Appended Notes AN 1.2.1 Handling of “services industries” and “professional and business services” in industrial classification This is an explanation regarding industrial classification methods used in handling statistical data in Chapter 1, Section 2. First, in handling US statistics, “professional and business services,” are, as a rule, “professional and business services,” classified by the US Department of Labor, Bureau of Labor Statistics as Super Sector 60. In the North American Industry Classification System (NAICS) codes, applied during a 1997 census, “professional and business services” correspond with the total of NAICS 54 (professional, scientific, and technical services), NAICS 55 (management of companies and enterprises) and NAICS 56 (administrative and support and waste management and remediation services) (Appended figure 1.2.1). Appended figure 1.2.1 Classifications of services industries according to the US Department of Labor and a comparative table of the NAICS Code Super Sector 07 Services Industries Super Sector 40 Trade, Transportation and Utilities NAICS Code 42 Wholesale Trade NAICS Code 44-45 Retail Trade NAICS Code 48-49 Transportation and Warehousing NAICS Code 22 Utilities Super Sector 50 Information NAICS Code 51 Information Super Sector 55 Finance NAICS Code 52 Finance and Insurance NAICS Code 53 Real Estate and Rental and Leasing Super Sector 60 Professional and Business Services NAICS Code 54 Professional, Scientific, and Technical Services NAICS Code 55 Management of Companies and Enterprises NAICS Code 56 Administrative and Support and Waste Management and Remediation Services Super Sector 65 Education and Medical Services NAICS Code 61 Educational Services NAICS Code 62 Health Care and Social Assistance Super Sector 70 Recreation and Customer Services NAICS Code 71 Arts, Entertainment, and Recreation NAICS Code 72 Accommodation and Food Services Super Sector 80 Other Services NAICS Code 81 Other Services (except Public Administration) Super Sector 90 Public Administration NAICS Code 92 Public Administration Super Sector 99 Non-classifiable NAICS Code 99 Non-classifiable Note: Bold lettering indicates the industry classifications according to the Bureau of Labor Statistics, US Department of Labor. Source: North American Industry Classification System (Census Bureau, US Department of Commerce). The “services industries” are “services industries” classified by US Department of Labor, Bureau of Labor Statistics as Super Sector 07. To see the relationship between “services industries” and their subdivisions according to the US Department of Labor, Bureau of Labor Statistics and the NAICS codes, please refer to Appended figure 1.2.1. -281- -282- 66-70 Hotels and other lodging 72 Personal services Finance, insurance, and real estate Hotels and other lodging Personal services 79 Amusement and recreational services 80 Health services 82 Educational services 83 Social services Sum total of the above Other repairs Film and video productions Entertainment Medical services Education Social services Service industries L L L L L L L G H L L J K L L L L L L L L L L L L L L L Service Service Service Service Service Service Service First division Electricity, gas, heat and water supply Transportation and communications Service Service Finance and insurance Real estate Service Service Service Service Service Service Service Service Service Service Service Service Service Service Service 76 79 79 88 89 91 90 75 72 74 79 79 79 82 83 84 86 92 77 79 78 80 Hotels, boarding houses and other lodging places Laundry, beauty and bath service Miscellaneous domestic and personal services Goods rental and leasing Goods rental and leasing Goods rental and leasing Information services and research Advertising Professional services (cannot be classified as other) Miscellaneous business services Scientific research institutes Automobile repair services Goods rental and leasing Machine, upholstery, furniture, etc., repair services Motion picture and video production Amusement and recreation services, except motion picture and video production Goods rental and leasing Goods rental and leasing Medical and other health services Public health services Education Social insurance and social welfare Sum total of the above 81 Broadcasting 87 Waste treatment services Japan Standard Industrial Classification (1993 revision) Second division Source: Japan Standard Industrial Classification (Management and Coordination Agency), North American Industry Classification System (Census Bureau, US Department of Commerce). 