Q: What is the difference between raw market data and derived market data in terms of licencing?
A: Raw market data is the original data that comes from the exchange/source. It is directly from the pricing and matching engines and emanates from the best bid/offers or executed trades depending on the policy of each exchange. Raw data has not had any calculations or changes applied to it, and hence it is called “raw”. In terms of licensing, this is the most expensive Licence Fee as it is the most valuable type of data. Derived market data is raw market data that has had a series or random calculations applied to it to. There are rules by all exchanges as to what a “derived price” means, the most important one is that the derived price cannot be reverse engineered (ie a reasonably skilled person or organisation cannot back calculate to the original raw market data).
Some exchanges, such as the London Stock Exchange and CME sometimes conduct a “pre-audit” via a questionnaire or requesting a daily file with the derived data in it; they will try to reverse engineer the derived data to its original raw format. If they cannot do this then you qualify for the derived data licence.
The licence fee for derived data is approximately 1/3 of the raw market data licence fee, though this varies for each exchange.
Q: How are exchanges, instruments and derivations managed in FinX?
A: FinX has an Administration system that clients login to. From within here they can add exchanges and instruments individually or via bulk upload. Derivations are configurable by exchange or instrument from within the Administration system.
Q: What sizes of adjustments do you add?
A: You can configure the size of adjustment by market / market segment including specifying how many decimal points it is applied over.
Q: How often do you add a price adjustment?
A: We can configure the frequency of adjustments. This is often set to 1 adjustment per second.
Q: How many instruments can FinX derive
A: We have tested FinX on 5000+ instruments with no impact on performance, even with exchanges such as BATS US that have a large number of price updates per second.
Q: What is the typical latency within FinX
A: Depending on the exchanges and instruments loaded into FinX over a 24 hour period we typically see 99.9% of derived prices <50ms and 98% <1ms.
Q: How can I demonstrate best execution if I am adding price adjustments to my prices?
A: To comply with the best execution rules firms are typically applying a common sense approach; they are putting in place best execution processes (which are explained in their terms and conditions) that demonstrate that they are not burning the client. They do this in a number of ways, for example demonstrating that they are applying asymmetrical slippage (ie that you are offering price improvement) and demonstrating that they have put in place a level of tolerance around the market price (eg +-0.4 bps) – this would include the derivation.
Q: Will my compliance department have an issue with us changing the market price through price adjustments?
A: You are offering a proprietary price which clients know about and is being made clear to clients from the beginning. Companies in this sector apply a suite of profiling tools to allow them to ensure that their product offerings and the data within them are competitive and relevant to the needs of users. You will always have access to the original raw data (the pre-derived price) should a client wish to raise a query. Also note that the derivation can be applied on an extra decimal which will leave the price displayed to the client by most firms unaffected.
Q: What if the adjustment triggers a rest order – or indeed doesn’t; the client is not getting the price they asked for?
A: The derivation can be applied on an extra decimal which will leave the price displayed to the client by most firms unaffected. Additionally in a ‘normal’ (not derived) product the same problem exists, eg:
- A client’s order gets triggered by a tick with quote A
- The order request goes to the server
- Server checks the market and quote A is already replaced by worse quote B
Furthermore, if a broker takes risk on their own book they should factor in minimal prices differences between request price and current market price.
Q: Can we adjust the price up only, down only or both up and down?
A: Technically we can apply adjustments up and/or down however it is not advisable from a regulatory standpoint. We recommend the derivation is random without an ability only to apply one side derivation ie bid only gets adjusted down and offer only gets adjusted up. Also if we ‘filter out’ one side of the price adjustments we’re essentially weakening the derivation process and potentially making it easier to reverse engineer.