Executing Statistics Calculation 

  1. Choose Financial accounting ® Treasury ® Market risk ® Tools ® Statistics calculator .

    2. The initial screen for calculating volatilities and correlations appears.

  2. Choose a Risk hierarchy.
  3. To restrict the calculation to one node on the risk hierarchy, choose Nodes.
  4. Choose a Reference currency.
  5. Choose a Volatility type.
  6. Choose a Correlation type.

The correlation type and volatility type are assigned a statistics type in Customizing. This describes the statistical calculation type and sets the parameters and basic estimation functions (sample range, confidence level, etc.).
  1. Choose a Calendar.. In doing this, all historical days are selected for which market data will be read.
  2. Choose the Start date (the date from which the market data should be imported).
  3. Choose a Holding period (in days). For this period, the volatilities and correlations are estimated. More precisely, by comparing the market data at the beginning of the holding period with the market data at the end, the system determines the market price changes.

If the holding period which you want to use for an evaluation is different to the holding period with which the statistic data is estimated, an adjustment is made. You define the type of adjustment in the Retention period calculation category parameter, which you use to define a statistics type. These settings are made in Customizing under Treasury ® Market data admin. ® Master data ® Define stat. type At present, statistic data based on price change logarithms is transformed using the root method. Statistic data based on relative or absolute price changes is transformed into the required retention period of the evaluation using interpolation.

Historical days are determined starting from the Start of history date and continuing backwards.

Say the beginning of the history is 03 March 1997 (Monday). With a holding period of one day, the market price change is the change from 28 February 1997 to 03 March 1997. The second price change is figured to be the change from 27 February to 28 February, and so on. The historical period reveals the number of examined market price changes.
  1. Enter the error tolerance (No. of missing rates allowed) to control the reaction of the system when data is missing.

    2. If there is no market data for historical dates (no quotation, no delivery via datafeed) the system has a replacement strategy. This involves using market rates from further in the past. Since this leads to a distorted statistical picture, you can use the error tolerance to determine the maximum number of such replacements allowed in an historical time sequence.

  2. You can make the following alternative calculations:
  1. With the indicator Test run (batch mode only), you can control whether the data is written to the database.
  2. Choose execute.

Result

Volatilities and correlations are calculated. The calculated values are output as a list, and in an update run they are stored in the database tables.