Productivity Measures and Related Variables - National (Annual)

Detailed information for 2002

Status:

Active

Frequency:

Annual

Record number:

1402

The annual program of the Canadian Productivity Accounts (CPA) produce indexes of labour productivity, multifactor productivity, and related measures (employment, hours worked, cost of inputs, measures of inputs) for the business sector, broad economic sub-sectors and their constituent industries at the national level. Each set of measures involves a comparison of the growth in output and input measures, but each relies on a different methodology.

Data release - July 10, 2003

Description

The annual program of the Canadian Productivity Accounts (CPA) produce indexes of labour productivity, multifactor productivity, and related measures (employment, hours worked, cost of inputs, measures of inputs) for the business sector, broad economic sub-sectors and their constituent industries at the national level. Each set of measures involves a comparison of the growth in output and input measures, but each relies on a different methodology.

The first set of measures covers labour productivity-real output per hours worked for the business sector and its constituent industries.

The second set of measures covers multifactor productivity. In these measures, a measure of real output is related to the corresponding measure of a bundle of inputs, such as capital, labour and intermediate inputs. Multifactor productivity estimates have been developed in recognition of the role capital input and intermediate inputs growth play in output growth. Note that these estimates published by the national program are experimental.

Productivity measures the efficiency with which resources are employed in economic activity. Annual productivities series are widely watched by analysts, government policymakers and researchers to quantify the extent to which productivity contributes to economic growth and the standards of living over the long-run.

Statistical activity

The Canadian Productivity Accounts (CPA) are responsible for producing, analyzing and disseminating Statistics Canada's official data on productivity and for the production and integration of data on employment, hours worked and capital services consistent with the System of National Economic Accounts. To this end, the CPA comprises three programs. The quarterly program provides current estimates on labour productivity and labour costs at the aggregate level for 15 industry groups. The annual multifactor productivity program provides yearly estimates on multifactor productivity and related measures (output, capital input, labour input and intermediate inputs) as they apply to the major sectors of the economy and to industries at the national and provincial levels (see Productivity Measures and Related Variables - National and Provincial [Annual], record number 1402). Lastly, the annual provincial program, as an integral part of the provincial and territorial economic accounts, provides estimates of employment, hours worked, labour productivity and labour costs at the industry level for each province and territory (see Labour Productivity Measures - Provinces and Territories [Annual], record number 5103).

Subjects

  • Economic accounts
  • Productivity accounts

Data sources and methodology

Target population

Most estimates are only available for the business sector and its constituent subsectors and industries. However, estimates for employment and hours are also available for the non-business sector.

Instrument design

This methodology does not apply.

Sampling

This methodology does not apply.

Data sources

Data are collected from other Statistics Canada surveys and/or other sources.

The measure of output that is used for the aggregate business-sector productivity estimates is based on the final demand gross domestic product available from the Income and Expenditures Division.

At the industry level, a variety of output measures--gross output, value added and sectoral output, each of which respond a different analytical need, are employed by the annual program of the CPA. Gross output and value added are produced by the Industry Accounts Division, while sectoral outputs are produced by the CPA for comparability with the U.S.

The CPA is responsible for constructing labour (employment and hours at work) data that accord with the CSNA production framework, using various data sources available at Statistics Canada. Other sources on labour data, available within Statistics Canada, do not completely satisfy this requirement. For more details see Economic analysis methodology paper series: National Accounts; catalogue no. 11F0026MIE, free.

The CPA also construct hours at work adjusted for changes in the composition of the workforce in terms of education and experience, using the Census of Population, the CPA data, and the Labour Force Survey.

Much like labour input, the CPA produces consistent estimates of capital services. Other sources are available within Statistics Canada for estimates of capital that do not completely satisfy the consistency needs of the CPA--partly because they provide only estimates of capital stock (not capital services) and partly because they are not fully integrated into the production framework.

Error detection

This methodology type does not apply to this statistical program.

Imputation

Unlike labour income data on paid jobs, data on the income of self-employed workers are not available from the Income and Expenditure Accounts Division. The income of self-employed workers is therefore established by imputation. The imputation is mainly based on the assumption that the value of an hour worked by a self-employed worker is equal to the value of an hour worked by a paid worker (paid at the average rate) in the same industry.

Estimation

Statistics Canada's productivity estimates are based on a bottom-up approach. Productivity indices are estimated with the most diaggregated data available. Prouctivity indices for 147 industries in the case of labour productivity and 123 industries in the case of multifactor productivity are then aggregated step by step to the total business sector.

