CHAPTER 4 DAMAGES VALUATIONS OF TRADE SECRETS
4.5 Value Estimation of Trade Secrets in EEA Cases: Survey of Models and Methods
As noted earlier, the models are grouped under three general titles: Income, Cost and Market models, as in Zwillinger and Genetski (2000.) The Income Models include Unjust Enrichment, Lost Profits and Reasonable Royalty and are based on discounted cash flows. The Cost models, Replacement Costs, R&D Costs and Actual Damages, are based on the costs associated with the trade secret. Finally, the Market models are based on the fair market value of the trade secret.
The intangible nature of trade secrets complicates its valuation in criminal cases.
With stolen physical property, the use of Fair Market Value presents a relatively
222 Moral hazard is “the tendency of agents who are insured to behave more recklessly because of their insurance cover.” Pearson Education (2005) Available from
http://wps.pearsoned.co.uk/wps/media/objects/2499/2559960/glossary/glossary.html
223 Bishop (1982), p. 29.
simple argument for the value of the stolen property (Green et al, 2000.) The stolen property is valued at the value a willing buyer and a willing seller would agree upon. However, with trade secrets, as the owner is not actually deprived of the property, the valuation becomes more complex. Often, the estimation is based not on the Fair Market Value of the trade secret, but on the losses suffered by the victim (as in Zwillinger and Genetski, 2000.) As the fair market value of the Trade Secret in EEA cases cannot be measured directly (due to information asymmetries, lack of a willing seller and a willing buyer etc.), loss estimation (damages) often provides a close substitute (Green et al, 2000.) In this section, the estimation methods used in EEA cases are described and their diverse results are then analysed in Chapter 5. In the EEA cases reported here, the valuations have been obtained from a combination of court documents, media reports and academic papers.
Due to the diverse nature of the valuations, a high and low value (as in Carr and Gorman, 2001) of the trade secrets was collected for the EEA cases. These values are reflected in Table 4-‐3. Due to the lack of a consistent official source for these estimates, estimates are only available for one third of the EEA cases.
Table 4-‐3: Summary Statistics of Value Estimates of Trade Secrets Value Estimates for Stolen trade secrets
EEA Cases 1996-2008
Low Estimate High Estimate
Mean $4,471,000 $26,338,000
Median $576,000 $1,916,000
Minimum $6,000 $12,000
Maximum $39,376,000 $271,736,000
Standard
Deviation 9,540,000 66,651,000
Number of
Cases 29 (1 outlier removed224)
Expressed in 2008 values
Note that the range of estimates from $6,000 to $272M is large. The mean of the
“Low” estimates, which are more conservative than their “High” counterparts, is
$4.5M albeit with a relatively high standard deviation of 9.5M. These values suggest that the applications of the valuation models results in highly diverse values for the same trade secret. Chapter 5 will provide a statistical analysis of these valuations.
4.5.1 Points of View and Concepts of Time in the Models
The criminal status of EEA cases means that these methods are calculated from the point of view of either the defendant or the victim, as shown in Table 4-‐4 below, which is an extension of Table 4-‐1 above to ‘point of view’ and ‘state.’ As Glick et al (2003) argue from a legal and fairness perspective, “Where the trade secret’s value has been destroyed, the secret should be valued from the owner’s perspective. … Conversely, the value of the secret to the misappropriator is usually the appropriate measure where the secret has not been destroyed.”225 With the exception of Reasonable Royalty and Fair Market Value, the methods applied in EEA cases present one-‐sided valuations. This creates an inherent
224 The outlier removed is discussed in further detail in Chapter 6 on page 242.
225 Glick et al (2003), p. 337.
conflict between the prosecution’s valuations and those of the defence, as Becker (1968) discusses.226 It also points further to the complexity of these valuation models, which gives rise to the highly diverse estimates for trade secrets.
