Revisions analysis – Canadian System of National Accounts 2012
Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Revisions to gross domestic product
Revisions to the growth in GDP
Revisions to income-based GDP components
Compensation of employees
Gross operating surplus
Gross mixed income
Taxes less subsidies on production and on products
Revisions to expenditure-based GDP components
Non-profit institutions serving households final consumption expenditure
Government final consumption expenditure
Gross capital formation
International imports and exports of goods and services
Revisions to the current and capital accounts
Household sector aggregates
Government sector aggregates
Corporate sector aggregates
This article is the sixth of a series of articles which have been prepared to help users understand the changes introduced as a result of the historical revision of the Canadian System of National Accounts (CSNA)—due to the implementation of the new international standards published in System of National Accounts 2008 (SNA2008). It provides users with a reconciliation between the previously published figures (CSNA97) and the revised figures (CSNA12). The changes are presented for the majority of major aggregates contained within the CSNA, such as gross domestic product (GDP), investment and household final consumption expenditure. For each of the aggregates, the previous estimate, the new estimate, the difference and a brief explanation of the change are presented. The tables referred to within this article are the national Income and Expenditure Accounts (NIEA) tables as they appear on CANSIM.1
The changes incorporated with CSNA12 have yet to be implemented for years prior to 1981. That work is on-going and Statistics Canada intends to release, at a later date, a consistent national accounts time series for the period 1947 to the present on a quarterly basis, and 1926 to the present on an annual basis.
Throughout this paper, the term CSNA12 is used. This refers to the vintage of the Canadian National Income and Expenditure Accounts that were released on October 1st, 2012. This includes data from 1981 to the second quarter of 2012. CSNA12 is based on the System of National Accounts 2008 international standard. Also used is the term CSNA97, which refers to the vintage of NIEA that were last released on August 31st , 2012, and which included data from 1961 to the second quarter of 2012. CSNA97 is based on the System of National Accounts 1993 international standard.
Revisions to gross domestic product
The implementation of CSNA12 did not result in a substantial change to the level, nominal growth rate or real growth rate of GDP. In 20072, the nominal level of GDP was revised upward by $36.4 billion. The main contributors to the change in the level of GDP were the broadening of the capital boundary (capitalization of government and business research and development activities and capitalization of military weapons) and the adoption of replacement cost measures of consumption of fixed capital, using geometric patterns of depreciation for all sectors of the economy.
Prior to CSNA12, business expenditures on research and development were treated as intermediate expenditures. SNA2008 recognizes that research and development activities represent an investment made by businesses and government that are used in the production of goods and services for more than one year (SNA2008p. 10:32). CSNA12 has reclassified research and development activities from intermediate expenditures to gross fixed capital formation. This revision increased the level of business investment in the calculation of expenditure-based GDP and gross operating surplus in the calculation of income-based GDP.
In CSNA97 government expenditures on research and development were treated as final government consumption expenditure. Similar to business investment in research and development, these expenditures are now treated as capital expenditures. The move from final government consumption expenditure to government gross fixed capital formation increases the level of GDP. Depreciation or consumption of fixed capital, as it is referred to in the CSNA, is a component of final government consumption expenditure on the expenditure side of GDP and gross operating surplus on the income side of GDP. With the capitalization of government research and development activities the value of government consumption of fixed capital has increased. Table 1 shows the impact of the capitalization of research and development on the level of GDP for the period 2007 to 2011.
Two other important changes had an impact on the level of GDP. Military weapons systems were re-classified from final government consumption expenditure to government fixed capital formation. The re-categorization of these transactions increased the level of GDP. The depreciation of these assets increased both final government consumption expenditure and gross operating surplus.
In addition, in CSNA97 estimates of government consumption of fixed capital were derived using a linear method of depreciation. In CSNA12, a geometric method of depreciation is used, which better reflects the decline of fixed asset prices through time from usage, and aligns with international practices. In general, this resulted in an upward revision over the entire time period to both final government consumption expenditure and gross operating surplus. The impact of both of these revisions is highlighted in Table 2 for the period 2007 to 2011.
In addition to the above conceptual and methodological revisions to the NIEA there were a number of statistical revisions. The details of these revisions are elaborated later in this paper. At an aggregate level, these revisions net to $11.1 billion in 2007. Table 3 provides a reconciliation between the previous level of GDP and the revised level of GDP, illustrated with the net conceptual revisions noted above and the net statistical revisions.
Revisions to the growth in GDP
Revisions to the growth in GDP over the revision period were not substantial. While the conceptual and statistical revisions resulted in a level shift upward in GDP they had little impact on the quarterly and annual growth rates. Over the 1981 to 2012 period, the mean absolute percentage point revision to the annual growth rate in GDP was .15 percentage points. The largest upward revision occurred in 2001 while the largest downward revision occurred in 1999.
