MONETARY POLICY, FINANCIAL CRISES, AND THE MACROECONOMY [1st ed.] 978-3-319-56261-2, 3319562614, 978-3-319-56260-5

This volume investigates different aspects of monetary policy and prevention of financial crises. It discusses some rece

291 94 9MB

English Pages [349] Year 2017

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

MONETARY POLICY, FINANCIAL CRISES, AND THE MACROECONOMY [1st ed.]
 978-3-319-56261-2, 3319562614, 978-3-319-56260-5

Table of contents :
Front Matter ....Pages i-viii
Monetary Policy, Financial Crises, and the Macroeconomy: Introduction (Frank Heinemann, Ulrich Klüh, Sebastian Watzka)....Pages 1-16
Front Matter ....Pages 17-17
Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard (Charles Goodhart)....Pages 19-26
Optimal Lender of Last Resort Policy in Different Financial Systems (Falko Fecht, Marcel Tyrell)....Pages 27-57
Network Effects and Systemic Risk in the Banking Sector (Thomas Lux)....Pages 59-78
Contagion Risk During the Euro Area Sovereign Debt Crisis: Greece, Convertibility Risk, and the ECB as Lender of Last Resort (Sebastian Watzka)....Pages 79-104
The Case for the Separation of Money and Credit (Romain Baeriswyl)....Pages 105-121
Front Matter ....Pages 123-123
(Monetary) Policy Options for the Euro Area: A Compendium to the Crisis (Sascha Bützer)....Pages 125-162
On Inflation Targeting and Foreign Exchange Interventions in a Dual Currency Economy (Ivana Rajković, Branko Urošević)....Pages 163-176
Macroprudential Analysis and Policy: Interactions and Operationalisation (Katri Mikkonen)....Pages 177-200
Are Through-the-Cycle Credit Risk Models a Beneficial Macro-Prudential Policy Tool? (Manuel Mayer, Stephan Sauer)....Pages 201-224
Assessing Recent House Price Developments in Germany: An Overview (Florian Kajuth)....Pages 225-235
Front Matter ....Pages 237-237
German Unification: Macroeconomic Consequences for the Country (Axel Lindner)....Pages 239-263
Approaches to Solving the Eurozone Sovereign Debt Default Problem (Ray Rees, Nadjeschda Arnold)....Pages 265-295
Appraising Sticky Prices, Sticky Information and Limited Higher Order Beliefs in Light of Experimental Data (Camille Cornand)....Pages 297-306
Rising Income Inequality: An Incentive Contract Explanation (Dominique Demougin)....Pages 307-323
No More Cakes and Ale: Banks and Banking Regulation in the Post-Bretton Woods Macro-regime (Moritz Hütten, Ulrich Klüh)....Pages 325-349
Back Matter ....Pages 351-351

Citation preview

Frank Heinemann · Ulrich Klüh Sebastian Watzka Editors

Monetary Policy, Financial Crises, and the Macroeconomy Festschrift for Gerhard Illing

Monetary Policy, Financial Crises, and the Macroeconomy

Frank Heinemann • Ulrich KlRuh • Sebastian Watzka Editors

Monetary Policy, Financial Crises, and the Macroeconomy Festschrift for Gerhard Illing

123

Editors Frank Heinemann Chair of Macroeconomics Technische UniversitRat Berlin Berlin, Germany

Ulrich KlRuh Darmstadt Business School Hochschule Darmstadt Darmstadt, Germany

Sebastian Watzka IMK - Macroeconomic Policy Institute Hans-Böckler-Foundation Düsseldorf, Germany

ISBN 978-3-319-56260-5 DOI 10.1007/978-3-319-56261-2

ISBN 978-3-319-56261-2 (eBook)

Library of Congress Control Number: 2017951194 © Springer International Publishing AG 2017 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

This volume contains invited contributions by (former) students, colleagues, and friends of Gerhard Illing, whose 60th birthday served as an occasion for collecting these articles. Nearly all contributions were presented in a special birthday symposium. Gerhard Illing’s research focuses on the relation between monetary policy, financial crises, and the macroeconomy. He has often argued that financial and macroeconomic instabilities are a key issue for our societies, an important research topic, and a challenge for macroeconomic policy. He encouraged students and colleagues alike to take the issues of financial crisis prevention and resolution seriously, even at a time when most macroeconomists believed that the great moderation had made crises in mature economies a thing of the past. His pioneering approach combines strong theory to explain causal relationships with a clear view on data and general macroeconomic developments. His proficiency with game theoretic and microeconomic methods has helped him (and others) to advance macroeconomics in novel and very fruitful directions. In particular, he contributed to making mechanism design an important tool for macroeconomic policy analysis. The editors owe Gerhard many thanks for his inspiring views. His open, curious, and analytical mind often pointed us to upcoming research topics, policy debates, and methodological innovations. Many chapters in this volume follow the approach of applying microeconomic and game theoretic methods to monetary policy and financial crises. They also contain interesting empirical results, reflecting Gerhard’s view that evidence antecedes any application of models. They discuss recently suggested measures for central banks’ responses to liquidity shortages and to the liquidity trap. They develop methods for assessing the potential of contagion via the interbank network and for capturing the interaction between micro- and macroprudential regulation. In addition, they contain empirical analyses of macroeconomic effects of German unification and current developments in the German housing market. A wider audience might be especially interested in the chapters that point to avenues for re-conceptualizing and renovating macroeconomics. One potential starting point for such renovation is the application of new microeconomic methods v

vi

Preface

to macro problems. This is reflected in an insurance-based approach to evaluate proposals for solving the sovereign debt problem in the Euro Area. It is also clearly visible in a new explanation for rising income inequality that is based on contract theory and advances in IT technology. Re-conceptualization, however, will also require a more fundamental, transdisciplinary critique of the current state of macroeconomics. Such critique is provided in a detailed analysis of the dogmatic superstructure of the process of financialization, which many believe has been an important driver of the developments in recent decades. The symposium on which this volume is based took place at LudwigMaximilians-University (LMU) in Munich from March 4 to 5, 2016. The conference was characterized by an extremely lively exchange between academics and practitioners, very much in the spirit of Gerhard’s approach to economics. We would like to thank all participants for their participation in the conference and their contributions to this volume. The atmosphere, depth, and policy relevance of the symposium greatly benefited from two policy panels. The panelists (Peter Bofinger, Charles Goodhart, HansHelmut Kotz, Bernhard Scholz, and Hans-Werner Sinn) have done a great job in translating research results into policy advice and to enliven the discussions during sessions and afterward. We thank them for their presence and their inputs. One secret of a successful conference is a generous host providing the necessary infrastructure and a committed team doing the background work. Many thanks go to the Ludwig-Maximilians-University (LMU) for its support and hospitality. It allowed all participants, many of who had spent an important part of their career at LMU, to feel very much at home and at ease. Our special thanks go to Mrs. Agnes Bierprigl and to the other team members at the Seminar for Macroeconomics. Their dedication and effort were crucial to make this event happen and to ensure its success. We also express our thanks to Mr. Alen Bosankic, Ms. Jasmina Ude, and Mr. Moritz Hütten for reading proofs and preparing chapter drafts. The team at Springer Publishing has not only been very patient but also very forthcoming with support and assistance. Finally, it is our pleasant duty to acknowledge financial support from Deutsche Pfandbriefbank and Cesifo. Berlin, Germany Darmstadt, Germany Düsseldorf, Germany

Frank Heinemann Ulrich Klüh Sebastian Watzka

Contents

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . Frank Heinemann, Ulrich Klüh, and Sebastian Watzka Part I

1

Liquidity From a Macroeconomic Perspective

Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . Charles Goodhart

19

Optimal Lender of Last Resort Policy in Different Financial Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . Falko Fecht and Marcel Tyrell

27

Network Effects and Systemic Risk in the Banking Sector .. . . . . . . . . . . . . . . . . Thomas Lux Contagion Risk During the Euro Area Sovereign Debt Crisis: Greece, Convertibility Risk, and the ECB as Lender of Last Resort .. . . . . . Sebastian Watzka

59

79

The Case for the Separation of Money and Credit . . . . . . .. . . . . . . . . . . . . . . . . . . . 105 Romain Baeriswyl Part II

Putting Theory to Work: Macro-Financial Economics from a Policy Perspective

(Monetary) Policy Options for the Euro Area: A Compendium to the Crisis .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 125 Sascha Bützer On Inflation Targeting and Foreign Exchange Interventions in a Dual Currency Economy .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 163 Ivana Rajkovi´c and Branko Uroševi´c vii

viii

Contents

Macroprudential Analysis and Policy: Interactions and Operationalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 177 Katri Mikkonen Are Through-the-Cycle Credit Risk Models a Beneficial Macro-Prudential Policy Tool? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 201 Manuel Mayer and Stephan Sauer Assessing Recent House Price Developments in Germany: An Overview .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 225 Florian Kajuth Part III

Re-Conceptualizing Macroeconomics: An Interdisciplinary Perspective

German Unification: Macroeconomic Consequences for the Country .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 239 Axel Lindner Approaches to Solving the Eurozone Sovereign Debt Default Problem .. . . 265 Ray Rees and Nadjeschda Arnold Appraising Sticky Prices, Sticky Information and Limited Higher Order Beliefs in Light of Experimental Data . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 297 Camille Cornand Rising Income Inequality: An Incentive Contract Explanation . . . . . . . . . . . . 307 Dominique Demougin No More Cakes and Ale: Banks and Banking Regulation in the Post-Bretton Woods Macro-regime . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 325 Moritz Hütten and Ulrich Klüh Greetings from Bob Solow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 351

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction Frank Heinemann, Ulrich Klüh, and Sebastian Watzka

Since the early 1970s, financial instability has been on the rise. For some time this trend had been mainly associated with emerging markets, even though there were occasional crises in some high-income countries as well. In the industrialized world, the increasing instability of economic systems had been masked by the fact that macroeconomic aggregates appeared to become more stable. The subdued fluctuations of the Great Moderation seemed to validate the view that crises and depressions were a thing of the past. This changed in 2007/2008, when a global financial crisis of yet unknown magnitude and character hit the U.S., Europe, and, through spillovers, the whole world. This crisis validated all those who had warned that depressions were still one of the main problems with which economics had to cope. It brought up many new and controversial policy topics that still are not resolved satisfactorily. Also, it has put into question many of the dogmas that had characterized macroeconomic thinking since the late 1970s. Gerhard Illing is at the forefront of those who have constantly argued that financial and macroeconomic instabilities are a key issue for our societies, an important research topic and a challenge for macroeconomic policy. Thus, he is one of those whose views have been validated by the crisis. This volume is a collection of contributions to a symposium held to celebrate Gerhard’s sixtieth birthday.

F. Heinemann Technische Universität Berlin, Berlin, Germany e-mail: [email protected] U. Klüh () Hochschule Darmstadt, Darmstadt, Germany e-mail: [email protected] S. Watzka IMK - Macroeconomic Policy Institute at the Hans-Böckler-Foundation, Düsseldorf, Germany e-mail: [email protected] © Springer International Publishing AG 2017 F. Heinemann et al. (eds.), Monetary Policy, Financial Crises, and the Macroeconomy, DOI 10.1007/978-3-319-56261-2_1

1

2

F. Heinemann et al.

Gerhard’s approach to macroeconomic analysis is unique in the way it balances different perspectives. He is one of the few German economists with an eye for the demand side of the economy. But he also looks at the supply side. He is a skillful microeconomist and he has used his microeconomic expertise frequently to illuminate macroeconomic puzzles. In spite of this ability, Gerhard is a macroeconomist by heart who does not force micro-foundations upon any macroeconomic problems. Finally, he is an economist with a strong preference for academic rigor and policy relevance, and wants to achieve both at the same time. Gerhard’s research interests are multifaceted. He has published and edited books and papers on diverse topics, such as game theory (Holler and Illing 2009), the digital economy (Illing and Peitz 2006), and spectrum auctions (Illing and Klüh 2004). But his main interest in recent years has been (i) the nature and role of liquidity for macroeconomic and financial policies; (ii) the design of policies, instruments, and strategies to cope with the macro-financial problems characterizing modern capitalist societies; (iii) the integration of new methods and views into macroeconomic thinking. This volume is organized along the above three lines of research. Part I deals with liquidity and contagion of liquidity crises. Liquidity becomes a relevant issue through frictions, in particular those analyzed by information economics (Illing 1985). It has many facets, ranging from market and funding liquidity to monetary forms of liquidity. And it has been at the heart of the analysis of financial crises and the optimal response to their occurrence (Illing 2007). Part II looks at policies, in particular those at the nexus between macroeconomics and finance. The crisis has brought about a revival of aggregate demand policies, a trend already foreseen in Illing (1992) and Beetsma and Illing (2005). It has put monetary policy in a very difficult position, caught between macroeconomic and financial stability (Cao and Illing 2015) and faced with the manifold challenges of the zero lower bound (Illing and Siemsen 2016). The crisis has made necessary a re-assessment of fiscal policy (Illing and Watzka 2014) and public debt (Holtfrerich et al. 2015), and it has raised the question of how to complete the re-regulation of the financial sector, with a view to strengthen its macroprudential dimension (Illing 2012). Part III presents approaches for a re-conceptualization and renovation of macroeconomics. The failure of large parts of the economics profession before and during the crisis has made such a re-conceptualization necessary. Economists have trusted too much in efficient markets. As a consequence, they did not warn sufficiently about the imbalances that were building up. During the crisis, they were not able or not willing to prevent the austerity backlash that has kept so many economies in depression mode. Looking for new approaches in macro-financial economics does not mean, however, that everything that has been done before should be disposed of. Those like Gerhard who have studied financial instability before the crisis have come up with important and often surprising insights (see, e.g. Heinemann and Illing 2002; Goodhart and Illing 2001). The problem has not been a lack of good theory, nor of good empirics, but a missing focus on relevant questions.

