Introduction

The years following the global financial crisis have given new energy and focus to understanding the nature and extent of financial industry power in the global political economy. Recent research has made important advances in our understanding of regulatory capture (Mattli & Woods Reference Mattli, Woods, Woods and Mattli2009; Young Reference Young2012; Carpenter & Moss Reference Carpenter, Moss, Carpenter and Moss2014), the structural and discursive power of the financial services industry (Bell & Hindmoor Reference Bell and Hindmoor2017; Winecoff Reference Winecoff2017) and financial industry lobbying influence (Woll Reference Woll2013; Pagliari & Young Reference Pagliari and Young2014; Young Reference Young2014; Chalmers Reference Chalmers2020; Keller Reference Keller2018; Pagliari Reference Pagliari, Helleiner, Pagliari and Spagna2018). Common to much of this research is a focus on explaining the conditions under which finance wins or loses in its efforts to shape financial regulation. This research has led to impressive cumulative findings and a central and innovative narrative that financial industry power is highly contingent (Young Reference Young2012, Reference Young2014; Goldbach Reference Goldbach2015, p. 17). Largely overlooked in recent research, however, is the related question of what it means for finance to win or lose. This is the focus of this paper.

Many scholars, either explicitly or implicitly, describe a win or a loss for finance in terms of preference attainment: the degree to which a firm's preferred regulatory outcomes are reflected in actual regulatory outcomes. Of course, this approach is not unique to those studying the politics of financial regulation. In fact, preference attainment has become the gold standard for measuring influence in the broader interest group literature (see Vannoni Reference Vannoni2017). Whether the focus is on finance or interest group influence in general, the problem with this approach is that it ignores a critical component of what it means to win or lose, namely when some proposed rule is excluded from finalized legislation. This is, in part, a methodological problem. Rules that are excluded from some legislative outcome cannot be matched to firm preferences. It is also partly, and more importantly, a theoretical problem. Rule omission, when it is examined, is done so when evidence clearly points to financial industry lobbying influence (Carpenter Reference Carpenter2010; Carpenter & Libgober Reference Carpenter and Libgober2018, p. 6; Libgober Reference Libgober2020b, p. 38). More importantly, rule exclusion is often assumed to be the pinnacle of lobbying power and influence (Baumgartner et al. Reference Baumgartner, Berry and Hojnacki2009; Nelson & Yackee Reference Nelson and Yackee2012; Young Reference Young2012). This seems to make intuitive sense. While determining whether regulatory rules are less stringent or more stringent or whether these rules impose smaller or greater costs on certain industry actors is important, having rules altogether removed from regulation arguably makes these other outcomes irrelevant.

Our aim here is to challenge this intuitive but overly simplified characterization of rule omission and thereby contribute further to the growing scholarship on the contingent nature of financial industry power. Our central argument is that to fully understand whether rule omission reflects a win or a loss for financial industry lobbyists we need to consider two things: (1) the discrete regulatory preferences of industry actors and (2) what happens to rules after they are omitted. This second step means systematically following rules after their omission. To the best of our knowledge ours is the first paper to do this. Our central contribution is to extend the preference attainment approach to understanding lobbying success and to bring it up to date with other recent developments regarding the complex and multilevel nature of financial regulation. First, this entails acknowledging the regulators and industry actors often work (simultaneously) at global, supranational and national levels in their efforts to shape regulation. For example, European Union rules for capital requirements are partially downloaded from the Basel Committee on Banking Supervision and, at the same time, take into consideration the domestic regulatory systems in the bloc's 27 member states. Second, regulations are comprised of multiple, discrete rules and not every rule is important to every industry actor. To really get a sense of how a proposed regulation differs from finalized regulation requires an examination at the level of the discrete regulatory rule. In fact, existing research shows that lobbyists rarely (if ever) lobby to shape an entire piece of legislation but rather focus their efforts on specific rules or issues (Baumgartner et al. Reference Baumgartner, Berry and Hojnacki2009). Finally, and bringing both points together, while proposed regulatory rules can certainly be shaped to mirror the preferences of industry actors, they can also be migrated to different regulatory venues, to different regulatory proposals and can even be completely omitted from the legislative agenda (i.e., where they neither end up in a different venue nor in a different legislative act). Critical to our argument is that none of these outcomes necessarily implies undue industry influence or regulatory capture. Only in mapping out industry preferences and carefully tracking what happens to rules following their removal can we say something about the extent to which finance wins or loses in its effort to shape regulation.