75 Automobile repair, services, and parking 76 Miscellaneous repair services 78 Motion pictures Automobile repairs, rentals, etc. Professional and business services 73 Business services 81 Legal services 87 Engineering and management services 60-65 Financial, insurance, and real estate Utilities, transportation and communications Definition US Standard Industrial Classification (SIC 1987) Code Industry 40-49 Transportation and public utilities Appended figure 1.2.2 Comparative table of the Japan Standard Industrial Classification (1993 revision) and the US Standard Industrial Classification System (SIC 1987) in the classification of services industries 795 Sports and recreation goods rental and leasing 799 Miscellaneous goods rental and leasing 794 Automobile rental and leasing 791 General goods rental and leasing 792 Industrial equipment and machinery rental and leasing 793 Office machinery rental and leasing Third division One must be careful regarding Figures 1.2.3, 1.2.5, 1.2.6, 1.2.7, 1.2.15, 1.2.16 and 1.2.17, however, in terms of ensuring comparability between Japan and the US and statistical data restrictions because industries are classified and defined individually as in Appended figure 1.2.2. Regarding the Japan Standard Industrial Classification, the reason that 1993 revised classification is used rather than the newest classification, which was revised in 2002, is that the acquired data is taken from the 1993 classification. For more information on industrial classification according to the North American Industry Classification System (NAICS), Standard Industrial Classification (SIC) and the US Department of Labor, Bureau of Labor Statistics and the Japan Standard Industrial Classification, please refer to the websites below. Also, please refer to the website below for more information regarding Standard Occupational Classification (SOC) used in Figure 1.2.14. (a) North American Industry Classification System (NAICS) and Standard Industrial Classifications (SIC) US Department of Commerce, Census Bureau website (http://www.census.gov/epcd/naics02/) (b) Industrial classification according to the US Department of Labor, Bureau of Labor Statistics US Department of Labor, Bureau of Labor Statistics website (http://stats.bls.gov/webapps/legacy/cesbtab1.htm) (c) Japan Standard Industrial Classification Ministry of Public Management, Home Affairs, Posts and Telecommunications, Statistics Bureau website (http://www.stat.go.jp/index/seido/sangyo/) (d) Standard Occupational Classification (SOC) US Department of Labor, Bureau of Labor Statistics website (http://stats.bls.gov/soc/home.htm) AN 2.1.1 Relationship between company sales and non-R&D intellectual assets 1. Method for estimating non-R&D intellectual assets1 (1) Basic model Y=company sales, K=capital, L=labor (number of employees), R=R&D (research and development expenditure), e=error term, A=regular non-R&D intellectual assets (can be used for all companies), FA=non-R&D intellectual assets unique to each company 1 Lev (2003). -283- [Production function] Y = f ( A, FA, K , L, R) = A( FA) KαLβRγe (1) (2) Estimation of non-R&D intellectual assets In estimating non-R&D intellectual assets, an annual growth equation (Equation 2) is used in order to estimate residual output components in the basic model (Equation 1). Log (Y / Y−1 ) = δ + sπD + αLog ( K / K −1 ) + β Log ( L / L−1 ) + γLog ( R / R−1 ) + Log (e / e−1 ) ( 2) * Here D is a dummy variable for non-R&D intellectual assets unique to companies, and this coefficient is the measured value of non-R&D intellectual assets unique to companies. (3) Monetary conversion of non-R&D intellectual assets The contribution of sales of non-R&D intellectual assets (RO) is equal to the difference of sales in the sales estimation in which non-R&D intellectual assets are included (2A) and the estimation in which non-R&D intellectual assets are not included (2B) (Equation 3). Y ∗ = Y−1exp(sπ )exp(δ )(K/K -1 )α (L/L-1 ) β (R/R -1 ) γ (2A) Y ∗∗ = Y−1 (K/K -1 )α (L/L-1 ) β (R/R -1 )γ (2B) RO = Y ∗ − Y ∗∗ (3) * The values estimated in Equation 2 are used for α, β, etc. 2. Estimation results (1) US2 (a) Data ・ This analysis covered approximately 250 companies listed in the Information Week 500 between 1991 and 1997. ・ Individual data (sales, capital, number of employees, research and development expenditure) was obtained from the Compustat Annual Database. ・ Regarding the company data sample used to estimate non-R&D intellectual assets, the sample that could be obtained for research and development expenditure data for the 1987-2000 period was 2 Lev (2003). -284- 1,952 and the sample that could not be obtained was 1,246. ・ Research and development expenditure is capitalized and depreciated over five years (annual depreciation rate=20%) (b) Results Appended figure 2.1.1 Estimation results for the US 1. Statistics Variables Y K L R US$ million US$ million 1,000 people US$ million R&D expenditure data available R&D expenditure data not available Average Minimum Median Maximum Average Minimum Median Maximum value value value value value value value value 9,123 146 4,678 101,781 6,532 3 3,104 191,329 3,433 17 1,302 51,161 1,808 2 632 40,934 42 2 25 813 39 1 16 1,244 1,036 1 323 16,439 2. Coefficient of correlation Variables Log(K/K-1) Log(L/L-1) Log(R/R-1) R&D expenditure data available Log(Y/Y-1) Log(K/K-1) Log(L/L-1) 0.61 0.70 0.67 0.41 0.34 0.33 R&D expenditure data not available Log(Y/Y-1) Log(K/K-1) 0.56 0.65 0.63 3. Estimation results R&D expenditure data available R&D expenditure data not available Coefficient t value P value Coefficient t value P value Intercept 0.03 5.41 0.00 0.02 5.58 0.00 Log(FA/FA-1) 0.02 5.61 0.00 0.02 6.16 0.00 Log(K/K-1) 0.20 8.87 0.00 0.12 4.39 0.00 Log(L/L-1) 0.38 12.47 0.00 0.44 16.12 0.00 Log(R/R-1) 0.19 6.61 0.00 0.6253 0.5813 Final variables (adjusted) Variables 4. Estimation results for monetary conversion of non-R&D intellectual assets Average Standard Minimum Median Maximum Variables deviation value value value value 251 777 -2,724 72 8,654 RO=Y*-Y** US$ million Y-Y-1 US$ million 576 1,876 -27,425 207 27,379 Source: Lev (2003). (2) Japan (a) Data ・ This analysis covered 964 listed companies in Nihon Keizai Shimbun’s corporate database of Nikkei NEEDS for which items used for estimation3 (sales and operating profit, total tangible fixed assets, land and the rest of the tangible fixed assets, number of employees, development expenses and experimental and research expenses) could be continuously obtained for the 1989-2002 estimation period. ・ Of the 964 companies analyzed, 402 were in manufacturing industries and 562 were in non-manufacturing industries. Of the companies analyzed in the manufacturing industry, 173 were in the machinery assembly manufacturing industries4, and of the companies in the non-manufacturing 3 Numerical values are on a non-consolidated basis. They cover the four sectors of machinery, precision instruments, electrical instruments and transport equipment under the industry classification of the Tokyo Stock Exchange. 4 -285- industry, 263 were in retail/service industries, etc.5 ・ The company data sample used to estimate non-R&D intellectual assets was 5,628 for manufacturing industries (including 2,422 for machinery assembly manufacturing industries) and 7,868 for non-manufacturing industries (including 3,682 for retail/service industries, etc.). ・ Research and development capital is defined as the average value of research and development expenses during the most recent five years. (b) Results Using Lev (2003) as reference, estimation of non-R&D intellectual assets in Japan is carried out applying company data for each year to Equation 2. Appended figure 2.1.2 Estimation results for Japan (manufacturing industries, non-manufacturing industries) 1. Statistics Variables Y K L R Million yen Million yen People Million yen Manufacturing industries Non-manufacturing industries Average Minimum Median Maximum Average Minimum Median Maximum value value value value value value value value 181,098 1,435 55,419 3,408,251 250,664 518 82,323 9,419,359 41,282 269 12,313 997,139 74,323 10 9,265 4,884,790 2,480 31 1,120 42,375 2,052 7 897 100,090 8,120 8 968 301,298 2. Coefficient of correlation Variables Log(K/K-1) Log(L/L-1) Log(R/R-1) Log(Y/Y-1) 0.30 0.35 0.11 Manufacturing industries Log(K/K-1) Log(L/L-1) 0.45 0.18 0.15 Non-manufacturing industries Log(K/K-1) Log(Y/Y-1) 0.24 0.36 0.32 3. Estimation results Manufacturing industries Non-manufacturing industries Coefficient t value P value Coefficient t value P value Intercept 0.003 2.046 0.041 0.018 13.081 0.000 Log(FA/FA-1) 0.005 1.280 0.201 -0.008 -5.448 0.000 Log(K/K-1) 0.142 12.495 0.000 0.077 12.193 0.000 Log(L/L-1) 0.332 19.102 0.000 0.358 28.574 0.000 Log(R/R-1) 0.030 3.060 0.002 Final variables (adjusted) 0.147 0.148 Variables 4. Estimation results for monetary conversion of non-R&D intellectual assets Manufacturing industries Average Standard Minimum Median Maximum Average Variables deviation value value value value value 784 2,121 -62 188 23,052 1,089 RO=Y*-Y** Million yen Y-Y-1 Million yen 1,178 39,188 -780,935 333 581,049 -2,458 Source: Estimates based on Lev (2003). 