Canada: Report on the Observance of Standards and Codes (ROSC)-Data Module Response by the Authorities

This Report on the Observance of Standards and Codes Data Module provides a review of Canada’s data dissemination practices against the IMF’s special data dissemination standard (SDDS), complemented by the in-depth assessment of the quality of six key sets of macroeconomic statistics: national accounts, consumer price index, producer price index, government finance statistics, balance of payments, and monetary statistics. Canada’s macroeconomic statistics are of high quality. They are comprehensive, timely, and accurate, and thus are adequate to conduct effective surveillance of economic and financial policies.

Abstract

This Report on the Observance of Standards and Codes Data Module provides a review of Canada’s data dissemination practices against the IMF’s special data dissemination standard (SDDS), complemented by the in-depth assessment of the quality of six key sets of macroeconomic statistics: national accounts, consumer price index, producer price index, government finance statistics, balance of payments, and monetary statistics. Canada’s macroeconomic statistics are of high quality. They are comprehensive, timely, and accurate, and thus are adequate to conduct effective surveillance of economic and financial policies.

A. Cross-cutting Recommendations

Across agencies

  • Articulate the roles of Bank of Canada and Statistics Canada in collecting, compiling, and disseminating statistics on the financial sector. Explore the possibilities of more data sharing between Statistics Canada and Bank of Canada on monetary and financial statistics.

The Bank of Canada and Statistics Canada agree it is important to review and clarify the roles and responsibilities for financial and monetary statistics. The two agencies have a long history of cooperation but as with all such collaborations, from time to time the roles and modus operandi should be reassessed.

Early in 2003, a Task Force on the Roles and Responsibilities of Statistics Canada and the Bank of Canada for the Collection, Compilation and Dissemination of Data was created to examine current practices and to develop guiding principles with respect to roles and responsibilities. The Task Force set up a working group that has already made an initial report to the steering committee.

The Task Force will produce a document recommending a clearer division of responsibilities. It has also been charged with exploring the possibility of more data sharing between the two agencies.

  • Keep the descriptions of Canadian dissemination practices on the DSBB up to date.

Canada recognizes the value of this source for metadata for international users and the importance of keeping these metadata as up to date as possible.

Fine tuning at the corporate level and across programs at Statistics Canada

  • Mention in a place widely accessible to Canadians, such as on the website, that a few government officials have access to certain economic statistics on the afternoon before their release for ministerial briefing.

Statistics Canada has accepted this recommendation and now has the following text on its website under the heading “About the Daily”:

Release times

“The Daily is released at 8:30 a.m. Eastern time each working day with the exception of data from the Labour Force Survey and the Consumer Price Index, which are available at 7 a.m. The Daily has been published since 1932 and posted on the Web since June 1995.

Given that Ministers’ comments on information released by Statistics Canada can have repercussions on financial markets, it has been deemed essential that a few key federal officials have advance access to a small number of economic statistics on the afternoon before their release.”

  • Consider a review of the terms of reference and operations of the 15 professional advisory groups, to help ensure the continued effectiveness of these important mechanisms.

Statistics Canada believes it is receiving valuable advice on both program priorities and professional and technical issues from its professional advisory committees. Periodically it reviews the operations of these committees to ensure that each is operating effectively and that they share good practices in the way they are run. The Agency will seriously consider another such review in conjunction with an overall assessment of any coverage and overlap issues among the committees.

  • Consolidate the many government-wide and agency-specific guidelines on ethical conduct into a one-stop shopping site.

While information on topics such as conflict of interest, preserving objectivity, confidentiality, privacy and security is already available on Statistics Canada’s Intranet site (employee access), Statistics Canada agrees that it would be beneficial to consolidate access to this information. Such a consolidation would make the information more visible and serve as an ongoing educational tool for employees.

  • Carry out revision studies more regularly and formally. For example, now that major methodological changes for national accounts and balance of payments are in place, resume analysis of measures such as bias and dispersion. For the consumer price index, as another example, monitor revisions to the seasonal factors and the relationship between seasonally adjusted and unadjusted series. Make such studies available to the public more fully as part of the information on accuracy.

Statistics Canada has re-established its previous revision study process after a hiatus related to implementation of SNA 93. Every year Statistics Canada revises the national accounts estimates for the previous four years and at the same time it publishes an article discussing the size and nature of the revisions. This serves largely as an information vehicle for explaining the nature and size of revisions to users. In addition, the SNA has currently embarked on a study of revisions focussing on the process, the integration since the implementation of SNA 93 and a large improvement to the survey taking infrastructure at Statistics Canada. This study will focus on the performance of the projection system relative to the newly developed annual surveys which feed the benchmark I-O tables.