Revisions to the real rate of growth over this period did not necessarily coincide with that observed for the nominal rates of growth. This is mainly due to the implementation of a number of new final demand classifications (household expenditures, imports of goods, exports of goods and investment) that affect the weighting of the elemental indexes used in the calculation of the volume measure of GDP and the GDP deflator. The mean absolute percentage point revision to the annual growth rate in real GDP for the 1981 to 2012 revision period was .14.
Of particular interest are the revisions during business cycles characterized by a recession and a recovery period. Charts 1 to 3 compare the growth in GDP for the periods 1981 to 1983, 1989 to 1991 and 2008 to 2009.
Revisions to income-based GDP components
Aside from revisions to the various components of income-based GDP, there has been a major change in its presentation. Prior to the release of CSNA12, income-based GDP was broken out into the following components:
- labour income3;
- corporation profits;
- interest and miscellaneous investment income;
- inventory valuation adjustments;
- net income of unincorporated businesses (including rental income);
- taxes on production less subsidies;
- taxes on products less subsides; and,
- capital consumption allowance4.
The old presentation of income-based GDP dated back to the 1950s and was a mix of national accounting concepts (labour income and taxes and subsidies), business accounting concepts (corporation profits and net income of unincorporated income), with adjustments to bring it all in line with national accounting practices (inventory valuation adjustment and interest and miscellaneous investment income). In order to better align the Canadian presentation with the international presentation and more clearly articulate the generation of income from the production of goods and services, CSNA12 has adopted the international conventions.
The income-based GDP measure now includes the following categories:
- compensation of employees, which represents the return to the labour factor of production used in the production of goods and services;
- gross operating surplus, which represents the return to the capital factor of production used in the production of goods and services;
- gross mixed income, which represents a combination of the return to the labour and capital factors of production used in the production of goods and services when they cannot be separately identified as is the case for the activities of unincorporated businesses;
- taxes less subsidies on production and taxes less subsidies on products.
Compensation of employees
Under the previous version of the Canadian System of National Accounts the term 'compensation of employees' was referred to as 'labour income'. There is no conceptual difference between the two measures—the change was made in CSNA12 to align with the international terminology—allowing users to more readily compare Canadian estimates with those found in other countries' national accounts. While conceptual changes were not made to the estimate of compensation of employees, a number of important statistical revisions were introduced with CSNA12. These revisions include:
- elimination of an estimate that was made in CSNA97 for wages and salaries (income-in-kind) paid to individuals employed (clergy) by religious organizations, eliminated from 1997 onwards;
- a reduced estimate for wages and salaries for persons employed by households (such as nannies, butlers, gardeners and babysitters), for the period 1981 to 2012; and,
- an increased estimate for wages and salaries paid to individuals working for aboriginal general governments, for the period 1981 to 2012.
In CSNA97, an estimate representing income-in-kind was made to source data used to compile wages and salaries of individuals (mainly clergy) employed by religious organizations, such as churches. The estimate was to account for income-in-kind deemed to be excluded from the source data. Conceptually, the estimate represented rent-free or reduced-rent housing and/or use of a vehicle provided by the employer. For those employees satisfying conditions set out by the Canada Revenue Agency (CRA) these items were identified as non-taxable, but for CSNA purposes included as part of the measurement of wages and salaries. Upon further review of the source data and CRA reporting requirements, this adjustment is no-longer required as it is being reported as part of employment income on the source data (T4 Statement of Remuneration Paid).
In CSNA97, an adjustment was made to the estimate of wages and salaries of employees for the private household industry. This industry "comprises private households engaged in employing workers, on or about the premises, in activities primarily concerned with the operations of the household." (NAICS 2007). The estimate of wages and salaries of employees for this industry comes from the T4 administrative data system. In 1997, these data were confronted with household spending estimates originating from household surveys and it was discerned that the T4 estimate was much lower than expected for the level of expenditure reported on the surveys. An adjustment was then made to the T4 data to bring it in line with the household survey estimate. After careful examination during the CSNA12 revision, it was determined that much of the spending being reported on the household survey was related to payments made by households to businesses (such as lawn care companies and snow removal companies) rather than to employees the household was directly employing. As such, the adjustment to compensation of employees for the private household industry has been substantially reduced for the entire period 1981 to 2012.
A significant data quality issue associated with the CSNA97 data was the estimate of compensation of employees in the Aboriginal general governments sector. In an effort to measure the under-coverage of wages and salaries of tax-exempt, or partially tax exempt employees, employed by Aboriginal general governments CSNA97 included an upward adjustment to estimates of compensation of employees derived from T4 administrative files. As a result of work completed on implementing an Aboriginal general government sector, including additional information gleaned from the T4 Statement of Remuneration Paid, an improved estimate of under-coverage has been established for CSNA12.
The trend and growth rate of compensation of employees has not changed significantly with the implementation of CSNA12. Table 5 shows the mean absolute percentage point revision in the quarterly and annual growth rates over the 1981 to 2012 period. The largest annual upward revision occurred in 1998, while the largest annual downward revision occurred in 2010.
