PARTE I: LA REGULACIÓN DE MERCADOS
CAPÍTULO 1. MARCO TEÓRICO DE LA ECONOMÍA DE LA REGULACIÓN
1.2. Intervención del sector público en el mercado de bienes y
1.2.1. Teorías del interés público: los fallos de mercado
D.1 Description
(a) Sources
This database was constructed from two sources, both of which relied on the China Credit Information Service (CCIS), located in Taipei, Taiwan:
(1). Business Groups in Taiwan, 1996/1997, published by CCIS (contains data for 1994) (2). Company annual reports to the Taiwan stock exchange, for 1994. This data was collected by the CCIS, and supplemented by interviews of firms conducted by CCIS.
Source (1) lists membership, sales and other financial information for the firms within 115 business groups in Taiwan, but provides only qualitative information on whether a firm transacts with another firm in that group or not. The quantitative information on the value of internal transactions was obtained from source (2).
A principal limitation of the data collected in (2) is that it did not cover exactly the same set of firms as indicated by group membership in source (1). For example, if a firm belonged to a business group, but the qualitative data in (1) indicated that it had little or no transactions with the other firm in that group, then no information for that firm was collected from source (2). This means that the set of firms for which information was collected in source (2) are only a subset of the firms in each group (namely, those which engage in transactions with other firms in the group). The information from source (2) was received in Chinese, and translated by Shunli Yao during 1996; this data appears in TORIG94.XLS.
This information from source (2) therefore had to be extensively supplemented with the data from source (1) in order to obtain a complete set of the firms in each group and their internal transactions and sales data. This additional data was added by Shunli Yao during 1996 and 1997, and the resulting dataset is in TGROUP94.XLS. The database includes the 80 largest business groups from the 115 groups identified in source 1, and there are 796 firms belonging to these groups in total. All data is for 1994, and measured in thousand of New Taiwan dollars.
The sector classification was based on the 1989 Input-Output Tables: Taiwan Area, the Republic of China, published by the Directorate-General of Budget, Accounting and Statistics, Executive Yuan, Republic of China. The 123 sector classification was used (the same
classification is also used the 1984 IO Table). The IO numbers were assigned based on the companies made products as identified in source 1. In most cases a firm produces several products, so the first product listed in source (1) was used. The list of IO numbers and names appears in T_IO123.XLS.
(b) Trading Companies
Trading companies are identified in the dataset. This designation was based upon two factors: if the name of the company from source (1) included the words “trading company”; or if the description of products from source (1) indicated “buying and selling” as a primary activity. The trading companies in most cases belong to the IO sectors 89 (domestic wholesale trade), 90 (domestic retail trade) and 91 (import and export trade). Note that only a subset of the firms with these sector classifications are designated as trading companies.
(c) Company codes and names
The groups and firms are identified by code numbers, which are the same as the codes used in the 1983 and 1986 Taiwan data. The English names for the groups in 1994 are contained in TGNAME94.XLS, while the English or Chinese names for the individual firms for 1994 or earlier years appears in the file TFNAME94.XLS.
D.2 Layout of TORIG94.XLS
This was the information received from the China Credit Information Service (source 2), which was translated into English. On the left of the sheet is a matrix for each group. This matrix gives internal sales from the firms listed in each row, to the firms listed in each column. Groups are stacked on top of each other in this matrix. All data are in NT$1,000, expect for a small number of firm sales that are in the foreign currencies, such as US$ or HK$. The last rows of the matrix list (i) purchases of that column firm from outside firms; (ii) total purchases of inputs by each column firm; and (iii) total group sales. It should be stressed that the information in (i) and (ii) is
incomplete: the data provided on total input purchases may not equal the sum of purchases from within and outside the group; and firms are likely to be purchasing additional products from other firms that are not included in the data. These cases are indicated with a * on the spreadsheet, and will be corrected in TGROUP94.XLS. Missing data may appear as either a "na" (not available), or as the number zero, or as a blank field.
On the right side of the sheet are listed additional data for each row firm. Thus, column AH lists sales outside the group, and column AI then sums each row to list total sales. Note that the data on total group sales in (iii) does not necessarily equal the sum of individual firm sales (as listed in column AI). This is because the value for total group sales in (iii) was not taken from the survey, but instead, was obtained from the published volume of CCIS. It may be inconsistent with the survey data of individual firms. This inconsistency is corrected in TGROUP94.XLS. Additional columns provide information on the firm's: export sales (AJ); capital value (AL) at historical cost; and the number of employees (AM). In a few cases, a currency other than NT$1,000 is used for the capital value, such as US$ or HK$.
At the far-right of the sheet is a correspondence between each group number, the location of this group in the CCIS book, and the file number of the original data from CCIS (these were computer spreadsheets in Chinese).
D.3 Layout of TGROUP94.XLS
Sheet 1 - With Trading Companies
This revises the data from TORIG94.XLS. Added to the information was:
• the sector number of each firm, taken from the 123 sector 1989 Taiwan IO Table.
• additional firms listed in the CCIS book, source (1), but not included in source (2). These firms have no internal transactions, but are included in terms of their total sales. The data in each group transactions matrix are not included for these added firms, but it can be generally assumed that there are zero transactions between these added firms and other firms in the group.
