ModelRisk 3.0

ModelRisk adds a range of advanced functions and windows to Microsoft Excel, with a focus on applications in risk analysis modeling for finance and insurance.

Designed to bring advanced risk analysis techniques to the spreadsheet user, ModelRisk works together seamlessly with any Monte Carlo simulation package for Excel (As risk analysis consultants, we use and teach these packages intensively ourselves.)

Are you new to ModelRisk? The ModelRisk PDF Quick start guide available through the Start menu provides installation instructions and a brief introduction tour for first-time users.

The main areas of ModelRisk functionality are:

• Distributions

• Probability calculations

• Fitting

• Time series modeling

• Aggregate modeling

• Copulas

Distribution objects play a central role in ModelRisk. To learn about modeling with objects, see the Modeling with objects topic.

From now on, and throughout the help, we will generally refer to functions from ModelRisk as VoseFunctions. Also, VoseDistributions, Vose Objects, etc.

Once Excel is started with the ModelRisk add-in loaded, you can access the different VoseFunctions through the ModelRisk menu, the ModelRisk toolbar or ribbon, and the context menu that appears on right-clicking a selected cell or range in Excel. All available ModelRisk functions are listed here.

Apart from using the windows you can type every VoseFunction directly into a spreadsheet cell, or multiple cells for array functions. You can directly open the window for a ModelRisk function in a spreadsheet cell by using View Formula.

The main areas of ModelRisk are briefly explained below.

Statistical distributions

In ModelRisk , over 70 distributions are included from which random values can be sampled. Random sampling is only the beginning though, each VoseDistribution (as we refer to them in the ModelRisk context) can be used and manipulated for modeling in many ways.

Probability calculations

ModelRisk has functions for calculating the joint probability density (or probability mass) f({x}) and joint cumulative probability F({x}) for a set of values {x} against a specified distribution.

These functions offer a simple way of calculating the likelihood of observations being drawn from a specified distribution, which is useful for various statistical models from distribution fitting to hypothesis testing, as well as predicting the likelihood of observing values in the future.

The functions are particularly efficient where you have a large set of values {x} as the required joint probability can be calculated in one single formula. However, the joint probability of probability density for a large set of values can quickly approach values too small for Excel to handle. Therefore ModelRisk has a parallel set of functions that return Log base 10 of the probability calculations.

Percentiles, statistical moments, parameters fitted to spreadsheet data, boundaries and many other quantities can be obtained from all VoseDistributions.

Time Series modeling

With ModelRisk you can model, fit and forecast with a number of commonly used Time Series, varying from simple AR, MA or GBM models to advanced Time Series like GARCH, EGARCH, etc.

Both multivariate and univariate versions of the most common time series are included.

Aggregate modeling

In insurance modeling, a central task is often to model the sum of a random number of claims, each of which has a random size. The sum can be modeled using Monte Carlo simulation, but advanced modeling for insurance often requires more accurate methods.

ModelRisk offers a number of tools for aggregate modeling using the most popular standard methods, including pure Monte Carlo, but also Panjer's recursive algorithm, Fast Fourier Transform...

Copulas

Copulas provide the most advanced way currently available for modeling (not necessarily linear) correlation - as opposed to most classical measures of correlation that are limited to linear correlation.

Copulas have become very popular in finance modeling, and ModelRisk has a number of Archimedean and Elliptical copulas included, as well as an algorithm for constructing a copula from spreadsheet data: we call this the empirical copula.

Modeling with objects

In modeling with ModelRisk distribution objects play a central role. It often makes logical sense to differ between a distribution as a whole, and a random value sampled from that distribution. It would be conceptually wrong and inconsistent to have the same function playing both roles. This is why ModelRisk has a VoseDistributionObject function for every distribution available.

Looking at a simple example, the sense of modeling with objects becomes clear immediately. To sample random values from a Normal(50,21) distribution - for example when doing MC simulation - you would insert this formula in a spreadsheet cell:

=VoseNormal(50,21)

On the other hand, if you need the statistical moments of that same distribution in your model, you would use

=VoseMoments(VoseNormalObject(50,21))

because of course you want the moments of the distribution, not a sample from that distribution, which would be implied by writing:

=VoseMoments(VoseNormal(50,21))

The use of Object functions avoids any ambiguity in how a function works and interacts with other functions.

Example models

There is a larger number of example spreadsheet models available (as Excel .xls files ready to use), illustrating the various ModelRisk functions and applications for modeling. To read more about the example models provided, click here.