Modelling

As the leading risk/reward management firm, Z/Yen helps clients to organise information and make decisions. Modelling is often key to both organising information and understanding the consequences of decisions. Z/Yen has had to develop, and keep developing, modelling skills in line with client-driven problem-solving. Z/Yen's modelling work, which flows from Z/Yen's proprietary risk/reward methodology:

  • Is based on sound stochastic principles - the shape of a distribution function is far more insightful than a single number;
  • Uses both quantitative and qualitative information in the framework, e.g. broker behaviour patterns or perceived rates of fraud detection can be as important as easily quantified financial numbers;
  • Helps clients learn the dynamics of their environment, rather than simply provide numbers for justification.

One of our financial services clients asked Z/Yen to help them estimate the reserves necessary for a new electronic product. Z/Yen worked with the client to establish the fundamental specification of the product, the parameters which were open to change (e.g. distribution, pricing, membership) and the marketplace in which it would be traded. Z/Yen then built a large computer model of the market and stress-tested it against a variety of scenarios. The new product appeared to be particularly vulnerable to client unfamiliarity. This led to further market-testing and product re-design in order to understand better the buyer's perceptions and to keep product familiarity high. Z/Yen's principal objective, the calculation of reserve levels, led to some suggestions, such as the use of re-insurance, which resulted in required reserves being a tenth of the original business plan forecasts. The model, its manuals and trained staff, were left to re-calculate market sensitivities quarterly.

Z/Yen draws upon a wide variety of skills in modelling - financial, scientific, computing, market research and behavioural science. Z/Yen's team of accountants, actuaries, scientists, researchers, psychologists, marketeers and even some `rocket scientists' uses a number of software packages from simulation modelling tools such as Simul8, Optima! or Instrata to risk management packages such as Crystal Ball, Predict! or @Risk, and genetic modelling or neural network software such as Evolver or Neuralyst, as well as our own risk/reward prediction engine PropheZy. Z/Yen's modelling work takes a number of forms:

  • Model specification and design, where other parties wish to develop the model themselves;
  • Model critique, testing or audit, where a client needs to validate a model with an outside organisation;
  • Model building, where Z/Yen assembles a team, manages the project and delivers a working model and its first set of results to a client.

Z/Yen's modelling clients come with a wide range of requirements and a variety of backgrounds - government organisations modelling fraud rates, financial organisations seeking model critiques of Value at Risk or portfolio models, venture capitalists looking for model validation, large institutions requiring a simulation of new products, processing houses looking for models to help them reduce costs, scientific organisations seeking to optimise research portfolios or service organisations seeking stochastic pricing models they can share with their clients. Z/Yen people have the background and the desire to show how stochastic models can markedly improve decision-making. Z/Yen has the resources to deliver models which help clients learn.