Statistics Canada has a policy of not revising the CPI. Thus the official unadjusted CPI data are published without subsequent revision. The seasonally adjusted data in contrast are produced using the X-11 ARIMA program and are subject to revision as actual subsequent data observations change the seasonal factors initially estimated by the X-11 model based on projected data. Thus changes to seasonally adjusted factors are not “revisions” in the same sense as revisions to unadjusted time series data. Unadjusted time series data are usually revised due to more complete information about the reference period becoming available, such as late reports from respondents or other relevant data series. These types of revisions do not apply to the unadjusted CPI data.

As part of its internal production processes, Statistics Canada already regularly reviews revisions to the seasonal factors and the relationship between seasonally adjusted and unadjusted series. In fact, the Agency made adjustments effective January 2003 to the specifications in the X-11 ARIMA model as applied to CPI data based on this monitoring. The Agency will consider making publicly available occasional analyses of revisions to seasonal factors and the relationship between seasonally adjusted and unadjusted series. Such studies might provide a useful quality indicator to users of the seasonally adjusted data series and could help indicate whether further fine-tuning of the specifications in the X-11 ARIMA model is required. However no significant demand from either the academic or user communities has been identified to date for such studies and information. The Agency intends to consult users on this matter.

  • Extend the work on documentation of methodology and quality, especially on, referenced on, or linked to the IMDB. Consider adapting the template so that it more readily fits datasets such as national accounts, government finance, and balance of payments statistics.

The documentation of methodology and quality in the IMDB continues to be a priority for all statistical programs. Documentation for all business statistics programs (over 200 records) has been reviewed and upgraded as part of a special project. Discussions have started to establish similar special projects for national accounts, government finance and balance of payments statistics. Elements of the collection template that are not suitable for these programs are being identified and more appropriate guidelines will be developed in conjunction with these program areas. Processes to maintain the level of quality achieved through these special projects are in place, as are mechanisms to measure the quality of the documentation in the IMDB on an on-going basis. Linkages to this documentation from various locations on the Statistics Canada website have been improved and will be extended.

B. National Accounts

Priority

  • Estimate consumption of fixed capital at replacement cost instead of the historic cost estimates now used for the corporate sector.

Statistics Canada has a highly developed capital stock program which produces replacement cost estimates of Capital Consumption Allowance. Currently, however, for the “business” sector of the national accounts the CCA is recorded at “Book Value” coming directly from companies’ fiscal records. This choice was made because other elements of the income statement and balance sheet are also being recorded at “Book Value”. Now that the Balance Sheet program is being converted to Market Value, it is an opportune time to change the CCA estimates. We are currently studying the various options (linear versus geometric depreciation and estimated shapes of depreciation functions versus surveyed expected asset lives). In a year, we expect to publish these new estimates as alternative data and do a complete conversion the next time we revise back beyond our normal four year revision period – the timing of which is yet to be determined.

Fine Tuning

  • Complete the updated review of the “exhaustiveness” of the GDP estimate. Consider use of the framework for such studies recently prepared by an international group and published by the Organisation for Economic Co Operation and Development (OECD).

Statistics Canada is in the process finalizing a study on the non-observed economy. We have used the framework proposed by the OECD: Measuring the Non-Observed Economy: A Handbook. The report will update a 1994 Statistics Canada study: The Size of the Underground Economy in Canada, (Cat. 13-603E, No. 2).

  • Adopt COFOG for classifying government expenditure in order to enhance international comparability and to permit measurement of consumption by “who consumes” as well as “who pays.”

Currently, government expenditures by function are being developed as part of a larger project to build GFS data for the public sector. These will be the basis for National Accounts estimates.

  • Review the range of methodological publications on national accounts in order to ensure that all datasets are covered and that they reflect the changes consequent on closer alignment of the CSNA with the 1993 SNA.

We have released several national accounts methodological documents over the last few years and are in the process of preparing others. Among those released:

  • — Gross Domestic Product by Industry: Sources and Methods, released November 2002

  • — A Guide to Deflating the Input Output Accounts, Sources and Methods, released February 2001

  • — Canada’s Balance of Payments and International Investment Position, Concepts, Sources, Methods and Products, released July 2000

  • — Guide to the National Tourism Indicators, Sources and Methods, released February 2001

  • — Productivity Growth in Canada: Appendix 1 Statistics Canada’s Productivity Program, Concepts and Methods released January 2001,

  • — Concepts, sources and methods of the Canadian System of Environmental and Resource Accounts (Econnections: linking the environment and the economy), (Cat 16-505-GPE).