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

3

1 Liquidity and Contagion of Financial Crises It is difficult to overestimate the role of liquidity as a key or possibly even paramount concept in macroeconomic thinking. Monetary macroeconomics as a discipline could not be constituted without the notion of liquidity. In the history of economic thought, liquidity has been central in constituting different paradigms of macroeconomics. It has informed early discussions of macroeconomic issues, such as in Gresham’s law. It has been central to physiocratic views of the economy, in which some see the beginning of economic thinking in circular flows. The concept of liquidity is closely related to Say’s law (Klüh 2014), and it is one of the main features of Keynesian economics and all “modern macroeconomic” DSGE models. The view on the role of monetary aggregates divides different schools and is a defining element of many controversies regarding monetary policy and financial market regulation. Proponents of real business cycle theory and perhaps growth economics might argue that liquidity and monetary effects are only temporary and the welfare losses arising from fluctuations are small in comparison to the long-run gains of real economic growth. Indeed, if one assumes complete markets and perfect rationality, liquidity is of no major concern. This view, however, has been largely knocked down by recent experience. As soon as one starts to look at the pathologies of capitalist societies, focusing on liquidity becomes inevitable (Goodhart and Illing 2001) because the long-run effects of misdirected investment activities, longrun unemployment, and high youth unemployment rates that are associated with financial crises are estimated to protract growth for several years with no chance of returning to the old growth path. In spite of its overwhelming importance, many economists perceive liquidity as a riddle within an enigma. Trained to think in models in which real exchange dominates, the importance of the nominal dimension of economics that directly follows from the notion of liquidity is often difficult to accept. More importantly, the frictionless or friction-poor world of many models provides only little space for a concept that is largely a consequence of frictions. These frictions are many and most can be traced back to incomplete information. But what is liquidity? And when does it (or a shortage of it) constitute a problem? Charles Goodhart (2017), in the first chapter of this volume, sets out his analysis by asking these fundamental questions. He contextualizes his analysis of lender-of-last-resort (LOLR) policies by first looking at the nature of liquidity problems. Liquidity shortages have a dual nature. On the one hand, a lack of liquidity in most cases reflects some kind of solvency concern: if payments and repayments are certain, both with respect to their incidence and with respect to the details of their occurrence, the ability to borrow ensures liquidity. On the other hand, illiquidity does not necessarily reflect actual solvency problems, because fundamentally solvent banks can become illiquid due to the network effects in financial markets. Goodhart argues that there is no clear cut distinction between solvent but illiquid and insolvent banks.

4

F. Heinemann et al.

Thus, the provision of liquidity during banking crises must compromise two goals: on one hand, systemic crises should be avoided because of the huge losses to society, on the other hand, any implicit guarantee for providing liquidity to banks in distress raises concerns that banks may game the rules and exploit tax payers. Moral-hazard should be avoided. Against this background, determining optimal last resort policies involves difficult judgements. Depending on which of the two views of liquidity shortages is emphasized, very different policy recommendations follow. If liquidity problems are mainly a reflection of solvency problems, policy should be more restrictive. If liquidity problems are a reflection of the inherent fuzziness and non-linearity of the liquidity-solvency nexus, central banks should have maximum flexibility to prevent unnecessary harm to the economy. The standard advice in the literature has been influenced strongly by the first view. To prevent lending to insolvent and thus likely irresponsible players, the central bank should mostly lend to the open market and not to individual banks via LOLR measures. The fear of unwarranted support to failed institutions has also dominated changes in crisis-management arrangements after the crisis, such as the Dodd-Frank act. As a consequence, there is a risk that central banks will have insufficient flexibility when the next crisis comes. Goodhart argues that this underestimates the importance of the second view, and in particular the dynamics of contagion. Provision of liquidity to the market is not helpful to stave off contagious banking crises, because the market allocates extra liquidity to those institutions who are not directly affected by the crisis. While openmarket operations may prevent a complete meltdown, they may leave us with a partial meltdown and severe macroeconomic consequences. Instead, Goodhart recommends that a central bank should treat the first bank to run out of liquidity most toughly up to letting the bank fail, but provide liquidity at more favorable conditions to other banks in distress that may have been affected by contagion. This mechanism raises incentives for banks to avoid illiquidity but saves them from the network effects and, thereby, avoids systemic crises. Nevertheless, any LOLR policy creates moral-hazard incentives. For Goodhart, the only way to properly take this into account would be a much more ambitious approach to change incentives. The rules should be such that they come as close as possible to an unlimited liability arrangement, for example through multiple liability schemes and a much stronger emphasis on bail-in-able debt. The question, whether central banks should provide liquidity to the market or to individual institutions in distress, is also analyzed by Falko Fecht and Marcel Tyrell (2017) in the second chapter of this volume. Building up on a model by Diamond and Rajan (2001), they ask whether the answer may also depend on the nature of the financial system. A key ingredient are the losses that arise if a bank needs to liquidate or sell projects that it cannot continue to finance. Fecht and Tyrell assume that in bank-based financial systems, such as continental Europe, intermediaries have more information about the profitability of projects that they are financing than in a market-based system such as the United States. Bank-based financing allows banks to extract a larger share of the liquidation value of a project, while the market

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

5

value to outside investors is higher in market-based systems, where information is less asymmetric. From these assumptions, Fecht and Tyrell derive a number of results that inform us about differences in LOLR policies between the two systems. They show that the provision of liquidity by open-market operations leads to inefficiencies that are more severe for a bank-based than for market-based system. Providing liquidity to individual institutions is more preferable in a bank-dominated system. The employed model does not account for moral hazard effects that may provide a general argument for open-market operations. LOLR assistance to individual institutions may also be more costly for the central bank. Assuming that these costs are comparable in both systems, Fecht and Tyrell conclude that in bank-based financial systems with their rather illiquid assets, LOLR assistance to individual institutions may be a more favorable instrument than providing liquidity to the market via open-market operations, while the opposite may be true in a marketoriented financial system. The model by Fecht and Tyrell considers contagion via the relative prices of assets in terms of liquidity, but it does not account for contagion arising from direct links between banks. These contagion effects are the reason why Goodhart rejects the clear distinction between insolvency and illiquidity. The dynamics of contagion that are at the heart of Goodhart’s analysis are largely a consequence of the fact that financial systems are complex networks. Should the central bank or supervisor have a very good grasp of the systemic consequences of a specific support measure or punishment, official responses to liquidity problems could be much more targeted. The degree of moral hazard would be reduced and the flexibility of the central bank increased. Moreover, one could start devising incentives to reduce systemically relevant network effects, for example through special rules for money-center banks. In his contribution, Thomas Lux (2017) argues that the pre-2008 mainstream approach to macroeconomic research had “deliberately blinded out” these issues, mainly because of the purported efficiency of financial markets. The post-crisis research on interbank networks and contagion dynamics is becoming more receptive to the alternative view, which emphasizes market inefficiencies, behavioral aspects, non-linearity, and non-standard probability distributions. Lux shows that this literature has yielded a set of important stylized facts ranging from topological features such as core-periphery structures to stability characteristics (such as the surprising persistence of certain linkages). He also recognizes first successes in explaining the self-organization of the system. However, attempts to theoretically measure and then internalize network externalities are in the fledgling stages, at least academically. Thus, the potential for informing policies to change the system’s structure in an attempt to contain contagion remains limited. Lux presents simulations of a stochastic model of link formation and spillovers. An individual default of one bank affects in most cases only few other institutions. But for a small number of banks, their default triggers a system-wide collapse. Most stress tests by monetary authorities have only considered the financial stability of individual institutions and neglected the propagation of liquidity shortages through the banking system. One reason is data limitations. Moreover, as contagion happens

6

F. Heinemann et al.

through a multitude of channels and because balance sheets change quickly, policies grounded in theory may quickly be outdated. What, then, is the role of the new generation of models described in the chapter? According to Lux, network models may help to get a better grasp of the capital cushions needed to prevent shocks and shock transmission in an otherwise fragile system. By focusing on capital buffers, Lux picks up an argument that has been crucial for the crisis response so far: more targeted measures focusing more explicitly on the structural problems would be desirable. However, a lack of knowledge about the impact of these policies precludes their implementation. The second-best method might be to focus on capital, an argument implicit in Illing (2012, p. 17). Sebastian Watzka (2017), in his chapter, considers the liquidity risk again from a different angle. He discusses the euro area debt crisis—and in particular the Greek tragedy—under the assumption that some of the risk premia in Greek government bond yields were due to what the ECB referred to as “convertibility” risk, i.e. the break-up risk of the euro area. This idea has forcefully been demonstrated by De Grauwe and Ji (2013) arguing that an individual euro area member country is naturally lacking a LOLR and this by itself would generate multiple equilibria with unduly high liquidity risk premia for countries of which investors believe that public debt is too high. To test for such effects, Watzka empirically assesses how important non-fundamental contagion was during the early phase of the Greek debt crisis. He concludes that Mario Draghi in his famous 2012 London speech reassured markets that the ECB was in fact acting as LOLR for euro area countries, if certain criteria were met. A crucial and usually innocuous assumption in most papers on banking crises is that money and credit are intrinsically conjoined. Does this need to be the case? The crisis shows that the pursuit of price stability (which had been achieved almost universally before 2007) does not imply financial stability. In contrast, there are important ways in which policies to achieve one can be detrimental to the other. An important reason for this perceived antagonism is the nature of money creation through credit markets. It is therefore not surprising that a radical departure from this approach has been envisioned by some. In his contribution to this volume, Romain Baeriswyl (2017) argues that the close connection between money and credit is a relic of the Gold Standard. With unredeemable fiat money, there are few reasons to stick with it. But what would be the inter-sectoral and inter-temporal implications of such a departure? Baeriswyl argues that the provision of liquidity via the credit market has the largest effects on private credit volume and primarily stimulates demand for goods that are bought on credit such as real estate. Hence, expansionary monetary policy fuels asset prices and may cause price bubbles, along with its stimulus effects for aggregate demand. For targeting consumer price inflation, lump-sum transfers of money to consumers are likely to be more effective. Lump-sum transfers from the central bank to the citizens sound radical at first, but has some important advantages. In Baeriswyl’s view, these advantages strongly outweigh the disadvantages. In particular, the pursuit of price stability would no longer require destabilizing the financial system through credit creation or contraction. Finally, there would be less interference with

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

7

inter-temporal decisions, because interest-rate policies prevent the free adjustment of credit markets to supply and demand of real resources as savings and investment. By contrast, lump-sum transfers of money stimulate demand without directly affecting interest rates. Baeriswyl’s analysis does not stop here. Separating money from credit would have far-reaching implications that go beyond monetary policy. For example, it seems to require a departure from fractional-reserve banking. Lump-sum transfers also require a re-assessment of the way central banks absorb liquidity. Proponents of credit-based money creation often raise three interrelated arguments against its abolition. First, they argue that lump-sum transfers constitute fiscal policy. Because of their distributional consequences, transfers need to be decided upon by elected officials, not technocrats. Second, they believe that the current system is better than often assumed in bringing investment and savings in balance. Has it not allowed economic growth for large spells of the last two centuries? Third, they question the need to focus so much attention on central bank policy. If fiscal policy is proactive, a credit-based monetary system can work smoothly. Fiscal policy takes center stage in absorbing excess liquidity and savings and in making sure that investment expenditures are sufficient. It can also take the necessary steps to prevent or escape a liquidity trap. Unfortunately, European fiscal policy currently appears rather dysfunctional: it neither uses the opportunity of a huge excess supply of savings and demand for safe assets to boost public investment, nor does it exploit the large multiplier effects of fiscal policy in a liquidity trap for stimulating demand. This has raised a discussion for helicopter money as an additional instrument for central banks. Baeriswyl just goes one step beyond and suggests to replace the credit channel completely by a helicopter.

2 Putting Theory to Work Macroeconomics is a policy-orientated science. A main challenge is to take theory and empirical scrutiny as far as possible while always having policy in mind. Bringing cognitive interest and policy relevance together has always been a hallmark of Gerhard Illing’s thinking. This has been most visible during the symposium that has given rise to this volume. A frequent comment of participating central bankers was that if academic conferences would always be so interesting, they would have rather remained in academia. While all three parts of this book reflect this practical side of macroeconomics, this section puts special emphasis on it. Financial markets and institutions are not just playing a dominant role in transmitting monetary policy to the real sector. In recent years, they have often absorbed policy impulses. Macroeconomic policy feeds into the peculiar logic of expansion and contraction that increasingly characterized the financial sector. From a certain point on, however, periods of financial contraction become a source of fiscal and growth risk. Finance, thus, simultaneously charges and discharges policy.