While our arguments about rule omission apply equally to all studies of lobbying influence that use a preference attainment approach, the empirical context of our study is financial industry lobbying in the European Union (EU). Our rationale for this case selection is twofold. First, our understanding of financial industry lobbying and power has seen important advances since the global financial crisis and focusing on finance allows us to speak to these recent developments directly. Second, the EU serves as an ideal case for this analysis in that the bloc's post‐crisis regulatory reforms occurred across different dimensions of financial industry lobbying and over a relatively short period. Specifically, we examine rule omission across six regulatory proposals and six pieces of finalized EU legislation as well as approximately 900 comment letters sent to the European Commission by financial industry actors in the context of large‐scale stakeholder consultations. Importantly, the administrative procedure of the EU codifies and facilitates stakeholder involvement through open public consultations at the final drafting stage of a legislative proposal. These documents allow us to map out actor preferences, trace rules from their proposal to finalized legislation and identify the prevalence of rule omission. In a second step, we present two in‐depth case studies where we aim to better understand the causal mechanisms explaining rule omission and to systematically trace out what happens to omitted rules after legislation is passed. Our first case study relates to solvency rules in the Institutions for Occupational Retirement Provision Directive (IORP II) and is an example where rule migration does reflect a strong case of industry influence where pension providers successfully lobbied to avoid costly solvency requirements. The second case study relates to regulations on short selling in the Alternative Investment Fund Managers Directive (AIFMD). In this case, short selling rules are first omitted from AIFMD but later form the basis for the EU's new Short Selling Regulation (SSR). We show how this reflects a second‐best outcome for the hedge fund industry. Short selling rules under AIFMD would have only targeted the hedge fund industry. However, the SSR applies to all financial service providers engaging in short selling activities. While the intense lobby efforts of the hedge fund industry did result in an instance of rule migration, it cannot be deemed an instance of industry influence. Taken together, our case studies paint a more complete picture of financial industry power and, in particular, the highly contingent nature of this power when it comes to shaping financial regulation.

Winning, losing and rule omission

We define rule omission as any instance where a discrete and meaningful aspect of a regulatory proposal is omitted from a regulatory outcome. Our definition starts from an understanding that regulatory proposals and regulatory outcomes comprise any number of discrete rules or issues. For instance, the European Commission's proposal for MiFID II comprises a series of specific rules on the types of financial instruments that should be regulated at the EU‐level, levels of capital requirements for these instruments, as well as rules for those working in securities markets, especially with regards to consumer protection and transparency. Industry actors rarely try to change an entire proposal. Rather, they focus their lobbying efforts on specific, discrete rules within a proposal. This has important implications on possible regulatory outcomes when financial power is exercised. Specifically, proposed regulation can pass into finalized legislation with any number of its originally proposed rules either (1) as they were proposed (i.e., no change between proposal and outcome), (2) altered (the outcome, for example, is less stringent or more stringent than proposed) or (3) omitted.

This third outcome, where rules are omitted, is something that is commonly acknowledged by scholars but is either under‐theorized or interpreted to be the pinnacle of industry power. Altering a specific proposed rule to fit industry preferences is important, but having rules altogether omitted from regulation arguably makes other regulatory outcomes irrelevant. Even still, existing studies on interest group influence tend to ignore rule omission, focusing instead on rules that have been altered or rules that appear in finalized regulation as they were proposed. Why is this the case? Why is rule omission both acknowledged to be an important aspect of the power of finance but is also swept aside when it comes to empirical analysis?

For much of the existing literature, the issue starts with the assumption that winning and losing can be conceived in a unidimensional policy space (see Dür et al. Reference Dür, Bernhagen and Marshall2015). In basic terms, a win or loss is reflected in how close an actor's ideal points on a specific rule are to what that rule looks like in a final regulation. Winning or losing is therefore constrained to what policymakers have proposed and winning is all about actors having their preferences regarding a rule reflected in the final regulation. Omitted rules are ignored because they are not observed in the final regulatory outcome. This approach is consistent with a growing body of scholarship in the broader interest group literature (McKay Reference McKay2012; Dür et al. Reference Dür, Bernhagen and Marshall2015) as well as studies focusing on financial industry lobbying success (Woll Reference Woll2013; Pagliari & Young Reference Pagliari and Young2014; Young Reference Young2014; Keller Reference Keller2018; Pagliari Reference Pagliari, Helleiner, Pagliari and Spagna2018). Chalmers (Reference Chalmers2020), for example, measures financial industry lobbying success using this assumption of unidimensionality. Lobbying success only relates to whether the lobbying demands are reflected in finalized regulation that is more stringent, less stringent or at the same level of stringency as proposed by regulators. Using a similar approach, other scholars replace rule stringency with more or less government involvement (Yackee & Yackee Reference Yackee and Yackee2006; McKay & Yackee Reference McKay and Yackee2007; Nelson & Yackee Reference Nelson and Yackee2012) or efforts to make final legislation more aggressive or to dilute it (Klüver Reference Klüver2011). In these cases, the policy space is unidimensional, and proposed rules are traced to their appearance on some legislative output but, crucially, ignored if they were omitted from that output. Goldbach's (Reference Goldbach2015) study of industry influence over Basel II rules reframes this approach in terms of rule integration and rejection, where rejection relates to the number of ‘preferred policies [i.e., rules] not integrated into the agreement’. Rejection is different from what we are calling rule omission in that it refers to industry preferences that do not make their way onto the final regulation as opposed to rules that are proposed by regulators but then omitted from the final regulation. In terms of preference attainment, rule rejection would be characterized as a complete loss for industry actors.

The problem of assuming unidimensionality is that it either obscures or excludes the possibility of rule omission. The approach taken in Chalmers (Reference Chalmers2020), for instance, can only measure whether lobbying has made rules more or less stringent, but cannot say anything about rules that have been removed from finalized regulation. Of course, this limitation is also found in the broader interest group literature using the same basic method for measuring preference attainment. Taking a less structured approach, like Goldbach (Reference Goldbach2015), does not eliminate the possibility of rule omission, but it also does not explicitly ‘observe’ omission. In other words, this approach obscures the question of rule omission by being less detailed about rule acceptance and rejection. More generally, conceiving of winning and losing on a unidimensional policy space tends to disallow consideration of rules that appear in a proposal but that are omitted in regulatory outcomes.

