Racing all over the place: A dispersion model of international regulatory competition

Introduction

What is the effect of economic globalization on levels of government regulation across countries? There is a large body of literature, both within academic International Relations (IR) and in popular discourses, about globalization and regulatory convergence. Opponents of globalization fear regulatory races to the bottom (e.g. Leonard, 1988; Rodrik, 1997; Rudra, 2002), while some proponents argue that globalization can actually put upward pressure on regulatory standards (e.g. Garcia-Johnson, 2000; Saikawa, 2013; Vogel, 1995). Still others argue that there is a regulatory race to the middle rather than to either extreme (DeSombre, 2008). The common theme throughout these arguments is the assumption that globalization exerts similar pressures, be they toward lesser or greater regulation, across countries. Meanwhile, empirical studies suggest that while there seems not to be a generalized problem with regulatory races to the bottom (Drezner, 2001a), regulatory havens cause some real problems (e.g Clapp, 2002; Palan, 2002). However, these studies do not identify the relationship between these havens and more generalized regulatory patterns.

I argue that globalized economic exchange in a world of sovereign states generates patterns of regulatory dispersion, particularly when regulations themselves are targeted primarily at international commerce. Classic examples of this kind of regulation can be found with regulatory offshoring, when economic actors are explicitly looking for ways in which to decrease their regulatory burdens, and potential host states are explicitly looking to attract their business (Drezner, 2001b; Palan, 2003). These sorts of situations should be prime examples of regulatory races to the bottom: if economic actors are looking to find lower levels of regulation than can be found at home, there should be a generalized pattern of downward pressure on the regulatory levels offered by countries that are attempting to attract their business. However, regulatory races to the bottom rarely happen. Instead, we often see patterns of regulatory differentiation, with some states finding niches as regulatory havens, but often with an overall tendency to upward drift in regulatory levels over time.

Two examples illustrate this pattern, one from the realm of financial regulation and the other from the realm of shipping regulation. In the financial example, a number of countries, usually relatively small ones, host offshore financial industries that service individuals and corporations looking to minimize their tax burdens. Within the more established and successful of such offshore financial hosts, there has been a pattern of increasing levels of regulation over time, particularly recently. Meanwhile, offshore financial jurisdictions are more likely to compete by finding new market niches than by simply undercutting incumbent jurisdictions on regulatory costs. A similar pattern can be found in international shipping regulation, in which ships often flag with, and are therefore regulated by, flags-of- convenience rather than by home-port countries. Flags of convenience compete with each other by establishing niches to attract particular parts of the shipping industry. The more successful flags have often increased their regulatory levels over time, and new providers of flags of convenience tend to look for open regulatory niches rather than competing directly with the most successful registries. How can we explain these patterns?

The argument here builds on the insights of Charles Tiebout about competition between governments in the provision of government services. His model of regulatory differentiation is one of the most frequently cited in the economics and public policy literatures (Fischel, 2006; Tiebout, 1956). The Tiebout model was developed to explain differing levels of municipal taxation and service provision, and is generally used at the domestic policy level. In the literature on international regulatory competition, it tends to be dismissed as inappropriate (e.g. Murphy, 2004: 37–38; Palan, 2002: 158–159). It does, however, yield insights into international regulatory competition that can help explain the pattern of regulatory dispersion seen in both the finance and the shipping examples. Furthermore, a model of international regulatory dispersion that draws on Tiebout can help to identify the key mechanism through which global standards are pushed, and can be encouraged, upwards: the use of club goods such as market access as leverage by onshore actors, whether state or non-state.

There are three key categories of actors in this model. Companies seeking regulatory homes are looking not only for lower costs, but also for governmental services and for access to key markets. Therefore, they will tend to cluster in those offshore regulatory homes that provide the appropriate services and access for their particular product. Offshore states — those offering their services as regulatory homes — should seek to avoid competing in crowded regulatory niches, where profit margins are likely to be competed away, and to focus on niches where higher margins can be found. The third category is those actors (including both onshore states and private actors, ranging from non-governmental organizations (NGOs) to insurers) interested in increasing overall regulatory standards, and able to use market access to affect companies’ choice of regulatory home. The interaction of these three types of actors results in regulatory dispersion, with different offshore states offering different levels and kinds of regulation to appeal to specific sub-industries. It also results in overall levels of regulatory stringency that reflect the extent to which onshore states (and other actors), both individually and collectively, use market access to force standards up.

I begin with a discussion of the existing literature on international regulatory convergence and dispersion. I then review the Tiebout model, present a general overview of its successes and failures, and discuss the ways in which it has been used with respect to IR. The following section, building both on Tiebout’s work and on critiques of it, presents the core argument of the article, building a model that helps to explain the pattern of regulatory dispersion seen in internationally mobile industries. As was the case with Tiebout’s original model, the argument developed here is a heuristic, rather than a predictive, one — it shows how patterns of international regulatory dispersion come to be, rather than predicting specific outcomes. The model is then illustrated with the two cases noted earlier: shipping and finance.

Globalization and regulation: Convergence or divergence?

The existing literature on regulatory responses to globalization falls into three general categories: first, arguments that globalization creates pressure on states to move their regulatory standards in the same direction; second, arguments that globalization creates regulatory havens that predatory economic actors can take advantage of; and, third, arguments that the response of national regulatory levels to the pressures of globalization depends upon various issue characteristics. The argument presented here differs from all three in that it posits heterogeneous state responses to globalization, resulting in regulatory dispersion.

The first category — arguments that globalization creates pressures on states to move their regulatory standards toward global norms — can be labeled the convergence literature. ¹ There are two distinct, and mutually incompatible, arguments within this category. The first is that globalization will create downward harmonization, or regulatory races to the bottom. The logic is that, to the extent that globalization frees capital to seek the lowest-cost locations in which to invest, states will use lower levels of regulation to attract investment. Other states will need to respond in kind in order to prevent capital exodus, creating continuing downward pressure on regulatory levels. While the logic of this argument may seem compelling (but see Kahler, 1998), existing research suggests that it rarely, if ever, actually happens in practice (Basinger and Hallerberg, 2004; Drezner, 2001a).

