Annual Survey of Manufactures (ASM)

Detailed information for 2001

Status:

Active

Frequency:

Annual

Record number:

2103

Data collected from the Annual Survey of Manufactures are used to measure production of the industrial sector in Canada.

Data release - May 2, 2003

Description

The Annual Survey of Manufactures (ASM) is a survey of the manufacturing industries of Canada conducted annually since 1917. It is intended to cover all manufacturing establishments as well as sales offices associated with these establishments.

Details collected include principal industrial statistics (such as shipments, employment, salaries and wages, cost of materials and supplies used, cost of energy and water utility, inventories, goods purchased for resale, etc.) and commodity data, including shipments or consumption of particular products.

Data collected by the Annual Survey of Manufactures are important because they measure production of the industrial sector in Canada as well as provide an indication of the well-being of each manufacturing industry, and its contribution to the Canadian economy. The data are used within Statistics Canada by the Canadian System of National Accounts, the Monthly Survey of Manufacturing (Survey ID 2101) and Prices programs. These data are also used by the business community, trade associations, federal and provincial departments and international organizations and associations to profile the manufacturing sector, to undertake market studies, to forecast demand and to develop trade and tariff policies.

Statistical activity

The survey is administered as part of the Unified Enterprise Survey program (UES). The UES program has been designed to integrate, gradually over time, the approximately 200 separate business surveys into a single master survey program. The UES aims at collecting more industry and product detail at the provincial level than was previously possible while avoiding overlap between different survey questionnaires. The redesigned business survey questionnaires have a consistent look, structure and content. The unified approach makes reporting easier for firms operating in different industries because they can provide similar information for each branch operation. This way they avoid having to respond to questionnaires that differ for each industry in terms of format, wording and even concepts.

Reference period: Calendar year

Collection period: November of the reference year to October of the following year.

Subjects

  • Business performance and ownership
  • Financial statements and performance
  • Manufacturing

Data sources and methodology

Target population

The target population consists of incorporated and non-incorporated establishments primarily engaged in manufacturing. The population size is approximately 100,000 manufacturing establishments.

Instrument design

Two different types of questionnaires are used in the Annual Survey of Manufactures: (a) a long form and (b) a short form. The long form is a fully detailed questionnaire sent to establishments with shipments above a certain threshold that varies by province, by industry and by survey year. Among the smaller manufacturers whose shipments fall below the established threshold, some get the long form while others receive a short form. The short form is an abbreviated version of the long questionnaire, bearing a close resemblance to a typical income statement.

Twenty-two long form questionnaire "templates" have been developed, one for each manufacturing sub-sector based on the 3 digit level of the North American Industry Classification System (NAICS). A separate template has been developed for the Sawmills Industry and for the Wood sub-sector excluding the Sawmills Industry. Each template contains questions asking for standard financial data and an extensive list of commodities consumed and produced by establishments in the relevant sub-sector. These questionnaires collect data for about 12,000 commodities classified according to the Standard Classification of Goods (SCG). The questionnaires include standard financial questions and personalized commodity detail relevant to the individual establishment based, normally, on its response to the previous year's survey.

The questionnaires were developed in collaboration with data users in order to meet their statistical needs. Respondents and industry associations were also consulted through focus groups and individual meetings to ensure the information being asked was available and that the questionnaire could be filled out within a reasonable timeframe.

Sampling

This is a sample survey with a cross-sectional design.

The frame used for sampling purposes is the Statistics Canada Business Register with the observed population of all incorporated manufacturing establishments with payrolls (employer) above a certain threshold. The statistical unit is the establishment. A two phase probability design is used for sample selection. Establishments are stratified by province by industry and by revenue. "Take-alls" are selected based on being large and dominant within their industry. The "take-all" units receive a "long form" questionnaire. A "take-some" sample is also drawn. A sub-sample selected from within this group of units receives "long form" questionnaires while the rest of the "take-some" sample units receive "short form" questionnaires.

The proportion of long form questionnaires, short form questionnaires and the use of tax data varies from one year to the next, from one province to another and from industry to industry. These proportions are based on the resources available as well as the survey's target coverage at the national, provincial and industry levels.

Data sources

Data collection for this reference period: November 2001 to October 2002

Responding to this survey is mandatory.

Data are collected directly from survey respondents and extracted from administrative files.

Data are obtained from two sources: questionnaires that are mailed out and administrative files. Since the survey collects a wide range of information for over 250 manufacturing industries (based on the NAICS), the response burden is substantial. Using short form questionnaires and administrative files, where possible, reduces both the survey response burden and data collection costs, while maintaining the necessary level of accuracy.

Mail out occurs in November of the reference year (for establishments with fiscal year-ends of April to October), in January (for establishments with fiscal year-ends of November and December) and March (for establishments with fiscal year-ends of January to March) of the year following the reference year. The survey is usually directed to a financial officer within the business.

Respondents are asked to return the completed questionnaires within thirty days of receipt. Fax reminders are sent to outstanding ones 45 days after the mail out. Collection is completed no later than October of the year following the reference year.

The Annual Survey of Manufactures employs a "score function" strategy for collection, based on the value of manufacturing shipments. Sampled units are divided into 3 follow-up categories (priorities 1, 2 and 3). The priority 1 records are re-contacted for non-response and all attempts made to elicit a response as their contribution to the estimates is significant. In the case of priority 2 units, re-contact is based on scores with highest scored units being contacted first. The units included in this category are substitutable for one another. When a unit is a reluctant respondent, that unit is skipped and units next on the scored list are followed-up while ensuring that these units represent the contribution to coverage of the skipped unit. As a result, efforts in terms of time and money involved in re-contacting a reluctant respondent are reduced when a replacement unit(s) representing the same coverage can be re-contacted with relative ease. Follow-up continues until the pre-determined coverage level is reached. Full edits are applied to the units required to attain the coverage rate.

The remaining respondents and non-respondents (some priority 2 and all priority 3 units) are subject to minimal collection follow-up and edits. Failed edits are handled by an automatic post-collection edit and imputation system.

View the Questionnaire(s) and reporting guide(s).

Error detection

Most reporting and data entry errors are corrected as a result of the computer capture and edit procedures applied to the data. Historical edits (year to year comparisons) and consistency edits (totals equal sum of parts, proper units used, etc.) to micro-data are applied at the collection and capture stage. The scope exists for coherence analysis, i.e. the comparison of results from different surveys. Subject matter specialists are able to perform analysis at the macro level for both financial and commodity data and can "drill down" and manually correct micro level records if problems are discovered.

Imputation

The ASM data is run through two different Edit and Imputation (E&I) systems: one for the financial data and the other for commodities.

The E&I system for financial data uses generic rules, which have been tailored to the needs of the ASM. The system is set up to impute data by section of the questionnaire. It uses monthly-annualized manufacturing data, historical ASM data, donor data and any data reported by the respondent. There is also the imputation of variables, collected on the long forms to the short forms, i.e. short forms are modeled to long form equivalents. At the end of the financial E&I, all forms have long form data reported or imputed.

The E&I system for commodities was developed using the Standard Economic Processing System (StEPS) software. StEPS is a product developed by the US Census Bureau written entirely in SAS and operating in a UNIX environment. Modules within StEPS include many interactive SAS/AF screens. Key modules within StEPS include Edit, Imputation, Estimation and Review and Correction. The commodity imputation process uses a number of methods of imputation such as raking, historical imputation and donor. Both values and quantities (where required) are imputed.

Estimation

Estimation is done for incorporated establishments with payrolls (employer) in the manufacturing sector by producing a census-like data base and then aggregating it. Modeling is used to achieve the same for non-sampled units. Modeling first starts by fitting, on the sampled units, regression equations that relate the four "financial section" totals to their tax counterparts. Those equations are then used to impute the section totals for non-sampled units. Each imputed financial total is then split into its components by (i) observing the patterns of presence or absence of the possible component variables among sampled units, (ii) selecting randomly one of the observed patterns with probability equal to its observed frequency in the sample, (iii) allocating the section total into the chosen components. The latter step uses allocation ratios derived from tax information and respondent data. This completes the financial sections of non-sampled units. Then the total input and output amounts need to be broken down into commodities. This process is similar to the process for financial totals except tax information is not used.

For non-employer incorporated and non-incorporated establishments in the manufacturing sector, only total revenue and shipments are estimated, by industry and province, through the aggregation of tax data.

Quality evaluation

Survey results are analyzed for comparability with observed industry trends and general economic conditions using surveys such as the Monthly Survey of Manufacturing (Survey ID 2101) and the Survey of Employment, Payrolls and Hours (Survey ID 2612).

Disclosure control

Statistics Canada is prohibited by law from releasing any data which would divulge information obtained under the Statistics Act that relates to any identifiable person, business or organization without the prior knowledge or the consent in writing of that person, business or organization. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

Confidentiality analysis includes the detection of possible "direct disclosure", which occurs when the value in a tabulation cell is composed of a few respondents or when the cell is dominated by a few establishments.

Data accuracy

The amount imputed for non-response in RY2001 represents about 17% of the total value of manufacturing shipments in Canada. This percentage, however, will vary from province to province or from industry to industry. As with any other survey, errors influence program quality.

Documentation

Date modified: