Income Research Paper Series
The proposed methodology for the 2023-base Market Basket Measure of poverty

The Market Basket Measure (MBM) was first introduced as Canada’s Official Poverty Line in 2018 following the release of Opportunity for All – Canada’s First Poverty Reduction Strategy. In addition to designating the MBM as the official metric to measure the Government of Canada’s poverty reduction targets, the Poverty Reduction Act also stipulated that the MBM should be reviewed on a regular basis to ensure it reflects the up-to-date cost of a basket of goods and services representing a modest, basic standard of living in Canada. This paper summarizes the proposed 2023-base MBM methodology for the third comprehensive review.

Statistics Canada and Employment and Social Development Canada (ESDC) initiated the third comprehensive review of the MBM, in June 2023.^Note  Since the launch of the latest review, numerous engagement and outreach activities have been undertaken, each of which were tailored towards capturing ideas, proposals and recommendations from different segments of the MBM user community (e.g., academic experts, non-governmental organizations, government officials [federal, provincial or territorial] and the public). The results of the third comprehensive review engagement phase are detailed and summarized in the What we heard report: Engagement activities for the third comprehensive review of the Market Basket Measure. The feedback and recommendations received during that phase lay the foundation on which the proposed 2023-base methodology was constructed.

This discussion paper outlines the proposed changes and updates for the 2023-base MBM methodology, provides preliminary 2023-base thresholds and poverty rates for 2023 reference year, and compares these with the 2018-base thresholds and poverty rates.^Note  In addition, the paper identifies research topics that are to be studied in preparation for the next comprehensive review. After the release of this discussion paper, a shorter consultation period will be held to discuss the results with the public, academic, non-governmental, and provincial and territorial partners. Finally, Statistics Canada and ESDC will review the feedback received and will determine whether corrections to the proposed 2023-base methodology are necessary, before it becomes final in fall 2025.

Introduction

The Market Basket Measure (MBM) establishes poverty thresholds based on the cost of a basket of goods and services that reflects a modest, basic standard of living. The poverty thresholds are defined for different regions across all Canadian provinces and territories and adjusted for different family sizes. Families’ thresholds are then compared with their disposable income to determine their poverty status.

Leading up to the third comprehensive review, research topics that were identified during the previous comprehensive review were analyzed and their findings were published.^Note  Each of these discussion papers explored its respective topic, aiming to foster engagement and debate with the public and stakeholders to help inform discussions for the comprehensive review, improve understanding of the MBM methodology, and potentially expand analytical tools that involve or rely on the MBM.

Following the official start of the third comprehensive review, Statistics Canada and ESDC have worked collaboratively to engage with Canadians, poverty experts, and officials from provincial, territorial and federal governments, and other stakeholders. During this outreach phase, each research topic was presented and discussed in varying degrees of detail. The feedback and advice received through the engagement activities have helped inform the decision-making process for the proposed 2023-base updates presented in this discussion paper. Following historical nomenclature standards, the current MBM methodology is referred to as the “2018-base MBM,” while the proposed methodology is referred to as the “2023-base MBM.”

This discussion paper will describe the proposed 2023-base changes for both the MBM and the Northern Market Basket Measure (MBM-N). Next, it will compare the official 2018-base MBM and MBM-N thresholds and poverty rates to the preliminary 2023-base thresholds and poverty rates. Finally, it will prioritize outstanding topics that require more research in preparation for the next comprehensive review. Following the publication of this discussion paper, a review period will be provided allowing time for feedback and any final recommendations to be received, after which the proposed 2023-base methodology would be updated (if necessary) and will become final in fall 2025.

Overview of the proposed changes to the 2023-base Market Basket Measure and Northern Market Basket Measure

The underlying structure of the proposed 2023-base MBM methodologies will remain the same as the 2018-base MBM methodologies.^Note  They will continue to establish poverty thresholds based on the cost of a basket comprising a food component; clothing and footwear component; shelter component; transportation component; and other expenses component. These components are based, where possible, on standards developed by experts in their field of study (e.g., the National Nutritious Food Basket developed by Health Canada). Moreover, where necessary, the expenditures and costs used in the calculations of thresholds and disposable income would reflect those of families in the second income decile.^Note  Families with disposable income less than the applicable thresholds, given family size and region of residence, are deemed to be in poverty. The following are the proposed changes to the 2023-base MBM and MBM-N methodologies. More detail on these proposals is provided in the appendixes.

Basket summary

The following section presents a summary of the methodology applied in the MBM and the proposed major changes to the basket components and other important methodological features that would impact the underlying standard used by the MBM.

 Regions

• 2018-base method: MBM regions in the provinces are either based on specifically defined communities or population centre size and province combinations. In the MBM-N regions for the territories they are based on specifically defined communities or Census Subdivision boundaries.
• 2023-base proposed method: The same standard would be used to redefine the MBM regions.
• Note: Even though the method has not changed, there are no longer any geographies that match the definition for Newfoundland and Labrador with a population between 30,000 and 99,999. Accordingly, this threshold is not produced for the 2023 base. The removal of this region would bring the total number of MBM regions in the provinces to 52, while there would be no change in the number of MBM-N regions for the territories which would remain at 13.

 Reference family size and composition

• 2018-base method: The reference family size and composition (i.e., sex and age of family members) would remain unchanged.^Note  For the provinces, Yukon and the Northwest Territories, the reference family consists of one male and one female adult aged 25 to 49 with two children (a girl aged 9 and a boy aged 13). The Nunavut family includes an additional child (a girl aged 4).
• 2023-base proposed method: The same standard would be used for the reference family and composition.
• Note: After the basket is calculated for the reference family, to arrive at thresholds for different family sizes, the MBM methodology would continue to use the square root equivalence scale.^Note 

Shelter component

• 2018-base method: The shelter component follows the Canada Mortgage and Housing Corporation’s (CMHC) National Occupancy Standard (NOS) for “adequate” (i.e., not in need of major repairs) and “suitable” housing (i.e., given the reference family’s size and composition). Therefore, a three-bedroom dwelling not in need of major repairs is the standard on which the shelter costs are estimated. Specifically, the shelter component costs presented in the basket represents the cost of a non-subsidized rental, while the costs for other tenure types (e.g., subsidized renters, homeowners with or without a mortgage) are accounted for on the disposable income side of the measure through tenure type adjustments (TTAs).
• 2023-base proposed method: The same standard would be used for the shelter component with minor changes to the estimation process.
• Note: Adjustments to the estimation model are being proposed to better reflect shelter costs. Also, while the 2018-base was benchmarked using data from the 2016 Census, the 2023-base would use data from the 2021 Census.

 Clothing and footwear component

• 2018-base method: The clothing and footwear component for the provinces reflects the latest Harvest Manitoba and Social Planning Council of Winnipeg’s Acceptable Living Level (ALL) 2012 clothing basket, while adjustments are made to reflect the colder climate and life in the territories.
• 2023-base proposed method: The same standard would be used for the clothing and footwear component with minor changes to the estimation process.
• Note: In keeping with methodological improvements to price collection implemented in the Consumer Price Index (CPI), the 2023-base MBM would now include web-scraping collection, where possible.^Note 

Food component

• 2018-base method: The food component uses Health Canada’s 2019 National Nutritious Food Basket (NNFB), which is based on the latest Canada’s Food Guide. The food component for Nunavut includes costs associated with country food.
• 2023-base method: The same standard would be used for the food component with minor changes to the estimation process.
• Note: Given improvements in food price collection methods, the MBM methodologies would follow Statistics Canada’s Consumer Prices Division’s (CPD) best practices by incorporating scanner data (where possible).^Note 

Transportation component

• 2018-base method: The transportation component reflects private and public transportation commuter patterns for people living in larger urban centres, while rural and smaller urban communities include costs associated with only private transportation needs.
• 2023-base proposed method: The same standard would be used for the transportation component with minor changes to the estimation process.
• Note: While the foundational methodology has not been altered, adjustments to the estimation process are being proposed to improve statistical quality and to better reflect commuter patterns of Canadians.

Other expenses component

• 2018-base method: The other expenses component^Note  uses a “fixed” multiplier methodology which is applied to the regional food and clothing component costs to arrive at a dollar amount for the other expenses component.
• 2023-base proposed method: The same standard would be used for the other expenses component with minor changes to the estimation process and adjustments to accommodate for the new communication services component.
• Note: Although fundamental changes are not being proposed, the expenditure values used in the multiplier would be updated to reflect the most recent Survey of Household Spending (SHS) data, and the population of interest would be updated to better reflect the MBM methodology and reference family. In addition, the proposed creation of a separate communication services component would impact the multiplier calculation since the communication services expenditure categories would need to be removed from the calculation to prevent double counting.

