Enterprise Applications, Markets and

Services in the Finance Industry: 9th

International Workshop, FinanceCom
2018, Manchester, UK, June 22, 2018,
Revised Papers Nikolay Mehandjiev
Visit to download the full and correct content document:
https://textbookfull.com/product/enterprise-applications-markets-and-services-in-the-fi
nance-industry-9th-international-workshop-financecom-2018-manchester-uk-june-22-
2018-revised-papers-nikolay-mehandjiev/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Enterprise Applications, Markets and Services in the

Finance Industry: 9th International Workshop,
FinanceCom 2018, Manchester, UK, June 22, 2018, Revised
Papers Nikolay Mehandjiev
https://textbookfull.com/product/enterprise-applications-markets-
and-services-in-the-finance-industry-9th-international-workshop-
financecom-2018-manchester-uk-june-22-2018-revised-papers-
nikolay-mehandjiev/

Enterprise Applications Markets and Services in the

Finance Industry 8th International Workshop FinanceCom
2016 Frankfurt Germany December 8 2016 Revised Papers
1st Edition Stefan Feuerriegel
https://textbookfull.com/product/enterprise-applications-markets-
and-services-in-the-finance-industry-8th-international-workshop-
financecom-2016-frankfurt-germany-december-8-2016-revised-
papers-1st-edition-stefan-feuerriegel/

Enterprise and Organizational Modeling and Simulation

14th International Workshop EOMAS 2018 Held at CAiSE
2018 Tallinn Estonia June 11 12 2018 Selected Papers
Robert Pergl
https://textbookfull.com/product/enterprise-and-organizational-
modeling-and-simulation-14th-international-workshop-
eomas-2018-held-at-caise-2018-tallinn-estonia-
june-11-12-2018-selected-papers-robert-pergl/

Security Protocols XXVI 26th International Workshop

Cambridge UK March 19 21 2018 Revised Selected Papers
Vashek Matyáš

https://textbookfull.com/product/security-protocols-xxvi-26th-
international-workshop-cambridge-uk-march-19-21-2018-revised-
selected-papers-vashek-matyas/
New Trends in Computer Technologies and Applications
23rd International Computer Symposium ICS 2018 Yunlin
Taiwan December 20 22 2018 Revised Selected Papers
Chuan-Yu Chang
https://textbookfull.com/product/new-trends-in-computer-
technologies-and-applications-23rd-international-computer-
symposium-ics-2018-yunlin-taiwan-december-20-22-2018-revised-
selected-papers-chuan-yu-chang/

Big Scientific Data Benchmarks Architecture and Systems

First Workshop SDBA 2018 Beijing China June 12 2018
Revised Selected Papers Rui Ren

https://textbookfull.com/product/big-scientific-data-benchmarks-
architecture-and-systems-first-workshop-sdba-2018-beijing-china-
june-12-2018-revised-selected-papers-rui-ren/

Computer Vision Applications Third Workshop WCVA 2018

Held in Conjunction with ICVGIP 2018 Hyderabad India
December 18 2018 Revised Selected Papers Chetan Arora

https://textbookfull.com/product/computer-vision-applications-
third-workshop-wcva-2018-held-in-conjunction-with-
icvgip-2018-hyderabad-india-december-18-2018-revised-selected-
papers-chetan-arora/

Cyber Physical Systems Model Based Design 8th

International Workshop CyPhy 2018 and 14th
International Workshop WESE 2018 Turin Italy October 4
5 2018 Revised Selected Papers Roger Chamberlain
https://textbookfull.com/product/cyber-physical-systems-model-
based-design-8th-international-workshop-cyphy-2018-and-14th-
international-workshop-wese-2018-turin-italy-
october-4-5-2018-revised-selected-papers-roger-chamberlain/

Communication, Networks and Computing: First

International Conference, CNC 2018, Gwalior, India,
March 22-24, 2018, Revised Selected Papers Shekhar
Verma
https://textbookfull.com/product/communication-networks-and-
computing-first-international-conference-cnc-2018-gwalior-india-
march-22-24-2018-revised-selected-papers-shekhar-verma/
 Nikolay Mehandjiev
Brahim Saadouni (Eds.)

Enterprise Applications,
LNBIP 345

Markets and Services

in the Finance Industry
9th International Workshop, FinanceCom 2018
Manchester, UK, June 22, 2018
Revised Papers

123
Lecture Notes
in Business Information Processing 345

Series Editors
Wil van der Aalst
RWTH Aachen University, Aachen, Germany
John Mylopoulos
University of Trento, Trento, Italy
Michael Rosemann
Queensland University of Technology, Brisbane, QLD, Australia
Michael J. Shaw
University of Illinois, Urbana-Champaign, IL, USA
Clemens Szyperski
Microsoft Research, Redmond, WA, USA
More information about this series at http://www.springer.com/series/7911
Nikolay Mehandjiev Brahim Saadouni (Eds.)
•

Enterprise Applications,
Markets and Services
in the Finance Industry
9th International Workshop, FinanceCom 2018
Manchester, UK, June 22, 2018
Revised Papers

123
Editors
Nikolay Mehandjiev Brahim Saadouni
Alliance Manchester Business School Alliance Manchester Business School
Manchester, UK Manchester, UK

ISSN 1865-1348 ISSN 1865-1356 (electronic)

Lecture Notes in Business Information Processing
ISBN 978-3-030-19036-1 ISBN 978-3-030-19037-8 (eBook)
https://doi.org/10.1007/978-3-030-19037-8

© Springer Nature Switzerland AG 2019

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

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

Advancements in information and communication technologies have paved the way to

new business models, markets, networks, services, and players in the financial services
industry, developments now labelled as “FinTech.” FinanceCom workshops have been
providing research leadership in this area well before the rise of the FinTech concept,
helping academics and practitioners to understand, drive, and exploit the associated
opportunities in the financial sector.
After very successful FinanceCom workshops in Sydney (twice), Regensburg,
Montreal, Paris, Frankfurt (twice), and Barcelona, FinanceCom 2018 was held in
Manchester (UK) for the first time. The workshop spans multiple disciplines related to
the use of technology for financial services, including analytical, technical, services,
economic, sociological and behavioral sciences, aiming to foster true cross-disciplinary
work for addressing research questions grounded in real industrial needs. Indeed, one
of the main goals of the workshop series has been to bridge the boundaries between
“traditional” business disciplines (such as finance and economics) and “ICT-based”
disciplines (such as software engineering and information systems), which demonstrate
inertia in their focus on isolated parts of a much larger picture in which diverse aspects
are intertwined in complex ways.
This inertia reduces the size of the FinanceCom community yet at the same time it
amplifies the impact of the work presented at our specialized interdisciplinary
workshop series. This year we received 18 submissions, of which we selected
11 high-quality papers to be presented at the workshop and to consequently be
reworked based on the feedback and published in this volume together with an addi-
tional invited paper. The selection was instrumented by a rigorous review process
implemented with the help of a Program Committee consisting of internationally
renowned researchers and practitioners.
The first part of these proceedings contains five contributions in the area of financial
innovation. The first paper is our invited paper “The MiFIR Trading Obligation: Impact
on Trading Volume and Liquidity in Electronic Trading” by Gomber et al. It tests
whether the migration of trading from the over-the-counter (OTC) markets to regulated
trading venues and on platforms of systematic internalizers (Sis) has any effect on
liquidity. The paper shows that liquidity on open limit-order book markets (OLOBM)
increases as a result of the additional turnover coming from the OTC market. However,
trading on SIs has an adverse effect on liquidity for OLOBM. The second paper, “Open
Innovation Effectiveness in the Financial Services Sector,” by Piobbici et al. argues that
an open innovation framework could be useful in analyzing innovation patterns that
involve both financial and non-financial institutions. The paper suggests that
market-based collaborations are significantly correlated with the performance in
relation to innovation. Overall, the paper demonstrates that market-based collaborations
are the most effective. Using a very large sample of 2469 observations and covering a
13-year period, the third paper, “Identification of Financial Statement Fraud in Greece
vi Preface

by Using Computational Intelligence Techniques,” by Chimonaki et al. reports that

computational intelligence (CI) techniques can enhance the detection of financial fraud.
The fourth paper, “What Sort of Asset? Bitcoin Analyzed,” by Corbet et al. examines
whether the introduction of futures contracts in bitcoin addresses the issues that have
inhibited it from being classified as a currency. The paper shows that spot volatility has
increased following the introduction of futures contracts, these contracts are ineffective
as hedging instruments, and price discovery is driven by uninformed investors in the
spot market. Overall, the paper demonstrates bitcoin is a speculative asset and the
introduction of futures contracts has failed to make it a currency. The fifth paper,
“Blockchained Sukuk-Financing,” by Shaikh and Zaka examines the use of sukuk in
Islamic finance as an alternative source of finance to conventional bonds. This source of
finance can involve a number of parties, especially when sale, lease, and agency
contract are combined. The paper presents a model for blockchained sukuk-issue,
highlighting the significant design features that are unique for this source of finance.
This approach should help the traceability of any asset transfers that will ultimately
enhance sukuk credibility and their valuation. Further, a better contract infrastructure
with blockchain security should also significantly reduce the execution time for
transactions involving sukuk.
The second part of the proceedings contains four contributions in the area of market
data analytics. The first paper in this section, “The Role of Customer Retention in
Business Outcomes of Online Service Providers,” by Assemi examines two competing
theoretical models using archival data from a leading crowdsourcing marketplace. The
findings show that a provider’s profile information is significantly associated with the
provider’s business outcomes, while customer retention partially mediates this
relationship. Furthermore, the results demonstrate the important role of new customers
in achieving better business outcomes on crowdsourcing marketplaces. The second
paper, “Using NSIA Framework to Evaluate Impact of Sentiment Datasets on Intraday
Financial Market Measures: A Case Study,” by Qudah and Rabhi examines the impact
of news sentiment on the intraday cumulative average abnormal returns (CAAR) using
cases from Australia, Canada, Germany, and the USA. The results show that in nine out
of the 12 cases examined, the release of the sentiment news has a significant negative
impact on the CAAR. The third paper, “Financial Data Visualization in 3D on
Immersive Virtual Reality Displays,” by Lugmayr et al. presents a design prototype for
the 3D visualization of financial data relating to Australia’s energy sector with a
large-scale immersive virtual reality environment. The paper attempts to develop a
proof-of-concept implementation referred to as “ElectrAus,” a tool to visualize publicly
available data from the Australian energy market. The primary aim of the proof-of
concept implementation is to demonstrate new concepts in data visualization such as
utilization of immersive environment and large screens for financial data visualization;
creating an understandable visualization for the general public; and additional insights
and analytics through 3D displays. The fourth paper, “Document Representation for
Text Mining: Opportunities for Analytics in Finance,” by Roeder and Palmer
investigates the utilization of a distributed representation of words and documents
referred to as “embeddings” for text analytics in finance. The results reveal a potential
application of the document representation techniques for text analytics in the area of
finance.
 Preface vii

