Metropolitan Gross Domestic Product: Experimental Estimates, 2001 to 2009

View the most recent version.

Archived Content

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.

Archived

This page has been archived on the Web.

by Mark Brown and Luke Rispoli^Note 1
Economic Analysis Division

Start of text box

End of text box

The growing concentration of Canada’s population in cities^Note 2 has been accompanied by requests for more extensive measures of city economies.

To date, most analyses have relied on employment and income to assess metropolitan economies. These indicators measure the amount of, and returns to, labour used to produce goods and services, but neither offers a measure of the production of goods and services or gross domestic product (GDP).^Note 3

GDP provides a means to assess the importance and performance of metropolitan economies—that is, how much they contribute to provincial and national GDP and how effectively inputs, like labour, are converted into output.

Presented here are experimental estimates of GDP over the 2001-to-2009 period for 33 census metropolitan areas (CMAs) and the non-metropolitan portions of the nine provinces with CMAs.

Methodology

Four guiding principles were used to develop more economically meaningful estimates of metropolitan GDP. Specifically, these estimates must be 1) consistent, 2) comprehensive and 3) comparable, while maintaining 4) “geographic fidelity.”

• Consistent. Sub-provincial estimates of GDP must add to known provincial totals. Industry-level estimates of GDP by income component^Note 4 must sum to provincial aggregates of current dollar GDP. This ensures consistency across the national accounting system.
• Comprehensive. Sub-provincial GDP estimates must encompass the entirety of the economy covered by the National Accounts, so that metropolitan areas with different economic structures are comparable.
• Comparable. Definitions of geography and industry must be consistent through time. This ensures that shifts in the size and industrial structure of economies are not due to changing definitions.
• Geographic fidelity. Income generated by the factors of production—land, labour and capital—is allocated to where the factor is employed, using records geocoded to that location. For instance, returns to capital are reported where the capital is used rather than where profits are reported.

These principles ensure that performance measures like productivity can be consistently estimated from these data. The Appendix contains further discussion of the methods used to produce metropolitan GDP.

Concentration of economic activity in metropolitan areas

Economic activity in Canada tends to be concentrated in cities. About half of Canada’s GDP is produced in the six CMAs with a population of 1 million or more—Toronto, Montréal, Vancouver, Calgary, Edmonton, and Ottawa–Gatineau. Even within this group, output is highly skewed. In 2009, about 1 out of every 5 dollars of the country's GDP was produced in the Toronto CMA (Table 1). Toronto accounts for less than 1% of Canada's land mass, but has an economy that is larger than that of every province except Ontario and Quebec.^Note 5

Growth through the 2000s shifted toward Calgary and Edmonton. The Calgary and Edmonton CMAs combined had less than half the population of Toronto, but gained close to the same amount of GDP ($62 billion versus $71 billion) from 2001 to 2009. Moreover, during the 2001-to-2009 period, only 9 of the 24 CMAs east of Ontario gained GDP share, while 8 of the 9 CMAs west of Manitoba increased their GDP share. See Appendix Table 1 for complete estimates of GDP by CMA and provincial non-CMA.

The share of GDP in non-CMA areas rose between 2001 and 2005, and then dropped. Because GDP is presented in nominal dollars, growth comes from changes in the volume and price of goods and services produced. The evolution of GDP shares in non-CMA areas coincides with commodity price shifts during the period.

Table 1
Gross domestic product, large census metropolitan areas, 2001, 2005 and 2009
Table summary
This table displays the results of Gross domestic product Gross domestic product , Share, 2001, 2005 and 2009, calculated using billions of dollars and percent units of measure (appearing as column headers).

	Gross domestic product			Share
	2001	2005	2009	2001	2005	2009
	billions of dollars			percent
Census metropolitan areas	741	894	1,064	71.8	69.8	72.2
Large census metropolitan areas	514	622	747	49.8	48.5	50.7
Toronto	202	242	274	19.6	18.9	18.6
Montréal	116	134	158	11.2	10.5	10.8
Vancouver	68	84	103	6.6	6.5	7.0
Calgary	43	57	75	4.2	4.5	5.1
Edmonton	39	50	69	3.8	3.9	4.7
Ottawa–Gatineau	46	55	68	4.5	4.3	4.6
Other census metropolitan areas	226	272	316	21.9	21.3	21.5
Non-census metropolitan areas	292	387	410	28.2	30.2	27.8
Canada	1,032	1,281	1,473	100.0	100.0	100.0
Note: Numbers may not add to total because of rounding. Sources: Statistics Canada, authors' calculations based on data from multiple sources.

The east–west pattern of growth is also reflected in the industrial structure of metropolitan economies. At the most aggregate level, the economy can be divided into goods- and services-producing^Note 6 industries. For the large, eastern CMAs, the goods-producing industries’ share of output declined throughout the period (Table 2). For the large western CMAs, goods-producing industries maintained their share of output until 2005, and then fell off relative to services as the recession in 2009 impacted goods- producing more than service-producing industries. This is consistent with the more pronounced decline in the volume of manufacturing industries in Ontario and Quebec through the 2000s (Brown 2014).

