... new processes, mergers & acquisitions). Best Data always reflects the most current conditions of companies' operational risk and contains all the information available at a given time.

... in different years, can not be compared with each other. Generally the price levels for adjusting claims change over time. Usually inflation is measured by the price development of a defined portfolio of common goods. In addition to general cost increases which affect each claim equally, liability claims are subject to a further systematic price increase, called social inflation.

Social inflation describes those price increases arising from the changing socio-economic environment. The main causes of this are increased jury awards, more liberal treatment of claims, legislated increases in benefit levels and new concepts of tort and negligence that emerge to increase insurance losses. In the Riskeeper^Toolbox these effects can be predicated into an annual price surcharge.

... the underlying circumstances are unknown in principle, the corresponding costs can be considered only on a portfolio basis. In connection with IBNR claims the term IBNER (Incurred But Not Enough Reserved) is often mentioned, which refers to future payments for known claims, that have not been considered in the current case reserves.

... probability that the expected loss can be exceeded. If, for instance, we take the 75th percentile of a simulated result and define it as a target budget, the corresponding amount would be the budget for the current risk and the probability of an insufficient budget would be limited to 25%.

Obviously, this risk measure mainly relies on the selected target percentile. The higher the target percentile, the lower the probability to exceed the corresponding value. Consequently, risks become more 'expensive' as the selected target percentile goes up. The definition of the target percentile can only be individually motivated and translates directly into the risk perception of a person or an organization.

Since the VaR is an absolute measure it has only limited use for a comparison of different structures. However, if we put the VaR in relation to the 'Mean', we receive a ratio that can be used again as a metric. Similar to the method for the 'coefficient of variation' we can then describe the inherent volatility. In order to calculate that ratio we take the difference between the 'Mean' and the selected percentile and set the result in relation to the 'Mean'. Also here a high ratio indicates a high volatility.

... to high security standards. These have been tested thoroughly and successfully within the scope of a so-called CERT project by a large German DAX company. Riskeeper meets the requirements of the international standard ISO/IEC 27001 and is certified regularly. The provider of our main server, as well as the provider of our backup server are also certified to this standard. All of our customers' data is transferred and stored only in encrypted and password-protected form.

In addition, Riskeeper^Documents allows users to securely upload data files to our server. Online access to Riskeeper^Toolbox is encrypted with the HTTPS (hypertext transfer protocol secure) standard. Each user has an individual password-protected login, which provides access only to data that is explicitly authorised to their user account.

... the data has been given to us as the process does not require any specific formats in order to proceed. Riskeeper brings together the data from different sources or formats and generates a consistent data table. During this process contradictory information, gaps or other inconsistencies are cleared up immediately by talking directly to the customer and the results are documented.

At the end of the process the claims data is made available via the Riskeeper tool set to be analysed by the customer. The consolidation and validation process is documented individually for every customer in a continually updated data manual. Software filters automatically adapt future data updates.

... risk class the appropriate distribution functions with adequate parameters are derived. The modelling of risks which have not yet occurred (for example major losses or natural catastrophe scenarios) is based on theoretical considerations. On the one hand this is carried out with the help of the individual customer's exposure information (insured values, insured risks, business activity, turnover, etc). In some areas, like liability insurance, the risks can also be estimated according to the recurrence periods of defined loss scenarios.

This exercise requires long-standing specialised knowledge and is therefore carried out entirely by Riskeeper's experts. The resulting mathematical studies are made available to the customer. This way, the complete process with all basic assumptions and underlying data is transparent, easy to follow and resilient. For the customer, however, the use of the simulation tool is absolutely intuitive.

... generic database systems without considering insurance-specific demands. Such tools cannot practically answer the specific questions which arise from the structuring of individual insurance solutions. This gap is closed by Riskeeper with decades of experience applied by experts for experts.

... login to our internet-based software tools. No additional software has to be installed on the customer's computer and no other maintenance has to be carried out locally. Riskeeper takes all the tedious and often complex effort of data processing and validation from you. This guarantees that your decisions are always based on Best Data, and alternative solutions (changed deductibles, new risk exclusions, etc.) can be evaluated in a timely and consistent manner.

... true for the choice of the most suitable distribution functions and adequate distribution parameters. This task is carried out by our experts. Therefore, the user can concentrate completely on the structuring of alternative insurance solutions and on the evaluation of simulation results. Riskeeper lets the user modify characterising simulation parameters with intuitive slide controls. Thus, for example, the effect of increasing the claim frequency or the effect of above-average occurrences of large losses can be tested directly and easily.