The third part of the proceedings contains three papers focusing on the use of
semantic modelling in supporting financial trading. In their paper, “Semantic Model
Based Framework for Regulatory Reporting Process Management,” Pilaka et al.
describe how semantic modelling can help in extracting instances of regulatory
reporting processes from event traces and help with compliance monitoring. The
proposed framework is tested through applying it to event traces from the Australian
Securities Exchange to extract instances of the “off market bid” regulatory process.
The second paper, “Applying Ontology-Informed Lattice Reduction Using the
Discrimination Power Index to Financial Domain,” by Quboa et al. describes the use of
semantically encoded knowledge about asset allocation to support the automatic
clustering of instances sampled from the domain and their tagging with semantic
information. The final paper in this section by Behnaz et al., “A Statistical Learning
Ontology for Managing Analytics Knowledge,” proposes an ontology development
process tuned to developing statistical learning ontologies that can support analytics.
Two case studies ground the research to the domains of commodity pricing and digital
marketing.
Special thanks go to Alliance Family Foundation and Alliance Manchester Business
School Strategic Investment Fund, which supported this event financially through the
Alliance MBS Big Data Forum. We are also grateful to Fethi Rabhi, who has guided us
all the way in getting this workshop and proceedings organized from start to finish. We
are grateful to our team of reviewers and Program Committee members, who worked
very hard with the authors to ensure the quality of the papers included in this volume,
and to Ralf Gerstner and Christine Reiss from Springer for their excellent support in
producing this proceedings volume.

June 2018 Nikolay Mehandjiev

Brahim Saadouni
 Organization

The workshop took place at the Alliance Manchester Business School in Manchester,
UK. Financial support by the Alliance Family Foundation through the Alliance MBS
Big Data Forum is gratefully acknowledged.

Organizing Committee and Program Chairs

Nikolay Mehandjiev Alliance Manchester Business School, UK
Brahim Saadouni Alliance Manchester Business School, UK

Program Committee
Marc Adam University of Newcastle, Australia
Madhushi Bandara University of New South Wales, Australia
Sonia Cisneros-Cabrera Alliance Manchester Business School, UK
Onur Demirors Izmir Institute of Technology, Turkey
Saif Dewan Australian National University, Australia
Stefan Feuerriegel University of Freiburg, Germany
Mahdi Fahmideh Gholami University of Wollongong, Australia
Peter Gomber University of Frankfurt, Germany
Nikolay Kazantsev Alliance Manchester Business School, UK
Stefan Lessmann Humboldt University of Berlin, Germany
Artur Lugmayr Curtin University, Australia
Jan Muntermann University of Göttingen, Germany
Dirk Neumann University of Freiburg, Germany
Maurice Peat University of Sydney, Australia
Helmut Prendinger National Institute of Informatics, Japan
Qudamah Quboa University of Manchester, UK
Fethi Rabhi University of New South Wales, Australia
Federico Rajola Catholic University of the Sacred Heart, Italy
Ryan Riordan University of Ontario, Canada
Michael Siering University of Frankfurt, Germany
Andrea Signori Università Cattolica del Sacro Cuore, Italy
Basem Suleiman University of Sydney, Australia
Christof Weinhardt Karlsruhe Institute of Technology, Germany
Axel Winkelmann University of Würzburg, Germany

Steering Committee for the FinanceCom Workshop Series

Christof Weinhardt Karlsruhe Institute of Technology, Germany
Dennis Kundisch University of Paderborn, Germany
x Organization

Federico Rajola Catholic University of Milan, Italy

Fethi Rabhi University of New South Wales, Australia
Jan Muntermann University of Frankfurt, Germany
Peter Gomber University of Frankfurt, Germany
Ryan Riordan Karlsruhe Institute of Technology, Germany
 Contents

Financial Innovation

The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity

in Electronic Trading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Peter Gomber, Benjamin Clapham, Jens Lausen, and Sven Panz

Open Innovation Effectiveness in the Financial Services Sector. . . . . . . . . . . 27

Francesco Piobbici, Federico Rajola, and Chiara Frigerio

Identification of Financial Statement Fraud in Greece by Using

Computational Intelligence Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Christianna Chimonaki, Stelios Papadakis, Konstantinos Vergos,
and Azar Shahgholian

What Sort of Asset? Bitcoin Analysed . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Shaen Corbet, Brian Lucey, Maurice Peat, and Samuel Vigne

Blockchained Sukuk-Financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Shazib Shaikh and Fatima Zaka

Market Data Analytics

The Role of Customer Retention in Business Outcomes of Online

Service Providers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Behrang Assemi

Using NSIA Framework to Evaluate Impact of Sentiment Datasets

on Intraday Financial Market Measures: A Case Study . . . . . . . . . . . . . . . . 101
Islam Al Qudah and Fethi A. Rabhi

Financial Data Visualization in 3D on Immersive Virtual Reality Displays:

A Case-Study for Data Storytelling and Information Visualization
of Financial Data of Australia’s Energy Sector . . . . . . . . . . . . . . . . . . . . . . 118
Artur Lugmayr, Yi Juin Lim, Joshua Hollick, Joyce Khuu,
and Felix Chan

Document Representation for Text Analytics in Finance . . . . . . . . . . . . . . . 131

Jan Roeder and Matthias Palmer
xii Contents

Semantic Modelling

A Semantic Model Based Framework for Regulatory Reporting

Process Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Manjula Pilaka, Madhushi Bandara, and Eamon Mansoor

Applying Ontology-Informed Lattice Reduction Using the Discrimination

Power Index to Financial Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Qudamah Quboa, Nikolay Mehandjiev, and Ali Behnaz

A Statistical Learning Ontology for Managing Analytics Knowledge . . . . . . . 180

Ali Behnaz, Madhushi Bandara, Fethi A. Rabhi, and Maurice Peat

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

Financial Innovation
 The MiFIR Trading Obligation: Impact
on Trading Volume and Liquidity
in Electronic Trading

Peter Gomber , Benjamin Clapham , Jens Lausen , and Sven Panz(B)

Goethe University Frankfurt, Theodor-W.-Adorno-Platz 4,

60323 Frankfurt am Main, Germany
{gomber,clapham,lausen,panz}@wiwi.uni-frankfurt.de
http://www.efinance.wiwi.uni-frankfurt.de

Abstract. The new financial market regulation MiFID II/MiFIR will

fundamentally change the trading and market infrastructure landscape in
Europe. One key aspect is the trading obligation for shares that intends
to restrict over-the-counter (OTC) trading to ensure that more trading
takes place on regulated trading venues and on platforms of Systematic
Internalisers (SIs). In this context, market experts often argue that SIs
might have a competitive advantage due to the best execution concept
in combination with the possible exemption of SIs from the tick size
regime. Applying scenario analysis, we determine the likely migration of
OTC trading volume to regulated trading venues and SIs. Based on our
data set, we investigate how changes in trading volume influence liquid-
ity on open limit order book markets (lit markets). The results of our
scenario analysis indicate that liquidity on lit markets might increase
due to additional turnover formerly traded OTC. However, also a nega-
tive liquidity effect for lit markets and for the price discovery process is
possible because of increased trading via SIs.

Keywords: MiFID II/MiFIR · Trading obligation · Liquidity ·

Trading volume · Electronic trading

1 Introduction

Electronic trading in European equity markets has changed significantly in the

last 15 years due to regulatory initiatives. The Markets in Financial Instru-
ments Directive (MiFID I) came into effect in April 2004 and has been applied
by investment firms and Regulated Markets (RMs) in Europe since November
2007. A central goal of MiFID I was to enable investors to “trade securities at
maximum efficiency and at minimum cost” [10]. This goal should be achieved by
increased transparency and accessibility of markets, investor protection, market

The authors acknowledge financial support from Deutsche Boerse AG.

c Springer Nature Switzerland AG 2019
N. Mehandjiev and B. Saadouni (Eds.): FinanceCom 2018, LNBIP 345, pp. 3–26, 2019.
https://doi.org/10.1007/978-3-030-19037-8_1
4 P. Gomber et al.

integrity, harmonized European regulation and a level playing field among dif-
ferent types of trading venues to assure competition and to foster innovation.
Furthermore, MiFID I increased competition between trading venues through
the introduction of multilateral trading facilities (MTFs).
The competition between trading venues fostered by MiFID I led to a highly
fragmented equity market in Europe. As of June 2017, new competitors of the
incumbent national exchanges together achieved a market share of more than
28% of total European electronic order book trading in equities [17]. Industry
studies (e.g., [32]) show that the service and fee competition triggered by MiFID I
reduced explicit transaction costs both at trading venues and post-trade infras-
tructures.
Today, new requirements laid down in MiFID II and its accompanying regu-
lation Markets in Financial Instruments Regulation (MiFIR) seek to extend the
benefits that MiFID I generated for equity markets to other asset classes and
to address problems caused by market fragmentation, dark trading, and over-
the-counter (OTC) trading. One of the key objectives of MiFID II and MiFIR,
which have to be applied from January 3rd, 2018, is that OTC trades (unless
they fulfill certain criteria as discussed in detail in Sect. 3) are being forced to
take place on RMs, MTFs, or Systematic Internalisers (SIs) due to the so-called
trading obligation for shares. The goal of this study is to describe the context
of the new trading obligation for shares and to investigate its potential effect
on trading volume and liquidity on lit venues, i.e. transparent open order book
markets that play a crucial role in the price discovery process. Thereby, this
study should serve as a toolbox for market participants to assess the impact
of the new regulation based on their own individual estimates regarding likely
scenarios and the thresholds for migrating OTC volumes. Based on data and
experts’ estimates collected by a questionnaire, this study provides a prospec-
tive assessment of potential effects due to the introduction of the MiFIR trading
obligation.
In Sect. 2, we discuss the drivers and goals of MiFID II/MiFIR against the
background of changes in European equity trading due to MiFID I, technological
developments in trading in the last ten years, and lessons learned from the
financial crisis. Section 3 describes the concept of the MiFIR trading obligation
and the related industry discussions concerning its implementation in detail. In
Sect. 4, we explain the scenario methodology and develop the scenarios to assess
the potential impact of the trading obligation on market share distribution and
market liquidity in European equity trading. Section 5 describes our empirical
analysis and results regarding the impact of the trading obligation on trading
volume and liquidity. Section 6 discusses the results as well as limitations of this
study. Finally, Sect. 7 concludes this paper.

