Selected Essays on Market Microstructure

Table of Contents

List of Tables

List of Figures

List of Abbreviations

List of Symbols

1 Introduction
1.1 Motivation
1.2 Overview and Organization

2 The Structure of the German Equity Market
2.1 General Overview
2.2 Trading Venues
2.2.1 Screen-based Trading at Xetra
2.2.2 Floor-based trading at the Frankfurt Stock Exchange
2.3 Intermediaries and Market Access
2.4 International Comparison of German Equity Market

3 Market Quality of Floor Trading when Screen Trading Closes Early
3.1 Introduction
3.2 Related Literature and Testable Hypotheses
3.3 Data
3.4 Liquidity and Trading Activity
3.5 Probability of Informed Trading
3.5.1 Methodology
3.5.2 Results
3.6 Informational Efficiency of Prices
3.6.1 Methodology
3.6.2 Results
3.7 Conclusions

4 Dual Trading in Anonymous Markets
4.1 Introduction
4.2 Relevant Literature
4.2.1 Dual Traders and Information Steaming from Order Flow
4.2.2 Dual Traders and Information Steaming from Hidden Order Volume
4.2.3 Advantage for Parasitic Traders at Xetra
4.3 Data
4.3.1 Descriptive Statistics
4.3.2 Trader Classification, Order Classification and Trading Profits
4.4 Front-running and Piggy-backing
4.4.1 Empirical Tests
4.4.2 Robustness Tests
4.5 Iceberg Orders and Order Book Information
4.6 Liquidity Provision
4.7 Conclusions

5 Intraday Pattern of Principal- and Agent-Account
Trading Volume
5.1 Introduction
5.2 Literature Overview
5.3 Institutional Details and Data
5.3.1 Subject of Study: Xetra vs. New York Stock Exchange
5.3.2 Descriptive Statistics
5.4 Trading Activities across Time and Trading Regimes
5.4.1 Sub-Sample Test
5.4.2 Regression Framework
5.4.3 Discussion
5.5 Summary

6 Conclusion and Outlook

References

List of Tables

Table 2-I: Index Market Capitalization across Time

Table 2-II: Market Share of Stock Exchanges across Time and Indices

Table 2-III: Additional Order Specification for Xetra Orders

Table 2-IV: Size of Financial Markets in Different Countries

Table 2-V: Absolute Transaction Costs across Countries

Table 2-VI: Relative Transaction Costs across Countries

Table 3-I: Descriptive Statistics – Floor Trading Sample

Table 3-II: Trading Activities at the Floor

Table 3-III: Probability of Information Based Trading at the Floor

Table 3-IV: Informational Efficiency at the Floor

Table 4-I: Detailed Description of Transaction Records

Table 4-II: Descriptive Statistics Total Sample

Table 4-III: Distribution of Traders across Securities

Table 4-IV: Descriptive Statistics Active Sample

Table 4-V: Piggy-backing and Front-running: Number of Significant Coefficients

Table 4-VI: Piggy-backing and Front-running: Mean Coefficients

Table 4-VII: Piggy-backing and Front-running: Robustness Tests

Table 4-VIII: Iceberg Orders: Trading Volume

Table 4-IX: Iceberg Orders: Trading Profit

Table 4-X: Dual Trader and Liquidity Supply: Originator vs. Aggressor Orders

Table 4-XI: Dual Trader and Liquidity Supply: Correlation Analysis

Table 5-I: Descriptive Statistics Total Sample

Table 5-II: Descriptive Statistics U-shaped Intraday Pattern of Trading Volume

Table 5-III: Differences in Intra Pattern during Continuous Trading

Table 5-IV: Differences in Intra Pattern - Robustness Test

Table 5-V: Differences in Intra Pattern during Auctions

Table 5-VI: Panel-Regression Results

List of Figures

Figure 2-I: Equity Indices in Germany

Figure 3-I: Tree diagram of the trading process

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

List of Symbols

Abbildung in dieser Leseprobe nicht enthalten

1 Introduction

1.1 Motivation

O'Hara (1995) defines market microstructure as the “study of the process and outcomes of exchanging assets under explicit rules.” Similarly, Hasbrouck (2007) notes that market microstructure is the “study of trading mechanisms used for financial securities.” Common to both definitions is the focus on trading rules or mechanisms. Taking a look at stock exchanges around the world, we see that trading mechanisms vary substantially.^[1] Trading platforms may involve an intermediary, such as a stock specialist at the New York Stock Exchange (NYSE), or a Kursmarkler on the floor of the Frankfurt Stock Exchange (FSE). Other trading platforms include a centralized location, such as a future pit at the Chicago Mercantile Exchange (CME), or an electronic bulletin board, like Euronext. Despite huge variations in trading mechanism design, all platforms involve prices emerging and buyers and sellers trading. So the basic questions arises, what is the impact of trading mechanisms on trading. Garman (1976) provides one of the first articles in that area and formulates the question as follows:

