Financial Stability Risk. Measuring the Illiquidity of Corporate and Sovereign Bonds

Contents

List of Figures

List of Tables

List of Abbreviation

List of Symbols

1 Introduction

2 Review of the Regulation Development of Financial Markets

3 Liquid Market and Market-Making
3.1 Market Liquidity
3.1.1 Swiss Franc Revaluation
3.1.2 The Taper Tantrum
3.1.3 Treasury Market Rally and Volatility of European Sovereign Bonds
3.2 Principles and Importance of Market-Making
3.3 Market-Making Versus Proprietary Trading
3.4 Modelling Market-Making
3.4.1 A Simple Model of Market-Making
3.4.2 The Case of Fixed Income
3.4.3 The Case of Lower Risk-Adjusted Return
3.4.4 The Case of Non-Falling Interest Rates
3.5 Change of Market-Making
3.5.1 Trends in Market-Making and the Behaviour of Market-Makers . .
3.5.2 Drivers of the Trend
3.6 Danger of a Dry-Up
3.6.1 A Model of Herd Behaviour
3.6.1.1 Herding Measure
3.6.1.2 Herding over Time and Price Impacts
3.6.2 A Model of Self-Fulfilling Liquidity Dry-Ups
3.6.2.1 Structure of the Model
3.6.2.2 Equilibrium
3.6.2.3 Externality and the Result of the Model
3.7 The Link between Liquid Markets and the Real Economy
3.7.1 The Households
3.7.2 A Household’s Decisions
3.7.3 The Equilibrium
3.7.4 Result of a Liquidity Shock

4 Data

5 Estimating Illiquidity
5.1 Related Literature
5.2 Bid-Ask Spreads
5.3 The Conventional Liquidity Ratio
5.4 Amihud Illiqudity Ratio
5.5 The Index of Martin
5.6 Marsh and Rock’s Liquidity Ratio
5.7 Change of Prices
5.8 Bao-Pan-Wang Model
5.8.1 Measuring Illiquidity
5.8.2 A Dynamic Approach

6 Conclusion

References

Appendix
A Datastream Information
B Calculations
C Further Results of Measuring Liquidity

Abstract

Market liquidity is most important for financial markets and thus for the real economy. Market-makers seem to provide less liquidity recently. The reasons of such a behaviour are shown within this work. It exhibits the regulations which have changed, the behaviour of market-makers and how financial markets are able to become illiquid. After this more theoretical framework, which refers to financial stability, several measures of liquidity are introduced and empirically tested on a dataset of about 60,000 corporate and sovereign bonds in 34 countries over a period of eleven years. The result, is that bond markets became less liquid within the last three years than during the financial and the following European debt crisis.

Keywords: Market Liquidity, Market-Maker, Measuring Illiquidity, Financial Markets,

Corporate and Sovereign Bonds

JEL Classification: C22, E44, G14, G15, G21, G23

List of Figures

1 Dimensions of Market Liquidity
2 Simplified Model of a Market-Maker’s Profit
3 Example of Multiple Equilibria
4 Spread of German Corporate Bonds
5 Spread of US-American Corporate Bonds
6 Gross Prices of German Corporate Bonds

List of Tables

1 Market Participants’ View on Regulatory Reforms
2 Results of Amihud and Conventional Liquidity Ratio
3 Results of the Index of Martin
4 Results of the Marsh and Rock’s Index for Corporate Bonds
5 Results of the Marsh and Rock’s Index for Sovereign Bonds
6 Results of the Price Change for Corporate, Sovereign and Supranational Bonds per Country, Period and Year
7 Result of the Measure γ for Corporate, Sovereign and Supranational Bonds per Country and Period
8 Dynamics of Illiquidity
9 Number of Corporate Bonds per Country, Period and Year
10 Number of Sovereign Bonds per Country, Period and Year
11 Number of Supranational Bonds per Country, Period and Year
12 Bid and Ask Prices of Corporate Bonds per Period
13 Results of the Index of Martin for Corporate Bonds for Years
14 Results of the Index of Martin for Sovereign Bonds for Years
15 Results of the Index of Martin for Supranational Bonds for Years
16 Results of the Measure γ for Corporate Bonds per Year
17 Results of the Measure γ for Sovereign and Supranational Bonds per Year

List of Abbreviations

illustration not visible in this excerpt

List of Symbols

illustration not visible in this excerpt

1 Introduction

The financial crisis was the largest crisis, since the “Great Depression” in the 1930s. Subsequently the European debt or Euro-crisis followed. As a result, politicians all over the world implemented new and intensified existing regulations. Credit access should be improved, trust between banks recovered and additionally, financial markets should focus more on their main task, serving the real economy. Unfortunately, many of those new or stricter regulations seem to worsen the problem of lacking market liquidity. In October 2015, the IMF warned in its Global Financial Stability Report of an increasing level of market illiquidity in both, industrialized and emerging markets. The concern is about corporate and sovereign bonds, especially where prices are difficult to observe, since most of these bonds are not traded at exchanges frequently.

The resulting question is, what is market liquidity? In addition, market-makers them­selves have to be explained as well as its role in financial markets. Then remains the question how financial markets are connected to the real economy. After that first part, which refers to financial stability, the second part clarifies, whether the markets become more or less liquid from a global perspective and within the countries. For this purpose several measures of liquidity are shown and empirically used on a microeconomic dataset, since the perfect measure does not exist.

The results are important for political decisions according to further and existing regulations of financial markets. They provide a hint whether the existing rules should be adjusted, eliminated or expanded. Moreover, the results are separated by countries, allowing to obtain differences, which are assumed between industrialized and emerging countries. In section two, a short overview of new and adjusted regulations due to the financial crisis is shown. After that in the third section, the term of market liquidity is explained. Furthermore, the concept of market-makers, their role in financial markets and the question why they are important will be clarified. Afterwards, a simple model of market-making is provided, followed by recent observed trends of market-making and the behaviour of market-makers. Then, two models are presented to exhibit how an illiquid market arise. Finally within this section, a model is presented to show the connection between financial markets and the real economy and what will happen if a financial crisis occurs.

