All the above mechanisms have in common a process of search for

A Survey of the Microstructure of Block Trading
in European Equity Markets
Chwen Chwen Chen*
Istituto di Finanza,
Università della Svizzera Italiana
May 2003
Abstract
This paper provides a survey of the microstructure of upstairs markets of the largest
five European stock exchanges. The explosive growth of institutional trades is the
main factor which leads exchanges to compete for optimal trading facilities for this
remarkable source of market order flow. We account for institutional similarities and
dissimilarities as a key determinant of regulatory arbitrage across European
exchanges and as one of factors that give rise to varying growth of institutional
trading on individual European exchanges. The final view is a very wide institutional
fragmentation, in particular with respect to the enforcement of interaction rules and
transparency.
* E-mail: [e-mail]
1 Introduction
In latest years institutional trading has become a consistent proportion of market
order ‡ow in terms of a steady growth both in the trading volume and
in the average size of orders. For instance, last year J.P.Morgan reported it
executed a $1.9 billion trade; other trades too were reported to involve more
than $1 billion.1 These …gures are expected to sharply increase in coming years.
Given this trend, a separate trading arrangement from the normal order ‡ow of
investors will always be needed, otherwise institutional trading would execute
in alternative trading venues. What will change is then the de…nition of large
orders (”block trades”); in other words, the size of trades de…ned as block trades
will increase.
Whereas in the US and UK markets institutional trading is quite consolidated,
most continental European exchanges are becoming very pro…table
markets to institutional investors mainly because of a relatively weaker public
demand; this in turn is re‡ected in a poorer o¤er of trading facilities for institutional
investors. Most important, at current state, despite the monetary
harmonization, there exists institutional variety in the handling of large trades.
This regulatory fragmentation leads to strong regulatory competition across exchanges
themselves and continously reshapes the European market structures.
From the supply side’s point of view, that is from the public exchanges’
point of view, the main issue concerning institutional trading is to reduce negative
externalities of large transactions on the central market (or downstairs) by
ensuring large trades adequate trading facilities. As being of size larger than
the normal market order ‡ow, the impact of these trades on the central market
may be considerable so as to drain liquidity from the public electronic order
book and create temporary liquidity shortage in the downstairs market and disorderly
price movementes as well. A typical situation is that, while downstairs’
investors e¤ort to search for a counterpart in the transaction of a stock, at the
same time a trade of larger size on the stock is being executed on the block
market. Market prices would consequently lose their signalling function. This
e¤ect would be ampli…ed in case large trades are even information-motivated,
with the consequence to distort the ”true” price of assets. Exchanges are therefore
called for the design of rules to avoid trade-throughs and ensure gains to
all traders on the marketplace.
Another issue to account for is that institutional trades represent a considerable
source of pro…ts for trading systems providers. Block traders may work the
block as whole at once or break block trades into smaller ones and work them
over several days to obtain the best overall price for the customer. With the
explosive development of information technology over past years, electronictrading
algorithms are actually replacing some of brokers’ activity; trades of
100,000 (or even a million) shares may soon be executed electronically.2 This
1 See Davidson et al. (2001).
2The size of trades that can be executed electronically is increasing thanks to innovations
such as ”volume-weighted average price” (VWAP) trading, which allows institutions to execute
a trade at the day’s average price, and time slicing, which breaks up a block and executes it
1
is already a reality in the U.S. and U.K. markets not only for equity markets.
In face of the competition from alternative trading systems or Electronic Network
Communications (ECNs), European public exchanges are chanllanged in
o¤ering more competitive trading environments.
The objective of this survey is to document the trading mechanisms for large
transactions in the largest …ve European equity markets. Our focus is therefore
on the supply side of block trading, that is on the distinctive set of rules that
govern the trading process for institutional trades. This set of rules is dissimilar
across European exchanges, although there exist overall institutional common
features. Among other factors, it may be that such institutional similarities
and dissimilarities are a key determinant of the varying growth of institutional
trading at individual European exchanges giving rise to regulatory arbitrage
across European markets. A relevant case is the enforcement of interaction
rules in the Paris Bourse in 1989, which had to be relaxed years later because
of institutional competition from the London stock exchange. The analysis of
the institutional framework is important because it is this set of trading rules
that in the end a¤ects the pro…tability of various trading strategies, and hence
a¤ects the cost of trading and the …nal allocation of …nancial resource.
Typically block transactions are conducted in o¤-exchange markets called
upstairs markets. These are not the only public trading venue for institutional
trading, though. Because of the growing importance of this source of market
order ‡ow and the e¤ort of order ‡ow consolidation in the central market, exchanges
too o¤er trading facilities to incentivate institutional investors to trade
in the central public order book. What is striking is that there are much common
e¤orts and views across national exchanges as for the integration of institutional
trading facilities into the central market rather than the search for common European
trading institutional requirements for markets for block transaction.
This paper is organized as follows. In the next section we present a statistical
comparative overview of block trading and electronic book order trading and
account for the state of development of institutional trading in Europe. Section
3 reviews the trading mechanisms for blocks in the major European stock
exchanges considers the possible changes. Section 4 presents the main common
and di¤erent institutional features. Section 5 concludes.
2 Institutional trading: the demand side of block
trading facilities
The trading of large amount (above 10,000 shares) accounts for roughly 50% of
total trading volume at the NYSE.3 Almost 67% of block trading volume is exover
a certain time interval.
3 Block trading on the NYSE amounted to 44.4% of total NYSE volume in 2002, a decline
from 48.1% the prior year. This proportion was 51.7% in 2000. This decline is mainly due to
the recent world economic conditions.
2
ecuted in the upstairs market at the Paris Bourse (Besseminder and Venkataraman,
2002). Table 1 and table 2 show the amount of block trades with respect to
downstairs trading volume in domestic and foreign equities on the major European
exchanges. Overall, we observe that trading in the central electronic order
book prevails in continental European exchanges than in the London Stock Exchange.
