Quarterly GDP compilation at NBS

1
Quarterly GDP compilation at NBS

• By Roger Jullion (Statistics Canada)
• OECD/NBS Workshop on National Accounts
• Haikou, China
• December 2009

2
Overview of presentation

• Background of Statistics Canada project
• Progress to date
• Production based GDP
• Expenditure based GDP
• Reconciling quarterly GDP estimates
• Seasonal adjustment and benchmarking
• Looking ahead

3
Background of Statistics Canada project

• Statistical Information Management Program (SIMP
  II)
• Seven year project, began in 2006
• Three sub-projects
• Environment statistics
• Social Statistics
• Economic Statistics
• Business survey component
• National accounts component

4
Background of Statistics Canada project

• National accounts component (objectives)
• Produce discrete quarterly seasonally adjusted
  production-based GDP estimates in both current
  and constant prices and, on a trial basis,
  quarterly discrete expenditure-based GDP
  estimates in both current and constant prices
• Reconcile annual and quarterly expenditure-based
  and production-based GDP such that NBS would
  eventually be able to publish expenditure-based
  GDP together with their official production-based
  GDP estimates
• Build analytical capacity for future improvements
  of the overall quality of the GDP estimates, both
  production and expenditure-based.
• Improve the overall quality of the GDP estimates
  through improvements to methodologies and price
  deflation

5
Progress to date

• We have been working with NBS national
  accountants for three years
• Four visits to NBS by Statistics Canada and one
  visit to STC by NBS officials
• NBS officials will be visiting Canada again this
  coming spring
• We have discussed our respective sources and
  methods and strategies for analyzing and
  balancing the accounts
• Suggested data coherence tools and data
  confrontation practices including supply use
  analysis
• Discussed revision policy, benchmarking and
  seasonal adjustment
• Workshop to enhance communication with data
  suppliers
• Three years remaining in the SIMP II project

6
Progress to date

• Production-based GDP
• Quarterly discrete estimates by industry have
  been developed
• A combination of quarterly indicators and the
  subtraction method have been used to derive the
  quarterly data
• Quarterly data have been benchmarked to the
  annual value added estimates and subsequently
  seasonally adjusted using X12ARIMA
• Nominal and price data are being analyzed by
  industry

7
Progress to date

• Production-based GDP
• Key activities planned over the next three
  months
• Re-estimate the industry sub-component at the
  level of 39 industries above the cut-off
• Seasonally adjust the industry sub-sector at the
  39 industry groups and then aggregate
• Disaggregate the Others industry grouping into
  the 9 industry sub-sectors using suggested
  indicators as discussed at our September STC/NBS
  meeting.

8
Progress to date

• Expenditure-based GDP
• NBS has estimated internal trial-based quarterly
  expenditure-based GDP
• We have discussed improvements to the compilation
  methods and how the expenditure-based data could
  be reconciled or compared with the higher
  quality production-based estimates
• Where data gaps exist in the estimation of
  production-based GDP, some of the consumption
  indicators could be used
• Household consumption uses the rural and urban
  household expenditure surveys
• Rents, financial services compiled in conjunction
  with production-based GDP estimates

9
Progress to date

• Expenditure-based GDP
• STC suggested that the quarterly consumption
  estimates be estimated at a more detail level
• This would allow for better analysis and more
  accurate constant price estimates. It would also
  allow for other data to replace or be compared to
  the household survey data e.g. motor vehicle
  sales, services spending
• Government current spending
• Budget vs non-budgetary
• Capital vs current

10
Progress to date

• Gross fixed capital formation
• Construction output vs capital spending on
  construction
• Supply use analysis for ME
• Change in inventories
• Price deflators
• Supply use analysis
• Exports and Imports
• Exchange rate adjustments
• Unit values vs Producer Price Indexes

11
Progress to date

• Reconciliation of production-based with
  expenditure-based GDP estimates
• Quarterly and annual GDP comparison tables have
  been prepared
• Inconsistencies between the datasets have been
  identified
• Data confrontation and data coherence techniques
  have been discussed
• Alternative compilation methods have been
  introduced which are narrowing the discrepancies
• Additional issues were identified at the workshop
  held with national accounts data suppliers

12
Progress to date

• We have discussed the use of supply use analysis
  for certain commodities such as vehicles and ME
• Improvements to price deflators as well as the
  level at which deflation is carried out has been
  discussed and NBS is looking at alternative
  approaches
• Comparisons of implicit prices between
  production-based and expenditure-based GDP have
  been prepared such that constant price
  discrepancies can be decorticated between nominal
  values and prices
• NBS has made considerable progress on
  implementing benchmarking and seasonal adjustment
  techniques

13
Looking ahead

• Publishing seasonally adjusted GDP is a priority
  for NBS
• While production-based GDP estimates are the
  priority, we are working in parallel on improving
  the quarterly and annual expenditure-based
  estimates and the reconciliation of the two sets
  of estimates
• We are working with NBS on improving the
  quarterly discrete GDP estimates, using all
  available survey and administrative information
• We are discussing various data validation and
  data coherence techniques that we use at
  Statistics Canada to compile quarterly national
  accounts

14
Looking ahead

• We are working via the Statistics Canada/NBS SIMP
  II business survey and household data improvement
  projects to ensure that national accounts needs
  and key data gaps are addressed
• While we are recommending that Chinas GDP
  estimates should be compiled at a more detail
  level we recognize the resource constraints
  within the national accounts group
• Compiling and publishing both production-based
  and expenditure-based GDP quarterly ensures data
  coherence while meeting policy and analytical
  requirements
• We are discussing revisions and revisions
  policies (timeliness and data quality tradeoffs)