NETWORK-SENSITIVE FINANCIAL REGULATION.

• Author: Enriques, Luca

JEL Classifications: G18, G28, K22, K29

1. INTRODUCTION 352 II. ATOMISTIC VERSUS NETWORK-SENSITIVE APPROACHES TO 356 FINANCIAL REGULATION A. Pre-Crisis Convention: Atomistic Microprudential 357 Regulation B. Post-Crisis Criticisms of Atomistic Microprudential 357 Regulation C. Macroprudential Policies 360 D. Network Theory 361 1. Primer: Nodes and Edges 361 2. Centrality 362 3. Resilience to Shocks: Density 364 E. Network-Sensitive Policies 365 1. SIFI Designation 366 2. The Risk-Based Capital Surcharge 368 3. Stress Tests 369 4. Optimal Network Topology Design Policies 371 F. Conclusion: A Stylized Taxonomy of Prudential 372 Regulation III. NETWORK-SENSITIVE GOVERNANCE RULES FOR SIFIS: HOW IT 373 WOULD WORK A. Corporate vs SIFI Governance 374 B. Personal Liability of Directors and Managers 376 C. Managerial Compensation 384 D. Failing SIFI Shareholder Rights in Shadow 387 Resolutions E. Bumpy Atomistic Rules Versus Smooth Network-Sensitive 389 Ones IV. POSSIBLE COUNTERARGUMENTS TO NETWORK-SENSITIVE REGULATIONS 390 V. RESEARCH AGENDA: THE CASE OF DISCLOSURE AND REPORTING 394 OBLIGATIONS VI. CONCLUSION 396 "Network analysis. . . has the potential to help us better monitor the interconnectedness of financial institutions and markets. "

   Ben Bernanke (1)

1. INTRODUCTION

   In the wake of the 2007-09 financial crisis, economists and policymakers alike have become increasingly aware that the structure of the financial system is a key determinant of systemic risk. (2) Shocks that hit part of the system, such as the subprime mortgage market in 2007, can propagate through a complex network of interconnections among financial and non-financial institutions and, ultimately, have a catastrophic impact on the entire economy. (3) In the absence of such interconnections, localized shocks hitting individual players or specific parts of a financial system would not propagate. To put it differently, systemic risk presupposes the existence of a system, which consists of a network of interconnected actors. The resulting "[c]omplex links among financial market participants and institutions are a hallmark of the modern global financial system." (4)

   Given the importance of networks to understand the financial system and the systemic risk it generates, the use of network theory to inform financial regulatory policy would seem a natural approach. (5) Yet, while leading researchers from many disciplines identify network theory as the natural framework for studying systemic risk, (6) legal scholars have largely overlooked this perspective. (7) More importantly, policymakers' use of network theory insights to curb systemic risk has been patchy so far, (8) despite the fact that, following the financial crisis, they have refocused their attention from regulations attempting to preserve the stability of individual banks ("microprudential" regulations or policies) to policies aimed at ensuring the stability of the system as a whole ("macroprudential" regulations or policies). (9)

   Our key argument is that the transition to a regulatory regime that can effectively mitigate systemic risk in the modern highly connected economy will not be complete until financial regulation fully accounts for the structure of the financial network and the interconnections among its components. That is true also for policies designed to improve the corporate governance of what have come to be known as systemically important financial institutions (SIFIs), which are the financial intermediaries that are so designated by regulators in light of their ability to destabilize the financial system in case of failure. (10)

   The goal of this Article is to show that network theory can also be of great help in the corporate governance domain. (11) To show how insights from network theory can make SIFI governance regulations more effective in taming systemic risk, we consider four policy prescriptions--three of which have been put forth by legal scholars, and one that has become law within the European Union (E.U.). To be clear, we remain agnostic about whether these reforms are, in principle, desirable. Our purpose is not to argue in favor of any specific prescription. Rather, we aim to highlight the advantages of a shift of paradigm in the direction of policies incorporating insights from network theory ("network-sensitive policies" or "network-sensitive regulations") and show how they can be implemented in respect of governance-focused regulations.