O¤-exchange trading is however consistent, and the growth has been
more remarkable in the Italian Stock Exchange. In any case, London remains
the most attractive center for institutional trading especially of foreign securities.
This may be largely due to more ‡exible rules enforced for block trading in
the London Stock Exchange than those in the continental European exchanges.
The main determinant for the increase of block trades is the sharp growth of
institutional trading in Europe, a process which is also favoured by the advent
of the single currency. Only in period 1990-1998, the value of assets managed
by institutional investors grew at an annual rate of 14%, and it is expected to
increase in coming years. The total value of funds under management in Europe
- including the United Kingdom - is over $10 billion, of which the United
Kingdom accounts for 40% (Table 3). This trend has been however more significant
in most continental European markets than in the U.S. and U.K. because
the former are still contestable markets for domestic institutions and foreign
intermediaries and therefore ensure excess pro…tability.
The main forces driving the steady expansion of institutional trading in Europe
include the trend toward professional management of household assets,
the shift of portfolios into equities and foreign assets, the move from de…ned
bene…t to de…ned contribution pensions in private sectors. Above all, the most
crucial factor is the aging of population and the recognition that pay-as-you-go
pension systems are unsustainable as they stand. Although these factors are
present elsewhere, European countries are among those that have the greatest
scope for growth in this respect having less well developed mutual funds sectors,
ongoing pay-as-you-go pension systems, and little funding except in the
United Kingdom, Ireland, Denmark, the Netherlands, Sweden and Switzerland.
More generally, owing to the dominance of pay-as-you-go pensions, scope for
expansion is arguably even greater than in the relatively mature markets of the
United States and the United Kingdom, where pension systems already have
major funded elements. Institutional saving is also likely to increase sharply
over the next 20 years as individuals seek for their retirements.
In the light of this trend, it is clear that European exchanges face a great
challenge to cope with the increasing demand of trading facilities from institutions.
Exchanges have moved along two directions. On one hand, they arrange
for a separate trading venue for the handling of block trades. Several surveys
show that institutions and their brokers will prefer to execute their orders away
from the public order book to prevent excessive front-running or mitigate price
impact. They are also willing to sacri…ce immediate execution if this sacri…ce
results in lower trading costs.4 On another hand, most exchanges - also in the
e¤ort of consolidating the order ‡ow into the central market - o¤er institutional
4 see Schwartz and Steil (1996).
3
investors to trade the total size of blocks to be concealed on the central trading
system by means of hidden orders.
3 Block trading mechanisms in Europe
In this section we outline the state of rules governing the trading of large orders
in the major European stock exchanges: the London stock exchange, the
Euronext markets, the Frankfurter Wertpapierbörse, the Milan stock exchange
and the Swiss exchange. Institutional details of the London stock exchange,
the Swiss exchange and the Euronext markets are mostly drawn from Demarchi
and Foucault (1998). Pagano and Steil (1996), Pagano (1998), Biais (1998)
and Demarchi and Foucault (1998) give an exaustive description of the most
important changes in the past decades in the microstructure of the respective
downstairs markets of these stock exchanges. Cybo-Ottone et al. (2000) discuss
about the general trends in force in securities markets, with a particular look
at the European Union.
3.1 The London Stock Exchange
The London Stock Exchange is the biggest European exchange for turnover
and the number of stocks listed. A quote-driven and an order-driven system5
are provided for both domestic and foreign securities. Right the coexistence of
both trading mechanisms leads to market fragmentation, with the conseguence
to divert order ‡ow from the order book. Actually, unlike in other European
markets, over 50% of volume is executed away from the electronic order book.
In this framework, dealers still play a fundamental role especially in the trading
of large size orders.
In the London Stock Exchange, block thresholds are de…ned on the basis
of the Normal Market Size6 (NMS). A transaction is considered a block and is
therefore allowed to use the block trade facility if it is: a) larger than 75 times
the NMS of 2,000 shares or above; b) 50 times NMS of 1,000 or 500 shares.
As for European securities, a transaction can be elected to use the block trade
facility if it is quali…ed as such on a regulated European market, or more in
general, for an international security traded on the International Order Book,
the threshold is above 50 times the NMS.
Block trades are mainly ”worked” over the market trading session (09.00-
17.00 hours). The reason is that over o¤-exchange trading hours there are few
reference prices for market makers, therefore blocks traded o¤ the exchange
hours take place on other markets which are open. Since the introduction of
5The quote-driven systems are SEAQ and SEAQ International for domestic and foreign
securities, respectively. SETS and International Order Book are order-driven trading systems
o¤ered for the trading of domestic and international securities. There exists also a hybrid
trading system (SEATS) for domestic securities.
6A Normal Market Size (NMS) for a stock is roughly 2.5% of the average daily trading
volume for the same stock.
4
SETS in 1997, when member …rms were permitted to act as principal for all order
sizes, large trades (> 8 Normal Market Size) in SETS are subject to a special
procedure, the Worked Principal Agreement (WPA). Under a WPA, a member
…rm acting as principal and its client agree to execute, at some point in time in
the future, a large trade. The price and size of the trade are determined at the
time of agreement but the member …rm can o¤er price and/or size improvement.
3.2 Euronext markets (the Paris Bourse, the Amsterdam,
the Bruxelles and the Li*****ona Exchanges)
The Euronext N.V.7 consists of an integrated trading platforms for the cash
markets of the Paris Bourse, the Amsterdam exchange, the Brussels and the Li*****ona
exchanges. This order-driven trading architecture is based on the French
trading system, the Nsc (Nouveau Systèm de Cotation). The underlying market
undertakings operate a series of regulated markets for …nancial instruments
which remain separate legally, subject to the laws and to the supervision of the
competent authorities of relevant jurisdiction.
The market rules that govern the Euronext markets are set forth in a ”Rule
Book”, one for each Euronext market undertaking, as well as in ”notices” issued
pursuant to such Rule Book. A number of these rules have been harmonized
across the Euronext market undertakings and apply equally to all Euronext
markets involved (Book I of Market Rules). Other market rules, pending in
view of a coming process of harmonization, remain country-speci…c (Book II).
