one important measure: the amount of counterfeit currency in circulation. This method uses a simplified model to map the flow of counterfeit currency from production to removal.

It is important to note that the committee does not suggest that this particular elementary model should be used to make policy decisions about adopting future features. Instead, the model is presented only as a device to provide a possible rationale for determining which features might be—in some sense—evaluated as being “better” than others. Moreover, for this (or any alternative) model to be used as an instrument for determining policy, it would have to be appropriately validated. As discussed below, this would require either estimating its parameters from existing field or experimental data, or running experiments specifically designed with this goal in mind. In any event, a truly useful model would have to be capable of considering sets of features together, since features do interact, in addition to overcoming the limitations of this quite-basic model. The model is presented as an aid in explaining the origins of the evaluation process adopted by the committee.

This said, the counterfeiting threat can be represented as a basic flow system, as shown in Figure 2-2, in which the boxes represent repositories of counterfeit currency (containing amounts measured in units of dollars). Counterfeit notes progress through the system from production, through stockpiling and, by means of a passing event, into circulation. These flows are indicated by the arrows in Figure 2-2. At each stage, disruptions to the system can be introduced by removing counterfeit notes or by deterring their production—shown as arrows pointing to the ovals that represent repositories of counterfeit currency removed from the system. The variables used to represent this system are defined in Table 2-5.

The topmost ovals in Figure 2-2 are special cases—representing counterfeits that are never actually made because of technology stoppers, or “blocks,” or owing to the difficulty of feature simulation. As such, these ovals are “virtual” repositories of counterfeits; no actual notes reside there. The effect of these virtual repositories is captured in ρ, the counterfeit production rate, which reflects the effects of technology stoppers or the difficulty of feature simulation for the purpose of feature evaluation. In particular, any increased flow to the virtual repositories will necessarily decrease the production rate, ρ.

In order to best analyze this system, the parameters for each feature and for each class of counterfeiter should be quantified. While the values of some of these are known, such as u (the amount seized by the Secret Service), many others remain unknown. However, a strong case can be made for estimating or measuring these variables, since they provide a basis for quantitative analysis of the effect of new features on counterfeit deterrence, as discussed below. Even for variables that are hard to measure in the real world—that is, s, the fraction of successful pass attempts—controlled experiments could provide some insight that would aid in modeling the system. With the aid of a fully characterized flow model, such as