8

F. Heinemann et al.

As fiscal policy has taken a backseat since the beginning of the 1970s, monetary policy has found itself in the center of this double role. It faces a difficult conflict. On the one hand, it tries to fulfill its role as a levee against the negative real consequences of financial contraction. On the other, it tries to enclose the dangers of excessive financial expansion. As the instruments to achieve the first may inhibit or even foil the instruments available to achieve the second, a conflict emerges. An intriguing analysis of this conflict and its relation to liquidity issues is provided in Cao and Illing (2010, 2011). The challenges for monetary policy are all the more acute when fiscal policy becomes increasingly passive. This is most obvious in the case of the Euro crisis, which is surveyed and analyzed in the first chapter of the second part. Here, Sascha Bützer (2017) first illustrates the dramatic failure of fiscal policy. The institutions of the European Monetary Union lack mechanisms to pool risks across its member states and put the burden of adjustment on these national states while stripping them of some of the most effective instruments to achieve these adjustments, like national interest and exchange rates. Integrated financial markets would be an alternative to fiscal risk pooling, but financial integration stopped short of the standards achieved in other currency areas. Apparently, several member states have been overcharged by these demands. An almost religious belief in austerity and structural reform has prolonged the recession. It has led to an increase in indebtedness and thus defeated itself. Finally, it has pushed monetary policy in a situation that is perceived as an overburdening of its possibilities and mandate. In Bützer’s view, monetary policy has been the victim of a cure that has nearly proven fatal. While the detrimental effects of fiscal contraction were recognized by many monetary policymakers, structural reforms have been viewed at as “a panacea to jump-start growth and generate employment” (p. 143). Against the backdrop of hysteresis, the combination of procyclical fiscal, impotent structural and insufficient monetary policy is now yielding medium- to long-term effects. After describing the current situation, Bützer looks at options available now. He analyses their potential in keeping the Euro area together and leading the way out of depression. Simultaneously, he asks whether the expansionary effects of these policies are outweighed by their disadvantages in terms of financial stability and redistributive effects. He concludes that conventional monetary policy and quantitative easing “have run out of steam at the zero lower bound and increasingly pose risks to financial stability, the outright creation of broad money through lump-sum transfers from the central bank to private households may well be the most effective measure to achieve the Eurosystem’s primary objective and lift the economy out of its slump” (p. 155). He recognizes that there are dangers associated with putting the central bank in such an exposed position. In the end, however, he favors managing credibility, independence, and financial stability risk to letting the Eurozone unravel. Bützer’s analysis illustrates the ever expanding universe of central bank instruments. This points to a policy challenge that emerging-market central banks have already faced long before the crisis. In these countries, monetary policy has often

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

9

been characterized by the use of multiple instruments. Sometimes this has been due to multiple objectives. In other cases, central banks have felt that a combination of instruments might be preferable to achieve a single goal. Using the example of foreign-exchange-market interventions, Ivana Rajkovi´c and Branko Uroševi´c (2017) develop a framework to analyze this multiplicity. The context is a small open economy with pronounced euroization. It follows an inflation-targeting strategy. In such a dual currency setup, the degree to which foreign currency is employed to store value or extend credit affects how the policy rate is set. If interest rates are the only instrument, monetary policy faces constraints that can be relaxed by foreign-exchange interventions. The responses to domestic and international shocks become less extreme and policy is less distortionary. However, to successfully operate with different instruments requires pre-conditions. In particular, central-bank risk management needs to be developed further to take into account the cost of foreign exchange interventions. Furthermore, monetary and macroprudential polices have to be calibrated jointly. This important take-away from the chapter of Rajkovi´c and Uroševi´c is further refined in the next three contributions of this volume that deal with the conceptual basis, measurement, and data requirements of macroprudential regulation. Katri Mikkonen (2017) reviews recent contributions to macroprudential policy analysis. She first looks at the relationship among macroprudential, monetary and microprudential policies, emphasizing synergies and the need to focus on comparative advantages. In a second part, she presents an operationalization of macroprudential policy. Recent work at central banks has come up with new ways of risk identification and assessment. With a view to get a holistic picture of macro-financial risks, qualitative and quantitative techniques have been married in innovative ways, for example in novel early warning systems. Recent work has also come up with new views on macroprudential instruments, for example countercyclical capital buffers, loan-to-income ratios, or a time-varying net stable funding ratio. Mikkonen concludes that much has been done to improve macroprudential policy. However, policies so far cannot be based on a stable set of stylized facts and instruments. The financial cycle has received less attention than the business cycles. Missing data and tools to model complexity in quickly changing systems limit the applicability of many models. “There is no universally accepted dogma for macroprudential policy” (p. 196). Trial and error will remain important elements of existing policy approaches. Much more empirical research needs to be carried out. Manuel Mayer and Stephan Sauer (2017), in their contribution, study macroprudential aspects of measuring credit risk. Though the practice is currently contested, banks use their own estimates for the probability of default and the loss given default. The respective models follow different approaches. Accordingly, an important distinction with macro-financial relevance is the one between point-intime (PIT) models (using all currently available information) and through-the-cycle (TTC) models (canceling out information that depends on the current position in the macro-financial cycle). TTC models are often viewed as favorable for

10

F. Heinemann et al.

macroprudential regulation, because credit risk estimates do not improve (deteriorate) in a boom (recession). Thereby, constant equity requirements are less procyclical than if risk weights need to be adjusted when risk is measured by PIT models. Mayer and Sauer question the perceived superiority of TTC, performing a range of empirical tests on the relative reliability of the two methods. They show that TTC are more difficult to validate. Having a theoretically good but empirically questionable method might do more harm than good. It also opens the door for misunderstandings between the supervisors and the supervised. Taken together, their arguments favor PIT models for measurement purposes. To compensate for the pro-cyclical nature of these models, the authors argue for a more extensive use of counter-cyclical capital buffers. Florian Kajuth (2017) concludes Part II with a discussion of a current macroprudential topic, the rise in house prices, in particular in German urban agglomerations. House price developments are crucial to understand macro-financial dynamics (Illing and Klüh 2005). The analysis looks at German house prices from at least two different angles. One the one hand, it discusses issues of data availability and quality, comparing parametric and non-parametric approaches. In this way, it raises awareness for an often neglected but extremely important issue: the availability (or lack thereof) of data for macroprudential and other policy purposes. On the other hand, the chapter asks whether there is reason for concern. Did expansionary monetary policy result in substantial overvaluations, thus giving rise to prudential concerns? Kajuth provides extensive evidence for the deplorable state of property price statistics in Germany. In particular, there is a lack of time series that go back sufficiently in time. Moreover, existing statistics lack comprehensiveness. It is therefore necessary to rely on cross-sectional variations of housing markets in Germany. Using this information and a range of other sources confirms that some urban areas do indeed seem to be overvalued. For Germany as a whole, however, there is no indication of a bubble, at least not yet.

3 Re-conceptualizing Macroeconomics The financial crisis has left a deep mark on the kind of topics that are on macroeconomists minds. New methods have evolved, and macroeconomic issues have become more interesting to those who were previously focused on microeconomics. Macroeconomics has changed quite a deal since Lucas’s now infamous quote that “depression prevention has been solved” (Lucas 2003, p. 1). The chapters in this volume reflect some of these developments. Macroeconomics is currently undergoing a period of re-conceptualization (Blanchard et al. 2010). This period started already before the crisis, but went largely unnoted, with few exceptions, such as the ones discussed in Beetsma and Illing (2005). The final section of this

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

11

volume looks at five elements of this trend: – A renewed focus on stylized facts, economic history and path dependence, – The application of established methods to new problems, such as the institutional structure of the Euro area, – The application of new methods to old topics, building in particular on insights from behavioral and experimental economics, – The resurgence of distributional issues as a topic of macroeconomic research, and – The emergence of inter-disciplinary work to re-embed economics in the social sciences and contextualize its findings. Axel Lindner (2017), in the first chapter of Part III, shows that going back a little further can yield important insights about the present situation. He looks at the macroeconomic effects of German unification and argues that the German economy had been off steady state already before unification. At the same time, Germany seems to have been on a trajectory that very much resembles the dynamics that we now associate with the anamnesis of the Euro crisis. In particular, investment was trending down already before unification, and continued to do so after a brief jump in the beginning of the 1990s. Moreover, the financial balance had been on an increasing trend during the eighties, a trend it returned to around 10 years after unification. The wage share in national income follows a similar pattern, yet with the opposite sign. These observations cast some doubts on the view that these developments were a consequence of introducing the Euro. The problems of the Euro area are at the core of the chapter by Ray Rees and Nadjeschda Arnold (2017). They ask whether insurance-based approaches can help solving the sovereign default problem and argue that the economics of insurance markets can guide a redesign of the common currency area. This redesign seeks to preserve decentralized fiscal policy. Its main idea is to use risk-based insurance premia as an instrument to increase fiscal discipline. Rees and Arnold encourage the creation of an independent insurance agency. This agency ensures incentive compatibility by promising to remove the threat of sovereign default if certain conditions are fulfilled. Its main instrument are risk-based premia “payable ex ante into a mutual fund that must at least break even in expectation” (p. 267). In case of a fiscal emergency, the mutual fund arranges automatic payouts. Regular reviews of fiscal plans, minimum insurance reserves, and reinsurance arrangements complement the set-up. Rees and Arnold compare this insurance-based approach with the existing European Stability Mechanism and different suggestions for Eurobonds. They conclude that none of these alternatives is incentive compatible, because they fail to make the costs of default risk accountable for governments ex ante. Camille Cornand (2017) shows in her contribution that new empirical approaches can yield important insights about macroeconomic phenomena. In an attempt to provide additional foundations for the non-neutrality of money, she compares the role of three potential explanations for nominal rigidities: sticky prices à la Calvo (1983), sticky information à la Mankiw and Reis (2002), and limits to the level

12

F. Heinemann et al.

of reasoning that price setters achieve. The latter is based on the observation that subjects in laboratory experiments fail to reach common knowledge when information abounds. Cornand uses the data from an experiment by Davis and Korenok (2011), in which subjects play the role of price-setting firms in a macro-environment with stochastic demand shocks. The data reveal a sluggish adjustment to shocks, even if these shocks are publicly revealed. Cornand investigates which model yields the best fit of these price adjustments and finds that the sticky-information model fits best. Selecting models on the basis of laboratory experiments provides an alternative to assuming artificial frictions in macroeconomic models. Experimental data also allow estimating behavioral parameters independently from other model parameters, while empirical tests with macroeconomic field data allow only a joint estimation of all model parameters. The estimated behavioral parameters may then be used for calibrations and as restrictions in the joint estimates of other model parameters with macroeconomic field data. One should not underestimate the significance of these and other behavioral insights into wage and price stickiness. The rejection of Keynesianism by Lucas and others was largely justified with the argument that Keynesians were unable to derive such stickiness from micro-founded models with optimizing agents. The data from experiments and the results from behavioral economics more generally show that nominal rigidities and non-rational expectations are just a fact of life. This makes pragmatic reasoning much easier, as it is not hampered anymore by the requirement that all macroeconomic variables need to be derived from rational choices. In the penultimate chapter, Dominique Demougin (2017) analyses an issue that more and more dominates the policy debate. After having long been relegated to the fringes of macroeconomics, the rising inequality of income and wealth now takes center stage. Using an incentive contract approach, Demougin provides a novel explanation for this trend. Information and communication technology allows managers to better monitor worker behavior. This redistributes informational rents from the bottom to the top of the income distribution. While middle management wins, firm owners win big. They do not just gain from a redistribution of rents from a given output: they also benefit from increased worker effort and productivity. The mirror image of this effect is that workers are penalized twice. They lose the rents that they had enjoyed before, and they suffer from a work environment that requires higher effort. Demougin uses a standard hidden action problem to explain increasing income inequality. The argument is solely based on the organizational structure of the firm and, thus, provides an alternative to standard explanations based on globalization or skill-biased technological progress. Demougin’s numerical exercise replicates a sizeable number of crucial features of the macroeconomic environment since the early 1970s. While technology advances, wages dynamics are at best subdued, if not stagnant. The wage share in income declines. Working conditions are increasingly resembling a treadmill with little space for discretionary decisions by workers. Certain groups of the society are able to keep up as middle managers, so the wage

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

13

distribution starts to become more uneven. But the strongest implication of the rise in information and communication technology is that the very top of the income and wealth distribution experiences large gains, a feature that cannot be explained by skill-biased technological progress. Moritz Hütten and Ulrich Klüh (2017), in the final chapter of this volume, pick up the fact that macroeconomic developments since the end of the Bretton-Woods regime display peculiar characteristics. Not only has there been a redistribution from the bottom to the top and from labor to capital, but in parallel, inflation has come down and is too often close to deflationary levels. Unemployment has become a constant feature of capitalist societies, while it was largely absent in the decades before. Public debt as a share of GDP has trended up, in part because the incidence of financial crisis has increased continuously. Exchange rates and other prices on financial markets have exhibited a degree of volatility seemingly unrelated to fluctuations in fundamental variables. All this has taken place in a context in which the task of stabilizing macroeconomic and financial fluctuations has been concentrated in the hands of central banks. These, in turn, have largely bought into the notion that some degree of unemployment is necessary to keep inflation in check, in particular the very low inflation targets that have become standard. Fiscal policy has been confined to implement a regime of institutionalized austerity (Streeck and Mertens 2010). And structural policies have often followed the prescriptions of the so called Washington consensus. Hütten and Klüh argue that the beginning of the 1970s is a watershed between two ways of organizing economic activity in capitalist societies. The end of the Bretton-Woods system did not only change the way exchange rate movements and international capital flows are organized. A “regime change” occurred that led to a dynamic adjustment of capitalism, in which finance becomes increasingly important (financialization). Regrettably, there have been only few attempts to characterize these two phases of economic history holistically. The chapter first introduces the concept of “macro regimes” as a framework for analyzing macroeconomic aspects during periods of large social transformations. Building on approaches from political science and sociology, macro regimes are defined as arrays of implicit or explicit principles, norms, rules and decisionformation procedures that lead to a convergence of actor expectations. Both the convergence of expectations (i.e. the emergence of regimes) and the divergence of expectations (which usually marks the beginning of a regime change) are reflected in specific characteristics of time series. In the view of many observers from other social sciences, a characteristic feature of the macro regime in the last four decades is the increasing role of finance in society. This element of the current macro regime, often coined financialization, is the focus of the chapter. Can the macro regime approach itself explain financialization? What does financial sociology contribute to its understanding? And how could financialization happen on the watch of economic experts that now frequently reject it? Thereby, this volume ends with a reflection on the roles that economics in general and macroeconomics in particular play in our society. This question has also been

14

F. Heinemann et al.

characteristic for the symposium held in honor of Gerhard Illing and for Bob Solow’s letter at the very end of this book. Solow asks: “Why is it so difficult?” referring to the combination of expert technique with common sense in economics. One explanation might be that economics is faced with a difficult double role. On the one hand, it can be considered a science (the objective of which is to distinguish true and false). On the other hand, it is a toolbox for policy. Put differently, it is a language that is employed within the economy to organize discourse about the economy. In this second role, it is highly political, applied, sometimes useful, and sometimes counterproductive. Gerhard Illing has taught many people how to walk on the fine line between academic scrutiny and policy relevance that emerges from this double role.