We plan to release, before the end of the year, a Guide to the Income and Expenditure Accounts as well as detailed methodologies underlying the estimates of Gross Domestic Product by industry. Plans are also under way to update the sources and methods used to construct Canada’s Input Output tables.

  • Provide a general note in statistical publications on the revision cycle of each national accounts dataset.

We have included in our published products a general note on the revision cycle of the national accounts dataset. The description of the revision cycle was highlighted in an electronic publication called “The Canadian Economic Accounts Quarterly Review” which encompasses all elements of the “National Accounts” data set for Statistics Canada.

C. Consumer Price Index

Fine tuning

  • Follow through on plans to develop tracking series for alternative treatments of owner-occupied housing; suggest extending it to producing a superlative tracking CPI with annual SHS data, which are available from 1997 onward.

The latest periodic analysis of the impact of alternative treatments of owner-occupied housing in the Canadian CPI was released in March 2003. Statistics Canada expects to produce the data based on alternative treatments of housing in the CPI on an annual basis by 2004.

Statistics Canada currently has no plans to produce a superlative index using annual household expenditure data. The Agency will consider producing an experimental superlative index as a possible research project subject to budgetary considerations.

  • Follow through on plans to implement checklists in the second phase of the CPI Sample Diversification Project.

Checklists are used selectively in the CPI. The current data processing infrastructure is restrictive in extending the use of checklists. The present phase of renewal of this processing infrastructure does not address requirements to permit extending the use of these checklists. Those requirements will be considered in a future phase of development, subject to budgetary considerations.

  • Review procedures to minimize use of the method of imputing the last observed price in price imputation methods for seasonally and temporarily unavailable items.

The CPI process renewal project is reviewing procedures such that carrying forward the last observed price in imputation will be minimized. This aspect of the project is expected to be fully implemented by 2005.

  • Provide more detailed documentation on how the areas and retail establishments are selected.

Statistics Canada has made extensive documentation on CPI concepts, definitions and methods available on its website and in detail in CPI reference manuals. More detailed documentation on how area samples and retail establishment samples are drawn will be included in future updates to documentation on the CPI.

  • Provide more detailed documentation describing how Statistics Canada has optimized the CPI sample regarding seasonal shopping patterns.

Statistics Canada will make available more detailed information on how the CPI deals with seasonal shopping patterns in its methodology in future updates to documentation on the CPI.

  • Compile and release indices on the Classification of Individual Consumption by Purpose.

Statistics Canada has examined the relationship of the COICOP classification to the Canadian CPI classification. The Agency intends to compile a supplementary series based on COICOP for analytical purposes, subject to budgetary considerations.

D. Producer Price Index

Fine Tuning

  • Follow through on plans to extend the PPI’s coverage of the service industries and continue to scan for ways to accelerate improvements in coverage of services.

Statistics Canada is developing more detailed plans on how it could extend the number of producer price indexes for service industries. These more detailed plans will be used in requests for funding an enlarged program of services price indexes.

  • Develop a summary index of the various output PPIs. Include crude materials coming from mining at basic price valuations in the IPPI as well as the RMPI.

A summary index of output PPIs can be considered but is not a current priority since there has not been demand for such an index to date. Users will be consulted and if demand warrants, the development of such a summary index will be considered, subject to budgetary considerations.

  • Review procedures to minimize use of the method of imputing the last observed price for seasonally and temporarily unavailable items.

In order to minimize response burden, most PPIs do not contact all respondents every month. The frequency of contact is based on prior information provided by respondents on the normal timing and frequency of their price changes. Thus for many price quotes, it is appropriate to impute for uncollected prices using the last observed price, as no change in price is a reasonable assumption in these circumstances.

Statistics Canada agrees with the goal of minimizing the use of last observed price imputation for cases such as seasonally or temporarily unavailable items and will further research appropriate methods, subject to budgetary considerations.

  • Continue efforts to update and keep current the content of PPI metadata and develop a portal for PPI series on the Statistics Canada website.

Statistics Canada recently expanded significantly the amount of PPI metadata available on the Statistics Canada website. Statistics Canada will update these metadata periodically and examine the implementation of a portal for PPI series on the Statistics Canada website.