2 Analysis of Drivers and Goals of MiFID II/MiFIR

Against the background of the financial crisis, new trading technologies, new
products, and the necessary reviews already pre-defined in Article 65 of MiFID I
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 5

[14], the European Commission triggered a massive overhaul of European secu-

rities legislation. This initiative resulted in the publication of MiFID II and the
accompanying regulation MiFIR in June, 2014 [15,16]. Starting on January 3rd,
2018, the new rules had to be applied by investment firms, market operators,
data reporting service providers, and specific third-country firms. MiFID II and
MiFIR aim at (i) assuring a safer, sounder, more transparent, and more respon-
sible financial system, (ii) contributing to the delivery of the G20 commitment
to tackle less regulated and opaque parts of the financial system, (iii) improving
the organization and transparency of markets especially in those instruments
traded mostly OTC, (iv) improving the oversight and transparency of commod-
ity derivative markets, (v) taking account of new developments in market struc-
tures and technology, like dark trading, algorithmic trading, and high frequency
trading (HFT), and (vi) minimizing the discretions available to member states
[11].
The regulatory changes introduced by MiFID II/MiFIR also reflect the results
of various academic studies that analyze the impact of MiFID I on market quality.
A comprehensive overview of these studies is provided by [22]. While the majority
of studies finds that MiFID I indeed broke national monopolies of exchanges
thereby lowering transaction costs and also spurring new technical developments,
most studies find no positive effect of MiFID I on transparency and integrity of
European financial markets [22].
Although MiFID I changed the competitive landscape in lit trading consider-
ably, OTC trading still represents a high and very relevant market share in total
European equities trading contradicting one of the key objectives of MiFID I,
which is to increase trading transparency. Analyzing a large intraday data set of
OTC transactions from January 2008 to June 2013 covering EURO STOXX 50
constituents, [23] finds that OTC trading accounts for more than 50% of total
trading volume with increasing shares during the sample period. Furthermore,
their results reveal that the majority of OTC trades (58%) are retail-sized trades
while less than 5% are above the large-in-scale (LIS) threshold according to the
definition of MiFID I. The fraction of LIS-trades has decreased from more than
15% in 2008 to even less than 2% in 2013. In 2013, more than 95% of OTC
trades would not have had a market impact if they had been submitted to the
main market. Contrary to the MiFID I spirit and text, these results show that
OTC trades are not infrequent and large, but are rather frequent and small. In
addition, the increased emergence of dark pools, especially OTC dark pools or
so-called broker crossing networks (BCNs)1 , harm liquidity since they do not
disclose pre-trade information [9].
To achieve the goals and to address the weaknesses of MiFID I, several
adjustments and extensions have been implemented within MiFID II/MiFIR.
In particular, transparency goals were not completely achieved with MiFID I
since trading systems such as BCNs emerged, which were not captured by the

1
CESR, the predecessor of ESMA, defined Broker Crossing Networks as “internal
electronic matching systems operated by an investment firm that execute client
orders against other client orders or house account orders” [4].
6 P. Gomber et al.

regulatory regime, and a large fraction of trading is still conducted OTC. Conse-
quently, MiFID II and MiFIR include several new regulations aiming at enhanc-
ing transparency, which are described in the following Sect. 3.

3 MiFIR Trading Obligation and Related Regulatory

and Industry Debate

A key concept of MiFID II/MiFIR is to require all organized trading to take place
on organized venues and to ensure that trading systems are properly regulated
to assure a higher level of transparency (MiFIR, Recital 6). To achieve this goal,
the new market framework introduces four key concepts: (i) a new category of
trading venues called Organised Trading Facility (OTF) for non-equity instru-
ments to be traded on a multilateral platform, (ii) the so-called “double volume
cap regime” for equity trading, (iii) a new trading obligation for derivatives, and
(iv) a new trading obligation for shares. Moreover, transparency requirements
that only apply to shares according to MiFID I are extended to equity-like and
non-equity instruments establishing uniform requirements for the transparency
of transactions in financial markets (MiFIR, Recital 1).
The trading obligation for shares intends to restrict equity trading conducted
OTC to ensure that more trading takes place on regulated trading venues and on
platforms of SIs in order to increase transparency and to improve the quality of
the price discovery process as well as liquidity on lit markets (MiFIR, Recital 11).
It requires investment firms to undertake all trades in shares on an RM, an MTF,
an SI, or an equivalent third-country trading venue. Exemptions to the trading
obligation only apply if there are legitimate reasons2 . MiFIR, Recital 11 explicitly
states in this context that “[...] an exclusion from that trading obligation should
not be used to circumvent the restrictions introduced on the use of the reference
price waiver and the negotiated price waiver or to operate a broker crossing
network or other crossing system”. Therefore, BCNs, that are a relevant part of
OTC trading in equities today, cannot exist any longer in their former set-up
after January 3rd, 2018. These systems have to be authorized and operated as
MTFs or SIs (MiFIR, Recital 6).
The introduction of the trading obligation and the prohibition of BCN trig-
gered an intensive industry and regulatory debate concerning the level playing
field and delineation between bilateral (SI and OTC) and multilateral trading
(RM and MTF). This debate centered around the possibility for SIs to trans-
act via riskless back-to-back transactions both by executing customer orders
2
According to MiFIR Article 23 (1), trades can only be executed on an OTC basis if
they are non-systematic, ad-hoc, irregular and infrequent, or are carried out between
eligible and/or professional counterparties and do not contribute to the price discov-
ery process. Article 2 of Commission Delegated Regulation (EU) 2017/587 (RTS 1)
lists seven circumstances where trades do not contribute to the price discovery pro-
cess: vwap-twap trades, portfolio trades, hedges, transfers among fund portfolios,
give-ups/give-ins, collateral transfers, deliveries in case of exercises, securities financ-
ing transactions and buy-ins.
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 7

within their own SI and with other SIs or HFT Market Makers via so-called
SI-Networks. Exchanges and various regulators argued that this might create
unfair advantages for SIs relative to multilateral venues, reduce market trans-
parency and impede the efficiency of the price discovery process. After an inten-
sive correspondence between ESMA, the European Parliament and the European
Commission in early 2017, the European Commission proposed a Commission
Delegated Regulation [12] to clarify and amend the definition of SIs by adding a
new article 16(a) to Commission Delegated Regulation No 2017/565 that states:
“An investment firm shall not be considered to be dealing on own account [...]
where that investment firm participates in matching arrangements entered into
with entities outside its own group with the objective or consequence of carrying
out de facto riskless back-to-back transactions in a financial instrument outside
a trading venue”. Recital 2 of the Commission Delegated Regulation applies this
restriction to both internal and external matching of trades, thereby referring to
the intensively discussed networks of SIs.
This debate reveals that MiFID II/MIFIR re-launch a fight for market shares
between trading venues and SI/OTC trading on the one hand and among the
different types of trading venues on the other hand. Therefore, it is by no means
a given fact that the trading obligation and the prohibition of BCNs will result
in higher market shares of multilateral venues relative to bilateral means of
execution.
An intelligent interpretation and implementation of the new MiFID II/MIFIR
set-up by the financial industry might even result in higher shares of non-lit
and non-multilateral executions in European equity trading from January 2018
onwards. Specifically, the tick size regime in combination with best execution
requirements and the double volume caps within MiFID II/MiFIR might lead to
a shift of trading behavior and resulting order flows.
In order to curb competition among trading venues based on tick size,
MiFID II (Article 49) imposes the adoption of a (minimum) tick size regime
that is defined in prescriptive and detailed tick size tables of Commission Del-
egated Regulation No 2017/588 [13]. While Article 49 MiFID II requires RMs
to comply with the tick size regime, Article 18 (5) extends this requirement to
MTFs (and OTFs). However, the tick size regime does not extend to SIs and
OTC trading. Empirical research of [24] reveals that 29.3% of all SI trades and
24.8% of all OTC trades in a 2.5 years data set of European top liquid stocks
violate the existing voluntary tick size agreements [18] among exchanges and
MTFs in Europe. The vast majority of those tick size violations are trivial price
improvements over the quotes of public lit markets. The ability to grant those
trivial price improvements in the SI space and thereby still meeting the require-
ments of the trading obligation makes the SI regime very attractive. As it enables
to execute within the public spread and to free-ride on the public price discovery
process, it reduces the incentives of liquidity providers to provide competitive
quotes in lit markets. Furthermore, the price improvements, even if economically
very small, will likely prioritize SIs over lit markets in order routing decisions
driven by best execution requirements. In response to ongoing debates about
the minimum tick size exemption for SIs, ESMA opened a public consultation in
8 P. Gomber et al.

November 2017 to clarify that SIs’ quotes should reflect the tick sizes applicable
to EU trading venues.
The dark pool caps will also result in a change of trading behavior. Market
operators (e.g., Cboe Europe3 offering a Periodic Auctions Book model) as well
as buy and sell side firms (e.g., the Plato Partnership project that is linked to the
Turquoise Block Discovery service) launch new services and functionalities that
help to avoid a classification of orders under the negotiated trade and reference
price waivers and classify them as executions under either the auction model
(and therefore limited pre-trade transparency) or under the LIS waiver (that is
not included in the double volume cap mechanism). While this will increase the
relevance of block trading in the future European execution ecosystem, investors
will still try to identify ways to execute below LIS trades without pre-trade trans-
parency and market impact. Against the background of the BCN prohibition,
SIs offer an attractive alternative for customers in this respect as the SI trans-
parency is limited to standard-market sizes (which is 7,500 e for most European
shares), SIs fulfill the trading obligation requirements, and SIs enable customers
to receive price improvements in combination with lower explicit fees compared
to trading venues. The concept enables SI providers to keep order flows currently
traded in their BCNs while at the same time fulfilling the requirements of the
trading obligation (without the need to become an MTF). Moreover, they are
able to tailor quotes to specific clients based on risk considerations, to cream-
skim the order flow, i.e. to send informed order flow to public markets, and to
attract smart order routers based on small increments in price improvements
over the public spread. SI executions (e.g., at midpoint) also do not count under
the reference price waiver and are not included in the double volume caps. The
benefits of the SI regime might additionally attract current liquidity providers
in BCNs to register as SIs in an attempt to retain their business model.
Thereby, the new regulatory framework and the combined effect of the trad-
ing obligation, the tick size regime, and the double volume cap will significantly
influence European equities trading and will likely result in a redistribution of
market shares between multilateral trading venues, SI, and OTC trading.