“We depart from the usual approaches of the theory of exchange by (1) making the assumption of asynchronous, temporally discrete market activities on the part of market agents and (2) adopting a viewpoint which treats the temporal microstructure, i.e., moment-to-moment aggregate exchange behavior, as an important descriptive aspect of such markets.”

As the research departed from the view that security prices are a macroeconomic phenomenon, two different classes of theoretical models emerged, namely, inventory models and information based models. The first class of models analyzes inventory problems of the specialist, or dealers, and execution problems for traders that arise with uncertainties in the order flow itself.^[2] One of the most important findings of these early models is that the bid-ask spread is caused by inventory holding costs incurred by the specialists.

Even though this implication is valid, it is not the entire story. An interesting extension is offered by information-based models. They assume that information asymmetries exist in markets, meaning that some traders have superior information on the fair value of an asset.^[3] These models employ insights from the theory of adverse selection to show how, even in competitive markets without explicit transactions costs, spreads would exist. They argue that the specialists take the bid-ask spread as compensation for the risk of trading with better informed counterparts. Apart from the fact that information-based models provide an alternative explanation for the existence of the bid-ask spread, they have also paved the way for a new way of thinking. Previous models characterize the behavior of security prices when all agents act competitively. The existence of private information, however, means that trader may have an incentive to act strategically to maximize profits. While it is quite natural to assume that the behavior of informed traders might be modeled using a rational expectation framework, depending on the market statistics about others’ information,^[4] it took some time to allow strategic models for uninformed traders. Usually, it was assumed that uninformed traders take a passive role by supplying liquidity, which is determined exogenously. However, allowing uninformed traders to become aware of their situation and let them try to reduce trading losses, introduces a number of interesting dimensions to the trading process. Around the assumption that informed and uninformed traders behave strategically in the market, a large body of literature developed.^[5]

Currently, we observe rapid changes in the trading industry due to technological developments, such as the rise of algorithmic trading, or institutional changes, as for example, the introduction of the Markets in Financial Instruments Directive (MiFID). Analyzing existing theoretical modes of (un)informed traders in the light of these recent developments are interesting for two reasons. Firstly, theoretical models might not be valid because assumptions do not hold true in reality any more. Secondly, available data increased in quantity and quality to a never expected high, allows new empirical tests to confirm existing theoretical models.

Most of these models and the including assumptions are formulated in a time, when market fragmentation was not an issue and market consolidation was expected to progress even more. It was a time when most trading was conducted by human interaction, and no electronic networks existed, to disseminate information in a split second. Therefore, it is questionable, how many of these models are still applicable in the world today. Currently, we observe a very strong growth in algorithmic trading. So far, it is not certain how algorithmic trading will change current markets. However, there is already first empirical evidence indicating that algorithmic trading will do more to financial markets then increasing its speed. Hasbrouck/Saar (2004) analyze limit orders that are cancelled in an extremely brief time (so called fleeting orders). They argue that fleeting limit orders are closer associated with market orders than standard limit orders, because fleeting orders search for immediacy. Their paper shows that technology changes the existing understanding on liquidity and that supply and demand will drastically change in the future.

Additionally, Goodhart/O'Hara (1997) notes that the introduction of high frequency databases and the dramatic fall in costs of gathering and processing data, allows for new empirical investigations of a wide range of issues in financial markets. However, while high frequency data has been available for some time now, the recent rise of purely electronic open limit order books has changed available data fundamentally. Previous empirical studies had to adjust data for obvious data errors due to human interaction (e. g. Wood et al. (1985)) or have some uncertainties with the quality of data and timestamps (e. g. Lee/Ready (1991)). But, electronic trading platforms do not face such problems, as all trading activities take place and are recorded on a computer. Thus, data samples become increasingly available with virtual no data error and trading timestamps exact to 1/100th of a second. These richer and more precise data samples are expected to suffer from less noise or systematic bias, and therefore, might provide new empirical evidence which could not have been observed before.

