Observations

The limited sample of case-studies necessarily renders the comparative find­ings preliminary; they are meant to stimulate further empirical and compara­tive work rather than provide definitive answers.

3.1 Principal Public Security Challenges

Security challenges in Canada, Germany, Spain and Switzerland tend to be latent (in the sense of unintentional and non-criminal). The USA faces a moderate risk, largely as a result of rates of violent crime that are above aver­age among democratic federations. Four federations in the sample face high risks: Brazil and Mexico from drug-related homicides and organized crime, South Africa from disproportionate rates of murder, rape and robbery that are exacerbated by racism and socio-economic conditions, and India from high levels of homicide and persistent political violence in different parts of the country (unodc 2019). Most of the countries in our sample confront the threat of transnational terrorism. India, by contrast, faces regular challenges from domestic, international state-sponsored and transnational terrorism. Given this panoply of challenges, theory would predict a more centralized approach to security in India and, to a lesser extent, South Africa, than among the remainder of the sample. Analogous dynamics informed the centralization of security measures in many countries in our sample in the aftermath of 9/11, resulting in a recalibration of self-rule in favour of shared rule with respect to public security.

3.2 Constitutional and Legal Frameworks

Countries' constitutional and legal foundations for legislative competencies and the provision of public security in the federal polity, including the divi­sion of powers along the centralization-decentralization continuum seem to be a function of both path-dependency and diversity. In particular, the dis­tinct legal histories of each country case-study have a path-dependent impact on contemporary policy choices as mediated by current constitutional and legal frameworks (Leuprecht, Kolling and Hataley 2015).

Heterogeneous federations tend to centralize public security more than homogenous ones, likely because of a latent threat that heterogeneity is per­ceived to pose to the territorial integrity of the federation as a whole. India and South Africa stand out for their high level of centralization on matters such as policy, investigation, arrest and public order. In India, this is manifest in an emphasis on the Indian Union as opposed to its federal aspects, seemingly driven by a concern about territorial integrity, heightened by the experience and aftermath of partition. Indian states have just a few formal powers over public security, although, constitutionally, they have control over ‘police' and ‘law and order'. In fact, the Indian constitution allows the president to dis­miss and dissolve constituent units' elected governments in emergency cases and bring them under direct central rule, as has happened in the Punjab, the North-east and Kashmir. Since these constituent units abut an international border, it stands to reason that concerns about territorial integrity continue to outweigh their autonomy in the delivery and administration of public security.

In South Africa, the centralized approach to security is a lag effect of the centralized security apparatus under the apartheid regime. Nowadays, the provinces are limited to oversight, consultation and social services; munic­ipalities, by contrast, have limited powers of arrest, along with control over policing traffic and by-laws. In all cases, suspects are handed over to national authorities for detention and prosecution. Mexico and Brazil show similar pat­terns of centralization, but to a lesser degree than India and South Africa. In Mexico, the legacy of decades of authoritarianism continues to shape public security. While the constitution adopts a balanced, decentralized approach, the Mexican regime is more centralized than it appears on paper. A similar sit­uation can be found in Brazil. Constitutionally, the federal government's con­trol over security in constituent units is actually relatively weak, but de facto it has control over powerful civil and military police forces and intelligence agen­cies, which it is at considerable liberty to deploy because of areas of shared jurisdiction with the constituent units.

In Spain, the legacy of decades of authoritarianism prompted a gradual process of decentralization. Today the State of Autonomies is characterized by a latent federalization and heterogeneity between the different constitu­ent units. Those with a long history of some degree of linguistic and cultural autonomy, notably the Basque Country, Navarre and Catalonia, have been especially proactive in demanding their own security forces. Regaining the autonomy over security that these regions had traditionally enjoyed as a result of the decentralization process seems to have had a contagion effect. Other Autonomous Communities also sought a transfer of responsibilities for pub­lic security. However, the economic and financial crisis (2008-13) halted their ambitions. Despite this strong decentralization trend, central government has maintained exclusive jurisdiction over public security, and both national police forces, the civil guard and the national police, are responsible for the devel­opment of security functions and the prevention and repression of crimes in rural areas and larger urban centres.

By contrast, centralization in matters of public security among the sam­ple of homogeneous federations is relatively low. The German arrangement is very decentralized—right down to domestic security intelligence agencies under the purview of each Land—but happens to be structured in such a way as to produce symmetrical outcomes. Legislative capacity and implementa­tion relating to public security resides largely with the Länder. This is less by intentional institutional design than by the strategic interests of strong local bureaucracies, and predates the current constitutional arrangement by several decades, autocratic interludes in the first half of the 20th century notwith­standing. Competencies of the federal government are limited to overcoming collective-action problems, such as might arise in national criminal investiga­tions, counter-terrorism or the armed forces.

