The Handbook of Global Shadow Banking, Volume II: The Future of Economic and Regulatory Dynamics

By Luc Nijs

This global handbook provides an up-to-date and comprehensive overview of shadow banking, or market-based finance as it has been recently coined. Engaging in financial intermediary services outside of normal regulatory parameters, the shadow banking sector was arguably a critical factor in causing the 2007-2009 financial crisis.

This second volume explores three particular domains of shadow banking. The first domain deals with the macro-economic fundamentals of the respective shadow banking segments: Why do they exist, what problems do they solve and why are some of their embedded risks so persistent? The second domain captures the global dimensions of shadow banking markets, reviewing the particularities and specifics of various shadow banking systems around the world. Volume II concludes with an extensive overview of how the sector has changed since the financial crisis, focusing on regulatory arbitrage, contract imperfection and governance.

Closing on unresolved issues and open-ended questions that will no doubt remain prominent in the shadow banking sector for years to come, this handbook is a must-read for professionals and policy-makers within the banking sector, as well as those researching economics and finance.

• Language: English
• Publisher: Palgrave Macmillan
• Release date: Jun 30, 2020
• ISBN: 9783030348175

Book preview

© The Author(s) 2020

L. NijsThe Handbook of Global Shadow Banking, Volume IIhttps://doi.org/10.1007/978-3-030-34817-5_1

1. The Macroeconomic Dimensions of Shadow Banking

Luc Nijs¹  

(1)

Hong Kong, China

Luc Nijs

1.1 Introduction

It is well understood by now that the shadow banking (SB) industry, its players, activities and evolution do not occur in isolation. In fact they are a direct function of all elements including the nexus that shadow banking has with the many policy domains including microeconomic, macroeconomic and monetary and fiscal policies as well as financial regulation and supervision. Also the evolution and innovation in the financial sector impact its outlook, design and content. A look back in recent history provides ‘fuel’ for that line of thinking. Starting in the 1990s and pretty much all the way into the start of the 2008 financial crisis, the financial system (in the US but also elsewhere) did go through a period of rapid change, growth and innovation. Banking as an industry did transform away from the traditional intermediary functions as loan origination and deposit-taking and engaged into a ‘securitized’ banking model through which loans were ultimately distributed to entities in the shadow banking sphere.¹ The implication of that happening is that shadow banking entities came to replicate or at least engage in traditional banking activities including credit and maturity transformation. The consequence was that these shadow banking activities took the same risks as banks but with a (very) limited capital base. That combined with overleverage of the financial system overall created the by now well-known consequences which included financial stability and recessionary conditions.

What happened in the period 1990–2008 has however been in the making for decades though. Also in the period before 1990 shadow bank credit expanded each time banks went through traditional cyclical contractions. Even more, while consumer credit and mortgages held by traditional banks were positively correlated with gross domestic product (GDP), those holdings, outside the banking sector, were negatively correlated. The two aggregates move in different directions following a monetary tightening.² It was Meeks et al.³ who demonstrated (through a general equilibrium model⁴) that the ability of traditional banks to securitize can stabilize the overall supply of credit in the system by offloading it in the shadow banking system, but that it (‘the risk taking by those shadow banking entities’) is also at the root cause of increased macroeconomic volatility.⁵ Their model doesn’t however capture all complexities of shadow banking activities,⁶ financial innovation (and its flaws⁷) and regulatory change (prudential regulation and financial system regulation) as well as the imperfect or suboptimal working of some asset markets (e.g. pricing in the collateralized debt obligation, or CDO, markets⁸). Also the dynamics of special purpose vehicles and their use to reduce the amount of capital required are left out of scope.⁹ Their model is built on the understanding that there are ‘two types of financial intermediary, each facing endogenous balance sheet constraints which depend on their net worth’.¹⁰ The commercial banks originate the loans and decide how much of those to keep on their balance sheet and how much to offload. From there the securitization process starts as described elsewhere in the book. As indicated there, securitization doesn’t mean that commercial banks are no longer exposed to risk. Commercial banks in their turn invest in securitized products as they are better quality ‘collateral’ material than an idiosyncratic loan book on a balance sheet of a commercial bank. They achieve that by ‘exchanging a direct exposure to the real economy for an intra-financial claim’,¹¹ thereby reducing their costs and possible constraints regarding funding and increase their leverage and profitability. The securitization process distributes that risk, often in a rather opaque way over multiple balance sheets of shadow banking entities creating vulnerability, but with the benefit that it expands the supply of credit by broadening the base of quality pledgeable assets. In case of an adverse shock in the system,¹² the traditional and shadow banking system moves in tandem as the supply of collateral from the banks dries up, reducing the shadow banking activities. That shortage of collateral also makes commercial banks reduce credit supply to the real economy.

Although shadow banks and traditional banks have separate economic functions, they both face financial constraints. The shadow banking market funds itself through the issuance of securitized products (among others). Those products were bought by, among others, commercial banks looking for higher-quality collateral. Nevertheless, shadow banks retain the more risky (equity or first loss tranches) of securitized products. The distribution of the remaining risk between shadow banking entities and investors relies heavily on the type of liabilities issued by the shadow banking entity. That can be: (1) asset-backed securities (ABS) may offer ‘pass-through’ exposure to an underlying collateral pool (risk sharing whereby the risk ‘contingency of cash flows in the underlying loan pool’ is shared between SB and investor)¹³; or (2) ABS represent fixed (non-contingent) claims,¹⁴ that is, brokers issue one-period discount bonds that promise a fixed return. The consequence of that is that the SB incurs all the risk embedded in the transaction. By doing so, the SB starts to incur bank-like risks (through credit and maturity transformation). This distinction is critical as Meeks et al. demonstrate that ‘the composition of the ABS portfolio turns out to be a crucial determinant of both the relative volatility of bank and shadow bank credit, and the co-movement between them’.¹⁵ That risk is withheld in the financial system and concentrated on the balance sheets of financial intermediaries and not distributed to ‘external’ unlevered investors.¹⁶