Even when scholars explicitly acknowledge some form of rule omission it is done in a rather limited way. Libgober and Carpenter (Reference Libgober and Carpenter2018, p.38) touch upon the issue of rule omission in their case study of the Volcker rule and acknowledge that existing quantitative measures may underreport lobbying success by ignoring rule omissions. However, they start with a rule (the deletion of Appendix B) where rule omission clearly reflects lobbying success (see also Carpenter & Libgober Reference Carpenter and Libgober2018, p. 6). Hence rule omission is necessarily limited to one outcome: lobbying success. In a different context, Young (Reference Young2012) uses process tracing to examine financial industry lobbying influence over Basel II regulations. Young considers rule omission as part of what it means for finance to win or lose and states: ‘Evidence of influence can be one of the following: either a specific proposal articulated as part of a lobbying effort can be shown to have made its way into actual regulatory policy, or, a regulatory proposal that was already on the agenda of policymakers is later removed as the result of private‐sector efforts’ (Young Reference Young2012). While Young allows for the possibility of rule omission as an outcome in his analysis, he characterizes this outcome as an instance of financial industry lobbying success. He does so, we would argue, because he does not follow omitted rules after they leave the regulatory agenda.

Conceptualizing rule omission within the context of financial industry power over financial regulation requires relaxing assumptions about unidimensionality (i.e., acknowledging that not all proposed rules end up in the finalized regulation) and engaging in more explicit theorizing than we see being used in process tracing and case study approaches as well as large‐n studies. We argue that understanding rule omission requires (1) mapping actor preferences as they pertain to discrete rules that comprise regulatory proposals, and (2) extending analyses beyond regulatory outcomes to trace out what happens to withdrawn rules after legislation is passed. The first task of mapping preferences, as noted above, is something that has increasingly been done with great success in recent work (e.g., Pagliari & Young Reference Pagliari and Young2014). It is the second task of carefully following omitted rules where we are breaking new ground in this paper.

Typically, omitted rules are assumed to have completely disappeared from the agendas of both regulators and industry actors. This may indeed be the case. Lall (Reference Lall2012), in his analysis of the impact of financial industry actors on the Basel III Accord, notes how so‐called ‘w factor’ rules were ‘dropped’ from the new Accord and that this reflects industry preferences. Part of our main argument is that omitted rules are not invariably relegated to the dustbin of regulatory history. Specifically, omitted rules can also migrate, moving to different venues, whether that is a different regulatory venue, a different regulatory proposal or both. To understand rule migration, it is useful to borrow some insights from research on venue shopping. Indeed, interest group researchers have long acknowledged that groups are adept at seeking out the most favourable venue for specific issues (Marshall & Bernhagen Reference Marshall and Bernhagen2017). In particular, interest groups facing local regulatory pressures are able to take advantage of venue shopping strategies. These insights about venue shopping can be adapted and applied to rule migration. Given the highly interdependent nature of international finance, regulation occurs at multiple and sometimes overlapping and competing venues, either at the sub‐national, national, regional and international levels (Farrell & Newman Reference Farrell and Newman2014). We contend that instead of groups simply mobilizing at different venues, as is the case in venue shopping, they can also work to move rules to different venues. From the perspective of industry actors, we would expect the aim of rule migration to be to move a rule to a venue that would give these actors better odds of seeing their preferences reflected in regulatory outcomes. This can be achieved in different ways. For example, within the context of the EU, a rule may migrate from an EU Regulation, which is legally binding for all EU Member States, to a Directive, which gives each Member State leeway in implementing the new rule into national law. Alternatively, rules can also be migrated from the European Union decision‐making institutions, which have far‐reaching powers over national laws, to some international regulatory agency, like the International Organization of Securities Commissions (IOSCO), which produces rules that are not legally binding. This can even happen within the EU. A rule may migrate from the Commission to a regulatory authority, like the European Banking Authority (EBA). While still part of the EU decision‐making apparatus, the EBA's remit is more technical thereby shielding industry lobbying activities from public view.

It is important to note that regulators can also be motivated to migrate rules to different venues. This can result from several factors. They may move rules at the request of industry actors, whether because they are ‘captured’ or because they coincidently share the same regulatory preferences. However, regulators may also do so to protect rules from undue industry influence, seeking out more protected venues, more public venues or venues that are less hospitable to industry lobbying.

The different scenarios sketched above clearly mean different things for understanding the power and influence of financial industry actors. This brings us to our final observation. Rule omission, and in particular migration, are part of the larger narrative of what it means for finance to win and lose, but in addition to following what happens to rules once they have been removed from regulatory outputs, we also need a specific sense of the regulatory preferences of industry actors to complete our picture of the nature of financial industry power over regulatory outcomes.

