Gambar halaman

steel castings during 1953 is about 5 percent of the total alloy and stainless steel, and less than 1 percent of the total carbon steel consumed in manufacturing. The industries for which the comparability is affected by more than 5 percent are footnoted in their respective tables.

The 1953 and 1950, data were collected on consumption of iron castings. These data are not available for 1952 and 1951, since establishments were not required to report consumption of iron castings in the NPAF-103 survey.

c. Consumption of Forgings:

Forgings are defined as a fabricated product in the surveys. Since consumption data were collected on mill shapes, and not on any fabricated products, the forgings industry (3391) was instructed to report its consumption of steel (ingot, billet, and bars) in the production of forgings, in order to account for this quantity of steel consumed.

d. Limitations of Data on Metals Consumed:

In making comparison with statistics on the production of mill shapes and castings, it should be noted that the metals consumed data shown in the tables are limited to the major metal fabricating industries, described in lla above. With the exception of the Iron and Steel Forgings Industry (3391), therefore, these tables exclude the consumption in those primary metals industries (major industry group 33) also engaged in fabricating metal products, as well as in such other significant metal consuming areas as Furniture and Fixtures (Major Group 25). In 1947, it is estimated that these industries consumed 10 percent of all carbon steel, and less than 5 percent of the total alloy and stainless steel consumed in manufacturing. Similarly, consumption in nonmanufacturing activities, as construction, transportation, mining, agriculture, and public utilities, and shipments for export are not covered. For each metal shown, it is believed that the estimates of consumption in the specified manufacturing areas represent between 60 and 70 percent of the total consumed in all areas.

12. Fuels, Electric Energy, and Water Use a. Fuels and Electric Energy Used:


Data comparable with those published in the previous Annual Surveys are presented for 1953, by major industry group, on the cost of fuels and cost of electric energy purchased for use manufacturing establishments. Data on quantity of electric energy, however, are not comparable with those shown for the prior Annual Survey years, as the 1953 figures represent electric energy used, while previous Annual Surveys collected statistics on electric energy purchased for use. The difference between these two totals is the generation of electric energy by manufacturing establishments for their own use (i.e., total generation by manufacturing establishments, less their sales


for the

Comparable figures are shown quantity of electric energy generated at manufacturing plants in 1947, 1950, 1951, 1952, and 1953. The quantity of electric energy generated data include electricity generated for use at the manufacturing plant and also electricity generated for sale or for transfer to other establishments of the same company. The figures for 1950-1953 are based on data collected by the Federal Power Commission from all known manufacturing establishments at which electric energy was generated. The 1947 figures on electric energy generated, as well as cost of fuels and electric energy, and quantity of electric energy purchased, were collected in the 1947 Census of Manufactures. The statistics on sales of electric energy by manufacturing establishments were also collected in the 1947 Census of Manufactures but not in the Annual Surveys.

The figures on "cost of fuels" exclude fuels used as a raw material, such as coal used in making coke. Also excluded, in order to avoid duplication in fuel and energy cost, are the values of fuels and electric energy produced and consumed in the same establishment. The cost of fuels used in generating electricity, however, is included.

b. Water use:

The 1953 Annual Survey included, for the first time, an inquiry on water use. Manufacturers having a total water intake during the year of 20 million gallons or more were asked to report the total quantity of water intake, the quantity of brackish water, the principal source of fresh water, and whether significant amounts of total water intake were recirculated or re-used. It is estimated that the statistics presented in this volume represent between 90 and 95 percent of the water used in manufacturing establishments.

13. Estimating Procedure Used in the 1953 Annual Survey

Most of the estimates for 1953 were calculated by the formula for simple unbiased estimates of totals; i.e., each sample establishment's data were inflated by the reciprocal of the establishment's sampling rate and the inflated figures were summed to publication levels. This procedure was used for all the 1953 estimates, except the data in Chapter III, Value of Shipments by Selected Product Classes.

