Determinants of royalty terms : Evidence from

3 downloads 0 Views 581KB Size Report
simple average of the values of 48 industries in the columns II. The latter average controls ... monopoly rights and the share of cross-licensing provision increased slightly. It declined from ... Figure 2 shows the frequency of patents in the licensing .... all payments are made by a running royalty and if it does not affect the sales.
(Preliminary) Determinants of royalty terms : Evidence from technology import contracts of Japan† Sadao Nagaoka*, Hitotsubashi University August 2003 Abstract This paper examines the changes of royalty terms of technology import contracts of Japan in 1980s and 1990s, based on the sector-level panel data. Major findings are the following: (1) The share of high royalty contracts increased significantly in 1980s and 1990s, which, however, mainly reflected the change of the composition of imported technologies, especially the increasing share of software licenses. (2) According to our estimation, controlling industry fixed effects, an exclusivity provision and trademark licensing are accompanied with more high royalty contracts. It also shows that the increase of the incidence of cross-licensing contracts and the increase of R&D intensity of the industry tended to be associated with high royalty contracts. (3) After controlling these factors, the share of high royalty contracts increased significantly in the latter part of 1990s in the industries for which intellectual property rights are important and the restriction of sales territory to Japan is frequently imposed in the contracts. They seem to reflect the impact of stronger IPR policy of Japan since the middle part of 1990s. Key words: licensing, intellectual property rights, technology import JEL classification: O34, F23 *) Institute of Innovation Research, Hitotsubashi University, 2-1 Naka Kunitachi Tokyo 186 Japan, Fax 81-425-80-8410, Email [email protected]

†I

would like to thank Professors Akira Goto and Akiya Nagata for allowing me to use

their survey results for constructing an index of the effectiveness of patent protection. 1

I. Introduction Information on licensing is expected to provide important clues on some of the basic issues of intellectual property rights (IPRs). For examples, how extensively IPRs are shared among firms? Are geographic or firm boundaries important for technology trade? How important are the effects of the stronger protection of IPRs on the price of technology?

Do stronger IPRs

promote or reduce licensing? Theory alone often cannot provide a final answer to most of these questionsi. Despite of such importance, empirical studies on licensing are scarce, although there are important exceptions such as Taylor and Silberston (1973), Caves, Crookell and Killing (1983), Davidson and McFeridge (1984), Anand and Khanna (2000) and Arora, Fosfuri and Gambardella ( 2001). This paper is an attempt to fill some gap. Specifically, it addresses the following two questions. First, it analyzes how royalty rates are related to basic contract characteristics such as exclusive right, cross licensing provision and the structure of IPRs specified in contracts. Second it analyzes whether stronger IPR policy in Japan in recent years may have increased royaltiesii. Since the data we can use is industry-level data, we can analyze only the aggregate characteristics of licensing contracts. This prevents us from controlling firm-level factors on royalty terms. On the other hand, the survey we use is very comprehensive, based on the reporting system which was compulsory. In addition, it has a long time horizon, so that we can use panel estimation in investigating the above questions. The paper consists of the following. Section II provides a brief

2

overview of the major features of the technology import contracts of Japan. Section II provides an analytical framework, and section III provides the framework of empirical analysis. Section IV provides empirical results and Section V concludes and discusses the implications. II. Brief overview of licensing contracts for technology import II.1 Number of contracts, price and non-price conditions The number of technology import contracts was mostly from 2,000 to 3,000 contracts annually, during the period from 1971 to 1998 (see Figure 1). The number of contracts dropped sharply in 1998, due to the reduction of reporting requirements iii . There has been a significant change in the composition of technologies which are imported. Most importantly, the share of software technology has increased significantly over time. It accounted for only less than 10% in 1981, but for about a half in 1990s (See Table A-1 in the appendix for details). The change of the aggregate picture of contract characteristics over time has also been closely related to this structural change. (Figure 1) Let us turn to the change of contract characteristics in both price and non-price conditions over the last two decades (see Table 1). Table 1 shows the average of these conditions for all contracts in the columns I and the simple average of the values of 48 industries in the columns II. The latter average controls the structural change of industry in terms of the number of contracts due to its fixed weight. In terms of the first average, 94.5% of the contracts are onerous, and 62.2% of the contracts have royalty provisions,

