Pricing Skew & Flow Imbalance
Flow imbalance can be an alternative/complementary technique to Liquidity Reduction. It can improve PnL by $5-50/M or more. Flow imbalance exploits retail sentiment and herding.
FLOW IMBALAnCE SKEW Benefits
- Flow Imbalance can add $5-$50/M to the amount of money B-Book brokers make on their flow. It is a value that scales with the trade volume they do.
pricing skew benefits
- Reduce overall company risk
- Reduce NOP held at LPs to make collateralisation cheaper.
- Help attract good client flow and defend against bad client flow
- Passive risk clearance (avoid paying other people's spreads)
Using the techniques described below we can help achieve a number of business benefits.
Clean or Neutral Mid PriceS
In price formation early on in the process the objective is to find what is the neutral mid price for an instrument, the "clean mid".
Even getting to this stage, involves an at times complicated process of filtering errant quotes, venue weighting, normalisation. For instance, FX is decentralised. There is no centralised exchange for all FX spot liquidity. A large portion of the FX market remains transacted in an opaque manner. The process of arriving at a clean mid varies by asset class also.
Providers such as NCFX exist to provide an independent benchmark mid. As they say their "prices are not subject to bias or potential manipulation by those who would profit from skewing the market in their favour".
What is pricing skew?
Price skew is an adjustment to the bid/offer from a clean mid.
This can be a very useful tool. Ultimately it helps people differentiate their rate from others and hopefully attract good flow and avoid bad flow.
There are many different types of skew people apply and there can be different objectives at play when doing so.
Some of these skew objectives can be at odds with each other and it's important to consider what is being optimised for, when introducing skew and also whether at all any skew should be applied at all (e.g. where that skew may be leaked and exploited)
Example Skew Objectives
- reduce overall risk a company has. Some large brokers have gone bankrupt due to the size of their positions.
- sometimes in doing so this can leak your position to the market.
- If you are a major investment bank which dominates order flow in a currency this may not be a smart move to make.
- reduce the NOP of externalised risk to reduce collateralisation obligations.
- Early stage brokers can be forced to take on higher un hedged risk due to the collateralisation requirements of hedging their risk at LPs.
- to work passive large orders within environments. clearing risk passively with a view to attracting flow of a particular direction. Why pay spread as a taker when, you can clear the risk passively.
- this can arise from agency or internal desk flow
- macro positioning from hedge funds
- as a result of machine learning or predictive based models (defend against bad + attract good flow)
- random skew
We support methods for applying all of the above styles of skew and can help tune a pricing model to reflect optimising for a particular objective for onwards distribution.
Flow imbalance skew & PRICING B-Books
We spend a lot of time researching, developing, assessing new features for inclusion in our pricing & trading signals. We typically evaluate them and assess their efficacy against where the price will be in 30, 60 or 180 seconds time. This is a core part of our predictivity framework.
However, for retail trading books another style of analysis can be helpful. What methods can we use to not predict where the price will be in 30, 60 or 180 seconds time, but what will predict the imbalance of flow (i.e. will there be more buyers than sellers in 30, 60 or 180 seconds time).
If we're able to do that well, we can skew our rate to make it more expensive for the higher volume side and cheaper for the lower volume side and capture more spread relative to a clean mid, whilst also showing a very tight spread to customers with zero slippage.
These kind of techniques exploit the herding rather than predictive nature of retail flow
This technique makes for more PnL being generated regardless of risk management technique used.
Success Criteria
We can observe it's success via two measures
- Observing no change in the pricing model spread distribution (ensure we maintain client spreads)
- Observing increased spread capture relative to a clean mid as we charge the more abundant trade volume side more spread)
Comparative Skew Style Commentary
A large B-Book position can take weeks or months to build up from flow.
Consider the scenario, where retail sentiment has built up a large position in say USDJPY. Then they reverse their decision and want to go the other way. Risk inventory skew makes it cheaper for them to exit their position.
Flow imbalance skew is more responsive to retail herding. Risk inventory skew is not very responsive
SIMULATION Support
Simulation support for applying various skew strategies within the Value Add Simulations. We're able to fine tune a strategy ahead of risking any capital. Track the cumulative PnL of B-Booking trades at their actual price vs different skew styles and sizes to help understand the value it can bring and fine tune the configuration.
Compass Product Support
We also have the option to perform real-time skew analysis within Compass.
We can produce comparative B-Book PnL assessments of what would be the mark to market of trades if they were B-Booked with and without skew.
This is made available within the Trading Dashboard. Please contact Mahi to explore more.
Skew Sizing
There is some consideration to how much you skew. You have to be careful that you are not exposing arbitrage opportunities to clients. We have inbuilt protections for this. Note that even a single price increment can add 5-8$/M on the value of your flow.
FAQ
Server Configuration
analytics.riskReporting.adjustmentSignals.signalsForSkew = list of two signals to be used for analytical skew source. If blank will use
analytics.riskReporting.adjustmentSignals.isSkewFixedSize = if this is true we will skew in price increments. If not we'll use a spread based measure
analytics.riskReporting.adjustmentSignals.fixedSizeIncrements = number of price increments to use if fixed skews are in play
analytics.riskReporting.adjustmentSignals.normalQuantities = skew by the difference between TOB + one normal qty spread.