The results of the equilibrium model exercise performed by Meeks et al. demonstrate that difference. While in the first scenario (risk sharing) commercial banks are exposed to aggregate risk through both loan and ABS (or in general securitized products) prices, whereby the losses they make on ABS reinforce the losses they make on loans which reduces their balance sheet capacity and forces them to rebalance their portfolios away from loans and toward ABS (they anticipate that the ABS value will appreciate after the shock). The SB party also undergoes a reduction in balance sheet capacity, but ‘the decline in the mark-to-market value of their liabilities as the price of ABS falls offers partial protection to their net worth’.¹⁷ Indeed, they can expand their loan holdings by increasing leverage and using the widened loan-ABS spread. In the latter model (where the SB incurs all the risk as commercial banks hold fixed claims). Meeks et al. conclude, ‘the decline in asset prices triggered by the adverse productivity shock is now fully absorbed by shadow bank net worth, which undergoes a substantial contraction.’¹⁸ Asset and ABS values decline consequently. In this scenario, whereby commercial banks’ net worth is (partly) protected, they are able to expand their loan holdings even when scaling back on their loan activities. But, as Meeks et al. further conclude, despite that the ‘total effect on aggregate credit is a substantially larger contraction than under risk sharing securitization, resulting in larger declines in investment, and so a deeper recession’.¹⁹ That can be explained as follows: in this scenario commercial banks are less exposed to aggregate risk and so their balance sheet is less impacted. But without the ability to securitize their loan book, total credit will fall by more (compared with scenario one). To put it in perspective, under scenario one ‘commercial bank net worth receives an additional boost from the revaluation of their ABS portfolios post-shock. They therefore reduce their overall demand for ABS, inducing the loan-ABS spread to fall to eliminate the resulting excess supply.’ In the latter case, ‘the higher leverage of the shadow banking sector tends to create a large ABS supply response which again leads the loan-ABS spread to fall.’²⁰

A further implication is that under scenario two, the macroeconomic volatility observed is larger compared to model one and is caused by ‘the higher effective leverage of the financial system when shadow banks issue debt-like claims’. It was already indicated that commercial and shadow bank credit tend to move in different directions in response to business cycle shocks. But ‘the cyclical behavior of credit components and spreads depend on the source of the shock, as they depend on the differential responses of aggregate investment.’²¹ Or put differently, heterogeneity within the financial system produces different macroeconomic outcomes²² at least in case of a shock that affects directly the leverage of financial intermediaries (‘the collateral value of assets held or issued by the shadow banking system became impaired’).

Securitization shocks lead to a material reduction in real-economy activity even when offset with cuts in official interest rates and enhanced provision of liquidity.²³ The reason for that is that the assets held by the shadow banking system become less effective for raising secured funding.²⁴ The immediate implication of that is a reduction in the supply of securitized assets (banks keep loan portfolios in their books, but selling by SB pushes prices downward impacting commercial banks—the SB balance sheets are highly levered and thus balance sheet adjustments are material). The secondary implication is ‘that shadow bank liabilities become less valuable as collateral for commercial banks’.²⁵ The implication is that commercial banks now have to hold more ABS in their portfolio and are less attractive compared to loans. This has material policy implications²⁶ as Meeks et al. conclude that ‘a policy that raises asset values through direct loan purchases is more effective than a policy that supports the price of ABS (by lending to shadow banks by purchasing asset-backed securities), reducing funding costs for shadow banks’²⁷ and that ‘the ability of banks to securitize loans when their net worth is impaired can have a beneficial effect on the macroeconomy, acting as a stabilizing force for aggregate activity and credit supply. But when securitization is accompanied by high leverage in the shadow banking system, as is the case when ABS have debt-like characteristics, the economy instead becomes excessively vulnerable to aggregate disturbances.’²⁸

1.2 Shadow Banking Dynamics and Monetary Policy

A further domain that deserves attention and that is related to the discussion above is the question that deals with the question to what degree monetary policy is instrumental to the size and dynamics of the shadow banking market, its players and the content of the transactions. The drivers behind the SB market were discussed, in particular to what degree monetary policy related to the growth or slowdown of commercial banking assets and inversely, as discussed above, the slowdown or growth of the shadow banking assets and level of securitization activity. Front-running the actual discussion in this matter, it can be reported that Nelson et al. have considered part of this question and conclude that ‘a contractionary monetary policy shock has a persistent negative impact on the asset growth of commercial banks, but increases the asset growth of shadow banks and securitization activity’²⁹ and ‘cast doubt on the idea that monetary policy can usefully get in all the cracks³⁰ of the financial sector in a uniform way’. Their starting point is what interests us the most; that is, was monetary policy an (important) driver of financial intermediaries’ balance sheet dynamics and evolution in the run-up to the 2008 crisis? Should monetary policy have been more considerate when it comes to the excessive ‘leverage’ buildup and respond in a countercyclical way? These are relevant questions given the fact that US interest rates have been (very) low for a protracted period of time. It also raises questions, to what degree ‘monetary policy’ has effects beyond the reach of the traditional regulatory tools, that is, the effects of monetary policy on the balance sheet growth of financial intermediaries (both commercial banks and shadow banking entities).