References Baeriswyl, R. (2017). The case for the separation of money and credit. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 105–121). Cham: Springer. Beetsma, R., & Illing, G. (2005). Revival of aggregate demand policies – Introduction. CESifo Economic Studies, 51, 497–509. Blanchard, O., Dell’Ariccia, G., & Mauro, P. (2010). Rethinking macroeconomic policy. Journal of Money, Credit and Banking, 42(s1), 199–215. Bützer, S. (2017). (Monetary) Policy options for the euro area: A compendium to the crisis. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 125–162). Cham: Springer. Calvo, G. (1983). Staggered prices in a utility maximizing framework. Journal of Monetary Economics, 12, 383–398. Cao, J., & Illing, G. (2010). Regulation of systemic liquidity risk. Financial Markets and Portfolio Management, 24(1), 31–48. Cao, J., & Illing, G. (2011). Endogenous exposure to systemic liquidity risk. International Journal of Central Banking, 7, 173–216. Cao, J., & Illing, G. (2015). ‘Interest rate trap’, or why does the central bank keep the policy rate too low for too long? The Scandinavian Journal of Economics, 117(4), 1256–1280. Cornand, C. (2017). Appraising sticky prices, sticky information and limited higher order beliefs in light of experimental data. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 297–306). Cham: Springer. Davis, D., & Korenok, O. (2011). Nominal price shocks in monopolistically competitive markets: An experimental analysis. Journal of Monetary Economics, 58, 578–589. De Grauwe, P., & Ji, Y. (2013). Self-fulfilling crises in the Eurozone: An empirical test. Journal of International Money and Finance, 34, 15–36. Demougin, D. (2017). Rising income inequality: An incentive contract explanation. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 307–323). Cham: Springer. Diamond, D. W., & Rajan, R. (2001). Liquidity risk, liquidity creation, and financial fragility: A theory of banking. Journal of Political Economy, 109, 287–327. Fecht, F., & Tyrell, M. (2017). Optimal central bank policy in different financial systems. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 27–58). Cham: Springer.

Monetary Policy, Financial Crises, and the Macroeconomy: Introduction

15

Goodhart, C. (2017). Balancing lender of last resort assistance with avoidance of moral hazard. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 19–26). Cham: Springer. Goodhart, C., & Illing, G. (2001). Financial crises, contagion and the lender of last resort: A reader. Oxford: Oxford University Press. Heinemann, F., & Illing, G. (2002). Speculative attacks: Unique sunspot equilibrium and transparency. Journal of International Economics, 58(2), 429–450. Holler, M., & Illing, G. (2009). Einführung in die Spieltheorie (7. Auflage). Berlin: Springer. Holtfrerich, C., Feld, L., Heun, W., Illing, G., Kirchgässner, G., Kocka, J., Schularick, M., Streeck, W., Wagschal, W., Walter, S., & Weizsäcker, C. (2015). Staatsschulden: Ursachen, Wirkungen und Grenzen (Bericht einer Arbeitsgruppe im Auftrag der Nationalen Akademie der Wissenschaften Leopoldina). Berlin: Union der deutschen Akademien der Wissenschaften e. V. Hütten, M., & Klüh, U. (2017). No more cakes and ale: Banks and banking regulation in the postbretton woods macro-regime. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 325–349). Cham: Springer. Illing, G. (1985). Geld und asymmetrische Information. Studies in Contemporary Economics 13. Berlin: Springer. Illing, G. (1992). Neue Keynesianische Makroökonomie. Tübingen: Mohr-Siebeck. Illing, G. (2007). Financial stability and monetary policy – A framework (CESifo Working Paper No. 1971). April 2007. Illing, G. (2012). Finanzmarktstabilität – die Notwendigkeit eines effizienten Regulierungsdesigns. In M. Held, G. Kubon-Gilke, & R. Sturn (Hg.), Lehren aus der Krise für die Makroökonomik “Jahrbuch Normative und institutionelle Grundfragen der Ökonomik” Band 11, 2012:283-306. Illing, G., & Klüh, U. (Eds.). (2004). Spectrum auctions and competition in telecommunications. Boston, MA: The MIT Press. Illing, G., & Klüh, U. (2005). Vermögenspreise und Konsum: Neue Erkenntnisse, amerikanische Erfahrungen und europäische Herausforderungen. Perspektiven der Wirtschaftspolitik, 6(1), 1–22. Illing, G., & Peitz, M. (2006). Industrial organization and the digital economy. Cambridge, MA: The MIT Press. Illing, G., & Siemsen, T. (2016). Forward guidance in a model with price-level targeting. CESifo Economic Studies, 62(1), 47–67. Illing, G., & Watzka, S. (2014). Fiscal multipliers and their relevance in a currency union – A survey. German Economic Review, 15(2), 259–271. Kajuth, F. (2017). Assessing recent house price developments in Germany – an overview. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 225–235). Cham: Springer. Klüh, U. (2014). Sismondis Spur: Krisen- und Selbstverständnis der Ökonomik. In M. Held, G. Kubon-Gilke, & R. Sturn. Normative und institutionelle Grundfragen der Ökonomik. Marburg: Metropolis Verlag. Lindner, A. (2017). German unification: Macroeconomic consequences for the country. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 239–263). Cham: Springer. Lucas, R. (2003). Macroeconomic priorities. Presidential address delivered at the one-hundred fifteenth meeting of the American Economic Association, January 4, 2003. Accessed August 24, 2016, from http://pages.stern.nyu.edu/~dbackus/Taxes/ Lucas%20priorities%20AER%2003.pdf Lux, T. (2017). Network effects and systemic risk in the banking sector. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 59–78). Cham: Springer. Mankiw, G., & Reis, R. (2002). Sticky information versus sticky prices: A proposal to replace the new Keynesian Phillips curve. Quarterly Journal of Economics, 117, 1295–1328.

16

F. Heinemann et al.

Mayer, M., Sauer, S. (2017). Are through-the-cycle credit risk models a beneficial macroprudential policy tool? In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 201–224). Cham: Springer. Mikkonen, K. (2017). Macroprudential analysis and policy – Interactions and operationalization. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 177–200). Cham: Springer. Rajkovi´c, I., & Uroševi´c, B. (2017). On inflation targeting and foreign exchange interventions in a dual currency economy. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 163–176). Cham: Springer. Rees, R., & Arnold, N. (2017). Approaches to solving the eurozone sovereign debt default problem. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 265–295). Cham: Springer. Streeck, W., & Mertens, M. (2010). Politik im Defizit: Austerität als fiskalpolitisches Regime (MPIfG Discussion Paper 10/5). Watzka, S. (2017). Contagion risk during the euro area sovereign debt crisis: Greece, convertibility risk, and the ECB as lender of last resort. In F. Heinemann, U. Klüh, & S. Watzka (Eds.), Monetary policy, financial crises, and the macroeconomy: Festschrift for Gerhard Illing (pp. 79–104). Cham: Springer.

Frank Heinemann is professor of macroeconomics at the Berlin University of Technology. His main research interests are monetary macroeconomics, financial crises, and experimental economics. Ulrich Klüh is professor of economics at Hochschule Darmstadt. His main research interests are macroeconomic theory and policy, central banking, financial markets and institutions, and history and theory of economic thought. Sebastian Watzka is senior economist at the Macroeconomic Policy Institute (IMK) at the Hans-Böckler-Foundation. Before joining the IMK he was assistant professor at the Seminar for Macroeconomics of the University of Munich, LMU. His research interests are monetary policy and financial markets, financial crises, inequality and unemployment.

Part I

Liquidity From a Macroeconomic Perspective

Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard Charles Goodhart

Abstract Solvency is rarely clearly defined, since it depends on valuations relating to future outcomes, which are themselves affected by policy decisions, including Central Bank Lending of Last Resort (LOLR). Positive LOLR may cause losses and moral hazard, whereas refusal could trigger a contagious panic. Measures to limit moral hazard, and hence allow more systemic protection include: (i) treating the first failure more strictly; (ii) involving other banks in any rescue; (iii) toughening the incentive structure for bank borrowers.

1 Introduction If an agent is certain to repay her debts, on time and meeting all the required terms and covenants, she can always borrow at current riskless market interest rates. So a liquidity problem1 almost always indicates deeper-lying solvency concerns. The solvency concerns that lenders may have about borrowers may, or may not, however, be well founded. I start in Sect. 2 by noting that the definition of solvency is fuzzy. The future likelihood of a borrower defaulting is probabilistic, and so the terms (the risk premia) and conditions on which a borrower can raise cash, her access to liquidity, are stochastic and time varying, Sect. 3. There is a common view that a Central Bank should restrict its activities in support of financial market stability to lending into the general market via open market activities, rather than lending to individual banks via Lender of Last Resort (LOLR) measures. I explain why I disagree with that argument in Sect. 4. Nevertheless the banks most in need of LOLR will generally be those that have been least prudent. Even though the Central Bank will choose not to support the most egregiously badly-behaved (and/or those whose failure is least likely to generate

1

That may be defined as an inability to access cash to meet due outflows, except perhaps at enhanced premia that reveal existing solvency concerns to a wider public. C. Goodhart () London School of Economics, Financial Markets Group, London, UK e-mail: [email protected] © Springer International Publishing AG 2017 F. Heinemann et al. (eds.), Monetary Policy, Financial Crises, and the Macroeconomy, DOI 10.1007/978-3-319-56261-2_2

19

20

C. Goodhart

secondary contagious-failures), the use of LOLR does entail a degree of insurance (against failure) and hence generates moral hazard. I discuss in Sect. 5 various ways of mitigating such moral hazard.

2 The Meaning of Solvency? The use of language in macro-economics is slipshod2 ; (perhaps this helps to explain our penchant for arid mathematical models). Solvency is just such a slippery term. We think that we know what it means, i.e. that the value of assets is greater than the valuation of the liabilities. But in practice we do not, because it all depends on how the assets (and liabilities) are valued, and that depends on the viewpoint of the valuer, and also on the (changing) conventions and practices of the accountant. Consider, for example, the British mortgage bank (Northern Rock) in September 2007, at the time when it asked the Bank of England for liquidity assistance. Looking backwards, to the prior bubble phase, it had very few non-performing loans, and was undoubtedly solvent (historic cost accounting). Looking forwards, to the likely future bust phase in housing, it was most probably insolvent (since it had expanded aggressively), as turned out later to be the case. Moreover, the assessment of the solvency of an institution, especially one seeking LOLR assistance from a Central Bank (CB), is not independent of the CB’s own actions and of the wider public’s (the market’s) interpretation of those same actions, (as in the case of Northern Rock).3 The valuation of a going concern (where any help has been covert) is much greater than that of a concern, which is either gone or

2

Examples are:

1. ‘Real’, as in real interest rates: Really means ‘adjusted for (expected) price changes’, but whose expectations and what prices? Not much ‘real’ about it; at best ‘uncertainly measured adjustment for future price changes’. 2. ‘Natural’, as in the natural rate of unemployment. Really means the level at which some other variable, e.g. inflation, would remain stable. Often treated as being synonymous with ‘equilibrium’, but equilibrium carries a connotation that there are forces restoring such an equilibrium, once disturbed. This latter remains contentious. 3

When LOLR assistance to a bank is revealed, the reaction can either be one of relief, i.e. the Central Bank will now restore order, or of greater concern, i.e. I did not know things were so bad. In the case of Northern Rock, Robert Peston of the BBC leaked that LOLR and had no incentive to calm the public. Moreover, Northern Rock had many depositors who interacted electronically. When a large number of these sought to withdraw simultaneously, the Northern Rock website crashed. The depositors interpreted this as a refusal of Northern Rock to allow withdrawals, and physically ran to do so from their nearest branch. Similarly when the authorities, e.g. the Treasury, guarantees the withdrawal value of an asset, as in the Irish bank deposits or US Money Market Mutual Funds, in 2008, this may calm the situation so that no further supporting action is needed. But alternatively, if the potential financial losses are feared to be large and the solvency of the guarantor is itself questionable, as it was in the Irish case, this can lead to both entities, guarantor and guarantee, dragging each other down, in a ‘doom loop’.

Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard

21

needed patent public help to continue; hence there is a serious stigma effect of being observed to need LOLR assistance from the Central Bank, with potentially severe effect in delaying and distorting recovery processes. Accountants have their own incentives. Although it is a crime to continue trading when knowingly insolvent, I am unaware of any bank having closed its doors because its accountant told them to do so. But, once a bank does close, most often because its liquidity problems become insuperable, the incentive of an incoming forensic accountant will be to exaggerate the potential scale of problems, thereby fuelling potential panic and risk aversion, because such an accountant will not want to have to claw back money from creditors on a future occasion, to meet further losses, if, indeed, such future claw back can be done at all. Too often creditors are originally told to expect large losses, whereas, after several years and a recovery from the crisis, they get paid back practically in full, (e.g. as in Lehman Bros London).

