took care of business. There is a code of silence, you don’t talk about all the stuff with others, the cops split up gangs if there’s a big problem so we keep to ourselves and mind our own business. (Trammell 2009, 755).

Conflicts do not only arise from the drug trade or the underground economy. They also occur because of simple interpersonal issues. Maintaining respect in prison is a high priority for most people, so when someone disrespects you, it is a serious issue. One person explains what happened when he first arrived to prison and did not yet know how the system works:

When I first got to prison, I said some shit to this white guy and the next thing I know, I’m told to make it right with him. I have to man up and take care of my shit. At first I thought, you gotta be kidding me. No way am I going to tell this guy that I’m sorry. Then they told me that I have no choice. That’s the rule, you do what you’re told. They made a very good argument about how I need to fall in line. Okay, so I made things right (Trammell 2009, 766).

What “they made a very good argument” means here is a bit ambiguous. It might genuinely be that they explained how the rules work and that if he doesn’t apologize, then he might be attacked by the other group. Or, they might have threatened to assault him if he did not. Another person describes what happens if no apology is given. He notes, “Usually, you settle the dumb stuff there. If someone disses me or someone takes my stuff then the leaders tell us to take it to the cell. We slug it out and get things taken care of” (Trammell 2009, 760-761).

Gangs govern noncriminal activities because it reduces the chance of large-scale disruptions and serious violence, both of which undermine their ability to profit in the underground economy. One gang leader explains,

We need to keep the boys in line. If one of our guys is a hothead or something and is always shooting off his mouth it can get everyone into trouble. We don’t want a lockdown, we don’t want a riot so I’ve had to beat down my own guys to control the bigger picture. If one of my guys is messing up then we either offer him up to the other guys or we take him down ourselves (Trammell 2009, 763-764).

There is obviously a great deal of illicit and criminal activity in prisons, and gangs are active participants in much of it. However, many of the most prominent, everyday activities of incarcerated people’s lives are also heavily and closely controlled by a system of criminal-civilian governance. Gang rule impacts where one lives, who one cells and eats with, where one can stand in the prison yard, who can use public property, and when, how, and who you can speak to. Gangs enforce norms around interpersonal and interracial interactions. Taken together, gang rule has a dominant effect on incarcerated people. They describe it as “gang politics” or “prison politics” because the system of gang rule fills a foundational role in the control of violence, regulation of social space, and adjudication of disputes for nearly all prisoners, both those actively engaged in crime and those who are not.

6 Conclusion

In this article, I have argued that political economy approaches provide strong explanations for why criminal governance emerges and what shape it takes. In this way, prison gangs are rational outcomes that “make sense.” Given the constraints that incarcerated people in California face—including a prison system that lacks sufficient resources, has limited administrative capacity, and inadequate official governance institutions—prison gang activity is understandable (Skarbek, 2016, 2020a).

These constraints are even more pressing because of the scale and scope of mass incarceration. An important explanation for the rise of mass incarceration in California and across the United States was the strikingly fast rise in violent crime in the 1960s, which stayed at a high level for thirty years. This was likely sparked by rising levels of concentrated urban poverty, resulting from people migrating to suburbs, employment opportunities dwindling, city tax coffers emptied, and physical and social infrastructure deteriorating (Clegg & Usmani, 2019). As levels of violent crime soared to unprecedented heights for those living through it, voters and politicians turned to increasingly punitive tough-on-crime policies.Footnote 15 Today, most voters know little about what happens to America’s incarcerated population. As a result, politicians in California have little incentive to reform prison governance. The costs of prison gang activity and violence are likewise not born by elected officials, and most state actors can only very indirectly be connected to those costs. The nature of a state prison system is also rife with externalities. All state residents pay for prisons, but only local voters pay for police, so this “corrections free lunch” leads local political decision-makers to lock up too many people and hire too few police (Zimring and Hawkins 1993, 211-215). Moreover, special interests have had a huge effect on expanding the criminal justice system (Page, 2011). The California Correctional Peace Officers Association spent decades lobbying for more prisons, tougher laws, and longer sentences. As a highly concentrated lobbying group, they deftly engineered a massive expansion in prison use. As a result, gangs rule in the face of the government failure that is the California prison system.

For research design reasons, it is difficult to evaluate how prison gangs in California have affected overall outcomes, such as levels of violence. Imagine an example where a study finds a cross-sectional correlation between being in a tier-one or tier-two gang and higher exposure to violence. This might suggest that gang involvement leads to more violence. Alternatively, it could be that individuals prone to violence are more likely to join gangs. In this scenario, it is also possible that someone who joins a gang might have experienced even more violence if they were not part of a gang. For example, we know that gangs often exert pressure to prevent the use of violence by their own members (Cook et al., 2007; Levitt & Venkatesh, 2000). Moreover, even longitudinal studies indicating a link between a rise in gang membership and increased violence could be influenced by various other factors and omitted variables. For instance, the formation of gangs in California’s prisons was a response to surges in the size and ethnic diversity of the prison population (Skarbek, 2012). In environments where prison populations are growing, gang affiliation could indeed elevate the risk of violence compared to periods with fewer prisoners and no gangs. However, this doesn’t reveal if levels of violence are higher or lower than they would be in large prison populations where gangs don’t exist. The difficulties in assessing prison performance because of this research design challenge is often overlooked in criminological studies (Byrne & Hummer, 2007; Bruinsma, 2016; Pyrooz & Decker, 2019). Better theory and design are essential to more carefully evaluate these types of relationships.