The aim of this thesis is to contribute to the existing empirical literature by investigating the strategic behavior of informed and uninformed traders under the light of recent developments. We observe their actual current behavior at financial markets and try to assess whether existing theoretical arguments and assumptions are still valid in the world today, or the newly available rich data samples provide new answers to old questions that researchers have not been able to answer before.

1.2 Overview and Organization

The structure of this thesis is as follows.^[6] In Chapter 2 we provide an extensive description of the German equity market. The aim of the chapter is to provide a comprehensive understanding of the markets, which are more closely investigated in the following sections. Firstly, we portray the structure of the equity market in general, followed by a description of the market microstructure of the most relevant stock exchanges in this thesis. Then, we explain how investors are able to access those financial markets. This is necessary to provide some background information on the source of data for Chapter 4 and Chapter 5. Finally, we conclude with an international comparison of the German financial markets. This chapter should help to put this thesis and its results in an international context.

Chapter 3 examines the effects of the reduction of the daily business hours of screen-based main trading system (Xetra) while a parallel floor-based trading system (FSE) keeps operating. Most stock markets are characterized by a number of parallel operating trading systems which interact intensively with each other. Usually, smaller trading platforms take the leading domestic main market as a benchmark in the price discovery process and for closing open trading positions. But what happens if the smaller trading systems suddenly have to act without this benchmark platform? While prior research on parallel trading focuses on changes due to a growing number of trading venues, we analyze market effects when the main trading platform reduces trading hours. We use this natural experiment, which is unique in its setting, to gain a deeper understanding on the importance of network externalities.

Chapter 4 analyzes dual trading at a completely anonymous equity market (Xetra). Dual traders are defined as market participants that trade for their own account and for the customer account parallel. The existence of dual traders is highly controversial in the academic, as well as the practitioner’s world. Opponents of the practice argue that abusive behavior of dual traders results in a decrease of overall market liquidity. Supporters, on the other hand, contend that they provide valuable liquidity to the market, as they are highly skilled and more flexible. We argue that the anonymous open limit order of Xetra provides a more favorable setting for abusive trading strategies, compared to any other market analyzed for those strategies. Our findings will provide valuable insight on the impact of full anonymity on the trading behavior of dual traders.

Chapter 5 studies the differences in intraday patterns of Principal- and Agent Account trading at Xetra. A trader uses his Principal (Agent)-Account when he trades on his own (customer) account. Based on extensive empirical literature, we assume that Principal-Account (P-Account) trading is a proxy for informed trading, while Agent-Account (A-Account) trading is a proxy for uninformed trading. So far, it is not certain whether the trading intensity of informed and uninformed traders differs systematically during the day. New findings have major implications for numerous theoretical and empirical models.

Finally, in Chapter 6, we summarize our findings of the three research essays. We conclude this thesis with an extensive discussion of the results and an outlook on future research.

2 The Structure of the German Equity Market

The aim of this chapter is to describe and classify the structure of the German equity market. Therefore, we provide an introduction to the German equity market, in general, and to the Deutsche Börse AG with its floor and screen-based trading systems for equities. Then, we will briefly discuss how investors are able to access those markets. Finally, we summarize this section in comparing the German equity market with financial markets in other countries.

2.1 General Overview

As described by Theissen (2003), the German equity market is horizontally and vertically fragmented. In the vertical dimension, the market is divided into three segments, namely Entry Standard, General Standard and Prime Standard. The Entry Standard is the least regulated segment, followed by the General Standard. It is designed for smaller companies with a domestic focus. The Prime Standard is the highest regulated segment, aiming at internationally visible companies. The structure of the most German indices is shown in Figure 2-I. The DAX is the blue chip index, including the 30 most liquid stocks. Next, 50 MDAX securities and 50 SDAX securities follow, each representing the “old economy”. The TecDAX comprises the 30 most liquid high tech firms, thus representing the “new economy”. Table 2-I shows the market capitalization of the indices and all domestic securities across time. The DAX index dominates the other indices. Interestingly, DAX is the slowest growing index, while SDAX has the largest relative increase.

Figure 2-I: Equity Indices in Germany

Abbildung in dieser Leseprobe nicht enthalten

* Automobile, Banks, Basic Resources, Chemicals, Construction, Consumer, Financial Services, Food & Beverages, Industrial, Insurance, Media, Pharma & Healthcare, Retail, Software, Technology, Telecommunication, Transport & Logistics, Utilities.