3 A Liquidity Problem Is a Solvency Problem So (forward-looking) solvency problems exhibit themselves as liquidity problems. Borrowers, including banks, would, as a generality, not have a problem in gaining (funding) liquidity from markets if they were perceived as absolutely certain to pay back in full as contracted. There are a very few technical exceptions, where timing, terrorism, IT breakdowns, as with Bank of New York in the 1980s, or some other exogenous event prevents access to markets; but the common, heuristic rule is that a shortage of liquidity presages (market) concerns about solvency. The CB then has to balance concerns both about potential loss from LOLR lending and the implications of being seen to support risk-loving, even reckless, management on the one hand (if it does support), against concerns about fuelling the panic, amplifying downwards pressures on asset prices and contagion on the other, (if it does not support). It is a difficult act of judgment, and there are no absolute clear rules. In particular, the ‘solvency’ of any potential borrower is not a deterministic, exogenous, knowable datum, but depends on many time-varying future developments, not least how the CB itself responds to requests for LOLR assistance, and whether (and how) that becomes known. There is a most unhelpful misinterpretation of Bagehot (1873) that contends that he claimed that the Bank of England should only lend to solvent institutions. But the Bank of England had no supervisory powers then, or the right to inspect other financial institution’s books. So how could the Bank of England know who was solvent, and who not? Instead, what he meant, and said clearly, in his second rule for LOLR is that the Bank of England should lend freely on all ‘good

22

C. Goodhart

securities’.4 The criterion for Bagehot was the quality of the collateral, which could be assessed,5 rather than the solvency of the borrower, which could not be.

4 Lend to the Market, not to an Individual Borrower? There is a common view, more prevalent in the USA than in Europe, that the authorities, including the CB, should intervene as little as possible in markets, and/or that markets are better informed (efficient market hypothesis) than any authority can be, (despite CB’s role as supervisor). If so, so it is asserted, in a panic the CB should provide liquidity to the market as a whole via open market operations, and leave the distribution of such liquidity to the market, which will sort out those deserving of support from those who should be let go, (better than CB). This is, I believe, wrong, because it fails to grasp the dynamics of contagion. In a panic, the weakest is forced to close. Its failure will worsen the crisis. The market will then withdraw funds from the next weakest, further amplifying the downwards spiral. To prevent total collapse at some point the authorities will have to step in to support every institution which can meet certain criteria, as the G20 did in October 2008. Bagehot’s criterion was the availability of ‘good collateral’. Such was the political revulsion from public sector support, ‘bail out’, of the banking sector in the USA, that the conditions under which the Fed could provide liquidity support to individual financial institutions were made somewhat more

4

‘The great majority, the majority to be protected, are the “sound” people, the people who have good security to offer’, p. 198, (1999 version: John Wiley: NY. HG3000. L8283). 5 But if the collateral was good, why could not a bank raise money on the open market? There are two answers to this, the first being more applicable to the nineteenth century, the second more to subsequent centuries, twentieth and twenty-first. First, during panics financial markets tend to become dysfunctional, with no one being prepared to part with cash at any reasonable price. In such circumstances, the Central Bank is not only the Lender of Last Resort, but also the market maker of last resort. In such a situation what interest rate should it set? As Bagehot states, a ‘high’ one, but obviously not a ‘penalty’ rate. Bagehot never uses the word ‘penalty’ in this context. Second, such has become the stigma of being seen to borrow on LOLR terms from the Central Bank that banks tend to use up all their good quality collateral to borrow from the market, before turning, if all else fails, to the Central Bank for succour. With banks also of the view, prior to 2007–2009, that they could always borrow cash in wholesale markets (funding liquidity), they had run down their holdings of high quality liquid assets to almost nothing at the start of the Great Financial Crisis. So amongst the various unconventional monetary measures then taken were those that swapped less liquid assets (held by banks) for more liquid assets, e.g. Treasury Bills. The Bank of England’s Special Liquidity Scheme is a prime example. In the aftermath of the Great Financial Crisis various requirements have been put in place, such as the Liquidity Coverage Ratio, to try to ensure that banks will always have enough high quality liquid assets to enable banks to be rescued from a panic, and associated liquidity troubles, without forcing the Central Bank to choose between accepting poor collateral, i.e. taking a credit risk, and letting that bank fail.

Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard

23

restrictive under Title XI, Sections 1101–1109, of the Dodd-Frank Act, passed by the House of Representatives in H.R. 4173, pp. 738–752, passed in 2011. Under this, A. Section 13.3 lending, under which previously the Fed could lend to anybody, not just to banks, under ‘unusual and exigent’ circumstances has been curtailed. In future the Fed can only lend to eligible banks, and/or to “any participant in any program or facility with broad-based eligibility”. What does this mean in practice? B. The Fed cannot now lend to ‘insolvent’ borrowers; though [the CEO of] the borrowing bank may certify the solvency of her bank, with a duty to update any material information on such solvency. C. More information on such emergency liquidity assistance has to be provided, and sooner. Also provision of additional guarantees to depositors and other creditors of financial institutions can only be provided after a ‘liquidity event’ is agreed by the Federal Reserve Board, Federal Deposit Insurance Corporation and the Executive (President and Secretary of the Treasury). This must then be accepted by both Houses of Congress. All this could make emergency liquidity assistance in a crisis less flexible, and make the Federal Reserve Board’s freedom of action constrained by legal interpretation of the Dodd-Frank Act. Would, for example, the Fed be expected to audit the books of a potential borrower, prior to granting emergency liquidity assistance, or could it rely on the borrower’s self-certification? If Bank A borrowed from the Fed in, say, May 2017 and then subsequently went into bankruptcy in July 2017, would there be (political) penalties, and, if so what, on the Fed and/or the self-certifier? It is my view that the Dodd-Frank Act has already imposed undesirably rigid constraints on the Fed’s flexible freedom of manoeuvre to respond to financial crises, though this is contentious. The Warren/Vitter Bill would have made such constraints much tighter, and the Fed has accepted, in November 2015, that ‘broadbased’ means at least five participants. The contrasting view is that the Fed used a legal loop-hole, in Section 13.3, to expand its powers to act in a way that was close to, if not beyond, its proper capacity, i.e. ultra vires, as even Volcker complained. Rules of behaviour and accountability should be made by the legislature. The problem with that is that no one can foresee the future. So, binding the hands of the authorities tightly in advance, ex ante, may force them to stand idly by as the financial system unravels, as turned out to be the case with the failure of Lehman Bros. Accountability to the legislature for actions already taken, ex post, is necessary, but tight prescription in advance overlooks the inherent uncertainty of an ever changing financial system. The future will not just be a re-run of the past, not just a different draw from an unchanging probability distribution.

24

C. Goodhart

A retort is that either the CB would have to lend in crisis conditions indiscriminately to everybody, which would generate moral hazard, or would have to rely on ambiguity, which in a crisis would be the reverse of ‘constructive’. So if one wants a CB to be flexible and accommodating not only in a crisis but even in the instance of a potentially dangerous disturbance, how might one proceed to limit ‘moral hazard’?

5 Limiting Moral Hazard One of the traditional roles of Central Banks has been to maintain financial stability, as emphasized in my 1988 book on The Evolution of Central Banks. But policies, such as LOLR and market maker of last resort to achieve this end, represent a form of insurance to commercial banks, and hence entail a risk of less prudent behaviour, since the banks believe that the Central Bank may save them from the adverse consequences of risk-taking, i.e. moral hazard. If we want, as I do, the Central Bank to continue to have responsibility for financial stability, and hence use its various policy instruments flexibly to this end, then what precepts can be applied to limit the accompanying moral hazard? There are some potential precepts:

5.1 Treat First Worst The first casualty of a financial downturn is likely to have been the most egregiously exposed and the most aggressive risk-taker. So a general precept should be to treat this most toughly, and become increasingly accommodating as contagion then threatens. Thus when the CB allows the first institution in a class to be liquidated, ‘pour encourager les autres’, it needs to stand ready to support the rest of that class. That principle was adopted by Governor Eddie George in 1995, when Barings was not supported, but steps were taken to prepare support for the remaining class of British merchant banks. The liquidity problems of foreign-headquartered banks, and the foreign-currency problems of domestic banks, within a global financial system involve somewhat separate issues which will not be tackled here.

5.2 Involve the Other Banks, if Possible While the CB should take the lead, it would be helpful to involve other banks in any process of cross-guarantees and mutual self-help. This not only reduces the publicsector burden, but also encourages additional information on status and reputation. This was done in the UK between Barings 1 (1891) and the Fringe Bank Crisis

Balancing Lender of Last Resort Assistance with Avoidance of Moral Hazard

25

(1974/1975), but collapsed under the strains of global banking (Johnson Matthey Bankers 1984). A US example is Long Term Capital Management (1998). With the potentially failing bank being a competitor, other commercial banks are not likely to agree to coordinated mutual assistance unless they are persuaded that the damage to the working and reputation of the banking system as a whole merits it. So a need to persuade other banks to participate in a rescue can act as a brake on a Central Bank which otherwise might seek to rescue a ‘bad’ bank whose demise would actually have benefitted the remaining system. How far can this be resurrected?

5.3 Change the Incentive Structure In the first half of the nineteenth century, most bank shareholders had unlimited liability, and this was a check on consciously risky behaviour. But when business borrowers became large and the efficient size of banks increased in line, in the latter half of the nineteenth century, it became necessary to attract bank equity from outside shareholders, who could not individually control or monitor the bank. Such outsiders could not be tempted to purchase bank equity if it had unlimited liability. But for many decades in the USA, until the 1930s, bank shareholders had double liability in the sense that, if their bank became distressed, not only would their share value go to zero, but also they could be required legally to inject further funding into their bank equal to the original par value of their shares. Clearly one could go beyond such double to multiple liability for certain shareholders, and require all designated insiders to hold, or to be allotted, shares with such multiple liability. Among such insiders could be: • • • •

Those with large share-holdings, Board members, Senior executives, Staff earning more than X over some period. Other steps to enhance more prudent behaviour might be:

1. Bail-inable debt for large creditors. 2. Require all bonuses to be paid in bail-inable debt. 3. Establish a Supervisory Board with a wider cast of stakeholders, including representatives of staff and creditors. Workers have more of their human capital tied up in their firm than more diversified shareholders and hence should be more risk averse. 4. Current regulatory proposals, e.g. on the use of bail-inable debt and claw-backs of prior bonuses in the event of failure, have gone some slight way in this direction, but not, in my view, far enough.

26

C. Goodhart

6 Conclusions My own preference would be to change the incentive structure dramatically, ‘make the punishment fit the crime’, but leave the CB with flexibility and leeway to balance the requirements of preventing financial instability against concerns about ‘moral hazard’. The Dodd-Frank Act has given too much ground to the moral hazard fundamentalists. We may live to regret that.

References Bagehot, W. (1873). Lombard Street: A description of the money market. In E. Johnstone & H. Withers (Eds.), The library of economics and liberty. London: Henry S. King and Co. http://www.econlib.org/library/Bagehot/bagLom.html (11 Sept. 2017) Goodhart, C. (1988). The evolution of central banks. Cambridge, MA: MIT Press. US House of Representatives. (2010). Dodd-Frank Wall Street Reform and Consumer Protection Act. Public Law 111–203 [HR 4173].

Charles Goodhart is a professor of banking and finance (emeritus) at LSE, and a former member of the Bank of England’s Monetary Policy Committee. His main research interests are central banking, monetary policy and financial regulation.

Optimal Lender of Last Resort Policy in Different Financial Systems Falko Fecht and Marcel Tyrell

Abstract In a framework closely related to Diamond and Rajan (J Polit Econ 109:287–327, 2011) we characterize different financial systems and analyze the welfare implications of different central bank policies in these financial systems. We show that in case of a large negative liquidity shock, liquidity demand has lower interest rate elasticity in a bank-based financial system than in a market oriented financial system. Market interventions, i.e. non-standard monetary policy measures to inject liquidity need to be much larger in a bank-based financial system in order to bring down interest rates to sustainable levels. Therefore, in financial systems with rather illiquid assets an individual liquidity assistance might be welfare improving, while in market oriented financial systems, with rather liquid assets in the banks’ balance sheets, liquidity assistance provided freely to the market at a penalty rate is likely to be efficient. While the costs of individual support might not be worthwhile in a market oriented financial system in which deadweight losses of market based support are small, in a bank based system the deadweight losses of unconventional monetary policy are large and thus individual support more efficient.

JEL Classification: D52, E44, E52, E58, G21

1 Introduction The way a central bank’s monetary policy in normal times is implemented reflects the financial structure of the country or currency area. That becomes apparent from a look at central bank policy frameworks before the onset of the great financial

F. Fecht Frankfurt School of Finance and Management, Sonnemannstrasse 9-11, 60314 Frankfurt am Main, Germany e-mail: [email protected] M. Tyrell () Witten/Herdecke University, Alfred-Herrhausen-Str.50, 58448 Witten, Germany e-mail: [email protected] © Springer International Publishing AG 2017 F. Heinemann et al. (eds.), Monetary Policy, Financial Crises, and the Macroeconomy, DOI 10.1007/978-3-319-56261-2_3

27

28

F. Fecht and M. Tyrell

crisis of 2007–2009. For instance, the Federal Reserve (FED) acted at these times by executing its open market operations (OMOs) with a small group of (nonbank) primary dealers as direct trading counterparts. Other banks only had indirect access to the central bank via Primary and Secondary Credit Facilities (PCF and SCF), the so called ‘Discount Window’, which however were left largely unused because the visibility of drawing on the discount window created a stigma effect. That stood in contrast to the European Central Bank (ECB) system where all financial institutions subject to reserve requirements had access to the regular ESCB auctions and standing facilities, and the recourse to a marginal lending facility (MLF) led to no stigma since it was not inferable. Thus, the FED intervened directly in financial markets by outright purchases and sales with primary dealer and relied on a proper functioning of financial market for the reallocation of liquidity from primary dealers to the rest of the financial system reflecting the market-based nature of its financial system. In contrast, the ECB allotted liquidity directly through repurchase agreements to large number of banks against collateral, reflecting the bank-dominated structure of the euro area. Hence the different central bank policies frameworks in place before the Great Financial Crisis (GFC) reflected the differences in the financial systems. During the GFC the need for broad liquidity support became obvious. An extraordinary liquidity shock affected the global financial system and impaired the functioning of national and international financial markets. In particular the FED reacted rapidly. Since the FED’s monetary policy framework relied on a functioning interbank market and on central bank’s counterparties being willing and able to lend to the institutions which are most liquidity-starved, the FED had to adapt its operational framework immediately when markets dried-up. The FED introduced the Term Auction Facility (TAF), available to all depository institutions, which previously had only access to the discount window. In addition, a Primary Dealer Credit Facility (PDCF) was created for securities dealers. Thus during the GFC the FED expanded its operational framework to provide liquidity directly to further counterparties, which made the FED’s framework effectively more similar to that of the ECB. However, in response to the enduring crisis central banks all over the world also adapted their operational framework and introduced non-standard monetary policy measures to inject liquidity on a large scale. In this respect the FED was much more aggressive and proactive. Already in 2008 the FED enacted it quantitative easing program centred on outright asset purchases of Treasury securities in particular. Only in July 2012 Mario Draghi announced to do “whatever it takes” and launched the Outright Monetary Transactions (OMT) program of the ECB, followed by an ‘expanded asset purchase program’ of euro-area bonds which was started in March 2015. Thus notwithstanding some differences in detail and motivation, the ECB copied the large scale liquidity allotment through outright asset purchases from the FED but with a significant delay. However, given the prevailing differences of the financial systems on the two sides of the Atlantic one might wonder whether this response by the FED is indeed also most efficient for the bank-dominated financial system of Continental Europe.