Insofar as Switzerland hitherto lacked a federal police force, it is even more decentralized than the German system, but has been subject to some centralizing security trends in recent years. In fact, the Swiss system takes the opposite approach to Germany: only those powers clearly vested in the federal authorities by the federal constitution, requiring a double majority by voters and cantons, are areas of federal competence. Switzerland's approach relies heavily on cooperation among the cantons, which enjoy a substantial degree of autonomy, particularly in matters of public security.

Federal countries with vast territories, such as the USA and Canada, have, historically, taken a more decentralized approach to matters of public secu­rity. However, in the USA that mindset is driven by scepticism of government in general, and of central government in particular, as exemplified by Posse Comitatus,^[55] which explicitly restricts the involvement of the armed forces in domestic law enforcement operations, along with other limitations on domes­tic operations by organizations such as the Central Intelligence Agency and the National Security Agency.

Although constituent units in Canada, the USA, Germany, and Switzerland enjoy considerable autonomy over public security, all four countries have seen sub-state autonomy compromised in the aftermath of 9/11 through anti­terrorism legislation with latent centralizing tendencies, including the cre­ation of federal police forces in Germany and Switzerland and more sweeping powers for security services in general and federal police in particular in the USA and Canada.

3.3 Public Security Operations

We now turn to the way each public security system actually works in practice, whether it is plagued by overlap and duplication, how and why it has changed over time, and where it sits on the symmetry-asymmetry spectrum.

Asymmetric approaches to public security may be more tolerable where public security is already inequitably distributed (for example, in India and South Africa). Moreover, asymmetry may actually be necessary to achieve just outcomes under certain conditions, including linguistic, cultural, ethnic or national diversity (India, South Africa and Spain). India—and, to a lesser extent, South Africa—is somewhat of an outlier in this regard: large and hyperdiverse, with a highly centralized approach to public security and little tolerance for asymmetry in its delivery. Public security in India is a task neatly divided between the union government and the states, but with overlapping and conflicting responsibilities. Despite a clear division of power, the consti­tution authorizes the central government to intervene in public security.

Brazil and Mexico face a dilemma in administering and delivering secu­rity: central intervention is driven by insufficient security capacity at the local level, yet calls into question the ability to bolster local capacity and reinforce local jurisdiction in the area, thus precipitating suboptimal outcomes for the federal arrangement as a whole. In Brazil public security has long been under­stood almost exclusively as protection against armed violence. However, a recent shift in public security policies has gradually been replacing militarized approaches based on repression with policies based on the rights of citizens to access services. This shift has begun to alter relations within the country's federal system of government, in the sense that the past strict focus on central­ized policing is slowly giving way to a more comprehensive understanding of security through education and health.

Mexico's public security crisis has given rise to tensions within the current arrangements of the public security system, leading to a re-centralization of authority and resources at the federal level. Recent major reforms seek to reshape the distribution of power and the governance of Mexico's pub­lic security system. It is a volte-face from reforms in the 1990s promoting decentralization and multi-level governance, which had been triggered by a democratization process after decades of a federation characterized by a strong central government and dominated by a single hegemonic political party.

In South Africa, institutions concerned with public security are located mainly at the national level, but fail to meet the challenge. The national gov­ernment's response to crime has focused largely on law enforcement, while neglecting primary, secondary and tertiary crime prevention. Against this backdrop, the provinces play a very limited supervisory role over the South African Police Service, although metropolitan municipalities are increasingly complementing the national force's efforts through their own metropolitan police forces.

The US case shows the perils of not finding an adequate equilibrium between shared rule and self-rule, and the disjunction between federal and local security priorities. In this sense, the challenges the founders faced in the 1700s are not so different from today's: how to balance security threats and the preservation of national unity with respect for autonomy among individual states, and protection of civil liberties and civil rights, especially given how decentralization has enabled past and present abuses of human rights in some states. Historically, public security concerns have trended towards increasing the role played by the federal government, at the expense of local communities and states.

Canada's asymmetric, decentralized approach to public security is emblem­atic of a system of local control that is flexible enough to meet diverse demands, yet centralized enough to benefit from federal support in times of need. An equilibrium of de/centralized service delivery makes possible relatively stan­dardized, albeit asymmetric, service delivery in day-to-day operations and during times of disaster or emergency relief (Leuprecht and Kasurak 2020). The functionality of the Canadian public security model is premised on local engagement and accountability. Apolitical federal bureaucratic coordination encourages and supports national standards and provides surge capacity in times of crisis. However, the multi-level provision of public security produces challenges for intergovernmental management and coordination, notably the difficulties of shared sovereignty in security governance. These are manifest in coordinating priorities and disconnects that arise when much of the emer­gency service response capacity resides with municipal and local governments, whereas jurisdictional authority is vested in provincial and federal authorities. The Canadian case demonstrates that, contrary to widespread belief, not only can decentralized and asymmetric delivery and administration of security yield effective and efficient security outcomes, but they may in fact be a pre­condition of such outcomes in large, diverse societies.