The opposite argument is that globalization creates an upward harmonization of regulatory patterns. There are three potential engines of such a phenomenon. The first is rich countries with large markets using market power to impose higher regulatory standards upon international economic (primarily corporate) actors (Gruber, 2000; Simmons, 2001; Urpelainen, 2011). The second is those actors themselves seeking to harmonize the standards to which they are subject across countries. Because these actors are usually headquartered in countries with higher regulatory standards, they tend to drive harmonization toward those standards they already meet in their home markets (e.g. Garcia-Johnson, 2000; Vogel, 1995). The third is the spread of normative and discursive preferences for higher levels of regulation through the mechanisms of global governance (e.g. Finnemore, 1996; Sharman, 2008; Strang and Meyer, 1993). Such patterns of upward harmonization are better supported empirically than race-to-the-bottom arguments, although studies tend to focus on specific issues rather than general patterns (e.g. Cao and Prakash, 2010; Mosley and Uno, 2007). However, they are nonetheless unable to say much about patterns of regulatory dispersion to be found in many industries, particularly many of those that are most mobile internationally.

An interesting version of the convergence literature can be found with the argument that regulatory standards are racing to the middle, rather than down or up. This argument, made by Elizabeth DeSombre (2008) in the context of international shipping regulation, notes that while high- and low-regulation sites persist, many of the major players in the industry are converging on a middle level of regulation. To a significant degree, DeSombre is using the racing metaphor as a rhetorical tool, and is, in fact, describing a pattern of regulatory dispersion. The pattern that she describes is one that fits in well with the modified Tiebout model presented later.

The second general category of arguments about the effects of globalization on regulatory levels focuses on heterogeneity. The fact of heterogeneity across countries in some regulatory patterns has been noted in a few different contexts (e.g. Falkner and Gupta, 2009), but the way in which patterns are heterogeneous has not been addressed systematically. The aspect of heterogeneity that has been discussed most in the literature is the phenomenon of regulatory havens, understood in this context as jurisdictions providing no effective regulation. An example of this phenomenon is pollution havens (Clapp, 2002). This is an example of regulatory heterogeneity because most jurisdictions, even low-regulation jurisdictions, are not havens in this sense — the regulatory havens are the outliers, not the norms. The literature establishes that these jurisdictions are a problem for maintaining international regulatory standards, but does not establish how widespread they are, or why some countries follow this route while others do not. Nor does this category of argument address the pattern of convergence or dispersion among non-haven jurisdictions.

The third general category is arguments that regulatory patterns depend upon certain economic or political characteristics of the particular states or industries being regulated. Ronald Rogowski (2003), for example, argues that convergence will happen in response to the globalization of capital movements when countries differ in their level of development, but globalization will increase policy divergence when countries differ in their political preferences. Daniel Drezner argues that it is the political characteristics of the issue-area that are key. When interests both among great powers and between them and lesser powers converge, regulatory standards will harmonize upward (Drezner, 2007). When those two conditions are not met, standards will not converge globally. His model predicts that: when interests among great powers converge but diverge from those of other actors, there will be club standards; when interests among great powers diverge from each other but converge with those of other actors, there will be rival standards; and when nobody’s interests converge, there will be what he calls ‘sham standards’ (Drezner, 2007: 72).

Conversely, Dale Murphy (2004) argues that it is the economic characteristics of the industry that drive global regulatory patterns. He points to three factors: the locus of regulation (whether it is the production process or market access of the good or service that is regulated); the concentration or oligopolistic characteristics of the industry; and the asset specificity of investment. Various combinations of these factors in his model can lead to upward convergence, to races to the bottom, or to regulatory heterogeneity.

Both of these arguments posit that some issue-areas will generate upward convergence, while others will result in regulatory heterogeneity or divergence. But, as is the case with the literature on regulatory havens, neither discusses what drives patterns of heterogeneity when they occur. There is, in fact, little literature on patterns of regulatory heterogeneity within the IR literature. However, there is a much more developed discussion of the subject to be found in the domestic policy literature, which can serve as a starting point for discussion of the issue at the global level. One of the seminal, and still one of the most common, models of regulatory heterogeneity in a non-hierarchical political setting was first proposed by Charles Tiebout in 1956, in an article entitled ‘A pure theory of local expenditures’ (Tiebout, 1956).

The Tiebout model

Tiebout’s model is a response to the argument, made two years earlier by Paul Samuelson in an article entitled ‘The pure theory of public expenditures’ (Samuelson, 1954), that there is no way for governments to figure out the economically optimal level of public services (Tiebout, 1956: 416). Different individuals have different preferences for levels of government expenditures for public goods, and the resultant necessary levels of taxation. Governments can charge equivalent levels of taxes to individuals with different preferences, but then some people will end up paying for more services than they want, with others not getting the services they want even though they are willing to pay more. Alternatively, governments can charge taxes according to individual preferences, but if they do so, individuals will have an incentive to act as free-riders by understating their preferences. In aggregate, this phenomenon will lead to the under-provision of public goods. By this logic, there is no way for governments to use market mechanisms to establish optimal levels of taxation — they can only use authoritative mechanisms.

Tiebout allows that this is true at the national level, but argues that it is not necessarily true at the municipal level. More broadly, his claim is that Samuelson’s argument holds in situations of political hierarchy, but not in situations of sovereign competition, where governments operate independently of each other’s control. His intuition here is straightforward. At the municipal level, different towns have the ability to provide different levels of public goods in exchange for different levels of taxation. Meanwhile, residents have the ability to move from one municipality to another. As such, individuals should self-select the municipality with the appropriate level of public goods for their preferences, at a level of taxation sufficient to support those services — people who want a higher level of municipal services will move to towns with higher taxes, and vice versa. Levels of government service will have been set by a market mechanism, rather than by government fiat.

In order for this intuition to hold, however, several assumptions need to be made. Individuals need to be rational actors, they must have complete and accurate information about tax and service levels, they must be free to move among municipalities, and the process of moving must be costless. In addition, there must be enough municipalities available that all individuals can find one that meets her or his needs. The model also assumes that municipalities have a geographically inherent optimal size, and that there are no significant externalities generated by levels of service provision. In other words, it assumes that low service levels in one town do not have a negative spillover effect on neighboring towns, and high levels do not have positive spillovers (Tiebout, 1956: 419–420).