Optimal Lender of Last Resort Policy in Different Financial Systems

29

More general, the main question we want to address in the paper is the following: Should a lender of last resort (LOLR) in a traditional bank-dominated financial system respond differently to liquidity shortages than in a more market-oriented financial system? Using a framework that strongly relates to Diamond and Rajan (2001) we argue that the relevance of relationship lending and securitization can be captured in differences in the pledgeability of bank returns and the liquidation value of bank assets. Thus we find that in bank-dominated financial systems liquidity shortages spike asset prices more severely given no governmental intervention. Therefore, asset purchase programs that provide liquidity through the asset market, i.e. quantitative easing, generate in our framework a windfall gain for liquidity rich banks. These windfall gains are the larger the less liquid bank assets are. Hence in a financial system like the one of the U.S. in which the bank assets are mostly securitized and tradeable, in which shadow banks play an important role and in which relationship lending is of minor importance, these windfall gains at the expense of the government are rather modest. In contrast, in the Euro area financial system in which firm bank relationships are essential to ensure financing of households and firms and in which securitization and loan sales to other financial intermediaries play a subdued role liquidity rich banks receive larger windfall profits from asset purchases. Consequently, governments confronted with the latter type of financial system have stronger incentives to avoid these windfall profits. Governments in those countries benefit from providing individually tailored credit lines that ensures that liquidity support is allocated efficiently but entails larger information costs on behalf of the lender of last resort. What is the intuition for our results? In our model banks serve as relationship lender. They must refinance themselves at least partially through demand deposits. Since households have no loan collection skills, they have to rely for efficient investments on the collection skills of a relationship lender, i.e. the bank. But banks can commit to repaying households only by issuing deposits. As Diamond and Rajan (2001) show, the demandable nature of deposits create a collective action problem for depositors. They will individually run to demand repayment in case they anticipate that the bank cannot, or will not, pay the promised amount. Since bankers will lose all rents when there is a run on the bank, they will repay whenever they can. For that reason deposits serve as a commitment device. Now consider a situation where a regional business cycle shock leads to a certain quantity of loans being overdue. Banks can either call loans due or rollover its loans. On the one hand, rollover requires further refinancing, which however can only be provided by entrepreneurs whose projects already were successful. On the other hand, collecting loans which are called due leads to firm defaults. Firm profits are lost, but the bank can seize the assets of the firm and redeploy them to the next-best use, resulting in a loss of overall surplus. Thus rolling over the loan would be the best solution. Yet the bank’s borrowing capacity against the pledgable returns on rolledover loans is determined by the interest rate prevailing in the financial market. For a sufficiently large liquidity shock a bank might be unable to borrow sufficiently to repay depositors and roll over its loans. Depositors anticipate that the bank obtains not enough liquidity. The collective action problem occurs and all depositors will

30

F. Fecht and M. Tyrell

start running on the bank. Depositors will seize all bank asset and call all delayed loans due. Firms with delayed cash-flow default and the entrepreneurs’ rents from their human capital are lost. Thus liquidity shocks with elevated interest rates can spark off banks runs and generate negative externalities for entrepreneurs. Inflicting negative externalities on entrepreneurs delivers a foundation for intervention by a LOLR. In principle, two options are possible for a central bank. First, the central bank can provide direct and discretionary liquidity support just to an ailing bank and second, the central bank provides liquidity to the market in order to stabilize the interest rate at a sustainable level, which might be interpreted as quantitative easing. Direct LOLR support only to ailing banks generates no windfall profits for other banks and therefore requires less central bank liquidity injection. On the other hand, for being effective it is essential that direct LOLR support is provided on the basis of very precise information. Otherwise, liquidity will be wasted even by using this policy option. How does the configuration of the financial system influences the choice between these two options? In our modelling framework we grasp the differences in financial systems by assuming that bank assets in a bank-dominated financial system are less liquid than in a market-oriented financial system. In our view bank-dominated financial systems are characterized by a strong relationship-lending orientation which typically leads to a higher illiquidity of bank loans. Furthermore the higher illiquidity of loans also affects the liquidity demand of banks which is more heterogeneous in bank-dominated systems. Therefore banks are more inclined to roll-over loans in a bank-dominated financial system. Liquidity demand increases and has a lower interest rate elasticity. As a result, market interventions need to be much larger in a bank-based financial system in order to bring down interest rates to sustainable level as compared to a market-oriented financial system, thereby generating more windfall profits for sound banks. Thus, the benefits of direct discretionary liquidity support relative to market liquidity interventions seem to be larger in bank-based financial systems. Of course, individual liquidity assistance, bailout policy, recapitalization or closure is more demanding for government and regulators. Furthermore, it might imply that liquidity assistance is unevenly provided to banks from different countries in a monetary union, which implies huge political costs. But unconventional monetary policy and the massive liquidity allotment by central banks also carry some well-established costs. To mention just a few arguments: It distorts bank investment incentives.1 It undermines market discipline.2 And, last but not least, it might cause financial repression by taxing savers. In fact, while the costs of individual support might not be worthwhile in a market-oriented financial system in which deadweight losses of market-based support are small, in a bank-based system the deadweight losses of unconventional monetary policy are large and thus individual support more efficient. However, the political costs from country specific

1 2

See for instance Drechsler et al. (2016) and Abbassi et al. (2016). See for instance Fecht et al. (2015).

Optimal Lender of Last Resort Policy in Different Financial Systems

31

liquidity support seems to be extraordinarily high for the ECB. And that might also be one important reason why in recent times the ECB in its crisis response more or less follows the blueprint of the FED.

1.1 Related Literature The basic consideration about an optimal design of lender of last resort policies goes back to the principles formulated by Bagehot (1873), based on the work by Thornton (1802). He suggested that in a crisis, the lender of last resort should lend freely, at a penalty rate, on the basis of collateral that is marketable in the ordinary cause of business when there is no panic. This doctrine apparently follows the view that interbank markets are not always efficient in reallocating funds to the most illiquid banks. This assumption is particularly criticized by Goodfriend and King (1988) who argue that given today’s repo markets banks that have sufficient collateral to turn to the LOLR should also be able to receive funding in the interbank market. Thus no LOLR is needed to overcome individual liquidity shortages that result from idiosyncratic shocks. In case of an aggregate liquidity shortage only liquidity provision to the market is required since the most illiquid banks will be willing to pay the highest rates to receive funding in the interbank market so Goodfriend and King’s reasoning. To ensure that the most illiquid banks can sustain the aggregate liquidity shortage the central bank has to provide sufficient liquidity to keep money market rates at a sustainable level. Our model is very much in line with this argument. However, we show that to contain the spike in the money market rates to level sustainable to the most illiquid bank the central bank has to inject liquidity that will be partially absorbed by banks that could sustain the liquidity shortage. Thus we show the liquidity provision to money markets as proposed by Goodfriend and King leads to a waste of liquidity. This might carry some cost as it inflates the central bank’s balance sheet which in turn might result in an inflation tax and excessive credit risks encountered by the central bank. The vast majority of the recent literature on the design of LOLR measures, however, challenges the applicability of the Bagehot doctrine because of the implicit assumption that solvent banks dispose of sufficient collateral. As argued, for instance, by Calomiris and Kahn (1991) and Diamond and Rajan (2000) the main reason why banks run a liquidity risk is because it serves as a commitment device which allows them to raise funding for opaque and thus illiquid assets. This however means that the collateral value of those assets is limited. Thus neither the repo market nor fully collateralized liquidity provisions might provide sufficient insurance for banks against adverse liquidity shocks. Our paper builds on this literature and studies what implications the differences in the collateral value of assets have for the functioning of the interbank market during an aggregate liquidity shortage. Some of our results are in line with those of Diamond and Rajan (2001, 2005, 2011, 2012). In particular, Diamond and Rajan (2011) also show that private

32

F. Fecht and M. Tyrell

and public benefits from rolling over overdue loans might deviate. They argue that impaired banks have private incentives to hold rather than sell illiquid assets even if they may be forced in the future to sell those assets. But in their theoretical analysis they look at the incentives of liquid buyers to lend to these ailing banks. Diamond and Rajan (2012) analyze the optimal interest rate policy of central banks and show that central banks should raise rates in normal times above the marketdetermined levels to offset reduced interest rates at times of financial stress. This is necessary to preserve bank incentives to maintain low leverage and high liquidity which otherwise, expecting that interest rates would be reduced in adverse times, would take on more short-term leverage and make more illiquid loans. But again they do not study to what extent the structural difference in the pledgeability of banks’ returns in different financial systems affects the severity of crises and the optimal policy design. Our analysis focuses on aggregate liquidity shocks and optimal measures to deal with them. If interbank markets are inefficient there might also be a role for an LOLR that contains individual liquidity shortages as pointed out by Freixas et al. (2004). For instance Rochet and Vives (2004) show in such a context that a lender of last resort can avoid inefficient liquidation of banks. Repullo (2005) investigates the question whether the existence of a lender of last resort really increases the incentives of banks to take risk. In our model, though, the interbank market is also not efficient in the sense, that it channels always the liquidity to those banks that are needing it the most. Freixas et al. (2004) discuss how the optimal LOLR policy is affected by moral hazard problems on side of the banks, an aspect not considered in our analysis.3 Our view of the differences of financial systems boils down the major insights from the extensive literature on comparative financial systems. This literature shows that there are many dimensions in which financial systems differ.4 It includes theoretical analysis, e.g. Allen and Gale (2000a) and, with respect to corporate governance systems, Magill et al. (2015), as well as more empirically oriented work such as Franks and Mayer (1995), Schmidt et al. (1999) and Levine (2002). Most interestingly, in a just published paper Langfield and Pagano (2016) provided strong empirical findings that differences in financial systems between the US and Europe are still persistent and even increasing in recent times. Furthermore, they documented evidence that the strongly bank-based financial structure in Europe increases systemic risk and lowers economic growth in comparison to economies which are characterized by a market-based financial structure, thereby resuscitating the debate on the relative merits of bank-based and market-based financing. However, we focus

3

For a discussion of the different lender of last resort function(s), see Freixas et al. (1999). We do not want to touch the issue if there should (and could) be an institutional separation between a central bank which is responsible for the conduct of monetary policy and a lender of last resort; on this topic see Goodhart (1995). Also we do not analyze the potential agency conflicts between deposit insurance fund, central bank and bank supervisors; on this see Repullo (2000) and Kahn and Santos (2005). 4 See Allen and Gale (2004) for a survey.

Optimal Lender of Last Resort Policy in Different Financial Systems

33

our very simple analysis on just one aspect, namely the differences in the importance of relationship banking, securitization and tradability of banks’ assets in marketoriented and bank-dominated financial systems and its impact on central bank policy measures in adverse times. Referring to the banking dimension, by conducting a meta-analysis Kysucky and Norden (2016) have shown that relationship lending is more prevalent in the bank-dominated financial systems in Europe and Japan. Furthermore, with respect to central bank policy, Cour-Thimann and Winkler (2013) recently emphasized in its analysis of the ECB’s non-standard monetary policy measures that the institutional set-up of the EMU and the mostly bank-based financial structure of the euro area economy are framing the ECB’s monetary policy.

