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.
- What is a self-contained labour area?
- Results: self-contained labour areas (SLAs) of Canada
- How "rural" are the self-contained labour areas?
- Strictly-RST SLAs as "rural" SLAs
- Mainly-census-rural SLAs as "rural" SLAs
- Overview and combining the two definitions
- Utility of self-contained labour areas
- An alternative to the MIZ classification
- Locate rural connections and integration
- SLAs as a platform for additional data
- Combine with other data to create functional areas
- Appendix tables
We defined a self-contained labour area (SLA) as a group of two or more census subdivisions (CSDs) where at least 75% of the workers both live and work in the area (Box 3). The SLAs were created by grouping together CSDs that presented reciprocally important commuting flows between themselves (Box 3).
The 'labour' in the title of the "self-containted labour areas" therefore refers to the movement of labour from the place of residence to the place of work (Box 2). Commuting flows are generally used to proxy other types of connections between CSDs. For example, the movement of workers can also be used to reflect other ties such as shopping patterns and the use of services (Tolbert and Sizer, 1990). In addition, commuting flows are used as the basis for the MIZ classification system (Box 1). Using commuting flows, therefore, allows for a comparison and possible integration of the classification of CSDs to MIZ and the classification of CSDs to SLAs. Several interesting applications will be discussed in more detail later.
Once it has been accepted that commuting ties can be taken to represent a degree of connection or integration between two areas, the question then becomes how to calculate the strength of that relationship. Various approaches have been taken by other researchers to create labour market areas (e.g. Puderer, 2008). These procedures have ranged from considering only the combined number of commuters moving between the two areas to creating a percent measure based on the size of the smaller of the two areas. We chose to use a clustering procedure that recognized the importance of the commuting flow to both the sending and the receiving location (i.e. "reciprocal importance," Box 3). Details of the methodology are presented in Munro et al. (forthcoming).
Four additional points need to be recognized. First, our delineation of SLAs is a descriptive and not a prescriptive concept. It is describing what is happening in terms of commuting flows. It is not describing what could happen (due to proximity or more jobs, for example). Second, SLAs are based on one single, albeit relevant, dimension of connectivity: the movement of workers. Thus, each SLA is 'self-contained' in terms of these movements. Other types of connectivity are suggested, but not measured. Third, the lack of measured commuting for some neighbouring areas is an important finding on its own, which invites further analysis for these specific CSDs (poor roads, no jobs, commuting flow is too small to provide a reliable estimate of the commuting rate, etc.). For this reason, we report this result here and we did not re-group these CSDs based on some additional criteria (like proximity or adjacency). Finally, our SLAs are delineated using 2006 commuting flows. The flows may be expected to change if other factors change (e.g. new roads, job growth, etc.).
The clustering methodology adopted in this analysis (Box 3) provided the following results. Among the 5,418 CSDs in Canada in 2006, 1,256 CSDs were "out-of-scope" (Box 2). The "out-of-scope" CSDs had 128,164 inhabitants (0.4% of Canada's population). The 4,162 CSDs that were in-scope were clustered into 685 self-contained geographic units. Among these 685 self-contained geographic units, 336 were formed by a single census subdivision that showed no in-commuting and no out-commuting (Box 2). The remaining 3,826 CSDs were grouped into 349 SLAs, formed by two or more CSDs.
These 349 SLAs are the focus of our discussion. They are 96% self-contained, on average, which is significantly higher than the minimum required level (75%). On average, the resident workforce is 36,000 workers and the resident population is 89,000 inhabitants. The average SLA is comprised of 11 CSDs.
Map 1 shows the spatial delineation of SLAs. White areas on the map reflect CSDs that were out-of-scope for this study plus the 336 CSDs that appear to be 100% self-contained. Appendix Table A7 provides a list of which CSD is assigned to which SLA.
There are two general features of the 349 SLAs at the core of this analysis. First, and most important, these SLAs generally 'make sense' in terms of their physical adjacency. As can be seen even from this very high-level map, the SLAs are generally made up of adjacent subdivisions1. This corresponds to a general tendency for commuting to be stronger over shorter distances and is a positive feature of the results. Second, most CSDs that were strongly influenced by a larger urban centre were grouped with the larger urban centre.