5 Non-manufacturing industries Standard Minimum Median Maximum deviation value value value 15,335 -270,831 1,252 45,952 176,304 -2,825,337 864 6,832,939 They cover the eight sectors of retail trade, services, land transport, maritime transport, air transport, warehousing/related transportation, information and telecommunications, and electricity and gas under the industry classification of the Tokyo Stock Exchange. -286- Appended figure 2.1.3 Estimation results for Japan (machinery assembly manufacturing industries, retail/service industries, etc.) 1. Statistics Variables Y K L R Million yen Million yen People Million yen Machine assembly manufacturing industries Retail/services industries, etc. Average Minimum Median Maximum Average Minimum Median Maximum value value value value value value value value 234,565 1,435 52,607 3,408,251 183,955 518 68,597 2,482,744 36,437 269 9,237 520,779 125,343 10 14,566 4,884,790 3,339 73 1,268 42,375 2,763 21 1,129 100,090 11,575 18 845 301,298 2. Coefficient of correlation Variables Log(K/K-1) Log(L/L-1) Log(R/R-1) Machine assembly manufacturing industries Log(K/K-1) Log(L/L-1) Log(Y/Y-1) 0.26 0.30 0.37 0.09 0.18 0.14 Retail/services industries, etc. Log(Y/Y-1) Log(K/K-1) 0.24 0.42 0.32 3. Estimation results Machine assembly manufacturing Retail/services industries, etc. industries Coefficient t value P value Coefficient t value P value Intercept 0.007 2.607 0.009 0.028 16.113 0.000 Log(FA/FA-1) 0.015 2.020 0.043 0.008 3.151 0.002 Log(K/K-1) 0.150 7.858 0.000 0.058 7.391 0.000 Log(L/L-1) 0.365 11.303 0.000 0.387 24.556 0.000 Log(R/R-1) 0.026 1.582 0.114 Final variables (adjusted) 0.114 0.193 Variables 4. Estimation results for monetary conversion of non-R&D intellectual assets Retail/services industries, etc. Machine assembly manufacturing industries Average Standard Minimum Median Maximum Average Standard Minimum Median Maximum Variables deviation deviation value value value value value value value value 2,037 5,330 -470 378 49,270 5,189 10,720 9 1,767 96,462 RO=Y*-Y** Million yen Y-Y-1 Million yen 2,942 53,532 -780,935 533 581,049 3,987 20,822 -249,346 980 468,117 Source: Estimates based on Lev (2003). AN 2.1.2 Provisional methods to evaluate intellectual assets 1. Companies analyzed The analysis was carried out using OSIRIS, a financial database by Bureau van Dijk of listed companies in countries around the world. Of the 8,437 manufacturing industry companies (in 21 countries) in OSIRIS, 7,897 companies were analyzed and 540 companies were excluded based on the standards below. Please refer to Appended figure 2.1.4 for the number of companies analyzed by industry and by country. ・ Data of companies with contradictions such as total capital being negative, etc. ・ Data of companies whose recent sales are unknown or below 100 million yen. -287- Food 121 113 Tobacco 1 6 Textiles 47 28 Clothing 28 48 Lumber and wood 14 23 Furniture and accessories 11 29 Paper and pulp 31 40 Printing and publishing 31 70 Chemicals 177 363 Petroleum and coal 5 20 Rubber and plastics 62 60 Leather 4 23 Ceramic, stone and clay 54 25 Iron and steel, non-ferrous metal 86 72 Metal products 92 71 General machinery 218 339 Electrical machinery 199 543 Transport machinery 94 126 Precision machinery 89 393 Other 19 66 (All manufacturing industries) 1,383 2,458 38 1 2 4 13 7 11 13 46 4 8 2 9 11 14 34 41 15 8 11 292 37 3 12 14 2 6 9 32 60 1 14 5 20 8 18 41 59 30 31 18 420 34 0 8 13 4 4 9 7 34 1 11 0 19 5 10 62 46 23 27 10 327 37 0 12 17 5 5 8 6 32 3 15 2 10 15 25 33 30 19 13 8 295 6 0 7 6 0 1 3 8 7 1 1 0 7 2 1 12 9 6 5 3 85 3 0 1 1 5 4 7 3 7 0 2 0 0 4 6 10 13 7 9 2 84 4 0 0 1 1 3 1 4 2 0 0 0 0 3 1 6 7 4 2 0 39 5 0 3 0 1 1 4 6 4 0 5 0 3 3 1 11 14 1 3 4 69 7 0 2 2 3 3 4 3 10 0 1 0 3 1 0 8 10 2 5 2 66 8 0 3 3 0 3 3 5 8 1 4 0 2 2 5 8 10 2 2 1 70 40 1 53 25 1 1 19 3 171 18 12 3 35 52 13 151 70 44 9 6 727 24 1 13 16 3 3 7 3 48 2 8 5 21 21 10 25 79 33 6 3 331 6 0 23 6 0 1 4 2 24 2 12 2 7 11 5 55 79 6 0 8 253 14 1 8 16 4 2 2 8 18 1 5 8 3 3 4 10 40 2 7 18 174 12 0 2 2 2 2 4 10 3 2 7 0 7 8 9 29 41 13 2 3 158 34 0 17 2 3 3 5 8 16 1 10 4 9 10 8 9 17 3 0 7 166 14 3 12 6 4 1 6 2 23 0 11 2 7 8 5 6 3 1 0 7 121 38 3 8 15 30 7 18 8 25 2 22 1 30 25 22 21 46 19 4 12 356 8 0 0 0 0 0 