From Market Data to Risk Factor Data

Computing Risk Factor Data has always been considered the most reactive task in the entire Financial Risk Management Process.

This section explains how [Unitized] Time-Series-Manager automates the process of converting prices, yields, rates or volatilities into  Stochastic Risk Factors on a Real-time or Snapshot (End-of-Day or Intraday) Basis.

 

[Unitized] Time-Series-Manager takes care of the entire Market-Data-Management process.

 

You provide prices or levels and it imports, scrubs, interpolates or extrapolates missing prices, converts and bootstraps yields to rates or prices, etc.

[Unitized] Time-Series-Manager then checks the quality and quantity of the new series against prior data to ensure consistency and validate Stochastic Risk Factor
computations.

[Unitized] Time-Series-Manager then proceeds to compute all requested Risk Factor Data needed by a stochastic calculation engine:

1. Mean-Zero or Drifted EWMA or plain Unitized/Standardized or Raw Returns.
2. Benchmarked and/or Indexed Weighted Returns (Country/Industry) Weighting for Asset/Obligor Correlation Computation.
3. Mean-Zero or Drifted EWMA or plain Volatilities.
4. Mean-Zero or Drifted EWMA Eigen-data for Market-Model volatility function injection.
5. Positive Definite or Semi Definite Corrector to ensure valid Correlation Matrices.
6. Mean-Zero or Drifted EWMA correlations.
7. Mean-Zero or Drifted EWMA Multivariate Regression for yield Curve mean-reversion / drift.
8. Mean-Zero or Drifted EWMA String like Forward Volatilities and correlations for Market-Model injection.
9. Higher order Series Computations. Volatility and Correlation of Volatility for Volatility Surfaces and advanced Scenarios.

 

 

 

 

EWMA: Exponentially Weighted Moving Average

The exponentially Weighted moving average is the industry standard methodology to give more weight to recent events.

The EWMA is usually defined with a Decay factor of

• 6% or a Lambda of 94% (0.94), for Daily Prices
• 3%, or a Lambda of 97% (0.97) for Monthly Prices.

 

if you defined 0% i.e. a Lambda of 100% (1.0) disables time-series weighting

The EWMA can also be defined with a Cutoff.

The EWMA MUST Not be used for BIS reporting.

 

 

 

The Cutoff

The Cutoff defines the number of data points in the Time-Series necessary to ensure results fall within a confidence band.

The Cutoff can be very useful when performing Real-Time computations as it provides consistent results with FEWER points than a no cutoff approach.