(New) Communication services component

• 2018-base method: Not applicable.
• 2023-base proposed method: The creation of a new communication services component.
•  Note: Following the announcement of a minimum communications service standard set by the Canadian Radio-television and Telecommunications Commission (CRTC), it is proposed that the MBM methodologies apply this standard in the costing of a separate communication services component. It is proposed that the following three sub-components make up the proposed communication services component: landline services, cell phone services and Internet access services.

Inuusiqattiarniq component (Nunavut only)

• 2018-base method: The MBM-N for Nunavut has an additional component used to estimate the costs of goods and services required to preserve Inuit knowledge, culture, traditions and way of life. Given the importance of maintaining and promoting these elements, these costs are presented in a separate component from the typical components found in the other MBM methodologies. The inuusiqattiarniq component costs use a multiplier methodology and are added to the cost of the basket for Indigenous families only.
• 2023-base proposed method: The same standard would be used for the inuusiqattiarniq component with minor changes to the estimation process.
• Note: The same fixed multiplier used in the 2018-base would be applied to the proposed 2023-base other expenses component value.

Disposable income summary

The following section presents a summary of the methodology and the proposed changes for the calculation of disposable income.

Tenure type adjustments

• 2018-base method: A tenure type adjustment (TTA) was applied to the disposable income of families to ensure that those living in different housing tenures^Note  than the one used to cost the shelter component (e.g., non-subsidized renters) are on an equal footing when considering shelter costs.
• 2023-base proposed method: The same standard would be used for the tenure type adjustments with minor changes to the estimation process.
• Note: Adjustments to the estimation model are being proposed to better reflect shelter costs. Also, while the 2018-base was benchmarked using data from the 2016 Census, the 2023-base would use the 2021 Census.

Medical expense imputation

• 2018-base method: Medical expenses are derived from administrative tax data and supplemented by survey data. When no medical expense data are available from tax data, a value is imputed using the latest SHS data for every adult in the family.^Note 
• 2023-base proposed method: The same underlying methodology would be applied with minor changes to the estimation process.
• Note: Changes are being proposed to the calculation of the imputation amount to better reflect the medical expenses of families.

Comparison between the preliminary 2023-base Market Basket Measure and official 2018-base Market Basket Measure thresholds and poverty rates

The following section describes the impact of the proposed 2023-base changes on the poverty thresholds and rates. To reiterate, the rebased poverty measure is referred to as the “2023-base,” while the existing measure is referred to as the “2018-base.” Also, as previously mentioned in this discussion paper, because the comprehensive review is not yet complete, these results should be treated as preliminary.

Table 1 shows the thresholds for the reference family for each MBM region under both the 2018-base and the 2023-base MBM methodologies for 2023. For the provinces, Yukon and the Northwest Territories, the MBM reference family consists of two adults with two children. For Nunavut, the MBM reference family consists of two adults with three children. Overall, the proposed 2023-base MBM methodology would have varying impacts on the different MBM regions. For example, under the proposed 2023-base methodology, the majority of the MBM regions in the Atlantic provinces, Alberta, the Northwest Territories and Nunavut would experience a relatively small decrease or no change in their total thresholds value compared with the 2018-base methodology. Conversely, the majority of the remaining MBM regions for the provinces and territories would see slight to modest increases or very little change. In general, on average, at the Canada-level there would be around a 2.0% increase in the proposed 2023-base total threshold compared with the previous base.^Note 

The preliminary 2023 poverty estimates, using the proposed 2023-base MBM methodology, are presented in Table 2. On average, the recalibration of the proposed 2023-base thresholds would change the Canada-level^Note  poverty rate, in 2023, by 0.7 percentage points, going from 10.2% using the 2018-base to 10.9% using the proposed 2023-base. For reference, in 2018, during the last rebasing the 2018-base poverty rate increased by 2.3 percentage points from the 2008-base. However, given the agreed upon duration between comprehensive review periods is every five years, relatively large increases between rebasings should no longer occur.

Table 2 shows that the majority of provinces and territories with lower proposed 2023-base thresholds compared to the 2018-base would have lower poverty rates between the bases and vice-versa. Although the levels for the select demographic groups are different, the trends over time and relative to each other are maintained when comparing the poverty rates between the proposed 2023-base and 2018-base methodologies.

Future research agenda

As was the case during the previous comprehensive review, the following topics have been identified by Statistics Canada and ESDC because they require further research or exploration given their complexity and the methodological challenges they present to the MBM. Similar to the last time, the results of the research will be published to allow for transparency and to solicit feedback and recommendations for next steps.^Note 

Different family compositions: Exploring methods for the derivation of poverty thresholds for families of a different composition than the MBM reference family (e.g., seniors, people with disabilities).

Income decile evaluation: Evaluating the appropriateness of the continued use of the second before-tax income decile to derive key parameters for the MBM.

MBM based on the household: Research will be undertaken to evaluate the impacts of changing the MBM standard from the economic family unit to the household unit.

MBM-N-specific topics: Exploring research topics previously announced, which are specific to Yukon, the Northwest Territories (Revised Northern Food Basket and Country Food and Harvesting) and Nunavut (community-level sharing, second-hand clothing and in-kind benefits).

Debt servicing: Exploring the feasibility and merit of including expenses related to servicing debt in the calculation of MBM poverty rates.

Material deprivation indicators: Exploring the development of material deprivation indicators as complementary indicators to the MBM.

MBM on administrative data: Continuing to advance previous research on deriving MBM indictors using administrative data.

Conclusion

This discussion paper described the proposed changes and updates to the MBM and MBM-N methodologies resulting from the third comprehensive review of the MBM. In addition, the official 2018-base and preliminary 2023-base MBM thresholds are presented for comparison, as are the impacts of the new base on the poverty estimates. Finally, a future research agenda has been created in preparation for the next comprehensive review.

The purpose of this paper is to foster engagement with the MBM user community, to explain the changes that have been proposed for the 2023-base MBM methodology and to provide users with the impact of the proposed changes on the thresholds. We encourage and welcome users to ask questions, provide feedback and make suggestions for future work.

Following the publication of this paper, a review period will commence, where Statistics Canada and ESDC will engage with experts; stakeholders; and federal, provincial and territorial officials to help validate the results. The review period is expected to end in fall 2025. Individuals or organizations interested in contacting us are encouraged to do so by sending an email to statcan.mbm-mpc.statcan@statcan.gc.ca

Appendix A: Proposed changes and updates to the components and disposable income of the Market Basket Measure methodologies

Proposed changes and updates to the components of the Market Basket Measure

In the following section, the changes and updates being proposed for the Market Basket Measure (MBM) methodologies are explained in more detail. In addition, preliminary 2023-base component costs estimates are compared with the 2018-base component costs.

Shelter component

2018-base Market Basket Measure methodology

In the 2018-base MBM methodology, the shelter component represents the annual shelter cost for the MBM reference family, which includes non-subsidized rent, appliances, basic tenants’ insurance and utilities.^{Note  Note  Note}  According to the CMHC’s NOS, the MBM reference family^Note  should be living in a three-bedroom dwelling not in need of major repairs. Finally, since the MBM methodology represents a modest, basic standard of living, the MBM family’s rent was estimated using median rent for households in the second income decile.

The 2018-base shelter methodology in the provinces used either a quantile regression model or a sampling median^Note  to estimate the median rent per MBM region using the 2016 Census of Population. In MBM regions where there were too few observations to have a high-quality sampling median, a quantile regression model, which borrows strength from a larger sample to improve the estimation process, was used to estimate the cost of a three-bedroom unit.

Once the shelter component costs were estimated for the base year (2018), the costs were updated annually using the provincial all-items Consumer Price Index (CPI).

2018-base Northern Market Basket Measure methodology

The underlying standard used in the province is also used for the territories (i.e., CMHC’s NOS). Again, according to CMHC’s NOS, the reference family for Yukon, the Northwest Territories and Nunavut^Note  requires a three-bedroom dwelling. Similar to the provincial MBM, the cost of the Northern Market Basket Measure’s (MBM-N’s) shelter component includes the rent of a non-subsidized three-bedroom dwelling not in need of major repairs, appliances, basic tenants’ insurance and utilities.^{Note  Note  Note}  The reference family’s rent was again estimated using households in the second income decile.

As for the provinces, the 2018-base MBM-N shelter costs for Yukon and the Northwest Territories used a quantile regression model to estimate the median rent for households in the second income decile of each MBM region.^Note  For the Nunavut MBM-N, shelter costs were estimated using the sampling mean of all rents for households in the first seven deciles of the income distribution. ^Note 

Once the shelter component costs were estimated for the base year (2018), the costs were updated annually using the territorial-level all-items CPI.

Proposed 2023-base Market Basket Measure methodology

In line with the 2018-base methodology, the proposed 2023-base shelter component would continue to follow the CMHC’s NOS for the MBM reference family. For the 2023-base MBM methodologies, the 2021 Census of Population would be used to cost the non-subsidized rent paid^Note  and utilities,^Note  and the component would continue to include appliance and insurance costs.