2 The Framework 2.1 The Setup Following Diamond and Rajan (2001) we consider an economy with three dates .t D 0; 1; 2/ and a large number of entrepreneurs, bankers and investors. Entrepreneurs are wealthless, however each of them has a project at his disposal which requires an investment I D 1 at t D 0. Each investor is endowed with a small amount of the consumption good in comparison to the required investment size, hence many investors are needed to fund a project. In addition, we assume that the aggregate endowment of all investors in the economy is lower than the total investment possibilities. Because of this shortage of investment capital at date 0 entrepreneurs and bankers must offer an expected return as high as possible to attract funding. Entrepreneurs, investors and bankers, whose role will be clarified below, are risk-neutral but differ in their preferences: Investors and bankers have a strong preference for consumption at date 1, i.e. they have a very high discount rate  for consumption at date 2, whereas entrepreneurs value consumption at each date equally. Investors can store their initial endowment earning a return of 1 for every unit invested, or they can invest it in the project. Financing the projects includes some difficulties which have to be overcome. Entrepreneurs have specific abilities vis-a-vis their projects, i.e. the cash flow each entrepreneur can generate from his project exceeds what anyone else can get out of it. But entrepreneurs cannot commit their human capital to the project, except on a spot basis. From this it follows that a lender can extract future repayment only by threatening to take away the project from the initial entrepreneur. The project returns C generated by the initial entrepreneur are uncertain in terms of their time structure. The project pays out C either at t1 if the project produces early or at t2 if the project is delayed. All uncertainty about projects is resolved at date 1. We consider two alternatives when taking away the project from an entrepreneur. The project can be restructured at any time until date 1 which will yield a payoff c1 immediately and nothing at date 2, or the entrepreneur can be replaced with assets

34

F. Fecht and M. Tyrell

redeployed to their next-best use, which does not change the timing of the produced cash flow but the level to  C with  < 1. Both alternatives result in a loss of surplus, since c1 < 1 <  C < C;

(1)

However, the big difference between these two alternatives is the following: The second alternative (replacement) can only be implemented by a bank who was the only initial financier of the project while restructuring can be done by any investor, irrespective of having been an initial financier of the project or not. How can we interpret these alternatives? Restructuring is an activity which can be understood as changing the original content of the projects so that some immediate cash can be produced without any specific knowledge. One may think of this strategy as abandoning the uncertain technology and using instead a commonly known technology that produces goods quickly or stopping half-finished projects and salvaging the production goods. All investors can realize this cash flow, hence c1 is the secondary market value of a project. On the other hand, replacing the entrepreneur and redeploying the assets to their next-best use, which yields  C is an activity which demands specific skills for replacing the entrepreneur but preserving the original content of the project. It may involve searching for a new entrepreneur who has similar skills to the original one, or abandoning only such aspects of the project that were particularly dependent on the old entrepreneur. Because this implies learning all about the project it takes time, effort and a constant close contact to retain these skills. Therefore, we assume that just one initial financier, effectively a “relationship lender” or banker who collect the savings of sufficient investors to become the sole initial financier, will undertake this costly activity. Accordingly, only the banker knows the next-best use of the project’s assets. To sum up, the bank can realize  C from the project, if it takes the project away from the initial entrepreneur, while other investors can only realize c1 . Therefore, the initial entrepreneur will offer to repay  C to a bank and only c1 to other investors. How can we grasp the differences between financial systems in this modelling structure? One obvious difficulty lies in the fact that this framework taken at face value allows only banks to exist as intermediaries. Capital markets in the literal sense as institutions, where firms issue stocks and bonds, households buy and trade these securities and the resulting prices incorporate valuable information, are not caught in our modelling structure. Yet what makes the framework attractive is the possibility to grasp certain consequences of market-based and bank-based financial systems. We view a bank-based system as a configuration with a relatively high  and a low c1 while the reverse, a relatively low  and a high c1 is true in a market-based system. A high  points out that usually in a bank-based system the intermediary has a great deal of information about her borrowers and their projects because of a long lasting and close relationship. As a consequence, she can enforce higher repayments from a borrower than a typical lender in a market-based system who does not collect as much knowledge and information. So the banker in a bank-based

Optimal Lender of Last Resort Policy in Different Financial Systems

35

system can “replace” the entrepreneur easier, thereby retaining much of the original strategy of the initial entrepreneur. This gives her bargaining power. In our opinion, this is an essential characteristic of a bank with typically firm-specific knowledge. On the other hand, c1 is the payoff of restructuring. Because this restructuring is the best alternative, publicly available use, it can be interpreted as the market value of these projects. A relatively high c1 indicates that much information about the best alternative use is released in the market. In sum, we conclude that the difference between  C and c1 is rather small in market-based systems.5 The assets are relatively liquid because a great deal of information gets “externalized” through the market activities. This reflects the notion that there are many analysts working for mutual funds, pension funds and other intermediaries who gather private information and incorporate these through their trading activities in market prices which is the general advantage of a market-based system. In bank-based systems assets are more illiquid. In countries with bank-based systems, relatively few companies are listed and accounting disclosure requirements are limited, so very little information is incorporated into stock prices. Also the number of analysts who follow stocks is small, so only limited private information is incorporated into stock prices. However, intermediaries have more information available in these systems. The greater prevalence of long term relationships, i.e. the “hausbank”-relationship, in bank-based systems means that the banks are able to acquire considerable information about the firm they lend to. Typically this information will not be released to the market; instead the information will be used internally to allow a smooth functioning of the long term financial relationship and allocate resources efficiently.6 Therefore information in a bank-based system is more or less “internalized”, outsiders to the financial relationship have only a small chance to get valuable information.7 Banks have strong incentives to acquire and use information because they can profit from information which doesn’t leak to outsiders. However, this creates the problem that most of the assets are rather illiquid because only the banker has the relevant information. This means c1 is small and the

Of course, we maintain the relation C > 1 > c1 for a market-based system. Only the difference is small. 6 See for instance Rajan (1992) and Gorton and Kahn (1992) for theoretical analysis and Elsas and Krahnen (1998) and Berlin and Mester (1998) for empirical analysis. 7 See Schmidt and Tyrell (2005) for a discussion how these two perspectives on information, i.e. externalization and internalization, can be mapped into two approaches to the role of information in financial systems, namely the rational expectations literature on the role of prices in resource allocation and the intermediation literature which is concerned with the role of banks as delegated monitors. See in addition Fecht (2004) where using a Diamond/Dybvig framework the stability of different financial systems is analyzed. Within that framework, distinct financial systems are characterized by the fraction of households with direct investment opportunities that are less efficient than those available to banks. If the fraction with inferior direct investment opportunities is relatively high, the financial system will be called a bank-dominated one, in the other case characterized by a high fraction of households with “efficient” direct access to investment opportunities, the financial system is market-oriented. Note that such a characterization of financial systems is complementary to the one given in the paper here. 5

36

F. Fecht and M. Tyrell

difference between  C, the payment a bank can extract, and c1 , the market value of a loan, is large. We feel that this parametrization captures one of the most important underlying causes of the observable differences between bank-based and market-based systems, namely the different ways of acquiring and using information in the respective systems.

2.2 Financial Structure of Firms and Banks What complicates the financial relations in this economy is the presence of specific skills at two different layers. First of all, original entrepreneurs with their specific abilities can generate a higher expected return from the projects than everyone else but they cannot commit this human capital on a long term basis to the projects. Thus, projects are illiquid in the sense that they cannot be financed to the full extent of their cash flows. The second layer causes the illiquidity of the loans. Only an initial lender has specific skills to extract high repayments from the entrepreneur but she also cannot commit her human capital to the loan. For these reasons the financial contracts we consider specify only who owns the physical assets conditional on the payments made.8 Let us turn to the resulting financial structure of a firm first. Initially the entrepreneur owns the blueprint of a project to produce goods. Since he has no endowment, he needs to borrow to invest and is obliged to pay back the credit later on. Hence, the contract signed by the entrepreneur specifies a repayment and the assets the financier gets in case of default. Because of his specific abilities and the limited commitment of human capital, an entrepreneur can threat to withhold his human capital at any time until the cash flows are produced. We assume here that the entrepreneur can make a take-it-or-leave-it offer, which gives him maximum bargaining power leaving financiers with their reservation return. Thus notwithstanding any ex-ante agreement between entrepreneur and banker, the most the banker can get as repayment for the credit is just her best outside option “replacement”, which yields  C. Only by threatening to take away the project and redeploy it to this next-best use, the banker as an initial financier can extract this amount as future repayment for the credit. In turn, this is also the maximum amount the entrepreneur can credibly pledge to an initial financier. Since the economy is short of investment capital at date 0, entrepreneurs are competing for the scarce resources and only a few of them get a loan by bidding the maximum amount they can credibly pay back. This means that in the financial contract the borrower promise to pay the banker Pt D  C on demand. If, however, the project turns out to 8

We assume a court system, which can enforce financial contracts and transfer assets to lenders when contracted repayments are defaulted upon, but cannot compel entrepreneurs or bankers to contribute their human capital. Thus the court can help to seize the project’s assets or the bank’s loans, respectively. However, the value of these assets depends on the cash flow the lenders can generate out of the assets.

Optimal Lender of Last Resort Policy in Different Financial Systems

37

be late and the entrepreneur cannot repay this amount and defaults, the bank has the property rights over the project’s assets and will decide what to do with them next. How can the banker refinance the project? Only the banker as an initial lender knows the next best use of the project’s assets. During the course of lending she acquired specific skills which she can use to collect more on the loan than other lenders could do. Similar to an entrepreneur the banker possesses human capital that she can threaten to hold back unless investors reduce the required payment. Thus, she cannot commit to repaying to outside investors the full amount that she can extract from an entrepreneur. This also implies that the banker may not be able to raise the full present value of the loan held. But bankers themselves have no endowment, so they have to find a way to refinance the loan through outside investors, otherwise they cannot persuade investors to entrust them with their goods in t D 0. As a consequence, the bank couldn’t act as the only initial financier of an entrepreneur and the projects wouldn’t be financed.9 As Diamond and Rajan (2001) show the bank can use a device to commit to repayment up to the full value of the loan. The bank should refinance lending by issuing uninsured demand deposits subject to a sequential service constraint. The sequential service constraint creates a collective action problem among depositors: If the bank makes an attempt to renegotiate deposit repayments she will cause a run. Rather than making concessions which may be in their collective interest, depositors find it in their individual interest to run immediately to capture full repayment of their deposits. Because of the “first come, first served” aspect of uninsured demand deposits, they cannot be negotiated down. Individually each depositor has an incentive to withdraw his claims as fast as possible because his payoff depends on his place in line. Thus withdrawing is a Nash equilibrium. In case of a run depositors seize the assets and restructure all the projects destroying any potential rent of the banker. It is not in the interest of a bank to renegotiate down an ex-ante agreed repayment because courts would enforce depositors’ demands, and the rents of the banker would be destroyed. Therefore, the bank’s ability to create liquidity is inseparable from its potential fragility.10 Hence in a world without uncertainty, a bank refinances entirely with demand deposits to maximizes the credit it can offer to entrepreneurs. The possibility of runs exerts market discipline on banks, although bank runs are never observed in equilibrium. Since the banker can threat not to deploy her specific collection skills on behalf of the investors at any point after the deposit is made, deposits must be demandable at any time to provide commitment value, even if consumption occurs only at date 1 or 2. But a bank’s capital structure typically involves (long-term) capital in addition to demand deposits. The reason is that capital represents a softer claim than demand deposits, i.e. a claim that can be renegotiated. In a world of uncertain project cash

9

Acquiring the specific collection skills to enforce repayment on the part of an entrepreneur is a costly activity which is not worth doing by a small investor in analogy to arguments given in Diamond (1984). 10 See Diamond and Rajan (2001) for a full analysis of this mechanism.

38

F. Fecht and M. Tyrell

flows, financing with only demand deposits carries a cost. It exposes the banks to destructive runs if they truly cannot pay because the realized project cash flows of entrepreneurs are too low. In this way, Diamond and Rajan (2000) show that with observable but not verifiable uncertainty in project returns, it may be optimal for a bank to partially finance with a softer claim called capital. Capital holders cannot commit not to renegotiate because they are not subject to a collective action problem. Thus capital acts as a buffer because its value adjusts to the underlying asset values and can prevent inefficient runs. On the other hand, this allows a banker to capture some rents in the future and therefore reduces its ability to raise funds and creates liquidity in the present. The optimal capital structure of a bank has to trade-off these costs against the benefits of capital. In the following we assume that banks face a capital requirement k, stating that a fraction k of the present value of a bank’s assets has to be refinanced using capital.11 By normalizing our financing problem and the capital structure of the bank on one investment project, we know that the bank assets are worth  C when the entrepreneur can repay at date 1: Owing to the capital shortage at date 0, the bank extracts all the rent from the entrepreneur that can be pledged, leaving the entrepreneur a rent of .1   /C. If D denotes the repayments on deposits, then  C  D is the surplus that can be split between the banker and the capital holder in the renegotiation process. Assuming equal division of the surplus, capital owners will be paid 12 . C  D/ and the same amount will be absorbed by the banker as a rent. It follows that D C 12 . C  D/ D 12 . C C D/ will be passed on as total pledgable payment per loan to depositors and investors holding a capital claim. Inserting this into the definition of the capital requirement (k D 1k  C. 1Ck

1 2 . CD/ 1 2 . CCD/

) gives

the maximum amount refinanced by deposits: D D Hence, the banker gets k a rent of 1Ck  C per finished project and capital owners get the same. Thus, the total C value that can be pledged to outsiders amounts to 1Ck .

2.3 Local Lending Markets and the Time Structure of the Model We argued in the last section that a banker acquires specific collection skills visa-vis entrepreneurs through her lending activity. But typically this experience or knowledge, which is costly to develop, can only be acquired for a subset of the date 0 project opportunities. For instance, a bank may only have experience in specific industries or possess knowledge about specific locations. From this it follows that each bank has a local monopoly in lending.

11

This requirement is either exogenously imposed by regulators or endogenously determined as a result of—in our case unmodelled—uncertainty along the line of Diamond and Rajan (2000).