Section four is simply the dataset. Its sources are mentioned, and the data usage as well as the separation is described. Section five provides seven measures of liquidity. It starts with an overview of the related literature and explains why certain measures cannot be used. After that, the bid-ask spread, several different ratios and a dynamic approach are shown and empirically used to measure the liquidity of bond markets in several countries and the changes of liquidity over time. Finally, the conclusion summarize the results and provides a short comment how one should react politically to such results.

2 Review of the Regulation Development of Financial Markets

Worlds asset markets increased and became more integrated since 1980, particularly during the 1990s. Global banks and asset managers have played a significant role according to this development (Miranda-Agrippino and Rey, 2015). Others reasons were the new economies which occurred at the end of the Iron Curtain and the deregulation of financial markets in Europe and the US. In Germany for instance, the social-democrats and the green party beneath chancellor Schröder were responsible for a more liberal financial system and established it.

During the Great Depression, the US began to regulate financial markets.^[1] The most popular regulation was the Glass-Steagall Act, which mainly separated commercial banking from investment banking (Carpenter, Murphy and Murphy, 2016). Now banks had to choose whether they want to lend money or to engage in securities underwriting and dealing. After that, regulation held for decades until the end of the 1970s. In 1978, the Supreme Court decided that banks can export the usury law of their home state nationwide. This triggered a competitive wave of deregulation with the result that South Dakota, Delaware and some others eliminated usury rate ceilings completely (Sherman, 2009).

1986, the Federal Reserve (FED) reinterpreted the Glass-Steagall Act. As result, banks were allowed to use at maximum five percent of their gross revenues into investment banking (Sherman, 2009). Later in 1996, the FED increased it to 25 percent (Carpenter, Murphy and Murphy, 2016). Finally in 1999, the Glass-Steagall Act was completely revoked (Sherman, 2009). Nevertheless, it is unclear whether this decision triggered the financial crisis in 2008 or attenuated it (Crawford, 2011). Together with the deregulation, global financial stocks increased sharply from $53 trillion in 1993 to $ 118 trillion in 2003 (Farrell, Shavers and Key, 2005). With the beginning of the financial crisis in 2008, the financial stocks had increased up to $ 175 trillion and grew continuously to $212 trillion at the end of 2010 (Roxburgh, Lund and Piotrowski, 2011).

Since the financial crisis and the following crisis of sovereign debt within the Euro-area^[2] regulations of financial markets have been adjusted. This was not an easy task, because financial markets in various countries are very different according to depth, access, stability and efficiency (Čihák et al., 2013). Additionally common international political decisions are complex and consume a lot of time. Claessens and Kodres (2014) lists some of the key reforms:

- First of all, Basel III was introduced. On the 6th of January 2013 (BCBS, 2013a) the Basel Committee on Banking Supervision (BCBS), published the full text of the global regulatory standards. Even though the committee cannot bind national governments, its rules are practically adopted everywhere, sometimes with modifications (Elliott, 2015). These rules include countercyclical capital buffer for global systemically important banks (G-SIBs).
- Liquidity Coverage Ratio (LCR): The LCR is a completely new requirement, which has been established in January 2013 (BCBS, 2013b). Its target is to ensure, that all institutions beneath Basel III have enough liquid assets, if their demand raises suddenly during a crisis. Additionally it shall ensure that institutions’ business models do not pass to an extremely large disequilibrium between assets and liabilities. In other words, the LCR is a stress test in order to make sure, that banks have the ability to handle a liquidity-crisis of 30 days. It follows, that banks have a higher demand for more long-term liabilities during such a period (Elliott, 2015).
- Identifying G-SIBs and adjust them to reduce the possibility of the status too- big-to-fail. In addition, higher capital requirements and a better supervision were implemented (Claessens and Kodres, 2014).
- Leverage Ratio (LR): Under the rules of Basel II, financial institutions needed more equity for riskier assets and less for safer ones. The equity share depends on the risk. Unfortunately, the risk has been measured mostly by the rating agencies Standard and Poor’s, Moodys and Fitch, which certificated many assets better than they really were. Nowadays a “leverage ratio” has been introduced, which requires the same amount of capital for all corporate assets, without consideration of the risk. For large banks operating in the US, the regulations are even more strict. This has impacts on trading, since assets with very low risks are often involved.^[3] As a result, these financial institutions can only invest on a lower level (Elliott, 2015).
- Net Stable Funding Ratio (NSFR): The NSFR is the second new requirement according to liquidity. It has been launched in October 2014 (BCBS, 2014) to ensure that banks do not have a large mismatch between their liabilities and their funding (Elliott, 2015).
- Single Counterparty Credit Limits: This point refers to the Dodd-Frank Act, and states that the largest banks must ration their credit exposures. They are not allowed to give credits as much as before the crisis. Certainly this produces opportunity costs (Elliott, 2015).
- The Volcker Rule: This regulations is a little bit unclear, because actually nobody knows what is the exactly meaning of it. It refers to “proprietary trading”. Since it is unclear, banks dealing carefully. They try to prevent that inventories rise too much or too quickly. Still it is too early to evaluate the Volcker Rule finally. Also, it refers to the situation when a bank is both, a market-maker and a proprietary trader. A detailed description of market-making and proprietary trading will follow in the next section.

Thus, after three decades of deregulation governments all over the world have begun to rethink about those developments and finally to tighten regulations, but not without consequences. On the one hand, some of these consequences were intended like more transparency or a better stability of financial markets. Also, the financial industry in the US has shifted. They invest more conservatively, capital has grown and leverage has fallen (Qualtieri and McCusker, 2014). On the other hand, negative repercussions appear as well. In particular, the improvement of the stability of financial markets could be a fallacy because these developments had potentially conflicting effects on market liquidity (IMF, 2015).