Similar to the 2018-base methodology, the calculation of shelter costs for the proposed 2023-base would use a combination of either a quantile regression model or a sampling median to estimate rent, depending on the number of observations available for each MBM region. The proposed 2023-base quantile regression model^Note  (used to estimate shelter costs in smaller geographic areas where small samples make it difficult to produce a good quality estimate) will be very similar to the 2018-base model in that it will estimate non-subsidized rent using the number of bedrooms, the household income decile, and the MBM region. However, for the 2023-base, it is being proposed that additional variables be considered for inclusion in the model. Relevant variables were identified based on feedback from stakeholders during the third comprehensive review outreach process. For example, household mobility status and dwelling type were variables that were stressed by stakeholders as being important determinants of rental costs. After a list of potential explanatory variables were created, a variable selection model (i.e., backward selection model) was used to choose the ones that had a large impact on improving the prediction capabilities of the model. The variables selected for the proposed 2023-base model are the following: number of bedrooms, income decile, MBM region, age group of the major income earner, household mobility status,^Note  structural type of dwelling, number of people in the household, and the index of remoteness. The provincial-level all-items CPI was then used to inflate the rent estimates derived from the 2021 Census to 2023 dollars (the use of this index is justified in the next paragraph). A supplement for appliances and an amount for tenant's insurance will continue to be included in the component cost.^Note 

As was done for the 2018-base, an index will be used to annually update the base year’s shelter costs to other years. An analysis was undertaken to find the best suited index to annually update shelter costs between census years. There are at least two important considerations when evaluating the appropriateness of any index. First, the index should not introduce unnecessary year-over-year volatility as this would directly impact poverty rates. Second, it is also important that the chosen index track shelter cost inflation adequately in all MBM regions. The analysis compared multiple indexes including the all-items CPI, the rented accommodation CPI, an index created with data from CMHC and an index created with data from the System of National Accounts (SNA). The average and median relative difference between the 2021 Census shelter costs and the indexed 2021 shelter costs of a three-bedroom dwelling, not in need of major repairs and rented by non-subsidized renter households with a total income in the second income decile were compared using all indexes. The results presented in Table C.1 in the Appendix C indicated that the CMHC rent index had high volatility and, on average, it underestimated the costs of rent. The analysis also found that on average the SNA rent index tended to overestimate rent and that the provincial rental accommodation CPI also exhibited high volatility, over time. Despite having the largest range, the all-items CPI was chosen to perform annual updating of the shelter component, because it produced the smallest average and median percent change between rents, when compared with 2021 Census data. Therefore, as was done for the 2018-base, once shelter costs are estimated for the base year, it is proposed to continue to use the provincial all-items CPI to adjust the shelter costs to other years.^Note 

Proposed 2023-base Northern Market Basket Measure methodology

The proposed 2023-base shelter component for the territories would continue to follow the CMHC’s National Occupancy Standard for their respective MBM-N reference family sizes. It would also be based on the cost of non-subsidized rental units, and include utility costs^Note  and insurance costs.

Given the high response rates to the 2021 Census long-form questionnaire^Note  and the high sampling rate^Note  in Yukon and the Northwest Territories, the proposed 2023-base methodology for estimating rents for these two territories would be to use a sampling mean. In the two capitals, households with a before-tax income in the first five income deciles of their MBM-N region income distribution would participate in the estimation process, while outside the capitals, households in the first seven income deciles would participate in the estimation process. This would bring the estimation process in line with the estimation methods used in Nunavut, which would continue to use a sampling mean for its proposed 2023-base MBM-N methodology.

Once the shelter costs are estimated for the base year, it is proposed that the 2023-base MBM-N shelter costs be annually updated using the territorial all-items CPI.

Market Basket Measure and Northern Market Basket Measure threshold comparison

In Table A.1, the differences between the 2018-base and proposed 2023-base shelter component costs are presented, by MBM and MBM-N region, for a reference family of four (five in Nunavut) in 2023.

What do the Market Basket Measure shelter costs represent?

The shelter component of the Market Basket Measure (MBM) represents the average shelter cost of all non-subsidized paid rents per MBM region (e.g., for families in the second decile, renting a three-bedroom unit, etc.). During the engagement activities for the third comprehensive review, it was frequently mentioned that the shelter costs were too low when compared with current asking rents. While true, it is important to remember that the shelter costs for people in and at-risk of poverty represents all paid rents, not just asking rents. Since established renters (i.e., someone who has been in their dwelling for at least 5 years) typically have lower rents and represent a significant proportion of all renters (e.g., 77.1% of renters did not move within the last year and 36.8% of renters did not move within the last five years), the MBM rent estimates will be lower than the average asking rent. For example, compared with the proposed costs presented in Table A.1, using only the rents of households that moved within the last 5 years to estimate shelter costs, the shelter component would increase by 20.7% in Toronto and 16.8% in Vancouver. However, using only these rents would diverge from the purpose of the MBM because the MBM’s component costs are meant to represent the average shelter costs in Canada. This is important in poverty measurement, since the thresholds are being applied to representative samples of the population. If an average based on asking rents is applied to the representative sample, the number of people in poverty would be overestimated because the average Canadian in the second income decile does not move every year. Or, alternatively, the reference family would need to be refined so that they are assumed to move every predetermined number of years.

Clothing and footwear component

2018-base Market Basket Measure methodology

The 2018-base methodology used the 2012 Acceptable Living Level (ALL) basket developed by Harvest Manitoba^Note  and the Social Planning Council of Winnipeg as the standard for the clothing and footwear component.^Note  At the time of its creation, the ALL basket of items received significant input from people living in lower and modest income situations, which reflected a similar standard of living to that of the MBM reference family. The basket itself contains items for common work, school and social occasions.

For the 2018-base, Statistics Canada’s Consumer Prices Division (CPD) collected in-store prices from outlets in 11 cities across the country.^Note  For that base, one city was used as the proxy for all the MBM regions in its respective province. For example, Montréal was used as a proxy for the other five MBM regions in Quebec. The only exception was Ontario where Toronto prices were used for the Toronto MBM region, while Ottawa prices were used as a proxy for the remaining MBM regions in Ontario. Next, the arithmetic mean of the three lowest price observations per year for three years (2016, 2017 and 2018) was calculated for each item.^Note  A three-year average was taken to reduce price volatility.^Note  The three-year average of clothing prices was then multiplied by the annual quantities from the 2012 ALL replacement schedule (i.e., average lifespan of the clothing item)^Note  and aggregated to obtain the cost of the component.

Once the component cost was estimated for the base year, the provincial-level clothing and footwear CPI was used to adjust the costs to other years.

2018-base Northern Market Basket Measure methodology

The 2018-base MBM-N methodology for the territories used the same standard as the provinces (i.e., the ALL), but adjustments were made to account for the colder climate in the territories. For example, certain items were considered unnecessary or not typically used in the territories while new clothing items were added. In addition, the replacement schedule (i.e., the lifespan of clothing items) for certain items was adjusted to account for the higher use of winter clothes in the territories. Unique to MBM-N for Nunavut, its clothing component also included costs associated with traditional Inuit clothing.

Costing the clothing component for Whitehorse and Yellowknife followed a similar methodology as the provinces (i.e., in-store collection, using an arithmetic average of the three lowest price observations for the year). For Yukon communities located outside the capital, a population weighted average of the costs associated with one shopping trip per year from each community to Whitehorse and online shopping was added to the clothing component costs. For Northwest Territories communities outside the capital but within driving distance of Yellowknife, the methodology was similar to the methodology for Yukon (i.e., a weighted average of a shopping trip and online shopping was added to the component cost), while for fly-in communities in the Northwest Territories, driving costs were replaced with the costs associated with flying to the most popular population centres.^Note  Finally, all communities in Nunavut, including Iqaluit, followed a similar methodology to the fly-in communities in the Northwest Territories.

Once the component cost was estimated for the base year, the territory-level clothing and footwear CPI was used to adjust the costs for other years.^Note 

Proposed 2023-base Market Basket Measure methodology

It is proposed that the 2023-base MBM continue to use the 2012 ALL basket as the standard for the clothing and footwear component. Although it is from 2012, it remains the most recent clothing standard published in Canada.^Note 

Given improvements in the price collection methods developed by the Statistics Canada’s CPD, the 2023-base price collection method would change from purely in-store collection to a combination of in-store and web-scraped collection.^Note  Web-scraping is an automated process through which data are automatically extracted from the web. This improvement in price collection yielded hundreds of thousands of additional price quotes paving the way for more precise estimation. This collection method is also cheaper, more frequent and requires fewer resources.

