Contributed by:
Keith E. Maskus Allan Webster
Dept. of Economics 9 Norton Road, Woodley
Campus Box 256 Reading RG5 4AH
Univ. of Colorado United Kingdom
Boulder, CO 80309-0256
E-mail: Keith.Maskus@Colorado.edu
Jose Campa Linda Goldberg
Stern School of Business Research Department, 3rd Floor
New York University Federal Reserve Bank - New York
44 W. 4th Street 33 Liberty Street
New York, NY 10012 New York, NY 10045
E-mail: jcampa@stern.nyu.edu Linda.Goldberg@frbny.sprint.com
The CD-ROM includes two sets of input-output tables for the United Kingdom and the United States. The first set uses an 80-sector classification for industries, and was contributed by Keith Maskus and Allan Webster. The U.K. table is for the year 1989 (IO1_UK89.ASC), while the U.S. tables are for the years 1987 and 1990 (IO1_US87.ASC, IO1_US90.ASC). These tables can be referenced to the papers:
Keith Maskus, C. Sveikauskas, and Allan Webster, "The Composition of the Human Capital Stock and Its Relation to International Trade: Evidence from the U.S. and Britain," Weltwirtschaftliches Archiv, 1994, Band 130, Heft 1.
Keith Maskus and Allan Webster, "Factor Specialization in U.S. and U.K. Trade: Simple Departures from the Factor-Content Theory," Swiss Journal of Economics and Statistics, 1995, Vol. 1, no. 131.
Descriptions of the files and their contents are below, followed by notes on how these data were collected and compiled. More detailed notes are available from Maskus.
This file contains first an (85x80) input-output direct use matrix, with entries in millions of 1989 pounds. There are 80 production sectors in this table, which was aggregated to be consistent with the industrial classification in the U.S. tables below. The first 80 rows provide data on intermediate inputs into each column industry. The final five rows provide further data for each sector on taxes, value added, total output, and so on as labeled. Industry names are provided in each row. Below this table is calculated, in a second (80x80) table, the intermediate coefficients matrix (often called the "A" matrix in IO work), which is simply each column entry of the use matrix divided by the column total output figure; thus these are input value shares or intermediate intensities. Those interested in using a "Leontief Inverse Matrix" would need to calculate (I-A)-1 themselves. Below this is a (9x80) table providing direct input coefficients (input use in pounds per pound of output) for certain primary inputs, made up of certain natural resources and physical capital types. These coefficients are calculated from the use matrix itself.
This file contains first an (83x80) input-output direct use matrix, with entries in millions of 1987 dollars. There are 80 production sectors, with the aggregation consistent with the UK matrix. The first 80 rows provide intermediate inputs and the final three rows provide total output, intermediate uses, and value added by column. Below this table is the (80x80) intermediate coefficients matrix and below that is the (9x80) matrix of direct coefficients for natural resources and physical capital types. Again, these last coefficients are calculated from the use matrix.
This file contains first the (82x80) input-output direct use matrix, with entries in millions of 1990 dollars. The 80 production sectors have been aggregated to be consistent with the 1987 matrix. The extra two rows contain value added and total output by column. Below this table is the (80x80) intermediate coefficients matrix. Below that is a (29x80) direct coefficients matrix, including again the nine natural-resource and physical-capital inputs (calculated from the use table) and an additional 20 categories of occupational employment. Thus, the coefficients are calculated as workers per million dollars of output in each sector, with the occupational employment data themselves listed in a final (20x80) table. These occupational employment figures by industry were aggregated from a different source as discussed below.
Input-output tables for 1987 and 1990 for the United States were purchased on diskette from Bureau of Labor Statistics (BLS). These tables include 228 sectors organized into a standard BEA classification system. The 1989 IO table for the U.K. was taken from Input-Output Balance for the United Kingdom, 1989, Central Statistical Office, London, 1992. This table contains 102 sectors. We developed a concordance between the IO classification systems for purposes of aggregating the IO tables into 80 sectors in each country. Details of this aggregation are available from Maskus; it is a painstaking task. Aggregation consists of performing column and row sums across sectors, with sometimes a need to assign portions of sub-sectors to different aggregate sectors. While we believe this aggregation is as careful as possible, errors undoubtedly were made in the aggregation and users should make allowance for measurement error arising from this procedure.
An even more difficult task was assembling the occupational employment data for the United States. Every five years BLS compiles highly detailed data on occupational employment by (approximately) 5-digit SIC sector (including manufacturing, non-manufacturing, and services). The data base includes 507 occupational categories in 240 industries. The occupational categories are summarized in BLS, Occupational Projections and Training Data (1992), which covers 1990. Readers should be aware that these categories change each five years. The actual data base (204x507) is available on diskette from BLS for 1990, 1985, 1980, etc. It is likely that the 1995 data will become available in 1997. We developed a concordance between SIC and BEA categories for industrial matches and aggregated the occupations into the 20 types listed in the 1990 IO tables. We also have aggregations for a finer set of 74 occupations, which we used in the first paper cited above.
