IPR enforcement is a subset of IPR protection. Measuring IPR ‘enforcement’ is a vital component of determining whether or not ‘protection is strengthened.

While each study makes important contributions in assessing the strength of IPR protection in countries, several conceptual and measurement issues are raised. The first conceptual issue is which IPR laws must be examined to gauge protection. Rapp and Rozek's measure assumes patent laws are ipso facto IPR laws. However, services increasingly are accounting for larger portions of economic growth. The value of a name or identification therefore has significant economic impact. Particularly in western countries, the economic contributions of copyrighted software production, entertainment products, and publishing are significant. To discount these two forms of IP protection would be to underestimate their economic impact. Similarly, it has been argued elsewhere (Ostergard 1999; Hettinger 1989) that nations do not treat all IP equally. Some forms of IP are more important to nations' economies than others. Using patents as a proxy for all IPR may overlook this issue. While Seyoum measures the strength of these laws individually, Sherwood's measure blends five different forms of IPR laws into the measure. As I noted earlier, the combining of various laws into one conceptual measure of IPR protection may present some theoretical and methodological difficulty.

A second conceptual issue relates to the enforcement of the laws. Rapp and Rozek and Seyoum do not include a component for enforcement in their study, though Rapp and Rozek recognize the issue (1990a, 7). The logic behind this proposition is intuitively simple. Nations may have in place the institutional structures and financial resources to enforce these laws, but others that do not have these structures and resources may not be able to enforce the laws adequately. Sherwood too recognizes this issue by including an enforceability component in his measure, but it does not significantly address the behavior of the governing regime in enforcing the intellectual property laws. The enforcement component Sherwood uses reflects an assessment of the potential to enforce IPR laws and not an assessment of the regime's actual performance.

From a measurement perspective, all three researchers assign scores based on the laws in force at the time they coded their data. Changes and amendments to the laws that may have occurred during any specific time period are not taken into account. The temporal element becomes critical particularly when using such measures to gauge the effect of IPR on foreign direct investment, technology levels and industry and economic growth. The missing temporal element means that their measure will not capture the impact of changes to these laws.

Another measurement concern is the replicability of these measures. For instance, Rapp and Rozek do not explain what constitutes the difference between a score of two and a score of three. What is the difference between "inadequate" laws and "seriously flawed" laws or the difference between "generally good laws" and laws that are "fully consistent" with the minimum standards? Without clearer criteria, an element of subjectivity is introduced, which casts doubt on the measurement's validity and reproducibility. In Seyoum's study, it is unclear as to how the raw data were reduced to a 0-3 scale. Similarly, because Sherwood's study is based preponderantly on the author's expertise, the ability to replicate the results without that vast experience is limited. Moreover, because the procedures use on-site interviews in the country, the practicality of duplicating these measures across time and across countries is limited as well.

A second concern with Sherwood's procedures is that because they are largely based on his experience, they may be more subjective than is desirable. For instance, when judging how many points to subtract for judicial independence, no set rules exist for what does constitute judicial independence. What criteria are used to subtract 5 points versus 6 points? We cannot be certain about distinctions given that the assessment is grounded in personal perception. Moreover, there seems to be some incongruity between the total points assigned to the major components and the total points that can be subtracted in the sub-categories. For instance, the enforceability component itself is worth 25 points, but the sub-category points that can be subtracted total 55 points. Is it possible that countries could have a negative score for enforceability? In theory, it is a possible scenario.

THE CONCEPT OF INTELLECTUAL PROPERTY RIGHTS PROTECTION

The measurement of IPR must incorporate both a statute component and an enforcement component. Roughly categorized, nations' laws can range from having no laws to having the strongest laws, and enforcement practices can range from having no enforcement of the laws to having strong enforcement of the laws. As Sherwood (1997) also notes, a nation can have strong laws or weak laws, but unless those laws are enforced, they are virtually non-existent in terms of the protection afforded IP owners and investors.

Law Measurement

In order to measure the strength of IPR laws, two issues must be addressed: what IPR laws and by what standards? I have settled for using just patents, copyrights and trademarks because the bulk of IP is protected under these laws (Benko 1987) and the information on cross-national patent, copyright, and trademark laws are more readily available than other types of IP laws. Following the practice of the other three studies, this study is based on the minimum criteria for IPR laws established by the US Chamber of Commerce. Using these criteria, separate code sheets for patent, copyright and trademark laws were created to conduct a content analysis of countries' IP laws.3 A number of legal sources were used to complete codesheets for each country in this study. All sources used pertained to IP statutes only, without reference to prior judicial review, trial histories or customs. The laws for seventy-six nations were coded twice with different coders utilizing the same code sheets.4 A comparison of the two codings showed intercoder agreement to be high; all inter-coder agreement scores were .90 or higher. This score reflects a 90% or higher agreement between coders on all items coded, which also indicates that these code sheets were constructed such that coding replication could be reasonably achieved. The laws for each nation were coded for three time intervals, 1988, 1991 and 1994. Three year intervals were used because the laws were not expected to change much on an annual basis. The time period was chosen because of the temporal limits of the enforcement component.

Enforcement Measurement

The enforcement score of the IPR protection measure was obtained through content analysis as well. Under the 1988 Omnibus Trade and Competitiveness Act, the United States Department of State was legislatively mandated to track the status of the IP laws and enforcement practices of the United States' trading partners. These assessments appear in the US State Department's book, Country Reports on Economic and Trade Practices. These reports are produced semiannually and began in 1989. Hence, the data set is temporally limited to the post1988 period. The early reports do not distinguish among the different types of IPR laws as the later reports do. The enforcement code sheet, which contains only one variable, was constructed to gauge a general assessment of the enforcement of IPR laws in the country. Coding of the enforcement practices was done for 1988, 1991 and 1994 (whereas the reports began in early 1989, it was assumed information gathered was for 1988 and reported in 1989). The laws for each year were coded twice by separate coders using the same code sheet with inter-coder reliability scores all over .90.

Scale scores of the IPR laws were obtained using Guttman scaling procedures (Garson 1976; MacRae 1970; McIver and Carmines 1981). Any polychotomous items were converted to dichotomous variables. A gamma matrix, measuring the strength of association among all variables, revealed that not all the variables clustered together (with a gamma score of .80 or higher) to form one scale of IPR Laws. In fact, the clustering that did occur was within each individual type of law. This finding has theoretical significance in that it confirms the idea that nations do treat these laws differently and that there is no unidimensional concept of IPR law. Each type of law must be treated distinctly from others.

The variables were then ordered and summated according to Gunman scaling procedures (Garson 1976; MacRae 1970; McIver and Carmines 1981). This procedure produced three scales: patents (0-5), copyrights (0-10) and trademarks (0-8). In these scales, zero represents no laws while the other ends of the scales represent the strongest laws. Tests of reproducibility and scalability were conducted on the scales. Coefficients of reproducibility were above .90 and all coefficients of scalability were above .60, the minimum acceptable levels for Guttman scaling.

To obtain complete IPR protection scores for each nation and each type of law, the scale score for each nation and law was multiplied by the enforcement score for each nation, which ranged from zero to four. Zero represents no enforcement of IPR laws while four represents the strongest enforcement. The multiplication of these two measures signifies that the law and the enforcement component are both necessary conditions. Because enforcement information could not be obtained for each type of law for each year, it is assumed that the enforcement practices apply to all the laws. In later years, the State Department reports assessed the enforcement of each of these types of laws, making individual enforcement scores for each type of law feasible in future studies.