Contingency in Capital Cost Estimating

Contingency in Cost Estimating

Owners add a contingency to a project to deal with project overruns due to uncertainties. A large number of projects are finished with costs and schedule overruns. This is why, as a cost estimator on a capital project, it is very important to understand how to use contingency and how to evaluate it.

Contingency is defined as funds added to the cost estimate to compensate for the cost estimate inaccuracies caused by uncertainties in the project definition. To define it further, the contingency is added to a capital cost estimate to cover certain uncertainties and risks.

AACE International defines a contingency as “An amount added to an estimate to allow for items, conditions or events for which the state, occurrence and/or effect is uncertain and that experience shows will likely result, in the aggregate, in an additional cost”. In a way, contingency is similar to allowances. The difference between the two is that contingency allows for unknown but expected, whereas allowances account for known and expected but what cannot be measured.

There is a large number of white papers written on the subject of contingency. An excellent discussion on the topic of contingency is included in the September/October issue of Pharmaceutical Engineering. Click here for the link to the document.

For the purpose of this article, we will look into practical ways to assign contingency to a cost estimate based on my experience in the field.

Types of contingency on a capital project

1. Project Contingency. This contingency is intended to cover the unexpected costs for a likelihood of 50/50 for the project to be over budget.
2. Owner’s Reserve. This contingency is intended to cover a lower probability of project overrun, 80/20, for example, and it is dependent on the owner’s approach to risk.

Contingency calculation methods

The Association for the Advancement of Cost Engineering gives us the following methods for calculation contingency:

1. Expert judgment,
2. Predetermined guidelines,
3. Simulation analysis, and
4. Parametric modelling.

Another common classification of contingency calculation methodologies is:

1. Deterministic methods, and
2. Probabilistic methods

In practice, the deterministic method is very common, from my experience. This methodology is used by owners who do not want to apply a risk assessment on the project for various reasons (time, funds, etc.). In the deterministic method category, we can mention:

1. A pre-set percentage for each type of cost estimate. This is a very handy methodology for owners/consultants with extensive experience in certain projects. For example, a class 5 cost estimate will be assigned a 40% contingency, whereas a class 2 cost estimate will include a 15% contingency. We can also calculate contingency as a weighted average. To do so, we apply contingency to various parts of the project scope based on the specific scope definition. I like this method because it is more detailed and considers risks and uncertainties with each WBS group.
2. Expert judgment. The difference between this methodology and the one above is that, in this instance, an expert, or a group of experts in the field of risk analysis and the specific type of project, determines the contingency amount. The expert or expert group can be a third party or an in-house resource.

Probabilistic methods for determining the contingency on a cost estimate are less common than the deterministic method. Probabilistic methods are usually employed by owners/contractors involved in very complex projects. Using this method, a range of probable costs is assigned to individual project items using the level of confidence against different values of the project cost. The overall contingency is determined based on an overall confidence level that the owner would like to have on the project.

The probabilistic methods can be broken down in:

1. Simulation and,
2. Non-simulation

The probabilistic simulation methods are the most expensive, and they involve the use of sophisticated simulation software. A very commonly used software is Monte Carlo, used for range estimating. Monte Carlo software can also be used for the integrated method of cost and schedule simulation. Most project owners/contractors will use the integrated method as it is more comprehensive than the range estimating, where only the risks associated with the costs are considered. For this simulation method to be successful and as reliable as it can get, the participating panel of experts should include a large representation of all involved in the project design, financing, management, future operations group, field experts, etc.

The non-simulation methods involve using analytical methods to calculate the risk and contingency without using simulation software. In this category, we have the following methods:

1. Probability Tree. This method uses a diagrammatic representation of possible outcomes based on the identified risks. When the risk register is very large, this method becomes unreliable.
2. First-Order Second-Moment. These are approximate methods to calculate the standard deviation of complex functions.
3. Expected Value.  In this method, all the risks in the risk register are assigned a value that would be spent if the risk actually takes place. This method is common among contractors who bid on large, complex projects. It is easy to complete, and there is almost no cost to it other than the time of the evaluation panel and the risk registry owner.
4. Program Evaluation and Review Technique. This method is used in project management and uses a three-point evaluation: most optimistic, most likely and most pessimistic. The approximation of the three points is then represented in a Beta distribution.
5. Parametric Estimation. A relationship between a cost overrun and an input which can be risk factors, is created using this method. The relationship is created using historical data. This method uses parameters which are risk factors such as project definition, size of the project, complexity, existing historical data, etc. This method uses linear relationships between dependent (cost increase) and independent variables (the risk factors).
6. Analytical Hierarchy Process. This method is about using a simple risk analysis by including subjectivity, experience, and knowledge to assess the cost effects of each risk (or set of risk factors) on specific work packages (discipline). A targeted cost is determined as a result of the risk analysis. The delta between the targeted cost and the base cost represents the contingency.
7. Reference Class Forecasting. This method uses historical data from completed projects. It classifies the projects into groups based on type, size, complexity, etc. It then analyzes the cost overrun data. The method assumes that similar projects will behave the same in the future as they behaved in the past from a budgeting point of view. A contingency amount is assigned based on past project performance for similar projects.

In my work, I often use non-simulation methods for calculating contingency. Not necessarily exactly as they are described, but rather a combination of a couple of methods for each project. I also took part in a few Monte Carlo simulations.

My takeaway from all that there is to be said about contingency is:

• Not to be confused with accuracy.
• A capital cost estimate should always show the cost ranges (the minimum possible cost and the maximum possible cost). The cost ranges are applied to the total cost, inclusive of contingency.
• Contingency calculation and the methodologies used are very specific to each project. There is an abundance of methodologies to calculate contingency. Deciding on which method is the most appropriate is a matter of time, money and available expertise.

What is your experience with contingency calculation? Please leave your comments below!

Sharing this article with your social network would be appreciated!

Thank you for reading!