Optimal Lender of Last Resort Policy in Different Financial Systems

39

To simplify our analysis we assume that the economy is divided into two regions of the same size. The two regions are ex ante at date 0 identical in every respect but can become heterogeneous at date 1 in the sense that the fraction of early projects in the two regions differ. More specifically, ex ante the regions are populated by many identical banks, each of them being a monopolist in their local market and facing an identical pool of (many) entrepreneurs. With probability p1 no macroeconomic shock occurs which means that all projects in both regions generate cash flows in t D 1. With a negligible probability 1  p1 a negative macroeconomic shock occurs which delays some projects.12 In one region only a fraction ˛ of the bank loans generates cash flows at date 1 while in the other region a fraction ˛ of projects financed by banks produce early cash flows with ˛ > ˛. Ex ante nobody knows which region will be hit by the more severe macroeconomic shock. Thus, while banks are identical ex ante, in t1 half of them turn out to be weak, i.e. having a higher fraction of delayed projects, while the other half turns out to be strong, which means having a high fraction of projects that generate an early return. The timing of our model is as follows: At date 0 the ex ante identical banks compete for the investors’ endowments. They issue a mix of deposits and capital to investors and promise them the maximum pledgable amount since consumption goods are short relative to projects at that date. Investors will invest as long as their opportunity rate of return, i.e. storage, is met. After raising cash, banks lend to entrepreneurs in their local lending market. We normalize without loss of generality the amount each bank can raise at date 0 to be 1. In lending to entrepreneurs the banks will charge the maximum repayment  C on demand. In t D 1 the regional fractions of early and late entrepreneurs becomes public information and each entrepreneur learns if his projects is early or late. Late entrepreneur demand a roll-over of their loan from the bank. Thus banks know the fraction of their loans that turns out to be early projects. As soon as a bank discovers that even with restructuring late projects or raising funds in the secondary financial market against late projects’ repayment it cannot generate enough liquidity to payoff depositors, the banker tries to renegotiate the deposit repayments. This will trigger a run and all the late projects will be restructured to yield c1 immediately. The repayments on early projects and the restructuring return on late projects will be seized by depositors. Early entrepreneurs retain .1   /C which they can either invest in the financial market with a bank or consume. Entrepreneurs with late projects will default. Then the bank decides how to deal with late projects. It can restructure the projects at date 1 to obtain immediate liquidity or it can reschedule the loan payment until date

12

As in Allen and Gale (2000b) we assign a nearly zero probability at date 0 to the occurrence of the negative shock. We perturb the model in this way in order to have no change in the allocation decision of the bank at date 0. Alternatively, Diamond and Rajan (2005) show that even with nonnegligible probability of an adverse shock the bank will lend all the funds they raise in case the expected return from lending is higher than the storage return. Thus, we could also assume a nonzero probability for the macroeconomic shock. It would not change the main results of our analysis.

40

F. Fecht and M. Tyrell

2, keep the project as a going concern and borrow against the future repayments. Which option is more preferable for the banker depends on the prevailing interest rate and its need for funds. A financial market, in which immediate consumption goods (liquidity) can be traded against claims on future repayments (assets) is open at date 1 to equate supply and demand. The bank itself uses repayments from the early entrepreneurs, from the restructured late projects, and the cash obtained from early entrepreneurs in the financial market (as deposits and capital) to repay investors at date 1. At date 2, the bank gets repayments from the finished late projects. Entrepreneurs will consume.

3 Stability of an Individual Bank The decision whether a loan is restructured or prolonged is taken in a first step by the bank manager. He maximizes his expected rent. If a loan is restructured the banker’s human capital is worthless and he is no longer able to extract any rent. Thus the bank manager would always prefer to continue late projects. Capital owners will, however, force the bank manager to maximize the net present value of the projects.13 Since capital owner have strong preferences for immediate consumption in t D 1 they will only allow the manager to continue late projects if the amount of cash he can raise against the future repayments on prolonged loans is larger then the immediate cash a restructuring generates. Thus the decision depends on the interest rate prevailing in the t D 1 financial market.14 They will force the banker to C restructure a project if c1 > .1Ck/r and let the manager continue late project otherwise, i.e. if c1 

C .1Ck/r .

We define the threshold rate as rQ D

C : .1 C k/ c1

(2)

The higher the interest rate for getting liquidity, the more valuable is restructuring because it generates liquidity immediately. But this restructuring decision is biased, because only part of late projects’ return is pledgable to outside financiers of the C k C bank. As long as c1 < .1Ck/r C .1Ck/ C .1   / C, it is socially inefficient to restructure late projects.

13

The bank manager will continue the project despite having a strong preference for date 1 consumption. This means that even with a high discount rate of date 2 consumption the present value of the rent she can earn is positive. 14 Strictly speaking, our usage of the term interest rate is a bit loose. To be precise, r D 1 C i with i as the interest rate in the liquidity market. Note, that we take as given that banks do not store but invest any funds in lending activity. Clearly, this is the optimal decision when the probability p1 for the state where all the projects in both regions are early, is sufficiently high.

Optimal Lender of Last Resort Policy in Different Financial Systems

41

Next consider the decision of depositors. A run is fundamentally unavoidable if the cash inflow the bank can obtain in t D 1 falls short of the face value of the outstanding deposits. The bank manager has to renegotiate with depositors which will trigger a run whenever the sum of deposits exceeds the net market value of the bank at date 1: D  V1 . Given that capital owners force bankers to restructure late projects, because r > C .1Ck/c1 , depositors will run if V1 D ˛ C C .1  ˛/ c1 < D D

1k  C: 1Ck

(3)

If late projects are continued because r  rQ depositors will run if V1 D ˛ C C .1  ˛/

1k C D. Thus, strong banks (those with the higher fraction of early projects) never depend on the liquidity raised in the t D 1-financial market or obtained through the restructuring of late projects to prevent a run. A run will never occur in a strong region. However, if the interest rate exceeds rQ capital owners of strong banks will restructure late projects. This implies that the critical interest rate level rOO with rOO D

1 ; 1  k 1C˛ 1˛

above which a strong bank would experience a run is higher than the rate rQ above which capital owners restructure late projects anyway. In contrast, we assume that weak banks are dependent on the liquidity inflow from financial market transactions to repay depositors: ˛ C < D. Consequently, weak banks will not suffer from a run if the interest rate is below the lower threshold: rO D

1 1  k 1C˛ 1˛

:

42

F. Fecht and M. Tyrell

Following Eq. (4) this means that we restrict our parameter space to ˛>

1k > ˛: 1Ck

(6)

Furthermore, we focus our analysis on cases in which weak banks’ capital owner prefer to continue late projects at the threshold rate above which depositors already run the bank: rO < rQ . In sum this implies that rO < rQ < rOO . Using (2) this gives: 1Ck 1

k 1C˛ 1˛


rQ r D rQ rO < r < rQ r  rO

(9)

Obviously, given this aggregate liquidity demand three qualitatively very different equilibria occur depending on the aggregate liquidity supply, which is given by the overall fraction of early projects in the economy.

Note that we assumed Qr always being below the interest rate level at which the strong bank cannot raise enough liquidity to repay deposits: rO < Qr < rOO. 15

44

F. Fecht and M. Tyrell

Fig. 1 Equilibrium in a slight liquidity crisis

Proposition 1 Depending on the aggregate fraction of late projects three types of financial crises may emerge. (1) Slight liquidity crises, in which no bank collapses, (2) moderate liquidity crises, in which only weak banks are subject to a run and (3) severe liquidity squeezes, which also destabilize stronger banks. If the overall fraction of late projects is rather limited, a slight liquidity crises occurs. This case is depicted in Fig. 1. Trying to attract new funds from the early entrepreneurs against the required mixture of deposits and capital banks bid up the interest rate only slightly to r D

 2˛˛ 1   ˛C˛ 1 1Ck

(10)

But this only reduces the rents of the bank manager and the return of capital owners. It does not destabilize any bank in the economy. Obviously, the interest rate in slight liquidity crises is the higher the larger the aggregate fraction of late projects relative to the fraction of early projects and the higher the relation of pledgable to non-pledgable income of finished projects, since both determine the relative scarcity of liquidity in t D 1. Moreover, the interest rate is higher if the capital requirements are smaller, since capital requirements increase the rents of the banker and thereby reduce the returns of late project that can be promised to new depositors and capital owners in t D 1. If the “cash in the market”-constraint is more restrictive, i.e. the aggregate fraction of early projects smaller, the economy ends up in a moderate liquidity crisis, in which part of the banking sector collapses. This case is depicted in Fig. 2. Here

Optimal Lender of Last Resort Policy in Different Financial Systems

45

Fig. 2 Equilibrium in a moderate liquidity crisis

the lack of liquidity leads to an elevated interest rate of r D

1  1˛   ˛C˛ 1 1Ck

(11)

At this level the liquidity inflow at weak banks is insufficient to meet the repayment to depositors. Therefore, the banks with the stronger liquidity needs will be run, whereas the stronger banks, which are less dependent on the liquidity inflow from transaction in the t D 1-financial market will not be destabilized by the liquidity squeeze and will continue all late projects. As the weak banks fail their depositors seize the late projects and restructure them. Since weak banks do not demand liquidity in the financial market at this interest rate levels, the equilibrium interest rate in a moderate liquidity crises only depends on the relation (1) of late projects at strong banks to the overall fraction on early projects, (2) of pledgable to non-pledgable income of finished projects and (3) of returns bank can pledge to new depositors and capital owners to her total return. So roughly speaking, in a moderate liquidity crises only part of the banking sector that is subject to a more or less idiosyncratic adverse liquidity shock will collapse. The other part of the banking sector that does not face a severe idiosyncratic liquidity shock, because only a limited fraction of its projects turns out to be late, can finish all projects (Fig. 2). In contrast, if the aggregate fraction of late projects is even higher the economy ends up in a severe liquidity crisis. This case is depicted in Fig. 3. In this case the equilibrium interest rate will reach its upper bound r D rQ

(12)

46

F. Fecht and M. Tyrell

Fig. 3 Equilibrium in a severe liquidity crisis

Obviously, at this interest rate level weak banks collapse. But what differentiates a moderate from a severe liquidity crisis is that in the latter even strong banks have to restructure part of their late projects. At the equilibrium interest rate rQ capital owners are indifferent between restructuring and continuing late projects. However, the available liquidity is insufficient to repay all depositors. Therefore, the bank manager, who only receives a rent if projects are finished, will restructure just enough late projects to produce sufficient liquidity to prevent a run. The fraction of late projects that can be continued in a severe liquidity crises is given in equilibrium by  D

˛ C ˛ .1   / .1  k/Qr ˛ C ˛ .1   /  C   D   1˛  1 1˛  c1

(13)

Apparently, this fraction will be higher (1) the larger the aggregate fraction of early projects relative to the fraction of late projects at strong banks, (2) the higher the non-pledgable returns of entrepreneurs in relation to the pledgable returns going to the banks and (3) the smaller the present value of the fraction of the banks’ returns that can credibly be promised to new capital owners and depositors at the given interest rate rQ . Inserting the equilibrium value for rQ into the last expression shows that this is just the relation between the pledgable return of late projects if continued to the return of these projects if restructured [see Eq. (13)]. Consequently, if continuing late projects gives a higher return to banks relative to restructuring, a higher fraction of late projects will be finished even in a severe liquidity shortage. To sum up, in a severe liquidity shortage it is not enough that weak banks fail and therefore stop demanding liquidity. If the aggregate fraction of late projects is too high, even those banks that have financed a comparatively small fraction of projects that turn out to be late will not be able to raise enough liquidity at the financial market. However, these liquidity rationed banks do not collapse, but they will have to restructure late projects to raise sufficient liquidity to repay deposits. That exactly

Optimal Lender of Last Resort Policy in Different Financial Systems

47

is the contagion effect which impose a negative externality on entrepreneurs and these other banks. Having described the equilibrium in the financial market it is straightforward to see which impact the particular type of the financial system has on the equilibrium. Obviously, the higher fraction of pledgable income ( ) in bank-dominated financial systems shifts the entire liquidity demand to the upper right. Because the higher the pledgable income the higher the present value of late projects and the more aggressive banks can bid for funds in t D 1 in slight and moderate liquidity crises. In severe liquidity crises the higher return on late projects makes capital owners more willing to accept a continuation of late projects even for higher interest rates. On the supply side a higher fraction of pledgable income reduces the return of early entrepreneurs, thereby lowering the liquidity supply in the economy. All these effects of a higher fraction of pledgable returns point in same direction: Fluctuations of the interest rate in case of a financial crisis are higher in bank-dominated financial systems than in market-oriented financial systems. This is also reflected in the respective equations of the equilibrium interest rate [see Eqs. (10)–(12)]. A lower return on restructured projects (c1 ), which we also characterized as being typical for a bank-dominated financial system only influences the equilibrium interest rate in severe liquidity crises. The lower the returns from restructuring late projects the higher the interest rate up to which capital owners will accept a continuation of late projects of the bank manager. Thus, as can also be seen in Eq. (13), the interest rate fluctuations in severe liquidity crises also increase with a lower c1 and are therefore higher in bank-dominated financial systems. It is interesting to note, that also the threshold level for the different financial crises with respect to a given liquidity supply depends on the type of the financial system. Inserting rQ into the liquidity demand one can derive the threshold level for aggregate liquidity supply between moderate and severe liquidity crises. This shows that if the aggregate liquidity supply falls short of .1  ˛/  c1 the economy ends up in a severe crisis. While this threshold level obviously is not influenced by the fraction of pledgable returns, it rises the higher the returns on restructured projects. Thus, in market-oriented financial systems, in which c1 is higher, the economy ends up more often in a severe liquidity crisis, while in bank-dominated financial systems given a certain level of aggregate liquidity supply moderate liquidity crises are more likely. Similarly, the threshold level between slight and moderate liquidity crises can be derived by inserting rO into the liquidity demand function showing that in bankdominated financial systems characterized by a high  it is more likely to be in a moderate than in a slight liquidity crisis. Proposition 2 In bank-dominated financial systems interest rate fluctuations are higher during financial crises than in market-oriented financial systems. Moderate liquidity crises are more likely in bank-dominated financial systems, while in market-oriented financial systems severe but also slight liquidity crises are more likely to occur.

48

F. Fecht and M. Tyrell

5 Optimal LOLR-Policy Restructuring late projects is always welfare reducing in this economy. If the interest rate is below rQ this is most obvious, since in that case the net present value of the pledgable income from late projects that can credibly be promised to capital owners and depositors of the bank is higher than the returns generated if the projects are restructured: c1