The population size across the 349 SLAs varies considerably (Table 1). Three SLAs have a population of 2 million of more. These are SLAs that are centred on Montreal, Toronto and Vancouver. These 3 SLAs comprise 40% of Canada's population. Within these three SLAs, there are 652,000 census rural residents – representing 5% of the population of these SLAs and 11% of Canada's total census rural population.
Using the rural and small town (RST) definition of rural (Box 1) and again looking at the 3 largest SLAs, they include 343,000 RST residents (largely from the surrounding Strong MIZ CSDs) – representing 3% of the SLA population and 6% of Canada's total RST population.
When we consider the SLAs with smaller populations, we see there were 162 SLAs with a population less than 10,000 in 2006 (Table 1). They represented 46% of the 349 SLAs and they represented 2% of Canada's population. Within this group of SLAs, 65% of the population resided in census rural areas (and 35% resided in population centres of 1,000 or more). However, all the individuals in SLAs with a population under 10,000 resided in CSDs classified as a RST area.
One of the major purposes behind the development of this project was the exploration of rural labour areas. The question of whether or not we have managed to locate rural labour areas is complicated by the fact that there are different ways in which the concept of rurality can be defined. For the purposes of this bulletin, two complementary definitions of rural will be used - the rural and small town (RST) definition and the census rural definition (Box 1).
One way to classify SLAs is in terms of the type of CSDs that are included in a SLA. We used the Statistical Area Classification (Box 1) (Statistics Canada, 2007) and we classified the SLA along the CMA to No MIZ gradient, according to the highest ranking of any component CSD. Hence we defined the following types of SLAs:
- a CMA SLA if the SLA has one or more component CSDs that are delineated as part of a CMA.
Among the remaining SLAs, we assign a SLA to be a:
- Larger CA2 SLA if the SLA has one or more component CSDs that are delineated as part of a larger CA.
Among the remaining SLAs, we assign a SLA to be a:
- Smaller CA3 SLA if the SLA has one or more component CSDs that are delineated as part of a smaller CA.
Thus, strictly-RST SLAs are SLAs comprised only of CSDs that are classified as part of a RST area (i.e. these are CSDs that are not part of any CMA or CA). Thus, strictly-RST SLAs exclude any SLA with any component CSD that is part of a LUC.
Our classification continues by considering the remaining SLAs. We assign a SLA to be a:
- Strong MIZ SLA if the SLA has one or more component CSDs that are delineated to be Strong MIZ (metropolitan influenced zone).
Among the remaining SLAs, we assign a SLA to be a:
- Moderate MIZ SLA if the SLA has one or more component CSDs that are delineated to be Moderate MIZ.
And our classification continues for the other MIZ groups.
Using this definition of a "rural" SLA, we found 229 self-contained areas to be "rural" among the 349 SLAs with more than one component census subdivision (Appendix Table A3). These SLAs contained 2.2 million residents in 2006. Among all RST residents in Canada, 39% resided in one of these "strictly-RST" SLAs in 2006.
The distribution of the population across this urban-to-rural gradient shows that:
- 77% of Canadians live in a CMA SLA;
- 5% live in a larger CA SLA;
- 11% live in a smaller CA SLA; and
- 7% live in a "strictly-RST" SLA (Figure 1).
Thus, 93% of Canadians live in a SLA that is centred on a CMA or CA and 7% live in a "strictly RST SLA". The spatial pattern for this typology of SLAs is presented in Map 2.
The mainly-census-rural SLAs are SLAs where a majority of the population lives in census rural areas (i.e. in the countryside or in small settlements with less than 1,000 inhabitants) (Box 1).