The common rules for block trading in Euronext markets are in e¤ect trading
rules that were enforced at the Paris Bourse and that are subsequently
extended to other markets. Before the creation of the common trading system,
block trades at the Amsterdam exchange were conducted in a wholesale market
completely separated from the retail segment and accounted for a relevant provision
of liquidity for large transactions after the Paris Bourse. Nowadays the
Paris Bourse remains the most important source of liquidity for institutional
investors since almost 67% of block trades still executes on the French upstairs
market. Table 4 shows the distribution of order book and block trades in cash
segment for individual Euronext markets.
Block trading rules are very detailed and composed. The amount of the
minimum block thresholds (the Normal Block Size8 , or ”NBS”) vary depending
on the phase the stock is traded (continous or auction) and its liquidity. Block
trades are prematched under two forms. First, prearranged trades can be executed
between or at the current best bid and ask prices on the order book.
These are called ”applications”. Second, a block also can be traded outside the
current spread, but then the ”clear the book” rule must be observed. This rule
7 Euronext also consists of integrated clearing and settlement systems through Clearnet SA
and Euroclear, respectively.
8The NBS is the minimum order size for an order to be eligible as a block trade. For
each stock, it is roughly at least equal to 2.5% of the average daily trading volume in the last
quarter and equal to 7.5 times the average depth at the best bid and ask prices in the last
quarter.
5
varies depending whether the block is ordinary or structural. Within ordinary
block trades, if the amount of the block trade is less than 5 times the relevant
NBS, block prices must be within the Weighted Average Spread (WAS), however
they are allowed to ‡uctuate from 5% below the best bid limit price to 5%
above the best ask limit price displayed in the central limit order book. In case
the amount of the block trade is equal or greater than 5 times the relevant NBS,
the WAS need not to be respected. Then the allowed price is 5% around the
last traded price on the central order book, during or not the trading session.
Structural block trades are de…ned as those block trades whose amount of the
block trade is equal or greater than 5% of market capitalization of the company
whose stock is being traded. In this case a spread of 10% (by referring to the
last traded price on the central limit order book) is allowed.9
The Weighted Average Spread is the di¤erence between a weighted average
of the best ask prices and a weighted average of the best bid prices up to
the Normal Block Size.1 0 For very large trades (>5NBS), the computation of
SuperWAS (the WAS enlarged) is allowed by request.
3.3 The Frankfurter Wertpapierbörse
A main feature of Germany is that trading still takes place in eight di¤erent
Stock Exchanges: Berlin, Bremen, Düsseldorf, Frankfurt, Hamburg, Hanover,
Munich and Stuttgart. Frankfurt (Frankfurter Wertpapierbörse, FWB) is the
largest of the eight German stock exchanges. For each exchange trading takes
place in a hybrid form: orders are routed to the ”Kursmakler”11 by brokers on
the ‡oor or through an electronic order routing system. Since 1996 the trading
of small-medium size orders takes place on Xetra, a pure order-driven market.
As far as large trades concerns, trading was characterized by o¤-exchange
and OTC transactions. In March 2001 the FWB introduced Xetra XXL, an
exchange-based, fully electronic block trading on Xetra with the view to reduce
o¤-exchange negotiations by cutting down trading costs. The current version
of Xetra XXL dated to November 2001 with major innovations in terms of
transparency.12 Since its launch Xetra XXL has been welcomed with market
9We are grateful to Charles Hoppmann at Euronext Paris for this clari…cation.
1 0 Suppose the NBS for a stock is 10000 shares and the limit order book is as follows:
Buy side (B) Sell Side (S)
Qt Price Price Qt
3 000 10 11 4 000
5 000 9 12 7 000
4 000 8 13 2 000
So the WAS will be :
B: (3 000*10 + 5 000*9 + 2 000*8)/10 000 = 9.1
S: (4 000*11 + 6 000*12)/10 000 = 11.6
WAS = 11.6- 9.1 = 2.5
1 1The ”Kursmakler” acts as the specialist in the NYSE. This broker-dealer manages a limit
order book for each stock.
1 2Another innovation concerns the number of instruments which are allowed to trade on
Xetra XXL. Besides DAX equities, block trades on MDAX and TecDAX instruments are
allowed to execute on Xetra XXL.
6
participants and trading volume on it is expected to increase.1 3 Because of its
innovative market features for block trading, we focus on Xetra XXL henceforth.
The German trading model for large transactions Xetra XXL represents
an innovation in the organization of trading systems for large blocks in the
framework of public exchanges, with respect to traditional broker-intermediated
negotiations, since it is nonintermediated and trades take place in a call auction
market where orders are routed to an electronic closed order book. The call
auction consists of prede…ned crossing periods based on a crossing schedule,
each consisting of a crossing’s pre-call phase and a crossing’s call phase with
random end. Between two crossing periods, the trading system is in the pre-call
phase. A crossing period starts after a scheduled point in time; over this period
the current crossing prices (the midpoint of the spread computed on the basis of
the order book information of the reference market) are continously displayed.
After the start of a crossing period, the change of a crossing pre-call phase
to a crossing call phase is triggered by the submission of market orders or if
the order limit for a buy (sell) order is equal to or higher (lower) than a predetermined
threshold de…ned by the exchange based on the current midpoint.
The change of the trading phase from pre-call to call may occur immediately
at the start of the crossing period or within the crossing period. If there are no
order entries, or orders not ful…lling the above criteria, the crossing period ends
without switching into the crossing call phase.
Transparency in this trading model is assured because the order book is
completely closed over the duration of the crossing auction. However, the phase
change is displayed to market participants as a tradable order is submitted and
real time information about the existence of a block order is available continously
both via Xetra and via information vendors.
A market design for large trades such as Xetra XXL shows many bene…ts.
First, search costs for counterparts and broker commissions are reduced at minimum
level. Xetra XXL is designed to explicitly reward the provision of liquidity
and incentivate the entry of block orders before the call auction by cutting exchange
transactions fees. Morevover, institutional traders are sure about the
prices at which their trades execute, since block trades execute at the midpoint
quote of the current spread of the reference market. In case of an imbalance
of buy and sell volume, orders are matched according to volume/time priority.