Patterns of rule omission in the EU

How prevalent is rule omission? To answer this question, we draw on a novel dataset of financial industry lobbying efforts to shape financial regulation in the European Union. We have collected data on six European Commission legislative proposals aimed at regulating finance in the post‐crisis period: CRR, CRDIV, IORP II, Solvency II, AIFMD, and MiFID II. Each legislative proposal comprises multiple rules, and interest groups lobby on the rules that matter most to them, rather than the proposal as a whole. Discrete rules largely follow the Commission's tendency to formulate consultation documents in terms of policy questions. Please see the Appendix for an overview of legislative proposals and EU legislation.

Having both the proposal and finalized legislation allows us to match discrete rules and how they change (or stay the same) across these processes. To compare proposed rules to finalized rules we code each rule in terms of regulatory stringency, consistent with preference attainment approaches used in existing studies. However, in this case, we also code for rule omission: rules that were proposed but removed from finalized legislation. This gives us four possible outcomes: the final rule is (1) the same level of stringency that was proposed, (2) more stringent than proposed, (3) less stringent than proposed and (4) omitted from the final regulation.

Figure 1 illustrates the results of this coding for all five pieces of finalized legislation. Note that Capital Requirements Regulation (CRR) and Capital Requirements Directive (CRD) IV are considered together since consultation proposals for both the directive and the regulation were paired. We can see that rule omission is an important outcome in nearly all five cases. Indeed, for CRR/CRD IV, rule omission is the most frequent regulatory outcome, constituting almost 60 per cent (9 from 14 rules) of all possible outcomes. It is also prevalent in IORP II, AIFMD and MiFID. The only case where we observe no instances of rule omission is Solvency II.

Figure 1.

Regulatory outcomes compared to regulatory proposals. [Colour figure can be viewed at wileyonlinelibrary.com]

A closer look at a few examples of omitted rules reveals several key patterns. First, rule omission can occur when a proposed rule is omitted from the finalized legislation and is simply phased into legislation at a later date. This was the case for new regulations regarding liquidity standards and leverage coverage requirements that were part of the European Commission's (2009) proposal and which only later become part of the CRR. It was only after an agreed‐upon observation period that these rules were introduced into the CRR via a delegated act several years after the ratification of the new regulation (European Commission 2015, 2016). Second, omitted rules can migrate to other EU decision‐making venues. The Commission's 2014 Review of the Markets in Financial Instruments Directive (MiFID), for instance, proposed a new definition regarding ‘admission to trading’ that gave the operator of the regulated market the authority to decide which financial instruments could be traded. The finalized rules outlined in MiFID II, however, omitted any mention of this new definition. Instead, the task of developing a new ‘admission to trading’ definition has been delegated to the European Securities and Markets Authority, the EU's chief regulatory authority for securities markets (MiFID II, L173/436 A51). Omitted rules can also migrate to other regulatory proposals and other pieces of EU legislation. Important new rules about short selling practices, put forward in the Commission's consultation on hedge funds, as well as new rules about foreign currency denominated mortgages, proposed in the 2009 review of the CRD, were omitted from finalized legislation but, critically, migrated to more targeted EU legislation. Specifically, short selling is addressed in the EU's 2012 SSR (which we discuss in detail below) while foreign currency denominated mortgages are part of the 2014 Mortgage Credit Directive. A fourth pattern occurs when omitted rules migrate beyond the borders of the EU, turning up in various non‐EU regulatory venues. This was the case for the Commission's 2009 proposal for far reaching rules on through‐the‐cycle expected loss provisioning (TELP) for credit risks on debt instruments. The Commission's proposal for TELP was abandoned when the International Accounting Standards Board (IASB) started developing their own expected credit loss model. The Commission decided to omit TELP rules from both the CRDIV and CRR and instead allow EU member states to use the IASB's 2014 standards (PricewaterhouseCoopers 2014). Importantly, these IASB rules were far less stringent than the Commission's original TELP rules. Finally, omitted rules can drop completely off the agenda neither to migrate to a new venue nor to a new legislative proposal. This was the case for proposed solvency rules in IORP II, which forms the basis of our second case study below.

Rule omission, rule migration and lobbying success

The patterns sketched above can only give us a sense of the prevalence of rule omission but cannot say much about what this means for the lobbying power of the financial industry in the EU. As we have argued above, doing so not only requires carefully tracing out what happens to omitted rules but also taking the preferences of the regulated industry into consideration and assessing who wins and who loses when rules are omitted from regulatory outcomes. This is what we do below. Specifically, we present two case studies centred around EU regulatory proposals: (1) the IORP II, and (2) AIFMD. These cases are two of the most important and far‐reaching proposals coming out of the EU's response to the financial crisis and have important variation in their outcomes. Our first case study is an example of rule omission where a proposed rule, Solvency II for insurance and occupational pension fund providers, is dropped from the final legislation and does not end up in a different venue or legislative act. We interpret rule omission in this case as a strong form of industry influence. Pension providers effectively lobbied to avoid costly capital requirements embodied in Solvency II rules. Our aim in this case study is to carefully trace the Solvency II rule from the proposal stage and examine when and why regulators dropped it from the agenda. We show how this resulted from (1) weak and sometimes equivocating support from EU institutions and (2) strong and united opposition from Member State regulators and the pension industry. Our second case study is an example of rule migration, where regulators omit restrictions on short selling activities for hedge funds from AIFMD (2011), but rather address short selling in a different piece of legislation, namely the SSR (2012). While hedge fund lobbying resulted in the removal of short selling from AIFMD, we cannot describe this as a lobbying success. Instead, hedge funds face more stringent regulations under SSR. This case study shows how this outcome is the result of: (1) a strong initial position of the Commission as well as a strong mandate to regulate hedge funds, and (2) a unified position on how uncovered short selling and short‐and‐distort practices were best regulated outside of AIFMD.