A more notable change is the share of compensation of employees as a percentage of total GDP. With CSNA12, the share of compensation of employees to GDP has shifted downward. This is a result of the statistical revisions noted earlier which tended to reduce compensation of employees. As well, the increase to the overall level of GDP resulted in a level shift in gross operating surplus, which has taken a proportionally greater share of total market-price GDP.
Gross operating surplus represents that part of income-based GDP which arises from the use of capital in production (SNA2008 p.7:12). The presentation of gross operating surplus has changed significantly with the implementation of CSNA12. In the CSNA97 presentation, it was not possible to derive an estimate of gross operating surplus since there was no breakdown between corporate and unincorporated consumption of fixed capital, inventory valuation adjustment or interest and miscellaneous investment income. In the CSNA12 presentation, gross operating surplus is estimated for both the financial corporations' sector and the non-financial corporations' sector. This change in presentation clarifies the distinction between the contribution of labour and of capital in the economic production process as well as between corporations and households.
As noted earlier, the capitalization of research and development resulted in an upward level shift in gross operating surplus between the CSNA97 data and the CSNA12 data. This is because, in CSNA97, business expenditures on research and development was treated as intermediate expense. In the CSNA12, purchases of research and development services are treated as capital expenditures and an imputation is made for own-account output of in-house research and development activities. Gross operating surplus represents the output of a firm, less compensation of employees and intermediate expenses. Reducing intermediate expenses, while increasing output, resulted in an upward shift in the gross operating surplus.
As mentioned at the beginning of this article, the estimation of consumption of fixed capital was revised, for all sectors of the economy, from a variety of valuations (book value and replacement cost) and methods (linear versus geometric), to one consistent methodology of replacement cost valuation using a geometric method of depreciation. This is an important improvement since it ensures that all income, saving and wealth measures are estimated using the same vintage of prices (referred to within the SNA as replacement cost). The implementation of the geometric pattern of decline in the prices of various vintages of assets best reflects the empirical evidence of used asset prices and is more in line with the national accounting practices of other countries.
The consumption of fixed capital related to government investment also forms part of gross operating surplus. The capitalization of government research and development, military weapons systems, and the move to a geometric rate of depreciation, resulted in an upward shift in the estimate of the consumption of fixed capital, further increasing the level of gross operating surplus6.
The change in method for the estimation of the consumption of fixed capital was also applied to the corporate sector. In CSNA97, consumption of fixed capital was calculated using a linear method of depreciation calculated from a capital stock at book value. With CSNA12 it is calculated using a geometric method of depreciation calculated from a capital stock at replacement cost. While this change does not result in a change in gross operating surplus, it does impact measures such as net operating surplus and net domestic product. Table 6 shows the impact of the move to corporate consumption of fixed capital at replacement cost for the years 2007 to 2011.
Aside from these conceptual changes, a number of statistical changes were made to gross operating surplus. These were primarily the result of incorporating statistical revisions made elsewhere in the calculation of GDP. This includes adjusting gross operating surplus for revisions such as new estimates of imports and exports of commercial services and investment in cars and trucks.
Gross mixed income is another measure that did not appear in CSNA97. Gross mixed income implicitly contains "an element of remuneration for work done by the owner, or other members of the household …. as well as the return to the owner as entrepreneur." (SNA2008p. 7:9). In CSNA97 this income was captured in the following income categories:
- Income of farm operators from farm production;
- Income of non-farm unincorporated businesses, including paid and imputed rent;
- Unincorporated portion of interest and miscellaneous investment income;
- Unincorporated portion of inventory valuation adjustment; and
- Unincorporated portion of capital consumption allowance;
CSNA12 follows the SNA2008 recommended presentation, grouping all of the income accruing to unincorporated businesses from their participation in production into a single category. In addition to the gross measure, CSNA12 provides a measure of net mixed income. Similar to the corporate sector, the consumption of fixed capital of unincorporated businesses is now measured at replacement cost using a geometric method of depreciation.
In addition to the presentational changes, a number of important statistical revisions were incorporated into the measure of gross mixed income. The CSNA97 measure has been revised to include: direct use of administrative data files, business information obtained from Statistics Canada's Business Register, and methodological improvements.
Tax data obtained from the Canadian Revenue Agency (CRA) and used in the derivation of estimates of mixed income include: the T1 Personal Master File (PMF), T5013 Partner Information Return and Statement of Partnership Income files, Assessed Record File (ARF), T1 E-File and T1 Barcode File. For the T1 PMF, Statistics Canada obtained historical files from CRA, dating back to reference year 19877. On average, the historical files add an additional 8% of mixed income to the original T1 PMF that was used to estimate mixed income of non-farm unincorporated business in CSNA97. The 8% difference represents late filers, amended records or dropped records.
Besides the data source changes noted above, an improved methodology was developed to estimate the undercoverage of activity in the construction, retail trade (tobacco smuggling) and health care and social assistance (child daycare) and other services (parking) industries. Triangulation of data from different sources such as tax files, business and household surveys was used in developing the new method.