In practice, of course, none of these assumptions describe actual conditions (as Tiebout himself admitted — his model is designed as a thought experiment, rather than a testable hypothesis). Nonetheless, the model can be seen as a useful heuristic for understanding heterogeneity in levels of government service provision in the absence of a central coordinating authority, and for understanding why municipalities do not necessarily compete with each other to lower taxes by reducing the level of services. Empirical tests of the model tend to show that it explains some, but not all, of the pattern of heterogeneity in service levels among municipalities, and works best in US suburbs, where the assumptions are often closer to being met than elsewhere (Dowding and John, 1994; Oates, 1969).

The Tiebout model is widely cited in the public policy literature. Despite its widespread use in examining domestic policy heterogeneity, however, it has not had much of an impact on the study of international relations. It has been used with respect to some particular issue-areas, such as international migration (e.g. Carruthers and Vining, 1982). In the realm of international regulatory competition specifically, however, the model has not been used to describe or explain outcomes. It has often been invoked in a general or metaphorical sense in the context of havens and of regulatory heterogeneity, but it has not been applied as a model to describe specific patterns of regulatory outcomes. When scholars have discussed the model specifically in this context, they have tended to dismiss it as having assumptions that are inappropriate to the international regulatory realm, and therefore as inapplicable (e.g. Murphy, 2004; Palan, 2002).

Not all of the assumptions are necessarily problematic, and some are less problematic in the international than in the municipal context. For example, corporations looking to arbitrage regulatory regimes to minimize both regulatory costs and taxes have far greater resources with which to research their options than the average suburban homeowner. As such, the assumption of perfect information may be less unrealistic in the international than domestic context. Also, it can be both simpler and cheaper to re-register a corporation or ship abroad than to move homes. Nonetheless, some of the other assumptions that Tiebout makes are, indeed, problematic in the context of international regulatory competition. Adjusting these assumptions in a way that makes them more appropriate to the study of international regulatory heterogeneity yields a model that is different from a standard Tiebout distribution.

The international model

Three assumptions that Tiebout makes are particularly problematic for the study of international regulatory competition, and need to be adjusted to successfully model the results of such competition. He assumes: first, that municipalities have an inherent optimal size; second, that there are no significant externalities generated by levels of service provision; and, third, that individuals have an exogenous preference for at least some level of municipal services. Modifying these three assumptions to fit international regulatory politics yields a model that allows for both regulatory havens and external pressure to increase overall regulatory levels, which are not predicted in levels of municipal service provision in the original model.

Tiebout’s model assumes that municipalities have an optimal size that is an exogenous given, at which there are enough taxpayers to provide adequate returns to scale in the provision of services, but not so many as to make those services overcrowded. He uses the example of a municipal beach — the town needs enough taxpayers to pay for lifeguards and facilities, but does not want to attract new residents beyond the point at which those facilities are at full capacity (Tiebout, 1956: 419). Optimal size is thus determined by physical characteristics: the borders of the town, the size of beaches and parks, and so on. These limits often do not exist in the realm of international regulatory arbitrage. This is particularly true in the examples of offshoring discussed here. Flagged ships need never come anywhere near the flagging country, and companies registered offshore need only a nominal address (if that).

Furthermore, the nature of the regulatory niche that a state chooses to occupy need not be determined by its geography, as is the case for municipalities in the original model. States thus have more flexibility in determining the regulatory niches they will compete for than municipalities do. In addition, the limit to their growth is provided not by an inherent optimal size, but by the demand for the regulatory niche that they choose to provide. States’ flexibility in determining regulatory niches does vary across issue-areas. It is highest in areas in which the state is offering to regulate an economic activity that is not primarily taking place within its borders. This is most clearly the case with the shipping example, in which ships flagged with a state need never visit that state (in fact, some flag-of-convenience countries, such as Bolivia and Mongolia, have no coast to visit). In areas in which the economic activity does happen within the regulating state, flexibility to determine regulatory niches can be substantially constrained by domestic economic capabilities. In the case of regulatory competition to attract manufacturing investment, for example, states are constrained in their choice of niches by the quality and quantity of their labor forces.

The second of the problematic assumptions of the original model is that there are no significant externalities, either positive or negative, created by differences in levels of municipal services (Tiebout, 1956: 419). This is unlikely to be the case in practice — better policing or flood control is likely to provide positive externalities to adjoining communities, while lower levels of both generate a risk of negative externalities. The implausibility of this assumption is, in fact, one of the key criticisms of the Tiebout model even on its home ground of municipal policy (e.g. Pauly, 1970; Williams, 1966). Similarly, levels of national regulation can generate international externalities. Lower levels of regulation offshore can create environmental risk or undermine tax collection efforts in higher-regulation countries. Conversely, higher levels of regulation, and, in particular, the enforcement of higher levels of regulation, can provide positive externalities, in the form of public goods, to other countries and to the international system as a whole. ²

Other things being equal, these public goods are likely to be under-provided. However, third parties that will benefit from higher regulatory standards have an incentive to pressure regulators to increase standards and the public goods they create. One way of doing this is to restrict access to club goods, from which it is possible to exclude certain actors. ³ High-regulation states may restrict access to their economies to those economic actors that meet high standards. In the case of banking regulations, these might be standards for transparency or knowledge about clients (Drezner, 2007: 119–148). In the case of shipping, it might be the physical condition of ships (DeSombre, 2006). To the extent that high-regulation states also have large markets, access to these markets is then dictated by compliance with the standards, irrespective of lower standards in regulatory havens. Therefore, economic actors may need to exceed the regulatory standards in some host countries in order to ensure access to lucrative high-regulation markets. The existence of such regulatory club goods provides one of the key logics for demand for particular levels of regulation by economic actors looking for opportunities for regulatory arbitrage.

