Re-thinking Speed in Financial Markets: Part 3
The IEX decision has re-ignited a polarizing debate about speed being good or bad for markets. What is the alternative to those two camps? If we think of speed as a proxy for price certainty, how would our perception of its role in modern market structure change?
We are entering a new chapter in speed where information about speed will become more valuable than speed itself.
In trading, two important attributes that market participants seek are liquidity and best execution. The ability to provide both of these is critically dependent on price certainty. We define price certainty as the probability of being able to access the price advertised by the market. Fast venues with fast participants generally operate with greater price certainty. However, slow participants on fast venues will have less price certainty compared to fast participants. Price certainty on slow venues will be less than that on fast venues. Fast participants on slow venues will have less price certainty advantage over slow participants as all participants have less price certainty. If only it was that simple.
In US Equities Markets we have a mix of fast participants and slow participants, fast venues and now slow venues (IEX), all interacting electronically based on a complex set of rules (e.g. order protection rule) and sophisticated order types — immediate (IOC) orders and now “extended life” orders (Nasdaq). We have co-location, speedbumps, random order processing, and now clock-synchronized multi-venue order submission (see US patent application14/451,356). All techniques used to leverage or de-leverage the impact of speed, or more accurately differences in speed, on execution outcome. Add to this, random traffic effects creating random delays that complicate prediction of execution outcome and quality. This is why I believe it will prove futile to regulate speed. It is impossible to eliminate the effects of speed from the equation unless you ban computers completely. What to do?
At the risk of simplifying, the faster you can receive market price and the faster you can respond to that price, the more likely you will be able to hit that price. The more a price is delayed and the slower you are in responding to a price, then the more likely the price is no longer available. In general, a fresh price is better than a stale price. Fresh prices are more certain. The “age” of a price (i.e. the difference in time between when a price was first created and when you received it) becomes a critically important metric to account for the potential effect of speed on execution quality.
Think of “age” as a risk coefficient to be used with “price”. It quantifies the probability of the price existing when you look to execute on it. We are all familiar with the “best before” date that typically accompanies perishable products. When we go to our local grocer we intuitively ask “is this today’s bread?”, “is this meat freshly cut?”. We always check the “best before date” on packaged perishables. In exactly the same manner, advertised market prices are perishable and need to be consumed before they go stale.
This then leads to the maxim I call the speed rule:
The Speed Rule — Decide to trade, once you know when the price was made.
Once we know the age of a price, we have transparency. We can then reasonably judge the validity of the price on offer and make an informed decision to trade or not trade based on our prior experiences of trading with prices of a certain age on certain venues. This can be applied uniformly across all venues irrespective of their speed. Of course, you may choose to weight your order routing decision to a fast venue or a slow venue depending on your execution experience and overall strategy. The point here is that once you observe the speed rule, then we no longer have to prescribe a specific speed or speed limit before we decide to trade. We simply deal with the speed environment that exists, and make informed decisions to trade that make the best of the situation. Just like how we deal with the weather.
So what is the catch? Unfortunately, there is one. We cannot accurately determine the age of a price on today’s markets because we don’t know time with sufficient granularity and accuracy. To determine the age of a price we need two things:
- A timestamp of when the price was generated
- A timestamp of when the price was received
The age is the difference between both timestamps. The problem is that timestamps used today by venues and participants lack the necessary granularity and are not synchronized to a common time reference across all venues and participants. Therefore, the age calculation is inaccurate. Sometimes the calculations using today’s venue timestamps are grossly inaccurate e.g. negative age — implying that the future happened before the past. This means we have to guess or deduce the real price of the market. This lack of precision synchronized time is the urgent problem to be solved by regulators. If we can’t tell time in the machine traded world, and can’t see exactly what is happening, we will never have a market structure that everyone accepts to be transparent, fair and robust.