Data Trends for Investment Professionals


Cutting-edge dataset opens new forex frontiers

More than 200 attendees – representing major FX trading desks, liquidity and execution providers, and buy-side firms – joined the Quandl-CLS “New Frontiers in FX Data” webinar to learn how the new data from forex settlement giant CLS have been changing forex trading, risk management, execution, and more.

Quandl Chief Data Officer Abraham Thomas and special guest Michael Melvin, Executive Director of the Master of Finance program at UC San Diego’s Rady School of Management and former senior executive at BlackRock and Barclays Global Investors, led an engaging discussion on how we can finally use data to gain true insight into the once-opaque currency market. They discussed the unprecedented coverage, granularity, and timeliness of CLS data, as well as how firms are employing the new FX datasets to gain an edge in prop trading, execution, operations and risk management.

For those who were unable to join, here’s a recap of some of the highlights:

CLS data bring new transparency to the FX market 

Melvin, who joined UC San Diego in 2006 after more than a decade as a managing director at BlackRock, started the webinar by reviewing the major data sources that have been used historically by FX traders – including Bank for International Settlements (BIS) and central bank surveys, exchange-traded data, banks’ internal flow data, Thomson Reuters/EBS data – as well as the benefits and disadvantages of each.

While praising each major data source, Melvin acknowledged that each has weaknesses: BIS triannual surveys are irregular, while the more frequent central bank surveys are regionally limited; bank data is idiosyncratic as each institution has a different clientele and specializations; and Reuters/EBS, while a great reference for pricing, only captures a limited slice of trade volume (which, additionally, has been becoming more narrow as market players shift to other platforms).

CLS forex data, which covers over 50% of global FX transactions and is updated hourly, has overcome many of the deficiencies in other data sources, said Melvin, making it an essential tool for large forex players. “When I look at CLS, and I have some familiarity and have been doing some work with [FX] data, there’s no question to me that that’s got the best single view of the overall market: It’s the most comprehensive. It’s global. It catches the different trading venues. It catches the different participant types. There’s no particular bias in terms of regions or slice of the market.”

Traditional sources of forex data

Traditional sources of FX data and what they capture in volume and timeliness

Thomas – who was a portfolio manager and head of US bond trading at a multi-billion-dollar hedge fund prior to co-founding Quandl – reiterated the unprecedented comprehensiveness of CLS data, noting that CLS is “by far is the biggest and the most comprehensive FX settlement service in the world with half a million trades per day on behalf of 25,000 clients, totalling well over a trillion dollars of average daily volume. It’s global, and it’s fast. The sheer depth and coverage, and volume that CLS has, enables accuracy and applications that you simply can’t get from other sources – what seem to be similar data is actually not at all.”

CLS Data is the new ‘table stakes’

When CLS began selling its forex data through the Quandl platform in 2016, the first adopters were quant hedge funds and prop-desk funds looking to achieve alpha by gaining insights from what initially were novel alternative data. However, the depth and breadth of CLS data quickly led to adoption throughout the financial industry, with it becoming a ‘must have’, rather than a ‘nice to have’, for firms dealing in forex.

“Today CLS data is already being used by many world-known hedge funds and an ever-widening list of investment banks, online dealers, and sell-side desks,” Thomas noted. “That’s a pretty big adoption rate for a new dataset that only became available at the end of 2016. I think this suite of CLS data is destined to become table stakes for the FX industry.”

As CLS data has become more widely adopted – and as new datasets have been added – Thomas noted that banks, hedge funds, and others have developed multiple ways to apply the previously unknown information in vital and valuable ways.

What data are the most important for successful FX trading?

Thomas reviewed the four different CLS datasets that are currently available through Quandl, some of the ways that they are being used to improve trading performance, and why market participants have been quick to adopt the products .

“Volume and price information has been available for a long time, so what’s so special about these datasets?,” he asked rhetorically. “The answer is the unique combination of data coverage and reporting speed that CLS brings to the table. If you look at all the other commonly used sources of FX transaction data, they all have been weak on one or the other of these attributes…. CLS is quite special; it has way more coverage and it’s also faster than everything else and that’s pretty powerful.”