The existence of such regulatory club goods speaks to the third of Tiebout’s assumptions that do not translate neatly to the international regulatory sphere. In his model, individuals have exogenous preferences for municipal services, with the implicit assumption that most individuals prefer at least some services (Tiebout, 1956: 418). The equivalent in the international regulatory sphere would be exogenous preferences by economic actors — the corporations and individuals who are players in the international regulatory arbitrage game — for specific, and differential, levels of regulation. But why would these players want any regulation at all? In some cases, the logic is congruent with that in Tiebout’s model. For example, companies looking to invest in manufacturing capacity should generally choose countries with higher levels of human capital investment as their need for skilled labor increases. Higher levels of state human capital investment must (in the long run) be funded by higher levels of taxation, so there should be a Tiebout sorting by companies investing in international production according to levels of social taxation.

However, this logic does not apply to the examples discussed here, in which economic activity is not particularly tied to the geography of the regulator. If companies are flagging their ships or registering shell companies abroad specifically to avoid home-state levels of regulation, one would think that they should, other things being equal, prefer less regulation to more. This is the logic of the regulatory race to the bottom. However, other things are not equal. More stringent regulation by offshore jurisdictions can generate the positive externality of access to club goods, such as rich-world markets or financial systems. To the extent that such access is predicated upon the certification of compliance with a certain regulatory standard, then the benefit of such certification can outweigh the cost of compliance.

Access to club goods can be restricted either by the beneficial home state of the economic actor seeking the offshore regulatory home or by third parties. In turn, third parties can be either other states or other parties whose actions can have a regulatory or even a de facto market access effect, such as labor or consumer organizations. In the examples of finance and shipping, the utility of the assets offshored is affected by restrictions on the movement of money and of ships, respectively. Economic actors should choose the level of regulation at which the marginal cost of complying with host-state regulation is equal to the marginal cost of exclusion from movement of assets to or through more heavily regulated locations.

To use the shipping case as an example, for well-capitalized operators using relatively new and costly equipment, the benefit of easy access to ports in high-regulation countries (such as the US or European Union (EU)) is worth the cost of complying with environmental and safety regulation. For more marginal operators using cheaper or older equipment, it makes more sense to avoid the cost of regulation, even at the expense of more restricted geographical access (DeSombre, 2006). The same sort of story can be told for the finance case. Large corporations using offshore financial services to reduce their tax burdens will nonetheless likely want to be able to move financial assets across countries as easily and quickly as possible. Tax exiles are more likely to trade-off ease of financial mobility for higher levels of bank secrecy (Palan, 2002). Higher regulation in third countries (or by actors within those countries) thus has the effect of increasing demand for regulation by users of offshore services, even as those actors are playing a game of regulatory arbitrage, but it does so differentially across these users.

The existence of regulatory club goods, along with the need for actual government services in some cases, explains the incentive for economic actors to look for regulatory hosts that match their preferred level of regulation. It explains, in other words, the demand for Tiebout sorting in international regulation. But what of the supply? What motivates states to set particular regulatory levels to attract economic actors engaged in international regulatory arbitrage? In a word, money. Hosting offshore finance, acting as a flag of convenience, or attracting any other form of offshored economic activity is a source of income, and can be an important source, particularly for relatively small countries. This income can be direct, through taxes and fees on bank and ship registration, or indirect, through the employment effect of the economic activity. Taxes and fees alone can contribute significantly to government budgets in offshore regulators. In the Cayman Islands, for example, fees on offshore finance have contributed more than 15% of government income (Roberts, 1995: 246), and in Liberia, the ship registry has contributed more than 30% at times (DeSombre, 2006: 78). To the extent that states choose to try to attract offshored business, such as banking or ship flagging, in order to generate revenue, it can be assumed that their goal is to maximize revenue by attracting as much of the business in question as possible.

This adds a complication to Tiebout’s model. He assumes that municipalities have an optimal size, beyond which they will not try to grow (Tiebout, 1956: 419). In the case of international regulatory competition, however, bigger is often better. States are limited by the availability of domestic resources or labor when attracting investment that draws on them. However, when states are primarily offering regulatory homes, as is the case with many forms of offshore finance and is entirely the case with shipping, there is no direct equivalent to the material limits to growth in Tiebout’s model. The limits to revenue faced by host countries are provided by the limits of demand for particular regulatory niches. However, some niches are more valuable than others. Other things being equal, bigger niches should yield more revenue to states, as should niches with a higher tolerance for taxation. The interaction of these two factors determines the yield of a given niche.

Therefore, Tiebout sorting in this environment introduces contradictory impulses: host countries should want to differentiate themselves in order to attract business, but, at the same time, should want to focus on those particular regulatory niches that offer the best returns. Economic theory suggests that in a perfectly competitive market in which sellers cannot differentiate their products from those of their competitors, they will compete on price. This will bring the price charged down to the marginal cost of production. In a perfectly competitive market for offshore regulation in which the quality of provision itself does not vary, therefore, states would be unable to generate higher revenues in those niches with a higher tolerance for taxation because other states would provide the same service for less. As such, we should expect states that are trying to attract revenues through the provision of regulation to economic actors looking for opportunities for international regulatory arbitrage to try to differentiate their product from other states in the same niche. To the extent that they can do so, and thereby offer a unique regulatory product, their ability to increase revenues without losing customers will increase. A state’s success in differentiating itself from its competitors in this context can be determined by the effects on its reputation as a provider of regulation within its particular regulatory niche.

In this context, reputation matters in two ways. The first way is a reputation for competent provision of service to the economic actors buying the regulation. To a certain extent, a reputation for competence reflects a first-mover advantage. A country that already occupies a niche might already have (but does not necessarily have) a reputation for competence, whereas a new entrant into that niche must build it anew. New entrants can compete on price, but buyers should be less sensitive to price competition the greater their need for competent regulation. This makes reputation, and first-mover advantage, more important to states that occupy more lucrative regulatory niches. Also, first-mover advantage is more valuable the higher the cost of changing regulators. If a change can be done at nominal expense, as is the case with some flags of convenience, then it is more likely to be cost-effective for buyers to experiment with new regulators. If a change is more expensive (requiring, perhaps, setting up physical offices in the regulating country), then it is less likely to be cost-effective to experiment. However, beyond first-mover advantages, reputation also depends on a history of competent provision, meaning that first-mover advantages can be lost through mismanagement, or reinforced through superior management, of a regulatory system.