3

26.5% of the royalty contracts have high royalty rates (8% or more of sales) and 61.3% of the contracts have initial payment provisions. (Table 1) The average price of imported technology seems to have increased during these two decades, although its extent is not large once we control the change of the composition of imported technologies. The share of high royalty contracts increased significantly from 13.4 % of the period of 1981-84 to 26.5% of the period of 1995-98. However, the increase is much less substantial, once we control the structural change. It increased only from 13.4% of the period of 1981-84 to 14.8% of the period of 1995-98, according to the second average. The share of the contracts with initial payments and the share of onerous contracts also increased from 57.6% to 60.4% and from 93.3% to 95.9% respectively according to the second average. As for non-price conditions (exclusive territorial provisions (or monopoly provisions), and cross-licensing), the share of the contracts with monopoly rights and the share of cross-licensing provision increased slightly. It declined from 51.0% of the period of 1981-1984 to 29.4% in the most recent period in the total average, and declined from 52.0% to 41.6% according to the second average. Cross-licensing became more prevalent according to the second average. II.2 Intellectual property rights specified in the contracts Table 2 shows the structure of intellectual property rights (IPRs) specified in the licensing contracts over the last two decades. Know-how is most frequently specified in the contracts. Close to 50% of the contracts specified

4

only know-how in these periods. More than 70% of the contracts specified know-how as one of IPRs in 1995-98. Thus, although it is often pointed out that know-how is difficult to trade due to information asymmetryiv, it is traded. Trademark has become more often specified. The frequency of the contracts with only trademarks increased significantly from 5.9% in 1981-84 to 19.4% in 1995-98. 32.4% of the contracts covered trademark in the period of 1995-98. Patents have become less specified, although this is entirely due to the structural change in the composition of imported technologies. The frequency of the contracts with only patents declined from 11.7 % in 1981-84 to 9.0% in 1995-98. 24.0 % of the contracts specified patents in the 1995-98 period. (Table 2) There exists significant heterogeneity in the importance of each intellectual property rights specified in the contracts across industry. Patents are more important in R&D intensive industry, with a major exception of computer industry. Figure 2 shows the frequency of patents in the licensing contracts in the most R&D intensive eleven sectors for the two periods: from 1981 to 1986 and from 1995 to 1998. As shown in this figure, patents were specified for more than 40% of the contracts in 8 out of 11 sectors for 1995-98 period and they became more important in 8 out of 11 sectors between the two periods. The frequency increased from 6.9 % to 9.6% in computer industry and from 47.6% to 69.2% in pharmaceutical industry. Figure 3 compares the changes of the frequencies of patents in licensing contracts with respect to software and hardware technologies (all

5

technologies excluding software and trademark). This shows the sharply rising tendency of software patents, beginning in 1996. (Figure 2 and 3) II.3 Licensing conditions in two-high tech industries This section takes a closer look at the developments of licensing conditions in the most R&D intensive two sectors: pharmaceutical industry and computer industries. Figure 4 and 5 show the changes over time of eight characteristics of licensing contracts in terms of the share in the total industry contracts: onerous contract, initial payments, high royalty contracts (the contracts with 8% or more royalty rates in all royalty contracts), exclusive marketing rights, cross-licensing, patents, trademarks, and know-how. The shares of high royalty contracts are twice more frequent in the computer industry than in pharmaceutical industries. The monopoly marketing rights are significantly less frequent in the computer industry. Know-how is most frequently traded in the computer industry, while patents are also very important in the pharmaceutical industry. (Figure 4 and 5) When we look at the changes over time, the price of technology in the pharmaceutical industry became more expensive: in particular, the share of contracts with high royalty rates increased significantly. As for the structure of IPR, patents became more important over time, while trademarks became less important. In contrast, there are no clear signs for the appreciation of price in the computer industry. This is so despite of the fact that patents have become more important in this industry, as seen in the above

6

subsection. III. Theoretical framework We adopt the following theoretical framework for the determination of royalty in licensing contracts. A licensor (X) and a licensee (Y) negotiate a licensing contract based on the Nash bargaining framework. We first consider the case of a bilateral monopoly and later discuss the extension. We denote the threat point for such negotiation by ( Π MX , Π YM ), where Π MX and Π YM are the levels of the profits of the licensee and the licensee respectively, when the license does not take place (Subscript M indicates that technology is used only by the licensor). If licensing takes place, the licensor and the licensee obtains the profit Π DX and Π YD respectively, both of which are the profits before the transfer of licensing fees (Subscript D indicates that technology is licensed). The Nash bargaining solution gives the following equilibrium payment (R) by Y to X:

R = (Π MX − Π YM ) / 2 + (Π YD − Π DX ) / 2

(1)

If we assume that the licensor X does not use its technology in the market of the licensee, this can be simplified into the following: R = −Π YM / 2 + Π YD / 2

(2)