Their findings demonstrate that the contribution of monetary policy shocks on asset growth in the financial sector as a whole has been small (less than 10% of the variation in asset growth of US commercial and shadow banks during the period 1966–2007). There is a direct link between the balance sheet dynamics of commercial and shadow banking entities and the role of monetary policy in ensuring financial stability.³¹ The literature on ‘monetary policy shocks’ is not new³² but has traditionally been focused on the impact of the macroeconomy³³ and concluded that the impact is relatively modest when measured at the level of GDP (e.g. impact on GDP of a 100 basis point shock). More recently larger effects were identified of monetary policy shocks on asset prices.³⁴ Since the outbreak of the 2008 financial crisis, the scholarly attention has shifted toward the question ‘how monetary policy may affect the balance sheet dynamics of financial intermediaries’. In short, the answer was that monetary policy might be an important factor vis-à-vis intermediaries’ balance sheets.³⁵ But until recently very little efforts had been going into quantifying that relationship.³⁶

Woodford indicates: ‘the increase in the riskless short-term rate did reduce demand and checkable deposits of households and firms, but this did not prevent a net increase in the overall liabilities of financial intermediaries, including shadow banks.’ Nelson et al. contribute to our understanding of the nature (and quantification) of the causal relationship between interest rate and balance sheet dynamics.³⁷ The increase (100 basis points) of the central banking rate has a persistently negative effect on commercial bank asset growth (−0.1%), while it had a positive effect on shadow banking asset growth (+0.2%). However, they also conclude that the impact of contradictory monetary policies was larger in the past (1970–1980s) than during the low interest rate environment post-9/11. After 9/11 ‘policy shocks contributed positively to commercial bank asset growth, but shadow banking activity that expanded rapidly due to increasing securitization seemed to have been curbed by expansionary monetary policy shocks’.³⁸ They therefore argued that the ‘overall importance of unexpectedly loose policy in the pre-crisis build up was small relative to other contributing factors’.³⁹ To the extent there is an effect it is the ‘financing and liquidity position’ of banks that seem to be the key determinants of the impact of monetary policy shocks on the balance sheets of commercial banks. They facilitate the policy transmission. This has important policy implications as the financial system may be unable, due to asymmetric information, to channel liquidity to solvent but illiquid intermediaries.⁴⁰

The countercyclical impact on shadow banking activity might point, according to Nelson et al., at a ‘waterbed effect’ whereby ‘commercial banks can circumvent tighter funding liquidity constraints following a contractionary policy shock by possibly increasing their securitization activity, leading to a migration of lending activity beyond the regulatory perimeter to the shadow banking sector’. It allows them to transform illiquid loans into more liquid assets and who, once removed from the balance sheet of the commercial banks, are financed by the issuance of tradable securities (rather than with bank assets).⁴¹

They strengthen their model by taking into account the role of asset prices and their impact on the ability of shadow banking entities to intermediate funds.⁴² Falling asset prices erode collateral value which shadow banking entities use to obtain short-term funding.⁴³ The credit crunch caused deleveraging and a further reduction of asset prices. Credit constraints in the financial system substantially amplify the impact of policy shocks on asset prices and on the balance sheet dynamics.⁴⁴ Reflecting asset prices in their model, Nelson et al. report that the 1% increase in central banking rate ‘continues to be pro-cyclical for commercial bank asset growth’ and ‘countercyclical for shadow bank asset growth’.⁴⁵ Over longer periods of time, however, they identify a ‘stable relationship between monetary policy shocks and commercial bank asset growth’ and ‘a countercyclical impact on shadow bank asset growth’.⁴⁶

One of the drivers behind that countercyclicality could very well be ‘securitization’. The knowledge we already had was that securitization made monetary less efficient (by lowering the interest elasticity of output).⁴⁷ Since securitization provides commercial banks with additional sources of funding, it makes commercial banks less prone to cost of funding shocks. The implication is that the regulator can less efficiently direct bank lending through monetary policies.⁴⁸ So, the bottom-line question is whether securitization activities enhance the countercyclical impact of monetary policy shocks on shadow banking or more precisely whether the countercyclical effect of monetary policy works through securitization. To answer that question Nelson et al. estimate ‘the impact of policy shocks on GSE (Government Sponsored Entities) asset growth’.⁴⁹ The answer to that question is positive⁵⁰ and supports the understanding provided by Loutskina that securitization has reduced the sensitivity of aggregate lending supply to traditional bank funding conditions and weakened the credit channel of monetary policy.

They conclude that indeed monetary policy is a powerful tool tackling financial excesses as it has a reach far beyond that of financial regulation, but also that ‘a monetary contraction aimed at reducing the asset growth of commercial banks would tend to cause a migration of activity beyond the regulatory perimeter to the shadow banking sector. The monetary response needed to lean against shadow bank asset growth is of opposite sign to that needed to lean against commercial bank asset growth’⁵¹ which reduces the effectiveness of monetary policy in this respect. That reinforces the need for prudent regulation as advocated by the Financial Stability Board (FSB)⁵² in order to ‘moderate excessive swings in risk taking by commercial banks’ and maintain monetary policy as a last line of defense against financial instability concerns.⁵³ The larger question at stake, given these findings, is ‘whether traditional interest rate policy is at all effective in curbing excessive credit booms fueled by shadow banks’.⁵⁴ The answer to that is far from certain positive, as was the case with macroprudential policies.⁵⁵

The aforementioned ‘waterbed effect’ has proven robust across different assumptions and model specifications. Securitization activity rises after monetary contractions and therefore challenges the effect of monetary policy to achieve financial stability in the market. It will force a disproportionately large size of the monetary policy response (needed to curtain rapid commercial bank asset growth) relative to the past, and it would potentially lead to asymmetric impact across the financial sector (nonuniform impact of monetary policy tools). As I advocate elsewhere in the book, regulatory tools are needed that ‘address the buildup of leverage in the regulated sector more directly than monetary policy does’.⁵⁶ Elsewhere in the book I will review the current regulatory instruments applied in both the traditional and shadow banking market and illustrate the ‘in my understanding’ material benefits of using a Pigovian type of instrument to tackle the issue. Monetary policy ultimately contributed little to the balance sheet expansion of US financial intermediaries post-9/11, leaving room for argumentation toward financial innovation and others as a potential driver.⁵⁷

1.3 Liquidity Transformation in the Shadow Banking Sector

The intermediaries in the shadow banking system aim to produce or maximize liquidity. They do so by issuing securities that behave as ‘cash’ (i.e. very liquid), but under distress when collateral becomes—as discussed—scarce, it turns illiquid. The need of investors, in times of distress, for crash-proof assets forces the shadow banking system to move toward ‘collateral-intensive’ assets. That in itself results in the fact that the SB system shrinks, reducing liquidity and increasing the risk premium, which make prices and investments to fall. Moreira and Savov⁵⁸ also indicate that it is often followed by a slow recovery and collateral runs. A scarcity of capital within the financial network aggravates the bust in any economic cycle.⁵⁹ The macroeconomic dimensions of a business cycle and accompanying regulation run through the balance sheets of financial intermediaries as they are the power box of the economy. The nexus between the macroeconomy and the financial sector seems to fall apart in times of distress. In fact, at those times it is demonstrated that liquidity and wealth generation deviate.