The product class totals were calculated by a ratio formula which took advantage of the fact that 1952 as well as 1953 product class data were reported on the 1953 Annual Survey report form. Inflated product class totals for both years were computed, in the same manner as for the other statistics, from the reports received for sample establishments that operated during both years. From these inflated totals for identical establish

1953 product class aggregates were then derived by multiplying the corresponding 1952 product class estimate, as derived from the 1952 Annual Survey, by these ratios. The 1953/1952 relatives of Chapter III were calculated by multiplying these estimated ratios by 100.

Some sample establishments submitted incomplete reports. In such cases the omitted data were inferred from related, reported figures for the same establishment, on the basis of historical average relationships. More importantly, a similar procedure was used for those sample establishments that did not submit Annual Survey reports. Most of these non-respondents had been omitted from the canvass. Some had been omitted because of mailing errors. The majority were manufacturers that had been temporarily inactive during the first quarter of 1951 and who could not be identified as currently active in time to mail them report forms. Special efforts were made to obtain reports directly from respondents for all large establishments, generally those with 100 or more employees, when the lists of omissions became available. For the smaller, sample nonrespondents, reports were imputed based on the BOASI industry classifications and the BOASI first quarter 1953 wages and salaries, and March 1953 employment data.

Product class data were not imputed in the same manner as missing data for other statistics. The ratio formula which was used to calculate the product class estimates automatically imputed the missing figures. In effect, it assigned the average ratio change, as estimated from the sample of reporting identicals, to the non-respondents. This method was preferable to individual establishment imputations, because the product class distribution may vary greatly from one establishment to another even within the same industry. Moreover, because the ratio formula took account of the correlation between individual establishments' activities in two successive years, a correlation which usually is high, it generally yielded substantially more reliable estimates of the 1953-1952 relatives than would have been obtained from the ratio of independent sample estimates of the two annual totals.

(NOTE: For a description of the estimation procedures used for the 1952 and earlier annual surveys, see the 1952 Annual Survey of Manufactures volume.)

14. Qualifications of the Published Data

All the Annual Survey estimates in this report are affected by random sampling errors, and to some extent by systematic reporting and procedural errors. The estimates, therefore, should be regarded as approximate rather than exact


The standard error columns and the explanatory headnotes and footnotes of the tables indicate the general magnitude of the of the random sampling errors. Specifically, the standard errors show the approximate range on either side of the published estimates within which a corresponding complete Census total would lie, if the Census was conducted under conditions that were strictly comparable to those of the Annual Survey.

The standard errors as published also cover those reporting and other non- sampling errors that would balance out exactly in a complete Census (when conducted under Annual Survey conditions). To the extent that undetected reporting or other systematic errors (clerical, tabulation, coverage, imputation, etc.) would not balance out in a comparable complete Census, the Annual Survey estimates tend to be too high or too low, but on the average by the same amount as such errors would affect the complete canvass totals. Measures of such biases' are not available. However, estimates which were subject to large uncorrected systematic errors were apt to have been detected and corrected or withheld from publication in the process of Census Bureau review of the data for reasonableness and for consistency with other published series. It is believed that the biases of the published estimates are small in comparison with the totals or with the standard errors of the published totals.

Imputed small establishment reports, which were developed solely from the BOASI wage and employment data accounted for less than 1 percent of the estimated total manufacturing employment etc. Imputation errors for this group, therefore, could have had only a trivial effect on most of the estimates. They had their most serious impact on the data for the apparel industries, and accordingly, on the data for New York, the only notable point of concentration.

Approximate formulas were used to estimate the standard errors which tend to overstate them except for the errors of the 1953/1952 product class relatives.

The 1953/1952 product class ratios are "biased,' inasmuch as they were derived only from the sample identicals, "' an incomplete sample which represented less than the entire universe. The corresponding standard. errors were estimated from the same incomplete sub-set of the sample. They do not reflect the bias due to incomplete representation, and therefore tend to understate the errors of the product class relatives. (In addition, ratio estimates of this type are mathematically biased, possibly seriously for small items..) However, in most cases the identical units accounted for an extremely high percentage of the totals so that the under-representation bias is likely to be small, and the published "standard errors" of the relatives are reasonably accurate.