The payment is high when the licensee Y can realize high profit using the licensed technology (a higher Π YD ), and when he can realize only low profit without licensing (a lower Π YM ). The profit which Y can realize using the licensed technology depends on the possibility of imitation. For simplicity, we assume that Y cannot make 7

a positive profit if imitation takes place, the probability of which is given by γ. If we denote the profit of Y with no imitation by Π YD, N , we have Π YD = (1 − γ )Π YD , N

(3)

In case Y cannot obtain a license, he can still realize some profit by doing its own R&D. In particular, it is assumed that Π YM = (1 − δ )Π YD − αRD X ,

(4)

where RD X is the amount of R&D expenditure, which X spent to invent that technology. The parameter δ

and α indicates the difficulty of

inventing- around. Combining the above three equations, we have

R = (1 / 2){δ (1 − γ )Π YD , N + αRD X }

(5)

Equation (5) suggests the following three determinants of the price of technology (R). First, R is high when the licensed technology enables the licensee to realize high profit. It may happen when the technology is advanced relative to prevailing technology, the technology licensed is comprehensive or highly developed so that its commercialization does not need a large additional expenditure. Second, if the technology to be licensed requires a large R&D expenditure for being invented around, the royalty rate would be high. Third, R would increase with stronger IPR protection, due both to smaller vertical competition and to smaller horizontal competition. Since stronger protection of IPRs makes inventing-around difficult and expensive (i.e. δ and α are large), it shifts the threat point in favor of a licensorv. In addition, it makes imitation more difficult ( γ is low), so that the

8

infringement by a third party is reduced. Both of these changes increase the payment for the technology. Let us consider the effect of ex-ante competition among licensees. In the case where the entire bargaining surplus belongs to a licensor due to such competition, we have the following:

R = δ (1 − γ )Π YD , N + αRD X

(6)

In this case, the payment (R ) would be twice as much as in the case of the absence of competition among licensees. A good indicator of ex-ante competition is exclusivity provision, since, if there is strong ex-ante competition among licensees, a licensee would not be willing to invest in commercializing the technology unless he were guaranteed by a licensor for an exclusive entry. Thus, we would expect that the royalty rate would be high when the licensing contract has an exclusivity provision. Let us then consider the effect of competition among licensors. In the case where the entire bargaining surplus belongs to a licensee due to the competition among licensors with identical technology, we have the following:

R=0

(7)

In this case, the exogenous changes affecting the ex-post profits of a licensee, including the level of protection of intellectual property rights, do not affect the price of technology. In reality, however, it is unlikely that competition among licensors are so severe, since patent protection restricts the possibility that two firms have the exactly the same technology among others. 9

In the empirical analysis, we depend on the data of royalty rates, Y

which are given in terms of the sales of the licensee ( S D ). If we assume that all payments are made by a running royalty and if it does not affect the sales of a licenseevi, we can obtain the following expression for the royalty rate:

Rate = R / S DY = δ (1 − γ )(Π YD , N / S DY ) + α ( RD X / S DY )

(8).

One noteworthy point is the fact that the reduced possibility of third party infringement ( lower γ) can reduce the royalty rate due to the expansion of the sales. This reflects the point that a part of the payment for technology depends on the threat point.

IV. Framework of estimation We postulate the following specification in order to examine how the royalty rates are related to the contract characteristics, as suggested by the above theoretical model, based on industry-level panel data.

( price) i ,t = α + β1 (rds) i ,t + β 2 (monopoly) i ,t + β 3 (br ) i ,t + β 4 ( pat ) i ,t + β 5 (kh) i ,t + β 6 (cr ) i ,t + β 7 (initial ) i ,t + β 8 ( propd ) i ,t + η i + ε i ,t

(9)

In the above specification, i denotes the sector and t the time period. We introduce 31 industry dummies to control industry-level missing variables, which are fixed over time, which may cover research opportunities, demand growth and market structure. For an example, we expect that a sector with rich research opportunities would have high level of R&D as well as high royalty rates, so that we may find spurious correlations between the two variables unless we control industry fixed effects. The variable η i is 10

unspecified fixed effects and ε i,t is a random term. We divide the 20 year time period into the following four periods: the first period is from 1981 to 1984, the second period is 1985, 1986 and 1989, the third period is from 1990 to 1994 and the fourth period is from 1995 to 1998. The dependent variable (price)i,t is the price of technology. We use the share of the licensing contracts with the royalty rate of 8% or more in all royalty based contracts (price), to measure the price level of imported technology. The variance of this variable is likely to be heterogeneous, given the difference of the number of contracts with royalty payments across sectors. We will use GLS estimation to take this into account. We also limit our sample to those industries which have ten or more licensing contracts with royalty payments in each period (we have 39 sectors satisfying this condition out of 48 sectors (see Table A-2), and the sample is further limited to 32 sectors, due to the availability of R&D data). Table A-3 in the appendix provides summary statistics. The first independent variable (rds) is the R&D intensity of domestic industry: the R&D expenditure of each industry divided by its sales. It aims at capturing the profitability of commercializing imported technology as well as the cost of R&D necessary for a licensee to invent around the technology in case licensing does not materialize. As suggested by equations (6) and (8), it would positively affect the royalty rates. The second independent variable (monopoly) is the share of the contracts with exclusive right. The presence of exclusivity provision shows the existence of ex-ante competition among licensees, as explained in the 11