In addition, the geographical coverage of in-store collection would also be expanded to include all cities covered by CPD in-store collection. More specifically, the previous 11 cities where in-store collection occurred for the 2018-base would be expanded to include an additional 11 cities for the 2023-base.^Note  Again, following CPD’s collection methodology, these in-store collected items would be augmented by using web-scraped data which comes from major retailers with physical locations across Canada. A national pricing strategy is assumed for the web-scraped data (i.e., the prices are consistent across the country and only differ by provincial tax), while the in-store collected prices still have regional variations.

Given the larger number of price observations (i.e., increased sample size) which allows for improved precision and reduces the impacts of seasonality caused by a smaller sample, the use of the three cheapest price observations and three-year average, as was done in the 2018-base, would no longer be necessary. Therefore, the 2023-base pricing methodology would use the first decile of all prices from all sources (i.e., in-store and web-scraped) for the base year (i.e., 2023) after removing outliers and luxury clothing items. As was done for the 2018-base, the final price estimates would be multiplied by the ALL replacement schedule and annual quantity allocations to arrive at the total cost for the clothing component. Given the larger number of price observations obtained by web-scraping, the larger portion of the clothing component cost would consist of one national price (i.e., one total cost for all of Canada). The provincial differences would be caused by the addition of provincial tax and the small portion of in-store collection.^Note 

Following the estimation of the clothing and footwear component in the base year, it is proposed that the 2023-base MBM continue to use the provincial-level clothing and footwear CPI to adjust the cost of the component to other years.

Proposed 2023-base Northern Market Basket Measure methodology

It is proposed that the 2023-base MBM-N continue to use the northern adapted 2012 ALL basket as the standard for the clothing and footwear component, with the addition of traditional clothing items for Nunavut.^Note  Given that the major retailers from which the web-scraped data are sourced have store locations in Yellowknife and Whitehorse, it is proposed that the costing methodology for these two territorial capitals use the same methodology as the one proposed for the provinces. Since web-scraped retailers do not have locations in Iqaluit, it is proposed that the 2018-base MBM-N clothing pricing methodology be used for the 2023-base MBM-N in Nunavut (i.e., in-store, arithmetic average of the three lowest price observations for the year).

For communities outside of the capital, the methodology used in the 2018-base MBM-N would be maintained, but the latest available price data would be used (e.g., flight costs, hotel costs, meal allowances, etc.).

Similar to the provinces, once the cost of the clothing and footwear component is estimated for the base year, it is proposed that the 2023-base MBM-N continue to use the territory-level clothing and footwear CPI to adjust the cost to other years in Yukon and the Northwest Territories and the all-items CPI will be used in Nunavut.

Market Basket Measure and Northern Market Basket Measure threshold comparison

In Table A.2, the differences between the 2018-base and proposed 2023-base clothing and footwear component costs are presented, by MBM and MBM-N region, for a reference family of four (five in Nunavut) in 2023.

The improvements to the price collection for clothing items would result in the proposed 2023-base clothing and footwear component costs being higher than the 2018-base component values for all MBM regions in the provinces and for the majority of MBM-N regions in the territories. Specifically, using in-store price collection for the 2018-base resulted in a limited quantity of price observations for each clothing item which constrained the estimation methodology used in the 2018-base. For the proposed 2023-base methodology, there are more prices to estimate the cost of the clothing and footwear component. Even after accounting for outliers and ensuring the clothing items represent a modest budget by taking items in the 10^th percentile of the price distribution, the average price per clothing item would be higher than the one obtained using the 2018-base methodology.^Note  As with the other components and across the different MBM methodologies (i.e., bases), improvements in data collection, methodological best practices and standards are a hallmark of the MBM (i.e., continually improving). Therefore, as with other updates proposed in this paper, these differences should not be viewed as one base being better than another one, but rather the bases should be viewed as reflecting the data and resources available at a particular point in time.

Food component

2018-base Market Basket Measure methodology

The 2018-base MBM food component is based on the 2019 National Nutritious Food Basket (NNFB), developed by Health Canada. The food prices for the items found in the 2019 NNFB were collected from 38 municipalities over a 12-month period by Statistics Canada’s CPD. The geometric mean of the monthly prices was then used to get the annual average price for a municipality. When an MBM region did not contain any municipalities providing prices, another municipality within the province was used as a proxy. Alternatively, when an MBM region had more than one municipality providing prices, the weighted average of the individual municipality prices was used.^Note  The 2018-base MBM would then use these prices and the NNFB weekly quantities required by the MBM reference family to estimate the food component costs for MBM regions. Finally, following the NNFB methodology, an additional 5% was added to the total component costs for miscellaneous foods and beverages (e.g., coffee, tea, herbs, spices and condiments).

Once the component cost was estimated for the base year, the provincial-level CPI for food purchased from stores was used to adjust the component costs for other years.

2018-base Northern Market Basket Measure methodology

The 2018-base MBM-N food component is also based on the 2019 NNFB. The collection method for the territorial capitals was similar to the one used for the provinces (i.e., the geometric mean of the monthly prices). For communities outside of the territorial capitals, spatial indexes were used to adjust the capital prices.^Note  These community-level adjusted prices, including the 5% addition, and the NNFB weekly quantities were aggregated and weighted by population counts,^Note  to their respective food component cost for their MBM-N regions.

For Nunavut, two large modifications were also made to the 2018-base MBM-N to reflect life in that territory. The first was to increase the quantities found in the NNFB so it would reflect the additional nutritional requirements needed for the fifth reference family member (i.e., a girl aged 4 years). Given the importance of maintaining Inuit culture and traditions and for practical requirements, the second modification to the NNFB was to include hunted and harvested country foods.^Note 

Once the component cost was estimated for the base year, the territorial-level CPI for food purchased from stores was used to adjust the component costs for other years in Yukon and the Northwest Territories, and the territorial all-items CPI for Nunavut.

Proposed 2023-base Market Basket Measure methodology

The proposed 2023-base food component methodology for the MBM would continue to use the same 2019 NNFB, the most recent standard currently available.

Scanner data

Statistics Canada uses mostly point-of-sale scanner data received directly from 21 grocery chains across all regions in Canada. Point-of-sale, or transaction, data are the highest quality price data available, since they track actual prices paid by Canadians at the till, rather than just advertised store prices. These data are collected on a weekly basis and include sales and promotions where applicable. To protect the confidentiality of those who make and sell the goods that are priced, Statistics Canada does not publish specific brand names or store names.

As indicated above, recently, Statistics Canada’s Consumer Prices Division has integrated data from point-of-sale systems (also known as scanner data or transaction data) into its processing environment. For more information on what scanner data is, see the Scanner data text box. Following CPD’s lead, it is proposed that the MBM also estimate food costs using scanner data. Scanner data obtained from Canadian retailers offers a significant quality advantage over traditional in-store collection. Scanner data provide a comprehensive electronic record of the transactions made through a retailer’s point-of-sale system. It includes prices from thousands of stores across the country, covering around a thousand communities and representing millions of weekly food prices and covering the majority of the Canadian grocery market. This would allow the MBM to use actual prices for municipalities in each MBM region, instead of proxies.

Additionally, instead of using the geometric mean of monthly prices, the 2023-base MBM would continue to follow CPD's lead and use the weighted arithmetic average of price observations.^Note  After the average price for a food item is calculated for a municipality, the individual municipality prices would be aggregated as was done in the 2018-base (i.e., using census population counts in a weighted average), but using the 2021 Census population counts. These annual MBM region prices would then be applied to the weekly 2019 NNFB quantities to arrive at the total food component costs. Finally, the additional 5% for miscellaneous items would continue to be applied.

Once the food component cost is estimated for the base year, the provincial-level CPI for food purchased from stores would continue to be used to adjust the component cost for other years.

Proposed 2023-base Northern Market Basket Measure methodology

The proposed 2023-base MBM-N methodology would closely follow the 2018-base MBM-N methodology. In Yellowknife and Whitehorse, transactional point-of-sale data are available. Therefore, it is proposed that these territorial capitals use the same pricing methodology that is being proposed for the provinces. Conversely, the scanner data are not available from any store locations in Iqaluit. Therefore, the proposed 2023-base MBM-N pricing methodology for Nunavut would be the same as the 2018-base pricing methodology but would use monthly prices collected in 2023.

As was done with the 2018-base MBM-N methodology, once the food component cost is estimated for the base year, it is proposed that the territorial-level CPI for food purchased from stores would be used to adjust the component costs for other years in Yukon and the Northwest Territories and the territorial all-items CPI be used in Nunavut.

Market Basket Measure and Northern Market Basket Measure threshold comparison

In Table A.3, the differences between the 2018-base and proposed 2023-base food component costs are presented, by MBM and MBM-N region, for a reference family of four (five in Nunavut) in 2023.

The improvements in the price collection methodology have allowed for better geographical disaggregation. For example, using the 2018-base methodology, in Prince Edward Island, there was only one food component cost for all the MBM regions, while there are different food component costs for all regions using the 2023-base. In general, recalibrating the food component did not have a uniform impact across MBM regions when comparing the two bases (i.e., some went up, some went down). Specifically, the improved sample size brought by the integration of scanner data for the 2023 base allows the production of good quality regional price estimates for each food item included in the NNFB. The proposed 2023-base updates would increase the overall Canada food component average cost by about 4.6%.^Note 

Transportation component

2018-base Market Basket Measure methodology

For the 2018-base MBM, the transportation component used a weighted average of two subcomponent costs: private and public transportation. For areas with a comprehensive public transit system both private and public subcomponent costs are used in the costing methodology. For areas without a comprehensive public transit system (e.g., small urban centres with fewer than 30,000 people and rural areas) only the private subcomponent cost is used.

The private transportation subcomponent was based on a basket of five used compact cars. It included the cost of purchasing, maintaining and operating an eight-year-old, four-door compact car, once every five years with a 36-month financing term. It also included 1 200 litres of gas per year for the four-person reference family. The private subcomponent used data from the CPD, registration data, CARFAX’s Canadian Redbook, and the Survey of Household Spending (SHS). As for the public subcomponent, its cost included the cost of two adult monthly passes and one child pass. Where monthly passes were not available, 40 ticket fares substituted one monthly pass. It also included 12 two-way taxi trips each year.

Once the costs of both subcomponents were estimated for the base year, the provincial-level CPI was used to adjust the costs to other years. For the private subcomponent, the private transportation CPI was used, while for the public subcomponent, the public transportation CPI was used.

Finally, a weighted average was used to combine the two subcomponents at the MBM region-level, based on the proportion of families where at least one family member indicated using a car to get to work on the 2016 Census.

2018-base Northern Market Basket Measure methodology

In the territories, the 2018-base MBM-N transportation component used only costs related to private transportation. Also, considering the road infrastructure in the territories, the transportation component for Yukon and the non-fly-in communities of the Northwest Territories (e.g., communities with road access) used a basket of vehicles that included one small sport utility vehicle (SUV) and one compact car.^Note  A weighted average, based on units sold, is used to determine the total costs for the basket of cars. It included the costs associated with the purchase of an eight-year-old vehicle, along with 1 200 litres of gas per year (or 15 000 kilometres driven per year at 8 litres per 100 kilometres). The value of the vehicle was amortized over five years with a 36-month financing term at a rate corresponding to the published consumer loan rate. Other costs, such as insurance, maintenance, registration, and driver’s licence renewals, were also taken into account.^Note 

For fly-in communities in the Northwest Territories, the transportation cost was based on the costs associated with purchasing a new mid-quality range all-terrain vehicle (ATV) and a snowmobile from Hay River and shipping them by barge to the fly-in community, as well as the costs associated with operating and maintaining them.

A similar methodology to the Northwest Territories’ fly-in communities is applied to the MBM-N regions in Nunavut. Therefore, the transportation component of the Nunavut MBM includes the cost of purchasing, operating and maintaining both a mid-quality ATV and a snowmobile in the territory.^Note 

Once the component cost was estimated for the base year, the territorial-level private transportation CPI was used to adjust the component costs for other years in Yukon and the Northwest Territories, and the all-items CPI was used for Nunavut.

Proposed 2023-base Market Basket Mesure methodology

For the proposed 2023-base, the transportation component would follow the same underlying methodology employed by the 2018-base (i.e., a weighted average cost for private and public modes of transport), but with the following methodological changes:

Average kilometres driven per year

It is proposed for the 2023-base methodology that the number of kilometres driven per year would differ by province. The proposed methodology would use Natural Resources Canada’s (NRCan’s) Comprehensive Energy Use Database (CEUD) data from 2017 to 2021 to obtain the average number of kilometres driven annually per province.^Note  This change in methodology better reflects the differences in commuting patterns across Canada. Based on this methodology, on average, Atlantic provinces would tend to have higher kilometres driven, as well as Manitoba and Saskatchewan, while the number of kilometres would be lower for Quebec, Alberta and British Columbia. For the proposed 2023-base MBM, the number of litres per 100 kilometres would also change from 8 litres to 7.3 litres per 100 kilometres since the cars in the basket have better fuel efficiency. This improvement decreases the average number of litres used annually, from around 1 200 to around 1 100 litres.

Maintenance costs

For the proposed 2023-base methodology, the tune-ups estimation would be simplified since SHS expenditure data would be replaced by CPD price data. In addition to simplifying the calculation, this change improves the quality of the estimates because it uses direct pricing. In determining the CPD-based replacement items, the age of the vehicle and typical maintenance needs would be considered. For example, the average lifespan of brakes is around 60 000 kilometres and given that 15 000 kilometres are driven per year on average, the break replacement costs would be amortized over four years. Similarly, the costs of all-season tires and their installation would be amortized over five years, based on the average lifespan of tires (about three to five years or 75 000 kilometres).^Note  While it is recognized that adjusting for the number of kilometres driven based on the province would impact these amortization periods, it is proposed to not adjust the maintenance costs based on these new kilometres because the added complexity would outweigh the marginal benefit of a slight difference in cost.

Ride-share inclusion

The increased use of ride-sharing services^Note  (e.g., Uber, Téo Taxi, Lyft, etc.) in many urban communities over the past few years has led to growing interest (as mentioned during the consultation process) in including ride-sharing services in the methodology of the transportation component. Thus, it is proposed that both ride-sharing services and taxi services be included in the public transportation subcomponent. Specifically, it is proposed that a weighted average of the two service costs based on the provincial use of each service be used to estimate the cost of a taxi trip.^Note 

For this new methodology, CPD and SHS data would be used to determine the costs of taxi and ride-sharing services in 2023. Statistics Canada’s CPD data would be used to obtain the average costs of a one-way taxi trip and a one-way ride-share trip, using 2023 CPD data, per province while the latest SHS data would be used to create the weights.^Note  The weighted average of both service costs would be used to estimate the cost of a two-way trip, before being annualized (i.e., multiplied by 12) to obtain the annual cost for each province. This proposed methodology has the advantage of using observed prices and better represents commuter patterns.

Apart for these three methodological updates, the other parts of the private and public subcomponents would only be updated to reflect the most recent price data available for the 2023-base.^Note  After the private and public subcomponents are calculated for the base year, the provincial-level private transportation index and public transportation index, respectively, would be used to make adjustments for other reference years. Finally, it is proposed that a weighted average would continue to be used to combine the two subcomponents, based on the proportion of families where at least one family member indicated using a car to get to work on the 2021 Census.^Note 

Proposed 2023-base Northern Market Basket Measure methodology

The proposed 2023-base MBM-N methodology would closely follow the one used by the 2018-base MBM-N.^Note  The only proposed change would be to update to the basket of vehicles for the capitals and communities with road access to the capitals in Yukon and the Northwest Territories to include the average cost for the five most popular SUVs and five compact cars instead of using only one of each type of vehicle. This change would make the basket more resilient to changes in vehicle production and would add stability to the estimates.

As was done for the basket of cars, registration data from 2017 would be used to obtain the most sold SUVs and compact cars for Yukon and the Northwest Territories.^Note  Then, to obtain the 2015 retail price, the Canadian Redbook (CARFAX) would be used. Finally, the weighted average of the baskets would be found by using the registration counts as weights.

Once the component cost is estimated for the base year, the territorial-level private transportation CPI would continue to be used to adjust the component costs for other years in Yukon and the Northwest Territories, and the territorial all-items CPI would be used for Nunavut.

Market Basket Measure and Northern Market Basket Measure threshold comparison

In Table A.4, the differences between the 2018-base and proposed 2023-base transportation component^Note  costs are presented, by MBM and MBM-N region, for a reference family of four (five in Nunavut) in 2023.

Other expenses component

2018-base Market Basket Measure methodology

The MBM has a final component that recognises the needs for other goods and services that are not captured under the shelter, clothing, food and transportation categories (e.g., household items, personal care, reading materials, etc.). The methodology for pricing the other expenses component (formerly known as the other necessities component) does not price individual items because too many items could potentially need to be priced and many of these items would be considered necessities for some families but not for others.

The derivation of the other expenses component used data from Statistics Canada’s SHS. In 2010, the SHS was redesigned, including changes to the questionnaire and the introduction of a two-week diary.^Note  These changes resulted in information on products and services being collected in relation to a reference period (e.g., last week, month or year) which made it easier for respondents to recall their expenditures more accurately. In addition, expenditure categories in the redesigned SHS were also updated to a new custom product classification. However, these changes to the concepts, recall periods and mode of collection raised questions regarding the suitability of the redesigned SHS data for the MBM methodology.