In line with Carpenter (Reference Carpenter, Carpenter and Moss2013, p. 58), we consider the counterfactual in each case to be the largely unchanged transposition of the proposed legislation. This is also consistent with an important study by Dür et al. (Reference Dür, Bernhagen and Marshall2015, p. 955) whose analysis of 70 proposals put forward by the European Commission between 2008 and 2010 shows that the Commission primarily proposes legislation with the anticipation that it will indeed pass.

IORP II

The global financial crisis revealed critical vulnerabilities on pension fund balance sheets. IORPs were exposed to financial markets in their capacity as institutional investors and had been taking on increasing risk through securitization. While securitization was intended to transfer credit risk to those better able to absorb losses (e.g., large institutional investors), it instead increased the fragility of the entire financial system by allowing banks and other intermediaries to ‘leverage up’ by buying one another's securities. Meanwhile, the biometric risk profile of pension plan holders was changing with an ageing population. The exposure of IORPs to financial markets meant that beneficiaries were at risk of losing their pensions, potentially leading to significant societal costs.

The Commission's main response to these problems was to address perceived fragilities in existing IORP rules. The process started with the 2008 consultation on the Harmonization of Solvency Rules Applicable to IORPs (European Commission 2008b). One weak point of IORP I was its use of Solvency I rules for both insurance and pension providers. Solvency I, the minimum amount of additional assets required to be held by insurance and pension providers, was criticized for being descriptive and allowing for heterogeneous approaches to solvency between Member States and market actors. Although IORP I had taken the first step to ensure prudential supervision, Solvency I, encompassed in article 17, it did not include capital requirements enabling institutions to absorb unforeseen losses and give assurance to policyholders. These concerns were addressed in Solvency II. Hence, when the time came to review the IORP directive, the logical implication was that the revised solvency reference would be transposed to IORPs, as done previously.

The Commission's initial position on adopting Solvency II for IORPs was far from unequivocal. In a speech in April 2008, Charlie McCreevy, Commissioner for Internal Market and Services, stated that the extension of Solvency II to pension funds would not be automatic, but would rather be left to a review of the IORP directive later in the year (McCreevy Reference McCreevy2008c). This same language was adopted in the official consultation document a few months later: The “Commission has no pre‐conceived ideas on the way forward at this stage and there will be no automatic extension of the Solvency II directive proposal to IORPs” (European Commission 2008b). Weakening the Commission's position even further was the subsequent advice given by the European Insurance and Occupational Pensions Authority (EIOPA), an EU regulatory authority. In this case, EIOPA maintained that using Solvency II for IORP II was not desirable (European Insurance and Occupational Pension Authority, 2012).

The consultation was organized along two sections, one about solvency requirements for IORPs, and the other about IORPs operating on a cross‐border basis. Replies to the solvency section reveal four main points of disagreement. These points relate to the difference between IORPs and life insurance, efficiency of Solvency I, potential costs and benefits and implications of moving away from defined benefit plans.

Support for Solvency II came from disparate sources. The time between IORP I and the 2008 consultation had seen international best practice make progress towards a Solvency II‐type approach for pensions: namely, a move towards a risk‐based approach. Proponents argued that while Solvency I had been adequate, new financial methods had increased the risk exposure to the extent that it needed to be replaced. This position was shared by insurance providers and their associations. For instance, the Association of Mutual Insurers and Insurance Cooperatives in Europe argued that “the uneven application of Solvency I, which penalizes operators and beneficiaries alike, is an additional reason why pension systems currently differ so much between the different Member States” (Association of Mutual Insurers & Insurance Cooperatives in Europe 2008). Other interested parties, namely the European Insurance and Reinsurance Federation, Swedish and Danish regulators, also expressed concern for potential market distortions if insurance products and IORPs would diverge in their solvency requirements.

Opposition to adopting Solvency II in new IORP rules came from a united coalition of pension providers, their associations and national regulators. Two main issues formed the basis for their opposition. First, pension providers started from the position that IORPs’ exposure to financial risk was difficult to quantify and asked the European Commission to carry out an impact study to establish a basis for discussion. They noted that studies to date had only looked at a blanket application of Solvency II without considering IORP specific adjustments. Aegon, a Dutch life insurance company, argued that the potential benefits outweighed any potential costs, and in particular that ‘a transition from more “optimistic” systems of determining assets and liabilities to a more economic approach does not create more liabilities but reveals the true state of affairs’ (Aegon 2008). Moreover, any extra cost would have to be assessed against the possible consequences of higher capital requirements. The lack of a cost‐benefit analysis of an IORP‐specific Solvency II meant that the benefits were undefined, whereas the costs, expressed as defunct pension funds and lower benefits, resonated strongly with regulators.