above section. It would positively affect royalty since it implies stronger bargaining power of a licensorvii. The third, fourth and fifth independent variables indicate the structure of intellectual property rights (IPRs) specified in contracts. The variables br, pat and kh respectively denote the share of the contracts with trade-mark, patents and know-how. Considering the possibility of collinearityviii, we also estimate equations by dropping one of the variables (kh). The sixth independent variable (cr) is the share of the contracts with a cross licensing provision. Cross-licensing in the context of technology import implies that part of the payment is made by technology. However, it may also indicate that the technology introduced is of high quality, since the patentee of a pioneer patent often requires the grant-back of derivative technologies. The seventh independent variable (initial) is the share of the contracts with initial payment. The total payment for technology consists of royalty payments and initial payments. The variable initial would control the substitutability between initial payments and royalty payments. The last independent variable (propd) is a dummy variable representing the effect of stronger IPR policy of Japan since the middle part of the 1990s. We evaluate whether royalty rates increased in the last period of 1994-1998 in a manner consistent with the impact of stronger IPRs. As a benchmark estimation, we use a time dummy variable (Time4) , which takes a value 1 for the last period, as one of the dummies. The effects of stronger IPR would differ across sectors, which have two sources: the effect of IPR on appropriability of R&D for a given market and the territorial focus of

12

licensing contracts on the Japanese market. As for the indicator of the appropriability, we use two variables. One indicator is the index of appropriability of patent protection (hapr, which have values of either 0, 1 or

2), which we assigned, based on the survey results of Goto and Nagata (1996) and Cohen, Walsh and Nelson (2000). The second indicator is R&D intensity of each sector, since stronger protection of IPR (larger α) has a larger effect on royalty in R&D intensive industry as shown in equations (6) and (8). The effects of stronger IPR protection would also depend upon the territorial scope of licensing contracts. If licensing covers only the Japanese market, its term is more strongly influenced by the IPR protection in Japan. We use the average percentage of licensing with its territorial scope restricted to Japan from 1990 to 1991 (mrkj) as an indicator of the territorial focus of the contract (see Table A-2 in the Appendix ). The appropriability (or R&D intensity) indicator and the focus of the territorial scope are likely to affect the royalty rate in a multiplicative manner, so that we use the following

variables

haprmrkd = hapr × mrkj × time4

(

and

rdsmrkd = rds × mrkj × time4 ).

V. Estimation results Table 3 shows the estimation results. Estimation 2 to 5 use the appropriability indicator to reflect the impact of stronger IPR policy of Japan in late 1990s. Estimation 6 to 9 uses the R&D intensity instead. Estimation 4 and 8 drop the share of contracts with know-how (kh) as an independent variable. Estimations 5 and 9 drop the share of contracts with initial payment (initial) as an independent variable. The Table omits reporting the

13

estimated coefficients of the industry dummies. R&D intensity (rds) has a positive and significant coefficient, either 1% or 5%, except for the last estimation. Thus, higher R&D intensity over time has resulted in higher royalty rates. The share of exclusive rights (monopoly) has a positive and highly significant coefficient at 1% level in all estimation results. Thus, consistent with the above theory, exclusivity increases royalty rate, although exclusivity became less important in the sample period (see Table 1). As for the impact of each IPR, in the estimations with all of the three IPR variables, the brand variable (br) has the largest and positive coefficient, followed by the patent variable (pat) and then the know-how variable (kh), although none of them are statistically significant. According to estimation 4 and 8, which leaves out kh as an independent variable, however, br is significant at 5% level and has a positive sign, while pat remains insignificant. Thus, the expansion of trademark relative to know-how is associated with more high royalty contracts. This may not be surprising, given that a licensing contract with trademark would enable a licensee to market a distinctive product, which would in turn indicate that the technology introduced is comprehensive. The cross-license variable (cr) has a significant and positive sign in all estimation. This indicates that a cross licensing provision means high quality of licensed technology more than payment in kind. Estimation 2 and 6, compared with estimation 1, 3 and 7, show that high royalty contracts increased significantly in 1995-98 more in the