Source: Theissen (2003).

Table 2-I: Index Market Capitalization across Time

This table reports the market capitalization of German indices in Million Euro for December 2003 through 2006.

Abbildung in dieser Leseprobe nicht enthalten

Source: Deutsche Börse (2003), Deutsche Börse (2004a), Deutsche Börse (2005), and Deutsche Börse (2006a)

In the horizontal dimension, the German equity market is fragmented between seven floor-based stock exchanges and one electronic trading system. The Deutsche Börse AG is hugely dominating the equity market by operating the electronic trading system (Xetra) and the floor-based trading platform (FSE). Additionally, there are six other stock exchanges located in Germany: Berlin-Bremen, Düsseldorf, Hamburg, Hanover, Munich, and Stuttgart. Table 2-II compares the relative and absolute market share of Xetra, FSE and all other stock exchanges, combined across time and indices. We see clearly that Xetra (Panel A) strongly dominates the German equity market, with a relative market share ranging between 64.30% (SDAX in 2003) and 97.74% (DAX in 2006). Additionally, we observe that the relative market share increase across time and increases with the size of the index measured as market capitalization. The second largest stock exchange, measured on order book turnover, is the FSE (Panel B). The relative market share of the FSE ranges between 1.33% (DAX in 2006) and 29.95% (SDAX in 2003). Even though the FSE is significantly smaller compared to the Xetra, it still has a higher order book turnover than all other six stock exchanges combined (Panel C) at any point of time and at any indices.

Table 2-II: Market Share of Stock Exchanges across Time and Indices

This table shows the market shares measured as the order book turnover of Xetra (Panel A), FSE (Panel B) and all other German stock exchanges combined (Panel C). We report the absolute order book turnover in Million Euro and the relative share compared to the total order book turnover across all stock exchanges.

Abbildung in dieser Leseprobe nicht enthalten

Source: Deutsche Börse (2003), Deutsche Börse (2004a), Deutsche Börse (2005), and Deutsche Börse (2006a)

2.2 Trading Venues

In the following section, we discuss the trading protocols of the screen-based trading system, Xetra, and the floor-based trading system, the FSE. Even though there are more stock exchanges in Germany, we will concentrate on those two for the following reasons: Firstly, those two trading platforms capture the vast majority of the market in equity trading. Secondly, our analysis takes place only at those two venues.

2.2.1 Screen-based Trading at Xetra

In the following section, we will describe the trading protocol of Xetra. This section is based on Deutsche Börse (2004b) and Deutsche Börse (2004c). However, these sources are rulebooks and tend to be quite complex. Overview articles on the structure of Xetra at a whole (or very specific issues such as the auctions) are provided by Reck (2001), Theissen (2003), Beltran-Lopez/Frey (2006).

Xetra is a fully automated trading system based on an open limit order book providing continuous trading (also referred to as continuous double auction). The computerized order book manages all incoming market and limit orders. Orders are automatically matched by the system based on price and time priority. Xetra operates fully anonymously since the introduction of the Central Counter Party (CCP). In this way, no trader will know who the counterparty of a trade is, pre-trade or post-trade.^[7]

In trading, there are three basic types of orders: market orders, limit orders and market-to-limit orders. Market orders are instructions to trade immediately at the price currently available. So market orders are executed right away, but sometimes at inferior prices. Limit orders are an instruction to trade at the best price available, but only if it is no worse than the limit price specified by the trader. If no trader is immediately willing to take the opposite side at an acceptable price, the order will wait in the open limit order book. Market-to-limit orders are limit orders with a limit price that is below (above) the asking (bid) price. All volume will be executed against standing orders until the order walked through the book and hits the limit price. The order is then converted to a standard limit order. These basic order-types can be further specified through (a) execution conditions, (b) validity constraints, and (c) trading restrictions. For a detailed description see Table 2-III.

Table 2-III: Additional Order Specification for Xetra Orders

This table provides an overview of additional order specifications used at Xetra.

Abbildung in dieser Leseprobe nicht enthalten

Besides the three standard order types with supplementary order specification, Xetra allows for two additional order types. One is called a stop order and the other an iceberg order. Stop orders enter the order book for execution once a specified price is reached. They are available as stop market and stop limit orders. Iceberg orders are undisclosed limit orders, where the trader only displays a fraction of the total quantity she wishes to buy or sell. In different context, they also might be called, hidden orders, or reserve orders.