More important, institutions may not be afraid of being front-running and possible
information leakage. However, it is questionable whether the execution
price for block trades - the midpoint quote - is e¤ectively a ”good” price for the
institution.
We note also that in this block trading model orders provide no contribution
to price discovery. In the context of block crossing, order limits serve only as
execution conditions and orders are only good-for-day.
1 3 In particular, with the new release of Xetra XXL, the display of existing block orders led
to an open interest of 247.8 million and a record sales of 35.4 million in December 2001 while
the amount of 4 million shares with a volume of 311,440,000 within a single trades in May
2002 was the largest trade executed so far.
7
3.4 The Milan Stock Exchange
The Milan Stock Exchange has been long characterized by a very large o¤-
exchange trading volume mainly because of the existence of statutory …xed
commission on the exchange. By the concentration rule introduced in 1991 (the
SIM law),14 order ‡ow has been consolidated in an electronic limit order book
for the trading of small-medium size orders, while large orders trade o¤-exchange
through broker intermediation. Table 5 provides a historical perspective of retail
and block trading at the Milan Stock Exchange after the 1991 …nancial reforms.
Actually, o¤-exchange trades, which include blocks and transactions called
"fuori mercato", must be viewed as an exception to the concentration rule. O¤-
exchange trades are allowed provided they be executed through a prior written
authorization by the client and at a better price than one could obtain on-market
(best execution principle). What is not immediately clear is the meaning of the
"better price". The implication is that there exists an objective and easily
veri…ed parameter with respect to which an investor can always compare the
price he obtains o¤-exchange with those existing on the central market. Of
course, as Steil (1996) notes, this parameter cannot easily be veri…ed, and is
certainly not objective. A question is how to compare prices across markets
unless the investor is demanding the same service. However, allowing for a
less restrictive interpretation of the "better price" principle, it may be that the
”better price” (even best price) should be meant to be better (even best) to the
client. This price may be worse than those prevailing on the central market,
however the client may held it as the better he can achieve for himself. The
motivation underlying the best execution principle is that better prices may be
attained through private negotiations.
Block thresholds, which are updated every six months on the basis of the
average daily turnover of the previous six months, have ever grown. This may
re‡ect on one hand, the increasing average volume and size of institutional
trades and on another, the e¤ort of Italian regulators to discourage trading o¤-
exchange in order to consolidate the central market order ‡ow. Up to date there
are three minimum block size thresholds:
- not less than 150,000 Euro for stocks with an average daily turnover of less
1,5 millions Euro
- not less than 250,000 Euro for stocks with an average daily turnover between
1,5 and 3 millions Euro
- not less than 500,000 Euro for stocks with an average daily turnover between
3 and 10 millions Euro
- 1,5 millions Euro for stocks with an average daily turnover greater than 10
million Euro.
1 4Amihud, Mendelson and Murgia (1990) provide a deeper insight into the Italian Exchange
prior to the reforms of Italian …nancial markets. Murgia (1993) makes a comparative study
of the market performance in the transition from the ‡oor to the electronic trading.
8
3.5 The Swiss Stock Exchange
The Swiss Exchange allows for individual orders larger than CHF 200,000 to
be executed o¤-system. For these trades, the Exchange o¤ers a trading facility
that provides for a form of electronic negotiation. Using this facility a member
can make a Statement of Interest that indicate to the other members, in a nonbinding
manner, that he or she would like to trade in a certain stock. Members
can also direct an Addressed O¤er to a speci…c member (or to several speci…c
members), which can then be accepted, ignored or rejected.
By exchange rule, o¤-system trades have to be executed at prices prevailing
in the central market at the time of the trade1 5 (Best Execution Principle). We
also observe that a condition for a large transaction be executed as a block is
that the participant itself assumes the position risk.
4 Institutional similarities and dissimilarities
In this section we consider the institutional common features and di¤erences in
the basic design features of block trading systems of the …ve European stock
exchanges surveyed in the previous section. The distinction between what features
are common and not is however hard in that common institutional features
themselves share remarkable institutional di¤erences. Rather than proceeding
by categorizing, we identify some basic aspects in block trading mechanisms
and for each design feature we outline similarities and dissimilarities across exchanges.
We identify mainly …ve basic areas:
- enforcement of speci…c arrangements for block trading
- enforcement of interaction rules
- block prices are not integrated in the formation of prices
- transparency
- toward nonintermediated block trading?
4.1 Enforcement of special procedures for blocks handling
All the …ve exchanges provide an ”upstairs market” for the handling of large
trades. However, some exchanges provide for special procedures that block
traders must observe over the negotiation of blocks. Overall, these procedures
consist of an agreement between the counterparties before a block transaction
actually takes place so that all the trade terms are …rm and well known also to
exchange o¢cials. Examples of these special procedures are the applications at
Euronext markets, theWPA at the LSE, the Statement of Interest at the SWX.
No special trading procedure is provided on Xetra XXL, whereas in the Milan
Stock Exchange block traders freely come to negotiate for large orders. The
upstairs market at the Milan Stock Exchange is somehow particular with respect
to upstairs markets of other European exchanges because that upstairs market is
1 5We thank Britta Hirschi at the Swiss Exchange for this detail.
9
actually not regulated, that is block transactions do not give rise to a pure block
market; rather, large transactions should be considered as individual operations
negotiated o¤ regulated markets in the context of market rules dictated in this
issue by the Italian regulators.1 6
A possible explanation for this built-in design feature - special procedures for
block trading within a trading system for block trades - is that in markets where
this procedure is provided block traders are often engaged as dealers.17 The size
and days taken for a block trade be ”worked” may have induced Exchanges to
provide special procedures for safe and good conclusion of blocks of considerable
size.
4.2 Enforcement of interaction rules
Most exchanges enforce some interaction between upstairs and downstairs markets.
