Steps Ahead

Demystifying the complex: how to gain insight from FICC derivatives transaction data

The pandemic has accelerated electronification of fixed income trading for many banks, and with that has come a renewed focus on innovation to help them keep pace with market structure changes. In an increasingly data-centric market, nowhere has innovation been pronounced than in the field of data analytics.

Banks are increasing their technology spend with a laser focus on solutions that will help them achieve a competitive advantage – and this means turning data into actionable intelligence. In a recent survey of c-suite executives at Fortune 1000 companies, 99% reported they had invested in data initiatives and 96% stated they had achieved successful outcomes from Big Data and AI.

But implementing a successful FICC data analytics regime means being able to gain insight from the simplest to the most complex of transactions – normalised, aggregated and delivered on one screen.

If deployed effectively, analytics is an area where banks can develop a true edge by uncovering opportunities and gaining a better understanding on when, where and how to trade. Against this backdrop, it’s unsurprising that more banks are deploying AI and machine learning to optimise workflows and enable trading desks to allocate more resource to higher-value trades, especially those that aren’t suited for electronification.

To deliver results at speed and with better ROI, many banks are partnering with service providers and specialist fintech companies to speed up digitisation.

Mosaic Smart Data’s cutting-edge analytics tools support fixed income and derivatives data, empowering banks to enhance their trading and risk management strategies. Mosaic is partnered with the European Space Agency (ESA) to apply some of the world’s most advanced data analytics models to the problems banks face in financial markets.

The following examples demonstrate how the MSX and MSX360 platforms can be used to analyse liquidity in difficult-to-price areas of the market, measure best execution, value securities and understand risk more effectively.

Real-time analytics

Real-time risk approximation: Mosaic’s ability to approximate real-time risk on derivatives

Ex 1:- Basis swaps (SONIA vs 6m Libor) or Treasury vs Futures spread sensitivity on real-time portfolios

Fig 1: Intraday net risk positions on Bonds vs Futures spread trades

Interpretation: Going into COVID, Banks/BKR/CLR were maintaining a net short basis (treasury vs futures), and are now following AMs views on basis movement views.

Ex 2: Real-time risk transfer on Compression swaps trades

Fig 2: Intraday net risk position on Compression USD swaps

Interpretation: There is a significant risk transfer on 21st of July in $ Compression swaps, driven primarily by Bank A

Profitability: Mosaic’s ability to calculate markouts/ market impact on derivatives (including off-market swaps/ packages etc) helps to identify optimal hedge timings and profitable clients

Fig 3: Profitability curves across different dimensions, overlaid with different hedge strategies (red plot)

MSX360+ Insights into client flow

Mosaic’s real-time analytics capabilities enable the data scientists to inject insights on client activities into the platform.

Ex: looking at USD switch flows over the last month and comparing to recent position for detecting shifts in behaviour

Fig 4a: Current month spread flow
Fig 4b: Previous months spread flow

MSX360 Alerts: In the current month, asset managers have been trading bear steepness in USD, which is a shift from previous month where they were trading bull flatteners

Anomaly detection on complex packaged products.

Ex: shown below is a Treasury vs Futures spread butterfly product handled by Mosaic.

Fig 5a: Structure of a Spread butterfly
Fig 5b: MSX360 algorithmic detection of butterfly strategy

MSX 360: asset managers are running positive butterfly on Treasury vs Futures butterfly, primarily using 10Y as the belly and 5Y/30Y as the wings



Our team will book in a analytics consultation with you and show you around the MSX360 platform.

Industry and academic partnership

FX Liquidity benchmark: Innovative ML based modelling of liquidity.

Developed in partnership with Oxford University (Prof Rama Cont), MUFG and CLS, the model creates a intra-day liquidity benchmark for G10 currency pairs. It also marks the shift in liquidity for a currency pair pre- and post-COVID.

The model is a multi-factor model and combines several factors like market impact, autocorrelation of returns, volume imbalace, order book depth – to name a few.

The sample chart below shows the optimal execution time for a £10mil GBP/USD back on 16th June, 2020 – along with how the liquidity had shifted after the onset of the pandemic.

Fig 6: GBP/USD liquidity benchmark asof 16th July, 2020

ML driven market impact model (Partnership with European Space Agency)

We are currently partnerning with European Space Agency to develop a predictive market impact model. The model looks at various market impact profile curves (both pre and post) and identifies patterns in these curves. This is then used for predictive modelling. Example business use-cases include:

  • reduce market impact
  • client classification
Fig 7: Algorithmic detection of different strategies (like momentum/reversion) in different asset classes like swaps, cash, futures





  • Issuer: G10 Government, Supra or Agency and denominated in G10 currency
  • Currency: G10 (USD, CAD, EUR, GBP, SEK, DKK, NOK, JPY, AUD, NZD)
  • Type: Bullet bonds, linkers, FRNs, strips, bills, notes, zero coupon bonds, Multi-leg (bond vs. bond)


  • Currency: G10 (USD, CAD, EUR, GBP, SEK, DKK, NOK, JPY, AUD, NZD)
  • Type: Vanilla Swaps, IMM Swaps, OIS Swaps vs FF, Forward Starting, MACs, Multi-legs (swap vs. swap, swap vs. futures, swap vs. bond, bond vs. future, Agency spreads)



  • Issuer: Sovereign Governments
    • EM LATAM = MX, CL, PE, BR, AR, CR, CO, DO, PE, UY, VE
    • EM CEEMEA = PL, HU, CZ, RO, ZA, RU, TR, IL, LT, LV, CY, EE, EG, GH, IS, KE, KZ, NG, RS, SI, SK, UA, VN, ZM
    • EM EMASIA = IN, SG, ID, MY, TH, KR, TW, CN, LK, PH
  • Currency: Local Currency or G10
  • Type: Bullet bonds, linkers, FRNs, strips, bills, notes, zero coupon bonds, Multi-leg (bond vs. bond)



  • Currency: All currencies and precious metals
  • Type: Spot, Forwards, Swaps, NDFs, NDF Swaps


  • Currency: G10
  • Type: Options (future release)