Trading phases change during the course of a trading day. The exact order of trading phases is determined by the liquidity of the stock. Given that we focus on liquid stocks in our empirical analysis, we describe only the trading model with continuous trading and auctions. In that trading model, the security is traded during 9:00 a.m. and 5:30 p.m. in a continuous trading environment.^[8] Furthermore, continuous trading is interrupted with a call auction at the beginning of the day, at noon and at the end of the trading day.

All auctions function according the same setting. The auction starts with a call of fixed duration, followed by a call with a random end (random phase). At the end of the second phase, the auction price is set and executable orders are matched. The execution price is determined by maximizing the executable volume and minimizing the surplus (exceeding volume that cannot be executed due to an overhang of volume on one side).

As a safeguard measure, Xetra allows for an additional type of call auction, the volatility interruption. A volatility interruption functions as any other call auction and should protects the market from excessive volatility. They are not scheduled for a fixed point of time during the day, but are triggered dynamically. A volatility interruption is triggered if the potential execution price lies outside a defined static or dynamic price corridor around a reference price.^[9]

All securities listed on the market are ranked according to their liquidity. Securities belonging to the low-liquidity category are only traded when they have at least one designated sponsor. Those designated sponsors are required to quote bid and ask prices and to participate in call auctions. They are responsible for providing the minimum requirements of liquidity. For quality assurance, all designated sponsors are regularly rated by the exchange and the results are made public. Designated sponsors have an additional order type available at Xetra, which has not been discussed so far. They may also post quotes in the limit order book and they are defined as simultaneous sell and buy limit orders.^[10]

2.2.2 Floor-based trading at the Frankfurt Stock Exchange

In the following section, we will describe the trading protocol of floor trading at the FSE. This section is based on Deutsche Börse (2006b). However, these sources are rulebooks and tend to be quite complex. Brief descriptions of the structure of floor trading are provided by Grammig et al. (2001), Theissen (2001a), Theissen (2002), Theissen (2003), and many others.

Trading at the FSE is organized similarly to the NYSE. A Kursmarkler (similar to the specialist) conducts trading, and has exclusive access to the information in the limit order book. Contrary to the NYSE, the Kursmarkler may also trade on his own account, but is not obliged to do so. He may quote Preistaxen (also called Pretrades) or Preisspannen. Preistaxen is not binding, as no quantity is publicly announced, in contrast to Preisspannen. Kursmarkler at the FSE mostly quote Preistaxen. Although no quoted depth is publicly announced, the depth at the quotes appears to be reasonably high, as pointed out by Theissen (2001a).

In contrast to Xetra, the floor is not fully anonymous. Firstly, a Kursmarkler may recognize specific floor traders. Secondly, a Kursmarkler may also request the identification code of any electronically submitted order.

Kursmarkler may earn profits from two sources. Firstly, they may earn on their market making activities. And secondly, Kursmarkler receive a commission called a Courtage. The Courtage is a fee that both buyers and sellers have to pay, depending on trading volume of the transaction. However, empirical evidence by Freihube et al. (1999) suggests that profits due to market making are close to zero and negligible.

The operating hours of the FTE are between 9:00 a.m. and 8:00 p.m. The Kursmarkler conducts an opening call auction at 9:00 a.m., goes on to a continuous trading session and finally ends with a closing call auction at 8 p.m.

2.3 Intermediaries and Market Access

In the following section, we will describe how investors may access the markets. We will focus our description on Xetra; however, the institutional setting is very similar to the FSE.

In order to trade in Xetra, every investor has to trade through a trading participant.^[11] Regulations for trading participants are defined in the Rules and Regulations for the Frankfurter Wertpapierbörse (See Deutsche Börse (2006b)). Trading participants are defined as any institution admitted by the exchange (FWB rules §14) and are required to fulfill a number of legal and technical prerequisites (FWB rules §16 and §16a).^[12] Each registered trading participant may install exchange traders, which execute the actual trading. Each trader may trade (a) for her own account, (b) in her own name for the account of third parties, or (c) as intermediaries for contracts to buy or sell (FWB rules §15). Option (a) is classified as the Principal (P)-Account. Options (b) and (c) are classified as the Agent (A)-Account. The trader itself specifies with each trade, for which account she trades.^[13]

For governance purposes, several traders may form a group. Some traders (referred to as senior traders) have access to all orders and trades of the traders within this group and are able to carry out “trading on behalf” activities. Finally, some traders are classified as security officers, which are responsible for managing rights and privileges of all traders of a Xetra Participant and communicating directly with the Deutsche Börse AG. A trader can never have more (but possible less) rights than the participating institution.

