In the process of inputting data to the financial model, several assumptions would have been made, and the model must be sufficiently flexible to reveal the impact on the final cash flows of changes in those assumptions.
All the sensitivity drivers are typically concentrated in a summary sheet that communicates the main assumptions adopted and their effects on the bottom line cash flows. Some typical drivers identified are:
- Variation in construction costs;
- Delays in completion;
- Eventual peaks in operational costs, in any given year of the contract;
- Variations in the demand;
- Fluctuations of revenue due to performance deductions or other changes in drivers of commercial revenue;
- Changes in the debt conditions;
- Specific risks with impact in cost overruns or delays; and
- The most relevant macroeconomic assumptions such as exchange rate fluctuations.
These sensitivities, or cases, create a range of possible cash flows, depending on the chosen assumptions, and the most probable case should be identified. This is normally called the base case. See box 4.6.
The base case is the model’s expected case, determined by using the assumptions that the project team consider are most likely to occur. The financial results from the base case should be better than those from conservative scenarios, but worse than those from upside cases.
In order to create a representative base case that reflects a realistic scenario, all inputs and assumptions must be defined. Starting with this base case, other possible scenarios could be defined and analyzed. These scenarios may vary depending on the objectives that are sought – some scenarios allow for the structuring of the PPP (for example, the nature of any government contribution, the payment mechanism, the contract term, and so on), other scenarios test or structure the level of risk (for instance, the demand level), and others are used to assess commercial feasibility.
BOX 4.6: Key Aspects of Sensitivities
Financial modelling is full of uncertainties. This is an inevitable consequence of the attempt to predict future events. The more the financial model is able to recognize this weakness, transforming uncertainties into variables within reasonable ranges, the more it will effectively translate the future reality.
Thus, a good financial model is not the one that produces one single, arguably precise number, but the one that identifies the main drivers and presents some reasonable ranges within which the model’s conclusions still hold true. These ranges are known as sensitivities and they mainly aim to assess the robustness of the project business or financial plan to material changes in their key assumptions.
The financial model is a very interactive tool in the sense that the model’s conclusions enable analysis that leads to a change in the assumptions. In turn, new conclusions are reached and another set of assumptions can be changed. Some of these interactions have been specifically mentioned above, but this is a very general and essential characteristic of the task of financial modelling: it is fundamentally circular!
It should be noted that the financial model will be used in some projects as a tool in bid evaluation, and also as a necessary support tool for managing the contract (while in this latter case this PPP Certification Guide considers it more appropriate to use the private partner´s financial model).
 The Feasibility Study Guideline for Public Private Partnership Projects, by the University Transportation Centre for Alabama (2010) presents examples of structuring different cases to test the commercial feasibility of PPP projects. The Municipality of Rio PPP Guide: Screening, Appraisal and Auctioning of PPPs (Volume 2, Section III) presents a practical approach to the design of cases.
 A “scenario” differs from a “sensitivity” in that the former represents a complete new or different “case” (different than the Base Case) with new values defined for one or more key variables duly backed up and supported by a specific analysis (for example, the “optimistic demand scenario”), while the latter only represents a switch in the value of one or more variables so as to observe the impact of that specific variable on the Financial Model key performance indicators (KPIs).
 See “Financial model issues” in chapter 4.8.1.