14

industries for which intellectual property rights are important in terms of the appropriability indicator or the R&D intensity, and for which the restriction of sales territory to Japan is frequently imposed in the contracts. According to estimation 1, the time dummy for the last period has a positive coefficient, but not significant, confirming that the inter-industry difference is important. In addition, estimation 3 and 7 show the variables which is not multiplied by the importance of the territorial scope of contracts (haprd and

rdsd respectively) is not significant, given the variable incorporating the territorial scope. These results provide further support to the view that the increase of high royalty contracts in the latter part of 1990s reflected the impact of stronger IPR policy in Japan. Finally, estimation 5 and 9 remove

initial variable out of the estimation. The results are largely unchanged. VI Conclusions This paper examined the changes of royalty terms of technology import contracts of Japan in 1980s and 1990s, based on the sector-level panel data. Major findings are the following. First, the share of high royalty contracts increased significantly in 1980s and 1990s, which, however, mainly reflected the change of the composition of imported technologies, especially the increasing share of software licenses. The share of the contracts with a trademark increased significantly both in individual sectors and in the aggregate.

The share of the contracts with a patent increased in individual

sectors but decreased in the aggregate, in contrast to the share of contracts with know-how, reflecting the above composition change. Secondly, our estimation, controlling industry fixed effects, shows

15

that an exclusivity provision and trademark licensing (relative to know-how licensing) are accompanied with more high royalty contracts. It also shows that a cross-licensing provision and the increase of R&D intensity of the technology importing industry tended to be associated with high royalty contracts. Thirdly, after controlling these factors, the share of high royalty contracts increased significantly in the latter part of 1990s in those industries for which intellectual property rights are important in terms of appropriability or R&D intensity and the restriction of sales territory to Japan is frequently imposed in the contracts. They seem to reflect the impact of stronger IPR policy of Japan since the middle part of 1990s.

16

Reference Anand, B.N and T. Khana, 2000, “The structure of licensing contracts,” Journal of Industrial Economics 48, 103-135 Arora A., A. Fosfuri and A. Gambardella, 2001, Markets for Technology, MIT Press Arrow, K, 1962. “Economic Welfare and the Allocation of Resources for Invention,” In R. Nelson (Ed.), The Rate and Direction of Inventive

Activity, pp. 609-625. Caves, R., Crookell, H. and Killing, P.J., 1983, ‘The Imperfect Market for Technology Licenses’, Oxford Bulletin of Economics and Statistics, 45, pp. 249-267. Cohen, W., Nelson, R. and J. Walsh, 2000, “Protecting Their Intellectual Assets: Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not),” NBER Working Paper No. W7552 Gallini, N. and Wright, B., 1990, ‘Technology Transfer Under Asymmetric Information’, Rand Journal of Economics, 21, pp. 147-160. Goto A. and Akiya Nagata, 1996, ”Study of innovation process by a survey data,” presented at the Workshop on Appropriability and Technological Opportunities at National Institute for Science and Technology Policy) Taylor, C. and Silbertson, Z., 1973, The Economic Impact of the Patent

System: A Study of the British Experience (Cambridge University Press, New York). Schankerman M. and S. Scotchmer, 1999, “Damages and injunctions in the protection of proprietary research tools,” NBER Working paper 7086

17

Appendix 1

Data Sources

The licensing contracts data set is from the annual issues on status of technology imports and technology-related imports compiled by the Science and Technology Agency (now by the National Institute for Science and Technology Policy). The tables with detailed industrial classification (39 industries) are available for the period from 1981 to 1998, except for the years of 1987 and 1988. It has compiled this report based on compulsory reporting requirements by the “Foreign Exchange and Foreign Trade Control Law” (hereinafter referred to as the “Foreign Exchange Control Law”). The R&D data is from the National Survey of Research and Development by Ministry of Public Management, Home Affairs, Posts and Telecommunications. Since the data published from this survey is less aggregated than the license data in terms of industrial classification, the matching between two data is imperfect in general machinery and electrical/electronics machinery. We classify industrial sectors into three group (high appropriability sector, medium appropriability sector and low appropriability sector), according to the survey results on appropriability of patents in the Japanese and US industries (Goto and Nagata (1996), and Cohen, Nelson and Walsh(2000)). We assigned each sector a number (either 2, 1 or 0), according to the average of the effectiveness of patents for appropriability in process and product innovation (see Table A-2 ).