4 Scenario Development
The scenario analysis is a qualitative forecasting technique that is useful for
strategic planning [3]. It is an effective methodology to deal with uncertainty
when forecasting complex developments or structures, which is regularly used
by academics and practitioners [34]. The scenario analysis represents an alterna-
tive to extrapolating past trends and relationships. In line with [29], we under-
stand scenarios as “a narrative description of a consistent set of factors which
define in a probabilistic sense alternative sets of future business conditions”. A
scenario analysis enables to capture a whole range of possible future outcomes
thereby revealing how different assumptions and uncertainties interact under
certain conditions [36]. Scenarios are suitable to capture new states after major
3
Formerly Bats Europe.
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 9

shocks, e.g., due to new regulation or technological changes. When conducting a

scenario analysis, it is essential to clearly define the scope, to identify trends as
well as key uncertainties, and to consider major stakeholders, who will be affected
by possible changes [36]. To derive realistic scenarios regarding the impact of the
MiFIR trading obligation on market share distribution and liquidity, we rely on
the one hand on historical trends and relationships and on the other hand, we
make additional assumptions to cover future developments.
We prospectively evaluate three different scenarios to assess potential migra-
tions of trading volume and related effects on liquidity due to the trading obli-
gation imposed by MiFIR Article 23. These three scenarios reflect different reac-
tions of relevant stakeholders of the financial industry to cope with the new
regulation from January 2018 onwards. This allows to assess the effect of the
trading obligation also with varying expectations in this regard. The three sce-
narios of our study are defined as follows:

– Scenario A (“Pro-Multilateral ”) assumes that OTC trades (including the

BCN share of OTC trading) will migrate to lit markets according to different
trade size categories and their probability of migration.
– Scenario B (“BCN Volumes Migrate to SIs”) assumes that the providers of
BCNs will manage to migrate the current BCN trading volumes into new
or existing SI setups. Therefore, in scenario B, we assume that OTC trades
(excluding the BCN share of OTC trading) will migrate to lit markets accord-
ing to different trade size categories and their probability of migration.
– Scenario C (“Internalization as in the US ”) assumes on the one hand that
new realizations and implementations of the SI concept will enable providers
of BCNs and other SIs to migrate the current BCN trading volumes into
new or existing SI setups. On the other hand, additional volume from lit
markets is expected to migrate into those SI setups, e.g., due to the possible
ability to price improve relative to the tick size regime and due to related
best execution obligations. Therefore, in scenario C, we assume that (i) OTC
trades (excluding the BCN share of OTC trading) will migrate to lit markets
according to different trade size categories and their probability of migration
and (ii) new SI set-ups will attract trading volume equivalent to the US
BCN share plus the US retail internalization level adjusted to the European
context.

In order to assess the likelihood of migration for different OTC trade size
categories, we conducted an online survey among industry experts on trading
and market structure. This survey was composed of 25 questions dealing with
different aspects of MiFID II/MiFIR that might have an impact on European
market structure. Access to the questionnaire started on November 29th, 2017
and closed on December 19th, 2017. This ensures that participants’ answers are
based on a prospective view and are not biased through first changes triggered by
the new rules. The questionnaire was distributed among 375 experts, whereof 111
experts submitted the online questionnaire before the deadline. This results in a
response rate of 29.6%. Among more general questions, the survey participants
10 P. Gomber et al.

were specifically asked to evaluate the scenarios described above and to estimate
the volume shifts of OTC volumes in different trade size categories to lit markets
due to the trading obligation for shares.
We assume that the probability of OTC volume migrating to lit markets
depends on the respective size of an OTC trade and 98% of the respondents
agreed to this assumption. The survey participants estimated that on average
39% of the OTC trades smaller than 500,000 e and 22% of the OTC trades
between 500,000 e and 50 mn e will migrate to lit markets. As OTC trades
larger than 50 mn e are very likely to account for portfolio transfers and other
non-trading related motivations, a vast majority of survey participants (98%)
expects that this kind of trades will remain OTC. Therefore, we assume that no
volume of this category will shift to lit markets.
Additionally, the participants were asked to determine which of the three
scenarios (A-C) appear to be most realistic. 17% of the survey participants
consider scenario A to be most realistic while 34% of the respondents believe
that scenario B is most realistic. With 43%, the largest group of industry experts
considers scenario C to be most realistic among the three different scenarios.
Only 5% of the respondents think that none of the scenarios described above is
realistic4 .

5 Empirical Analysis

5.1 Data Set

In the next steps, we determine the likely change in trading volume on lit mar-
kets due to the trading obligation for OTC trades respectively the BCN trading
prohibition for the three scenarios described above. Furthermore, we investi-
gate the liquidity effect due to these volume shifts in public limit order books
on the main market. For this analysis, we use Thomson Reuters Tick History
(TRTH) as the primary source of data. Our sample comprises constituents of the
EURO STOXX 50 as of January 2013. The sample period covers all trading days
between 1 January 2008 and 30 June 2013. The sample period is suitable for the
research question at hand since it ends one year before MiFID II came into force.
Thus, it excludes any possible announcement effects or changes in trading pro-
tocols to cope with the new regulatory environment such as new mechanisms to
handle the double-volume cap mechanism that were already introduced shortly
after MiFID II came into force in mid-2014 (e.g., the Volume Discovery Order,
Cboe Europe’s Periodic Auctions Book, or the Plato Partnership). Nevertheless,
market share distribution and overall trading volume in the last quarter of our
sample (2nd Quarter 2013), which serves as the reference for our scenario anal-
ysis, were similar to the same quarter in 2017 (2nd Quarter 2017) before the
application of MiFID II/MiFIR.
Our coverage of venues includes the main market for each stock of the EURO
STOXX 50 index. Therefore, we include data of the following main markets:
4
The complete analysis of the survey results can be found in [21].
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 11

Paris, Frankfurt, Amsterdam, Milan, Helsinki, Madrid, Brussels and Dublin. The
data contains trades on all these main markets as well as order book updates at
a millisecond precision. In addition to the main markets, we include all trades
from alternative venues such as Bats, Chi-X5 (now Cboe BXE and Cboe CXE)
and Turquoise, thus covering more than 95% of overall trading volume of these
stocks on lit markets. Besides the trades executed in open limit order books,
we include all off-exchange trades reported to Markit’s trade reporting platform
BOAT.

5.2 Volume Migration According to the Scenarios

Our empirical analysis comprises two consecutive steps. In step one, we set up
three different scenarios (as outlined in Sect. 4) and determine the likely increase
in trading volume on lit markets due to the trading obligation for OTC trades
and the BCN trading prohibition. As noted before, our data set features the
intraday turnover for each single trade on a single stock basis. This special
characteristic enables us to distinguish between different trade size classes. The
distinction of different trade sizes within our analysis is of high importance to
determine the expected turnover which will likely migrate from OTC to lit mar-
kets dependent on the probability to classify for the exemptions of the trading
obligation. Due to the difficulty to predict the exact amount of volume migration
as well as the actual change in fragmentation, we perform a scenario analysis.
For our subsequent analysis, we divide OTC trades and the corresponding
turnover in three disjoint classes representing different probabilities to migrate
to lit venues. The first class (OTC Small) represents the OTC trades below
LIS6 . The daily aggregated turnover in this category amounts to 17.70 mn e in
the second quarter of 2013 for an average EURO STOXX 50 constituent. The
second class (OTC Medium) includes the OTC trades between 500,000 e and
50 mn e, featuring a significantly lower probability to migrate. Trades within this
category amount to the highest turnover of the three classes representing a daily
average turnover of 101.15 mn e per instrument. The third class (OTC Large)
incorporates the remaining OTC trades (daily average turnover of 55.25 mn e
per instrument) with very high probability to remain OTC after the introduction
of the trading obligation.
Overall turnover in EURO STOXX 50 constituents amounts to 1,294 billion e
in the second quarter of 2013. Thereof, 48.90% were traded on lit markets, 2.64%
via SIs, 0.92% in regulated dark pools (i.e. dark pools provided by RMs or MTFs)
and 47.54% were conducted OTC (see Fig. 1). In our data set, BCN trades and
other OTC trades are not distinguishable as they are both flagged as OTC.

5
Cboe Europe operates both markets, Cboe BXE and Cboe CXE.
6
Under MiFID I, orders that are above the LIS threshold can benefit from a waiver of
pre-trade transparency. This waiver is intended to protect these orders from adverse
market impact and to avoid significant price movements that can cause market
distortion. For our sample of EURO STOXX 50 constituents, this LIS threshold is
500,000 e.
12 P. Gomber et al.

Since BCNs are prohibited under the new regulatory framework, it is necessary
for our analysis to determine the fraction of BCN trading volume. Therefore, we
rely on the market share of BCNs provided by Rosenblatt Securities, Inc. who
approximate that BCN turnover accounts for 4.55% of overall trading volume
in Europe.7 Consequently, the remaining market share of OTC trading without
BCN volumes amounts to 42.99%. Concerning lit venues, main markets account
for 70.14% of all lit trading volume, i.e. main markets (alternative lit venues)
have a market share of 34.29% (14.60%) of overall turnover in EURO STOXX 50
stocks.

Fig. 1. Turnover distribution in EURO STOXX 50 constituents as of Q2/2013.

Due to the trading obligation introduced by MiFIR, the turnover distribution

will likely change since BCNs are prohibited and since trades have to classify
for certain exemptions from the trading obligation in order to still be conducted
OTC. Moreover, we analyze possible changes in the European equity trading
landscape based on the scenarios outlined in Sect. 4. The three trade size cate-
gories necessary to determine migrating volumes are based on the whole OTC
category including BCN trades since they are not distinguishable in our sam-
ple. Therefore, we assume the trade sizes to be equally distributed in the pure
OTC as well as in the BCN category. The volume migration effects for the three
developed scenarios are depicted in Fig. 3 in the appendix.
In Scenario A, we assume that the OTC category as a whole including BCN
volumes migrates to lit venues according to the identified trade size categories

7
An overview of the Rosenblatt Securities, Inc. estimates for BCN market shares in
Europe over time is available at
https://www.bloomberg.com/news/articles/2017-07-10/dark-pool-traders-find-
mifid-workarounds-to-stay-in-the-shadows.
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 13

and their respective migration likelihood (as derived from the expert estima-
tions). On average, this results in a volume migration of 29.16 mn e per day
and instrument from OTC including BCNs to lit markets, which equals a mar-
ket share increase of lit markets amounting to 7.96 % points resulting in a mar-
ket share of 56.86%. The BCN volume that does not migrate to lit markets is
assumed to remain OTC and the SI market share to stay constant. The remaining
OTC turnover (including BCN) represents a market share of 39.58% compared
to 47.54% before the trading obligation.
Scenario B differs from scenario A in the sense that SIs are able to capture
the entire BCN volume so that only the OTC volume excluding the BCN volume
migrates to lit venues according to the identified trade size categories and their
respective migration likelihood. This results in a volume migration of 26.37 mn e
per day and instrument from OTC excluding BCNs to lit markets. The market
share of lit markets, therefore, increases by 7.20% points to 56.10%. Volume
traded via BCNs is completely captured by SIs leading to an increase of their
market share by 4.55% points to 7.19%. The remaining OTC turnover represents
a market share of 35.79%.
Different to scenario B, scenario C assumes that the market share of SIs rises
to a level equivalent to the US share of internalization. However, scenario C
assumes the identical volume to migrate from the OTC volume excluding the
BCN volumes to lit markets as scenario B. In addition, SIs (i) capture the entire
BCN volumes prior to the trading obligation and (ii) are able to attract turnover
from lit venues having advantages due to their possible independence from the
minimum tick size regime up to the derived internalization level comparable to
the US. The US volume traded in investment bank dark pools equivalent to
BCNs in June 2017 accounts for 8.79% of overall turnover [35]. At the same
time, the US retail internalization level is approximately 16.00% [8,35]. Due to
the fact that the percentage of equity holdings relative to total financial assets for
households in Europe is approximately half of the percentage of equity holdings
in the US8 , we derive a predicted market share of SIs = 8.79% + 51% × 16% =
16.95% for Europe. This scenario C is assessed to be the most realistic among
the participants of the survey and results in a net negative volume migration
effect for lit markets of −9.38 mn e per day and instrument, leading to a (net)
decrease by 2.56% points of lit markets’ market share to 46.33%. SIs completely
capture BCN volumes and additionally gain turnover from lit markets leading
to an increase of SI’s market share by 14.31% points to 16.95%. The remaining
OTC turnover represents a market share of 35.79%. The resulting market shares
of the different trading forms according to all three scenarios are reported in
Table 1.