Moreira and Savov document that ‘securities are liquid only to the extent that they are backed by sufficient collateral to make their payoffs insensitive to private information’.⁶⁰ ‘A security is liquid if its expected payoff does not depend too much on private information about the state of the economy. This makes it immune to adverse selection and allows it to trade without incurring price impact or other costs.’

Providing liquidity is clearly linked to and constrained by the supply of collateral as that liquidity promise must be backed by assets. The scarcity of that collateral under distress is fueling the growth of the shadow banking sector. The shadow banking sector engages in liquidity transformation as it allows greater liquidity for each dollar/euro of available collateral. Moreira and Savov comment, ‘[w]hereas an always-liquid, money-like security requires enough collateral to remain informationally insensitive at all times, even in a crash, a near-money security that is only liquid absent a crash uses collateral mainly when it is more abundant, making it cheaper to produce.’⁶¹ As such, the liquidity frontier is created by the amount of collateral available and the demand determines the distribution of securities at any given point in time. It also explains why shadow banking securities are less in demand as a crash is increasingly likely—that is, those shadow banking securities will become less liquid as the likelihood of a crash increases. As such investors will behave on the trade-off between sensitivity of the liquidity supply and the demand for liquidity, as the likelihood of a crash ‘drives a wedge between the current value of an asset and its collateral value’.⁶² That creates the following understanding as reflected in Table 1.1.

Table 1.1

Shadow banking dimensions under different macroeconomic phases

aMoreira and Savov, (2014), ibid., p. 19

b‘By increasing the supply of liquidity for a given amount of collateral, shadow banking lowers discount rates and pushes up prices, investment, and growth’; Moreira and Savov, (2014), ibid., p. 20

cMoreira and Savov, (2014), ibid., p. 3. The link between liquidity transformation and economic fragility is well documented in recent years; see: (1) M. Baron, and W. Xiong, (2014), Credit Expansion and Neglected Crash Risk, Working Paper, (2) J. Bai, Jennie, et al., (2013), Measuring the Liquidity Mismatch in the Banking Sector, Working paper, (3) M. Schularick, and A. M. Taylor, (2012), Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008, American Economic Review, Vol. 102, Issue 2, pp. 1029–1061

dMoreira and Savov comment: ‘[c]ollateral runs are a side effect of shadow banking and the fragility it generates. At times of high liquidity transformation, an uncertainty shock not only contracts liquidity provision, increasing discount rates and reducing prices, it also increases the volatility of liquidity provision going forward. The heightened exposure to future uncertainty shocks makes discount rates more sensitive to crashes. As a result, collateral values drop faster than prices (haircuts rise), further amplifying the contraction in the supply of liquidity’; see Moreira and Savov, (2014), ibid., p. 4

eJ. Bai, et al., (2013), Measuring the Liquidity Mismatch in the Banking Sector, Working Paper

fMoreira and Savov, (2014), ibid., p. 20

The collateral decelerator implies the slowdown of the economic regression but also makes that a future economic recovery is more protracted. The already highlighted ‘flight to quality (assets)’ following an uncertainty shock and the resulting rise in collateral premia make ‘safe, collateral-rich assets appreciate even as overall prices are falling’⁶³ (collateral mining).⁶⁴

The change in the capital mix during an uncertainty shock (liquidity provision contracts) leads to a protracted slow growth scenario even after the uncertainty has receded. The reverse occurs in the buildup phase, where economic growth continues to fuel liquidity transformation (and not economic activity), which is neglected for longer periods of time even when vulnerability is building up.

Collateral runs (aka margin spirals) ‘are episodes during which liquidity creation requires progressively greater amounts of collateral, which is equivalent to a rise in haircuts in financial markets’.⁶⁵ A collateral run occurs when ‘a simultaneous rise in the level and volatility of discount rates’ (caused by an increased level of uncertainty during the peak of the SB market)’…‘The higher level puts downward pressure on prices, while the higher volatility depresses after-crash collateral values, increasing haircuts’.⁶⁶ The higher haircuts create a contraction in liquidity transformation, shifting the market toward more expensive funding, which leads to falling prices. When the recession bottoms out, the opposite occurs; that is, asset prices become less sensitive to uncertainty reducing haircuts or, as Moreira and Savov indicate, ‘the haircut-price dynamic reverses so that a collateral decelerator eventually puts a floor under asset prices’.⁶⁷ Building on that they conclude that ‘haircuts initially rise as uncertainty begins to subside, which slows down the recovery of asset prices. The same mechanism that amplifies downturns also prolongs their aftermath.’

The ‘flight to safety’ occurs also when uncertainty recedes. The demand for safe assets means collateral dries up and liquidity transformation nosedives. The premium for collateral goes up and so does the discount rate. The yield between shadow banking assets and safe assets widens. The higher the level of liquidity transformation (and therefore collateral is scarce), the more the delta widens. That phenomenon will occur until ‘uncertainty rises sufficiently (or the capital mix becomes safe enough) so that shadow banking shuts down, flight to quality disappears. Thus, flight to quality results from the acute shortage of collateral that occurs when uncertainty rises suddenly after a shadow banking boom’, they conclude.⁶⁸

1.4 Policy Implications of the Macroeconomic Understanding of Shadow Banking Dynamics and Activities

After the 2008 financial crisis, governments across the world have turned to ‘unconventional monetary intervention’. Notable in this respect are (1) the Federal Reserve Bank’s (Fed’s) Large Scale Asset Purchase Program (LSAP, 2008–2010) and (2) the Maturity Extension Program (Operation Twist, 2011–2012). The objective was for the former to support prices of mortgage-backed securities by buying them in large quantities. Under the latter, the Fed purchased long-dated Treasuries and sold short-dated ones, with the stated goal of reducing long-term interest rates. The regulator enacted the Volcker Rule (separation of commercial banking and proprietary trading) and the Basel Committee put forward the Basel III package. The question on the table is how these initiatives relate to the ‘liquidity transformation discussed.’⁶⁹ Table 1.2 tries to illustrate, in a condensed format, the nexus between the components and the liquidity transformation channel.