18

For an example, with respect to the effects of IPRs on technology transfer, a view focusing on the profitability of imitation suggests that strong protection of IPRs in a technology recipient country would reduce such transfer, since a larger imitation cost due to stronger IPRs would increase the payment by the licensee when imitation cost is binding on license terms (see, for an example, Gallini and Wright (1990)).On the other hand, a view focusing on the profitability of a licensor in the investment for technology transfer suggests that strong protection of IPRs may facilitate the licensing of know-how (Arora et. al. (2001)). ii IPR protection has become significantly strengthened in Japan since the middle of 1990s. The major developments include the following. The possibility of compulsory licensing was significantly restricted, based on the Japan and US governmental agreement in 1994. The scope of patent protection to software has been extended through several steps. The TRIPS agreement became effective in 1995. The doctrine of equivalent was established in the Supreme Court decision in 1998 (the first decision by the lower court in 1994). iii The reporting requirement was narrowed down to the contracts with total payments more that 30 million Yen. iv See Arrow (1962) for a seminal discussion. v However, there is a theoretical possibility that stronger protection of IPRs shifts the threat point in favor of a licensee. As discussed by Schankerman and Scotchmer(1999), larger damage for infringement can induce a licensee to switch its conduct from the strategy of infringement (and payment of damage) to that of non-infringement when it faces refusal of license. vi Such would be the case if the licensee faces Bertrand competition in the product market. In this case his price is entirely determined by the product quality or cost of the competitor with the second-best technology. vii The exclusivity provision may also indicate the quality of technology, since a licensor is more concerned with preventing competition among licensees when the technology is drastic over the prevailing technology. viii If each contract has only one type of IPR (i.e only one of patent, trademark and know-how), the sum of br, pat and kh is unity. As shown in Table 2, 80% of the contracts have only one type of IPR. i

19

Table 1 Contract characteristics over time I: Average of all contracts

II. Simple average of industry values

The proportion in the contracts (%) 81-84

price

85,86,89

90-94

95-98

81-84

85,86,89

90-94

95-98

Onerous contracts

94.0

94.5

94.2

94.5

93.3

92.6

93.9

95.9

Royalty contracts

68.7

50.3

56.0

62.2

73.3

63.4

70.1

71.9

High royalty contracts

13.4

14.9

23.7

26.5

13.4

11.0

17.5

14.8

Initial payments

58.2

70.6

69.6

61.3

57.6

66.1

63.2

60.4

Monopoly rights

51.0

44.1

36.7

29.4

52.0

48.7

45.8

41.6

4.0

3.5

2.7

4.1

4.4

4.7

4.1

5.6

non-price Coss license

Note:

% of high royalty contracts are with respect to the royalty contracts, not with respect to all contracts.

Table 2 Structure of IPRs over time

The proportion in the contracts (%)

I: Average of all contracts 81-84

85,86,89

90-94

II. Simple average of industry values 95-98

81-84

85,86,89

90-94

95-98

Only patents

11.71%

10.38%

8.15%

9.01%

14.35%

15.50%

12.46%

15.42%

With patents

36.47%

33.89%

24.30%

23.96%

41.08%

44.04%

44.23%

43.06%

Only trademark

5.90%

9.01%

12.01%

19.44%

6.80%

11.25%

19.95%

17.69%

With trademark

24.20%

21.49%

21.01%

32.39%

23.61%

25.87%

31.04%

38.51%

Only knowhow

48.78%

52.84%

59.20%

48.71%

41.66%

36.20%

32.09%

28.96%

With knowhow

69.34%

73.05%

90.04%

70.97%

72.42%

73.06%

74.02%

60.36%

Table 3 Estimation results (Ppanel estimates, Fixed effects GLS estimation) Number of obs = 128 , Number of groups = 32, ***: significant at 1%, **: significant at 5%, and * significant at 10% Independent variables

rds monopoly br pat kh cr initial time4 propd haprd haprmrkd Log likelihood Independent variables

rds monopoly br pat kh cr initial time4 propd rdsd rdsmrkd Log likelihood

Estimation 1 Dependent variable (price) 31 Industry dummies Coef. Std. Err. 2.107 0.667 *** 0.107 0.040 *** 0.052 0.052 0.033 0.058 -0.004 0.053 0.501 0.191 *** -0.106 0.044 ** 1.670 1.048

Estimation 2 Dependent variable (price) 31 Industry dummies Coef. Std. Err. 1.752 0.622 *** 0.099 0.035 *** 0.068 0.049 0.030 0.055 0.005 0.049 0.472 0.178 *** -0.090 0.042 ** 5.882

-358.79

1.591 *** -353.96

Estimation 3 Dependent variable (price) 31 Industry dummies Coef. Std. Err. 1.799 0.645 *** 0.100 0.035 *** 0.081 0.051 0.036 0.055 0.016 0.051 0.503 0.182 *** -0.085 0.042 ** -1.570 8.461

