What Kind of Stocks Fragment the Most?

Overview

As most people reading this article will know, the percentage of trading volume on the main market LSE for FTSE-100 stocks is currently hovering at around 60%. So about 40% of the volume in these stocks is executed on the other lit MTF venues.

However, within the 102 stocks of the FTSE-100 index, how much does the level of fragmentation differ for individual stocks? Assuming there are significant differences in market shares for LSE for different stocks in the index (there are), a more intriguing question is what 'kind' of stocks might the MTF venues have gained most market share in?

For instance, does the LSE or Euronext do relatively better or worse (in market share terms) with stocks that have higher or lower tick sizes relative to their spreads(strictly speaking 'spread leeways'). Have a guess now and check your answer later in this article.

Aside from the simple market share the primary venue retains, one can also pose MTF venue specific questions such as:

• Does CHIX do better or worse on stocks with higher/lower intraday price volatility?
• Does TRQX do better or worse with stocks with a higher/lower level of total European trading volume?
• Does BATS do better or worse on stocks with a higher/lower tick sizes?

Bringing all of this together, can we build a profile of the kinds of stocks that seem most prone to high degrees of fragmentation? Also can we say what types of stocks the different current venues seem to do best in from a market share point of view?

Knowing answers to these questions not only helps explain the current spread of market shares in different instruments, it may also provide some insight into what might happen to fragmentation if, for instance, market conditions (trading volumes /volatility) or microstructure rules (tick sizes) were to alter in the future.

Indices Covered

It would be nice to do this analysis for all European market indices; however, that would make for a rather long article! So we will focus primarily on the UK FTSE-100 index, which is the index which currently shows the highest degree of fragmentation. We will also, however show results from other indices (AEX, MIB, FTSE-250) at various points.

To get a rough idea of the current state of fragmentation of the FTSE-100 index, below is a LiquidMetrix Battlemap for the Week of 8th February 2010. The Battlemap summarises market shares, spreads, book depths and 'best price' availability for the FTSE-100 index on the different venues.

(click the image to enlarge)

At a high level:

• LSE has around 58% market share, CHIX around 27% etc.
• CHIX and LSE are closely matched in terms of spreads, book depths and 'Best Price' percentages for smaller orders 'At Touch'
• For larger order sizes (25KEUR+), LSE has an advantage over CHIX.

To see up-to-date Battlemaps with other indices and a full description of what all the numbers mean see the LiquidMetrix website www.liquidmetrix.com/LiquidMetrix/Battlemap

Now let's start drilling inside the FTSE-100 index and see what is happening on an instrument by instrument level.

Basic Breakdown of Market Shares Stock by Stock

Market Share for FTSE 100 constituents (randomly ordered)
(click the image to enlarge)

• The Average value of 60% market share for the LSE translates into a range of 50-80% at the stock by stock level.
• The MTFs also have quite significant stock by stock ranges, for instance, market shares for instruments on BATS vary from about 2% (Schroeders) to 15% (RSA).
• The data shown in this chart is averaged for the whole of January 2010. However, a similar chart for December 2009 shows that whilst there is a similar range of values, the market share for individual instruments does not vary greatly month to month. So for instance, the stock 'RSA' has had a lower than average (i.e. below 60%) LSE market share consistently week on week for a long period of time. There is something specific to this stock that is leading to a higher degree of fragmentation from the primary venue than other stocks in the index.

(click the image to enlarge)

So the question is why might different stocks have different degrees of fragmentation? To answer this we will try looking at how the market share of different stocks correlates with different characteristics of the trading activity of that stock such as total value traded, price volatility and spread leeway. We will start with total trading activity.

Breakdown By Total Trading Activity.

(click the image to enlarge)

This picture gives a good overview but is a little hard to read, so the table below shows summaries of market share for each quartile of stocks. i.e. the first quartile averages market shares for the 27 most actively traded stocks, then the next 26 etc.

(click the image to enlarge)

• LSE has a higher market share in the less actively traded stocks (62.92%).
• CHIX and BATS behave in the opposite way, i.e. they have higher shares in the more actively traded stocks.
• Interestingly, TRQX behaves similarly to the LSE i.e. it has more market share in the less heavily traded stocks. NURO also, if anything, does slightly better on the less actively traded stocks

A similar picture emerges when looking at the AEX index, the trends for Euronext, CHIX and BATS are if anything stronger than for the FTSE-100.

(click the image to enlarge)

Breakdown By Intraday Volatility

Next, we will try ordering the stocks based on their intraday 1-minute price volatility. This is the standard deviation of minutely mid price 'returns' on the primary venue expressed in basis points. Using volatility the instruments are ordered from the most volatile, Royal Bank of Scotland with a volatility of 15 BPS/minute to Reckitt Benckiser with 1.6 BPS/min.

(click the image to enlarge)

(click the image to enlarge)

This time there is no clear up/down trend in market shares. LSE does perhaps slightly better on the two extremes, i.e. for the most and least volatile stocks, but looking at all the instruments individually the market shares seem fairly randomly distributed. So, price volatility does not appear to be a good predictor of market share / fragmentation.

