Integration of earned value and risk management using contingency reserves
This paper will demonstrate a simplified approach for integrating risk management into earned value through the use of contingency reserves. Most project managers who use earned value performance management consider risk as a somewhat separate process with risk registers, mitigation, and contingency planning. This presentation will use risk work packages with budget and schedule values entered into the project planning process. As risk events occur or not, we will describe alternate ways to measure project performance. Some real-life examples of projects are shown to prove how effective the system is in integrating risk and earned value.
Most of us know about the fundamentals of managing risk in our projects. We understand the importance of having a risk management plan so that we will have a process that continually investigates uncertainty and responds in ways that increase the probability of success in reaching our project goals, both early and late in the process.
Most of us have been trained on using a risk register to capture and track the risks identified. We might analyze and prioritize the impacts and likelihood of risks and develop mitigation plans and contingency plans to increase our likelihood of success.
However, in my own experience in managing projects, risk events and work packages appear to be related, although not directly. It was not until my experience in 2007, on behalf of the University of Texas at Dallas, when we teamed with CH2M Hill to do some training in project management with the Panama Canal Authority, that I began to see a much stronger connection.
We were preparing our risk management materials that we were going to use to teach project management to the Panama Canal Authority project managers. This was an important topic, as the personnel we would be teaching would eventually be responsible for executing the estimated US$5.25 billion expansion program to add an additional set of larger locks in Panama. There were two absolute key points that my co-teacher from CH2M Hill thought we should emphasize. The first was that risks need to be understood as they relate to specific work packages, not just the project in general. The idea was that we were to analyze the work that we had constructed in our work breakdown structure at the work package level or lower level. Hence, in risk management, the uncertainty in these elements is analyzed relative to the project objectives of scope, schedule, and cost on this work package. The risk manager would determine what “risk” events might occur that would effect a particular work package. These events would then be studied and stored in a risk register. This is a really important issue, which we emphasized over and over in our class and that makes perfect sense.
The second idea that my co-teacher brought up was the idea of integrating risk, schedule, and cost as they relate to earned value. This idea was quite new to me, but sounded like it had significant merit. After a few examples, it started to make even more sense and I have since discovered that others in the industry are using something similar. We decided to include it in our curriculum for the Panama Canal Authority and now it is taught in all my graduate classes on risk management—a way to increase understanding and apply the concept of integrating schedule and cost risks into an approach in using earned value.
There are many baselines in project management. A baseline is an approved plan for a project, plus or minus approved changes. Baselines in project management include but are not limited to the scope baseline, cost baseline, and the schedule baseline. In earned value, these are sometimes combined and referred to as the performance measurement baseline (PMB). Exhibit 1 shows the earned value parameters that are used to measure schedule and cost performance: Planned Value (PV), Actual Cost (AC), and the Earned Value (EV). The variances are the Cost Variance (CV) = EV – AC and the Schedule Variance (SV) = EV – PV. Each of the parameters will be used to describe an example and explain the concept.
Here is the problem setup. In a traditional project management implementation, risks are identified, mitigation strategies are implemented, work breakdown structure (WBS) is modified if needed, and contingencies for schedule and budget are identified.
But how do I include this information in the PMB and account for risk
events during execution, if certain risk events either occur or do not occur?
A work package has scope, budget, and cost associated with it, as well as other knowledge areas, such as quality, communications, and so forth. This paper deals with scope, budget, and cost as they relate to a work package.
Risk Work Packages
The question is then if there is uncertainty in the project, doesn’t that really mean that there is either more or less work that may be required that may take more or less cost or time to complete?
The approach here assumes that this potential uncertainty is captured by a risk work package. It is really nothing more than a work package that is capturing this uncertainty. A risk work package is associated and tracked with another work package. Hence, for that work package, it represents the uncertainty in work that might exist because of the uncertainty around this work package. Risk work packages have probability incorporated into their schedule (durations) and budget (costs).
The amount of schedule or budget contingency must be determined. For instance, if an analysis of a particular work package revealed the possible occurrence of an event that, if it occurred would cost 15 extra work days and US$300,000 to the work package, how much contingency should be associated to the work package? The approach here is to use similar concepts as used in the risk topic of computing the expected monetary value. In this case, if there is a 25% chance that the particular risk event might occur, then one should add .25*us$300,000 = US$75,000 of budget contingency and .25*15 days = 3.75 days. Following the same approach, if a particular work package had multiple uncertainties, then the amount of contingency associated with the work package would be weighted by the probability of each event occurring.
Risk Work Packages and the PMB
The approach is to link the risk work package to the associated work package and include the risk work package into PMB. Also, the duration contingency of the risk work package is included as float, either positive or negative. The work package is then treated as any other work package.
Execution and the PMB
During project execution, one of the primary goals of the project manager is not only to keep the project “on track” but to also generate forecasts at completion. I remember when I worked at Sandia National Laboratories as a “rocket engineer” in the late 70s and early 80s, prior to my experience as a project manager. One of my primary goals as a Member of Technical Staff was to create a simulation of a missile system and predict missile performance by evaluating its ability to navigate and predict how close the missile would come to “hitting” a target. In retrospect, project management is a lot like this! How can we manage a project not only at the planning stage but also during execution to forecast schedule and cost performance, even as changes occur.
Undistributed budget — Undistributed budget (UB) is a temporary holding account for the budget for authorized work that has not yet been planned in detail at the control account (CA) or summary level planning package (SLPP). Undistributed budget, or UB, is not allocated to either control accounts or to management reserves. It represents the budgets for authorized work that has yet to be planned, and as such, it should be allocated as quickly as possible. Since undistributed budgets are not time phased, UB for near term authorized work should be moved to control accounts, and far term UB should be moved to summary level planning packages. Ideally, at an integrated baseline review there should be no or very, very little budget in the undistributed budget account. Undistributed budget is part of the performance measurement baseline.
Consider the problem setup in Exhibit 2. The project consists of three regular work packages (3.1, 3.3, and 3.4) and two risk work packages (3.2 and 3.5). The total cost based line or PMB is US$865,000 which is the sum of all these work packages. The risk based contingency included in the PMB is US$90,000. In addition, and totally separate, the project plans include a US$100,000 management reserve to account for unknown unknowns. The total contract value is US$965,000. The undistributed budget is equal to the sum of the risk based contingency, which is US$90,000.
We propose two methods here for the project manager and team to decide on how to incorporate contingencies during execution.
Consider the case in which the risk event 3.2 associated with work package 3.1 does not occur. What should be done with these contingency amounts? One approach would be to incorporate these amounts into the management reservcs now. Notice that if this were to happen, something a little surprising might occur. Note that the actual costs associated with the work package were US$255,000 or US$5,000 over budget. The performance values then would show that we are over budget at this point, resulting in a fractional CPI of .98. However, in reality, we have increased our management reserves by US$50,000 or now have a total of US$150,000 in management reserves. If this contract had been managed by an owner contract that was looking for ability to capture these reserves back at some point, the owner might like this one. However, note that the contractor is actually now managing a project that is over budget—not such a great plan.
The other approach to consider in this case is to actually include the risk-based contingencies in the baseline. Hence, in this approach, as shown in Exhibit 4, would be to increase the PV (basically the cost PMB) by the risk contingency amount. In this case, the PV becomes US$300,000, which when compared with the actual costs of US$255,000 shows a CPI of 1.18, which shows that the contractor is actually under budget. No money is moved into the management reserves, and the risk based contingency stays at US$90,000. Typically, a contractor would prefer this approach because they would be reporting a more favorable result. (The National Defense Industry Association provides the industry with some great guides to EVM application guides with concepts similar to those proposed here.
Let’s consider the case in which the risk event actually occurs as shown in Exhibit 5.
In the case, in which a risk event actually occurs, the results are similar. In the case shown in Exhibit 5, risk work package 3.5 actually occurs. Following the approach used earlier, the PV at this point is US$515,000 and the actual costs incurred as a result of the occurent of 3.5 is US$60,000 compared with the US$40,000 budgeted. The actual cost of work package 3.4 was equal to the budget in this case of US$225,000. In this case, the CPI is once again even worse than before, at .95, even more over budget. The management reserves are reduced in this case by US$20,000 which is the over budget amount for the risk work page 3.5 (US$60,000 versus US$40,000).
Following the alternate strategy of including the risk work packages into the performance measurement baseline, the PV in the second month is US$565,000. The actual costs at this point are US$540,000. Using this approach, the CPI is now 1.05, indicating that the project is still under budget. Once again, a contractor might prefer this approach.
Some organizations might have real issues implementing the first approach, because the baseline actually changes during the course of the project, depending on whether a risk event occurs. Some government contractors might have real issues with a fluid PMB; hence, many times, the second approach might be more suitable.
The examples shown here are more for cost performance and do not look at risks and durations, but integrating the occurrence or not into the float values in the contract would provide similar results.
This paper has presented an approach to integrating risk based contingencies into earned value. Two methods were shown for reporting performance measures. Processes like these are very important in order to enable more reliable predictions of project outcomes.
Earned Value Management Implementation Guide (October, 2006). Defense Acquisition University, www.acc.dau.mil/CommunityBrowser.aspx?id=19557.
National Defense Industrial Association (2011) Earned Value Management Systems Application Guide, Association Program Management Systems Committee, May 4, 2011, Sec 3.4, http://www.ndia.org/Divisions/Divisions/Procurement/Documents/PMSCommittee/CommitteeDocuments/ComplementsANSI/NDIA_PMSC_Application_Guide_Rev_050411.pdf
© 2011 Unstuck Company
Originally published as a part of 2011 PMI Global Congress Proceedings – Dallas/Ft. Worth, Texas