The interaction (or crossing) rules (”best price”, ”cross reference”, ”clear
the book” rule) stipulate the allowable price range for upstairs trades and
whether downstairs orders that o¤er superior prices for smaller quantities will be
allowed to participate in the transaction.1 8 The reference price at which block
prices refer to is usually the midpoint quote or an weighted average spread; in
any case, block prices are allowed to ‡uctuate at maximum within a 10% range.
The enforcement of interaction rule is commonly presented as an issue of
fairness regarding order book users in order to prevent trade-throughs and freeriding
of upstairs trades on price discovery. Further, enforcing block prices to
interact with the downstairs market may in part attenuate the impact of block
trades on the market and preserve the signalling function of prices. Whatever
the merits of these positions, price enforcement is equivalent to immediate publication
of blocks done outside the bid ask spread and represents a clear subsidy
by o¤-exchange traders of on-exchange traders.1 9
Paris Bourse interaction rules, created in 1989 after members were permitted
to trade for their own account, were signi…cantly relaxed in 1994 owing to the
e¤ect of regulatory arbitrage in favor of London. Block traders in Paris routinely
executed their block traders in London to avoid having to expand capital or leak
information by obeying the Bourse’s interaction rules. Even under the current
Paris regime, block trades are still often executed in London via screens in Paris
to avoid the market impact risk that a dealer might take on in trading within
the Paris spread limits.
So far empirical evidence shows that relatively stringent interaction rules
have the bene…t of reducing price di¤erence between downstairs and upstairs
1 6We thank Luca Filippa at the Borsa Italiana for this clari…cation.
1 7 In the Milan Stock Exchange block traders rarely commit their own capital in block
transactions because of the considerable capital amount required (Murgia andGottardo, 2000).
1 8At the NYSE, for example, upstairs trades must typically be completed at prices at or
within the downstairs Best Bid-O¤er (BBO) quotes, and downstairs participants are allowed
to take a portion of the block. At the Toronto Stock Exchange, upstairs trades need to be
executed at or within the BBO quotes in the downstairs market at the time the order is
received.)
1 9 see Davis and Steil (2001).
10
markets.20 However, the debate over enforcing or not interaction rules remains
open. We would point out two aspects related to crossing rules. First, they are
not enforced in all exchanges. Although the IMilan Stock Exchange statutes
the best execution principle for block trades, block prices are actually far away
from current market prices. The lack of an interaction rule in the London
Stock Exchange is due to the remarkable importance of dealers in negotiating
large orders; competition among dealers will keep prices to be closer to one
another. Another point concerns the extent of interaction rules’ restrictiveness.
In Euronext markets there is a cross-sectional variation of the rule depending
on the order size with consistent e¤ects on execution costs. Besseminder and
Venkataraman (2002) …nd that for the so called eligible stocks (the most liquid
stocks in CAC) trades occurring outside the quotes pay execution costs that
are about 40 to 50 basis points higher than for upstairs trades executed at or
within the quotes. They show that more ‡exible crossing rules reduce incentives
to manipulate downstairs spreads.
4.3 Block prices excluded from price discovery
In all exchanges surveyed, block trading arrangements are not integrated into
the price discovery mechanism of the trading system. This objective is partially
obtained by making block prices to interact. Where there are no interaction rules
as in the Milan and London Stock Exchanges, block prices are not considered in
the computation of market indexes, although there is an implicit price impact
on the market due to these large trades.
The issue about the entering of block prices into price discovery is of concern
because of the size of these orders and the motivation behind these trades,
whether liquidity- or information-motivated. In the former case, large trades
cause only a temporary impact on market prices which then reverse to their
level prior to the block trade. In the latter case, market prices are alterated
and the resulting prices may not be ”true” prices. Many exchanges therefore
require some price interactions and other bland requirements in terms of transparency.
In exchanges where no price interaction is provided, block traders -
acting as dealer or/and broker - are somehow entitled to reduce or eliminate
pricing anomalies.
The extent to which upstairs prices enter the price discovery process has been
the focus of latest empirical works. For instance, using the models of Gonzalo
and Granger (1995) common factor and Ha*****rouck (1995) information share
methods and data of the Helsinki stock exchange, Booth et al. (2002) assess
the relative importance of the upstairs and downstairs markets’ contribution to
price discovery. They …nd that the downstairs market contributes more to price
2 0 see Madhavan and Cheng (1997) for the NYSE and Booth et al. (2002) for the Helsinki
stock exchange. The latter …nd that pricing di¤erence between upstairs and downstairs market
at the NYSE is smaller than that at the HSE because of an NYSE regulation which requires
that an upstairs trade has to be exposed to the public in the downstairs market, as opposed
to Finland’s "best price" rule which only requires brokers make an upstairs trade at a price
that is the best price for the customer.
11
discovery than upstairs market, however the downstairs market’s dominance
is not prevalent. Upstairs prices consist of downstairs prices plus a transitory
component.
4.4 Transparency
Transparency covers three aspects of markets - pre-trade, post-trade publication
and anonymity of counterparties. The issue concerning transparency has been
long debated by academics and practioneers since too much transparency may
jeopardize the willingness of traders to o¤er liquidity, especially in case large
volumes are traded. In e¤ect, institutions concern that their orders may be
front-run once counterparties …nd that a broker who is known to have links
with institutions is in the market. The combination between transparency and
liquidity is ultimately left to the decision of national exchanges.
At European level there has been an e¤ort to set up a minimum common
transparency regulatory requirement under the Investment Service Directive.2 1
The result of that regulation is that national exchanges have actually adopted
their own rules. The main di¢culty to …nd common rules in transparency
arises from di¤erent market structures in Europe, since London is a quote-driven
system while continental stock exchanges are order-driven.
4.4.1 Anonymity
Anonymity is maintained before and after execution of blocks. The identity of
counterparties is unknown, although in intermediated block markets it is likely
that block intermediaries know the parties involved in the transaction.