Breakdown by Spread Leeway

Spread leeway is a measure of the size of the smallest price increment (tick) of a stock compared with its average quoted bid/offer spread. As an example, if the tick size of a UK stock is 0.05p and its average spread is 0.125p then the spread leeway would be (0.125/0.05) - 1 = 1.5. Possible values range from 0, where the bid and offer spread of the market is always exactly one tick wide to in theory any value, though values above 10 for liquid stocks are rare.

A small spread leeway indicates it will be more difficult for a potential liquidity provider to post a new better bid or offer price inside the current best bid/offer prices. In other words there is less 'leeway' for a market maker or other liquidity provider to improve the existing best bid/offer price. It also means, in a fragmented market, it will be more common for there to be more than one venue offering the best European Bid or Offer price at the same time.

Prior to the tick size changes in the summer of 2009, spread leeways in European markets tended to be rather low. Since the summer spread leeways in some of the more liquid instruments have increased as smaller tick sizes were introduced.

BATS and Deutsche Bourse have expressed the opinion (http://www.batstrading.co.uk/resources/participant_resources/BATSEuro_Tick_Size_Paper.pdf) that the best ranges for spread leeway are around 3-10 [BATS ] or 5-19 [Deutsche Bourse]. Most stocks in the major European indices tend to have somewhat lower spread leeways than this.

As an illustration of what spread leeway means in practice, below we show two CHIX order books from FTSE 100 stocks, Liberty (on the left) has a fairly high spread leeway (3.64) and one can see small numbers of orders at each price level and a choice market (i.e. 'empty' price slots between best bid and offer). Diageo on the other hand has a low spread leeway (0.12) and one can see there are many orders at each price level and the best bid and offer are separated by a single tick.

Liberty OrderBook (high spread leeway)
(click the image to enlarge)

Diageo OrderBook (low spread leeway)
(click the image to enlarge)

So, if we look at the market shares of stocks in the FTSE-100 ordering the stocks by spread leeway, we get the chart below.

(click the image to enlarge)

• LSE does considerably better on stocks with a higher spread leeway, its market share rises from 57% to 63% as leeway increases.
• CHIX and TRQX do slightly better on stocks with lower spread leeway.
• BATS does much better on stocks with a low spread leeway. Market share almost doubles from the top to bottom percentiles. BATS shows this strong preference for low spread leeway across a number of indices including AEX (shown below, where for instance BATS has 4.49% market share for stocks with the lowest spread leeway), MIB and CAC.

(click the image to enlarge)

Conclusion

• Within indices different stocks have different levels of fragmentation that are fairly persistent over time. It seems that some venues do consistently better on certain stocks.
• Generally speaking there is most fragmentation (volume traded away from the main venue) for stocks with:
  • High Daily Value Traded
  • Low spread leeway.
• Differing price volatilities show little effect on the degree of fragmentation.

• On UK stocks TRQX's and NURO's market share actually show the same pattern as the primary (LSE) in terms of market share versus value traded i.e. they have higher market share for instruments with lower overall trading activity. But their fellow MTFs, CHIX and BATS behave in the opposite way.
• BATS does extremely well on stocks with low spread leeways.

Here we show what aspects of a stock tend to make a particular venue have a much higher(++), higher(+), lower(-) or much lower (--) market share.

Caveats and other Points

This is not a rigorous analysis. In particular, the patterns we have found are correlations not causations! For instance, it could be argued that more fragmentation leads to tighter spreads which leads to lower spread leeways. Alternatively, it could be suggested that a low spread leeway means that the depth/price advantages for the primary venues are neutralised, as there are lots of bids and offers just one tick apart and all that is left to compete on in terms of which venue to trade on is the exchange fee and so the MTFs have an advantage here that leads to more fragmentation. Our data could support either argument!

What we can say is that low spread leeways certainly correlate with higher fragmentation, which is not necessarily what many people might guess at first, especially as the drive to lower tick sizes (and hence raise spread leeways) over the summer was initiated largely by the actions of the MTFs.

Our main analysis has focussed on the more liquid stocks in the major indices. It's interesting to look further down the list of stocks quoted on MTFs. For instance the graphs below show the FTSE-250 stocks ordered by Daily Trading Volume (top) and Spread Leeway (bottom). This shows the extent to which fragmentation drops off as we reach stocks that are less actively traded and/or have high spread leeways.

(click the image to enlarge)

(click the image to enlarge)

It will be interesting to see if in the coming months any further significant fragmentation can happen in these 'smaller' stocks currently mainly traded on the primary market. Whether or not it is possible for the MTFs to capture market share in lower value traded/high(er) spread leeway stocks will determine to what extent market share fragmentation in Europe will remain limited to the few hundred most liquid stocks.

If you have any comments / suggestions about this article and in particular if you think we have got anything wrong please let us know! You can contact us at liquidmetrix@if5.com