Given that neither gross operating surplus nor gross mixed income were identifiable in CSNA97, it is not possible to provide a detailed reconciliation between the old measures and the new measures. This reconciliation can only be performed when looking at gross operating surplus and gross mixed income together. Gross operating surplus plus gross mixed income as presented in CSNA12 is equivalent to the sum of the following income components in CSNA97:
- corporation profits
- income of farm operators from farm production;
- income of non-farm unincorporated businesses;
- interest and miscellaneous investment income;
- inventory valuation adjustment; and
- capital consumption allowance;
Table 7 provides a reconciliation between the old measure and the new measure for the years 2007 to 2011.
With CSNA12, there were a number of revisions that were made to the estimate of taxes less subsidies on production and products. Most of these were the result of re-classifying certain fees from sales of goods and services to taxes or from taxes to the sale of goods and services. The specific changes include:
- Probate fees: Probate fees are the responsibility of provincial and territorial governments and are charged differently depending on the jurisdiction. In Newfoundland and Labrador, Nova Scotia, Prince Edward Island, New Brunswick, Saskatchewan, Manitoba and British Columbia, probate fees are calculated as a percentage of the value of a person's estate, with no limit to the fee charged. In CSNA97, these were classified as sales of goods and services while in CSNA12, these are classified as tax revenue, since the amount charged tends to exceed the cost of providing the service. This is consistent with the treatment of probate fees as recorded for Ontario in both CSNA97 and CSNA12. In the case of Alberta, Quebec, Northwest Territories, Yukon and Nunavut, probate fees will continue to be recorded as sales of goods and services since the fees are capped and do not exceed the cost of providing the service. These revisions were applied from 2008 onwards.
- Deposit insurance premium revenue in Ontario and Alberta: The deposit insurance premium revenue of the Deposit Insurance Corporation of Ontario and the Credit Union Deposit Guarantee Corporation of Alberta was reclassified from insurance premium taxes (indirect taxes) to other miscellaneous indirect taxes, since the taxes charged do not relate to the insurance service. This change was made from 2008 onwards.
- Reallocation of Quebec Provincial government revenues: Revenues for Quebec allocated to provincial income tax, payroll tax, health insurance premiums, interest and Quebec Pension Plans (QPP) contributions were revised for 1999 to 2003. This corrected a mis-allocation of these components in this period. The total revenue is unchanged.
- Revised mining tax, royalties and capital tax in Newfoundland and Labrador and Saskatchewan: Two statistical changes were made to taxes in Newfoundland and Labrador and Saskatchewan. In Newfoundland and Labrador, mining tax was classified as resource royalties and resource royalties as mining tax for the period 1999 to 2003. This has been corrected. In Saskatchewan, for the period 1988 to 2005, a resource surcharge was reclassified from corporate capital tax to taxes on production—natural resource revenues.
- British Columbia local government tax revenue – The South Coast British Columbia Transportation Authority (SCBCTA) is treated as a government business enterprise (GBE) within the CSNA. As such, the tax revenue collected by the SCBCTA is treated as a subsidy revenue of the GBE, rather than tax. The tax revenue collected by the SCBCTA was attributed to the BC provincial and local governments and then flowed as a subsidy to the SCBCTA. For the period 1999 to the present, tax revenue associated with property taxes, hydro levies and parking lot or site improvements were not recorded in the CSNA. This has been revised with CSNA12 and consequently both provincial and local government tax revenue and subsidies have been revised upward.
- Subsidies on products: In the process of developing the non-profit institutions serving households (NPISH) sector account it was discovered that a portion of the government transfers to non-profit institutions serving households appearing in CSNA97 were in fact directed to non-profit institutions serving businesses. In the case of funds flowing to non-profit corporations serving business, these transactions are recorded as subsidies on products within the CSNA. The result is that subsidies on products for the entire time period 1981 to 2012 shifted upwards. This revision was offset by an increase in corporate surplus.
In CSNA97 the expenditures of households were grouped together with non-profit institutions serving households and Aboriginal general governments. With CSNA12 this is no longer the case. Transactions of non-profit institutions serving households and Aboriginal general governments have been recorded in their own sector, in order to more clearly identify the final expenditure of households. The result is a large downward revision to the level of household final consumption expenditures. Table 8 shows the value of non-profit institutions serving households and Aboriginal general government expenditures that were removed from the previous measure of household final consumption expenditure.
In addition to the removal of the expenditures of these institutional units from household final consumption expenditure, a number of methodological improvements were made to selected household expenditures. The main series impacted include:
Tobacco: Household final consumption expenditure on tobacco products have been revised for the period 1981 to the present. The revisions are explained by the following methodological improvements:
- there was more illegal contraband tobacco than previously estimated as a share of total tobacco consumption;
- the provincial and territorial distribution was revised to reflect more current information;
- the phenomenon of discount cigarettes offered by manufacturers, starting mostly in 2003, is reflected in the revision to prices.