4.4.2 Pre-trade transparency
Unlike the full access to the electronic order book to the market, in all exchanges
the upstairs market remains opaque during pre-trade period. In exchanges where
interaction rules are enforced, what is displayed over the trading phase is the reference
price (the WAS or the midpoint of the spread). On Xetra XXL although
the electronic order book is closed, transparency has been increased through the
indication on the existence of block orders per stock to the market.
4.4.3 Post-trade transparency
There is a convergence in post-trade transparency rules as for tight reporting
deadlines, the reporting of all trade details to exchange o¢cials and the release
of some trade details to the market (instrument identi…cation, date and time of
execution, price and quantity traded in the instrument). The post-trade information
however di¤ers in the time of publication to the market. The complex
2 1 See Steil (1996) for a discussion of the European Investment Services Directive and its
e¤ects.
12
procedure set forward in the European Investment Services Directive had been
almost entirely ignored.
The publication of large trades is problematic. Actually it jeopardizes the
ability of a market maker involved in the block trade to unwind his position in
good conditions. Thus quick publication can ultimately result in bad prices for
block trades. However, large trades are informative. For prices in the central
limit order market to re‡ect the information contained in these trades, a quick
publication of the price and the size of the block trade is necessary. Timely
publication is also necessary in order to prevent counterparties in a block trade
from trading on their superior information. This can ultimately impair the
liquidity of the central limit order book, by increasing the adverse selection risk
faced by limit order traders. Given these costs and bene…ts of quick publication
of large trades, exchanges allow traders to delay publication of the large trades
in which they are involved. The conditions under which this delay can occur
and its duration vary across exchanges, however (see Table 1 in the Appendix).
The delay allowed varies from hours to even several days later. Delays are most
likely in markets where dealers take position to facilitate investor business: in
the LSE, for instance, publication of large trades occurs …ve days following the
execution of the transaction or when the block trade has been 90% o¤set.
4.5 Toward nonintermediated block trading?
The SWX and the FWB are the …rst European exchanges to introduce an electronic
trading platform for the trading of large transactions. With respect to
human intermediation nonintermediated electronic auction markets yield trading
cost savings. Besides the elimination of broker commissions, electronic trading
of large trades ensures immediate execution so to cut o¤ execution costs,
market impact and information leakage and front-running risk intrinsec in a
broker-intermediated negotiation.
Today block trading is actually distributed among di¤erent execution mechanisms.
The typical way remains telephone trading consisting of direct bilateral
trading or brokered trading via telephone. Another venue is exchange trading
through speci…c non automated trading systems such as upstairs markets. The
rapid development of technology enables block trades to be executed through
separated platforms such as POSIT, E-crossnet.
All the above mechanisms have in common a process of search for counterparties.
The major risk of the search process is a likely information leakage as
when a block is being ”shopped” around the market and potential counterparties
may learn of it so that the block is priced worse than there is no information
leakage. In case of a large amount being traded, an institution wanting to trade
even anonymously, will move prices adversely merely in revealing its interest.
Knowledge of this interest in the market leads participants to infer that the
current market price does not accurately re‡ect demand. Bids and o¤ers will
thus adjust accordingly even without any transaction taking place, thereby precluding
the trader from …lling only part of the order at the price that prevailed
before the block order was revealed. The automation of block trading may re-
13
duce market impact and risks related when submitting large orders. However,
in a completely automated trading system, what is then the economic function
that upstairs markets accomplish, that is the role played by block intermediares
- acting as principal or agent? The importance of block intermediaries is
expressed in the superior knowledge the intermediaries have of the order ‡ow
and that enables them to tap into unexpressed trading interests as discussed
in Grossman (1992). Brokers play also a screening function by discriminating
among informed and uninformed traders and imposing a sort of reputation to
all parties. Empirical evidence shows that, in this respect, uninformed (institutional)
traders prefer the upstairs market, while informed traders randomize
their trades between the upstairs and downstairs markets.
In general, the trend is evolving to increasing disintermediation of block
trading. It will likely be that human intermediation of block trades will not
disappear completely. The implication is rather that intermediation that is built
into trading structures by design, rather than being chosen by the investor, is
unlikely to withstand increasing contestability in the market for trading systems.
Competition on liquidity at di¤erent block sizes may increase from alternative
trading mechanisms.
5 Conclusions
In this paper we survey the microstructure of block trading in the major European
stock exchanges. We focus speci…cally on the tradingmechanisms for block
trades because of the increasing importance of institutional trading in Europe
and the deriving demand for trading facilities. Our study is about the supply
side to analyze how the major European exchanges respond to this consistent
proportion of market order ‡ow.
We identify …ve basic design features that exchanges share and that at the
same time make them di¤erent from the regulatory point of view. Common
institutional features include the handling of block trades in a separate trading
arrangement, typically called the upstairs market; the e¤ort to not integrate
block prices into price discovery, and bland requirements in terms of regulation
(transparency). The means adopted by exchanges to pursuit these objectives
are several. However, the …nal view is a wide institutional fragmentation, which
mostly determines and favors regulatory arbitrage across European exchanges.
In particular, institutional dissimilarities concern the enforcement of interaction
rules and the level of post-trade transparency.
Institutional di¤erences have been the main cause of failed attempts for the
creation of a common European market structure, attempts initiated both by
national exchanges or at European level. The regulatory di¤erences remain
also the major obstacle to remove in the process of designing a common market
structure for block transactions. However, despite this fragmentary institutional
environment, it will be the competition for and arbitrage in trading rules which
bring European exchanges to search for common trading rules. Actually, as
14
surveyed by Schwartz and Steil (1996), European institutional investors seem
to prefer these rules be set up by individual national exchanges rather than at
European level.
15
References
[1] Amihud, Y. and H. Mendelson (1991), How (not) to integrate the European
capital markets, in A. Giovannini and C. Mayer (eds.), European Financial
Integration, Cambridge University Press.
[2] Amihud, Y., H. Mendelson and M. Murgia (1990), Stock Market Microstructure
and Return Volatility, Journal of Banking and Finance 14,
423-440.