Still, there are a number of evaluative claims that we can make based on observational data focusing on outcomes related to (1) effectiveness, (2) adaptability, resilience, and robustness, (3) equity, (4) accountability, and (5) general morality (Ostrom 2009, 66-67). On effectiveness, gangs are surprisingly successful at getting contraband inside prison walls. They have also devised a wide-range of clever and relatively effective rules, constitutions, and procedures to protect property rights and to adjudicate disputes. In this sense, it is another example of the successful spontaneous emergence of market activity in prison (Radford, 1945). The system of gang-based governance is also remarkably adaptable, resilient, and robust. Decades of investigation and suppression by correctional officers and gang investigation units has done little to eliminate their presence. This is even more impressive given the selection bias in who goes to prison, with people being disproportionately from socioeconomically disadvantaged backgrounds. Incarcerated people often lack resources to promote collective action. There is also a selection effect that brings less cooperative and trustworthy people into the community. In addition, many gang leaders are isolated in solitary places miles removed from the people and places they govern. Nevertheless, the system of gang-based governance is able to incorporate with ease first-time residents who have little knowledge of the inner-workings of this extralegal system, and there is little that officials can do to stop them.

On the other hand, prison gangs’ effectiveness in promoting the underground economy undermines the operation of the prison in several ways. First, the prison economy is plagued by negative externalities. For example, correctional officers spend time and effort to battle it, often become corrupted participants themselves, and get hurt attempting to break up fights and riots over illicit deals gone bad. Likewise, prisoners uninvolved in these illicit transactions often get forced into such battles, or face punishment themselves for refusing to. Imperfect information, insufficiently secure property rights, and liquidity constraints are pervasive. Countless acts of violence against prisoners are the direct result of drug debts. Second, prison gang membership undermines rehabilitation and increases recidivism (Dooley et al., 2014). Gang membership often reduces law-abiding human capital and social capital, making legal employment more difficult to find after prison. For example, having a racist gang tattoo on one’s hands, neck, or face makes it difficult to find a good job. Such tattoos might be helpful or even necessary to reach the free world, but it also leaves a person severely disadvantaged once free.

In terms of equity, prison gangs wield tremendous power but deviate — often substantially — from widely agreed upon principles of the rule of law (Bingham, 2011). The rules created by gangs are often applied with substantial discretion rather than in a direct, transparent, and objective way. For example, friends of gang leaders might be able to violate gang rules, while leaders apply the rules strictly to others. Some people also have different and greater rights than others do. Some incarcerated people, such as sex offenders and former law enforcement officers, have essentially no rights at all in this system. Gang rule has little to no due process, where people can face their accusers, see the evidence against them, and present evidence of their own. In fact, most people will not even know that they had been “on trial” and “convicted” by a prison gang until its too late.

Likewise, gang rule lacks robust systems of accountability. For legal systems in many states, when a person receives a death sentence, his case is immediately sent to an appeals court for review. The severity of the punishment demands further scrutiny. By contrast, there are no appeals procedures when gangs hand down their own death sentences. Finally, we should be especially cautious in assessing social outcomes when people have few exit options or choices (Holcombe 1994, 72-91).Footnote 16 Coercion permeates many aspects of prison life, whether from the hands of the state or incarcerated people. Residents are forced to go to prison, forced to live with other people, forced to affiliate with gangs, and often forced to fight—and possibly take a life. Unlike some of the classic cases of self-enforcing exchange mentioned in the introduction, this is the dark side of private ordering (Milhaupt & West, 2000).Footnote 17

This broader, more multidimensional evaluation should also push us toward greater interdisciplinary and multi-method forms of scholarship in the study of criminal governance. Since its beginning, public choice has always been an interdisciplinary research program, and not merely economic imperialism (Amadae 2003, 133-156). In addition to working across disciplinary boundaries, public choice also has a long record of using quantitative and qualitative research methods in tandem (Ostrom, 1990; Poteete et al., 2010). Qualitative evidence expands the type of evidence available for studying criminal governance. It helps us to identify causal mechanisms in addition to causal effects. In many instances, qualitative evidence is simply better suited for more fully observing and describing multifaceted institutions of criminal governance (Skarbek, 2020b). Given that millions of people live under criminal rule today, studying the political economy of criminal governance has never been more important.