3 Liquid Market and Market-Making

A key role for the stability of financial markets is liquidity (Borio, 2009). It is closely related to market-making, meaning it is necessary to explain both. This section firstly offers a description of both terms, liquidity and market-making . In addition, recent changes of market-maker behaviour are presented, next to an explanation of the importance of market-making. Finally, three models are shown, a simple one of market-making and two of a dry-up.

3.1 Market Liquidity

The term liquidity refers to at least three different things. It has a tendency to be slippery in meaning (Hicks, 1962). When people speak about liquidity mostly they mean monetary supply, which can be named “global liquidity”, when referring to the world (Bank of England, 2007). Another notion is “monetary liquidity” which is associated with monetary aggregates (IMF, 2015). The second term is known as “funding liquidity”, which is defined as the ability of a bank or another financial institution to settle obligations with immediacy (Drehmann and Nikolaou, 2009). In other words, a bank has different assets in its balance sheet and sells them to customers, who hold them to maturity (de Haan, Oosterloss and Schoenmaker, p. 58, 2015). Of course, it is problematic, when a large part of assets remains in the financial sector, due to the remaining risk. Acharya and Schnabl (2009) state that this happened during the financial crisis in the US, when 30% of all AAA asset-backed securities^[4] remained within the banking system. This fraction rises to 50%, if one includes ABCP^[5] conduits (asset-backed commercial papers) and SIVs^[6] (structured investment vehicles) with recourse.

The third term is known as “market liquidity” — the ability to rapidly execute large financial transactions at a low cost level and with limited price impacts — and it is very important for financial stability and the real economic activity (IMF, 2015 and CGFS, 2014). Nevertheless, these three different liquidity concepts are related. When companies and households hold more money on their balance sheets because the supply of money has grown sharply, they will probably buy more financial assets. As a result, it is possible that market activity is stimulated and market liquidity increases (Bank of England, 2007).

Market liquidity is a three-dimensional concept. The first one is “tightness” (sometimes called breadth), which means the cost of turning around a position (Kyle, 1985). It can be measured by the bid-ask spread^[7] (Kerry, 2008). The second concept is “depth”, that refers to the volume which can be traded without or at least minimal price impacts (Sarr and Lybek, 2002). The third concept “resiliency” is defined as the ease with which prices returns back to normal (Qaultieri and McCuster, 2014). It refers to the time, how fast prices reach their equilibrium (Bervas, 2006). Often a fourth concept is mentioned named “immediacy”, which is the speed with which trades of a given size can be dealt to given costs (Kerry, 2008). The dimensions depth, tightness and resilience are illustrated in Figure 1. The bid-ask spread shows the transaction costs (i. e. the difference between the buy and sell price). Within this spread, a given amount can be traded without any impacts on the price. This is represented by the segment [illustration not visible in this excerpt] for sales and the segment [illustration not visible in this excerpt] for purchases. Beyond point Q1 and Q2 the price is going to change, because transactions have reached a critical amount.^[8]

Prices of assets changes over time, which is not unusual, but if prices changes suddenly adversely impacts can occur. They can destabilize financial systems by generating problems of key intermediaries, like banks. These intermediaries do not only make losses in their trading books, but also hamper them to hedge market risks effectively, by affecting their ability to distribute risks. Furthermore, they can make losses by trading with counterparties, which are active on financial markets as well (Kerry, 2008). In addition, it is possible, that a bank for instance, needs to raise cash quickly, perhaps because of a large

illustration not visible in this excerpt

maturity mismatch of their assets and liabilities. If the market for those assets is illiquid, then a quick sell to significantly lower price is likely. Unfortunately other intermediaries are also affected, because they must sell their assets, too, since prices deteriorate or maybe because they use “mark to market” accounting (Elliott, 2015). In the worst scenario a liquidity crisis can arise from such market events. It is defined as “sudden and prolonged evaporation of both market and funding liquidity, with potentially serious consequences for the stability of the financial system and the real economy” (Borio, 2009).

If a liquidity crisis occurs, trading is costly or even impossible (Kristoffersen and Pedersen, 2015). This is a dangerous situation, because return for investors are affected negatively. In opposite to their image, financial markets are not only useful for investment banker, hedge funds, shadow banks and others, but also mainly for investors who save for retirement or college. In addition, it affects the costs of corporations, governments and other borrowers, in particular in the US, where most of the credit provided to businesses and households comes from financial markets (Elliott, 2015).

All these credits are supplied by insurers, pension funds, mutual funds, individual investors, and others, though the primary source are households. Assuming a liquidity crisis, tremendous consequences on the economy might arise, depending on the strength of the crisis (Elliott, 2015). Thus, financial markets are crucial to establish greater efficiency between investors and companies, the savers and the borrowers. Well-functioning financial markets are a key factor for economic growth (Mishkin and Eakins, p. 42, 2012). Currently, they became more fragile and a heavy liquidity crisis can occur suddenly according to the IMF and the Bank for International Settlements (BIS), for instance (IMF, 2015 and Fender and Lewrick, 2015). They refer to market liquidity for financial assets, particularly to bonds issued by companies and governments, known as “fixed income” instruments. Since 2013, which means after the financial crisis and after the peak of the European debt crisis, it came at least to four occasions which may indicate a greater vulnerability of fixed income markets. Likely, they are a result of significant market and regulatory changes (Elliott, 2015).