8
Based on OECD data regarding household financial assets, equity holdings in the US
represent 34.96% of a household’s total financial assets while this number amounts
to 17.83% in Europe, which is 51% of the US figure. The European number is a
weighted mean considering only those European countries where corporations listed
in the EURO STOXX 50 are headquartered.
14 P. Gomber et al.

Table 1. Market shares (Each line adds up to 100%) of the different venues and forms
of trading before and after the introduction of the trading obligation.

Lit venues Dark pools SI OTC excl. BCN BCN

Q2/2013 48.90% 0.92% 2.64% 42.99% 4.55%
Scenario A 56.86% 0.92% 2.64% 39.58% -
Scenario B 56.10% 0.92% 7.19% 35.79% -
Scenario C 46.33% 0.92% 16.95% 35.79% -

Table 2 summarizes the resulting changes in trading volume and market share
of lit venues according to the three scenarios. It also reports the results for the
main markets and alternative venues separately. Therefore, the fraction of main
market turnover of overall lit turnover is assumed to remain constant at 70.14%
as in the second quarter of 2013.

Table 2. Volume migration and market share changes (average per day and instru-
ment) of lit markets in total as well as split according to main market and alternative
venues (assuming a constant fraction of main market turnover of overall lit turnover
amounting to 70.14%).

Volume Change in Volume Change in Volume Change in

Migration Lit Market Migration Main Migration Alterna-
to Lit Shares to Main Market to Alter- tive
venues [percentage Market Share native Venues’
[mn e] points] [mn e] [percent- Venues Market
age [mn e] Share
points] [percent-
age
points]
Scenario 29.16 7.96 20.45 5.58 8.71 2.38
A
Scenario 26.37 7.2 18.49 5.05 7.87 2.15
B
Scenario −9.38 −2.56 −6.58 −1.8 −2.8 −0.76
C

5.3 Measuring the Effect of Volume Migration on Liquidity

Having determined the potential volume migration effect due to the trading
obligation, this result is applied as an input to derive the impact of the new
regulation on liquidity. By quantifying the actual liquidity effect due to a volume
increase, we follow a similar approach as [7,27,31]. In line with [6], we follow
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 15

the reasoning that additional turnover increases liquidity. According to [2], it is

doubtful whether there exists a single measure that is able to capture all aspects
of liquidity. In general, liquidity turns out to be a multi-dimensional variable and
can therefore hardly be captured in a single one-dimensional measure. For this
reason, we investigate different liquidity indicators: the relative bid-ask spread
as the most common measure, Depth (10 bps) (see [9]) representing the euro
volume 10 bps around the midpoint, and XLM50k as the round trip costs for
trading 50,000 e (see [25]).
To determine the liquidity effect arising due to a volume increase in public
limit order books, we calculate the elasticity of additional trading volume by
applying an ordinary least squares regression on the above mentioned liquidity
measures. For our independent variables, we rely on turnover, the inverse of the
average stock price, market capitalization, the number of order book updates
(tick speed), market fragmentation, and volatility. The regression equation is
the following:
Liquidity measurei,t = α + β1 · Turnoveri,t + β2 · Price−1
i,t + β3 · Market Capitalizationi,t
9

+β4 · Tick Speedi,t + β5 · Fragmentationi,t + β6 · VolatilityM P
i,t + βk · Controlk + i,t ,
k=7

where i represents the respective constituent of the EURO STOXX 50 as of

June 2013 and t is a time variable for each quarter from the first quarter of
2008 until the second quarter of 2013. Controlk represents dummy variables for
stocks, markets, and quarters. i,t are normally distributed residuals.
To run the actual regression, we log-transform our dependent variables to
make the residual a relative instead of an absolute value and to avoid unneces-
sary heteroscedasticity. Furthermore, this transformation results in an increase
of the adjusted R2 by more than 53% on average. To additionally avoid poten-
tial problems related to autocorrelation patterns, we aggregated the intraday
observations to quarterly values. For each of the dependent variables, we run
two regressions, differing in the existence of control variables. In the following,
we explain why we are referring to these exogenous variables.
Since we investigate the impact of additional stock turnover on lit markets
because of the trading obligation introduced by MiFIR, we include Turnover
as the first independent variable in the regression model. Consistent with prior
evidence, we include the inverse of stock prices (P rice−1 ) as in [27]. For low-
priced stocks, the minimum tick size may constrain spreads. The inverse of the
stock price captures the tendency for this effect to diminish as stock prices rise
(see [20]). Moreover, we use the Market Capitalization of each firm to control for
firm size [5]. As analysts tend to follow larger firms more closely, these stocks may
feature higher trading volumes and (therefore) perhaps better liquidity indicators
[30]. The variable Tick Speed represents the number of order book updates and
can be seen as an approximation for order book activity as it measures the total
number of messages submitted to the order book. As shown in [9], fragmentation
has a significant impact on liquidity. To account for this effect, we rely on the
inverse of the Herfindahl Hirschman Index [28], which is in line with the Fidessa
16 P. Gomber et al.

Fragmentation Index [19]. Within our setup, Fragmentation takes into account
all lit venues and is determined on a quarterly basis for each stock separately. To
account for general uncertainty, we include the order book midpoint volatility
(V olatility M P ) as a measure of price deviation. In general, high volatility is
associated with low liquidity and vice versa. Due to a high multicollinearity to
the tick speed and turnover, we do not include the number of trades or a dummy
for years in our regression setup.

5.4 Description of Regression Results

In general, the relative spread and the XLM should show consistent signs of the
beta coefficients for all independent variables as visible in Table 3. In contrast to
the order book depth, a high relative spread and high values for XLM indicate
lower levels of liquidity, whereas higher values of order book depth indicate higher
liquidity levels. The adjusted R2 in the full models including control variables
(models 1, 3 and 5) ranges from 63% to 85%. Therefore, the explanatory variables
capture a high fraction of the overall variation in liquidity of European equity
markets.
Under all considered models, the beta coefficients of Turnover are consistent.
Each additional million of turnover decreases the spread by 0.038% (model 1)
and 0.26% (model 2), while round trip costs measured by XLM50k are reduced
by 0.14% (model 5) and 0.32% (model 6). Model 3 and 4 quantify the impact
of an additional million of turnover by an increase in depth of 0.3% and 0.47%.
Regarding the inverse of the stock price (P rice−1 ), our results almost consis-
tently show (except model 4) that stocks with low prices are less liquid, which
might (abschwächen, da nicht getestet) be explained by constraints due to min-
imum tick size requirements. Through higher analyst and media coverage of
stocks with high market capitalization, trading interest in these stocks might be
higher thus improving liquidity. However, the effect of Market Capitalization is
only significant for the order book depth and the relative spread in model 1.
The number of order book updates (Tick Speed ) is significant for the models 2,
4, and 6. By considering the full models, the significance of this effect vanishes.
Therefore, this effect seems to be captured by other variables within the full
model. By investigating the effect of Fragmentation, we obtain consistent results
(apart from model 1 and 5) and give additional empirical evidence that fragmen-
tation improves liquidity, which is in line with the studies presented in Sect. 2.
As V olatility M P is a measure of uncertainty, it is not surprising that high levels
of stock volatility are associated with lower liquidity since liquidity providers
face higher risks. Volatility as an explanatory variable is highly significant in
all models. The effect of each additional basis point of volatility increases the
relative spread by 5.25% (model 1) and 4.40% (model 2). XLM50k is expected
to increase by 7.42% (model 5) and 7.26% (model 6). The effect on order book
depth is even greater. For each additional basis point of volatility, the order book
depth will decrease by 10.95% (model 3) and even up to 11.84% in model 4.
 The MiFIR Trading Obligation: Impact on Trading Volume and Liquidity 17

Table 3. Regression results for the EURO STOXX 50 sample based on data for the
main markets. Endogenous variables are different liquidity measures. Exogenous vari-
ables are turnover, the inverse of the stock price, the market capitalization, the tick
speed of the order book, fragmentation, and volatility. The full models include addi-
tional controls for stock, market, and time specific effects. We apply robust standard
error estimations to correct for potential heteroscedasticity and autocorrelation biases.
Please note: *p < 0.1, **p < 0.05, ***p < 0.01.

Dependent variable
log(Spread) log(Depth10bps) log(XLM50k)
(1) (2) (3) (4) (5) (6)
Constant −8.651 -6.884 −1.542 −1.699 1.463 2.629
t= t= t= t= t= t=
−36.435*** −46.290*** −7.966*** −12.017*** 7.239*** 17.187***
Turnover [mn] −0.0004 −0.003 0.003 0.005 −0.001 −0.003
t = −1.239 t = t= t= t= t=
−10.525*** 11.144*** 15.788*** −4.532*** −12.737***
P rice−1 1.633 1.4 −0.97 0.287 2.002 1.183
t= t= t= t = 2.272** t = t=
4.791*** 7.857*** −3.180*** 5.312*** 6.048***
MarketCap [bn] −0.003 −0.0003 0.011 0.006 −0.002 0.0002
t= t = −0.269 t = t= t = −1.160 t = 0.178
−1.760* 5.330*** 6.658***
Tick Speed [k] 0.0003 0.002 0.0001 −0.001 0.0001 0.002
t = 1.109 t= t = 0.499 t= t = 0.377 t=
9.889*** −3.257*** 8.763***
Fragmentation 0.321 −0.187 0.533 0.705 0.149 −0.157
t= t= t= t= t = 2.281** t =
3.638*** −2.431** 8.339*** 12.237*** −2.086**
V olatility M P 524.91 440.217 −1,095.38 −1,183.61 742.487 726.289
t= t= t= t= t= t=
7.264*** 6.965*** −19.707*** −18.092*** 11.660*** 12.839***
Controls:
Stocks Yes Yes Yes
Marketplace Yes Yes Yes
Quarters Yes Yes Yes
Observations 1,034
R2 0.654 0.348 0.858 0.725 0.713 0.417
Adjusted R2 0.634 0.345 0.851 0.724 0.697 0.414
Residual Std. Error 0.417 (df = 0.559 (df = 0.331 (df = 0.45 (df = 0.392 (df = 0.546 (df =
978) 1027) 978) 1027) 978) 1027)
Max VIF 4.88 2.14 4.88 2.14 4.88 2.14
Mean VIF 2.37 1.65 2.37 1.65 2.37 1.65
18 P. Gomber et al.