Table 2
Gross domestic product shares of goods- and service-producing industries, by large census metropolitan areas, 2001, 2005 and 2009
Table summary
This table displays the results of Gross domestic product shares of goods- and service-producing industries Goods-producing industries, Service-producing industries, 2001, 2005 and 2009, calculated using percent units of measure (appearing as column headers).

	Goods-producing industries			Service-producing industries
	2001	2005	2009	2001	2005	2009
	percent
Census metropolitan areas	27	25	22	73	75	78
Large census metropolitan areas	25	23	21	75	77	79
Toronto	26	23	20	74	77	80
Montréal	29	25	22	71	75	78
Vancouver	19	19	17	81	81	83
Calgary	29	29	26	71	71	74
Edmonton	31	31	29	69	69	71
Ottawa–Gatineau	15	12	11	85	88	89
Other census metropolitan areas	31	29	24	69	71	76
Non-census metropolitan areas	49	52	43	51	48	57
Canada	33	33	28	67	67	72
Note: Numbers may not add to total because of rounding. Sources: Statistics Canada, authors' calculations based on data from multiple sources.

Nominal gross domestic product per capita

GDP per capita is a measure of the value of output per person living in a metropolitan area. While it is tempting to think of it as a measure of labour productivity (GDP per hour worked), this is only part of the picture. GDP per capita in a metropolitan area will be higher when labour productivity is higher; each worker, on average, works more hours; more workers are employed; or the working-age population is larger. This can be expressed as:

$$\underset{\begin{array}{c}\text{Per capita }\\ \text{gross domestic product}\end{array}}{\underbrace{\frac{GDP}{Pop}}}\equiv \underset{\begin{array}{c}\text{Labour }\\ \text{productivity}\end{array}}{\underbrace{\frac{GDP}{Hours}}}\times \underset{\begin{array}{c}\text{Average hours }\\ \text{worked}\end{array}}{\underbrace{\frac{Hours}{Employment}}}\times \underset{\begin{array}{c}\text{Employment }\\ \text{rate}\end{array}}{\underbrace{\frac{Employment}{Po{p}^{15-65}}}}\times \underset{\begin{array}{c}\text{Working age }\\ \text{population}\end{array}}{\underbrace{\frac{Po{p}^{15-65}}{Pop}}},$$

where:

GDP = Gross Domestic Product
Hours = Total hours worked
Employment = Number of workers employed
Pop^15–65 = Working age population (aged 15 to 65)
Pop = Total population

Therefore, GDP per capita reflects not only labour productivity, but also, labour market conditions and demographics. This is an important distinction. Metropolitan GDP is a measure of where output takes place, but it does not take into account where workers live. If a significant portion of a CMA’s working-age population is employed outside its CMA of residence (for example Oshawa), the ratio of employment to working-age population will be lower, and so, too, GDP per capita.^Note 7

Despite its limitations, GDP per capita reflects the underlying dynamics of the Canadian economy through the 2000s. Of the CMAs in the top 10 in terms of GDP per capita in 2001, Kitchener–Waterloo, Halifax, Windsor and Oshawa were no longer in the group by 2009, replaced by St. John’s, Saskatoon, Victoria and Vancouver (Table 3). This pattern is consistent with a broad-based shift from manufacturing towards resource-based production. Of the nine CMAs with 25% or more of their output in manufacturing at the start of the period, six fell in rank, all of them in Ontario (Chart 1). By contrast, CMAs serving regions with expanding commodity-based economies increased. For example, Saskatoon rose 14 places, from 20th to 6th, in tems of GDP per capita, and St. John’s rose 10 places, from 15th to 5th. All the large eastern metropolitan areas lost relative ground. Ottawa–Gatineau fell 2 places (2nd to 4th); Toronto, 4 places (3rd to 7th); and Montréal, 6 places (11th to 17th). See Appendix Table 2 for complete estimates of GDP per capita by CMA and provincial non-CMA.

Table 3
Gross domestic product per capita, census metropolitan areas ranked in top 10 in 2001, 2005 or 2009
Table summary
This table displays the results of Gross domestic product per capita Nominal gross domestic product
per capita, Census metropolitan area rank, 2001, 2005, 2009 and Rank change, 2001 to 2009, calculated using dollars and number units of measure (appearing as column headers).

	Nominal gross domestic product per capita			Census metropolitan area rank
	2001	2005	2009	2001	2005	2009	Rank change, 2001 to 2009
	dollars			number
Regina	38,737	47,465	65,404	6	4	1	5
Calgary	44,438	52,681	61,246	1	1	2	-1
Edmonton	40,355	48,268	59,941	5	3	3	2
Ottawa–Gatineau	41,643	47,176	55,506	2	5	4	-2
St. John's	31,385	37,994	49,844	15	14	5	10
Saskatoon	30,572	38,220	49,213	20	12	6	14
Toronto	41,397	46,001	48,532	3	6	7	-4
Victoria	30,640	37,149	46,763	19	15	8	11
Vancouver	32,680	38,822	44,249	12	11	9	3
Guelph	41,143	48,410	44,217	4	2	10	-6
Kitchener–Waterloo	35,258	40,824	43,989	8	8	11	-3
Halifax	32,982	39,182	43,471	10	10	13	-3
Sudbury	28,727	42,162	42,138	24	7	14	10
Windsor	34,739	39,567	36,194	9	9	24	-15
Oshawa	37,551	32,507	28,918	7	25	32	-25
Sources: Statistics Canada, authors' calculations based on data from multiple sources.