0 1 1 1 2 0 3 0 1 2 3 1 0 0 23 Overall amount Philippines Malaysia Indonesia Thailand Singapore Hong Kong Taiwan Korea China Netherlands Denmark Finland Norway Sweden Italy France Germany UK Canada US Japan Appended figure 2.1.4 Number of companies analyzed by industry and by country 603 20 261 225 118 97 195 233 1,079 65 272 61 274 350 321 1,090 1,359 451 615 208 7,897 2. Handling of data in the process of indexation of proxy variables (1) Deviation values of proxy variables Indexes (deviation values) are calculated for each proxy variable according to procedures shown in (a)-(j) below, with the average value 50 and changes of 10 for each standard deviation. In the calculation results, when the deviation value is 75 and above or 25 and below, the deviation value sought is either 75 or 25. Please refer to Appended figure 2.1.5 for the average values of proxy variables and cover rates, etc. Appended figure 2.1.5 Basic data for each proxy variable Component Business structure reform capacity Business efficiency Proxy variable Changes in turnover ratio of total capital Changes in the ratio of operating profit to sales Inventory asset turnover ratio Turnover ratio of tangible fixed assets Cumulative R&D expenses Technological capacity Average value Standard deviation Maximum value Minimum value Number of Coverage covered rates companies 0.02 times 0.27 times 1.08 times -1.03 times 7,052 89.3% -2.3% 10.3% 39.0% -44.5% 6,477 82.0% 8.78 times 7.60 times 61.34 times 0.81 times 7,458 94.4% 6.97 times 53.61 times 0.23 times 7,694 97.4% 159.729 bn yen per 2.126578 bn yen per 2 m yen per company company per year company per year per year 2,269 28.7% 5.54 times 36.517 bn yen per company per year Cumulative sales / Cumulative R&D expenses 48.18 times 435.50 times 17,312.35 times 0.04 times 1,635 20.7% Credibility 0.72 times 0.65 times 5.44 times 0.06 times 6,146 77.8% Sales share 0.25% 1.19% 46.76% 0.00007% 7,897 100.0% 26.039m yen/person 21.323m yen/person 192.595m yen/person 139,000 yen/person 5,565 70.5% 4,635 58.7% Marketing capacity Organizational capacity Employee productivity Changes in number of 1.01 times employees Note: Denominator for "coverage rate" is the 7,897 companies that are analyzed. 0.37 times 12.60 times 0.00 times (a) Changes in turnover ratio of total capital (the latest term, two terms before the latest term) ・ 655 companies for which the turnover ratio of total capital in both the latest term and two terms before the latest term could not be calculated were excluded from the analysis. -288- ・ The average value and standard deviation were sought for the numerical value of the turnover ratio of the total capital two terms before the latest term subtracted from the ratio in the latest, and the samples of 190 companies for which there was a divergence of 2.5 times or more the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 7,052 companies remaining after the above process, and the deviation value of each company was calculated based on this. (b) Changes in the ratio of operating profit to sales (the latest term, two terms before the latest term) ・ 652 companies for which the ratio of operating profit to sales in both the latest term and the two terms before the latest could not be calculated were excluded from the analysis. ・ The samples of 487 companies for which the numerical value of the ratio of operating profit to sales two terms before the latest term subtracted from the ratio in the latest is not within the range of ±100% were regarded as outlying values and excluded. ・ The average value and standard deviation were sought for the numerical value of the ratio of operating profit to sales two terms before the latest term subtracted from the ratio in the latest, and the samples of 281 companies with a divergence of 2.5 times or more the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 6,477 companies remaining after the above process, and the deviation value of each company was calculated based on this. (c) Inventory asset turnover ratio (the latest term) ・ 208 companies for which inventory asset turnover ratio in the latest term could not be calculated were excluded from the analysis. ・ The average value and standard deviation were sought for the logarithmically transformed numerical value of the inventory asset turnover ratio in the latest term, and the samples of 231 companies with a divergence of 2.5 times or more the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 7,458 companies remaining after the above