2.043 3.607 ** -353.79

Estimation 4 Dependent variable (price) 31 Industry dummies Coef. Std. Err. 1.727 0.610 *** 0.101 0.036 *** 0.065 0.031 ** 0.027 0.050 0.517 -0.080 5.783

0.172 *** 0.039 ** 1.577 *** -352.67

Estimation 6 Dependent variable (price) 31 Industry dummies

Estimation 8 Estimation 7 Dependent variable (price) Dependent variable (price) 31 Industry dummies 31 Industry dummies

Estimation 9 Dependent variable (price) 31 Industry dummies

Coef. Std. Err. 1.420 0.621 **

Coef. Std. Err. 1.480 0.622 **

Coef. Std. Err. 0.826 0.661

0.116 0.065 0.047 0.002 0.329 -0.087

0.034 *** 0.049 0.054 0.049 0.184 * 0.041 **

0.118 0.088 0.054 0.025 0.366 -0.082

0.034 0.051 0.054 0.052 0.185 0.041

1.922

0.469 ***

-0.458 2.834

0.375 0.883 ***

-351.58

Coef. Std. Err. 1.402 0.614 **

*** *

0.113 0.063 0.045

0.035 *** 0.029 ** 0.049

** **

0.379 -0.080

0.179 ** 0.038 **

1.838

0.469 ***

-351.62

-351.15

0.120 0.069 0.033 -0.028 0.324

2.174

0.036 *** 0.045 0.052 0.043 0.178 *

0.508 *** -355.02

Estimation 5 Dependent variable (price) 31 Industry dummies Coef. Std. Err. 1.264 0.644 ** 0.105 0.038 *** 0.071 0.046 0.015 0.053 -0.027 0.044 0.486 0.174 ***

6.492

1.618 *** -353.11

Fiscal year 19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

19

98

97

96

95

94

93

92

91

90

89

88

87

86

85

84

83

82

81

80

79

78

77

76

3,500

75

74

73

72

71

4,000

19

19

19

19

19

Number of contracts

Figure1 Number of technology import contracts

4,500

Others Software

3,000

2,500

2,000

1,500

1,000

500

0

Figure 2  Frequency of patents in licensing contracts in R&D intensive industry

Radio & television receivers and electric audio equipment Other electronic equipment Industrial organic chemicals Electronic parts and devices Other electric machinery Drugs and medicines Radio & television receivers and electric audio equipment Electricity generation, transmission & distribution equipment etc.

95-98 81-86

Precision machinery Oil & fat products, soaps, etc.

(%)

Computers

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0 100.0

Figure 3 Frequency of patents in licensing contracts 70 60.9

60

59.3

58.8 55.2

58.3

62.3

60.5 56.5

54.6

50 40 %

Frequency of patents in the license of hardware technology Frequency of patents in software license

30

22.6

20 15.5 10

9.9 0.9

0 89

90

1.8 91

2.3 92

3.0 93

3.6 94

3.2 95

96

Note. Hardware technology covers all technologies, excluding software and trademark.

97

98 FY

Figure 4 Licensing conditions in pharmaceutical industry 100.0 81-84 85,86,89 90-94 95-98

90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0

ho w ow Kn

Tr ad em ar k

nt Pa te

s Cr os

on op ol y M

y ro ya lt Hi gh

In iti al

No nf

re

e

0.0

Figure 5 Licensing conditions in computer industry 110.0 81-84 85,86,89 90-94 95-98

100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0

ho w ow Kn

Tr ad em ar k

nt Pa te

s Cr os

on op ol y M

y ro ya lt

Hi gh

In iti al

No nf

re

e

0.0

Table A-1 Industry composition of technology import contracts (%)

81-84 Total number of contracts

8,895

85,86,89 7,695

90-94 15,800

94-98 11,258

Clothing and textile products Drugs and medicines Other chemicals General machinery and tools Transportation equipment

11.7% 3.2% 5.8% 18.8% 3.7%

8.0% 3.9% 5.0% 12.6% 3.9%

5.3% 3.3% 2.8% 8.0% 2.1%

12.6% 2.8% 2.1% 7.3% 1.8%

Computers

17.4%

32.9%

53.0%

46.7%

Other electric/electronics machinery Precision machinery Others

11.7% 2.9% 24.8%

11.8% 2.5% 19.5%

11.6% 2.3% 11.7%

13.4% 1.7% 11.6%

Table A-2 Industry characterristics of licensing (average of the four periods) ,%