These valuation models incorporate various concepts of time and states. The dominant use of time is the number of time periods, t, which is used to account for the time value of money (Slottje, 2006.) The other concepts of time deal with the states as they relate to the timing of the theft of the trade secret, as modelled in Werden et al (2000.) These concepts play a large role in the application of Income Models.
One factor to consider in assessing valuation models is the number of time periods, t. Unlike patents, which have a fixed time limit, trade secrets have no legally defined expiration date (Scotchmer, 2004.) The courts must, instead, consider alternate forms of estimating t. One is to estimate the time it would have taken the defendant to develop the trade secret independently. This can be called the “head start” or “lead time” advantage.227 Another is to estimate when the trade secret would have become public knowledge without the theft, or when the trade secret became obsolete (e.g. replaced by a new technology), as in
Slottje et al (2006.) These methods are imprecise but necessary in the absence of a legally defined time period.
The valuation models also incorporate different concepts of the state (status) that deal with the sequence of time periods associated with the theft of the trade secret. One concept looks at the status quo in sequence as two time periods that are state 1 – before the theft, and state 2 – after the theft, as in Hall and Lazear (1994.) This real concept of time, which I will refer to as “factual” time, looks at the actual value of variables before and after the theft. For example, in Unjust
226 Becker (1968) develops the argument for the differing values of criminals, victims and society for criminal cases in general.
227 Glick et al (2003), p. 338, defines the head start period as “the period of time required to eliminate any competitive advantage obtained by the misappropriator” and allows defendants to limit damages based on the amount of time it would have taken them to discover the trade secrets via legitimate methods.
Enrichment, the thief has per unit production costs of c1 before the theft and costs of c2 after the theft (Slottje, 2006.)
An alternate concept of time, found in the Lost Profits Doctrine, again divides time into two states: state 1 – “but for”, and state 2 – “actual” or the status quo.
This is the state of the variables with the theft (actual) and “but for” the theft (what the state of variables would have been if the theft had not occurred.) I will refer to this theoretical concept of time periods as the “counterfactual” concept of states as it compares actual events to the theoretically modelled events
described in the “but for” analysis. In Reasonable Royalty and Fair Market Value, the counterfactual state is used to purport hypothetical agreements between defendant and victim. As Slottje (2006) notes, the use of counterfactual time can be problematic:
The methodology incorporates fantasy and flexibility – fantasy because it requires a court to imagine what warring parties would have agreed to as willing negotiators; flexibility because it speaks of negotiations as of the time infringement began, yet permits and often requires a court to look to events that occurred thereafter and that could not have been known to or predicted by the hypothesized negotiators.228
The use of factual time allows for a valuation that compares actual values to actual values. However, the use of counterfactual time periods requires the theoretical prediction of the counterfactual (profits without the theft) and
compares this to the factual (profits with the theft.) This presents a challenge for courts as the use of counterfactual states gives less weight to objective,
evidenced values and more weight to subjective, theoretical models, as discussed in Green et al (2000.) From the perspective of justice, the use of factual states may be more appealing due to its more objective nature.
The definitions of time and states in the valuations of trade secrets influence the valuation of the trade secret and lack certainty. Despite these weaknesses, the models are appealing in that they incorporate the long-‐standing principles of the time value of money (Block and Hirt, 2002.)
228 Slottje (2006), p. 85.
Table 4-‐4 presents a summary of the characteristics of the estimation groups.
Table 4-4: Characteristics of Estimation Models
Value Estimation of trade secrets: Models and their Characteristics
Group Model
Point of
View State
Unjust Enrichment Defendant Factual
Lost Profits Victim Counterfactual
Income Models
Reasonable Royalty Both Counterfactual
R&D Victim Factual
Replacement Costs Defendant Factual or Counterfactual Cost Models
Actual Damages Victim Factual
Market Models Market Models Both Counterfactual
The following section details these models and their use in EEA cases. The foundations of the methods are based in economic reasoning, however, as
Chapter 5 will discuss, can result in a wide range of valuations for the same trade secrets.