E. Government Finance Statistics

Priority

  • Disseminate information on the sources and methods used in compiling quarterly public sector statistics for the quarterly CSNA.

Statistics Canada is currently putting together general documentation for dissemination concerning sources and methods used in compiling statistics for all sectors for the quarterly CSNA. Documentation for public sector statistics will be part of this.

  • Expand on the information about the main concepts used in compiling government finance statistics in the manual for the Financial Management System.

The Financial Management System (FMS) manual is updated annually. We will ensure that in future editions passages describing the main concepts are included.

Fine Tuning

  • Continue to implement the plan for migration to GFSM 2001, including the plan to implement quarterly Financial Management System estimates. Disseminate, in advance, information on the changes in concepts and methodology.

The fundamentals of the methodology for producing GFSM 2001 data have been established and much of the necessary historical data has been assembled. Since we plan to replace the FMS with the GFSM 2001 in our published output, we intend to launch a user-education campaign in advance to facilitate the transition.

  • Disseminate reconciliations of the annual statistics with the quarterly sum of the four quarters in the CSNA for the general government sector.

We will begin to produce and disseminate such reconciliations once the GFSM 2001 has been implemented.

F. Balance of Payments Statistics

Priority

  • Reclassify transactions that are not recorded in line with the BPM5. These reclassifications include (a) identifying financial derivatives in a separate functional category, (b) recording foreign money market instruments in portfolio investment, and (c) recording various services in their specific categories as part of the ongoing process of improving the quality of services data.

Canada recognises the usefulness and validity of these classification recommendations. However, to address a number of these issues will require redesigns of one or more survey questionnaires and associated systems. A general review of many of the questionnaires used in the BoP program is about to be initiated and these recommendations will be fully considered in this review. In the case of derivative financial instruments, survey redesigns will be coordinated across Statistics Canada and the Bank of Canada.

Work is well under way to address the reclassification of foreign money market instruments and it is expected this classification change will be made in 2004. Canada is reviewing the treatment of insurance services for both the SNA and the BoP but will await the decisions of the international working party before implementing any change.

  • Adopt international standards for classifying goods, breaking them into merchandise, goods for processing, repairs on goods, goods procured in ports by carriers, and nonmonetary gold.

While recognizing the utility of moving to these standards in terms of international comparability, there has not been a strong national demand for these classes. As the survey redesigns mentioned above go forward opportunities will be sought to move toward these standards.

  • Classify portfolio investment and other investment transactions in accordance with the institutional sector classification of BPM5. (This would be in line with the institutional sector classification prescribed by the SDDS for the external debt category, soon to be implemented by Canada.)

Gross Canadian external debt by institutional sector will be published for the first time in September 2003. Following this, additional work will be required to classify the full range of Balance of Payments and International Investment Position data by the institutional sector classification of BPM5.

Fine Tuning

  • Improve the coverage of financial derivatives and “e-commerce” transactions.

It is accepted that both phenomena have been increasing steadily over time and deserve additional attention. Improvements in the coverage of derivatives will require redesign of collection activities. As mentioned earlier, such work is about to be initiated. The latest figures show that e-commerce represented only about 0.3% of total personal expenditures in Canada (reference year 2001) and roughly 0.6% of total operating revenues of Canadian businesses (reference year 2002). Overall, Statistics Canada would like to improve coverage of these activities; however, significant resources will be required and these demands must be weighed against other priorities.

  • Pursue plans to publish IIP data at market value as recommended by BPM5 by mid-2004.

Canada has under way a project to move the IIP to a market value basis but due to resource constraints the exact timing is uncertain. Results will be released over the next year or two as they become available, and eventually the complete IIP will be made available at market value.

  • Continue efforts to identify mechanisms and implement specific measures to improve the coverage of goods exports to countries other than the United States not fully recorded in customs documents.

The Chief Statistician has met the Commissioner of the Canada Customs and Revenue Agency (CCRA) on this key issue. Plans are under way both to increase coverage and to improve its measurement as it decreases.

There has been progress in solving the problem of Canadian export undercoverage, through promotion of the Canadian Automated Export Declaration (CAED). The number of participants has grown from about 2,500 to over 10,000 in fiscal year 2002. In addition, CCRA is currently negotiating MOUs with carriers that would implement a “no load” policy to ensure the appropriate export documents are filed.

Canada and the US are considering modifications to the existing Customs document sharing arrangement that will allow access by Statistics Canada to U.S. in-transit documents for the purpose of matching with Canadian export filings (a non-match may represent undercoverage). Analytical work drawing on documents held at the Canadian ports, in cooperation with CCRA, will also allow improved estimates.