Participants may also implement order routing systems (in accordance with § 1 paragraph 12 and §2 paragraph 5 of the Implementation Regulations of the Frankfurt Stock Exchange Concerning Technical Equipment for the Electronic Trading System (Implementation Regulations) [See Deutsche Börse (2006b)]). Each order routing system has its own Trader-ID and all his trades are marked as routing system orders. Nevertheless, all this information is only for internal use, as Xetra is a completely anonymous trading platform. Participants remain responsible for compliance issues, regardless of whether it is an order routing system or a trader. Groups of traders do not consist of traders using order routing systems and regular trading systems. Traders of each type are grouped together differently.

Traders, which trade for personal and customer accounts in parallel, are defined as dual traders. The German Securities Trade Act (Wertpapierhandelsgesetz - WpHG) §32 states that dual traders are required to ensure that the interests of clients come first and that proprietary trades that cause a disadvantage for customer orders are illegal. Hence, trading on information provided by customer order flow via proprietary account is illegal and should not be observed.

2.4 International Comparison of German Equity Market

In the following section, we present descriptive statistics on the liquidity of the German Stock Market. The German financial system is characterized as a typical bank dominated system with financial markets playing only a smaller role.^[14] Supporting that notion, we compare the German financial markets with other countries. The selection of comparable countries is based on qualitative criteria, such as size of economy or size of stock market. The selection is certainly not exhaustive, but includes the most important stock exchanges around the world. Table 2-IV reports (a) total value of share trading, (b) domestic market capitalization, and (c) number of listed firms for the Deutsche Börse AG and comparable stock exchanges. We see that the Deutsche Börse is strongly dominated by total value of share trading (Panel A), domestic market capitalization (Panel B), and number of listed firms (Panel C) across nearly all countries.

Table 2-IV: Size of Financial Markets in Different Countries

This table compares the size of financial markets across countries. We report total value of share trading (Panel A), domestic market capitalization (Panel B), and number of listed firms (Panel C). Total value of share trading is defined as the total number of shares (single-counted) multiplied by the respective matching price.

Abbildung in dieser Leseprobe nicht enthalten

Source: World Federation of Exchanges (2007).

Theissen (2003) points out that financial markets might be less developed, in terms of volume, or in terms of organization of trading and operational efficiency. Even though financial markets are less dominant in the economy, they can still be highly efficiently operated. Thus, in the following pages, we will focus on the comparison of transactions costs in financial markets across countries. It is argued that lower transaction costs are usually associated with higher operational efficiency. However, studies of transaction costs across countries are highly sensitive to the quality of data and choice of methodology. Therefore, results should be interpreted with caution.

Two of the most recent studies of transaction costs across stock exchanges are Jain (2002) and Swan/Westerholm (2004). They analyze the impact of institutional features on transaction costs in the cross-section, using different data sources, time periods and methodologies. Table 2-V summarizes some of their results on transaction costs. Column I (quoted spread) as well as Column II (effective spread) reports results presented by Jain (2002) and Column III (adjusted effective spread) shows results by Swan/Westerholm (2004). Findings indicate that Germany has large transaction costs, compared to most of the other countries. Furthermore, we see that the UK and Germany tend to have similar transaction costs, while at the NYSE and NASDAQ, transaction costs are considerably lower. Additionally, Jain (2002) is confident that the results are robust as they are in line with empirical findings of Domowitz et al. (2001).

Even though results appear reliable, comparing transaction costs on an absolute basis has some methodological shortcomings. Venkataraman (2001) points out that transaction costs systematically differ across firm specific characteristics such as market capitalization, stock price, trading volume, or volatility. He concludes that comparing transactions costs across stock exchanges without adjusting for those factors might be misleading. Thus, Aitken et al. (2006) suggests the following approach. They define a sample of stocks traded at the NASDAQ and find a matching sample for any other stock exchange using price, trading volume and industry. Although this approach does not allow comparing transaction costs directly to each other, they can compare them relative to the NASDAQ sample. Some of their findings are presented briefly in Table 2-VI. We report the time weighted mean of the quoted spread for the specific stock exchange sample and the corresponding NASDAQ sample. Furthermore, we report whether those differences are statistically significant at a 1% level in mean or median. We see that the Australian Stock Exchange (ASX), NYSE and Toronto Stock Exchange (TSX) exhibit lower transaction costs, compared to NASDAQ, while the London Stock Exchange (LSE), Paris and Xetra exhibit significant larger transaction costs. Thus, these results are fairly consistent with Jain (2002) and Swan/Westerholm (2004).^[15]