- Used motor vehicles: Estimates of used motor vehicles were revised for the period 1981 to the present. The methodology that was previously used to estimate the net household spending on used motor vehicles had been relatively good at predicting the importance of used motor vehicle sales, economy-wide. Nonetheless, it was inadequate at taking into account some historical changes in the transfer of ownership from one sector to another. In addition, the estimation of gross margins associated with the activity of selling used cars and trucks by vehicle dealers has been improved. The method has now been adapted to account for these and other types of purchases of used motor vehicles by institutional units in one sector from institutional units in another sector. When households lease a new vehicle, the entire value of the vehicle is recorded in household expenditures. When the vehicle is returned to the business at the end of the lease it was assumed that the vehicle would be sold back to the household sector at a later date. While this is in fact the case for many leased vehicles, a careful examination of the data showed that some of these vehicles were exported. As such, expenditures on used vehicles were revised downward by the value of leased vehicles returned to businesses that were subsequently exported.
- New passenger vehicles: This statistical revision employs the provincial registration data to better allocate the purchase of new passenger vehicles by sector. In July of 2006, Statistics Canada acquired access to a database on registration of new motor vehicles by province and territory. The information reported in this database originates in motor vehicle registration offices across Canada. Updated on a monthly basis, it presents detailed information on every new car and truck registered in Canada, including the model of the vehicle; its segmentation (i.e., car or truck); its model-year as well as specific information on its ownership (whether it was acquired by a household, a business or a government). The database also distinguishes between vehicles that have been purchased or leased. The CSNA had previously used statistics on unit sales from the Canadian Vehicle Manufacturer's Association (CVMA) to derive a household and business split for new vehicle purchases. Those statistics were only partial, however, as the breakdown between the household and the business or government acquisitions were available only for a few manufacturers. In addition, the business purchases available from that source were based on fleet sales only and did not represent all motor vehicle purchased by the business or government sector. The new data source provides the CSNA with this breakdown for all manufacturers of new motor vehicles. This revision was somewhat offset by a revision to business investment in new motor vehicles and trucks.
- Charitable gaming: A review of household spending on charitable gaming events within each province and territory was completed in June 2011. Charitable gaming events include bingos, raffles, break-open tickets, charity casinos and other games of chance organized by groups licensed to hold such events by provincial or territorial governments. All household expenditure on charitable gaming events will now be included in household final consumption expenditure. As with other games of chance, estimates of household expenditure on charitable gaming are calculated as wagers minus prizes, on a calendar year basis. Estimates for the charity portion of gambling will be combined with current estimates for government controlled gambling.
- Expenditure by Canadians abroad and expenditures by non-residents in Canada (net travel expenditures): Another important change in the measurement of household final consumption expenditures was the treatment of the expenditures of foreign students, as well as expenditures on international medical services. With CSNA97 these purchases were assigned to the individual expenditure categories. For example, if a resident was studying in New York, the resident's expenditures would have been recorded under the individual expenditure categories such as tuition fees, food expenditures, rent, transportation, etc. Likewise, if a resident purchased medical services abroad, these expenditures would have been recorded under medical services, food, lodging, transportation, etc. This treatment differed from the Balance of International Payments treatment, where all of these expenditures are treated as imports and exports of travel services. In order to align with the Balance of International Payment treatment within CSNA12, all expenditures of foreign students and international medical services are now recorded as either expenditures by Canadian residents abroad (C2511) or as expenditures by non-residents in Canada (C2521). These revisions were offset in the various expenditure categories such as tuition fees, food and accommodation and medical services.
Financial intermediation services indirectly measured
'Financial intermediation services indirectly measured' represents the cost of financial intermediation services households purchase when incurring loans or making deposits with financial institutions. Revised source data improved the value of this estimate.
- Household purchases of goods and services from non-profit institutions serving households: Some of the goods and services produced by NPISH are purchased by households. These purchases are reflected in household final consumption expenditure. Within the implementation of the CSNA12 and the exploitation of new NPISH data sources, the CSNA was able to confront aggregate estimates of household purchases from NPISH originating from household surveys with those directly reported by NPISH from administrative filings with the Canadian revenue agency. The administrative data indicated that the CSNA97 estimate of household purchases from NPISH were underestimated and was therefore adjusted upward with the implementation of CSNA12.
Table 9 shows the revision for each of the above noted expenditure categories for the period 2007-2011.
One of the more notable changes associated with the CSNA12 revision is the level shift downward in household final consumption expenditure's share of GDP. The removal of the NPISH final consumption expenditure and Aboriginal general governments' final consumption expenditure resulted in a substantial downward level revision to household final consumption expenditure. This served to reduce household expenditure's share of GDP. In addition, the upward shift in the level of GDP was mainly due to the capitalization of investment in research and development—increasing the overall share of investment in GDP, further driving down the household share.
The CSNA12 historical revision did not substantially revise the household final consumption expenditure growth rates. The mean absolute percentage point revision to the annual growth rate was .15 percentage points. The largest upward revision to the quarterly growth rate occurred in 2010, while the largest downward revision to the annual growth rate occurred in 2009.