[3] Besseminder, H. and K. Venkataraman (2002), Does an Electronic Stock
Exchange Need an Upstairs Market?, working paper, Southern Methodist
University.
[4] Biais, B. (1998), European Stock Markets and European Uni…cation, in European
Capital Markets with a Single Currency, J. Dermine and P. Hillion
(ed.), Oxford University Press.
[5] Booth, G.G, J-C. Lin, T. Martikainen, and Y. Tse (2002), Trading and
Pricing in Upstairs and Downstairs Stock Markets, Review of Financial
Studies 15, 1111-1135.
[6] Borsa Italiana, Istruzioni al Regolamento dei Mercati Organizzati e Gestiti
dalla Borsa Italiana S.p.A., July 2002.
[7] Borsa Italiana, Regolamento dei Mercati Organizzati e Gestiti dalla Borsa
Italiana S.p.A., July 2002.
[8] Burattelli, S. (1999), Ordini e Contratti di Borsa, Milano, Il Sole 24 Ore.
[9] BusinessWeek (2002), Battle of the Bourses, October 28.
[10] Cybo-Ottone, A., C. Di Noia and M. Murgia (2000), Recent Development
in the Structure of Securities Markets, Brookings-Wharton Papers on Financial
Services, 223-273.
[11] Davidson, J., L.M. Grépin and C.M. Hogg (2001), Remaking Market Making,
The McKinsey Quarterly 3.
[12] Davis, P.E. and B. Steil (2001), Institutional Investors, MIT Press, Cambridge.
[13] Demarchi, M. and T. Foucault (1998), Equity Trading Systems in Europe.
A Survey of recent changes, working paper presented at the 1999 EFMA
Conference in Paris.
[14] Domowitz, I. and B. Steil (2001), Automation, Trading Costs, and the
Structure of the Securities Trading Industry, in Davis P.E. and B. Steil,
Institutional Investors, MIT Press, Cambridge.
[15] Euronext Rule Book, Book I, Harmonized Market Rules, July 2001.
16
[16] Federation of European Securities Exchanges, European Stock Exchange
Statistics, May 2001 and 2002.
[17] Gonzalo, J. and C.W.J. Granger (1995), Estimation of Common Long-
Memory Components in Cointegrated Systems, Journal of Business and
Economic Statistics 13, 27-36.
[18] Grossman, S.J. (1992), The Informational Role of Upstairs and Downstairs
Trading, Journal of Business 65, 509-528.
[19] Gruppe Deutsche Börse, Xetra XXL. Market Model, version 2.0, September
2001.
[20] Gruppe Deutsche Börse, About Xetra XXL. Release 6.5, November 2001.
[21] Institutional Investor (1999), Euro 100 cross border express - Europe’s
largest money managers, Institutional Investor, November.
[22] Ha*****rouck, J. (1995), One Security, Many Markets: Determining the Contribution
to Price Discovery, Journal of Finance 50, 1175-1199.
[23] London Stock Exchange, Rules and Regulatory Guidance, July 2002.
[24] Lucarelli, C. (2001), Gli accordi internazionali fra mercati, in Anderloni L.,
I. Basile and P. Schwizer (eds.), Nuove frontiere dei mercati …nanziari e
delle securities industry, Bancaria Editrice, Milano.
[25] Madhavan, A. (2000), Market Microstructure: A Survey, Journal of Financial
Markets 3, 205-258.
[26] Murgia, M. (1993), Il Mercato Telematico Azionario, IRS, Quaderno n. 7.
[27] Murgia, M. and P. Gottardo (2000), The Upstairs Market for Large-Block
Trades: An Empirical Analysis of Price E¤ects on Italian Shares, Working
Paper.
[28] New York Stock Exchange, Fact Book 2001, Year in Review.
[29] New York Stock Exchange, Fact Book 2002.
[30] Pagano, M. (1998), The ChangingMicrostructure of European Equity Markets,
in The European Securities Markets: The Investment Service Directive
and Beyond, edited by Guido Ferrarini, Kluwer Law International.
[31] Pagano, M. and B. Steil (1996), Equity Trading I: The evolution of European
trading systems, in B. Steil (ed.), The European Equity Markets. The
State of the Union and an Agenda for the Millenium, Royal Institute of
International A¤airs, London.
[32] Schwartz, R.A. and B. Steil (1996), Equity Trading III: Institutional Investor
Tradig Practices and Preferences, in B.Steil (ed.), The European
Equity Markets.
17
[33] Seppi, D.J. (1990), Equilibrium Block Trading and Asymmetric Information,
Journal of Finance 45, 73-94.
[34] Smith, B.F., A.S. Turnbull, R.W. White (2001), Upstairs markets for Principal
and Agency Trades: Analysis of Adverse Information and Price Effects,
Journal of Finance 56, 1723-1746.
[35] Steil, B. (1996), Equity Trading IV: the ISD and the Regulation of European
Market Structure in B. Steil (ed.), The European Equity Markets.
The State of the Union and an Agenda for the Millenium, Royal Institute
of International A¤airs, London.
[36] SWX Swiss Exchange, Directive 11: Reporting requirement with respect
to Sensitive Block Transactions, Version 01.01.2000.
[37] SWX Swiss Exchange, SWX Trading info, at website:
www.swx.com/market/trading_organisaion_en.html
[38] Trovatore, G. (2001), Concentrazione degli scambi e competizione tra sistemi
di trading, in Anderloni L., I. Basile and P. Schwizer (eds), Nuove
frontiere dei mercati …nanziari e delle securities industry, Bancaria Editrice,
Milano.
[39] Wells, S. (2000), Price Discovery and the Competitiveness of Trading Systems,
FIBV Annual Report.