  • Continue work to improve the recording of portfolio investment transactions, based on the results from the Coordinated Portfolio Investment Survey.

Canada has played an active role in the preparation and the development of the IMF Coordinated Portfolio Investment Survey (CPIS) conducted for the first time in 1997. The survey has been conducted in Canada on an annual basis since that time and for Canadian assets is now an important data source for the ongoing program. The higher number of participating countries in the 2001 edition, the greater coverage achieved and the decision to conduct the survey on an annual basis at the international level makes this survey a precious vehicle to assess and improve the estimates currently gathered from Canada’s Balance of Payments (BOP) surveys.

The CPIS results for other participating countries will be employed systematically in the production of the liability estimates for Canada’s International Investment Position in the near future. The use of data sources such as the CPIS is crucial for Canada in order to preserve high-quality estimates and ensure international comparability.

G. Monetary and Financial Statistics

Priority

  • Conduct surveys of users of statistics on the financial sector to enhance understanding of how the data are used and how they might be further developed.

Statistics Canada and the Bank of Canada employ a variety of means to ensure the data they provide are relevant and responsive to changing needs. User surveys are one of the tools used by Statistics Canada to assess user needs and satisfaction; whereas the Bank of Canada relies on user feedback provided directly to the Bank from its many regular contacts with major users. Both agencies recognise the need to keep the data on the financial sector up to date and relevant for current issues and pride themselves on being open to user input and value such input in setting priorities for data development.

  • Develop a plan for the implementation of the MFSM, especially with respect to the following:

    • Compilation of consolidated balance sheets for the financial corporations sector;

The Statistics Canada - Bank of Canada Task Force on Roles and Responsibilities has been charged with exploring the rationale for developing consolidated balance sheets for the financial sector and its subcomponents and exploring the costs of such development. The Task Force will compile information on the views and practices of other national data producers and international organizations in the context of emerging data needs related to evolving financial markets.

Specific decisions on the Canadian response to this recommendation will be made following the completion of the Task Force’s report.

  • Classification of transactions, including financial derivatives, repurchase agreements, income trusts, and accrued interest;

The above-mentioned Task Force will also review the costs and benefits to Canada of implementing further improvements to our data through the use of these classifications. The Task Force will build on past research and current initiatives at the two agencies related to new financial instruments to define future initiatives responding to the international standards.

  • Valuation of financial assets and liabilities according to market value.

Statistics Canada has initiated a project to develop market value estimates for the National Balance Sheet Accounts and the International Investment Position. Plans are also being developed to collect additional market value information as existing surveys come up for redesign.

Canadian financial institutions currently value financial assets held for trading purposes according to market value. Assets held for asset and liability management or for other purposes are typically valued according to amortized value. In Canada, the Accounting Standards Board (AcSB) is initiating a limited move towards fair value accounting with respect to financial instruments. These proposals are expected to be adopted for fiscal years commencing on or after October 1, 2004. This means that all financial assets other than those that are to be held to maturity, and loans and receivables that are not intended for trading or securitization, will be valued according to fair value.

  • Reconcile monetary statistics compiled by the Bank of Canada with statistics on the financial sector compiled by Statistics Canada. Undertake consistency checks with other macroeconomic statistics (national accounts, government finance, and balance of payments) to eliminate or reduce potential inconsistencies in coverage and classification.

Reasons why data collected by Statistics Canada do not reconcile with data collected by the Bank of Canada are understood by both agencies and principal users of the data. The key reasons are:

  • 1) Statistics Canada requires and collects data on a locational (booked in Canada only) basis and the Bank of Canada requires and collects data on a fully consolidated booked worldwide basis.

  • 2) The Bank of Canada emphasizes daily average data rather than month-end data.

  • 3) Industry totals cannot generally be compared because some financial institutions are classified differently under the North American Industrial Classification System (NAICS) used by Statistics Canada and the Banking regulations under which the Bank of Canada collects data. The differences in classification combined with the confidentiality provisions of the Statistics Act prevent detailed reconciliations between datasets.

  • 4) Some of the concepts and definitions vary between the two agencies, for example those relating to consumer credit.

Nonetheless, analysts in both agencies regularly do use data from a wide range of sources and do communicate any inconsistencies in the results via various formal and informal communication mechanisms. The Task Force will examine ways in which this information could be made available to a wider community of users.