The final consumption expenditure of NPISH in CSNA12 is presented separately in the final demand expenditure category providing quarterly information on this small but important sector of the economy. In addition to separating these estimates from those of households, there are a number of statistical revisions to this series.
As previously noted, with the implementation of CSNA12, the final consumption expenditures of NPISH have been removed from household final consumption expenditure and are now presented as a separate final demand expenditure category. Additionally, there are a number of statistical revisions to this series. In 2004, Statistics Canada released a satellite account highlighting the non-profit and volunteering sector. In doing so, Statistics Canada was able to invest in the exploitation of a large amount of administrative data. This data source resulted in a revision to the estimates of NPISH final consumption expenditures when compared to the estimate that was included in CSNA97.
Government final consumption expenditures were revised with CSNA12. The capitalization of government research and development and weapons served to lower government final consumption expenditure by approximately $10.5 billion in 2007. This was partially offset by the increased levels of consumption of fixed capital accompanying the larger capital stock and the methodological change in the calculation of government consumption of fixed capital. Changes to government consumption of fixed capital increased government final consumption expenditure by $15.0 billion in 2007. Finally, the addition of the Aboriginal general governments sub-sector increased total government final consumption expenditure in 2007.
In addition to the conceptual, methodological and sectoring changes, there were a number of statistical revisions to government final consumption expenditure. The total net change to government expenditure in the 2007 to 2011 period is highlighted in Table 10.
Gross capital formation represents both gross fixed capital formation (investment in building, roads, houses and machinery and equipment) and inventories. Like other components of expenditure-based and income-based GDP, gross capital formation has undergone a number of presentational changes.
The capitalization of research and development activity has resulted in a new investment category called intellectual property products. According to SNA2008, intellectual property products "are the result of research, development, investigation or innovation leading to knowledge that the developers can market or use to their own benefit in production because use of the knowledge is restricted by means of legal or other protection."(p.10:98) The CSNA12 has assigned three types of investment to this new investment category: research and development; software; and mineral exploration. In CSNA97, software was included in the machinery and equipment investment category while mineral exploration was included in the non-residential construction category. Table 11 shows the impact of this reclassification on the machinery and equipment and non-residential investment categories.
In CSNA97, investment was presented for two sectors—government and business. With CSNA12, investment is presented for three sectors—government, business and non-profit institutions serving households. In CSNA97, investment in the business sector included investment of corporations, unincorporated businesses, non-profit institutions serving households and Aboriginal general governments. In CSNA12, business investment only includes the investment of corporations and unincorporated businesses. NPISH investment is shown separately in CSNA12 and Aboriginal general governments' investment is shown in the government sector.
Government investment was revised upward with CSNA12 for three reasons. First, the capitalization of research and development reclassified government current expenditures to capital expenditures. Secondly, the capitalization of military weapons systems also reclassified government current expenditures to capital expenditures. Finally, the reclassification of Aboriginal general government investment from the business sector to the government sector resulted in another upward adjustment. A reconciliation between the CSNA97 and CSNA12 government is provided in Table 12.
Overall, gross fixed capital formation increased $18.9 billion in 2007. The majority of the increase was due to the capitalization of both business and government research and development with the capitalization of military weapons systems contributing $1.3 billion.
The major statistical revision relates to an improved methodology for measuring business investment in cars and trucks. As noted earlier, Statistics Canada obtained new vehicle registration data which provided an improved measure of the sector making the purchase. Another important statistical revision involves estimates of the dis-investment associated with the sale of used vehicles by the business sector to the household sector. Gross fixed capital formation in the SNA is "measured by the total value of a producer's acquisitions, less disposals, of fixed assets during the accounting period plus certain specified expenditure on services that adds to the value of nonproduced assets." (SNA2008 p.10:32). In CSNA97, the disposal of used vehicles by the business sector to the household sector was underestimated—this has been rectified with CSNA12. The net result of the upward revision in the purchase of new cars and trucks by the business sector, and the upward revision in the dis-investment in cars and trucks by the business sector, was an overall downward revision in the net investment in this asset category.
The statistical revisions to inventories are mainly a result of capturing statistical revisions elsewhere in the system, such as changes in household expenditures, statistical revisions to imports and exports of services and statistical revisions to gross fixed capital formation. Table 13 presents a reconciliation between the CSNA97 estimate of gross fixed capital formation and the CSNA12 estimate
The CSNA12 revision did not change the overall trend in investment for the period 1981 to 2012. The mean absolute percentage point revision to the annual growth rate was .70, with the largest upward revision occurring in 2009 and the largest downward revision occurring in 1997. As noted earlier, investments share of overall GDP increased for the entire time period, due primarily to the expansion of the asset boundary associated with CSNA12.
International imports and exports of goods and services
CSNA12 introduced both conceptual changes to estimates of international imports or exports of goods and services as well as revisions that resulted from improved data sources or changes to the services and goods classifications.
The only conceptual change is a result of implementing the new Balance of Payments Manual, 6th edition (BPM6). In BPM6 international trade in patents is recorded as international trade in research and development services. This is recorded under international trade in commercial services. Previously, patents were considered non-produced assets with the resulting transactions recorded in the capital account.