3.1.1 Swiss Franc Revaluation

One of the four events which could be a hint of more vulnerable fixed income markets with excessive volatility is the Swiss Franc Revaluation (SFR). By ISO 4217, 162 currencies exist worldwide (XE, 2016). Important (in a sense of international trade and currency reserves) are just a few, namely the US-Dollar, the Euro, the Japanese Yen, the British Pound Sterling, the Swiss Franc and increasingly the Chinese Yuan^[9] (Auboin, 2012). Throughout the European debt crisis, the Swiss Franc appreciated almost to Euro parity, since it carries the strongest safe haven attributes (Ranaldo and Söderlind, 2007), which prompted the Swiss National Bank (SNB) to intervene (Grisse and Nitschka, 2015). In order to alleviate the repercussions of appreciation for the export oriented companies (Elliott, 2015), a minimum exchange rate of 1.20 Swiss Franc per Euro has been implemented from September 2011 until January 2015 and from the very beginning it had been questioned (Jermann, 2016).

Thus, the SNB was forced to buy more and more Euro (mostly the SNB bought sovereign bonds of Germany) and had to consider that the large amount of Euro would create losses, due to its ongoing depreciation and, of course due to the Quantitative Easing (QE) program of the European Central Bank (ECB). As the SNB gave up this policy, it took 13 minutes for the Swiss Franc to appreciate by 30% against the Euro. Some observers believe that this effect would have been weaker, but it is hard to judge (Elliott, 2015).

3.1.2 The Taper Tantrum

In May 2013, an event occurred which is known as “Taper Tantrum”. FED chairman at that time was Ben Bernanke and on the 22th of May he spoke at the United States Congress. He announced that the FED would begin to reduce the purchase of assets, if warranted by economic data. On the 19th of June, Bernanke suggested that the bond-buying program could end in 2014, if the economy will improve (Kurov and Stan, 2016). That was quicker as some financial markets participants expected and prices of sovereign bonds decreased, the 10-year Treasury bond of the US lost about 3% in two days for instance (Elliott, 2015).

Though the brunt was felt acutely by emerging markets. Next to some macroeconomic reasons, like weak fiscal balances, high inflation, low GDP growth, a high share of debt holding by foreigners or large current account deficits, reduced market liquidity contributed to the increased volatility in emerging markets (IMF, 2014). Furthermore, the Taper Tantrum was not just an effect of market uncertainty, but also a fundamental effect, rather than a financial effect (Lee, 2016).

3.1.3 Treasury Market Rally and Volatility of European Sovereign Bonds

The third event has actually no name. Some newspapers simply called it “Flash Crisis”, unfortunately the stock market crashed on the 6th of May 2010 has the same name. Here a different event is described. It occurred on the 15th of October 2014 and following Elliott (2015), it will be called “Treasury Market Rally”, to avoid confusions.

On that day, the market for US Treasuries, futures and other closely related financial markets were very volatile and liquidity became significantly strained. The market depth for 10-year securities was 20% below the year-to-date average between a 15 minutes period. That sounds negligible, but in one of the world’s most liquid market it is highly unusual. At the end of the day, the yield was 2.14% and only six basis points below the closing price of the previous day. Such high amount is unusual. Since 1998, larger ranges occurred only three times. There are several reasons to explain this event, like record trade volumes, a decline in order book depth, changes in order flow and liquidity provision, and unusual market activity. Nevertheless, again a decline of liquidity was observable (US Department of the Treasury et al., 2015).

In the first half of 2015, prices of sovereign bonds fluctuated sharply. Also, there was a lot of volatility during that period. It is less clear compared to the Treasury Market Rally or the Taper Tantrum, whether this is a result of lacking liquidity or not (Elliott, 2015).

Overall, these four events provided some evidence, that market illiquidity is a growing problem, not only in the US, but also in Europe and in emerging markets. Knowing these events and next to the explanation of the development of the financial system it is necessary to examine the role of market-makers to understand the present situation of financial markets and the problem of lacking liquidity. This role is going to be explained in the next point.

3.2 Principles and Importance of Market-Making

Transactions in financial markets mostly require specialized intermediaries. These intermediaries, often banks, are known as market-makers (O’Hara and Oldfield, 1986). They buy immediately a fixed amount of securities at a given bid price or sell them at a given ask price (Ho and Stoll, 1981). Market-makers or security dealers stand ready to buy or sell assets at any time, which is the reason why they are crucial. A simple example clarifies this statement. Imagine an investor who wants to buy assets. He has an interest of returns and in addition, he wants the opportunity to sell assets quickly. Thinking the other way around, consider a small, unknown company which tries to sell bonds to the public. An investor could be tempted to buy such bonds, still, he has an interest to have the opportunity to sell it immediately. If it cannot be resold easily, the probability that an investor will buy such bonds is low. That is the reason why the importance of market-makers, to smooth the functioning of financial markets, cannot be overemphasized (Mishkin and Eakins, p. 595, 2012).

In bond markets, market-makers deal the vast majority over the counter^[10] (OTC) rather than on the centrally limited order books of exchanges. Hence, a number of bond characteristics exist. At first, a large amount of issued bonds reduces competition between investors. Second, bonds have a fixed maturity, so they can be recouped without trading in secondary markets. That means these bonds are often less traded, when the end of maturity is coming. Thirdly, a lot of institutional investors, require execution of large-volume transactions. These transactions can have strong impacts (CGFS, 2014).

Market-makers have different business models, nonetheless, they have much in common. Usually, they have a large client base, which enable them to have a good view of the flow of orders. Furthermore, they can take large principal positions, because of their balance sheets capacities. Moreover, they have continuous access to multiple markets, including those for funding and hedging instruments. Finally, they have the ability to manage risk and market expertise in providing competitive quotes for a range of securities (Fender and Lewrick, 2015).