Description for Chart 1

GDP per capita also follows a distinct pattern across non-CMA regions (Chart 2), with a growing difference between regions that are oil- and gas-producing and those that are not. The rising volume and/or price of oil and gas production is evident in the non-CMA regions of Alberta, Saskatchewan and Newfoundland and Labrador^Note 8 between 2001 and 2009. By the end of the period, non-CMA regions in Alberta and Saskatchewan, and to a lesser degree Newfoundland and Labrador, had significantly higher GDP per capita than other non-CMA regions.

Description for Chart 2

One of the more consistent features of urban economies is that the larger they are, the more productive they tend to be.^Note 9 Per capita GDP, while confounded by labour market and demographic effects, tends to be higher in larger metropolitan areas, particularly those with a population greater than 1 million (Chart 3). GDP per capita also tends to be higher in CMAs than non-CMAs, but this distinction is only revealed when regions that specialize in oil and gas production are excluded—namely, Alberta, Saskatchewan and Newfoundland and Labrador (see Charts 2 and 3).

Description for Chart 3

Conclusion

This paper employs a new experimental metric to measure the contribution of GDP by CMA from 2001 to 2009. The analysis uses data sources and methods similar to those used in the Canadian System of National Accounts to estimate GDP across CMAs and non-CMAs. The estimates reveal an economy that is highly concentrated in cities, particularly in the large eastern metropolitan areas, but also one that experienced significant geographic shifts through the 2000s, with output, as measured by GDP, shifting toward the cities of western Canada.

Appendix: Methodology

Census metropolitan area (CMA)^Note 10 gross domestic product (GDP) is estimated by income component (wages and salaries + supplementary labour income + mixed income + operating surplus^Note 11 [primarily corporate profits] + indirect taxes on production less subsidies) across 20 goods- and service-producing industries.^Note 12 These income components by industry are then benchmarked to published provincial-level GDP totals from the input-output accounts. ^Note 13

The estimate of metropolitan GDP developed here allocates output to locations where economic activity takes place. For the business sector, wages and salaries and operating surplus, which together accounted for 80% of GDP in 2008,^Note 14 are allocated to locations based on firm-level microdata. The structure of firms and the location of their production units are defined using the Business Register. For simple firms with one location, wages and salaries and surplus are directly assigned to the location of the production unit. For firms with more than one production unit (complex enterprises), employment in production units is used to allocate wages and salaries and surplus to locations, after adjusting these to the average wage rate and average profit per worker of the industry of the production unit.

In most industries employment and capital are located jointly, but this is not the case for utilities and the oil and gas industry. Consequently, in these industries operating surplus was allocated to where the capital goods are located.

GDP estimates for the non-business sector were based on labour income from the 2001 and 2006 censuses for the non-profit and government sector. The estimates for owner-occupied dwellings were based on a combination of average income of owner-occupied dwellings by CMA, as derived by Brown and Lafrance (2010), and the number of dwellings by CMA, from the 2001 and 2006 censuses.

References

Brown, W.M. 2014. Testing for Provincial Industrial Change. Economic Analysis Research Paper Series, no. 92. Statistics Canada Catalogue no. 11F0027M. Ottawa: Statistics Canada.

Brown, W.M., and A. Lafrance. 2010. Income from Owner-occupied Housing for Working-age and Retirement-age Canadians, 1969 to 2006. Economic Analysis Research Paper Series, no. 66. Statistics Canada Catalogue no. 11F0027M. Ottawa: Statistics Canada.

Brown, W.M., R. Chan, and L. Rispoli. 2014. Census Metropolitan Area Gross Domestic Product Methodology. Ottawa: Statistics Canada. Discussion paper.

Lemelin, A., P. Mainguy, D. Bilodeau, and R. Aubé. 2012. “GDP Estimates for Regions within the Province of Quebec : The Changing Geography of Economic Activity.” In Defining the Spatial Scale in Modern Regional Analysis: New Challenges from Data at Local Level, ed. E. Fernandez Vazquez and F. Rubiera Morollon, p. 107–137. Heidelberg: Springer.

Panek, S.D., F.T. Baumgardner, and M.J. McCormick. 2007. “Introducing New Measures of the Metropolitan Economy. Prototype GDP-by-Metropolitan-Area estimates for 2001-2005.” Survey of Current Business 87 (11): 79–114.

Puga, D. 2010. “The Magnitude and Causes of Agglomeration Economies.” Journal of Regional Science 50 (1): 203–219.

---

Notes