As a result, during the 2018-base MBM comprehensive review, the 2008-base fixed percentage used in the other expenses component was reviewed, and it was held constant at 75.4%.^Note  This fixed percentage is referred to as the “multiplier” and was established based on an analysis of historical spending patterns on a set of selected expenditure categories deemed important for a modest, basic standard of living. Then, the multiplier was applied to each MBM region’s food and clothing component costs to derive a dollar amount for the MBM region’s other expenses component. Finally, an amount for cellular telephone services was also priced separately and added to the other expenses component.

Once the cost of the other expenses component was estimated for the base year, the cost was updated using the annual provincial all-items CPI to make yearly adjustments.

2018-base Northern Market Basket Measure methodology

The 2018-base MBM-N methodology applies the same methodological underpinnings present in the provincial MBM. It applies the same fixed multiplier, but against the food and clothing component costs for each MBM-N region. The methodology for the other expenses component in and outside of the capital is the same.

Once the other expenses component cost was estimated for the base year, the territorial-level all-items CPI was used to adjust the component costs to other years.

Proposed 2023-base Market Basket Measure methodology

The first proposed update to the 2023-base consists of using more recent expenditure data to recalculate the other expenses component multiplier. For this update, it would be assumed that while total household expenditures may change over time, the expenditure categories that families typically spend money on are fairly static (e.g., personal care products, children's toys, admissions to movie theatres, etc.). Therefore, the 2023-base would use the historical household expenditure categories to derive the multiplier value. The only exception would be for expenditure categories that are associated with communication services, which would be removed from the numerator to prevent double counting with the introduction of a new and separate communication services component (described further below).^Note  Based on this updated expenditure list, the most recent SHS data would be used to calculate a new multiplier (50.3%) as was done in previous MBM bases (i.e., the sampling mean of pooled data from the 2017, 2019 and 2021 reference years).^Note 

The second proposed update would be to better align the SHS population used to calculate the multiplier to the MBM reference family. It is proposed to use only SHS families with two adults aged between 25 and 64 years and two children aged between 6 and 16 years^Note  with a before-tax household income in the 20^th to the 50^th percentile in their MBM region.^Note 

These two updates would ensure that the historical standard is preserved (i.e., one multiplier for all the MBM regions), while allowing for the expenditure values to reflect more closely those of the MBM reference family. Using the historical expenditure categories when updating the expenditure values to the most recent SHS data available would allow for consistency between MBM bases and maintain the underlying structure of this component in the MBM basket.

It is proposed that the provincial-level all-items CPI would continue to be used to adjust the 2023-base other expenses component costs for years other than the base year.

Proposed 2023-base Northern Market Basket Measure methodology

The proposed methodology for the 2023-base MBM-N would be the same as the one proposed for the provincial MBM. The updated multiplier (i.e., derived using the new SHS universe, latest SHS data and excluding the communication service expenditure categories) would be applied to the food and clothing cost per MBM-N region to estimate a dollar amount for the other expenses component.^Note 

As was done in the 2018-base, the territorial-level all-items CPI would be used to adjust the 2023-base other expenses component costs to years other than the base year.

Market Basket Measure and Northern Market Basket Measure threshold comparison

Given the redesign of the other expenses component caused by the introduction of the communication services component threshold, comparisons are not possible between the 2018-base and proposed 2023-base. Therefore, Table A.5 only presents the other expenses component costs, by MBM and MBM-N region, for a reference family of four (five in Nunavut) in 2023.

(New) Communication services component

Proposed 2023-base MBM methodology

There was no separate communication services component in the 2018-base methodology. Costs associated with these services were implicitly accounted for in the numerator of the other expenses multiplier. In addition, since the MBM 2018-base uses data from the 2008 and 2009 reference years of the SHS, a period when expenditures on cell phones were relatively low compared with today, an additional amount for cell phone services was added to the 2018-base’s other expenses component.

For the 2023-base, it is proposed that a separate costed communication services component be created and comprise of the following three subcomponents: landline services, cell phone services and Internet access services.

The proposed communication services component aligns with the intent of the MBM to reflect the costs of good and services required for a modest, basic standard of living. The proposal also aligns with the universal service objective established by the Canadian Radio-television and Telecommunications Commission (CRTC) whereby Canadians, in urban areas as well as in rural and remote areas should have access to voice services and broadband Internet access services, on both fixed and mobile wireless networks.^{Note  Note}  To recognize costs associated with landlines, which are not included in CRTC’s basic affordable standard but are still used by a substantial portion of the population, it is proposed that a weighted^Note  cost of landline services be included in the communication services component. Furthermore, given the varying costs between rural and urban communities, a separate communication services component cost would be calculated for rural and urban regions per province. Finally, given that there is no standard available for communications equipment (e.g., cell phones, routers, modems, etc.), it is proposed that costs for communications equipment continue to be included in the other expenses component.

1. Landline services

   While landline usage has been declining over recent years, a large proportion of Canadian households in the lower end of the income distribution still reported having expenditures on landline services and equipment in 2023.^Note 

   It is proposed that the costs for landline services be based on the annual price for basic landline services in each province, including a one-time installation fee.

   The costs of these service plans would be based on the average price for basic wireline telephone services in rural and urban areas of each province in 2018, using CRTC’s 2019 Communications Monitoring Report.^Note 

   It is proposed that the installation fees be based on an average cost collected from a variety of service providers and be amortized over five years.

2. Cell phone services

   Given the composition of the MBM reference family,^Note  it is proposed that the communications component should cost out cell phone plans for the two adult members of the reference family.^Note  The proposed cost for cell phone services would include the annual price for two cell phone plan subscriptions including two subscriber identity module (SIM) cards and their activation fees.

   The proposed method would use the standard established by the CRTC for a bring-your-own-device (BYOD) plan for the cell phone plan subscription cost. The CRTC mandated that, by July 14, 2021, all major cell phone service providers must provide a minimum plan that costs $35 per month that includes three gigabytes of monthly data, unlimited Canada-wide incoming and outgoing calls, and unlimited text messages. All geographies (urban and rural, by province) had a mobile plan that met this minimum of $35. If the average price of the three lowest plans for a geography was less than $35, the average price would be used.^Note 

   The cost of the SIM card and activation fee for the cell phone services was calculated using the fees listed on selected cell phone service provider websites and amortized over five years. ^Note 

3. Internet access services

   The proposed methodology for estimating residential Internet access service costs would also be based on the CRTC standard established in 2016. The total cost for Internet access services would include an amortized one-time installation fee and be based on an Internet plan that has a download speed of at least 50 megabits per second (Mbps), an upload speed of at least 10 Mbps and unlimited data usage.^Note 

   The costs of these plans were estimated using the lowest reported prices for residential broadband Internet access service (download speed of 50 Mbps, upload speed of 10 Mbps, unlimited gigabytes per month) in rural and urban areas for each province in 2022.^Note  To ensure that the costs reflected a modest, basic standard of living, the average price of the lowest three prices per geography (urban and rural, by province) was used, when possible. The one-time internet installation fees were estimated by taking the average posted installation fee from major internet service providers^Note  and were amortized over five years.

It is proposed that the provincial-level communication CPI be used to adjust the 2023 communication services component costs for other years.

Proposed 2023-base Northern Market Basket Measure methodology

The 2023-base MBM methodology would be similar to the one proposed for the provinces, specifically, the same three subcomponents (i.e., landline services, cell phone services and Internet access services). As well, it would be estimated and aggregated in a similar manner but using prices in the territories.^Note 

It is proposed that the territorial-level communication CPI be used to adjust the 2023 communication services component costs for other years in Yukon and the Northwest Territories while the territory-level all-items CPI would be used in Nunavut.

Market Basket Measure and Northern Market Basket Measure threshold comparison

Although no historical comparison is possible, Table A.6 presents the proposed communication services component costs for urban and rural areas by province and territory.

Using the 2023-base proposed methodology, rural regions in the province would have slightly higher communication service component costs compared with the urban regions within their respective province.

Inuusiqattiarniq component (Nunavut only)

2018-base Northern Market Basket Measure methodology

The inuusiqattiarniq component, unique to the Nunavut MBM-N, is based on the Nunavut MBM-N reference family and only included costs incurred by residents of Nunavut. Based on the 2019 Nunavut Roundtable for Poverty Reduction and follow-up outreach activities, it was proposed that the inuusiqattiarniq component be informed by the following four pillars: (1) Culture and language, (2) Healthy personal relationships, (3) Trauma services and (4) Addiction treatment. Since the majority of costs, associated with programs and activities that support these pillars, are covered through grants or other territorial or federal funding, the majority of costs implicated for the MBM are in relation to the culture and language pillar. Furthermore, despite the coverage of costs, it was acknowledged that families in Nunavut still had out-of-pocket expenses when engaging in these activities. For example, not all cultural and language expenses were covered by a funding model, so an additional amount needed to be added.