The second way in which reputation matters is the reputation of a regulator with third parties, in particular, with third parties that control regulatory clubs and club goods like market access. These parties are generally states, which control access to their markets, but can also include labor or consumer groups, which can, in practice, prevent economic actors that do not meet certain regulatory standards from accessing specific markets. To the extent that an offshore regulator has a reputation for enforcing a certain level of standards, economic actors subject to its regulation may well be able to gain easier access to markets and regulatory clubs than if they were subject to regulation by a state with a reputation for a low level of, or shoddy, regulation. In fact, this reputational effect is often formalized, as when the third parties that run regulatory clubs approve some offshore regulators for easy access to the clubs, and reject others. Examples of such formalization can be found both in the financial and in the shipping cases. In the financial case, the Financial Action Task Force was created explicitly as a club to threaten sanctions against countries that did not adequately enforce regulatory standards with respect to money laundering (Drezner, 2007).

In the shipping case, most rich-country governments belong to a system of Memorandums of Understanding (MOUs) that code flag states according both to their regulatory requirements and to the extent to which ships flagged there actually meet those requirements. Ships flagged in countries with a history of flagging substandard vessels are more likely to be inspected in rich-country ports, a process that is time-consuming and expensive for the ship owner. A flag state with a reputation for higher levels of regulation will therefore be able to provide better access to developed-world ports to ships flying its flag than a flag state without. Flag states can increase their levels of regulation, but since, in this case, official reputation is based on a lagged moving average of past results, there will be a significant time lag before these higher levels are reflected by better access (DeSombre, 2006).

The effects of reputation mean that the more lucrative regulatory niches are likely to be filled by single providers with established reputations, rather than being perfectly competitive. This explains the pattern in both international banking and shipping of established providers holding large shares of the market over time, and being able to effectively dominate lucrative niches (examples include Switzerland and Bermuda in finance, and Panama and Liberia in shipping), with new entrants often having considerable difficulty establishing themselves. These effects may be reinforced by returns to scale as a provider of regulatory services. To the extent that regulators need to provide actual services, such as ship inspections or banking oversight, increased scale can allow bureaucratic overhead to be spread across more clients, lowering costs per client for equivalent services.

The final question to be addressed in this discussion is that of where the most lucrative regulatory niches are to be found. The answer is that, for the most part, they are to be found toward the higher end of the regulatory spectrum. Countries with lower levels of regulation are likely to remain open to economic actors that conform to higher regulatory standards, but the opposite is not true — countries with higher standards may well be closed to economic actors that fail to conform to those standards. This means that the farther up this spectrum one goes, the fewer restrictions there are on the global movement of the assets of economic actors, be those assets physical or financial. The idea of a single regulatory spectrum, it should be noted, is a simplification. There is regulation of various aspects of an industry, and the level of regulation need not covary across these aspects. In the shipping example, regulations can affect various aspects of safety standards, as well as labor standards and environmental standards. By focusing on tight regulation in some of these areas and looser regulation in others, flag-of-convenience states can specialize in, say, oil tankers, or bulk carriers, or cruise ships. However, the broader point remains valid. Some of these niches are more lucrative than others because the total value of shipping in them is higher. And within each of these niches, there is a reputational hierarchy.

In practice, the third parties most likely to be able to enforce access to club goods to force up regulatory standards are the governments of large, rich economies, particularly the US and the EU, which constitute almost half of the global economy between them. Host countries with levels of regulation high enough to get their clients’ assets into these clubs easily are likely to be able to provide access across most of the global economy, whereas those with levels that are not high enough to do so are likely to be able to provide access only to a significantly smaller part of the global economy. To the extent that this is the case, we should expect the host countries with the best regulatory reputations in particular industries ⁴ to have some of the highest levels of regulation among offshore regulators, and to be able to charge a premium for their regulatory services. It is this ability to charge such premiums that, other things being equal, make higher-regulation niches more lucrative for host states than lower-regulation niches.

This model of international regulatory dispersion thus suggests several patterns in international regulatory competition to attract income from economic actors looking for regulatory arbitrage opportunities. These patterns are most likely to predominate when the industry being regulated does not require productive economic activity (value-added activity) within the regulator country, and become less likely to predominate the more such activity is required. Four particular patterns are key.

First, regulatory patterns among those countries actively seeking to attract offshored business will tend toward dispersion rather than convergence. This is in contrast to patterns of regulation among countries seeking primarily to regulate domestic economic activity, which may well harmonize over time. The dispersion of regulatory offerings results from attempts by those states seeking offshored business to establish strong positions for themselves within particular regulatory niches, allowing them to charge premiums over what a perfectly competitive market for regulators would allow.

Second, first-mover advantages and, where applicable, increasing returns to scale mean that, other things being equal, the most established regulators within a particular industry will occupy the most lucrative regulatory niches. This pattern can be reinforced by a history of sound management of regulation by first-movers, and can be undermined by ineffective management and regulatory practices by regulating states.

Third, the most lucrative regulatory niches will generally be those at the most heavily regulated end of the market for offshore regulation. Regulation in these niches will be significantly lighter than in states that regulate primarily for domestic purposes; otherwise, there would be no gain to economic actors in international regulatory arbitrage. However, regulations will be sufficiently tight to allow for easy access to most international markets. Meanwhile, states entering particular markets for offshore regulation will generally be forced either to do so in the least lucrative niches or to compete head-to-head with established players. Even if particular offshore regulators leave these niches, either by moving to more lucrative ones or by ceasing to offer offshore regulatory services, the niches themselves will remain, and will be open to new entrants.

Finally, fourth, the distance, in terms of the stringency of the regulations in question, between onshore standards in the largest markets and the standards in the most lucrative niches of the market for offshore regulation is determined by the effectiveness of third parties in using access to club goods as leverage to force up regulatory standards. These third parties are generally developed economies, either individually or cooperatively, and, less frequently, other economic actors. The greater the extent to which these actors are willing and able to exclude from profitable activity those economic actors who do not meet certain standards, the greater the extent to which those standards will be reflected at the top end of the offshore market. Upward movement in regulatory standards across the offshore elements of an industry, therefore, depends upon increases in either the willingness or the ability of third parties to use this leverage.