Opposition to Solvency II rules also centred on the question of subsidiarity. The UK Department for Work and Pensions, for example, expressed a strong aversion to adopting Solvency II in the IORP directive: ‘there is currently no evidence to suggest pension schemes are inadequately protecting their members’ (UK Department for Work & Pensions 2008). It went on to refer to an EIOPA survey concluding that, although the implementation of solvency requirements across Member States varied, all respected the agreed requirements (Committee of European Insurance & Occupational Pensions Supervisors 2008). Moreover, the department continued by calling upon the principle of subsidiarity for IORPs, arguing that the Commission should have no place in regulating this policy area (UK Department for Work & Pensions 2008). Several other opponents to Solvency II, including AGE, The European Association of Paritarian Institutions and the European Association of Public Sector Pension Institutions, argued that the current requirements had not yet been put to the test, and could thus not be so quickly discarded (AGE 2008; Association of Mutual Insurers & Insurance Cooperatives in Europe 2008; European Association of Paritarian Institutions 2008).

A few weeks after the consultation was launched, McCreevy, speaking at the Life Trust Foundation Seminar, expressed awareness about the strong reactions to the launch of the consultation and was sure to reiterate its purely consultative nature (McCreevy Reference McCreevy2008a). It was, at this point, not clear what the Commission would do. There was still an option of putting forward a revised version of Solvency II, tailored for IORPs. However, as the consultation deadline was reached on 28th November, 2008, the strong opposition made it politically unfeasible for the Commission to move forward with a proposal where Solvency II would be applied to IORPs. Meanwhile, the financial crisis that had begun in the United States spread to financial markets in Europe, and bank rescue was at the top of the agenda. The question about whether to adopt Solvency II for IORP was subsequently delayed.

Though the solvency requirements in the pre‐crisis directives for IORPs and life assurance had been the same, the legislative revisions during the crisis led to divergence. Solvency requirements for IORPs remained unchanged, whilst the life assurance directive adopted a risk‐based approach with capital requirements. One possible explanation is found in the reply to the consultation on the solvency issue for IORPs. It resonated strong support against harmonizing requirements for IORPs with Solvency II. The coalition of a few strong Member States, pension funds, as well as representative groups for both employer and employees meant the Commission faced a broad base of resistance to a possible change. Member States with compulsory schemes and a high level of defined benefit plans were able to favour their domestic regulatory status quo and emphasize their prerogative on the social dimension of occupational pensions. The Commission backed down from any change to the existing solvency requirements for IORPs and instead included language eliminating any possible interpretation of Solvency II applying to IORPs. Not only was Solvency II omitted from IORP II, but the Commission committed to language dismissing its possible future inclusion.

AIFMD

Even more so than with the case of IORPs, the global financial crisis shed light on hedge fund activities. A rapid increase in the number and size of hedge funds starting in the late 1990s drove concerns about their regulation and market impact. Though it was acknowledged that financial markets can benefit from hedge fund activities under normal conditions, the role of hedge funds, especially short selling, came into question (McCreevy Reference McCreevy2008b). The Commission set out from a strong initial position, intending to legislate on hedge funds in its 2009 Consultation Paper on Hedge Funds that, as part of a comprehensive legislative review for all financial market actors (European Commission 2009, p.12), resulted in AIFMD 2011 (European Commission 2008c). AIFMD was the Commission's earliest response to the crisis and benefited from the need to give the appearance of acting quickly to mend fragilities in financial markets highlighted by the crisis. Giving further impetus to the Commission's work was the existing hedge fund regulation, being a patchwork of Member State rules and self‐regulatory initiatives driven by a pro‐market stance that promoted cross‐border development and the efficiency of the investment fund industry (European Commission 2009, p.15; Pagliari & Helleiner Reference Pagliari, Helleiner, Helleiner, Pagliari and Zimmerman2009).

Concerns about hedge fund activities crystallized around the issue of short selling: the practice of selling an asset the seller has borrowed to profit from a potential fall in the price of the asset. Short selling was described both by legislators and hedge funds as a defining characteristic of hedge fund activities, whilst acknowledging that other investors could also engage in the practice (Hedge Fund Working Group 2008, p.12; European Commission 2009, p.58).

Within the context of the Commission's consultation, a central argument revolved around how short selling plays a vital role in price discovery. Temporary bans in the autumn of 2008 on short selling in United States and United Kingdom markets had been followed by a few rapid downward spirals in price developments for selected stocks. The Commission's consultation, seeking replies about the usefulness of short selling, asked about these recent bans and if ‘the recent reduction in hedge fund trading (due to reduced assets and leverage, and short selling restrictions), affected the efficiency of financial markets’ and whether the restrictions had ‘led to better/worse price formation and trading conditions’. The majority of responses argued either that the ban had been negative, or, that not enough data was available to make a conclusive statement. The European Central Bank (ECB) gave further credence to the price discovery function of short selling, stating that ‘authorities and the academic literature have acknowledged that short selling plays a positive role in the market in the long run’ (European Central Bank 2009). Similar views were expressed by other government actors such as Swedish financial regulators, the French Ministry for the Economy and Finance and the German Federal Ministry of Finance (Bundesministerium der Finanzen 2009; Finansinspektionen & Sveriges Riksbank 2009; Ministere de L'Economie des Finances et de L'Emploi 2009). Even stronger sentiments against the short selling ban were expressed by financial industry associations, like the International Swaps and Derivatives Association, stating that ‘there has been a clear loss of liquidity in the market due to short selling restrictions’ (International Swaps & Derivatives Association 2009). The Chamber for Workers and Employees in Austria argued along similar lines. Their reply acknowledged short selling to lead to price distortions in some cases, but also that suspending such activities could lead to an even worse situation (Bundesarbeitskammer 2009). With key EU Member States, several interest groups and even labour market representatives holding a negative view on temporary restrictions on short selling, the burden to provide evidence to the contrary was heavy on stakeholders favouring restrictions.