Figure 2 offers a simplified overview on the market-maker’s revenues and costs. A market-maker buys assets for a given bid price and sells them to a given ask price. The difference, called spread, is his profit. Certainly, this is not the only source of his profits. He also gains from holding assets in his balance sheets. He profits therefore in two ways: First, as the owner of the assets, he gets the interest which must be paid from the issuer (sure, some assets do not deliver income, like zero-coupon bonds^[11] or some sovereign bonds, as described in footnote 3). Second, on the one hand, if asset prices increase during time the market-maker holds them, he will profit. On the other hand, losing capital is possible. The riskier an asset is, the higher must be the compensation which the market-maker demands in form of the bid-ask spread. It is crucial to understand, that market liquidity depends on the willingness of the market-maker to provide the necessary capital to purchase assets. If

Figure 2: Simplified Model of a Market-Maker’s Profit

illustration not visible in this excerpt

Notes: Figure 2 shows a simplified model of a market-maker’s profit. The bid-ask spread is the price, which the market-maker postulate for buying or selling assets. He must also hold a sufficient inventory of securities on his balance sheet in order to facilitate market liquidity. As long as he holds these assets he gains an interest income. Finally he can profit from increasing prices, which is represented in the third box as capital gains. On the other side, the market-maker loses, if prices decline. In addition he must pay for funding or opportunity costs when he hold an asset instead to sell it and offers less market liquidity. Also his costs of hedging increase if he is more risk averse.

Source: Kristoffersen and Pedersen, 2015 he is more risk-adverse, efficiency gets lost and the market become less liquid (Kristoffersen and Pedersen, 2015).

Next to the costs of capital losses, a marker maker is confronted with funding or opportunity costs. The higher the funding costs, the more expensive is holding an inventory. The funding costs can be seen as opportunity costs for market-making. Market­making itself is a risky business, in particular in abnormal times like the several last years. Thus, the market-maker has an incentive to avoid risks, which he does by hedging. The degree of hedging depends on the risk behaviour. That means that the bid-ask spread depends on other financial markets, such as the future and derivative market (Kristoffersen and Pedersen, 2015).

3.3 Market-Making Versus Proprietary Trading

Next to market-makers, other participants influence market liquidity. Temporary market imbalances can also be alleviated by proprietary traders. These traders work closely to market-makers, often within the same institution. Proprietary trading is similar to the market-makers business, but can be distinguished by four points according to the CGFS (2014):

- The first point is called objectives. It refers to the fact, that market-making serves the customer relationship, meaning that the success of market-making is not simply measured by the balance of profits and losses, but also to the business of the clients. This helps to understand, why banks still act as market-maker, although, markets are less profitable in volatile financial market environments. In contrast, proprietary traders mostly pull back from providing liquidity in such uncertain financial markets.
- The next point refers to asymmetric information. Market-makers typically do not trade on any informational advantages, whereas proprietary traders profit from them and often search for such informational advantages. This implies, that proprietary traders, reach better positions, to achieve gains against the trend in financial markets.
- The third points refers to the risk. The risk profiles of both can be very similar, but especially in illiquid markets, positions must be held by market-makers over a long period of time. In addition purchasing assets for their inventory for future demand of their clients is also necessary. Hedging these positions is often expensive or impossible.
- Finally, there exist regulatory definitions. Since the financial crisis, several regulations have been introduced. Often, they include definitions to separate market-makers from proprietary traders. These definitions can be found in many jurisdictions and are often used in banking laws or financial markets regulations. The CGFS (2014) states that “one core element of market-making definitions is the quoting of firm two-way prices on a regular and ongoing basis.” In contrast, proprietary trading is often defined as trading activities only to achieve gain for traders themselves and without any involvement of clients.

3.4 Modelling Market-Making

This section shows a simple model of market-making at first. Afterwards, it is adjusted to fixed income securities. Beyond that, it is going into more detail, dealing with deviating return compared to the bond market in general and the result of a risk-averse market-maker dealing closely on a lower limit.

3.4.1 A Simple Model of Market-Making

Kristoffersen and Pedersen (2015) offer a theoretical model to show, how the bid-ask spread under given, stated assumptions is determined. It is based on the model of Shen and Starr (2002), which is extended by taking discount into account. That means liquid markets are characterised by a small spread and illiquid ones by wide spreads. Further, the spread is the price of the market-maker’s services. Some assumptions have to be taken into account within this model. First the market-maker is a monopolistic competitor, a profit maximizer subject to an (average) zero-profit condition due to the threat of entry. His costs are the opportunity costs of his capital and his marginal costs are increasing as his financial position is extended. The market-maker finances it by debt or by placing increasing equity at risk (Shen and Starr, 2002).

According to Kristoffersen and Pedersen (2015), the market-maker’s net position, Nt, at the time t, of the security, which is a bond in this case, is represented as:

illustration not visible in this excerpt

whereby Vts denotes the sell and Vt the buy volumes at the end of period t. The market- maker acts passively and accept all orders. That turns into a symmetrical spread St. Furthermore, the market-maker has an initial cash position, represented as Mt, arising from previous periods. The costs of providing liquidity are shown as Ct and the value of the market-maker’s position as nt. The price is denoted as Pt. Together they can be written as,

illustration not visible in this excerpt

The evolution of the cash position is represented as,

illustration not visible in this excerpt

whereby [illustration not visible in this excerpt] denotes the relevant discount factor (the inverse interest rate) for the market- maker. In addition, it is the rate at which the market-maker obtains funding. Per assumption, the market has no obstacles of entry and perfect competition, so that the profit, with E denoting expected, is zero in equilibrium:

illustration not visible in this excerpt

where [illustration not visible in this excerpt] is the expected flow of securities in the inventory assumed equal for buy and sell orders. The asset price and the net position is assumed to be uncorrelated, which means that equation (5) is stable, if additionally buy and sell orders follow known distributions (Kristoffersen and Pedersen, 2015).

Next Kristoffersen and Pedersen (2015) assume that

illustration not visible in this excerpt

meaning that the price of the traded asset follow a martingale in discounted terms. As a result, if equation (4) is fulfilled, an equilibrium of the bid-ask spread builds up. Formally it is described as:

illustration not visible in this excerpt

Within a market of perfect competition and without entry obstacles, it follows the basic theories in economics. The price, here the bid-ask spread, is equal to the (expected) marginal costs.

