Nostalgia. Ah yes, things were always better before. Well, sometimes, but perhaps not always. The key thing which was different in the past, is of course us: we were younger. Perhaps a yearning for past isn’t so much for a world that existed before, but for our younger selves?
When it comes to markets, it isn’t always the case that the past was better too. If you were trying to execute a foreign exchange spot trade at the start of my career fifteen years ago, it’s quite likely it would have been a more costly process. Today, spreads have compressed significantly. The problem is that FX markets have become much more complicated. There are so many different types of venues, many different algos etc. The market you see is not the same as what others see (ok, maybe that hasn’t quite changed!) – it is complicated to navigate your own unique path to lower transaction costs.
Today, we have a lot more computing power to crunch your trade data combined with market data to understand the true cost of your trading through transaction cost analysis. The key question though is who should pay for transaction cost analysis?
You could get this service for “free”, paid by your bank. Of course when it comes to trading, you and your bank have different objectives..! You want to minimise your trading costs, and banks want to generate P&L from your trades. By all means having TCA provided by banks can be helpful, and definitely a starting point, but it shouldn’t be the only TCA you look at. Furthermore it might be challenging to reconcile the TCA provided by many banks, without spending a lot of time aggregating and parsing the data, given it’s in very different formats.
The alternative is for you to pay for TCA, by using an external vendor. You pay them on an ongoing basis for an external TCA service. This way it’s more independent. However, it involves giving out your full trade data to that external party. In many instances, you might want to keep this data private for a number of reasons. You are also basically locked into your provider. If you stop paying the service is turned off and all that integration, legal work etc., needs to be done all over again. That’s a cost in both time and money.
What if you want to do your own TCA internally to avoid this lock in and also to truly customise your TCA? This is big undertaking and is often expensive. But you get to define exactly which TCA metrics you want. You can keep your trade data internal too within your own firewall, which is much easier from a compliance perspective. On the flip side, there’s also the issue of ongoing maintenance, which you have to deal with.
However, there’s another way: get Cuemacro’s Python based FX tcapy library on an enterprise licence. You get all the source code so it’s 100% transparent (and yes we actually like our clients to see how we’ve implemented things – we’re proud of our code!). You can run tcapy on your own hardware (or AWS) and all it’s dependencies are open source too. You can define any metric you want in Python code, you aren’t at the mercy of your vendor to implement it (or not). For initial clients we are also offering the license as a one off cost, which is much cheaper than the years of coding hours it’ll take to write it yourself. There’s no implicit payment by giving us all your trade data, like Facebook etc. We take none of your data. You’ll also tie up your quants and developers to reinvent the codebase, when they could instead be spending that time and money, to extend and customise tcapy.
If you’d like technical support/updates there is a much lower annual support cost to help with maintenance, but this is entirely optional. You can continue using the library without any further costs if you want – there is no lock in, you can keep the library and use for as long as you want. We are also looking to create a cut down open source version (albeit without all the high performance features and GUI etc.) if we can find a sponsor who really wants to make TCA more accessible.
Plus, you can keep your trade data internal (we never see any of your trade data), which is easier to sign off from a compliance perspective. You own the process: there’s no conflict of interest too with any other party, and you are free to define metrics that you want, which suit your trading.
Finance needs to become more transparent, when it comes to TCA. If you can’t see the source code to see how metrics are calculated, can’t modify what market data is used in your benchmark, the whole TCA process is not in your hands. It’s time it was! If you’re interested in making TCA transparent and letting your firm own the TCA process, drop us a message and we can arrange a tcapy demo for your firm.