While a consensus view formed around the general merits of short selling, the hedge fund industry found it more difficult to make similarly compelling arguments about so‐called uncovered short selling and ‘short‐and‐distort’ practices. The question of uncovered short selling – that is short selling an asset that the seller never holds – faced strong opposition from a broad range of actors. Some, like the ECB, German Insurance Association and Irish Funds Industry Association argued for restrictions on uncovered short selling insofar as these practices constitute a form of market abuse and, as a form of speculation, could threaten market stability (European Central Bank 2009; German Insurance Association 2009; Irish Funds Industry Association 2009). Others, including INVERCO, a Spanish Association of Collective Investment Schemes and Pension Funds, argued in favour of a complete ban on uncovered short selling (INVERCO 2009). Only a small minority was against banning uncovered short selling. The Bundesverband Alternative Investments (BAI), a German investment fund, for instance, argued in favour of uncovered short selling due to the speed through which positions can be acquired, the limited fees incurred and the compatibility with algorithm‐based investment strategies (Bundesverband Alternative Investments 2009).

The second critical issue was about short‐and‐distort, the practice of short selling an asset and then spreading rumours to ensure its price drops. Such practices had already been identified by the Securities Exchange Commission and the Financial Services Authority on several occasions in the months leading up to the financial crisis as instances threatening the stability of individual assets, and to some extent, financial markets at large. One example of such a case was Halifax Bank of Scotland, which was targeted by short sellers in spring of 2008. False rumours about its solidity began circulating after the stock had been subject to relatively large short positions, eventually forcing it to seek emergency funding from the Bank of England (Bank of England and Financial Conduct Authority, 2015). For proponents of short selling, the issue was spreading rumours rather than shorting. Ernst & Young, for instance, argued that restrictions on market abuse were already in place through the Market Abuse Directive (MAD), for which implementation should be improved before new regulation was formulated (Ernst & Young LLP 2009). Spreading false rumours would fall under article 5 of the MAD, which addresses market manipulation, including transmitting false or misleading information (European Parliament & The Council of the European Union 2003). Other important industry actors, including the Association of British Insurers, Morgan Stanley and PricewaterhouseCoopers, argued in favour of revising existing legislation on market abuse to better monitor and address instances where short selling was involved (Association of British Insurers 2009; Morgan Stanley 2009; PricewaterhouseCoopers 2009).

For the Commission, set on addressing financial market fragilities, the signal was clear: short selling, a defining characteristic of hedge funds, might be flawed but these flaws would be better legislated away from AIFMD, whether that meant bolstering existing rules, like MAD, or creating new and perhaps bespoke rules. For the hedge fund industry, moving the issue of short selling away from AIFMD had the benefit of ensuring new short selling rules would be applied evenly to all actors engaging in these activities, not just hedge funds. Indeed, the hedge fund industry was effective in making a strong argument that short selling was not specific to hedge funds, even though the Hedge Fund Working Group, an international standard‐setting body for the industry had previously claimed the practice to be a defining feature of the industry (Hedge Fund Working Group 2008, p.12). The overwhelming number of replies to the Commission's consultation argued that short selling was not specific to hedge funds.Footnote ¹ Again, an opposing view came from a small number of actors. Chief among them was the European Trade Union Confederation (ETUC), which expressed a strong negative view of short selling, arguing that it poses a systemic risk and should be regulated under AIFMD (European Trade Union Confederation 2009). Nevertheless, the ETUC's argument was drowned out by the previously mentioned outright rejection of the idea of limiting the regulation of short selling to the hedge fund industry. For instance, the Ministry of Finance in Finland argued that it did not see the need for tight regulation on short selling and that existing supervision was sufficient (Ministry of Finance 2009). Other replies supported addressing concerns about market abuse through the revision of other legislation or even through stand‐alone legislation on the practice. One important concern was about how proposed restrictions on short selling by hedge funds would create arbitrage opportunities for other investors (Bundesverband Investment und Asset Management 2009; European Banking Federation 2009; European Federation for Retirement Provision 2009). Fears of other financial products, such as credit default swaps, being used to circumvent short selling restriction were also voiced. The Commission could thus not, because of the consultation, legitimately include short selling in the final hedge fund specific legislation.