Kristoffersen and Pedersen (2015) follow Shen and Starr (2002) by assuming a quadratic cost function. They do so because holding securities requires financing. Furthermore, the average cost of capital differs, depending on the inventory position of the market-maker. Finally, a quadratic cost function has the advantage, that it can not only represent the capacity to hold risky assets, but also that this capacity is limited and costs increase faster as linear. The costs are defined as

[illustration not visible in this excerpt]

where [illustration not visible in this excerpt] denotes the risk of an asset, which increases the bid-ask spread. Combining now equation (7) and (9) leads to

illustration not visible in this excerpt

with [illustration not visible in this excerpt] as volatility and [illustration not visible in this excerpt] as the covariance of the trading volume. While the volatility increases the equilibrium bid-ask spread, the covariance decreases it. At last, the equilibrium spread increases if the net position Nt increases, implying that costs and risk are functions of inventories.

3.4.2 The Case of Fixed Income

Fixed income markets are difficult to compare with equity markets, since they have some unique attributes. Well-known is the limited volume of fixed income securities, which are traded on public exchanges (Qualtieri and McCuster, 2014). As a result, the simple model has to be adjusted.

As before, I follow Kristoffersen and Pedersen (2015). For the case that a fixed income security is traded, here a bond, equation (1) is adjusted as

illustration not visible in this excerpt

whereby [illustration not visible in this excerpt] denotes the price of a zero-coupon, non-defaultable bond at time t with time-to-maturity z. Moreover, the initial net position is assumed to be positive, Nt > 0. It is assumed for the reason, that market-makers are warehousing assets on average and aggregated. Equation (3) changes a little to

illustration not visible in this excerpt

Since the asset is a bond, the profit function with perfect competition is

illustration not visible in this excerpt

If equation (13), the zero-profit condition is fulfilled, the equilibrium of the bid-ask spread is

From equation (14) one can see, that it is adjusted to cover two terms, the expected costs at expected trading volume, which is the first term, and second, the expected risk- adjusted holding period return from holding the inventory. After the period, it will be sold (Kristoffersen and Pedersen, 2015).

Since equation (14) holds for every cost function, the first term is going to be viewed as given in the further course. When the bond price today and the costs are given, as well as

Nt > 0, the equilibrium of the bid-ask spread is decreasing in the expected future bond price, since given expected future costs are

illustration not visible in this excerpt

As a result, the spread will be higher, if the market-maker expects losses and lower if he expects profits. In addition, he does not only act in this manner for one bond, rather for his total portfolio. Therefore, the market-maker will gain on his market-making. Since perfect competition does exist in the market, but no entry border, the spread tendency is moving downwards, because the market-maker does not want to lose clients. However, he will increase the spread if he expects capital losses (Kristoffersen and Pedersen, 2015).

3.4.3 The Case of Lower Risk-Adjusted Return

Offering market liquidity is the purpose of market-makers, so they are less risk averse, compared with non-financial companies or households who actively participate in bond markets, because of a larger capacity of bearing risk. This implies that market-makers’ demand returns with a lower adjustment of risk. Due to this, Kristoffersen and Pedersen (2015) argue that an adjustment is necessary, so that

illustration not visible in this excerpt

even without any arbitrage opportunity. As a result, every term in equation (14) influence the equilibrium of the bid-ask spread. Since market-makers are less risk averse, they discount the future on a level below that of the bond market. Equation (16) evolves to[illustration not visible in this excerpt].As proof, the model of Cox, Ingersoll and Ross (1985) is used. In the following it is labelled as CIR-model. It has several advantages, since it is well-known in the literature and provides a simple expression for the risk-adjusted return. In addition, a positive interest rate results for sure, even if it is not necessary needed within this analysis (Kristoffersen and Pedersen, 2015).

The CIR-model has an intertemporal approach and is an general equilibrium-asset­pricing model. Its purpose is to study the term structure of interest rates (Cox, Ingersoll and Ross, 1985). Sun (1992) adjusted the CIR within a framework of discrete time, whereas Backus, Foresi and Telmer (1998) introduced stochastic volatility. Together, they lead to the following short rate, which is assumed to follow a stochastic process

illustration not visible in this excerpt

The short rate is denoted as rt, the long-run mean of the short rate as [Abbildung in dieser Leseprobe nicht enthalten¹² ] and φ determines the persistence of rt. Furthermore, the volatility is denoted as [illustration not visible in this excerpt]. The volatility depends on the short rate, because a higher level of the short rate is connected with higher volatility. Equation (17) now allows to rewrite the price of the non-defaultable bond, which is also a zero-coupon, at time t and time-to-maturity z, to

illustration not visible in this excerpt

whereby Az denotes the shape of the average yield curve and Bz denotes the size of the change of the yield curve, if the short rate changes. The yields themselves are denoted by

illustration not visible in this excerpt

Of special interest is Ao. It is the price of risk which the market demands for holding one unit of risk and is a proxy for the risk aversion of the bond market. It is assumed to be negative ([illustration not visible in this excerpt]), because it is necessary to match an upward sloping yield curve within the CIR-model (Kristoffersen and Pedersen, 2015).

Finally with the CIR-model, the second term of equation (14), by taking the first term, the costs as given, can be rewritten^[13] to

illustration not visible in this excerpt

with [illustration not visible in this excerpt]. This implies that the discount rate of the market-maker is the risk­free short rate. It confirms the assumption of a lower demand for risk-adjusted return of the market-maker, because the discount rate of the bond market is now denoted as rt — AoBz-içrt, which is larger than r™, as long as A < 0. The term [illustration not visible in this excerpt]represents an asset-specific time-varying risk premium, whereby[illustration not visible in this excerpt] shows the degree of how much the bond depends of this value of risk. Market-makers are characterized as risk neutral (Kristoffersen and Pedersen, 2015).

