The Comparative Level of GDP per Capita in Canada and the United States: A Decomposition into Labour Productivity and Work Intensity Differences

1
The Comparative Level of GDP per Capita in Canada
and the United States A Decomposition into
Labour Productivity and Work Intensity Differences
By Jean-Pierre Maynard Canadian Productivity
Accounts Micro-economic Analysis
Division Statistics Canada OECD Seminar, Bern,
October 2006
2
Context

• Mandated in the fall of 2003 to attempt a
  comparison of productivity levels between Canada
  and the United States
• Published two studies in January 2005
• Baldwin al, A comparison of Canada-US
  Productivity Levels An exploration of
  measurement issues
• Baldwin, Maynard et Wong (BMW), The output gap
  between Canada and the United States the role of
  productivity (1994-2002)
• Canada and the United States are neighbours
• The two economies are highly integrated
• About 80 of Canadian external trade are with the
  United States.

3
Objective of this presentation

• Presenting the results of a third study
• With emphasis on the comparability of labour
  measures between the two countries
• Updating the level comparison to 2005
• The selection of labour sources matter
• Sometimes the best practice to compare levels
  between two countries consists in using a
  combination of labour data from different surveys
• For level comparisons, we should care about the
  comparability of the concept, the coverage and
  the accuracy or quality of the labour estimates
  used.

4
Three different measures gap in favour of the
United States () - 2000
LP Work Intensity Work Intensity Work Intensity Work Intensity
GDP / Pop Y / H H/ Pop H / E E / Pat Pat / Pop
M1 BMW Study -20 -7 -13 -6 -10 3
M2 LFS / CPS -20 -11 -9 -8 -5 4
M3 Official - Productivity Accounts -20 -14 -6 1 -10 3
5
Focus is on the relative labour market performance

• Source of these differences
• Average Hours worked
• Employment
• Population
• Methodology used for BMW
• S S(Jin x Hin) Vhin
• J Number of jobs
• H Average hours worked per job
• Vh Volume of hours worked
• Where i industry and nclass of worker

6
Similar surveys in both countries

• Canada
• Household surveys Labour Force Survey (EPA)
• Establishment survey (SEPH)
• Other surveys and administrative data

• United States
• Household surveys Current Population Survey
  (CPS)
• Establishment survey Current Employment
  Statistics (CES)
• Other surveys and administrative data

7
but different measurement challenges

• Canada and United States dispose of very similar
  surveys to measure their population and labour
  market.
• However, the statistical agency/ies of each
  country face different challenges
• Geography (more borders, weather, migration)
• Different labour regulations

8
Impact on level comparisons when labour and
population estimates are taken from the LFS for
Canada and the CPS for the U.S. - 2000
M Item Y / H H / POP H / E E / PAT PAT / Pop
1 Canada 33.0 864.6 1766 60.6 80.8
1 United States 38.1 980.3 1871 66.70 79,0
1 Diff (Can-US ()) -7 -13 -6 -10 3
2 Canada 32.5 877.4 1824 61.3 78.5
2 United States 36.2 959.1 1979 64.4 75.3
2 Diff (Can-US ()) -11 -9 -8 -5 4
9
(No Transcript)
10
(No Transcript)
11
From persons employed (LFS / CPS) to the number
of jobs used in productivity accounts
  In thousands  Canada USA Canada as a of the USA
  Persons employed LFS / CPS (1) 15 310 136 485 11.2
Multiple job holders LFS / CPS 756 7 691 9.8
- Unpaid absentees 674 2 076 32.5
Military personnel 82 1 464 5.6
Other adjustments for coverage 87 2 386 3.6
Number of SNA jobs (2) 15 559 145 950 10.7
  Percentage change (2 / 1) 2 7 -5
Note Employees includes incorporated self employed (SNA concept) Note Employees includes incorporated self employed (SNA concept) Note Employees includes incorporated self employed (SNA concept)
Self-employed includes unpaid family related workers (SNA concept) Self-employed includes unpaid family related workers (SNA concept) Self-employed includes unpaid family related workers (SNA concept) Self-employed includes unpaid family related workers (SNA concept)

12
(No Transcript)
13
(No Transcript)
14
Consequences

• The CPS measure of employment and working age
  population would be underestimated for part of
  the period
• The characteristics of the population
  underestimated is different than the overall
  population
• Highly hispanophone, many unauthorized migrants
  and temporary migrants.
• Because of this underestimation, it can
  potentially bias downward the employment to
  population ratio.

15
(No Transcript)
16
(No Transcript)
17
Impact on level comparisons when hours worked
from the BLS productivity growth are used for the
U.S. - 2000
M Item Y / H H / POP H / E E / PAT PAT / Pop
1 Canada 33.0 864.6 1766 60.6 80.8
1 United States 38.1 980.3 1871 66.70 79,0
1 Diff (Can-US ()) -7 -13 -6 -10 3
3 (1) Canada 33.0 864.6 1766 60.6 80.8
3 United States 38.1 913.1 1743 66.7 78.6
3 Diff (Can-US ()) -14 -6 1 -10 3

18
(No Transcript)
19
Comparing apples and oranges

• What do we get for Canada if we use a similar
  methodology than the US productivity program?

Hours per job, 2003 Hours per job, 2003 Hours per job, 2003 Hours per job, 2003
U.S. Canada Difference
Adjusted CPS-LFS (BMW) 1,844.4 1,733.1 111 hours
BLS-SEPH 1,714.8 1,607.8 107 hours
Difference 130 125
20
More comparable and better data

• Many studies comparing Time Use Survey data
  collected through time diary seem to confirm that
  household surveys like the Labour Force Survey
  produce reliable and comparable estimates between
  countries. (UK, Canada, USA, Finland)
• Some of these studies indicates as well that
  deriving hours worked from hours paid
  underestimate the hours worked. (UK study)
• Our adjustment to CPS hours data reduce the
  annual estimate by 4,7. An ATUS-CPS study
  mention that the CPS weekly data on actual hours
  overestimate the annual hours worked by 5.
  Frazis and Stewart, MLR, Dec 2004

21
(No Transcript)
22
Concluding remarks

• What is the level of GDP per capita in Canada?
  How does it compare to that of the U.S.? To what
  extent labour productivity or work intensity
  contributes to it?
• Answering these questions involves an empirical
  exercise that seems simple since it depends only
  on a small number of variables GDP, population,
  employment and hours that have been published on
  a regular basis since World War II by most
  statistical agencies.
• But in reality, making international comparisons
  even between countries as similar as Canada and
  the United States is not simple. Statistical
  agencies produce different variants of these
  primary indicators for different purposes and
  analysts that focuses on international
  comparisons should be very careful about the
  international comparability of these indicators.