At first glance, it may appear as though hedge funds benefited from short selling migrating from the AIFMD and that the consultation on hedge funds played a key role in this. Had legislation on short selling been restricted to AFIMD, hedge funds would have been disproportionately hit and arbitrage opportunities for other actors engaging with short selling would have arisen. Moreover, the Commission's press release for the consultation had stated that the consultation was to be part of a review ‘which is to be finalised in 2009’ and that ‘views and evidence are sought […] so as to guide on appropriate regulatory initiative’ (European Commission 2008a). Given that the AIFMD regulation was finalized in 2009 whilst the SSR regulation was finalized in 2012, it is reasonable to assume that the Commission intended to propose legislation on the ‘market integrity and efficiency’ of hedge funds, which almost exclusively deals with short selling. Yet, the final legislation does not mention short selling, ‘market integrity’ or ‘market efficiency’. Moreover, McCreevy discussed short selling legislation as a subject pertaining to hedge funds in his speech at the European Parliament before the consultation on hedge funds was launched (McCreevy Reference McCreevy2008b). This position was then revised at the delivery of another speech, following the closing of the consultation where McCreevy made the point that ‘short selling is not only used by hedge funds’, implying that legislation on short selling should not be included in hedge fund specific proposals (McCreevy Reference McCreevy2009). The Commission asked the European Securities Markets Expert Group (ESME), an independent advisory group to the Commission, to prepare a report on short selling after the consultation on hedge funds had closed. The ESME report makes reference to response letters to the hedge fund consultation when discouraging any specific regulation against short selling (European Securities Markets Expert Group 2009, p.12). This supports the argument that short selling was intended to be legislated under AIFMD, but that consultation replies managed to move the issue off the final legislation.

A closer look at the ensuing legislation, the previously cited SSR, reveals how hedge funds only achieved close to a second‐best outcome. Although, the rule migration mechanism remains opaque as the Commission impact assessment accompanying the SSR does not specify the preparatory steps in sufficient detail (European Commission 2010a), the Commission proposal for the SSR does mention that short selling was considered in the context of AIFMD (European Commission 2010b). It continues by outlining how a series of consultations in different contexts considered alternative venues, such as the Market Abuse Directive, for short selling legislation, before ultimately settling on an independent regulation. A Commission communication released on 2 June 2010, which references short selling of sovereigns, is the first known mention of a separate legislation for short selling, and was followed by a public consultation on policy options for possible legislative initiatives on short selling (European Commission 2010c). The final SSR heavily restricts the use of uncovered short selling and regulates possible bans of short selling. The regulation format itself is much stricter than a directive since regulations have binding legal force throughout the EU and enter into force simultaneously in all Member States, whilst directives are looser in their form and stipulate certain results that must be achieved by national legislation.

Conclusions

We have argued that the tendency for existing research to ignore rule omission is partly related to how lobbying success and preference attainment are operationalized. Starting from the position that winning and losing lobbying battles can be conceived on a unidimensional policy space helps scholars match regulatory preferences to regulatory outcomes. However, this approach currently fails to capture omitted rules, which simply fall out of the dataset (there is no outcome to which we can match the demand). What can remedy this problem? First, we need to include the possibility for omitted rules on our unidimensional policy space. In the first instance, this requires carefully cataloguing industry actors’ preferences on discrete rules, identifying rules that are omitted, and then following them to see where they end up. Quantitative approaches to preference attainment would therefore be applied across different venues for rules to migrate (see Vannoni Reference Vannoni2017). Critically, this would need to be carried out at the level of the individual rule and not the proposal in its entirety. While acknowledging that lobbying happens at the level of the rule (or issue), scholars still tend to look at the aggregate policy proposals or legislative outcome. This is a sure way to overlook omitted rules and/or overestimate the extent to which lobbyists obtain their preferences. A fruitful research question that could be addressed using this kind of quantitative approach would be to assess the determinants of rule omission. Have interest groups successfully lobbied rules ‘off the agenda’ or have regulators strategically moved rules to different venues in order to protect them from powerful groups? Finally, a mixed‐methods research approach combining a quantitative analysis of preference attainment (including a consideration of omitted rules) and process tracing would seem best suited to addressing the issues raised above. Quantitative analysis can identify and then indicate the prevalence of omitted rules using a preference attainment approach, and process tracing is needed to the do the work of following omitted rules after their deletion and explaining why they were omitted in the first place.

A chief takeaway from our analysis is the importance of following rules that are omitted from regulatory outcomes, especially if the goal is to better understand lobbying success. Ours is the first attempt to do this. However, our specific focus was on financial industry regulation. But how does rule omission and rule migration look for other regulatory public policy sectors or distributive and redistributive policy sectors? A sector‐level comparison of rule omission and rule migration would be one interesting extension of our research. Differences in salience, conflict and competition in policy sectors may change patterns of omission and dynamics of migration. One further possible extension of this analysis is to engage in more theorizing about the incentives of policymakers and regulators to migrate rules. We speculate that policymakers may shield sensitive rules by moving them to more protected venues. Indeed, our empirical analysis suggests that this might be the case. As a recent contribution by Libgober (Reference Libgober2020a) suggests, these incentives may already be shaped in meetings between rule‐makers and lobbyists before the formal proposal process has even begun. As such, future research may want to focus on this early agenda setting stage of the legislative process. Finally, providing a more compelling set of expectations about the role of policymakers would need consideration of incentives to migrate rules to shield them from lobbying pressures or because policymakers are already captured by the industry.

Acknowledgements

This work could not have been completed without generous funding from the King's College London Social Science and Public Policy (SSPP) Faculty Research Fund. We would also like to thank the team of research assistants who helped with coding, interviews, and primary research: Allana Yurko, Natalia Stachowicz, Mathilda Loussert, Adela Iacobov, Silvia Conti, Francisco Rodrigues and Suzanne Kluft.