Table 1.2

Policies enacted and the ‘liquidity transformation channel’

aConclusion of Moreira and Savov, (2014), ibid., p. 26

bSee R. Greenwood, (2014), A Comparative-Advantage Approach to Government Debt Maturity, Journal of Finance, Vol. 65, pp. 993–1028

c‘The EMP is based on the idea that risky productive assets are exposed to duration risk just like long-term bonds, so that reducing the supply of long-term bonds might free up balance sheet capacity for risky investment. In our economy the opposite happens because duration becomes a hedge when flight to quality is strong. This makes long-term bonds complements rather than substitutes for risky investment.’ See Moreira and Savov, ibid., p. 27

dJ. Stein, (2013), The Fire-Sales Problem and Securities Financing Transactions, At the Federal Reserve Bank of New York Workshop on Fire Sales as a Driver of Systemic Risk in Tri-party Repo and other Secured Funding Markets, New York, 4 October 2013

e‘The reason is that the cap on liquidity provision curtails the principal advantage of shadow money, the ability to create liquidity with less collateral’; Moreira and Savov, (2014), ibid., p. 28

fMoreira and Savov, (2014), ibid., p. 28

gMoreira and Savov, (2014), ibid., p. 29

1.5 Liquidity and Economic Growth

A further question that requires some answers is to what degree liquidity relates to economic growth. That liquidity in macroeconomic terms is supplied by (short-term) real interest rates that move throughout the economic cycle. That cyclicality of the interest level will determine the credit- and liquidity constraints in any given industry. The interesting question can then be framed how the interaction between the fluctuating short-term interest rate (i.e. the sensitivity of real short-term interest rates to the business cycle) and the credit and liquidity constraint in industries impacts the long-term growth in industries. Aghion et al.⁷⁰ have been looking into this matter and conclude: ‘(i) the more credit-constrained an industry, the more growth in that industry benefits from counter-cyclical interest rates; (ii) the more liquidity constrained an industry, the more growth in that industry benefits from more counter-cyclical interest rates.’⁷¹

Their study is remarkable in the sense that typically a clear distinction is drawn between long-term growth of an industry and the macroeconomic (fiscal and monetary) policies put in place to achieve financial stability. In contrast Aghion et al. argue that stabilization can affect growth in the long run. Financial constraints at the industry level were measured by them either ‘by the extent to which the corresponding industry displays low levels of asset tangibility (this measure captures the extent to which the industry is prone to being credit constrained), or by the extent to which the corresponding industry in the United States features high labor costs to sales (i.e. the extent to which the industry is prone to being liquidity-constrained)’.⁷² Their main finding is that ‘the interaction between credit or liquidity constraints in an industry and real short-term interest rate counter-cyclicality in the country, has a positive, significant, and robust impact on the average annual productivity growth rate of such an industry’.⁷³ The lower the asset tangibility of the sector, the more growth inducing the countercyclicality of the short-term interest rate. The more liquidity-dependent the industry, the more growth enhancing that is when the real short-term interest rate is more countercyclical. Their contribution is that they illustrate that there is a significant growth effect from more countercyclical monetary policies on industries. That complemented our existing understanding of an inverse relation between volatility and long-term growth, and the negative impact on growth of both bad institutions and low financial development.⁷⁴

1.6 Liquidity Risk Transmission in (Shadow) Banking

One of the major concerns that exist among regulators and supervisors is the ability of the connected globalized financial industry to ‘facilitate’ contagion. To be more precise the concern is built around the contagion effect, especially in a day and age when the credit-to-GDP ratio has been structurally growing in recent decades. In the case of the shadow banking segment, that contagion effect has as an additional complication that the contagion effect occurs throughout (and often simultaneously) a cascade of balance sheets. It is an understatement up till this day, we know very little about this effect, about what are the drivers and how they interact, as well as the question what the effectiveness is of central bank and government interventions aiming to neutralize or mitigate the ‘liquidity risk transmission’. One of the major problems is that, in contrast to many related issues, it doesn’t suffice to analyze data from a macro- or aggregate perspective but that a micro-level analysis is needed to advance our understanding regarding this matter.

During the period 2011–2014 a number of projects have been executed that (start to) help us understand this matter in greater detail. More specifically, the question regarding the exposure of (shadow) banks to liquidity shocks and the resulting impact of liquidity shocks on bank lending deserves some further attention. The latter, however, in this context, is only to a minor degree. The prime question is what ‘balance sheet characteristics’ shape the transmission. Buch and Goldberg⁷⁵ have been looking into the matter surveying dozens of studies conducted in the aforementioned period. What can be noted upfront is that the liquidity risk transmission does not depend, in a consistent way, on balance sheet characteristics, but that across entities and countries there are many interesting variations and distinctions in the way liquidity risk is transmitted into lending (‘there is substantial heterogeneity in the balance sheet characteristics that affect banks’ responses to liquidity risk’). In general it turned out that the balance sheet structure of banks and the liquidity management between the bank holding and the subsidiaries appear to be important for the transmission of shocks. Another important feature for lending was the liquidity window provided by central banks, who often replace the funding channels that were temporarily available in times of market distress. Also important in understanding the wider context is that domestic and international banks behave differently.⁷⁶ Domestic banks during market distress show more stable lending patterns than their international counterparts. Those international banks use internal capital markets to manage liquidity, and during times of market distress, they reduce funding to subsidiaries, leading to reduced lending. Further, ‘the structure of host country markets and the bank relationship with that market matters as well for the transmission of shocks. Foreign banks remained more committed to countries hosting an affiliated subsidiary, that are geographically close, and that have developed relationships with local banks.’⁷⁷ The empirical research can be summarized as illustrating the importance of the following features regarding the nature, size and intensity of the shock transmission being built around ‘the nature of the shock (bank-specific versus global), the balance sheet structure, the degree of internationalization of the bank under study, and the role of public sector liquidity support’.⁷⁸

