Risk is pivotal and takes a central position in business management. The detection, the assessment, the mitigation of risks and eventually the insurance for the leftover risks are essential and main components of a sustainable business process.
Risk management applies on different -rather known- fields such as (but not limited to) safety, security, hygiene, nuclear technology, but also in areas as innovation, product development, production, logistics, finance and general industrial projects, it is at stake. It is one of the ultimate decision supporting methodologies, insuring maximum sustainability of the decisions.
In this course we depart from a general approach towards risk modelling based on the concept of uncertainty. Uncertainty constitutes the basis of every risk. A general and dynamic risk model is being developed based on both a theoretical framework and a lot of practical cases.
A. Measurement, error and error propagation
Measuring is a central activity in every management system or in every decision process. Nevertheless, not much attention has been paid to the phenomenon of measurement in a economic context. In this first part (A) we will deal with the issue of measurement and look into the sometimes surprising translation of known measurement concepts into economics and -in some cases- alter or at least nuance significantly the decision making; aspects such as robustness, sensitivity, repeatability are explained based on the fields they emerged from and are transposed into economics.
Besides the measurement the different aspects of error will be dealt with such as systemic and stochastic error, type I and II error and the accompanying error propagation. Each time those concepts will be focused on the economic environment.
Finally we want to link risk, information and decision analysis by exposing the underlying fundamental relations.
B. The history of Risk Management
In this short part we will deal with the historical evolution of the field of risk management, which has in fact a very recent history, boosted by changes implied by primarily World War II on the field of operations research and the intrinsically linked quantitative approach of risk.
Next the failure of existing risk management systems will be highlighted, the way in which they may inflict erroneous decisions or give a false feeling of mastering.
Most of the current risk models do not comply with elementary mathematical rules and should be completely banned of a sound and sustainable risk management system.
C The Flaw Of Averages
The problem of averages and qualitative scoring methods are discussed in this third part. Especially the effects of erroneously dealing with averages and the consequent need for dealing with probabilities are being analysed. Also the boundaries of expert opinions are discussed, a technique so often used for estimation of hard to measure variables.
In a second part of this chapter we will prove the failure of scoring methods in general and especially in risk management.
D. Process mapping
The construction of the process map and accompanying process variables seems to be a pivotal step for building an appropriate risk-analysis and -risk management system. Therefor e the important concept of BPMN will be rehearsed.
E. Risk and the generalised use of probabilities
Mainly based on the foregoing proven failure of most of the current risk management systems, the build up of appropriate, broadly applicable and quantitative risk model will be started.
In a first phase we will design a process scheme and fill out the relevant variables, including their intrinsic uncertainty range. Then we will make a link between the uncertainty on the different process variables and the overall risk of the process and or project. We will also deal with the frequentist and the subjectivist view on probabilities.
F. Estimation of variables and calibrated estimation
When only small amounts of data are available, regrettably often the case when dealing with a economic context, we often are obliged to descent to the lowest level of data quality: the estimation. In spite of the negative image estimation often has, we will see that when using the technique of "calibrated estimation", the resulting data often approach high quality surveys. The calibrated estimation will be often incorporated in the risk model.
G. Risk Assessment
In this part we will deal with the general risk models and see examples in different risk applications (within and outside of the economic context) to conclude to a generalised risk model. Armed with Monte Carlo techniques (Excel ® and Matlab®), those risk models will be calculated to determine the final risk level. The concept of Markov Chains will be used to calculate the complex risk models.
Also in these new type of models error scan prevail an d we will discuss the most prevailing errors and how to avoid them or correct them.
H risk Management and risk management system
Finally besides the risk assessment and modelling different components of the management of risks will be discussed such as the mitigation and limitation of risk, the early detection and prevention of risk. We will briefly deal with insurance of leftover risks.
As risk is per definition a dynamic phenomenon, we will look into the dynamic dimension of the risk models; using risk simulation the different measures can be assessed. A trade off between investment (cost of assessment and measurement) and leftover risk -having a negative monetary impact too- will have to be determined. All those ingredients, including the implementation of procedures for monitoring and control, constitute the generalised risk management system.