   We first discuss the idea, advocated by John Armour and Jeffrey Gordon, of imposing personal liability on managers and directors of SIFIs. (12) Armour and Gordon argue that managers and directors of SIFIs lack the incentives to account for the systemic relevance of their firm when they devise and implement strategies, manage risk, and so on. On the one hand, compensation of managers and officers is generally tied to their firms' performance. On the other hand, U.S. corporate governance tends to push managers to focus on stock price maximization. (13) As a result, managers of SIFIs have every incentive to maximize shareholder value, even when so doing creates significant systemic risk. (14) Armour and Gordon claim that a tighter personal liability regime would be an effective means to counter these perverse incentives, because it would induce managers and directors to internalize the systemic relevance of their firm. We show that their proposal may better achieve that goal if amended to incorporate the insights of network theory. In fact, accounting for a SIFI's interconnectedness-via available measures of its "centrality" in the network (15)-better ensures that managers and directors of firms that can impose higher losses on the economy also face higher expected liability. (16)

   Second, some scholars have argued that the compensation structure for managers of systemically important firms-and in particular banks-should be regulated. (17) The basic idea here is that, if bank managers are compensated predominantly with stock and stock options, they will share all the potential gains from successful investments, but will be insulated from part or all of the potential losses in the event of negative outcomes. (18) As a consequence, bank managers will take excessive risks and impose negative externalities on the economy as a whole. (19) For this reason, Bebchuk and Spamann have suggested that part of bank managers' compensation should be based on a broader basket of securities that includes preferred stocks and bonds, thus inducing managers to internalize a larger fraction of potential losses from risky projects. (20) Similarly, but more radically, post-crisis European banking regulations have set a cap on variable compensation equal to 100% of the fixed compensation elements (200% with shareholder approval). (21)

   Without a network-sensitive component, however, these policies are not sufficiently flexible to accommodate the specific characteristics of the various SIFIs and, therefore, would place excessively stringent constraints on some SIFIs and too lax on others. Instead, we show how network theory would allow policymakers to tailor the structure of managers' compensation to the specific features of the various SIFIs. Holding other factors equal, managers of SIFIs that create higher systemic risk should receive less shares and more bonds than managers of SIFIs that pose a relatively smaller threat to the stability of the system.

   Finally, we discuss a proposal by Yair Listokin and Inho Mun. They contend that "regulation by deal," in which a solid firm acquires a defaulting SIFI, is an important tool for mitigating systemic risk. (22) Yet, as they show, regulation by deal is problematic because it allows the shareholders of the defaulting SIFI to hold the economy hostage: (23) shareholders are aware that by opposing the merger deal they can impose a significant externality on society, which allows them to extract rents from the buyer's shareholders and/or taxpayers. (24) For that reason, Listokin and Mun suggest that merger voting rights attaching to the target SIFI shares should be replaced by appraisal rights. (25) This proposal entails a trade-off. Because appraisal rights are very favorable to shareholders, (26) managers might be induced to act in a reckless way and shareholders might refrain from monitoring them, given that shareholders will be made whole even if the firm goes bankrupt. (27) That is why Listokin and Mun argue that shareholders should be awarded only a fraction of the appraisal value. At the same time, however, insufficiently compensating shareholders for their voting rights would amount to expropriation and would create problems in capital allocation. More specifically, if investors know that they can be stripped of their voting rights without receiving adequate compensation it will be harder for a SIFI to raise capital, especially if it is close to insolvency but still viable. As we show in Part II.D, it is easier to address this trade-off if one builds upon the insights from network theory.

   To summarize, we show that network-sensitive regulatory tools can be used to improve traditional atomistic policy prescriptions designed to mitigate systemic risk, even in an area--corporate governance--where such ideas have so far not been deployed.

   The Article proceeds as follows: Part II describes the traditional microprudential approach to financial regulation, its shortcomings and the rise of macroprudential regulation. Next, it introduces the concept of network-sensitive regulations. Part III develops the core argument by showing how network theory can make SIFI governance prescriptions more effective in curtailing systemic risk. Part IV addresses some potential counterarguments to the use of network theory in that domain. Part V briefly discusses the implications for disclosure and reporting requirements of a network-sensitive approach to financial regulation. Part VI concludes.

2. ATOMISTIC VERSUS...