Code

Industry

High royalty Onerous initial contracts contracts

price 4 Construction 11 Food and tobacco 12 Textiles 14 Outer garments 15 Other clothing textile products 16 17 22 24 25 26 31 32 33 35 36 41

Sawing/planing mill products, wood products and furniture Pulp and paper Industrial organic chemicals Oil & fat products, soaps, etc. Drugs and medicines Other chemicals Rubber products Tanned leather, leather products and fur skins Ceramics Non-ferrous metals and products Fabricated metal products

Boilers and engines Agricultural, construction and mining machinery and 42 equipment 43 Metal working machinery

45 47 48 49 50 51 52 53

Special industrial machinery Pumps, compressors and blowers Prime mover Chemical machinery and equipment Other general industry machinery Other machinery Transportation equipment Precision machinery

Electricity generation, transmission & distribution equipment and industrial electrical machinery, 61 equipment & supplies 64 Wired and radio communication equipment Radio & television receivers and electric audio 65 equipment 68 Computers

69 Other electronic equipment 70 Electronic parts and devices 71 Other electric machinery Precious metal products, costume jewelry, etc.

81 82 Leisure activity equipment 83 Plastic products

notfree initial

territorial restriction to Japan

mrkj

8.6 4.8 17.0 8.9 14.1

100.0 89.3 96.6 98.6 99.2

66.3 41.2 28.3 27.0 11.8

88.1% 81.7% 78.2% 82.3% 96.8%

5.2

96.8

46.8

85.0%

17.6 5.2 10.6 24.1 16.0 13.3 15.8 14.4 7.9 8.4 5.3 8.6

93.8 90.1 91.6 87.6 86.2 94.1 99.2 93.6 89.1 95.6 94.3 98.7

53.0 74.7 51.0 72.4 49.4 42.4 16.8 68.1 78.7 65.1 89.7 64.0

6.4 17.1 5.4 2.3 13.2 9.5 11.6 12.1 25.0

95.6 97.2 97.1 95.8 98.2 99.0 92.8 96.7 94.1

12.6

Approp R&D riability sampl intensity index e

rds

hapr

0.48 0.87 1.50 1.50 1.50

0 1 1 0 0

* * * * *

81.3% 42.5% 77.8% 56.5% 60.7% 40.0% 90.4% 50.0% 43.8% 49.3% 87.0% 25.0%

0.79 3.50 3.76 7.35 4.14 3.03

0 1 2 2 1 1

* * * * * *

2.43 1.98 1.44 2.92 2.92

1 1 1 1 1

* * * * *

70.8 65.7 64.5 61.2 76.1 75.8 72.4 72.1 58.6

43.1% 32.0% 33.3% 24.2% 55.9% 42.9% 36.2% 52.9% 43.4%

2.92 2.92 2.92 2.92 2.92 2.92 2.92 3.23 4.99

1 1 1 1 1 1 0 1 1

* * * * * * * * *

93.1

70.6

37.8%

5.21

1*

9.1

92.2

82.9

15.2%

5.61

0*

3.0

90.2

63.1

8.3%

5.61

1*

62.8 10.5 8.0 3.7 25.9

95.7 92.9 88.9 94.0 94.0

78.2 76.7 75.3 78.8 16.8

71.9% 10.7% 12.5% 17.6% 68.2%

5.61 5.61 5.61 5.61

2 0 1 0

* * * *

19.2 93.4 40.6 66.7% 5.7 89.7 65.9 52.2% Manufacturing industries not classified elsewhere 27.9 94.8 44.3 88.2% 84 90 Other industries 12.6 96.8 62.8 100.0% Note. The proportion of the contracts with the territorial restriction to Japan (mrkj) is based on the contracts in 90 and 91.

Table A-3 Summary Statistics Mean 1981-84 1985,86,89 1990-94 1995-98

initial rds monopoly br pat kh cr N of ob price 32 12.17 58.76 2.60 52.97 11.22 39.52 80.82 5.19 32 10.37 69.60 3.33 50.28 14.82 43.29 75.61 4.62 32 12.97 63.82 3.73 47.35 24.53 43.85 69.35 4.18 32 14.10 61.01 3.79 38.78 26.51 43.77 68.58 5.84

Standard deviation No of ob price initial rds monopoly br pat kh cr 1981-84 32 11.22 14.77 1.41 22.70 8.01 20.45 13.85 5.85 32 10.10 15.74 1.73 21.32 12.11 20.58 15.08 4.53 1985,86,89 1990-94 32 13.41 19.97 1.92 21.70 26.05 22.93 24.78 5.14 1995-98 32 13.07 25.79 1.89 23.11 25.19 25.70 24.33 5.07