The most important change is the revision associated with imports and exports of commercial services. There are three components to the revision: a statistical revision, a change in the treatment of insurance; and the reclassification of financial intermediation services indirectly measured (FISIM) from its own commodity group to commercial services.
In 2006, a project was undertaken to capture and compare all available sources of data related to trade in commercial services. The project took advantage of a linking exercise between the Balance of Payments survey frame and Statistics Canada's Business Register, to improve the delineation of the population of enterprises engaging in imports and exports of commercial services. During the same period, Statistics Canada's Unified Enterprise Survey program added a specific module on international trade to several of its surveys. Despite some limitations in the details available, that module covered both exports and imports, an improvement over previous modules that collected very little information related to imports. Information from these surveys was used to confront the previous estimates of commercial service imports and exports, as well as to improve the population of enterprises engaging in commercial service imports and exports. The last major change to the service program was an increase in the survey's sample size. The number of units surveyed by the International Trade in Commercial Services Survey (BP-21S) was increased from 1,800 to 3,500 between reference year 2006 and reference year 2007. Also, starting with reference year 2007, it was decided to establish a scientific sampling methodology to increase the coverage and to increase the representation of every economic sector. The universe of enterprises was defined by including all units that have reported trade in service activities (above a certain threshold) and it was complemented by all other enterprises that responded to some criteria about their size and industrial sector. The impact of this revision was backcasted to 1997, the time period at which statisticians in the Balance of Payments program felt the undercoverage problem started. Table 14 presents the revision to international imports and exports of commercial services between 2007 and 2011.
Imports and exports of insurance services are now presented on a net basis in CSNA12. Previously, imports and exports of insurance services were recorded on a gross basis where all receipts (either premiums or claims) were exports and all payments were imports. This was inconsistent with the international standards which treat insurance services as the 'margin' between premiums and claims. CSNA97 did not follow International Monetary Fund's Balance of Payments Manual, 5th edition (BPM5) and considered all premiums and all claims as services. CSNA12 approaches the recommendation laid out in the international standard, by making the imports and exports of insurance services the difference between the premiums received and the claims paid (after adjusting the claims for unusual low or high values, which was also an international recommendation). The adjustments are then added to current transfers).8
Financial intermediation services indirectly measured were shown separately from commercial services in CSNA97. In CSNA12, FISIM and other commercial services have been combined into the commercial services export and import category. This reclassification is an additional change to the level of commercial services but has no impact on the total level of trade in services.
Another important change to the international trade data was the reclassification of repairs from trade in goods to trade in services. Following the BPM6 change to international standards, values of international repairs are now classified as services. BPM6 emphasizes the distinction between goods and services. With previous standards, the values of repairs (not the values of goods being repaired) were included under goods. Adding repairs to services rather than to goods is in line with the change of ownership concept. There is no change to the total trade in goods and services. Table 15 shows the revision to trade in goods and the offset in trade in services, resulting from this decision.
There were also a number of statistical revisions made to the international imports and exports of goods. Most of these centered on the treatment of military goods that were moved from Canada to other countries for military operations. Under the Balance of Payments concepts the shipment of these goods should not be recorded as Canadian imports and exports. In some cases, over the last number of years, the shipment of these goods was not removed from the customs data used to compile Canadian imports and exports. A thorough review of the shipment of military goods was conducted as part of the historical revision and these transactions have now been removed.
Revisions to the current and capital accounts
As noted, one of the major changes associated with CSNA12 is the articulation of new sectors, including non-profit institutions serving households, non-financial corporations, financial corporations and Aboriginal general governments. Many of the changes noted above filter down into the current and capital accounts for these sectors, as well as the other CSNA sectors such as the household sector, general government sector and non-resident sector.
An important change to the current and capital accounts is their presentation. CSNA12 uses the SNA2008 presentation to show the flow of income, expenditure, saving, investment and net lending of each sector. This new presentation and the rationale for the change are explained in the article A new presentation for the quarterly National Accounts (Wilson, 2012). Aside from the presentational change, there were some notable changes to a number of the macro-economic indicators that emanate from the sector accounts.
Household sector aggregates
Household primary incomes represent "incomes that accrue to [households] as a consequence of their involvement in processes of production or ownership of assets that may be needed for purposes of production." (SNA2008 p.7:2).
The majority of this income represents compensation related to the labour services provided by households. As previously noted, there were a number of statistical revisions that impacted the CSNA measure of compensation of employees (COE). These revisions served to reduce household primary income. In addition to these statistical revisions, another important revision involved the recording of non-resident COE payments to residents and resident COE payments to non-residents. In CSNA97, it was assumed that these payments cancelled out and therefore they were not articulated in the sector accounts. A review of this methodology indicated that this was not the case and in fact resident COE payments to non-residents are significantly larger than non-residents COE payments to residents. This served to further revise downward household primary incomes. The total impact of the revisions to COE in the household sector is reflected in Table 16.