Table 2-V: Absolute Transaction Costs across Countries

This table summarizes transaction costs measures reported in Jain (2002) and Swan/Westerholm (2004). Jain (2002) report quoted spread (Column I) and effective spread (Column II). Swan/Westerholm (2004) report trade weighted relative effective spread including exchange charges and taxes (Column III). All columns are ranked, starting with the lowest transaction costs.

Abbildung in dieser Leseprobe nicht enthalten

Source: Jain (2002) and Swan/Westerholm (2004).

Table 2-VI: Relative Transaction Costs across Countries

This table summarizes transaction costs measures reported by Aitken et al. (2006). For both the full sample and the sample of thinly-traded stocks (trade value deciles 6 through 10), we present the mean quoted spread for the indicated market and for NASDAQ in basis points. The NASDAQ costs differ from example to example because the matched stocks differ. We also indicate whether the mean and median, in turn, are statistically different at the 1% level in rows 3 and 6, denoted by (*). The Euronext results for the median in row 3 are significant at the 10% level.

Abbildung in dieser Leseprobe nicht enthalten

Source: Aitken et al. (2006)

Finally, Kasch-Haroutounian/Theissen (2006) compares transaction costs at Xetra to Euronext. They define a stock based matching sample based on market capitalization, trading volume, price level and volatility. They find that both quoted, effective, and realized spreads are lower in Xetra. However, differences are not caused by the adverse selection component, by differences in the number of liquidity provision agreements nor differences in the minimum tick size or in the degree of domestic competition for order flow. Therefore, Kasch-Haroutounian/Theissen (2006) conclude that Xetra is the more efficient trading system.

Summarizing the results, we find empirical evidence that Germany has larger transaction costs compared to the USA (NYSE and NASDAQ), Canada (TSX), and Australia (ASX). However, level of transaction costs at LSE, Euronext or Xetra appear to be so close that it is difficult to find persistent and statistically significant differences.

3 Market Quality of Floor Trading when Screen Trading Closes Early

3.1 Introduction

Most highly developed stock markets are characterized by a number of parallel operating trading systems or stock exchanges. Despite that, strong arguments for consolidation markets are fragmented and remain so for long periods of time. Madhavan (2000) describes this phenomenon as the “network externalities puzzle”. On the one hand, advocates of parallel trading markets name positive effects such as decreasing monopolistic rents due to increased competition.^[16] On the other hand, opponents point out that consolidation leads to higher liquidity externalities.^[17]

Another string of literature concentrates on the changes in market quality by the closing of the main trading system and the beginning of aftermarket trading or the ongoing trading on electronic communication networks (ECNs), or regional stock exchanges and find that trading during after-hours is mostly illiquid. The price discovery is lower than during the regular trading day, and consequently, price efficiency is comparably poor.^[18] The results are driven by two factors. Firstly, the main stock exchange does not operate during after hours. Secondly, results are influenced by the intraday pattern of trading.^[19]

All these studies either analyze events oriented at the entrance of a new trading platform or over a long time period in the co-existence of established markets. In both cases, the number of trading alternatives is stable or growing. Market participants here decide whether to stay or go. We hardly ever had the opportunity to analyze an event where market participants suddenly had to switch because the main market reduced its trading hours. This happened in Germany in 2003^[20] and offers the opportunity to study how traders changed their behavior in response to this reduction. Traders have three alternatives: (a) to finish their activities earlier, (b) to wait until the next trading day or (c) to switch to another trading platform. The screen-based trading platform Xetra is the main trading system in Germany, concentrating more than 90% of the overall trading volume on this platform. This huge market share brings us to the assumption that the early closing must have some impact on traders’ activities in other trading systems.

Given the ongoing discussion about the consolidation of trading venues across Europe, it is important to examine the behavior of investors when an existing main trading system stops to operate and investors have to adapt. The effects of reducing the operating hours of a main trading system are expected to be much stronger than those of a small trading platform. The major differences are liquidity externalities, which could be destroyed by this event, while it is unclear how much of the pre-event conditions have to be attributed to positive liquidity externalities. If liquidity externalities are a major contributor to trading, it is possible that the floor-based trading platform is significantly negatively affected. This again could start up a vicious circle, which might force the floor-based trading venue to stop operating during the evening as well.