But can the heterogeneity of banks to liquidity transmission shocks across countries be explained?⁷⁹ We know already by now that international banks are more prone than domestic banks. We also know that the pattern of balance sheet drivers co-determines liquidity transmission risk, whereby banks with foreign affiliates are more prone to the liquidity risk than those that don’t. Buch and Golberg add to that understanding that ‘banks with more cross-border lending growth’ are more sensitive to liquidity shocks. Their interpretation is that ‘banks may subordinate cross-border lending relative to domestic lending activity as stress conditions change. This same pattern of results did not in general show up in the changes in bank lending conducted through the foreign branches and subsidiaries of these same banks. In general though, no single balance sheet factor affects banks’ exposure to liquidity risk in a consistent way across time and across countries (or banks).’⁸⁰

1.7 What Predicts Financial (In)stability?

Following the 2008 financial crisis, the search has been on not only to explain contagion and systemic risk and what is driving its size and magnitude, but also what early warning indicators can help explain (or better predict) contagion risk and more specifically financial instability. They should help to develop a set of matrixes in order to identify vulnerabilities in the financial sector. Accordingly, a set of macroprudential tools should be and are developed. In my understanding tax instruments deserve a more prominent role. I consolidated my thinking in the chapter dealing with a Pigovian approach to the financial industry elsewhere in the book (Vol. I, chapter 11) where I further hammered down on the issue regarding the relationship of command and control, quantity regulation and tax instrument in this complex and un-ended debate. Most quantification of those early indicators has been performed using what is known as a Bayesian model, that is, where a set of most probable models are averaged and where the search is for those indicators that are most frequently included. Although a number of indicators seem to appear regularly in stress indexes, ‘excessive credit growth⁸¹’, ‘unstable funding of banks (loan-to-deposit ratio)’ and ‘high return of bank stocks’ seem to the best (or the most consistent) early indicators.⁸²

The main purpose of developing a set of early indicators is ‘to quantify the current state of instability in the financial system, i.e. to summarize the level of stress stemming from different sources into one single (usually continuous) statistic’.⁸³ Different systemic and stress events are sought to be made comparable. That exercise is fairly new⁸⁴ and many different models exist which tend to vary materially in their set-up. The biggest variation in the models is the fact that they differ with respect to correlation between factors being considered or not. For example, some analyses use only levels or growth rates of variables, while some also take the correlation between the different variables into account.⁸⁵ What they have in common is the fact that they all try to capture risks for the respective financial system in three main segments: (1) the equity market (often measured through an index), (2) the money market (often measured as the spread between uncollateralized and collateralized interbank loans) and (3) the sovereign bond market (often measured as the spread between the bond yields of the countries involved in the study), often with equal weighting.

Three main approaches can be identified with respect to early warning indicators: ‘(1) the signal extraction approach’, (2) discrete choice models and (3) the index-based approach.⁸⁶ The second method is used the most and risk to financial stability is broken in a number of dimensions.⁸⁷ In Table 1.3 the risk dynamics of the different channels used are reviewed.

Table 1.3

Different risk channels

aD. Karim, et al., (2013), Off-Balance Sheet Exposures and Banking Crises in OECD Countries, Journal of Financial Stability, Vol. 9, pp. 673–681

bM. Behn, (2013), Setting Countercyclical Capital Buffers Based on Early Warning Models: Would It Work? ECB Working Paper, Nr. 1604, and M. Drehmann, (2011), Anchoring Countercyclical Capital Buffers: The Role of Credit Aggregates, International Journal of Central Banking, Vol. 7, pp. 189–240

cC. Detken, et al., (2014), Operationalising the Countercyclical Capital Buffer: Indicator Selection, Threshold Identification and Calibration Options, ESRB Occasional Paper Series, Nr. 5

dSee: M. Lo Duca and T. Peltonen, (2011), Macro-Financial Vulnerabilities and Future Financial Stress – Assessing Systemic Risks and Predicting Systemic Events. ECB Working Paper Nr. 1311; M. Behn, (2013), Setting Countercyclical Capital Buffers Based on Early Warning Models: Would It Work? ECB Working Paper, Nr. 1604; M. Brunnermeier et al., (2013), Assessing Contagion Risks from the CDS Market, ESRB Occasional Paper Series, Nr. 4

eSeems less relevant as a predictor Eidenberger et al., (2014), ibid., pp. 17–18

fSeems less relevant as a predictor Eidenberger et al., (2014), ibid., p. 17

gJ. Eidenberger et al., (2014), conclude in their analysis that there is a negative correlation; that is, a high share of interbank assets indicates positive market sentiment, that is, a well-functioning (short-term) interbank market (p. 17)