Another important source of household primary incomes is property income. Property income represents the income received by households through their ownership of assets that were needed for the purposes of production. These represent income flows such as interest on investments and dividends. In the past, a large number of these income flows for the household sector were derived residually. For example, with respect to dividends flowed to the household sector, first the CSNA would estimate the total amount of dividends declared, from this dividends flowing to corporations and non-residents would be subtracted and difference (residual) would represent the dividends received by households. With CSNA12, a large amount of research was undertaken to see if these flows could be directly measured using administrative data (such as T3, T5 and T1 tax forms). As a result, with CSNA12 many of the household property income flows are directly computed from the administrative data files. This research indicated that the level of property income in CSNA97 was underestimated beginning around the year 2000, largely due to the misallocation of the investment flows associated with income trusts.
The removal of non-profit institutions serving households and Aboriginal general governments (AGG) had a significant impact on the household sector's current and capital accounts. A significant amount of the income received by NPISH and AGGs are in the form of transfers from the federal and provincial general governments. In CSNA97, these transfers were shown as income flows into the household sector. With CSNA12, they are shown as income flows into their proper sectors. In addition, households transfer a substantial amount of their income to NPISH. In CSNA97, these flows were not articulated since they represented both households and NPISH, which were in the same sector. With CSNA12, these income flows are shown explicitly. As a result, the CSNA12 level of current transfers received by household has been reduced substantially as shown in Table 17.
In CSNA97, household disposable income was defined as total household income (compensation of employees, net mixed income, property income received and current transfers received) minus current transfers paid to general governments. This differs from the international definition where household disposable income equals compensation of employees plus net mixed income plus property income received less property income paid plus current transfers received, minus total current transfers paid (to all sectors not just general governments) equals household disposable income. CSNA12 has adopted the SNA2008 definition of disposable income, resulting in a downward revision in the level of household disposable income. Table 18 provides a reconciliation between the previous estimate and the revised estimate of household disposable income, accounting for all types of revisions (statistical, conceptual and definitional).
Household savings is calculated by taking household disposable income minus household final consumption expenditures. As noted previously, both household disposable income and household final consumption expenditures were revised downward with CSNA12, with the majority of this explained by the removal of NPISH and Aboriginal general governments. Given that the downward revision to household disposable income was larger than the revision to household final consumption expenditure, the level of household saving was revised downward from 1981 to 2005. The overall trend remains the same, with the household saving rate falling progressively from the 1980s through the 1990s, until the late part of the 2000s, where it has since begun an upward trend. Chart 4 plots the previous household saving rate against the revised household saving rate.
Government sector aggregates
General government incomes were not revised substantially with CSNA12. While there were some statistical revisions to taxes and subsidies, these did little to change the overall level of general government income. While the level of government gross saving (which includes net saving plus consumption of fixed capital) was not revised significantly the level of net saving was revised downward due to the upward revision to consumption of fixed capital which increased overall general government final consumption expenditures.
The level of investment increased due to the capitalization of research and development and military weapons systems. The overall trend in net lending or borrowing remained the same, with the general governments borrowing in the 1980s and 1990s, lending to the other sector through most of the 2000s and then borrowing again in the 2008 to present time period.
Corporate sector aggregates
In CSNA97 current and capital accounts were only produced for the total corporate sector. With CSNA12 current and capital accounts are produced for both the financial corporations and non-financial corporation's sectors. These accounts represent a new source of information for users of Canadian System of National Accounts data series. These accounts show the flow of income to both the financial corporation and non-financial corporation's sectors, the net savings of these sectors, as well as their net lending or borrowing.
CSNA12 represents the most substantial change to the Canadian System of National Accounts since 1997. It provides users with the most up to date portrait of the Canadian economy. This is achieved through the implementation of new classification systems, an increased level of detail and a presentation that aligns with the international standard.
It is envisioned that in the future these types of revisions will be undertaken on a more frequent basis, to ensure that the Canadian accounts always remain relevant and internationally consistent. The next historical revision is planned for 2014, where the envisioned changes are minor in comparison to those associated with CSNA12. Statistics Canada will continue to keep users abreast of any proposed changes, in advance of the release of any revised data, to ensure they are well prepared to use and take advantage of the new data.
- CANSIM tables 380-0063 to 380-0084.
- The year 2007 is used throughout this article as a year of comparison between CSNA97 and CSNA12, as it represents the new base year and reference year for CSNA12.
- Referred to as compensation of employees in CSNA12
- Referred to as consumption of fixed capital in CSNA12
- Note that this compensation of employees is included in the household sector's current and capital account as these payments constitute a part of Canada's gross national income.
- Note that in CSNA97, government consumption of fixed capital was measured at replacement cost but employed a linear method of depreciation.
- However, only data originating from files starting with reference year 1997 were incorporated into the CSNA12 estimates.
- Occasionally the claims are adjusting for unusual low or high values as recommended by the international standard. In these cases these adjustments are recorded as current transfers.