[blocks in formation]



(DOLLAR FIGURES ARE IN THOUSANDS. Data for 1947-1953 include all establishments employing one or more persons at any time during the year; for 1921-1939, those with products valued at $5,000 or more; and for 1849-1919, those having products valued at $500 or more. For basis of revisions beginning with 1899, see 1947 Census of Manufactures volumes, GENERAL EXPLANATIONS - Statistics for 1939 and earlier years.)

[blocks in formation]

1The figures for 1939-1953 include data for employees at manufacturing establishments who were engaged in distribution and in construction work. The extent to which data for such employees were included in the figures for earlier years is not known.

2The 1949 thru 1953 "Average number of employees" is based on reported employment totals for the payroll periods nearest the 15th of March, May, August, and November. For 1947, the number of "All employees" represents the average of 12 monthly figures; for 1939 and earlier years, it represents the sum of the average number of production workers for the year and the number of nonproduction workers reported for one payroll period (usually in October). See 1947 Census of Manufactures volumes, GENERAL EXPLANATIONS - Monthly and average employment.

3Value of products shipped less cost of materials, supplies, fuel, electric energy, and contract work. For 1849-1933, cost of contract work was not subtracted from value of products in calculating value added by manufacture. See 1947 Census of Manufactures volumes, GENERAL EXPLANATIONS

Value added by manufacture.

4Data for 1949 thru 1953 are estimates based on establishments' reports received in the Annual Survey of Manufactures. The survey estimates, of course, vary from the totals that would have been obtained from a complete canvass as conducted in prior years; the relative magnitude of this sampling variation is indicated in Chapter I, table 2, in the column captioned "Standard error of estimate."

"The number of establishments classified as engaged primarily in manufacturing activities is estimated at 285,000 for 1953, 267,000 for 1952, 262,000 for 1951 and 260,000 for 1950. (These estimates have a standard error of 2 percent.) Because of sampling variation, and some differences in definitions, scope, etc., these totals may differ to some extent from the establishment counts presented in the joint Census-BOASI publication County Business Patterns.

The totals shown for 1949 thru 1953 include employees separately reported at central administrative offices and auxiliary establishments. "The figures for 1939, but not for earlier years, have been revised on the basis of retabulation of the returns to exclude data for establishments classified as manufacturing in 1939 and prior years but classified as nonmanufacturing beginning with 1947. Value added by manufacture for 1939, prior to revision and on a basis comparable with 1937 and previous years, was $24.7 billion. See 1947 Census of Manufactures volumes, GENERAL EXPLANATIONS - Statistics for 1939 and earlier years.

Not revised to exclude data for establishments classified as manufacturing in 1939 and prior years but classified as normanufacturing beginning with 1947 (See footnote 7).

The 1939 figures for "All employees" were revised on the basis of estimates rather than by retabulation of the 1939 reports. The estimates were made in the following manner: for number of employees, by multiplying the retabulated figure for number of production workers by the ratio of all employees to production workers computed from the unrevised 1939 statistics; for salaries and wages, by multiplying the retabulated wage figure by the ratio for salaries and wages also derived from the unrevised 1939 statistics.

10Not including data for salaried officers of corporations and their salaries and, therefore, not strictly comparable with figures for other 11Reduced to gold basis.



Table 2.--GENERAL STATISTICS, BY INDUSTRY GROUPS: 1953, 1952, 1951, 1950, 1949, AND 1947

added by manu

(DOLLAR FIGURES AND MAN-HOURS ARE IN THOUSANDS. Only selected sample error percentages are shown; standard errors of the other general statistics estimates for an industry group are usually of the same general magnitude as the standard errors shown for employment and value facture.)

[graphic][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][ocr errors][ocr errors][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][ocr errors][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][ocr errors][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed][subsumed]
« SebelumnyaLanjutkan »