Additional industrial detail (203 industries from 1961 to 1980 and 243 industries from 1981 on) is produced and disseminated for the number of jobs and hours worked series for both the business sector and the non-commercial sector. In order to produce productivity growth estimates, various data sources from survey areas and the system of national accounts divisions are integrated.

Data that come from these different sources are conceptually adjusted to the CSNA framework and reconciled for accuracy and consistency in the estimates of inputs and outputs.

Quality evaluation

Three criteria are used to evaluate the quality of the estimates:

A) the quality of the data sources including deflators;
B) the nature of breaks (if any) in the series; and
C) the volatility of the series.

The combination of these criteria should be considered as a necessary condition for quality; however, no criterion on its own can be a sufficient condition for quality. Based on these criteria, variables are assigned a quality rating using a three-point scale: 1 - reliable; 2 - fairly reliable; and 3 - unreliable.

A) Quality of data sources and deflators

A variable cannot receive rating 1 (reliable) if:

1) the series in current prices are unreliable (no surveys or administrative sources) because they are based on imputed data; and/or

2) the variable has no appropriate deflator for constructing a reliable series in constant prices.

If the series look "reasonable" despite weaknesses 1 and 2 described above, it would be given a 2 (fairly reliable) rating. Otherwise, it would be considered unreliable (3) by default.

B) Breaks in the series

A few examples are: macroeconomic or industry-specific shocks whose impact is temporary or permanent, changes associated with the environment within which the industry operates (regulation, etc.), and changes associated with the statistical infrastructure (classification, sources, concepts and methods).

If breaks that affect the series are explicable - in other words, if they are covered by one of the above examples - then the series could receive rating 1 (provided the other quality criteria are satisfied).

If the breaks are not explicable, the series cannot obtain rating 1.

C) Volatility in the series

Does the change in the volatility of series over time suggest a change in cyclical fluctuations or a change in data estimation methods? Economists seem to have reached a consensus: the cyclical fluctuations have been greater since 1973. We therefore expect that the series will show more volatility during the period from 1973 to the present compared with the pre-1973 period but not the reverse.

We should also expect that, in general, estimates of capital would be less volatile than labour estimates. Capital is conceptually a stock, therefore, it changes sluggishly. However, there are times where investment shows discrete changes and, as a result, we might expect to see substantial changes in capital stock.

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.

The Canadian Productivity Accounts use the confidentiality model developed by the Industry Accounts Division. This model uses a method which suppresses sensitive cells.

Data are also suppressed if they have been assigned an unreliable quality rating.

Revisions and seasonal adjustment

The release of multifactor productivity, labour productivity and related measures estimates occurs twice a year. In the second quarter of each year, estimates of labour productivity and multifactor productivity for the aggregate business sector are released for the previous year. Data for these estimates are revised for the last four years, following the CSNA procedures. Data of the most two recent years are always considered preliminary.

Industry labour productivity and multifactor productivity measures are released in the first quarter for the previous year. Data of the most three recent years are always considered preliminary.

Data accuracy

A) Judgement - expert approach

All variables in the KLEMS have been assigned a quality rating using a three-point scale: 1 - reliable, 2 - fairly reliable, and 3 - unreliable.

At the P level of aggregation (i.e. 122 industries) the results weighted by the output share of each industry in the business sector are the following for the multifactor productivity estimates:

48% of the estimates are reliable,
24% of the estimates are fairly reliable, and
28% of the estimates are unreliable.

Interestingly, the industries that received rating 1 for their multifactor productivity estimates are those that post the highest multifactor productivity (weighted) average annual growth rate for the period 1961-1996 (1.05%). This growth rate falls to 0.38% for industries with a multifactor productivity rating of 2 and deteriorates (-0.21%) for those with a rating of 3.

B) Parametric approach

The results generally confirm those obtained by the judgement - expert approach.

For additional information and detailed accuracy measures, see:
"The Precision of Productivity Estimates" in "Productivity Growth in Canada, 2001", available through the on-line catalogue no. 15-204-XIE, Chapter 3, and "Assessing the Data Quality of Statistics Canada's Productivity Program", available in the Documentation section below, Productivity Growth in Canada, appendix 3.

Overall, the results of the quality evaluation can be summarized as follows:

First, the quality ratings that have been assigned indicate that the multifactor productivity data are of acceptable quality for a majority of industries.

Second, the error measurement results echoed those of the quality rating; multifactor productivity estimates had error estimates that were of intermediate relative quality.

Third, the productivity program's preliminary estimates provide the correct direction of change at least 8 times out of 9. The preliminary estimates of the productivity program are reliable and any subsequent revision provides a marginal gain in terms of accuracy.

Documentation

Date modified: