Cost Risk Analysis with Risk Drivers
A second method to determine estimate uncertainty is the risk driver approach. It quantifies the probabilities of risks from the risk register occurring, and what their effects on WBS element costs will be if they do. With this approach, a probability distribution of the risk impact on the cost—expressed as a multiplicative factor—is estimated and the risks are assigned to specific WBS elements. If a risk does not occur in an iteration of the Monte Carlo simulation, then the cost does not change for that element. Using this method, cost risk is estimated from the identified program risks and their expected effects on WBS elements.
The risk driver approach focuses on risks and their contribution to cost contingency, as well as on risk mitigation. Analysts can assign a risk to multiple WBS elements and the costs of some elements can be influenced by multiple risks. In this way, the risk driver method is used to examine how various risks may affect the program cost estimate. Table 13 shows a subset of possible risks associated with the construction of the UAV.
Table 13: Air Vehicle Risks, Likelihood, and Cost Effects
| Risk | Likelihood of risk | Minimum | Most likely | Maximum |
|---|---|---|---|---|
| Material is late or defective | 80% | 100% | 105% | 130% |
| Complex airframe producibility will lead to increased manufacturing time | 25 | 100 | 110 | 125 |
| GFI deliveries are delayed | 20 | 90 | 100 | 115 |
| Hiring and retention are affected by changing economy | 15 | 95 | 110 | 135 |
Source: GAO. | GAO-20-195G
According to table 13, the most likely risk in the program involves timely delivery of quality material and the least likely risk involves a changing economy that may affect the contractor’s ability to hire and retain employees at the prevailing wage.
In addition to including discrete threats and opportunities, analysts can include risks that represent ambiguity about the future. The existence of these ambiguities is known (that is, their likelihood is 100 percent) but their effects are uncertain. For example, the productivity of labor will affect the cost of many elements in some way, but whether the overall effect is positive (an opportunity) or negative (a threat) is uncertain and depends in part on bias in the point estimate. Cost analysts can also include some element of general uncertainty. For example, to account for natural variability around each of the element estimates, analysts can include an uncertainty to represent a global estimating error. Table 14 identifies some uncertainties for the UAV program.
Table 14: Air Vehicle Uncertainty and Cost Effects
| Uncertainty | Likelihood | Minimum | Most likely | Maximum |
|---|---|---|---|---|
| Labor productivity | 100% | 97% | 100% | 105% |
| Cost estimating errors | 100 | 97 | 100 | 106 |
| Software sizing errors | 100 | 76 | 100 | 140 |
Source: GAO. | GAO-20-195G
With the risk driver method, the risks and uncertainties shown in tables 13 and 14 will appear as factors that multiply the costs of the elements they are assigned to if they occur in the iteration. Once the risks and uncertainties are assigned to WBS elements, a simulation is run. The results may be similar to those in figure 16.
Figure 16: Air Vehicle Cost Cumulative Probability Distribution from a Risk Driver Simulation
In this instance, there is about a 1 percent probability that the program will cost $2.47 billion or less. If the program manager chose the 80th percentile, the estimated program cost would be $3.08 billion, including an amount of $610 million for contingency. In this case, the risk driver method has caused a wider spread of uncertainty compared to the three-point method. By combining the two methods, three-point estimates may be used to represent bias and uncertainty, while risk drivers are used to represent identifiable risk events that may be mitigated. When combining methods, it is important that risks not be double counted in the simulation.