5.5 Quantifying the Effect on Liquidity

In a final step, we combine the preceding analyses and estimate the liquidity
effects due to changes in trading volume. Since the regression analysis of turnover
and liquidity is based on data from the main markets, we calculate the liquidity
effects for the main lit markets. However, the liquidity effect for the alternative lit
markets should be of similar magnitude. For the analysis, we use the determined
volume migrations to main markets of 20.45 mn e (scenario A), 18.49 mn e
(scenario B) as well as −6.58 mn e (scenario C) as shown in Table 2 and combine
them with the beta coefficient of the full models (1, 3 and 5) representing the
liquidity sensitivity. Because the beta coefficients are point estimates and we
do not neglect their variability, we also compute the 95% confidence interval
represented by error bars in Fig. 2, which summarizes our results regarding the
effect of the trading obligation on main market liquidity.

Fig. 2. Effect of turnover variations on liquidity on the main market measured by

spread, Depth(10 bps), and XLM50k.

Assuming that the trading obligation changes the European market land-
scape according to scenario A and taking the full regression models as the model
of choice, the effect of additional turnover amounting to 20.45 mn e per stock
and day (see Table 2) will lower the relative spread on the main market by 0.78%.
The effect on XLM50k is slightly higher. Due to the additional turnover on the
main market, XLM50k will decrease by 2.82%. In terms of order book depth,
liquidity will increase by 6.32%.
The impact on liquidity measures in scenario B is slightly lower than the effect
within scenario A due to the assumption that BCNs will completely migrate to
SIs. The additional volume of 18.49 mn e per stock and day on the main market
results in a decrease of the relative spread of 0.71%. XLM50k reduces by 2.55%
and order book depth increases by 5.70%. Consequently, the trading obligation
Another random document with
no related content on Scribd:
spines helping to retain them in position on the back. It is said to be
the male that thus carries the eggs. This species is able to stridulate,
and when doing so vibrates its antennae with excessive rapidity. We
have only about a score of species of Coreidae in Britain, and none
of the remarkable forms of the family are among them.

Fam. 3. Berytidae.—Very slender Insects with the first joint of the

antennae and the femora thickened at the tips.—This small family
was not distinguished from Coreidae by the older authors. It consists
of about fifty species, eight of which are found in Britain.

Fam. 4. Lygaeidae.—The characters are the same as those

mentioned for Coreidae, except as regards the insertion of the
antennae; the upper surface or face of the head is not so flat, but is
transversely convex, so that seen in profile the antennae appear to
be inserted well down on the sides of the head.—The name
Infericornia was formerly applied to these Insects. They are on the
average of smaller size than the members of the Coreidae or
Pentatomidae, and are much less conspicuous in colour and form; a
good many of the larger Lygaeids are, however, variegate with black,
yellow, and red. The family is very numerous in species, about 1400
being known; they are arranged in thirteen sub-families; we have
about sixty species in Britain, nearly all small. Eremocoris lives,
when immature, in the nests of the wood-ant, according to
Wasmann. The family includes some notorious Insect-pests. The
Chinch-bug, Blissus leucopterus, commits very serious ravages on
corn and grasses in North America. The Cotton-stainer, Dysdercus
suturellus is also very injurious to cotton in certain parts of the New
World: its growth has been described by Riley,[485] who thinks a dye
valuable for commercial purposes might be procured from the Insect.
This bug has recently developed the habit of sucking oranges, and
has thus become injurious in Florida, as the fruit readily decays after
it has been punctured by these Insects. The phenomenon of
"micropterism" is exhibited by numerous Lygaeids, as well as by
Pyrrhocoridae.
Fam. 5. Pyrrhocoridae.—Distinguished from Lygaeidae only by the
absence of ocelli, and not recognised as a distinct family by all
Hemipterists. About 300 species are included. Our only British
member is the notorious Pyrrhocoris apterus; it is, however, very rare
in this country, though it abounds on the Continent, and has been the
object of investigation by embryologists and others. It displays in a
most marked manner the curious dimorphism as to the alar organs
that is so common in certain divisions of Hemiptera; the elytra and
wings being sometimes normally developed, while in other cases the
wings are entirely absent, and the horny, basal part of the elytra only
is present. In some localities, and in some years, only the
micropterous form is found, while on other occasions there may be a
large percentage of the macropterous form. The abundance of this
Insect has enabled the French chemist Physalix to obtain an amount
of its colouring matter sufficient for analysis; as the result he
procured a substance, insoluble in water, very closely allied to
carotine.[486] The Oriental Insect Lohita grandis is one of the most
remarkable of Bugs, the male of the Sumatran variety being over two
inches in length, having enormously long antennae, and the
abdomen extended to about twice the normal length, while the other
sex is in the usual condition in these respects. The species is said to
be injurious to the cotton-plant in India.

Fam. 6. Tingidae.—Tarsi two-jointed. Elytra more or less reticulate,

consisting of strong, irregular, thick lines forming a framework of
cells, the enclosed part of the cell being of different texture and
frequently transparent; antennae with terminal joint more or less
knob-like, the preceding joint very long; ocelli wanting; pronotum
prolonged behind, covering the scutellum; front coxae placed at the
lack of the thorax.—This is the first of a series of families with only
two joints to the feet. These little bugs are very remarkable objects,
and exhibit much variety in their peculiar sculpture, which in
numerous forms attains a condition of elegance well worthy of
attention. There are nearly 300 species known, and in Britain we
have about a score. The characters we have given above do not
apply to the genus Piesma, though it is usually placed in this family;
its scutellum is not covered, and ocelli are present. Although but little
is known as to the nature of the lives of Tingidae, yet it was pointed
out long ago by Réaumur that a species of the family (probably C.
clavicorne, Fig. 263), lives in deformations of the flowers of the
Labiate plant now called Teucrium chamaedrys; Frauenfeld has
more recently confirmed this observation, and shown that the closely
allied C. teucrii affects the flowers of T. montanum in a similar
manner.[487]

Fig. 263—Copium clavicorne. Europe. (After Rübsaamen.)

Fig. 264—Aradus orientalis. Siam.

Fam. 7. Aradidae.—Very flat, broad; scutellum exposed, large or

moderate; abdomen broader than the alar organs, which it frequently
encases like a broad frame. Front coxae placed in the middle of the
prosternum.—These very flat Insects, of obscure colour, have
frequently very peculiar sculpture. They live under bark, or on fungi
growing from bark, and are supposed to draw their nutriment from
the fungi, though but little is actually known as to their natural history.
The family is almost cosmopolitan, and includes about 300 species,
of which five occur in England. The small sub-family Isoderminae
consists of a few species that are placed only provisionally in
Aradidae; they differ from the normal members by there being no
groove on the breast, so that the rostrum is free. Of the five species,
three occur in Chili and Patagonia, two in Tasmania, and one in
Australia.
Fam. 8. Hebridae.—Minute bugs, of semiaquatic habits, clothed
beneath with a dense, minute, silvery pubescence; antennae five-
jointed; legs of not more than average length; elytra in larger part
membranous.—This small family consists altogether of only about a
dozen species; we have two species of the genus Hebrus in Britain;
they are usually found in very wet moss.

Fig. 265.—Halobates sobrinus. Under surface of a female carrying

eggs. Pacific Ocean (Marquesas).

Fam. 9. Hydrometridae.—Form very diverse; antennae four-jointed,

tarsi two-jointed. Coxae usually widely separated. Either wingless or
with elytra of one texture throughout, having no membranous part.
Under surface with a minute velvet-like pubescence. In many forms
the legs are of great length.—Although of comparatively small extent
—scarcely 200 species being at present known—this family is of
great interest from the habit possessed by its members of living on
the surface of water. In the case of the notorious genus Halobates
(Fig. 265) the Insects can even successfully defy the terrors of
Neptune and live on the ocean many hundreds of miles from land.
There is great variety of form among Hydrometridae. The European
and British genus Mesovelia is of short form, and but little dissimilar
from ordinary land-bugs, with which, indeed, it is connected by
means of the genus Hebrus, already noticed. Mesovelia represents
the sub-family Mesoveliides, which, though consisting of only four
species, occurs in both hemispheres, and in the tropics as well as in
the temperate regions. Our species, M. furcata, walks on the surface
of the water, the movements of its legs and the position of its coxae
being those of land-bugs. Another British Insect—the highly
remarkable Hydrometra stagnorum—is of excessively slender form,
with long thin legs, by aid of which it walks on the surface-film of
water, above which its body is held well separated. It is easily
drowned, and if submerged it has great difficulty in escaping from the
water. This genus represents the sub-family Hydrometrides, and is
apparently almost cosmopolitan. Velia currens is another common
British Insect; it loves the eddies and currents of backwaters on
burns and streams, and is very abundant in Scotland. An American
ally, Rhagovelia plumbea, appears to be not uncommon on the
surface of the ocean in the Gulf of Mexico, near the shores. The
great majority of the family belong to the division Gerrides, of which
the curious, long Insects that float so lazily and skim so easily on the
surface of quiet streams are typical. The species of the genus now
called Gerris, but formerly known as Hydrometra are apparently
distributed all over the world; we have ten in Britain. They have very
long legs, and on being alarmed move away with the greatest ease.

