Financial Analytics

These days operating financial services firms can be challenging and complex, therefore it is important to have the right tools for decision making and management. At Fuzzy Logix, our team has over 30 years combined experience in investment, commercial and consumer banking. Our experience is deep in areas such as:

• Building an origination strategy for mortgage lending that led to 100% growth in 4 years to $22B.
• Developing the strategy for portfolio optimization and re-balancing to manage one of the world’s largest portfolios of commercial loans.
• Implementing a process and the corresponding models to value thousands of companies daily to assess lending risk. These tools included the valuation of cash flow, receivables, inventory, forward contracts and organic growth.
• Building the technology platforms for modeling massive quantities of data using complex quantitative methods to simulate the behavior of financial instruments.

Based on our industry experience and customer feedback, we’ve developed in-database financial analytics that are listed below.  In addition, we build and deploy custom models.

Functions

• Financial measures – the same financial functions that are available in MS Excel
• Fixed income – bond math, volatility, convertible security analysis, various risk assessments and spread measures, MBS and ABS analysis
• Portfolio management – return, volatility and correlation analysis, CAPM and APT models, PCA, simulation of correlated returns, mean-variance optimization, CVaR based optimization, performance measures
• Derivatives – futures and options, exotics, interest rate derivatives, credit derivatives
• Time series – ARMA and ARIMA, exponential smoothing forecasting (e.g, Holt Winter’s), ARCH and GARCH, Cointegration, regime switch models

We’ve leveraged our function library to build and deliver the following types of solutions:

Solutions

• Simulate interest rate – Rather than simulating millions of interest rate paths in Matlab or c++, analysts can now do it in the database and save them to a table. These paths then could be used for various analysis or pricing purposes.
• Options pricing – Trading desks need to price various options contracts quickly and accurately, this can now be done in the database using different models, from Black Scholes to binomial trees to PDE grids.
• Value at Risk – Risk management needs to calculate VaR or CVaR for the portfolio on a daily basis, and since all the positions and market data are already in the database, they would like to run the daily VaR process all in-database, which saves a lot of time by not having to move large amount of data around.
• Sales forecasting – Financial analysts in a large retail company need to regularly forecast the sales of different products in different regions, they can use now choose between various time series models to do that. One particularly simple yet effective model is triple exponential smoothing, or Holt Winter’s model. This has enabled one of our clients to consistently forecast within 95% accuracy.