Since enumerating and costing individual items deemed necessary to maintain knowledge, language and cultural activities is highly impractical, the inuusiqattiarniq component pricing methodology was determined based on expert judgement. A fixed multiplier (25%), serving to approximate the overall cost for possible items that could potentially support the above-mentioned activities, was applied to the cost of the other expenses component to estimate thresholds for the inuusiqattiarniq component.

The inuusiqattiarniq component cost was applied only to Indigenous families when determining their poverty status.^Note  For the determination of poverty status among non-Indigenous families, the inuusiqattiarniq component was removed.

After costing the inuusiqattiarniq component for the base year, the territory-level all-items CPI for Nunavut was used to adjust the component cost for other years.

Proposed 2023-base Northern Market Basket Measure methodology

For the 2023-base MBM-N, in the absence of new data sources that can inform the component’s costing approach, it is proposed that the inuusiqattiarniq component follow the 2018-base MBM-N methodology. Therefore, the fixed multiplier used in the previous base would continue to be used for the 2023-base.

Similar to the 2018-base, after the cost of the inuusiqattiarniq component is estimated for the base year, the territory-level all-items CPI for Nunavut would continue be used to adjust the component cost for other years.

Northern Market Basket Measure threshold comparison

In Table A.7, the differences between the 2018-base and proposed 2023-base inuusiqattiarniq component costs are presented, for the Nunavut reference family of five in 2023.

Proposed changes to the Market Basket Measure disposable income

For a family or individual to be in poverty, according to the MBM their disposable income must be less than the basket threshold amount for their family size and region.

MBM disposable income will continue to be defined as total income (including government transfers) after deducting income tax, Canada Pension Plan and Quebec Pension Plan contributions, Employment Insurance Plan and Registered Pension Plan contributions, union dues, child care expenses, spousal support payments paid, public health insurance premiums, and direct medical expenses, including private insurance premiums. In addition, adjustments for different tenure types and for taxes paid on capital gains would continue to be made.

Tenure type adjustments

Beginning with the 2018-base methodology, tenure type adjustments (TTAs) for housing tenure types were created for homeownership with or without mortgages and for those residing in subsidized dwellings, again following the National Occupancy Standard (NOS) standard set out by the Canada Mortgage and Housing Corporation (CMHC) for the reference family size and composition. This was done to place families in all tenure types on a more equal footing to those families in non-subsidized rental units (the tenure type presented in the shelter component), when determining their poverty status.

The TTA represents the difference between the shelter costs of non-subsidized renters (e.g., cost of the shelter component) and the shelter cost of a particular tenure type (e.g., subsidized renters, homeowners with a mortgage and homeowners without a mortgage). If the difference is less than $0, the TTA is set to $0.

Equivalent TTA values for other family sizes are computed using the square root method. These TTAs are then added to disposable income when evaluating poverty status.

Summary of the process

2018-base tenure type adjustment methodology for the Market Basket Measure

The calculation of individual tenure type shelter costs in the provinces was done using the median cost for a three-bedroom dwelling for households in the second income decile.^Note  The methodology used either a quantile regression model or a sampling median to estimate the median cost per MBM region. In MBM regions where there were too few observations to have a high-quality sampling estimate, a quantile regression model was used to estimate the cost of a three-bedroom unit. Finally, once shelter costs were estimated for the different tenure types, the TTAs were calculated.

Once the TTAs were estimated for the base year (2018), the costs were updated annually using the provincial-level all-items CPI.

2018-base tenure type adjustment methodology for the Northern Market Basket Measure

The calculation of individual TTAs in the territories involved different methodologies to cost shelter based on the territory. In Yukon and the Northwest Territories, a quantile regression model was used to estimate the median shelter cost per tenure for three-bedroom dwellings per MBM-N region and households in the second income decile. The TTAs were then derived by subtracting the shelter cost of a particular tenure from the non-subsidized renter threshold (shelter component cost). Once the TTAs were estimated for the base year (2018), the TTAs were updated annually using the provincial all-items CPI. In Nunavut, shelter costs per tenure were estimated using a sampling mean using households with a before-tax household income in the first seven income deciles of the income distribution of their region.

Proposed 2023-base tenure type adjustment methodologies

As was done with the 2018-base MBM and MBM-N methodologies, it is proposed that both the MBM and MBM-N 2023-base methodologies continue to use the TTAs to account for differing shelter costs based on the tenure type.^Note 

Proposed 2023-base provincial tenure type adjustment methodologies

For all tenure types, the median shelter cost will continue to be estimated per MBM region for households in the second income decile living in a three-bedroom dwelling. As was done for the shelter component on the basket side, the calculation of tenure type thresholds for the 2023-base would use a combination of a quantile regression model and a sampling median to estimate the monthly payment amount depending on the number of observations available for each MBM region. The 2023-base quantile regression model^Note  (used to estimate shelter costs in smaller geographic areas where small samples make it difficult to produce a good quality estimate) would be very similar to the 2018-base model in that it would estimate the monthly amount using the number of bedrooms, the household income decile, and the MBM region. However, additional variables were added to the 2023-base model to improve its prediction capabilities.^Note 

Homeowners with mortgages

For owners with a mortgage, the threshold would still consist of the sum of utilities (e.g., electricity, fuel, water, and other municipal services), municipal taxes, condo fees (if applicable), appliances, insurance, and annual interest costs.

For the 2023-base, it is proposed that a new estimation methodology be applied when calculating the annual interest amount. The new methodology allows for the calculation of interest costs to account for the number of bedrooms and allows for an estimation of interest costs per MBM region.^Note  The annual interest cost would be calculated per MBM region and per age group using the 2023 Canadian Housing Survey (CHS) and the Census of Population. First, the interest portion of the monthly mortgage amount would be estimated per age group and region using the 2023 CHS and a composite mortgage interest rate.^Note  Because of sample size constraints, the provinces were divided into 12 geographic regions^Note  and three age groups^Note  for a total of 36 domains. The CHS proportions were then applied to the Census mortgage payment amount to obtain an estimate of the annual amount of interest paid on the mortgage per MBM region and age group.

For reference years other than the base year, the non-interest portion of the TTA would be adjusted using the provincial-level all-items CPIs, while the interest portion would be re-calculated using an interest rate based on all 2023 existing loans in Canada as is done by Statistics Canada’s CPD when deriving the Mortgage Interest Cost Index (MICI).

Homeowners without mortgages

For owners without a mortgage, the threshold would still consist of the sum of utilities (e.g., electricity, fuel, water and other municipal services), municipal taxes, condo fees (if applicable), appliances and insurance.

After calculating the threshold for the base year, the provincial-level all-items CPIs would be used to adjust them to other reference years.

Subsidized renters

The subsidized renter threshold would still be comprised of the cost of rent, utilities (e.g., electricity, fuel, water and other municipal services), appliances and insurance.

Adjustments to other reference years would be based on the provincial-level all-items CPIs.

Proposed 2023-base territorial tenure type adjustment methodologies

For all tenure types, it is proposed to follow the MBM-N shelter component methodology for the territorial TTAs, i.e., to use the sampling mean to cost a three-bedroom dwelling per tenure. For subsidized renters, all households with a before-tax household income in the first five income deciles of the income distribution of their region would be used, while households with an income within the first seven income deciles would be used for homeowners.^Note  The use of the sampling mean is proposed for two reasons. The first reason is that all households outside the capitals completed the 2021 Census long-form questionnaire (e.g., no sampling), which improved the quality of the sample estimates. The second reason is that the response rates to the census long-form questionnaire in the territories were high.^Note 

Homeowners with mortgages

For homeowners with a mortgage the cost would continue to include utilities (e.g., electricity, fuel, water and other municipal services), municipal taxes, condo fees (if applicable), appliances, insurance and annual interest costs.

As for the 2023-base TTA for homeowners in the provinces, it is proposed to use the same new estimation methodology to calculate the annual interest amount paid on the mortgage, but using two age groups instead of three because of the smaller sample sizes. The two age groups are created using the age of the major income earner. The first group would include households where the major income earner is younger than 50 years old, and the second group would include households where the major income earner is 50 years old or older.

For reference years other than the base year, the non-interest portion of the TTA would be adjusted using the territorial-level all-items CPIs, while the interest portion would be re-calculated using an interest rate based on all 2023 existing loans in Canada as is done by Statistics Canada’s CPD when deriving the Mortgage Interest Cost Index (MICI).

Homeowners without mortgages

For homeowners without a mortgage, the cost would continue to include utilities (e.g., electricity, fuel, water and other municipal services), municipal taxes, condo fees (if applicable), appliances and insurance.

Once the TTAs are calculated for the base year, it is proposed that they be annually updated using the territorial-level all-items CPI.