The two cases noted earlier — offshore finance and ship flagging — illustrate these patterns. In both, we see offshore regulators specializing in specific regulatory niches, with the most successful incumbents generally being those established relatively early, and new entrants in smaller niches. The higher-margin ends in both cases tend to be those that meet the standards of exclusionary clubs, which, in turn, are made up primarily of states, but also include other international actors. Finally, both cases display an upward movement in standards in the most successful incumbents when the major market powers — the US and EU — choose for exogenous reasons that international standards in the relevant industry have to improve.

Shipping

Ships that travel internationally are required, by international maritime law, to be registered with a sovereign regulator. However, this regulator need not, in practice, be a country that the ship visits, or one from which it is owned. It can be any state that is willing to register the ship. The state of registry of a ship is called its flag state (because the ship flies the flag of that state), and when states offer to register ships that need neither visit them nor be owned by their nationals, they are called flags of convenience. It should be noted that the existence of flags of convenience is not an inevitable outcome of economic globalization. That ships need a sovereign home in the first place is a historical quirk. Until relatively recently, the sovereign home was generally the state in which the ship’s home port was located (e.g. Boczek, 1962). This is still the case with commercial air transport — airplanes must generally be registered in the state in which the airline that operates them is headquartered (e.g. Chang et al., 2004). Flags of convenience as we now know them evolved between the World Wars, with the encouragement of the US government, with the explicit intent of lowering shipping costs.

Flags of convenience are an excellent example of pure regulatory offshoring. Ship owners flag for the express purpose of finding a regulatory home that best suites their needs, and the ships themselves need never even visit their flag states. Reflagging is generally both cheap and easy relative to the cost of moving other offshore activities to new regulatory hosts. Since ship owners need nothing physical from the flag state other than (literally) the flag, and since owners avoid registering ships in their home states to avoid regulatory costs (as well as taxes and fees) there, the logic of regulatory races to the bottom suggests that owners should choose flag states that do not impose standards. And some do. However, most flag their ships in states with regulatory standards that, while significantly lower than those of most Organisation for Economic Co-operation and Development (OECD) states, are much higher than those at the regulatory bottom. For example, the International Chamber of Shipping and International Shipping Federation do a rough assessment of the regulatory performance of flags of convenience, scoring them as positive or possibly negative on 18 indicators. Some flags of convenience, such as Bolivia, score more than a dozen negatives, but of the top three flags by registered tonnage, Panama received only three negatives, Liberia none at all, and the Bahamas only one (ICS and ISF, 2012).

As noted earlier, regulatory standards for ships can be divided into three categories: environmental, safety, and labor standards. Elizabeth DeSombre rates flags of convenience according to the percentage of international agreements with respect to each category that they have signed, and finds that there is no clear pattern (DeSombre, 2006: 41–50). The biggest flags of convenience tend to be near the higher end of the scale for flags of convenience, although much lower than rich-world registries. However, different registries focus on different categories of agreements, depending upon the niche in which they specialize. Panama, as the biggest registry and therefore aiming for the biggest niche, shows a relatively even pattern with respect to environmental, safety, and labor standards. It is more prone to signing labor agreements than the other two because it hosts many bulk carriers, which depend more on dockworkers than other kinds of ships and are, therefore, most sensitive to labor clubs. Liberia specializes in oil tankers, and shows higher environmental and safety standards, but much lower labor standards. The Bahamas, with a much higher proportion of cruise ships, has the lowest labor standards, and has environmental and safety standards somewhere in between Panama’s and Liberia’s.

The hierarchy of flags of convenience suggests a significant first-mover advantage. The largest (of 32, by a count of the International Transport Workers Federation (ITF); ITF, 2010), Panama, was the first. The second-largest, Liberia, was the first to specialize in tankers, and still dominates that end of the industry. However, being among the first is not sufficient for success. The second-oldest flag of convenience, Honduras, is barely in the top 20 by registered tonnage, and has fairly low regulatory standards (DeSombre, 2006: 121–124) — it specializes in a low-regulation niche, which attracts less business than some higher-regulation niches. A key distinction between it and the two more successful registries are that the Liberian and Panamanian registries are operated by professional managers, at arm’s length from their respective governments. The Honduran registry is operated directly by the government (e.g. Winchester and Alderton, 2003: 186). This suggests that a reputation for competent management does affect the ability of flags of convenience to attract business.

The flags-of-convenience case is somewhat complicated by the phenomenon of international registries. Some states, generally from the developed world, have two ship registries. Their traditional registries require that ships meet domestic regulatory standards, including labor standards (such as minimum wages); the international registries, aimed at international shipping only, require that ships meet some but not all domestic standards. These international registries generally have safety and environmental standards that approach those of the first registry, but lower labor standards, allowing ships to register with a home flag and still be able to pay ship workers international rather than local wages (DeSombre, 2006: 81–83). Overall, these registries tend to have regulatory standards above those of the big flags of convenience, but below developed-country first registries. In other words, while the biggest flags of convenience tend to have the highest regulatory standards among offshore flags, they tend not to have the highest standards when international registries are taken into account. International registries are a special case, neither domestic nor purely offshore, which explains their peculiar regulatory patterns.

Shipping is also an interesting case of regulatory dispersion because while some of the key clubs that push up offshore regulatory standards are run by states, not all of them are. The key point of regulatory control for non-flag states is port access. The largest developed states, including the US, the EU, and Japan, have agreed among themselves that they will enforce generally accepted environmental and safety standards on ships that use their ports. Furthermore, they have committed themselves to inspect ships based on the regulatory records of their flag states. Since inspections are time-consuming and expensive for ship owners, it makes sense for them to flag with states that have good-enough regulatory records to prevent constant inspection. Furthermore, since regulatory records take time to catch up with improvements in regulatory practice, reputation in this context necessarily changes slowly (on Port State Control, see Hare, 1997; Özçcayir, 2001; DeSombre, 2006).