The result is of special interest, if one compares it to the situation on financial markets at the moment. As described in the previous sections, institutions like the IMF (2015) or the BIS (CGFS, 2014 and Fender and Lewrick, 2015), as well as private companies such as the State Street Global Advisors^[14] (SSGA) (Qualtieri and McCuster, 2014) are concerned about market liquidity. Since interest rates are on very low level in (almost) every important market, caused by the lax monetary policy of the FED, ECB, SNB, the BOE (Bank of England) and the BOJ (Bank of Japan), it is interesting to observe the connection between interest rates and market liquidity, which is shown by bid-ask spreads.

Low interest rates correlated negatively to a high bid-ask spread. Therefore, in an economy with low interest rates and a purchasing program of bonds by central banks, market liquidity is declining. The reasoning is that the market-maker earns an excess risk premium —[illustration not visible in this excerpt] which lowers the bid-ask spread and the risk premium is increasing in the level of the short rate, as well (Kristoffersen and Pedersen, 2015).

3.4.4 The Case of Non-Falling Interest Rates

From the previous sections of the model, one experienced that future bond prices do not affect the equilibrium of the bid-ask spread directly. They are only affected by an expected change in risk-adjusted bonds. This is still true, if the market-maker is a risk taker or at least risk neutral, but it does not hold if the market-maker is risk averse. So far, this point was neglected, but now it will be also taken into account. In such a case, the future interest rates cannot fall beneath a certain boundary leading to an asymmetric distribution. It results from the fact that market-makers inventories which lead to capital losses, are extra costly. Within the model, the cost function has to be readjusted in a way that inventory losses will be penalised. It can be interpreted as an issue of market-maker’s solvency. In this sense, the cost function shows loss aversion. It expected form is

illustration not visible in this excerpt

with [illustration not visible in this excerpt] as density function for bond price changes (Kristoffersen and Pedersen, 2015).

Assuming the market-maker realises a loss because of Nt > 0 and [illustration not visible in this excerpt] then costs are higher. Therefore, one can show a negative relationship between the bid-ask spread and the level of the short rate, even if the market-maker expects the same return of bonds like the bond market in general. For this purpose, the first term of equation (14) is used and combined with expression (21). The second term must be ignored, since the market-maker expects the same risk-adjusted return as the bond market. The spread is now

illustration not visible in this excerpt

Kristoffersen and Pedersen (2015) simulated equation (22) with interest rates on different levels and states, if the level of the short rate is far away from the boundary, bond prices are almost as likely to increase as to fall. Despite the closer the short rate comes to the boundary, the higher is the probability that bond prices will fall. So, the bid-ask price is higher in the neighbourhood of the boundary. Therefore, loss aversion leads to market illiquidity at the boundary, even in the case of zero-coupon-bonds. In such a case, duration would have rise and thus leading to more sensitive bond prices (Kristoffersen and Pedersen, 2015).

3.5 Change of Market-Making

So far changes of regulation has been introduced on some points. They should be adjusted in some details to improve market-making, since they affect the costs of market- makers heavily. Afterwards, the term market liquidity has been explained and four events were presented which suggest that bond markets became less liquid. Since the importance of market-makers and the liquidity they provide is clarified, as well as the coherence between interest rates and bid-ask spread and therefore liquidity, it is time to go deeper into the question how market-makers changed their behaviour since the financial crisis and determine its drivers.

3.5.1 Trends in Market-Making and the Behaviour of Market-Makers

According to the IMF (2015), market-making has reduced. However it must be distin­guished between countries and the kind of bonds. In most sovereign bond markets, as a result of QE, liquidity has greatly improved (CGFS, 2014), whereas banks in the US hold less and banks in Germany more (IMF, 2015). Holding corporate bonds had decreased in general (IMF, 2015), but with limitations according to Europe.

Before the financial and the European debt crisis, companies took advantage of bank credits or equity markets for refunding. Today, it is more difficult to get a credit from the bank, particularly in the periphery of the Euro-area, such as Italy or Spain. In core-countries like Germany or the Netherlands the situation is better. Overall within the Euro-area, companies have debts in form of corporate bonds with an amount of € 1.017 trillion. By comparison, the amount of credits from banks is € 5.624 trillion and market capitalisation itself, has an amount of € 12.597 trillion (Kaya and Meyer, 2013).

The financial crisis hit almost every financial market, but hardest the market of deriva­tives and fixed income securities. Corporate bond markets are less liquid than those of equities, resulting from a lower dynamic of trading, since they are often traded just once a day or even less (Friewald, Jankowitsch and Subrahmanyam, 2012).

In sovereign bond markets, the situation was worse after the financial crisis as the risk premium for such bonds widened (Sgherri and Zoli, 2009). The spreads could be explained by economic principles, but countries with larger fiscal imbalances got penalised (Schuknecht, von Hagen and Wolswijk, 2010). Nevertheless, nowadays it seems that market liquidity has improved in sovereign bond markets. Searching for reasons, it looks that other factors like the current low-yield environment together with the lax monetary policy and the implementation of QE cover up the full impact of underlying trends in market-making (CGFS, 2014). The CGFS (2014) identified several trends in global financial markets. It observed widened bid-ask spreads, in particular after the 15th of September 2008, which was the day of the breakdown of Lehman Brothers. Markets calmed down after a while, but still today the spread is larger than before the crisis, reflecting higher volatility. Furthermore, costs have increased sharply, measured by US Treasury securities in price coefficients after Fleming (2003), but have almost returned to the pre-crisis level. Next, sovereign bonds turnover ratios have recovered in several emerging markets,^[15] whereby secondary market volumes shrunk (CGFS, 2014).

Another less obvious point is liquidity bifurcation. The CGFS (2014) interviewed several market participants about their concerns. They said, that they fear a situation of liquidity bifurcation, meaning a concentration of liquidity to the most liquid instruments and a negligence of the less liquid ones. Analysing the situation at bond markets, in particular at the sovereign bond market, it is possible that such a bifurcation has happened at least in the US. Nevertheless, it is not clear, since the bid-ask spread is generally wider for less liquid securities. So, unfortunately until now a clear trend evident cannot be confirmed (CGFS, 2014).