On the left of the sheet is a matrix for each group. This matrix for each group gives sales from the firms listed in each row, to the firms listed in each column. Groups are stacked on top of each other in this matrix. All data in this sheet (and the following sheets) is in NT$1,000, for 1994. Missing data may appear as either a "na" (not available), or as the number zero, or as a blank field. The first two columns list the group number and firm number, and the third lists the sector. The firms numbers also appear within the first row for each group matrix. Generally, the ordering of firms in the rows and columns are the same, except that the firms that have been added to the rows from source (1) often do not appear in the columns (since these firms do not buy anything from other firms in the group). On the right of the worksheet are the following columns:
AG - Internal sales, obtained from summing values on each row AH - External sales outside the group, revised from TORIG94.XLS AI - Total sales of each firm, obtained by summing AG+AH
AJ - Group internal sales, obtained by summing down column AG for each group AK - Group total sales provided in TORIG94.XLS, or revised using source (1) AL - Group total sales (with trading co.), obtained by summing down AI
AM - The maximum of AK and AL. These rows can differ due to missing observations on firm sales. In most case the difference between AK and AL is small.
AO - Capital value, at historical cost AP - Number of employees
AQ - Whether the firm is designated as a trading company, indicated by the value 1. AR - A count of the number of groups with trading companies
AV - An estimate of the value-added for each firm, equal to AP4*31.833*12+AO4*0.35, where 31.833 is the annual wage for each employee (in NT$1,000), and 0.45 is the return to capital. See the notes on this formula below.
AW - Total purchases of all inputs (from within or outside the group) by each member firm. This is taken from TORIG94.XLS.
It should be noted that the total purchases of all inputs in column AW, taken from
TORIG94.XLS, is missing for quite a number of firms, and may be an underestimate even for the firms for which it is included. Thus, when the total purchases of inputs are summed across firms in each group, as in column AZ, the resulting “total input purchases of the group” is very likely to be an underestimate. According, a second estimate of this variable was constructed in column BA by taking the difference between estimated value added for the group (column AZ) and total sales of the group (column AM). The maximum of these two estimate for “total input purchases of the group” is contained in column BB.
The group value-added figure used in this calculation (column AY) is summed from the firm-level value added estimate (column AV), which uses the formula indicated above. The wages used in this formula are the average monthly earnings of employees for total industry in Taiwan, 1994. The return to capital used – 0.35 or 35% – is an estimate that is constructed to that the ratio of value-added to sales over all the Taiwan firms is the same as that for all Korean firms, or 21%. This return to capital is rather high because the capital values represent the historical cost of capital rather than the current cost.
Sheet 2: Sales by Group
This sheet provides summary information on the internal sales by business group. This is compared to the two estimates of the total group sales (taken from either the total group sales figure provided in column AK of sheet 1, or alternatively from summing the sales of the individual firms for each group, in column AL of sheet 1). The internal sales ratio listed in column AH of sheet 2 is computed using the maximum of these two estimates.
Sheet 3: Without Trading Companies
This is the same to sheet 1, With Trading Companies, except that the values in each column are set equal to zero for the trading companies. Thus, these values are not included in the internal purchases or in total group sales, recorded in columns AG, AI, AJ, AK or AL, which are constructed using the transactions matrix for each group. As in sheet 1, column AM gives the maximum of the constructed and provided groups sales, which are usually quite close.
Column AN in sheet 3 in another calculation of the internal sales of the groups, which not only sets the internal purchases of the trading companies at zero, but also sets the purchases of all other wholesale/retail firms in sectors 89, 90 or 91 at zero. This is constructed using the
information in columns AO, which gives the sales of each non-trading company in sectors 89, 90 or 91, and column AP which sums these sales over each group.
Sheet 4: No Trading Companies, Sale by Group
This is computed in the same manner as sheet 2, Sales by Group, except that the purchases of the trading companies are excluded.
Sheet 5: Group Summary
This summarizes information from sheets 2 and 4. Sheet 6: Sector Subtotals, and Sheet 7, Sector Totals
These sheets total up the sales of the business groups, by IO sector. Sheet 8: Manufacturing
This sheet takes the total sales of the business groups by broad manufacturing sector, and compares it to the total sales from all manufacturing firms in each sector. The total sales figures are taken from the 1995 Statistical Yearbook of the Republic of China, Executive Yuan,
Republic of China, p. 191, which lists 1994 data.
6.4 Errors and Corrections
i. The following groups had the values for total sales in TORIG94.XLS replaced with those from source (1), in order to obtain the values in TGROUP94.XLS: 1026, 1063, 1087, 1142, 1062, 1082, 1115, 1120.
ii. Group 1119, firm 1394, has a sales figure of -149,985 listed in source (1). This was revised to 149,985, but appears to be too low since the calculated value of sales for the group,
obtained by summing the individual sales of the firms, is substantially lower that the provided total sales of the group from source (1).
iii. Group 1123, firm 1425, has a sales figure of US$10.4 billion listed in source (1). This appears to be incorrect (far too high), so NA was entered instead in TGROUP94.XLS. iv. The following groups had missing or incorrect value for firm sales in TORIG94.XLS, which
were filled in or corrected using the data from source (1): 1056, 1092, 1142, 1025, 1028, 1030, 1058, 1081, 1115, 1127, 1128, 1131.
v. Group 1062, firm 445 had NT$ 198.6 billion listed in “other purchases” in TORIG94.XLS. This appears to be incorrect (far too large), so NA was entered instead in TGROUP94.XLS. vi. Group 1140, firm 1284, is the Aurora Mechatronics Corp. It was missing a total sales value
from source (1). CCIS provided its 1991 sales as NT$ 1.6 billion, which is used in TGROUP94.XLS. CCIS also informed us that firm 1283 is a trading company.
vii. Group 1190, firms 1607 and 1608, are Shanghai Vedan and Vedan Vietnam, respectively. They have internal purchases listed in TORIG94.XLS, but were missing total sales values. Since these could not be obtained from sources (1) or elsewhere, the internal sales and purchases for these two firms were changed to NA in TGROUP94.XLS.