Using this definition of a "rural" SLA, we found 197 "rural" SLAs among the 349 SLAs with more than one component CSD (Appendix Table A4). These SLAs contained 2.6 million residents in 2006 which represents 8% of Canada's total population (Appendix Table A4 and Figure 2, where the sum of all bars for SLAs that are "mainly-census-rural" is equal to 8%). Among all census rural residents, 29% resided in one of these mainly-census-rural SLAs in 2006. The spatial pattern for this typology is presented in Map 3.
Each definition provides an alternative perspective that can be used to examine the degree of rurality of the SLAs. When they are examined together, we can find the degree of overlap of SLAs that are designated as "rural" SLAs by each measure.
Specifically, there are 182 SLAs which we have classified as "rural" when both criteria are applied at the same time (Appendix Table A5). They are both "strictly-RST SLAs" and "mainly-census-rural SLAs". These 182 labour areas contain 1.8 million residents, representing 6% of all residents in Canada (Appendix Table A6).
Thus, we conclude that "rural" self-contained labour areas:
- do exist; and
- they represent a significant category for analysis.
This suggests that the use of the periphery-to-core commuting pattern for analysis may not always be appropriate. Specifically, many rural workers reside in "rural" SLAs and these labour markets are distinct from and not connected to the labour markets of larger urban centres.
At the same time, it is important to recognize that the majority (over 60%) of the "rural" population (by either definition) is located within SLAs with a lower degree of rurality. Thus, over 60% of rural Canadians are residing in a labour area that is not strictly or not mainly rural. It is also relevant to note that the lack of employment integration between a rural and an urban area does not necessarily indicate no interaction for other purposes – the residents of these "rural" SLAs may still patronize urban services.
The delineation of SLAs is a complement to the MIZ classification. It permits labour areas to be created based on rural-to-rural connectivity whereas the MIZ classification is based on the connectivity of a RST CSD to a LUC.
When we compare the MIZ and the SLA classification for southern Ontario, we see that the MIZ coding illustrates broad bands of colour radiating outward from the CMAs and CAs (Map 5). By comparing this to the SLA map (Map 4), it is possible to expand our knowledge of these broad bands of MIZ .
As a specific example, note the CSDs in the upper right corner of Map 5 (from Belleville towards Montreal). Looking at these CSDs in terms of their MIZ coding tells us that most of these CSDs are moderate or strong MIZ and that therefore more than 5% of their resident workers are commuting to work in a LUC. However, these CSDs are located between several urban centres – which ones are they connected to most strongly?
Turning to the SLA pattern (Map 4) provides an answer to this question, showing the ways a given type of MIZ group splits apart and joins with different LUCs. It is immediately apparent which CSDs are strongly associated with which LUC. Without the SLA system, this would only have been possible by selecting a particular CSD and manually locating its commuting flows.
This is an individual case, but the real strength of the technique is its ability to be expanded to provincial or national levels. This makes it easy to locate comparable groups or to track changes over time without initially selecting a site of interest.
One way to re-group SLAs for analytic purposes is by their degree of rurality, as discussed above. Compare the SLAs (Map 6) and the pattern of SLAs according to share of their population that is census rural (as one indicator of the degree of rurality) (Map 7). This clearly shows the SLAs where most of the population is living in census rural areas. These SLAs therefore constitute sparsely settled SLAs (because the census rural population lives in low density or sparse areas, by definition).
The sparse nature of rural settlement patterns is problematic in and of itself. By proposing a way to classify SLAs in terms of their degree of rurality, we are able to reduce the analysis from thousands of CSDs that are mainly census rural to a smaller number of SLAs that are similar in terms of their degree of rurality.
By using CSDs as building blocks, the SLAs provide a platform for the tabulation and presentation of additional data. For example, the data to indicate the demand for road infrastructure may be tabulated (such as the number of workers who drive their vehicle to work). The demand for community college infrastructure may be indicated by the number of individuals in high school who are residing in this "functional" SLA. Since our SLAs, by construction, have stronger "within-group" ties and weaker ties with neighbouring jurisdictions, the SLA structure provides a platform to assemble and to present these data.