For corporate bonds the situation is unclear, as well. As mentioned before, the situation has changed in Europe, so that corporate bonds are issued in a larger volume, because of credit restrictions, especially in the south of Europe. In contrast, market participants considered trading of larger volumes as increasingly difficult. Bid-ask spreads have widened and remained a little larger than before the financial crisis (Fender and Lewrick, 2015). It seems, despite data of debt markets are rarely available, that trading is focused on some few, more liquid bonds (CGFS, 2014).

Moreover, market-makers changed their business model. They focus more on core markets and clients. These clients shall generate more income in other fields. In addition, many market-makers shifted their business model by moving to more market-making and less proprietary trading. Actually, that means better market liquidity but it leads to a lower capacity on balance sheets with the result that larger volumes cannot be traded or need more time. Taking Australia as an example, one can observe that foreign banks have ceased market-making in corporate bonds. Even domestic market-makers in core markets, such as the sovereign bond market, have closed down their activities (Fender und Lewrick, 2015).

Unfortunately, this behaviour of market-makers is nothing else than getting more risk­averse. From the model of Kristoffersen and Pedersen (2015), one knows that might lead to less market liquidity. The situation is getting worse, since the demand for fixed income instruments is growing.

3.5.2 Drivers of the Trend

Drivers of the trend arise from regulation changes. Not surprising, new regulations sometimes move in another direction as planned. In an extreme case, it is possible that a regulation which should improve a situation, makes it worse at the end. Table 1 provides an overview of key regulations and their expected impacts on the profits and losses (P&L) of market-makers (CGFS, 2014).

Table 1: Market Participants’ View on the Potential Impact of Regulatory Reforms

illustration not visible in this excerpt

Notes: Table 1 shows the concerns of market participants according to regulations which were introduced or tighten in the last years. It must be taken into account that this table just represents the expected impacts, not the actual ones.

Source: CGFS, 2014

It is important to take into account, that only the expected impacts are shown. Many market-makers expect less liquidity, because of higher costs for market-making. In addition, they expect the less liquid assets to be affected the most. Another step into bifurcation could follow from this (CGFS, 2014). Furthermore, higher capital requirements for banks result into credit restrictions, which is especially observable in southern Europe. It is possible, that market liquidity depends on the decisions of some large, but few institutions (Fender and Lewrick, 2015).

The CGFS (2014) found next to structural changes, resulting mostly from changing regulations, some cyclical and market driven adjustments. Market-makers act more risk­averse by a reappraisal of their activities. The risk premium increased and the value of trades was assessed far more on a case by case basis (Fender and Lewrick, 2015).

Moreover, risk measures have become more sensitive to changes in volatility. Financial participants focus more on recent developments, whether volatility rises. Another issue is a more aggressive behaviour of market-makers to reduce their exposures, even if volatility and thus risk-aversion decrease. They are concerned about a quick dry-up in the markets and how severely prices can reverse, since they experienced it during the financial crisis. To make matters worse, underlying markets such as the derivatives markets became less liquid, too. So, some hedging strategies are more difficult to implement.

[...]

---

^[1] A good explanation of the Great Depression can be found by Ben Bernanke (2004).

^[2] The countries who use the Euro as common currency and are members of the European Union.

^[3] That means also assets with low capital returns and sometimes even worse. Meanwhile bonds with the best rating are traded, returning a negative yield. An example are sovereign bonds of Germany with a maturity of 2, 5 or 10 years (Bloomberg, 2016a) or sovereign bonds of Japan also with a maturity of 2, 5 or 10 years (Bloomberg, 2016b).

^[4] Asset-backed securities are securities which generate income from a pool of assets. Alone these assets are to small and illiquid to sold them separately. For more information see Fabozzi (2002).

^[5] ABCP are assets, which are issued short-term to finance medium- or long-term assets (Acharya, Schnabl and Suarez, 2009).

^[6] A SIV is typically a company, often incorporated in a tax and regulatory haven, like the Cayman Islands, issuing debt liabilities and purchasing debt assets. It earns the credit spread between these liabilities and assets (Partnoy, 1999).

^[7] This is the difference between the highest price of an assets at which an investor buys and the lowest price where the seller sold.

^[8] Transactions just reached an amount, where prices begin to change. It is absolutely possible, that there is rather a small amount than a large one. It depends on the asset and the market situation.

^[9] Actually that is not perfectly right. The official name of the Chinese currency is Renminbi. A good explanation of the difference is offered by Mulvey (2010).

^[10] The most sales and purchases of financial assets are traded directly between the market participants, because it has two advantages. Firstly, these trades are cheaper, because a placement fee must not be paid. So, especially larger trades will be sold outside of the exchange. Secondly, every exchange only offers standardized assets without any possibility for individual modification, which is often wished. Whereby the second point plays almost no role in bond markets and absolutely no role in equity markets. But it is important in markets of derivatives (Beike and Schlütz, p.31 f., 2010).

^[11] Actually they generate income, but subtler as normal bonds. For more information of the types of bonds, see section data.

^[12] For proof see the appendix of Kristoffersen and Pedersen (2015).

^[13] For proof see appendix A of Kristoffersen and Pedersen, (2015).

^[14] The State Street Global Advisors is a division of the State Street Corporation and third largest asset manager. The assets under management have a size of $ 2.3 trillion on 31th of March 2016 (SSGA, 2016).

^[15] The question is, which countries are labelled as emerging markets and which as industrialized? The CGFS (2014) enumerates as examples Brazil, China, Korea and Mexico. With three of them I agree, but I see it totally different in the case of Korea. According to the World Bank (2016), the GDP per capita of Korea is about $ 27,221.5 in 2015, which is more as in Spain ($ 25,831.6), close to Italy ($ 29,847) and not far from the mean of the European Union ($ 31,843.2).