1.8 Traditional Banks and Shadow Banks: Connected at the Hip

From various angles we have already concluded that traditional banks and shadow banking entities coexist in the marketplace and that monetary, fiscal and macroprudential policies and financial regulation co-determine the size, nature and magnitude of both. They are truly communicating vessels: bank lending versus market-based lending. In that sense traditional commercial banks compete with shadow banks. They both are intermediaries that create safe ‘money-like’ claims, but in different ways. Hanson et al. indicate: ‘[t]raditional banks create safe claims by holding illiquid fixed-income assets to maturity, and they rely on deposit insurance and costly equity capital to support this strategy.’ Deposit holders are assured as they aren’t hindered by fluctuations in the market value of the banks’ assets. Shadow banks, however, ‘create safe claims by giving their investors an early exit option requiring the rapid liquidation of assets. Thus traditional banks have a stable source of funding, while shadow banks are subject to runs and fire sale losses’.⁸⁸ Traditional banks therefore have the advantage of holding illiquid fixed-income products (asset-backed securities, corporate bonds and mortgage-backed securities) with modest underlying risk. They tend not to hold liquid products such as money-market paper and T-bonds or equities. In contrast the shadow banks hold relatively and in mainstream liquid assets. What is remarkable in the analysis of Hanson et al. is that they demonstrate that commercial banks followed shadow banking–type strategies before the introduction of the federal deposit insurance system.⁸⁹ They took in deposits and held assets with shorter duration maturities. That caused more runs on banks than is the case today. The turning point in their asset allocation strategy turned out to be the introduction of the deposit insurance system, as Hanson et al.’s model predicted. They substantiate their models with an analysis of the aggregate asset and liabilities structures of financial institutions that indeed show that contemporary commercial banks hold a larger proportion in their portfolio that are illiquid. The structure of the financial intermediation has changed over time, they conclude, and makes that shadow and commercial banks coexist in the spectrum of providing ‘money-like claims’ to customers that can be used for transaction purposes.⁹⁰ They coexist but with a different portfolio of assets backing their activities. Traditional banks hold mostly longer-term financial (FI) products that, although subject to price fluctuations over time, are relatively risk-free at maturity. They rely on the deposit insurance system and their equity position to do so. It allows for deposit holders to ignore the temporal market price fluctuations. But that funding stability comes at a cost in terms of equity and compliance. The shadow banking system relies mainly on a public backstop to ensure money-like claims and a portfolio of mainly liquid assets.

They look at ‘fire sales’ (forced sale below the fundamental value of the asset) as a key source of illiquidity. The ‘fire sale’ of a certain type of asset is larger when the asset is held more by shadow banks and less when the asset is held by commercial banks. The (il)liquidity dynamic and the endogenous fire sale discount reflect an equilibrium. Hanson et al. report, ‘when an asset is held by both intermediary types, the relative holdings of banks and shadow banks must be such that the expected loss to a shadow bank from liquidating an asset at a temporary discount to fundamental value is just balanced by the added cost a traditional bank pays for more stable funding’⁹¹ ‘transitory non-fundamental movements in asset prices are central to understanding financial intermediation, and especially the connection between the asset and liability sides of intermediary balance sheets’.⁹² A stable funding basis is key when your asset portfolio is predominantly made up out of illiquid assets.

Although most of the findings of Hanson et al. are properly embedded in existing literature and research, it also includes some interesting normative features, in particular regarding the migration of intermediation activity from the traditional banking sector to the shadow banking sector. That is a point of contention which is permanently on the radar of regulator and supervisors. In particular the question arises to what degree supervisors should be concerned regarding a migration from traditional to shadow banking activities. The regulatory concern is clear: shadow banking creates negative externalities, as the social cost of fire sales exceeds the private costs⁹³—that is, ‘an intermediary that switches from traditional to shadow banking fails to internalize how this switch reduces liquidation prices and, thus, the feasible amount of money creation by other shadow banks’.

In a traditional setting when the shadow banking is too large and the commercial banking market is too small, from a social optimum point of view, the regulator can restore that optimum by imposing a set of minimum required haircuts (e.g. capital requirement⁹⁴) on shadow banks, and in an effort to push money creation back to the traditional banking sector. But also traditional banks create social costs that are not fully internalized, as safe bank deposits imply that the taxpayer absorbs tail risk.⁹⁵ As a result the structure of ‘financial intermediation may be shaped in important ways by the non-fundamental movements in asset prices—due to fire sales, noise trading, slow-moving capital, and other frictions’.⁹⁶

It is widely acknowledged that the shadow banking sector grew rapidly before the 2008 crisis, to a large degree driven by the demand for the above-discussed ‘money-like’ claims. That dynamic implies that investors treat short-term debt generated by the shadow banking market as money-like claims or investments. That makes sense but has until recently not been properly documented. What we did now was that, and as was discussed at large in the chapter on securitization, the securitization market grew rapidly before the 2008 meltdown because it allowed intermediaries to supply more money-like claims.⁹⁷ These claims (characterized by short-term safety and liquidity) were backed by highly rated long-term securitized bonds as collateral. It was Sunderam⁹⁸ who demonstrated that investors, using the asset-backed commercial paper (ABCP) as a proxy,⁹⁹ indeed treated the short-term debt produced by the shadow banking system as a ‘money-like claim’.¹⁰⁰ He does so by documenting the ‘high-frequency, micro-evidence of tight interlinkages between the markets for Treasury bills, central bank reserves, and short-term shadow bank debt’. Those interlinkages help us learn about the pricing and quantities of those claims cause ‘if deposits, Treasury bills, and shadow bank debt all deliver monetary services, then the behavior of prices and quantities across the markets for these claims will be interlinked’.¹⁰¹ Two important considerations following the analysis are the following: (1) low yields on T-bills are to be associated with the issuance of shadow bank debt (shadow banking debt and T-bills are substitutes),¹⁰² and (2) the shadow banking response is concentrated in the issuance of short-term debt.¹⁰³ Sunderam’s extrapolations indeed indicate that the increase of ABCP in the run-up to the financial 2008 crisis could account for up to 50% by the increased demand for ‘money-like claims’ by investors. The increasing supply of collateral (to the securitization market) in the post-2008 crisis world has made the ABCP supply more elastic, that is, ‘the same demand shock (for money-like claims (ed.)) now produces a larger increase in the quantity of short-term debt.’¹⁰⁴ So, the growth in shadow banking activities was driven by the demand for claims that provide ‘money-like claims’ rather than just short-term debt (T-bills and shadow banking claims are treated as substitutes).