The genus Halobates includes at present fifteen species. They are

found on the ocean, where the surface-water is warm, in various
parts of the world. They are destitute of any trace of alar organs, the
meso- and meta-thorax are closely united and large, while the
abdomen is very small, so that the body is of oval form; the middle
legs are thrown so far back that they are placed immediately over
the posterior pair. When the sea is calm these Insects skim over the
surface with rapidity, but disappear as soon as it becomes agitated.
They are believed to feed on small animals recently deceased;
Witlaczil says on the juices of jelly-fish. The young are frequently met
with, and there can be no doubt that the whole life-cycle may be
passed through by the Insect far away from land. The Italian ship
Vettor Pisani met with a bird's feather floating on the ocean off the
Galapagos Islands, covered with eggs which proved to be those of
Halobates in an advanced stage of development. It was formerly
believed that the female carries the eggs for some time after their
exclusion, and although this has since been denied, it is
nevertheless an undoubted fact, for it was observed by Mr. J. J.
Walker,[488] to whom we are indebted for a specimen having the
eggs still attached to the body, as shown in Fig. 265. Mr. Walker
believes the bugs shelter themselves when the sea is at all rough by
keeping at a sufficient distance below the surface; they can dive with
facility, and are gregarious. They are frequently found close to the
shore, and Mr. Walker has even met with them on land. The stink-
glands of other Hemiptera are said by Nassonoff to be replaced in
Halobates by peculiar ventral glands. An allied genus, Halobatodes,
was supposed to be oceanic, but this is not the case, some of the
species having been found recently in fresh water in India, and
others in estuaries at Port Darwin. A remarkable allied form,
Hermatobates haddoni, was recently discovered by Professor
Haddon in Torres Straits. Apart from the oceanic life, Halobates is by
no means the most extraordinary of the Hydrometridae. The
Javanese Ptilomera laticaudata repeats some of its peculiarities, and
is of larger size, with the sexes very different. The most remarkable
of the family is perhaps the fresh-water genus Rheumatobates (Fig.
266), in which the males have peculiar prehensile antennae that look
like legs. These curious Insects inhabit North America and the West
Indies.

Fig. 266—Rheumatobates bergrothi. × 10. West Indies. (After Meinert.)

We may here notice an enigmatic Insect called Hemidiptera haeckeli

by Léon. From the single specimen known it is concluded that the
Insect has only one pair of wings, and that they are attached to the
metathorax. It is, however, possible, as suggested by Bergroth,[489]
that the anterior pair have been detached by some accident.

Fam. 10. Henicocephalidae.—Head swollen behind the eyes so as

to form a sort of globe, on the anterior part of which the ocelli are
placed. Rostrum extremely short. Elytra rather large, of one
consistence throughout; conspicuously veined.—There is only one
genus; it is very widely distributed, about a dozen species being
known; one of these occurs in the south of Europe. These curious
little bugs appear to be most nearly allied to the Reduviidae.
According to Westwood and others they are somewhat gregarious; a
Tasmanian species dances in the air after the fashion of midges or
May-flies, and dispenses an agreeable, musk-like odour.

Fig. 267—Carcinocoris binghami (Phymatidae). Burma.

Fam. 11. Phymatidae.—Front legs of peculiar structure, short and

stout, with long coxae, short thick femora, and tibiae curvate,
pointed; frequently without tarsi.—The Insects of this family are
believed to be predaceous, the structure of the legs being such as is
called raptorial, and one species, Phymata erosa, being known to
capture and suck honey-bees in North America. There are only
about seventy species of Phymatidae known. We have none in
Britain, though there are a few in Southern Europe; one of these, P.
crassipes, extends as far north as Paris. The distinction of the family
from Reduviidae is doubtful.[490] There are a few very rare forms
(Fig. 267) in which the front tibia is articulated to the femur in such a
way that a pair of pincers is formed: the tarsus is in this form, as well
as in some other Phymatidae, absent.

Fig. 268—Ghilianella filiventris. Brazil. A, the female Insect. B,

extremity of the body of the male.

Fam. 12. Reduviidae.—Head more or less elongate, very movable,

eyes placed much in front of the thorax, ocelli, when present, behind
the eyes. Proboscis short, or moderately short, not extending on to
the breast, in repose curved under the head so as to form a loop
therewith. Elytra, when present, consisting of three divisions. Tarsi
three-jointed.—This is one of the largest and most important families
of Hemiptera. Upwards of 2000 species are already known; the
habits seem to be chiefly of a predaceous nature, the creatures
drawing their nutriment from the animal rather than from the
vegetable kingdom, and their chief prey being in all probability other
kinds of Insects. There is, perhaps, no family of Insects exhibiting a
greater variety of form and colour. The Emesids are amongst the
most delicate of Insects, equalling in this respect the daddy-long-leg
flies; they are, however, highly predaceous; their front legs are
peculiarly formed for capturing and holding their prey, and have long
coxae, like Mantis, so that these Insects are commonly mistaken for
small or young Mantises, from which their sucking proboscis at once
distinguishes them. This curious starved-looking form of bug reaches
its maximum of peculiarity in the South American genus Ghilianella
(Fig. 268). According to Pascoe the linear form enables the young
larva to be carried about by the mother, the long slender abdomen of
the larva being curled around the thorax of the parent. Ploiaria
pallida, from Woodlark Island, is an Insect of excessive fragility and
elegance, with the long thin legs coloured with alternate patches of
black on a white ground, giving rise to a very curious appearance
remarkably analogous to what we find in some of the equally delicate
daddy-long-leg flies.

Fig. 269—Nabis lativentris, young. Cambridge. A, Insect seen from

above; B, profile.

We have three species of Emesides in Britain, but most of our

Reduviidae belong to the sub-family Nabides. These approximate to
ordinary bugs in appearance and characters more than do any other
of the Reduviidae. One of our indigenous Nabides is of great interest
from the curious resemblance it has to an ant (Fig. 269). The
likeness is brought about by the sides of the base of the abdomen
being very pallid in colour, except a dark mark in the middle; this
mark is in shape like the pedicel of an ant. Viewed in profile it is
found that on the base of the abdomen there is an elevation like the
"scale" in this position in ants, and that the abdomen is extremely
ant-like in form. This resemblance is quite parallel with that of an
Orthopteron to an ant (see Vol. V. p. 323); the Insect is by no means
uncommon, and it is strange that this curious case of resemblance
should hitherto have escaped notice. The bug runs about on plants
and flowers, and is frequently in company with ants, but we do not
know whether it preys on them. Not the least remarkable of the facts
connected with this Insect is that the resemblance is confined to the
earlier instars; the adult bug not being like an ant. We may here
mention that there are numerous bugs that closely resemble ants,
and that on the whole there is reason to believe that the resembling
forms are actually associated during life, though we really know very
little as to this last point.

Fig. 270—Ptilocnemus sidnicus. Australia. (After Mayr.)

Fig. 271—Myiodocha tipulina. China.

The little sub-family Holoptilides, with twenty-five species, but widely

distributed in the Eastern hemisphere, is remarkable on account of
the feathered antennae and legs of its members (Fig. 270).
Altogether fourteen sub-families are recognised, the most extensive
one being Harpactorides, including a great variety of remarkable
forms; in the South American genus Notocyrtus (better known as
Saccoderes, Fig. 257), the prothorax is swollen and covers the body
to a greater or less extent in the fashion of a hood. In Yolinus and
Eulyes the coloration is the most conspicuous system that could be
devised, the sides of the abdomen (connexivum) being expanded
into bright-red lobes on which are placed patches of polished-black.
The most remarkable form of Reduviid is, perhaps, one from China
(Fig. 271) of considerable size, of great fragility, and greatly
resembling, like some Emesides, a daddy-long-legs fly, though it
does not belong to the Emesides. It is an altogether anomalous form.
According to Seitz there is found on the Corcovado in Brazil a
Reduviid that exactly resembles one of the dark stinging-wasps of
the genus Pepsis, and the bug makes the same sort of movements
as the wasp does, though these are of a kind quite different from
those of ordinary bugs.[491]

Fig. 272—Eggs of Endochus cingalensis. "The eggs are attached to a

leaf and to each other by a viscid substance; eggs red, the cover
pale yellow, with the club white at the tip."—MS. note of E. E.
Green.

Although the attacks of Reduviidae on animals are usually confined

to the smaller and more defenceless kinds, yet this is by no means
invariably the case; there are in fact numerous species that do not
hesitate to attack man himself. Several species of Reduvius do this
in Southern Europe, and are frequently met with in houses. R.
personatus is the only species of the genus in England; though far
from common anywhere, it is sometimes found in houses, and is
said to destroy the common bed-bug; it is able to pass its whole
existence in our habitations, for the young are found as frequently as
the adult, and are usually concealed by a quantity of dusty matter, or
refuse, adhering to the body. This habit of covering the body with
some foreign substance is natural to the Insect, the young that are
found on trees being covered with matter derived therefrom. Darwin
has given us an account of the Benchucha,[492] a bug an inch long,
which in South America attacks human beings after the fashion of
the common bed-bug. In this case no ill-effects follow the attack, but
in the case of Conorhinus sanguisuga in Arizona, great pain and
inflammation ensue and may end in the gathering and discharge of
pus.

Not the least remarkable of characters of Reduviidae is the form of

the eggs of some of the species (Fig. 272, and Vol. V. Fig. 78, C);
the egg bearing a peculiar operculum, the purpose of which is at
present quite mysterious.

Fam. 13. Aëpophilidae.—A single species forms this family. It is of

considerable interest, as it is incapable of flight, passing a large part
of its life covered by the sea. Aëpophilus bonnairei is a small Insect
with quite short head, without ocelli, and with the organs of flight
represented by a pair of very short elytra, with rounded hind-
margins. It is found on the shores of Western France, and, as a great
rarity, on our own south coast. It no doubt sucks small soft animals.
In the Channel Islands it occurs in spots where it is nearly always
covered by a considerable depth of water.

Fam. 14. Ceratocombidae.—Minute bugs with ocelli and elytra.

Rostrum free. Head not broad, somewhat prolonged in front; eyes
close to the thorax. Elytra usually without a distinctly separated
membrane. Tarsi three-jointed.—This family includes at present only
a few, minute, fragile bugs, that have often been classified with
Cimicidae or Anthocoridae. We have only two British species, one of
which, Dipsocoris alienus, is common amongst the damp shingle at
the margins of the burns and waters of Scotland.

Fam. 15. Cimicidae.—Ocelli absent; elytra very short and broad, so

that the broad abdomen is left uncovered. Head short and broad.
Rostrum received in a groove beneath the head. Tarsi three-jointed.
—Although this family consists of only a dozen species, it is the most
notorious of all the Order, as it includes the detestable Cimex
lectularius or common Bed-bug. This Insect is now peculiar to the
habitations of man, and is said not to trouble savage races; or rather
it is supposed to be present only when the habitations have a certain
degree of comfort and permanence. It has no fixed period of the year
for its development, but the generations succeed one another so
long as the temperature is sufficiently elevated; during too cold
weather the Insects merely become stupefied, their lives being as it
were interrupted till warmth returns. It is a favourite food with other
Insects, and is destroyed by cockroaches and ants as well as by
Reduvius; the small black ant Monomorium will, it is said, clear a
house of the bed-bug in a few days. Nothing is really known as to the
origin of this Insect; it is now very widely distributed. The other
species of the family frequent birds and bats, and are very similar to
the common bug. The genus to which the bed-bug belongs is in
many works called Acanthia instead of Cimex. Other authors apply
the term Acanthia to Salda, but it is better to allow the name
Acanthia to fall into disuse.