Along with this club process, know as Port State Control, there are two kinds of non-state clubs that push up flag-of-convenience regulatory standards. The first is a labor union, the ITF. This union organizes action by dockworkers to delay the loading and unloading of ships that do not meet its labor standards for international ship workers. For ships that need to meet ITF standards anyway, there is no incentive to register with flags that have significantly lower labor standards, and registering with flags with labor standards that approach those of the ITF can be a useful mechanism to signal intent to comply (DeSombre, 2006; Northrup and Scrase, 1996). The second is a process of industry self-regulation through the mechanisms of classification and insurance. While, in principle, ships need be neither classified by classification societies nor insured, failure to do either limits a ship’s freedom of action and increases the financial risk of operating it. To the extent that ships need to meet standards anyway for purposes of classification and insurance, there is no gain to owners in registering them with flags with lower standards that make the ship a bigger target for port inspections (DeSombre, 2006; Furger, 1997).

As a final note on the shipping case, Liberia provides an interesting example of the interaction of first-mover advantage and upward pressure on regulatory standards. In the 1970s, Liberian regulatory standards for tankers were quite low because there was little international pressure to force them up. However, a series of environmental disasters, beginning with the wreck of the Torrey Canyon off the coast of England in 1967, led to the formation of state-based clubs that ultimately yielded the current Port State Control system. Rather than re-register tankers elsewhere, or stay registered with the under-regulated Liberian flag, many owners of Liberian-registered tankers chose instead to pressure the Liberian registry both to increase and to enforce its regulatory standards (Carlisle, 1981; DeSombre, 2006; Perkins, 1997). In other words, standards in an offshore site were forced up by the economic actors being regulated, in response to the creation of clubs designed to put upward pressure on global regulatory levels.

Finance

The offshoring of a set of financial services ranging from banking to corporate registration can happen for a number or reasons, but primary among these reasons is tax minimization. For this reason, many (if not most) offshore financial regulatory centers are referred to as tax havens. They are effective at reducing the tax burdens of both corporations and individuals through a combination of low (or non-existent) taxation, secrecy, and ease of incorporation. While these regulatory foci, particularly secrecy, can abet activities such as money laundering (Blum et al., 1999), they are mostly focused on reducing taxes on legal income (in ways that may themselves be either legal or illegal in home countries, or in the grey area in between). Tax havens are not as distinct a category of offshore regulation as flags of convenience, and can cover a range of activities, including traditional money management, the creation of shell companies, and the booking of financial transactions. As such, they can be even more prone to specific market niches than ship flagging.

Ronen Palan, Richard Murphy, and Christian Chavagneux (2010), in their comprehensive treatment of the topic, stress at a number of different points that tax havens focus on a variety of different strategies, and compete with each other through the creation of market niches in which they hold specific advantages (e.g. Palan et al., 2010: 35, 135–139, 160, 183). One typology of niches they identify includes incorporation locations, registration centers, secrecy locations, specialist service providers, market entry conduits, fund managers, and tax raiders (Palan et al., 2010: 36–38). Since havens can specialize in specific niches that combine these types, the typology suggests a wide variety of possible niches, many of them filled. Havens can also specialize within niches. For example, both Bermuda and the Cayman Islands act as specialist service providers, but Bermuda specializes in reinsurance (Seessel, 2001) and the Caymans in hedge funds. Market entry conduits — offshore sites through which investment onshore is routed — generally specialize in specific geographical destinations. This is particularly true for ‘round-tripping’ sites, where domestic investment is routed through a tax/regulatory haven, because these can require specific agreements with the onshore country. Examples include Hong Kong for China and Mauritius for India.

Different tax havens can provide different combinations of low tax, secrecy, and ease of incorporation to best fit different niches, in a way that is analogical to different levels of environmental, safety, and labor regulation in the shipping case. For example, Ireland, which specializes in tax raiding, does not provide notable levels of either banking secrecy or ease of incorporation. Furthermore, there can be a trade-off between ease of incorporation and a well-regulated financial sector of the kind that is likely to attract fund management. Switzerland, a leader in this latter niche, has a reputation for prudent oversight of its banking industry, creating the sort of environment in which high net worth individuals will be comfortable parking their money. However, this oversight is expensive, and may not, therefore, be compatible with the ease of incorporation — high volumes of new incorporation may incur greater costs in new oversight than they yield in government revenue. At the other end of this particular spectrum, the British Virgin Islands are home to some 800,000 corporations, more than 20 per resident of the islands. This ratio suggests that very little oversight happens, and, in fact, the Islands’ government does not even keep track of which companies are active and which are not (Palan et al., 2010: 57).

Many niches have become quite competitive, and different tax havens have had different levels of success with their niche strategies. Some of the difference can be attributed to first-mover advantages — Switzerland, for example, got in the game quite early. However, first-mover advantage is not by itself sufficient. While some Caribbean islands have been very successful, others that started just as early or earlier, such as the Bahamas or Jamaica, have not. And some of the earliest tax havens, such as Lebanon and Uruguay, have had less success than their first-mover advantage would suggest. Part of the explanation for differing levels of success is accidents of history and location. For example, British and Crown dependencies ⁵ seem to be disproportionately successful as tax havens, particularly as registration centers and specialty service providers. Similarly, round-tripping favors adjacent or nearby territories, and European havens such as Switzerland and Luxembourg have land borders with some of the world’s largest economies.

However, a significant part of the explanation is also reputational. There are several important elements to reputation here, for political stability, good government, and service provision. Political stability matters more for most kinds of financial offshoring than it does for flags of convenience because a ship will continue to exist even if the flag state disowns it, but a shell company or a bank (and the money in it) can disappear through policy change. Successful havens that have experienced political instability, such as the Bahamas or Lebanon, have seen their franchises significantly weakened. More broadly, well-governed countries are more likely to become tax havens, and are more likely to be successful at it, particularly countries with strong and well-enforced property and creditor rights (Dharmapala and Hines, 2009; Houston et al., 2012). Furthermore, some kinds of niches, particularly those such as wealth management or entrepôt finance, require, in addition to a well-managed regulatory structure, considerable on-site financial talent. This combination is generally not found in poor countries (which cannot afford the regulatory structure) or small islands (which have populations too small to generate the pool of talent).