FX Volume FX Flow FX Price FX Forecast
Description Transaction volumes for 33 currency pairs representing 50% of global FX activity Transaction volumes aggregated by type of market participant and side of trade taken Transaction prices weighted by volume (VWAP) and by time (TWAP) Forecasted FX volume for 8 currency pairs
History From 2011 From 2012 From 2015 From 2017
Granularity Hourly aggregates Hourly aggregates 5-minute aggregates Hourly aggregates
Delivery Frequency Available in hourly or end-of-day variants Daily
Reporting Lag Within 30 minutes of the end of each hour N/A (forecast)
Coverage 33 FX pairs covering 50% of global FX volume 8 currency pairs
CLS Datasets available on Quandl and their attributes

How subscribers use CLS data

Thomas explored some of the ways that Quandl subscribers have been using the CLS datasets by showing how they inform the big questions that every market participant should ask:

What should I trade?
“This is a classic alpha-seekers question. You can use a historical investing data to design a currency-investing strategy, and we’ve seen systematic and macro funds build this in a few different ways,” Thomas said. “One idea is that volume confirms price, so you can put on continuation on currency moves on high volume, or more effectively on removal on low volume. Another idea is that actions of different participant types have different information content, and that determines if the effects are permanent or temporary.”

When should I trade?
“Volume, flow and forecast data can help answer this. We’re seeing buyside firms using CLS information very tactically in their pre-trade decision making, if they seek best execution, but we’re also seeing FX desks and liquidity providers incorporate much more formally by making it an input into their execution algorithms, and all of this is pre-trade,” said Thomas.

How well did I trade? Did I trade at the best possible time, did I get the best possible price?
“A standard approach to TCA [transaction cost analysis] is looking at the VWAP [volume-weighted average prices]. If you traded at the same average price that the rest of the market, then you probably did ok. The catch is that most TCA systems are using the VWAPs that are based on just a tiny volume of transactions – 5% of the market or less. That’s really misleading since you are getting the illusion of precision, but without accuracy,” said Thomas. “This is where the CLS dataset is especially valuable to clients, because it’s based on a much larger and more representative volume of actual market activity. You’re looking at actual trading volumes at each price. It’s indisputably the most accurate VWAP metric on the market. We are seeing all sorts of TCS vendors and internal groups on buy- and sell-side to start using this kind of data to benchmark their execution and staying on the right side of compliance.”

Who else is trading?
“You know we all love to know who else is active in the market. The motivation here is kind of two-fold: there’s an offensive or proprietary motivation when you want to position alongside smart and informed flows and position against so-called ‘dumb money’; there’s also a defensive motivation where you want to avoid being run over by large one-way activities, especially if you are on the sell side. CLS makes it possible because its updates are so fast you get the intraday data. That’s really valuable because you could see some of it in price action. But understanding who was behind the flows, there’s value there and that’s where the CLS data is it’s really useful.”

What’s my market share?
“Banks typically want to be large to improve their franchise ranking; the more market the can have the better execution they can offer. Conversely, funds want to be small, they don’t want to be too large relative to the market,” said Thomas. “In addition to the execution side of it, there’s a calibration side of it. You can estimate how much your actions moved the market. How you can estimate the things you are seeing at your desk against the whole market picture. “

How much risk am I taking?
“Modern risk systems don’t just measure, say. mark-to-market price volatility. You’re also concerned with how much liquidity is available in the market – so tracking the level and volatility of volume is an essential component of this,” said Thomas. With CLS data “you can use the volume data and float it to pre-emptively identify crowded positions which are more likely to see extreme movements. And, of course all of this applies not just to FX portfolio managers but to anyone who holds multi-currency or global assets.”

What’s next for CLS FX data?

CLS started selling their data from late 2016, and their data is distributed exclusively by Quandl. The take-up of the datasets has been exceptionally fast and widespread, reflecting value to a wide spectrum of market participants. “Today CLS data is already being used by many of the world-known hedge funds and ever-widening list of investment banks, online dealers and sell-side desks – that’s a pretty big adoption rate for a new dataset that only became available at the end of 2016,” said Thomas.

“But I don’t think it’s surprising at all. I think that the suite of data products that the CLS has released and continued to release is destined to become table stakes for the FX industry. Frankly, there’s no excuse for not using this data, especially if your competitors are using it.

If you are interested in learning more about CLS data and how they can benefit your organization, get in touch with one of Quandl’s subject matter experts.

Fix This
Created with Sketch.