The second set of input-output tables uses a 20-industry classification based on the two-digit U.S. SIC categories, and were contributed by Linda Goldberg and Jose Campa. The U.K. table is for the year 1990 (IO1_UK90.ASC) while the U.S. tables are for the years 1982 and 1987 (IO1_US82.ASC, IO1_US87.ASC). These files also contain annual time series on sectoral export share, import share, and a constructed measure of imported input share. These data span about 1970 through 1995. The methodology for constructing the imported input share series is described and applied in the papers:
Campa, Jose and Linda Goldberg, 1995, "Investment in Manufacturing, Exchange Rates and External Exposure," Journal of International Economics, 38(3/4), May, 297-320.
Campa, Jose and Linda Goldberg, 1996, "Investment, Endogenous Markups, and Exchange Rates: A Cross-Country Comparison," March (revision of NBER working paper # 5139, June 1995).
Campa, Jose and Linda Goldberg, 1996, "The Evolving External Orientation of Manufacturing Industries," November.
File Layout:
The layout of the files is self-explanatory, with documentation provided within them. Each column in an input-output table corresponds to a purchasing industry, and each row corresponds to a supplying industry. Details for each country and data sources are provided below.
Missing values:
For non-manufacturing industries, the data in the U.S. input-output tables may be missing. For example, in the 1987 U.S. table, the upper-left element represents the purchases of industry 1a from industry 1a (non-manufacturing), and is reported as zero. This is actually a missing value.
The input-output table was aggregated from the 1990 U.K. Table, as described in the concordance below. . Each coefficient in the input-output table represents the value of materials purchased (in millions of pounds) by the column industry from the row industry. This data is combined with import share measures (from the OECD STAN database, but not reported here) to compute the imported input into production series for 1970 to 1992 The imported input series captures the share of imported inputs from agriculture, mining and manufacturing, but does not capture the role played by other non-tradable products. The export to production and imports to domestic supply are also computed using information from the OECD STAN database (not reported here).
INDUSTRY CONCORDANCE FOR THE UNITED KINGDOM: ISIC INDUSTRY 1990 UK I-O TABLE CODE 311-312 Food 58-67 313 Beverages 68-69 314 Tobacco 70 321-322 Textiles and Wearing Apparel 71-76, 79-80 323-324 Leather & Leather Products 77-78 331 Wood Products 81 332 Furniture & Fixtures 82 341 Paper & Products 83-84 342 Printing and Publishing 85 351-352 Chemical Products 20-29 353-354 Petroleum & Coal Products 6 355 Rubber Products 86 356 Plastic Products 87 361,362,369 Non-metallic Products 15-19 371 Iron & Steel 11 372 Non-Ferrous Metals 12-13 381 Fabricated Metal Products 30-33 382 Non-Electrical Machinery 34-43 383 Electrical Machinery 44-52 384 Transport Equipment 53-56 385 Professional Goods 57 390 Other Manufacturing 88-90
Sources:
OECD, Industrial Structure Statistics, several issues. Data on exports, imports, production, value added, wages & salaries to employees and employers' social security costs.
Central Statistical Office of the United Kingdom (1993), 1990 Input-Output Balances for the United Kingdom, Table 3.
The input-output tables are aggregated from the 1982 and 1987 Benchmark U.S. Input-Output Tables, as described in the concordance below. Each coefficient in the input-output table represents the value of materials purchased (in millions of dollars) by the column industry from the row industry. Cost of materials is defined as the sum of total purchased inputs from manufacturing industries and by the total costs of compensation to employees. The measure of costs of materials does not include the costs of purchases from: agriculture, mining, wholesale and retail trade, public utilities, finance and insurance services and other services. The reason for excluding these industries is the lack of reliable international trade information for these industries.
INDUSTRY CONCORDANCE FOR THE UNITED STATES:
SIC INDUSTRY I-O TABLE (2-digits)
20 Food and kindred products 14
21 Tobacco products 15
22 Textile mill products 16-18(excluding 18.0400)
23 Apparel and other textile products 18.0400 and 19
24 Lumber and wood products 20, 21 and 61.0602
25 Furniture and fixtures 22, 23
26 Paper and allied products 24, 25
27 Printing and publishing 26
28 Chemicals and allied products 27 to 30
29 Petroleum and coal products 31
30 Rubber and misc. plastic products 32
31 Leather and leather products 33 and 34
32 Stone, clay, and glass products 35 and 36
33 Primary metal products* 37 and 38
(*excluding 37.0300 and 38.1400)
34 Fabricated metal products 39 to 42 and 37.0300 and 38.1400
35 Industrial machinery and equipment 43 to 52
36 Electronic & other electric equipment 53 to 58 (excluding 53.0100)
37 Transportation equipment 59 to 61 (excluding 61.0602)
38 Instruments and related products 62, 63 and 53.0100
39 Miscellaneous manufacturing industries 64
Sources:
Benchmark Input-Output Accounts for the U.S. Economy, 1982, Survey of Current Business July 1991, 30-71 and Appendix A.
Benchmark Input-Output Accounts for the U.S. Economy, 1987, Survey of Current Business, 74(4), April 1994, 73-115.