I discussed the information insensitivity of the debt markets already before.¹⁰⁵ The discussed findings can be connected with Sunderam’s analysis regarding securitization markets. Specifically that implies that originators issue too many informationally insensitive securities in good times. That blunts the incentive for investors to become informed. That endogenous scarcity of informed investors fosters markets to collapse when inefficiency takes over from information-ignorance. The need for ‘safe assets’ is often driven by the need of investors to minimize the reliance on costly informed capital and product. That would point at the need for regulation to limit (in a countercyclical way) the issuance of safe securities.¹⁰⁶ Securitization and the accompanied tranching have been argued as being instrumental to economizing information costs but have as a downside that is equally instrumental in facilitating market collapses when the scope for adverse selection rises.¹⁰⁷ Hanson and Sunderam document that securitization as such blunts ‘investor incentives to build the information production infrastructure needed to analyze securitization cash flows…’ ‘causing inefficient collapses arising from the interaction between issuer security design decisions and investor information acquisition decisions’.¹⁰⁸

There are essentially two ways to increase the informational sensitivity of securities issued: (1) ‘limiting the issuance of informationally insensitive debt in good times would raise the demand for informed capital to purchase equity in the primary market, increasing the returns to being informed. This would induce more investors to become informed ex ante and alleviate underfunding in bad times’, and (2) ‘constraining originators to issue riskier debt in good times would raise the adverse selection profits available to informed investors trading in the secondary market, again increasing the incentives to become informed ex ante’.¹⁰⁹ Therefore there is the need for regulatory intervention in this space as the private market is not able to overcome this friction itself. That is caused by (1) a commitment problem (originators cannot commit to limiting their use of informationally insensitive debt in good times as they at that point maximize their profits by issuing large amounts of informationally insensitive debt) and (2) an issuer that issues informationally sensitive securities in good times would induce investors to build ex ante information infrastructure capacity. However, that issuer would not necessarily receive funding from informed investors in bad times, implying that infrastructure is a ‘public good’ from the issuers’ perspective (diffuse costs and concentrated benefits). The consequence is that issuers avoid the cost of issuing information-sensitive securities to informed investors as the risk of underfunded loan pools is material in bad times.¹¹⁰ The information infrastructure of investors therefore is limited to the short term. Hanson and Sunderam therefore advocate that the collapse of the primary market for securitization in 2008 was due to a buyers’ strike (due to adverse selection and lack of information infrastructure) among the uninformed investors upon which the primary market tends to rely.¹¹¹ Their conclusion implies that haircut regulation (in order to limit fire sales) might not suffice to reduce the vulnerabilities of the securitization markets. The fire sale approach¹¹² assumes that the externality is created by the capital structure of the investors; their model suggests that the externality derives from the capital structure of the originators. Regulation that would structure the capital balance sheet of originators (trusts, special purpose vehicles or SPVs, etc.) would then be needed to avoid that ‘near-riskless securities encourage investors to rationally choose to be uninformed’.¹¹³

Although there was an understanding about the demand for ‘money-like claims’ as T-bill substitutes as discussed before in this section, there was little known about the underlying forces that drove demand for securitized products. Chernenko et al.¹¹⁴ identified recently ‘beliefs’ (a misunderstanding of the risks of investing in securitizations helped drive investor demand) and ‘incentives’ (agency problems between professional investors and their principals) as major drivers of demand for these products¹¹⁵ and further report as well that the heterogeneity in demand initially reported seems more uniform at closer look. To start with the latter first, they identify at a macro-level, that within the institutional investor pool (who are the (only) buyers with deep enough pockets to be active in the securitization market) the insurers and mutual funds increased their share in nontraditional securitizations in the run-up to the crisis (2003–2007) but that the growth rate did not keep pace with the broader credit market (and so they became increasingly underweight to the market). On a micro-level, the variation in holdings in the institutional pool was due to variation across investors and opposed to variation over time. The characteristics that can be attributed to the mutual funds and their managers holding proportionately more of the nontraditional securitized products¹¹⁶ boil mainly down to experience—that is, ‘inexperienced managers invested significantly less in nontraditional securitizations than inexperienced managers’.¹¹⁷ Although tested on many different features that could explain the difference in behavior between (in)experienced managers, the one argument demonstrating convincing (historical) evidence is differences in ‘belief’. Those that went through earlier and protracted periods of market turmoil seem to have invested less in nontraditional securitized products than those that didn’t. Previous personal experience seems to matter in this case.¹¹⁸ Overall the heterogeneity picture seems to be that for mutual funds ‘belief’ seemed to have been the largest driver whereas for insurers ‘incentives’ appear to have played an important role.

Their findings fit nicely in the growing theoretical framework that helps to explain the boom and bust in the securitization market. That theoretical framework mainly consists of two wings, that is, those that see ‘distorted beliefs’ and those that see ‘distorted incentives’ as an explanation of the boom in securitized products. That is, on top of the demand for money-like claims as discussed above. In parallel the bust following the boom is explained within that theoretical framework by referring to ‘fire sales’ and ‘buyers strikes’.¹¹⁹ Table 1.4 brings the dynamics of the growing body of literature together.

Table 1.4

Dynamics in explaining the boom and bust pattern of the securitization market

aSee: (1) N. Gennaioli, et al., (2012), Neglected Risks, Financial Innovation, and Financial Fragility, Journal of Financial Economics, Vol. 104, pp. 452–468, and (2) U. Rajan, et al., (2015), The Failure of Models That Predict Failure: Distance, Incentives and Defaults, Journal of Financial Economics, Vol. 115, Issue 2, pp. 237–260

bSee: (1) J. Chevalier, and G. Ellison, (1997), Risk Taking by Mutual Funds as a Response to Incentives, Journal of Political Economy, Vol. 105, pp. 1167–1200; (2) M. Jensen and W. H. Meckling, (1976), Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics, Vol. 3, pp. 305–360; (3) V.V. Acharya, M. Pagano and P. Volpin, (2014), Seeking Alpha: Excess Risk Taking and Competition for Managerial Talent, European Corporate Governance Institute (ECGI) – Finance Working