Overall, Palan, Murphy, and Chavagneux find that the most successful tax havens fall into three categories: mid-sized Western European countries such as Switzerland and Ireland; Asian entrepôt city-states such as Singapore; and European (primarily British) dependent territories such as the Cayman Islands, Bermuda, and the Channel Islands (Palan et al., 2010: 182). The least successful havens include Pacific island states, such as Palau or the Cook Islands, which do poorly on all fronts — they are relatively recent entrants, are far from the sources of money, and lack the requisite pool of talent.

An interesting side-note here is the international finance equivalent to international registries in the flags-of-convenience discussion. Many of the world’s primary financial centers have created facilities that mirror the lighter regulatory requirements of tax havens, without providing all of the tax benefits. Key examples include International Banking Facilities (IBFs) in the US and the Japanese Offshore Market (JOM). These allow money that would otherwise leave New York and Tokyo for regulatory rather than taxation reasons to stay in place. As is the case with international registries, these facilities can be seen as indicators of the standards that major onshore players see as most important, in this case, tax revenue rather than regulatory oversight.

Different niches in the tax-haven world demand different regulatory strategies, some of which require more regulatory effort (and competence) than others. Palan, Murphy, and Chavagneux argue that:

[c]ompetition and niche strategies should not be confused with a race to the bottom. The ‘consumers’ of tax havens — firms and wealthy individuals — are sensitive to tax and regulatory differentials when choosing where to invest, but only up to a point. Logistics, habit, and most important, political stability and reputation play at least an equal role. (Palan et al., 2010: 160)

Nonetheless, there are at least two key areas in which onshore actors and clubs have an interest in forcing up standards: in combating crime and in sharing tax information. The key mechanism through which onshore actors, primarily the US and EU, force up standards in tax havens is by interfering with the ease with which money from offshore sites can come back onshore.

Daniel Drezner (2007) discusses the use of clubs of states and state-sanctioned actors to force up international financial regulatory standards with respect to criminal activity. A recent example of this process can be found in the Financial Action Task Force (FATF). This organization, currently with 33 state members, including the world’s major financial powers, is designed to combat money laundering and (particularly since 2001) terrorist financing. In some ways, it has been successful — it claims that of 23 jurisdictions it classified as non-cooperative in 2000, all had made significant progress by 2006 (FATF, 2007). It is worth noting that the FATF mostly affects standards in the middle to lower end of the offshore range. Regulators nearer to the top of the regulatory spectrum, some of which are in fact members, already met FATF standards, and were thus unaffected. ⁶

More recently, the focus of both the relevant international regulatory clubs and the biggest onshore actors — the US and EU — has shifted from terrorism and organized crime to tax evasion. ⁷ Both the OECD and the G-20 have recently focused increasingly on tax evasion, in part through the promotion of tax information exchange agreements (Genschel and Schwarz, 2011). More recently, the US has begun using the threat of criminal prosecution to chip away at Swiss banking secrecy rules, and the Swiss government has responded by encouraging banks to cooperate with US efforts (e.g. Bart and Hirt, 2013).

Conclusions

Both the shipping and banking cases show a pattern in which the biggest offshore sites have significant regulatory capacity. They are selling relief from specific regulatory costs, but they are also selling the services needed to connect with the rest of the global economy. The regulatory bottom tends to be made up by more recent entries and relatively small players. The cases also show a tendency for states in the broad middle of the range of offshore regulators to aim for particular regulatory niches rather than to try to compete for business across the board. In other words, regulators in the middle work to develop reputations for dealing well with particular market segments specifically. The cases also show offshore regulatory standards on the whole increasing in response to pressure from onshore actors using club goods as leverage. In contrast to much of the literature on globalization and offshoring, this suggests that regulatory standards are not racing anywhere. They are dispersing, and are on average slowly increasing over time, when and because the major onshore regulators cooperate to push them up.

This model of international regulatory dispersion does, however, apply best to pure examples of offshore regulatory arbitrage. The farther away one gets from pure regulatory arbitrage, the less solidly the assumptions of the model hold. This is particularly true of offshoring to take advantage of cheaper labor, because labor is limited in a way that regulatory services per se are not, creating a crowding-out effect. In other words, beyond a certain point, offshoring labor-intensive activities to a particular country drives up the cost of labor there, offsetting any benefits to scale and putting a (flexible) upper limit on the amount of offshored business a regulator can attract. Nonetheless, even in the case of labor-intensive industries, actors that control access to major markets should be able to force average global standards upward.

From a policy perspective, this model tells us several things. It tells us that the logic of regulatory races to the bottom can be trumped by the logic of regulatory clubs. The strength of clubs affects patterns of offshore regulatory dispersion in two ways. First, it affects the distance between regulatory levels in the major markets and those in the leading offshore sites. The bigger the key clubs, and the closer the rules they enforce are to regulatory levels in major markets, the smaller the regulatory distance between onshore and the top tier of offshore is likely to be. Second, the willingness of onshore states and other actors, both individually and through clubs, to both enforce access standards and to raise those standards over time is a key factor affecting the upward pressure on offshore regulatory levels, both at the high end and across the regulatory spectrum. In other words, the upward pressure on offshore regulatory levels is largely determined by how willing clubs (both state and non-state), as well as individual onshore states, are to keep pressure on them.

However, it is worth noting two things in this context. The first is that, as a general rule, clubs, and, even more so, individual onshore actors, have much less leverage with respect to offshore sites at the very bottom end of the regulatory ladder than they have farther up. While the dispersion model suggests that regulatory races to the bottom are unlikely, and that offshore sites that populate the regulatory bottom are likely to be relatively few and small, it also offers little guidance about what to do about them. The second is that the regulatory distance between the major onshore and offshore regulators is not determined by the forces of globalization generally, or (beyond a certain point) by the nature of the industry being regulated. Rather, it is in large part determined by the choices made by onshore states — those that are the major markets for the industry — and is a trade-off between the costs and benefits of internationalizing domestic regulatory levels. Offshore regulation on a large scale exists because these states let it, not because globalization demands it.