Fam. 16. Anthocoridae.—Minute bugs, usually with ocelli and with

elytra; the latter occasionally abbreviated, but usually fully
developed, with membranous tip. Head prolonged in the middle in
front much beyond the insertion of the antennae; eyes not far from
the thorax. Rostrum free.—These small and obscure Insects appear
to be rather numerous in species, and to be chiefly connected with
woods and forests. Some of the species live in ants' nests. We have
27 British species belonging to 11 genera. About 200 species of the
family are known. The members of the sub-family Microphysides are
remarkable from the great dissimilarity of the sexes, for which it is
not possible to assign any reason.
 Fig. 273—Polyctenes fumarius. (After Westwood.)

Fam. 17. Polyctenidae.—Proboscis-sheath three-jointed, tarsi four-

jointed, antennae four-jointed. Tegmina quite short, of one
consistence.—The four or five anomalous species forming this family
are parasites on bats of the genus Molossus, and have been found
in both the Eastern and Western hemispheres. Westwood, who first
described them,[493] treated them as aberrant Anoplura or Lice, but
there do not appear to be any sufficient grounds for removing these
parasites from Hemiptera-Heteroptera. The condition of their alar
organs reminds one of what exists in Cimex and Aëpophilus, and the
mouth is not known to possess any very peculiar structure. We have
had no opportunity of making a thorough examination of Polyctenes,
and therefore speak with some diffidence.

Fig. 274—Helopeltis sp. East India.

Fig. 275—Section of a stem with egg of a Capsid bug allied to

Helopeltis (Moesa-blight). × 58. (After Dudgeon.)

Fam. 18. Capsidae.—Moderate-sized or small bugs, of delicate

consistence, without ocelli; the elytra and wings usually large in
proportion to the body, the former with two cells (occasionally only
one) in the membrane. Antennae four-jointed, the second joint
usually very long, the terminal two more slender than the others. The
proboscis not received in a groove. Scutellum exposed, moderately
large. Tarsi three-jointed. Female with an ovipositor capable of
exsertion.—This family is one of the most extensive of the
Hemiptera; we have about 170 species in Britain, where they are
most abundant in the south. The exotic species have been but little
collected. Their colours are usually delicate rather than vivid, and are
never metallic. They frequent plants of all kinds, and many of them
skip by the aid of their wings with great agility in the sunshine. The
majority probably suck the juices of the plants, but some are known
to prey on other Insects. The species of the Indian genus Helopeltis
(Fig. 274) are remarkable by possessing a knobbed spine projecting
straight up from the scutellum, making the individual look as if it were
a specimen with a pin through it: they attack the tea-plant and do
considerable damage. They are known as Mosquito-blight. The egg
is of comparatively large size, and is placed by the bug in the stems
of the tea-plant, but attached to one end of the egg are two long
slender threads that project externally. A similar egg (Fig. 275) and
method of oviposition have been described by Mr. Dudgeon as
occurring in another species of Capsidae, called Moesa-blight, in
India.[494]

Fam. 19. Saldidae.—Head short and broad, with large, prominent

eyes. Ocelli present. Proboscis not applied to under surface of head
or breast in repose. Scutellum large, not covered. Elytra covering the
upper surface of the abdomen, formed of three distinct parts. Tarsi
three-jointed.—These little bugs run with velocity over mud in damp
places, or live in wet moss; some of them can jump; they are all of
dark or obscure colour. There are only three genera: Salda, of which
we have numerous British species, being the principal one.

Series 2. Cryptocerata.
The remaining families of Heteroptera are of aquatic habits, and form
in nearly all works a separate division called Hemiptera Cryptocerata
(or Hydrocorisae, or Hydrocores), distinguished by the antennae
being apparently absent; they are, however, really present, being
situate on the under side of the head, to which they are closely
pressed, or in some cases placed in a pocket in front of each eye.
There are six of these families. Schiödte is doubtless correct in
treating this division as an unnatural one; it is, however, generally
adopted, and is convenient for the purposes of nomenclature and
arrangement.

Fam. 20. Galgulidae or Pelogonidae.—Form short and broad; head

very broad, with prominent eyes, ocelli present. Hind legs thin,
formed for running.—The Insects of this family are but little known;
they are only sub-aquatic in habits, frequenting damp places at the
margins of streams and waters. The presence of ocelli distinguishes
them from other water-bugs, with which indeed the Galgulidae
appear to be but little related. There are only about twenty species of
the family known. We possess none in Britain; but one, Pelogonus
marginatus, occurs in South Europe. The other members of the
family are very widely scattered over the surface of the earth.

Fam. 21. Nepidae.—Abdomen furnished behind with a long slender

siphon; front legs more or less elongate for capturing prey, placed
quite at the front edge of the prothorax.—This family consists of two
interesting but very dissimilar genera, Nepa and Ranatra. Both are
widely distributed over the earth, and are rather numerous in
species.[495] We have one species of each genus in Britain. Nepa
cinerea, the common "water-scorpion," is one of the commonest of
Insects in Southern Britain, living concealed in shallow waters when
nearly or quite stagnant. Ranatra linearis (Fig. 276) is much less
common, and appears to be getting rarer; it is not recorded from
farther north than Cambridge.
 Fig. 276.—Ranatra linearis, with the two portions, a, of the respiratory
siphon separated. Cambridge.

The nature of the respiratory arrangements in these Insects is of

considerable interest; the long tube at the extremity of the body
consists of two parts (as shown in Fig. 276) brought together in the
middle, one from each side. Lacaze-Duthiers states that the
processes are elongated pleurae, but in the young it is far from clear
that this is the case. However that may be, they seem to convey air
to the true breathing organs, situate inside the cleft on the apical part
of the abdomen itself; but details as to the way in which transfer of
air is effected along this very protracted passage are not
forthcoming. The development in Nepa has been studied to a certain
extent. The apical stigmata are the only pair of the abdominal
stigmata that exist in the imago of Nepa, the other six pairs being
obliterated; the third, fourth, and fifth, according to Schiödte, in a
very peculiar manner: hence, as Martin says,[496] the respiratory
system is metapneustic. In an earlier stage of the life, however,
these six pairs of stigmata exist in functional activity placed in a
groove on the under surface of the body; so that the condition is that
termed peripneustic, and remains so till the final moult, when the
long siphon appears. In the early life there is a short prolongation
from the end of the body in connection with the pair of grooves
alluded to, but it is a single unpaired organ, and does little therefore
to explain the appearance of the siphon, which must, at present, be
considered as being suddenly developed at the last moult.

Fig. 277—Egg of Nepa cinerea. (After Korschelt.)

The eggs of Nepidae are remarkable objects; that of the common

water-scorpion bears seven filaments at one end (Fig. 277); while
that of Ranatra is more elongate, and bears only two, very elongate,
threads. These eggs are deposited in the stems of water-plants,
being introduced therein, so that the body of the egg is concealed
while the threads project: those of Ranatra are placed in stems
floating on the water, and in consequence of the threads the stems
look as if they were infested by some fungus. The structure and
formation of the eggs have been investigated with considerable
detail by Korschelt.[497] He looks on the filaments as pneumatic, and
considers that they supply a coating of air to the body of the egg;
they consist of a spongy mass encircled by two layers of egg-shell,
both of these latter being peculiar in structure; the spongy mass is
continuous with a layer of the same kind of substance placed on the
interior of the shell of the body of the egg. It will be recollected that
we have described (p. 562) an egg, apparently of the same nature,
deposited by Capsids in the stems of land plants, so that it is very
doubtful whether the threads are really connected with the aquatic
development of the embryo in Nepidae. But the most interesting
feature connected with these eggs is, according to Korschelt, the
mode of development of the filaments, which is sui generis; the shell
of the egg is developed in the ordinary manner as an exudation or
excretion from epithelial cells; but the shell of the filament is formed
as an intracellular product; a mode of chitin-formation that appears to
be peculiar to this structure. Korschelt remarks that "it is in the
highest degree worthy of attention how by any process of
development through a large number of successive generations so
complex a condition could be established as the result of adaptation
to external conditions; and this becomes even more interesting when
we remember that highly peculiar special processes and departures
from the usual modes of tissue-formation are necessary to permit the
development of this apparatus."[498]

Fam. 22. Naucoridae.—No ocelli, and no terminal process to the

body; front legs inserted on or near the front of the prosternum.
Anterior femora usually broad and flat.—The members of this family
are truly aquatic, and swim readily in the water. The family is small,
including about nine genera and thirty species, but, like many water-
Insects, the genera are widely distributed. We have two in Britain—
one of them, Naucoris, common; the other, Aphelocheirus, rare.

Fam. 23. Belostomidae.—No ocelli, and no long terminal tube to

the body; front legs inserted near the front of the prosternum.
Posterior tibiae not spiny; flattened and provided with swimming
hairs.—Although these Insects have been classified with Nepidae
they have but little relation therewith; on the other hand, the
distinctions from Naucoridae are far less important. The family
includes some of the largest Insects. The South American
Belostoma grande attains a length of four or four and a half inches.
Notwithstanding their considerable size Belostomidae exist in very
large numbers in some localities, and frequently destroy young fish
by aid of the powerful though short rostrum.

Fig. 278—Zaitha anura, carrying eggs on its back. West Indies.

 Fig. 279—Antenna of Belostoma sp. A, One side of the under surface
of the head, with antenna, b, extended; B, with the antenna
retracted, a, Side of head; c, pocket for antenna; d, position of the
eyes. The corresponding joints of the antenna are numbered 1, 2,
3, 4 in each figure.

They appear to be unable to resist the attraction of artificial light, and

are consequently sometimes destroyed in large numbers. It has long
been known that species of the genera Diplonychus and Zaitha carry
their eggs on their backs. There is no special receptacle for the
purpose, but the eggs are kept in their peculiar position by means of
a cement insoluble in water. It has been stated by Dimmock that they
are placed in position by means of a long, flexible ovipositor.
Schmidt, however, found that a specimen of Diplonychus, bearing
eggs and examined by him, was a male, and he subsequently found
that this was the case with other egg-bearing individuals of other
species, so that the mode in which the eggs are placed in this
position and the object of so curious a habit, remain uncertain. The
species of Belostoma are highly remarkable on account of the
curious and complex structure of their antennae, in respect of which
the nearest analogy is to be found in the large Coleoptera of the
genus Hydrophilus. A very deep, ear-like pocket, exactly suited to
the form of the antennae, exists on the under side of the head;
hence in repose no sign of the peculiar shape of the antennae exists.
When the antennae are placed in this ear-like pocket only the one
side of the basal joints is exposed, the long processes being
received into the deep pocket. In Hydrophilus the antenna is used as
an accessory organ of respiration, and it will be interesting to learn
whether this is also the case in Belostoma. Belostomidae have
patches of air-carrying pubescence, analogous with those of
Hydrophilus, on the under sides of the body, elytra and wings, but we
do not know how they are charged. Another extremely interesting
analogy is found in the manner in which the elytra are locked to the