In addition, other data may be overlaid on the SLA structure. Map 8 shows an overlay of the highway road networks and the SLAs. In larger cities, commuting flow data is combined with other information and similar road network files in order to create projections of traffic patterns and utilization of services. The SLA classification allows a similar procedure to be followed in rural Canada. Census data provides one measure of the strength of commuting between rural areas. To the extent that commuting patterns are similar to shopping patterns and the pattern of use of other services (such as hospitals and post-secondary educational institutions), the SLA delineation is one place from which to build an understanding of these patterns.
SLAs offer a pattern of "functional areas" based on commuting patterns. Analysts may use these patterns plus other information to create their own functional areas. As an example of one potential application, one might compare our SLAs (Map 9) and health regions (Map 10) for southern Ontario. This comparison may be helpful for planners of health service delivery4. Note the weak correlation between the SLA boundaries and the boundaries of a health region (or a group of health regions).
This is a particularly interesting case because population health outcomes and health service utilization are potentially connected to an individual's workplace as well as to their residence. That is to say, people may be expected to use the services near to where they live or the services accessible on their way to their place of work. Also, the health regions and the SLAs may be examined in conjunction in order to understand natural disaster or health emergency situations.
This example illustrates one way a SLA classification might be used as a basis for comparing and understanding already established systems.
In this analysis we delineated 349 self-contained labour areas (SLAs) using commuting data from the 2006 Census of Population. These SLAs are clusters of two or more CSDs with strong and reciprocally important commuting flows. SLAs allow us to gain a better understanding of the labour market context within which workers live and work.
The driving idea that underpins this analysis is the evidence that rural-to-rural commuting is a key feature of some rural areas. The mapping of commuting linkages was therefore intended to further our understanding of labour market linkages across different types of regions. Statistics Canada's Statistical Area Classification delineates "rural and small town areas" (RST) in terms of the influence of a "larger urban centre" (LUC). The delination of SLAs presented in this analysis is less urban-centric and more sensitive to the multi-directional nature of commuting flows, compared to the MIZ (Metropolitan Influenced Zone) classification which is based on the degree of influence of LUCs. Thus, our delineation provides a framework that includes urban connections without being defined by them. The SLAs presented in this analysis show the existence of predominantly rural labour market areas where there are relatively higher linkages within the SLA and relatively lower linkages across SLAs.
Examining these SLAs in terms of the degree of rurality provided important results. Between 29% and 39% of Canada's rural population, depending upon how rural is defined, resides in SLAs that are rural. This is consistent with an earlier study that documented the importance of within-rural commuting flows.
The insights from this analysis and the use of this type of geographic delineation can further contribute to an understanding of income flows or patronage of various services (for instance retail, health or recreational services). At the same time, it is important to recognize that the majority of the rural population is located in a SLA with stronger connections to a larger urban centre. Thus, rural-urban linkages are important for these rural residents.
Our study represents an initial delineation. Many census subdivisions were too small to provide reliable estimates of "commuting rates" (or had no commuting flows) and these census subdivisions were not assigned to a self-contained labour area for the purposes of this study. Additional criteria (e.g. road networks, geographic proximity, etc.) could be used to create custom areas.
It is suggested that our pattern of SLAs will be useful for analysts to combine with their own data to build "functional areas" suitable for their specific purposes. Examples include road network patterns and the provision of health services.
Appendix Table A7 Concordance of each census subdivision to each self-contained labour area, Canada, 2006 (We show four selected self-contained labour areas; the complete table is available from the authors upon request.)
- A small group of CSDs are not contiguous to the SLA to which theyare assigned, as noted in Appendix Table A.7 (full table available upon request from the authors). Specifically, twenty CSDs are morethan 50 kilometres apart from the SLA to which they are attached.
- A "larger CA", as defined for this study, is a census agglomeration with more than 50,000 residents. These CAs have census tracts designated within the CA and are also known as "tracted CAs."
- A "smaller CA", as defined for this study, is a census agglomeration with less than 50,000 residents. These CAs do not have census tracts designated within the CA and are also known as "non-tracted CAs."
- The lines on Map 10 show each health region. The colours indicate which set of health regions are in the same peer group (Shields and Tremblay, 2002).