The WIA (Work and Income Act) and the Ziektewet (Sickness Benefits Act) regulate additional income for an employee who has become ill. RiskCo’s WIA model calculates how much additional income the employee receives and what this means for his or her total income. This can be calculated at any moment during the period in which the employee receives additional income.Read More
Short-rate interest models
Short-rate interest models use stochastic differential comparisons for the short-rate in order to model the interest term structures. The short-rate interest models of Vasicek, Cox-Ingersoll-Ross (CIR), Ho-Lee and Hull-White have been implemented. It is possible to calibrate these models to data to predict and simulate interest curves.
Interest Rates Generator
This software module applies the Monte Carlo approach. The module, based on the model developed by The Academy and the Society of Actuaries (SAO), provides ready-made generated interests. These generated interests are fit on the basis of the Nelson Siegel 2-factor model. This model assumes that the parameters are already known; estimating the parameters is done outside of this model.
Dynamic Nelson-Siegel models
The Nelson-Siegel model uses four parameters to fit the complete interest term structure. By modelling three of these parameters in time as time series, new interest term structures can be predicted. According to Diebold and Li* this gives better long-term predictions than do the usual models.
*Diebold, F.X. en Li, C., Forecasting the term structure of government bond yields, Journal of Econometrics 130 (2006) 337-364.
Minimization and optimization
This model uses the Nelder-Mead-method. a frequently used minimization method for functions with several variables. Because this method does not need the derivative from the function, it is easy to use.
This model consists of various functions for regression analysis: linear regression, multidimensional linear regression, and non-linear regression. Regression analysis is used to model dependencies between variables. It can also be used to make forecasts.
Interpolation is used to find values for variables that are not in a given dataset. It is possible to model just one or several variables. In addition to linear interpolation the model also contains cubic spline interpolation , so that a number of boundary conditions can be adjusted.
Time Series Analysis
This model contains diverse methods for time series analysis. In addition to several statistical methods, the model also has methods to fit autoregressive models (AR and VAR) and to use these models to make forecasts. Both one-dimensional and multidimensional time series can be modeled and forecast.
This model contains the calculations needed to use matrices. For example, the model contains the calculation of the determinant, the inverse, matrix multiplications, vector calculations and the concordance matrix. The model serves as a basis and can be used again in other models.
RiskCo has implemented full “one push of a button reporting systems” for defined benefit pension funds. The challenge involved in these types of reports is that the regulator wants to know what the liability is on eg the 1st of January 2012 with the knowledge of the 1st of april 2012 as well as the value of the liability is on eg the 1st of January 2012 with the knowledge of the 1st of January 2013. RiskCo’s solution consists of chains of calculation models that produce results to be consolidated into reports. In order to make reports flexible, the typical business intelligence tool can usually be used that is already employed by the pension fund.Read More