One of the core objectives of modelling is to provide the information that decision-makers will need. There are several aspects to this:
In addition to direct decision-makers, there may be multiple other stakeholders or influencers: Each of these people or groups may have different objectives, criteria, values, priorities and constraints. Further, some of these may be hidden or not made explicit for political or social reasons, or only become apparent when the “hard reality” of a decision strikes.
The modelling process can provide a framework which allows for various interests to be brought together in an integrated way. Therefore, it is important to try to identify the relevant actors or decision influencers, and to elicit their interests, objectives and decision-criteria as explicitly as possible.
One of the core objectives of modelling is to provide the information that a decision-maker will need: This requires that that the modeller should identify and evaluate the decision criteria, and establish which can be included within a quantitative model, and how non-modelled (out-of-scope) criteria should best be dealt with.
Of course, many decisions involve multiple criteria, and some of these may not always be able to be evaluated quantitatively or objectively. The model should aim to evaluate as many as are realistic; however, the model results will be an input into a decision-making process that may take other factors into account, and may involve making judgements in relation to moral, political or other issues.
In general, the process to evaluate a specific decision involves some assessment of its advantages and disadvantages. Of course, there are many specific ways that these can be expressed and compared (or weighed) against each other, resulting in many different approaches. This section begins with a brief introduction to decision-making criteria in general and then mentions some key criteria used in economic modelling.
In general, the starting point for identifying possible decision criteria is based on the general concept of “advantages-disadvantages”. This can be expressed in a variety of similar ways, including:
These types of analysis resemble each other closely, and when applied to a specific situation, each could give the same results as the other. For example, the advantages identified could be categorised into those based on strengths and those based on opportunities
In practice, each method may result in slightly different items being identified, simply due to the various focusses that the wording of each implies. Indeed, all methods could be considered in parallel, as a way to try to make this important phase of the analysis as complete as possible.
Given a set of decision criteria, one would need a way to compare or evaluate them. The following mentions some key approaches.
The use of “gut feel” means that a decision is taken with no or limited analysis, other than intuition or what one feels (or judges) to be right.
Of course, there are several positive points about such approaches:
However, in situations which are complex or with large and irreversible investments to be made, this approach is likely to be sub-optimal at best. The modelling process can help to create insight and overcome an excessive focus on gut feel; it can bring an alternative perspective and help to balance the use of judgment and analysis – when both are used correctly together, the result is generally a more powerful and robust decision process.
These methods are based on the application of some dominant principle, such as:
The key aspect of these approaches is that the method chosen relates to a single criteria which is considered to dominate all other considerations.
For example, if using cost-minimisation, one would choose the cheaper of two contractors (or training courses, or investment projects), even if the more expensive one could have greater benefits. Similarly, a risk-minimisation approach would choose a course, contractor or project that has the the smallest risk or downside outcome, even if the other options could have far greater benefits or upside returns in almost all scenarios.
Whilst a principles-based method could be used to choose a decision which “maximises the benefit”, such an approach would not take into account the cost of doing so. Thus, often it makes sense to use integrated approaches, where multiple items can be compared or traded-off within a single framework. The decision is taken based on finding the optimum balance.
These approaches are inherently quantitative, in order to be able to compare and integrate the items. Much economic analysis is conducted in this way. For example, the concept of “benefit” could be translated into sales revenue, and “cost” could be operating costs or investments, and so on. The cost-benefit analysis could measure profit (net benefit), cash flow, and so on.
The particular decision is then made by choosing the one which maximises an objective (such as profit) subject to some constraints (such as total investment). For example, when buying a car, one could fix a cost budget, and search for the best car that one can find for that money. Alternatively, one could first identify the cars that meet our requirements and then buy the one which has the lowest cost.
There are many possible ways to evaluate a decision from an economic perspective. In principle, most of these involve using quantitative (integrated) methods. These include:
Often, it is also required to also perform a sensitivity analysis of such items, to see the effect of different scenarios or assumptions on these measures. These are all covered in detail in later materials.
A bias is a general outlook, preference or inclination that is held by an individual or group, and which may influence the way that the person or group evaluates information that is provided to it. In other words, a bias is a lens through which information is viewed, and which may result in some types of information being given an inappropriate weight or importance (range from important information being discarded or ignored to unimportant factors being given prominence). Therefore biases act as a form of filter, and so “objective” information that is provided to decision-makers (such as the numerical evaluation of decision-criteria using a model) will be interpreted or used by them in accordance with their biases.
The effect of biases can be partly reduced if one is aware of them, and if there is a process to recognise potential biases and make them more transparent.
Biases can broadly be classified into three categories: motivational, cognitive, and structural biases.
Motivational biases are those which result from personal or group incentives, corporate politics, and general situations where there may not be full alignment between overall corporate goals and the goals of participants in the decision process.
Cognitive biases are those which are of a psychological or evolutionary nature. There are many sub-categories of these, including:
Cognitive biases can result in insufficient analysis being conducted, or the results ignored or de-emphasised.
Structural biases are those created by the process or methodology used. For example, and over simplified methodology may overlook certain key elements or risks and biases the results in a favorable way.
Modelling can help to reduce biases by increasing transparency, and by creating the capability to challenge assumptions with logic and robustness, and to test hypothesis or the effect of alternate assumptions.