Galen Stops takes a look at the new data service launched by CLS Group.
In September, CLS announced the launch of its new data service, CLSData. Speaking to Profit & Loss at the time of
the launch, the firm’s CEO, David Puth, explained that CLS was “now entering the data market space in a very
Since its launch in 2002, CLS has recorded every single transaction submitted to it, and considering that an average daily volume (ADV) of over $1 trillion is processed by CLS each day, this represents a massive and rather unique data set.
“If you compare this to some of the trading platforms out there, it’s a far larger data set, and because of our central place in the market, we think our data provides a very unique view. It’s the only consolidated view of the market, outside of the BIS survey,” says Robert Franolic, head of data and analytics at CLS.
In 2016, CLS began providing volume data by currency pair and products, which was made available via a platform provided by Quandl, a firm that specialises in delivering financial, economic and alternative datasets to investment professionals.
“CLS is traditionally a trade settlement organisation that was never really in the data business. They realised recently that they are holding data that’s of value to professional investors and they want to bring that to market as a product, but they lacked a capacity to do that in their organisation. So they were looking for a partner to be their data arm, and that’s exactly what we do. We work with firms that aren’t traditionally in the business of selling data,” explains Tammer Kamel, CEO and cofounder of Quandl.
With this new launch, CLS has developed the volume data offering by providing the data on an intraday hourly basis. It’s also now offering subscribers to the service VWAP and TWAP data; again both will be delivered hourly by currency pair and then by order flow.
Franolic predicts that these three different data sets can and will be used for a variety of use cases by market participants.
“If we talk about the volume data, one of the use cases there is as an input into developing trading and algo strategies,” he says. “One thing that you see a lot of in other asset classes is firms using the volume data in combination with the price data. The adage is that volume confirms price, so if the price moves up on a significant volume then that price is more likely to remain high, but if it moves up on thin volume then it’s more likely to reverse.”
In particular, this volume data could be of significant benefit to firms utilising momentum or mean reversion trading strategies.
“In most asset classes, traders are not working in the dark. The equity market is a classic example of this, everyone knows the volume all the time and it makes everyone a better trader, the same is true in futures markets. But the FX market has always been opaque because it’s OTC. A few big players have a good idea of volume, but most market participants are flying blind, they don’t know what the volume is.
“So this is like turning on a light switch, if you’re trading a momentum strategy you need to know the volume, because you’re that much more certain the trend will continue if it’s backed by high volume, whereas if you’re doing a mean reversion strategy, your conviction that there will be a reversion is higher when volume is low. It’s that simple,” says Kamel.
Another potential use case for this volume data could be as an input to improve transaction cost analysis (TCA) being conducted by trading firms.
“You can certainly use volume data to help inform that type of analysis. When volumes are thin, then obviously you might expect prices to move against you when you’re trading, so volume is clearly a denominator in the calculation of the price impact of executing a particular trade,” says Franolic.
He adds: “These are more buy side use cases for this data, but there’s also value in it for sell side firms. Some of these firms might use it to understand when there’s liquidity in the market and adjust their spreads based on the volume information, tightening when there’s more volume and widening when volumes are thinner. They might also want to use it for fairly simple purposes, such as understanding how much penetration they have in a particular currency pair. Our volume data can be a denominator in that.”
In terms of the pricing data, Franolic suggests that firms will primarily be using this for TCA, comparing the prices that they got against CLS data.
“Because we have this broad coverage and our data is matched executed trades, our VWAP is probably the most robust one you can get. It is probably the best measure of what the price was in an interval,” he comments.
Complementary Data Sets
CLS has been storing all the trades submitted to it in a data warehouse since 2002 and had already made investments geared towards capturing, cleaning and structuring the data that it receives in a way that makes it easy to report. However, until recently the data was only fully captured on an end-of-day basis and in order to get this data to a point where it would be useful to the market, CLS had to upgrade its infrastructure to allow the capture and cleaning up of data on an intraday basis.
Although he insists that there is considerable value in providing all this data on an hourly basis, Franolic concedes that the faster CLS can deliver the data, the more value it will have and says that the firm is therefore looking for ways to increase the frequency by which it can deliver this data.
There is, of course, a natural limit to how quickly CLS can provide this market data considering that FX trades only come to it after the execution of a trade has taken place. This limit means that CLS is unlikely to be on a collision course with the major FX trading venues, for whom selling market data accounts for an increasingly large percentage of their revenue. Indeed, Franolic says that because of this, the CLS data is likely to prove complementary to the data being provided by the platforms, which see a smaller slice of the market, but are able to provide data on a more frequent basis.
“For certain use cases, their data will be better, for others ours will be. Together, I think that they’re complementary,” he comments.
“You can always get super high frequency price data,” adds Kamel. “Based on this CLS data, you could execute strategies where you would hold the trade for several hours and if we increased the frequency, then you could use it to execute strategies with a shorter hold time. Even providing this volume information on an hourly basis adds a lot of value compared to what FX market participants are used to right now.”
Franolic says that the current focus is on building the CLSData subscriber base, but looking ahead, he reveals that the firm is looking for ways to enrich its current data set and develop new products.
“Part of my role is developing new data products and ways in which we can apply analytical techniques, machine learning and other tools to develop newer data products,” he says.