If you are interested in financial markets, then you may have come across references to the FIX (Financial Information eXchange) protocol. So what is it, and why is it so important?
Let’s start with a bit of history.
27 October 1986 was a pivotal day in the history of the UK; a day known as “Big Bang”. It is a day that saw massive, sweeping deregulation of the way that financial markets operated that forever changed the way stocks and shares were traded; away from the ‘in-person’ trading on the floor of the London Stock Exchange, towards electronic and phone-based trading. The trading floor fell silent almost overnight as traders instead retreated to their offices and started receiving market data on computer terminals, and trading via the telephone.
I refer to that as a pivotal moment for the UK as a whole (as opposed to just the financial heart of London), because the changes were part of a wider package of policies from Margaret Thatcher’s government to open up financial markets, which also included a programme of extensive privatisation, opening up the ability for ordinary investors to own a piece of well-known companies such as British Telecom (BT). Not only did Big Bang dramatically increase the City of London’s influence on the global stage, but it also gave the British people the ability to participate in that growth.
This is all very interesting you might say, but what does it have to do with FIX?
Well, as trading moved onto the phones and trading volumes increased, a very simple, logistical problem arose. Data needed to be re-keyed again and again into various systems, a task that was both manually intensive and error-prone. So what if there was a way that firms could send the message directly from one firm’s computers to another?
Today, there are a lot of ways we could achieve this. We could do something really simple like send a structured email or chat message, or even an API call. But this is 1990, and none of those things existed. Indeed the internet itself didn’t exist! The technology back in those days consisted of land-line telephones and fax machines (remember them?!)
And so the FIX project was designed to achieve what must have seemed almost like science fiction at the time — two computers can “talk” to each other and send/receive electronic messages that didn’t require a human to re-key the information.
The very first version of FIX was published in 1992. It was initially limited in scope to a few key workflows related to instructions to buy or sell securities, and trade notifications. The founding parties behind the protocol were Fidelity Investments and Salomon Brothers. At the time, that was groundbreaking stuff.
FIX adoption grew rapidly in the 1990s and by the end of the decade the protocol enjoyed widespread adoption, there had also been several version updates aimed at modelling additional workflows and asset classes.
By 2001, the protocol had matured to such a degree for equity products, that the version that was released in that year — FIX 4.2 — is still widely used to this day. FIX 4.4 and FIX 5.0 followed, as the protocol pushed into additional asset classes, and spread from the pre-trade (i.e. orders and trades) into post-trade activities (i.e. allocating trades between funds and even into settlement).
Somewhat remarkably, the core tenets and structures of the FIX protocol remain almost unchanged from the original, which makes it something of a living dinosaur. Its DNA pre-dated the internet, and yet it still exists today and is widely used throughout the industry. Why? Because it was designed for a specific purpose (which it still does very well), is quite simple, and is flexible enough to allow firms to customise it to their needs if required.
So what might the future hold for FIX?
I’ve heard people predict the imminent demise of FIX for over a decade now, but to date, it is very unclear what would replace it. Instead of being replaced, I believe that FIX will simply morph into something slightly different.
You see, one of the most valuable assets of the FIX project is there is now a globally agreed “dictionary” of known messages, fields and valid values. The most likely avenue for FIX going forward will be to keep this same “dictionary” but to make the messages go faster to keep up with the massive year-on-year growth in message traffic that we see. One of the most promising areas of future change is Simple Binary Encoding, which — at a very high level — means that the content of the FIX message would no longer appear as a human-readable piece of text (albeit a series of codes that require a level of understanding to read), but would instead be organised as a smaller set of non-readable characters arranged in a way to make it easier for the sending and receiving computers to read and write these “packets” of information. By making it significantly easier for them to read and write these packets, the amount of data the same hardware can handle increases sharply. (See https://dzone.com/articles/simple-binary-encoding for more information).
I hope this has been useful history on FIX. Please check out out other articles on this topic, and please leave a comment or drop me a line if you have any questions.