The German federation has experienced the creeping centralization of security in a country whose security structure was intentionally designed, through decentralization, to curtail the central government's purview of security. Similarly, security has also seen (limited) centralization in the Swiss federation. The Spanish case illustrates the problems of inefficiency to which duplication and lack of coordination give rise, in a polity that decentralized very rapidly and where the division of powers and labour between central and constituent governments with respect to security are often unclear and com­petitive. Owing to the decentralization process over the past 30 years, the pow­ers of both national police forces overlap with those of the constituent units' security forces.

Switzerland's pluralist cooperative federal system reflects diversity in unity of languages, religions, cultures and traditions. Swiss security policy is gener­ally perceived as adequate, but with significant weaknesses: law enforcement at the federal and cantonal levels is insufficiently staffed to meet the grow­ing demands posed by legal and illegal immigration, trafficking, crime, emer­gencies, and so on. At the same time, the division of competencies between federal and cantonal authorities is not always clear-cut and thus can occasion uncertainty and conflict between jurisdictions.

The nature and territorial concentration of distinct linguistic, cultural and ethnonational groups is likely to influence the need and extent of asymmetric public security arrangements. Irrespective of the degree of asymmetry, how­ever, decentralization emerges as a key determinant of the legitimacy and efficacy of public security. The institutional logic that informs this finding is a function of territorial differentiation of security priorities, values and interests across constituent units on the one hand, and subsidiarity on the other.

3.4 Intergovernmental Relations

As for intergovernmental mechanisms, their effectiveness, and whether they set appropriate incentives for cooperation, our sample divides into cases where intergovernmental relations in public security are heavily politicized and those where they are less politicized. Politicization seems to be determined more by institutions than by political development. Relations in Brazil, India and South Africa are heavily politicized; in Mexico less so. In Spain, dynamics between police forces are highly politicized and no formal mechanisms of cooperation among the different police forces have been established. The South African police is heavily influenced by a federal government controlled by the African National Congress. It has been politicized, it is demoralized by corruption and its public order policing is ill-equipped and inadequately trained to deal with public disturbances. The National Prosecuting Authority, too, has been politi­cized, and its success rate is declining. The court system is backlogged, and the national Department of Correctional Services has been reduced to warehous­ing a large and growing prison population. Moreover, intergovernmental coor­dination is limited and ineffective. However, growing collaboration between the South African police and municipal forces shows that the relationship may become less politicized.

In India, security rhetoric has legitimized strong central federalism, much to the chagrin of the constituent units—and to the detriment of intergov­ernmental cooperation. However, politicization is not so much a function of federal political parties as of minorities who resent growing oppression by the central state and its security agents. In Brazil, regional elite interests and concomitant corruption reinforce the paternalism and clientelism lingering from its oligarchic past. Although not partisan per se, these dynamics compli­cate intergovernmental relations.

In other words, politicization of intergovernmental relations tends to be driven by grievances over the use, provision and resourcing of public security at the level of constituent units and the lack of institutionalized cooperative bodies at the vertical and horizontal level. But politicization takes different forms: in South Africa especially, and to a lesser extent in Spain, it is driven by partisan politics, unlike in Brazil and Mexico. Ergo, partisan politics are neither a necessary nor a sufficient condition for the politicization of intergovernmen­tal relations. Conversely, however, the absence of such partisan politics may not be a sufficient condition, but appears to be a necessary condition for objec­tive intergovernmental relations. As in Brazil and Mexico, the administration and delivery of public security in Canada, Germany and Switzerland are largely above the fray of partisan politics. In the USA local sheriffs are elected officials, which can place them within the political arena.

The Canadian and US experiences in particular suggest that lapses in pub­lic security are prone to exploitation for partisan gain by local and opposition politicians. Similarly, ideology has an impact on intergovernmental relations, as some parties see more or less of a role for the state in public security. In Germany and Switzerland, for instance, strong horizontal and vertical institutional structures are pivotal to intergovernmental relations and often see relatively little political involvement. Swiss intergovernmental relations on public security seem to have benefited from the opportunity to mature over a long time without significant disruption, accounting in part for their strong institutionalization within a robust rule-of-law framework. Germany's Conference of Interior Ministers is a highly institutionalized and effective intergovernmental mechanism, but its name belies the fact that its primary purpose is to operate as a vehicle for horizontal and vertical coordination and cooperation among the bureaucracies of the Länder and the federal government.

Intergovernmental relations in Mexico are equally institutionalized, but mechanisms such as the National Public Security Council, the Attorneys General Conference, the Minister of Public Security Conference, the Prisons Conference and the Municipal Public Security Conference operate primar­ily at the political rather than the bureaucratic level. In Canada, by contrast, intergovernmental relations on matters of public security tend to be fairly ad hoc and premised on political involvement, with bureaucracies usually acting at the behest of politicians rather than on their own initiative. The USA and Spain are outliers: formal vertical intergovernmental institutions are all but non-existent, and horizontal ones are ad hoc, and often bilateral, with rela­tively little sway.

4