process, and the deviation value of each company was calculated based on this. (d) Turnover ratio of tangible fixed assets (the latest term) ・ 22 companies for which the turnover ratio of tangible fixed assets in the latest term could not be -289- calculated were excluded from the analysis. ・ The average value and standard deviation were sought for the logarithmically transformed numerical value of the turnover ratio of tangible fixed assets in the latest term, and the samples of 181 companies with a divergence of 2.5 times or more the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 7,694 companies remaining after the above process, and the deviation value of each company was calculated based on this. (e) Cumulative R&D expenses (three latest terms) ・ 5,628 companies for which the numerical value of R&D expenses in one or more of the latest three terms could not be obtained were excluded from the analysis. ・ The average value and standard deviation were sought for the numerical values of the 2,269 remaining companies, and the deviation value of each company was calculated based on this. (f) Cumulative sales (three latest terms) / cumulative R&D expenses (three terms before-five terms before) ・ 6,262 companies for which one or more data was lacking for the data on sales from the three latest terms and data on cumulative R&D expenses from three terms before to five terms before were excluded from the analysis. ・ The average value and standard deviation were sought for the numerical values of the 1,635 remaining companies, and the deviation value of each company was calculated based on this. (g) Credibility=trade payables (the latest term) / trade receivables (the latest term) ・ 1,583 companies for which credibility could not be calculated in the latest term were excluded from the analysis. ・ The average value and standard deviation were sought for the logarithmically transformed numerical value of credibility in the latest term, and the samples of 168 companies with a divergence of 2.5 times or more the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 6,146 companies remaining after the above process, and the deviation value of each company was calculated based on this. (h) Sales share (the latest term) ・ Companies with the same two-digit US-SIC code are placed into groups of companies in the same -290- industry, and sales share is a company’s sales ratio of the total sales in the group of companies in the same industry in the latest term. ・ The average value and standard deviation were sought for the sales share of the 7,897 companies analyzed, and the deviation value of each company was calculated based on this. (i) Sales per employee (the latest term) ・ 2,307 companies for which sales per employee could not be calculated in the latest term were excluded from the analysis. ・ The average value and standard deviation were sought for the logarithmically transformed numerical value of sales per employee in the latest term, and the samples of 25 companies with a divergence of 2.5 times the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 5,565 companies remaining after the above process, and the deviation value of each company was calculated based on this. (j) Changes in number of employees (Number of employees <the latest term + one term before the latest> / number of employees <one term before the latest+ two terms before>) ・ 3,259 companies for which the number of employees could not be calculated from the latest term to two terms before were excluded from the analysis. ・ The average value and standard deviation were sought for the numerical value of changes in the number of employees, and the samples of three companies with a divergence of 2.5 times the standard deviation from the average value were regarded as outlying values and excluded. ・ A recalculated average value and standard deviation were sought for the numerical values of the 4,635 companies remaining after the above process, and the deviation value of each company was calculated based on this. (2) Integration of components of each proxy variable A simple average of the proxy variables from which deviation values were derived was taken for each company by component (business structure reform capability, business efficiency, technological capability, marketing capacity, and organizational capacity), and a score is given for each component. If two of the proxy variables for a component were not complete, then the score was excluded only in the component. -291-