To our knowledge, this situation has hardly ever been subject of any research. One exception is a recent analysis by Hendershott/Jones (2005) on the market of three exchange-traded-funds when the Island electronic communication network stopped displaying its limit order book, while it was the dominant venue. The study documents that trading activity and price discovery decreases, leading to an overall worsening of market quality. After one year of non-display, Island redisplays quotes again. Hendershott/Jones (2005) show that the second event is much more modest and does not simply reverse the effects. In contrast to our study, Island did not stop to operate, but only denied displaying quotes to “any” market participant. In our German case the screen-based trading venue completely stops operating during the evening.

[...]

---

^[1] Commerton-Forde/Rydge (2004) and Demarchi/Foucault (2000) provide useful summaries of trading procedures in many securities markets and countries.

^[2] For an basic understanding of inventory models see Garman (1976), Stoll (1978), Ho/Stoll (1981) or Cohen et al. (1981). A comprehensive overview on inventory models is provided by O'Hara (1995).

^[3] For a starting point of information-based models, see Bagehot (1971), or Easley/O'Hara (1987). An comprehensive overview on inventory models is provided by O'Hara (1995).

^[4] See Kyle (1985) among others.

^[5] See Admati/Pfleiderer (1988) or Admati/Pfleiderer (1989), among others.

^[6] This thesis is presented as a collection of research papers. As a consequence, the notation may differ in each chapter and some definitions may be repetitive. Chapter 3 (Chapter 5) is based on a working paper published by Schiereck/Voigt (2007b) (Schiereck/Voigt (2007a)).

^[7] For a discussion on the introduction of the CCP, see Hachmeister/Schiereck (2007).

^[8] Before November 3rd, 2003 Xetra used to open until 8:00 p.m.

^[9] The specific thresholds for triggering volatility interruptions are undisclosed. This is a safeguard measure, because other traders might trigger volatility interruptions on purpose, in order to earn profits (see Cho et al. (2003)).

^[10] Empirical literature on designated sponsors is quite limited. But Anand et al. (2006) and Menkveld (2008) examined the impact of paid liquidity providers and concluded that they are beneficial to market quality and of performance of stocks.

^[11] The Deutsche Börse AG lists all current Xetra Participants on the webpage. As of April 30th, 2007, a total number of 256 institutions were named.

^[12] In general, only credit institutions and investment companies are admitted as trading participants.

^[13] As discussions with officials of the Trading Surveillance Unit at the Frankfurter Wertpapierbörse reveal, the classification is reliable for the most part. Some minor problems are with the classifications by foreign traders.

^[14] Rajan/Zingales (1995) or Hackethal/Schmidt (2000) provides supporting empirical evidence for the notion that Germany is a bank dominated system. An extensive overview/ summary on the German financial system is provided by Schmidt/Tyrell (2004).

^[15] We point out that Booth et al. (1999) compare quote spreads for the 30 most liquid stocks in Germany (traded at IBIS) to the 30 most liquid stocks at NASDAQ. They conclude that IBIS stocks have lower spreads. However, Theissen (2003) warns that the results should be interpreted rather carefully, as the data are from 1991 and DAX stocks had more trading volume, in general, at that time, which was not accounted for.

^[16] See Boehmer/Boehmer (2003), Battalio (1997), Wahal (1997), Mayhew (2002), Fong et al. (2001), DeFontnouvelle et al. (2000), Conrad et al. (2003), Biais (1993) or Chowdhry/Nanda (1991).

^[17] See Bernhardt/Hughson (1997), Amihud et al. (2003), Arnold et al. (1999), Madhavan (1995), Jain (2002), Pagano (1989a), Pagano (1989b).

^[18] See Barclay/Hendershott (2003), Barclay/Hendershott (2004) and McInish et al. (2002).

^[19] Numerous studies provide evidence that trading activities display intraday pattern, such as: Wood et al. (1985), Jain/Joh (1988), McInish/Wood (1992), Abhyankar et al. (1997) and Chan et al. (1995).

^[20] The trading system Xetra of Deutsche Börse AG opened every business day from 9:00 a.m. to 8:00 p.m. in the past. Since Nov. 3rd, 2003 Xetra operates every business day only until 5:30 p.m. For further details see Section 3.2.