top of page

Will Machine Learning disrupt soon Capital Markets? Definitely YES

Contrary to Blockchain, Machine Learning will TODAY disrupt the Capital Markets and not in the next 5 to 10 years.

We hear most of the time about using ML in KYC or AML, nevertheless if these are the obvious areas where investments are being made, the Capital Markets have opportunities much larger than these two.

The beauty of Machine Learning is that it is a non-invasive technology, that can be added to the existing solutions to enhance productivity and reduces cost with immediate results.

These are areas I identified (not an exhaustive list) where ML will completely change the game:

  • Processing: Analyzing the processing KPIs and identify the best workflow rules’ enhancements to reduce cost and optimize STP.

  • Corporate Actions: For the non-elective side, identify the impact of a CA and help decide what to do.

  • Market Data Cleansing: Identify stale prices, null prices, illogical prices, … clean-up and enhance the rules.

  • Agreements Management: Scan and identify relevant clauses to feed trading and back-office solutions like: Netting, Collateral thresholds, …

  • Matching: re-adjust matching sensitivity, number of tries, type of message to send, … and adapt the matching rules accordingly.

  • Reconciliation: Identify the most relevant criteria, adapt the reconciliation rules and workflow statuses accordingly.

  • Risk Management: Take the output results for Market Risk, Credit Risk and Liquidity Risk after running “What-if” simulations for Market Data and Behavior Assumptions. Identify the clients that will be most impacted by such simulations.

  • Collateral Optimization: Analyze the security balances (internal and external), compliance constraints, the cost, liquidity, … and come up with the cheapest collateral to deliver

  • Allocation Optimization: Analyze cost, availability, liquidity, strategies, … and come up with the best allocation possible.

  • Financing Venues Optimization: Analyze positions, financing venues (CCP, Central Banks, Repo, Tri-party, Internal, …), financing constraints, liquidity, … and come up with the best financing venue.

  • Regulatory Reporting: Analyze positions, trades, products, counterparts, regions, … clean, map and enrich the data, then decide what type of information to feed a specific Regulatory jurisdiction.

  • Searching for Alpha: Scraping the web and market data providers to analyze zillions of unstructured data fed into an ML engine in order to identify trends in a region, sector, stock, ... This will definitely help the portfolio manager enhance his trading decisions.

  • Hedging Strategies: Feed the positions, trades and constraints to an ML engine to analyze the topography of the portfolio. The output will be: what are the best instrument to hedge with, timeline, and propose the optimal hedging strategy that respects the constraints. And maybe execute the strategy in Real-Time.

We can continue in the areas of Balance Sheet Optimization, P&L Optimization, Liquidity Optimization, Research, Trading Strategies Optimizations, Disputes, … the opportunity is huge, and we did not yet talk about what can be done in Wealth Management, Asset Management and Corporate Banking areas.

#MachineLearning #AI #DeepLearning #BigData #BackOffice #RiskManagement #CapitalMarket #Treasury #Banking #WealthManagement #AssetManagement #collateral

bottom of page