18
Table 1 - Domestic Equity Turnover - end May 2001 and May 2002
All market segments, Euro millions (single counted)
May 2001 May 2002
Exchange Trading Electronic Negotiated Trading Electronic Negotiated
days Order Book Deals days Order Book Deals
Transactions Transactions
LSE 21 89 803.5 96 997.5 22 90 243.5 90 338.2
Euronext¤ 23 153 295.0 66 666.1 22 137 567.0 61 434.0
DB AG¤¤ 22 73 008.0 68 730.3 22 70 283.0 38 144.0
BIt 22 57 670.7 7 219.0 22 52 346.6 4 021.9
SWX 21 38 880.0 10 722.2 22 2 397.2 634.8
Source: FESE
Notes - LSE: the London stock exchange; DB AG: the Deutsche Börse AG; BIt:
Borsa Italiana (the Italian stock exchange); SWX: the Swiss exchange. ”Negotiated
Deals” refer to deals negotiated by brokers and reported to the Exchanges.
¤Euronext markets statistics refer to the Paris Bourse SBF, the Amsterdam and
the Bruxelles Exchanges.
¤¤Note that …gures about the DB AG include all the German stock exchanges.
Table 2 - Foreign Equity Turnover
All market segments, Euro millions (single counted)
May 2001 May 2002
Exchange Trading Electronic Negotiated Trading Electronic Negotiated
days Order Book Deals days Order Book Deals
Transactions Transactions
LSE 21 530.8 319 351.6 22 2 400.0 217 031.9
Euronext¤ 23 1 102.0 735.0 22 1 628.0 524.0
DB AG¤¤ 22 2 350.0 11 979.6 22 2 934.0 6 213.0
BIt 22 3 203.2 2.7 22 6 513.2 59.4
SWX 21 1 609.7 132.1 22 879.4 176.1
Source: FESE
Notes - LSE: the London stock exchange; DB AG: the Deutsche Börse AG; BIt:
Borsa Italiana (the Italian stock exchange); SWX: the Swiss exchange. ”Negotiated
Deals” refer to deals negotiated by brokers and reported to the Exchanges.
¤Euronext markets statistics refer to the Paris Bourse SBF, the Amsterdam and
the Bruxelles Exchanges.
¤¤Note that …gures about the DB AG include all the German stock exchanges.
19
Table 3 - European Money Management Centers
$ billion Percent of total
United Kingdom 4132 40
Switzerland 1997 19
Germany 1456 14
France 938 9
Netherlands 936 9
Italy 306 3
Sweden 257 2
Spain 154 1
Belgium 82 1
Ireland 38 0.4
source: Institutional Investor (1999).
20
Table 4 - Euronext markets (equity segment)
Euro millons as from Dec 1999 to Dec 2001
Trading Value of Domestic Shares Value of Foreign Shares
Year Location days Electronic Reported Electronic Reported
Order Book Deals Order Book Deals
1999 Amsterdam 22 34 047.0 19 318.0 78.8 373.7
Bruxelles 22 3 499.3 96.6 356.2 11.6
Paris 22 67 676.5 n/a 920.8 n/a
Li*****on and Oporto 19 4 597.3 n/a n/a n/a
2000 Amsterdam 19 41 057.6 29 275.3 10.4 77.3
Bruxelles 18 3 722.4 94.8 251.9 7.3
Paris 19 89 463.8 5 675.1 1 041.1 13.6
Libon and Oporto 17 4 581.2 n/a 5.0 n/a
2001 Amsterdam 18 33 865.5 15 506.2 14.8 5.2
Bruxelles 18 3 018.1 16.2 214.6 59.3
Paris 18 65 410.9 12 055.5 621.5 28.8
Libon and Oporto 17 2 061.2 n/a 2.5 n/a
Source: Paris Bourse SBF.
21
Table 5 - Milan stock exchange: upstairs and downstairs trading
Equity segment
Order book trading Block trades
Trades Turnover Trades Turnover
Number ML Euro Number ML Euro
1995 4 867 774 72 721.3 1 062 5 396.0
1996 5 494 904 81 129.1 1 220 3 168.3
1997 11 880 773 175 370.1 1 539 7 271.6
1998 24 884 935 424 852.8 3 126 18 775.6
1999 28 236 736 502 990.1 3 128 28 370.1
2000 50 687 351 838 491.7 4 796 46 544.3
2001 36 740 354 637 074.9 4 583 45 492.6
2002 37 042 552 622 897.8 3 804 28 904.3
Source: Borsa Italiana, Fatti e cifre 2002.
22
APPENDIX
TABLE 1 Price and Publication of Large Orders
Publication delay for Block Trades Block Price
NSC Order size i) 2 hours if broker acts as counterparty better prices in the limit order book
ii) Immediate if broker acts as agent ii) Block price must be inside the weighted
average spread, that is computed using the
Order size > 5 NBS best ask and bid prices in the limit order
i) Next morning if broker acts as counterparty book up to NBS¤¤
ii) Immediate if broker acts as agent
Structural blocks¤ iii) Structural block prices can be within +/
Immediate or T+2 if member acting as 10% of the best ask and bid quote
principal has not o¤set his position
SETS¤¤¤ Ordinary risk trade: Immediate publication i) No obligation to execute o¤ers posted at
Worked Principal Agreement: End of the a better price in the limit order book
trading day or once 90%of the transaction is ii) No price link with central limit order boo
o¤set
SWX 30 minutes i) No obligation to execute o¤ers posted at
better prices in the limit order book
ii) Rule of Best Execution: same execution
prices as those that could be realized in lim
order book.
XETRA At the end of the trading day (8.30 pm) Reference price: the midpoint price in the
XXL XETRA at prespeci…ed times
BIt i) 90 minutes of the execution of block trades Wholesale orders can be executed at any pr
if blocks traded during the trading session
ii) within 9.00 am next day if executed o¤-
trading hours
Source: Demarchi and Foucault (1998) and national stock exchanges’ rules.
* The value of a structural block must represent at least 2% of the company’s capital or
be greater than FF50m for a stock with a market capitalization larger than FF1bn. It must
be at least 5% of the company’s capital otherwise.
** A larger spread (SuperWAS) is computed for block trades larger than 5NBS on request
to SBF. These trades can be executed at prices within +/-5% of best bid and ask prices.
*** Di¤erent publication rules are used for stocks that trade in SEAQ (see Pagano and
Steil (1996)).
23