Subsidized renters

The subsidized renter cost would continue to include the paid rent, utilities (e.g., electricity, fuel, water, and other municipal services), and insurance.

Once the TTAs are calculated for the base year, it is proposed that they be annually updated using the territorial-level all-items CPI.

Market Basket Measure and Northern Market Basket Measure tenure type adjustment comparison

Tables A.8.1 and A.8.2 present the 2018-base and proposed 2023-base TTA amounts by MBM and MBM-N region.

Medical expense imputation

Medical expenses are one of the non-discretionary expenses that are deducted from after-tax income to derive disposable income for the MBM. These medical expenses include out-of-pocket expenses, and health care and dental plan insurance premiums.^Note 

The MBM methodology first uses medical expenses amounts declared on tax forms,^Note  then if a respondent aged 15 years or older^Note  is not linked to tax data or if their medical claim for tax purposes is zero,^Note  an imputation value is used.^Note  This section explains how the imputation amount is calculated.

2018-base Market Basket Measure and Northern Market Basket Measure methodologies

For the 2018-base MBM and MBM-N methodologies, the medical imputation value for the base year was derived using 2014-to-2017 pooled SHS data and adjusted for other reference years using the provincial-level health care consumer price CPI.^Note  The 2018-base amount corresponds to the provincial or territorial sampling mean of medical expenses among people in the second income decile not claiming an amount on their tax return.

Proposed 2023-base Market Basket Measure and Northern Market Basket Measure methodologies

During outreach activities, participants mentioned that medical expenses should differ by age (e.g., seniors compared with young people). Furthermore, internal research demonstrated that medical expenses differ by age group in the provinces. Therefore, for the MBM, it is proposed that new imputation values for the 2023-base take into consideration the age of the individual not claiming medical expenses. For the MBM-N, it is not possible to calculate imputation amounts per age group because of data restrictions.

The proposed 2023-base imputation amounts would also be updated to use the most recent SHS data available and would be adjusted using the same price inflator as in the 2018-base. In summary, the new imputation amount would differ from the 2018-base in that it would correspond to the provincial average of medical expenses by age group^Note  among people in the second income decile who are not claiming an amount on their tax return, and by age group.

Similar to the 2018-base, the proposed 2023-base medical expenses imputation amount for people who did not claim a medical expense on their tax form would be estimated using a two-step process. First, the average per province and per age group medical expenditures^Note  would be calculated using the 2017, 2019 and 2021 SHS data. Next, using the 2019-to-2023 Canadian Income Survey data, the average expenditures for medical expenses per person and age group who do not claim a medical expense is estimated.

Market Basket Measure and Northern Market Basket Measure medical expense imputation amount

Given the changes in the imputation methodology, direct comparisons between the 2018-base and proposed 2023-base imputation amounts are difficult. In Tables A.9.1 and A.9.2, the proposed 2023-base imputation amounts are presented by age groups and different geographies. In addition, the average proposed 2023-base imputation amounts are presented for comparison to the 2018-base imputation amounts.

Appendix B

Appendix C: Summary of analysis on annual shelter component adjustor

As part of the 2023-base Market Basket Measure (MBM) review for the shelter component, an analysis of alternative annual adjustors was conducted. Four alternative adjustors were evaluated based on the following data sources:

1. Canada Mortgage and Housing Corporation: Provincial index created using data from the Canada Mortgage and Housing Corporation (CMHC) Rental Market Survey (RMS). Data for the primary rent market was obtained from the CHMC for 2016 to 2021. A provincial annual index for year t was created by dividing the median for three-bedroom dwelling for year t by the median rent for year t-1. The median rent corresponds to a weighted average of the median rents for census agglomerations (CAs) and census metropolitan areas (CMAs) per province where the weight corresponds to the number of three-bedroom rentals in the primary rental market that exists in each universe.
2. System of National Account: An index was created by first calculating the geometric mean of provincial monthly costs obtained through the National Accounts Report Master file to obtain annual costs, and then dividing the cost of year t by year t-1.
3. Consumer Prices Index, rented accommodation: Consumer Price Index, annual average, not seasonally adjusted. [Table: 18-10-0005-01 (formerly CANSIM 326-0021) for rented accommodation].
4. Consumer Prices Index, all-items: Consumer Price Index, annual average, not seasonally adjusted [Table: 18-10-0005-01 (formerly CANSIM 326-0021) for all-Items].

After the annual adjustors were constructed, they were applied to the 2016 Census shelter component costs per MBM region and inflated to 2021 dollars. Next, the inflated shelter costs were compared with the 2021 Census shelter component costs. Table C.1 presents the results of this analysis. A negative difference indicates that the index overestimated the shelter cost, while a positive difference indicates that the index underestimated the cost. Comparing all the indexes, the CMHC rent index has shown high volatility over the years, and underestimates rents on average. The SNA rent index tended to overestimate rents. The provincial-level rented accommodation CPI also exhibits higher volatility, throughout the years, while the provincial-level all-items CPI produces the smallest average and median differences when compared with 2021 Census data. However, it is also accompanied by greater variation in the distribution of 2021 adjusted costs.

Based on this analysis, it is proposed to continue to use the provincial-level all-items CPI for the 2023-base MBM shelter methodology. On average, the all-items CPI tends to estimate the shelter costs more accurately than the other indexes and is also less volatile annually. Lower annual variability, which may result more from sampling and non-sampling error than true changes in prices over time, is as important as the comparison between the inflated 2016 Census values to 2021 Census values, because volatility could cause interpretability issues when analyzing annual poverty rate changes. For example, large volatility could be interpreted as changes in the housing cost distribution, but, the changes may be nothing more than statistical noise.

Appendix D

Appendix E

Appendix F

References

Benjamin, Wesley, Chanel Christophe, Nancy Devin, Sarah Maude Dion, Éric Dugas and Burton Gustajtis (2022), “Market Basket Measure Research Paper: Poverty Index”, Catalogue no. 75F0002M.

Byers, Jing, Chanel Christophe, Sarah McDermott and Natalie Simeu (2023), “Market basket measure research paper: Child care expenses”, Catalogue no. 75F0002M.

Christophe, Chanel, Nancy Devin, Burton Gustajtis, Benjamin Kow, José Mendoza Rodríguez, Lucie Raymond-Brousseau and Madeleine Steinmetz-Wood (2024), “What we heard report: Engagement activities for the third comprehensive review of the Market Basket Measure”, Catalogue no. 75F0002M.

Danieles, Prince Kevin, Andrew Heisz and Keith Lam (2024), “Market Basket Measure research paper: An analysis of the equivalization method”, Catalogue no. 75F0002M.

Devin, Nancy (2024), “Market Basket Measure research paper: Applying the Market Basket Measure methodology to an administrative data source”, Catalogue no. 75F0002M.

Devin, Nancy and Madeleine Steinmetz-Wood (2024), “Market Basket Measure research paper: Examining shelter and transportation costs within census metropolitan area Market Basket Measure regions”, Catalogue no. 75F0002M.

Devin, Nancy, Burton Gustajtis, Mackin Liu, José Mendoza Rodríguez, William MacMinn and Myron Wasylko (2023), “Market Basket Measure Research Paper: Options for updating the other necessities component and the creation of a communication services component”, Catalogue no. 75F0002M.

Devin, Nancy, Eric Dugas, Burton Gustajtis, Sarah McDermott and José Mendoza Rodríguez (2023). “Launch of the Third Comprehensive Review of the Market Basket Measure”, Catalogue no. 75F0002M.

Devin, Nancy, Burton Gustajtis and Sarah McDermott (2023), “Construction of a Northern Market Basket Measure of poverty for Nunavut”, Catalogue no. 75F0002M.

Devin, Nancy, Burton Gustajtis and Sarah McDermott (2022), “Technical paper for the Northern Market Basket Measure of poverty for Yukon and the Northwest Territories”, Catalogue no. 75F0002M.

Devin, Nancy and Raphaël Langevin (2022), “Market Basket Measure Research: Additional income inequality indicators using the Market Basket Measure”, Catalogue no. 75F0002M.

Devin, Nancy, Burton Gustajtis, Keith Lam and Sarah McDermott (2021), “Construction of a Northern Market Basket Measure of poverty for Yukon and the Northwest Territories”, Catalogue no. 75F0002M.

Djidel, Samir, Burton Gustajtis, Andrew Heisz, Keith Lam and Sarah McDermott (2019), “Towards an update of the Market Basket”, Catalogue no. 75F0002M.

Djidel, Samir, Burton Gustajtis, Andrew Heisz, Keith Lam and Sarah McDermott (2019), “Defining disposable income in the Market Basket Measure”, Catalogue no. 75F0002M.

Djidel, Samir, Burton Gustajtis, Andrew Heisz, Keith Lam, Isabelle Marchand and Sarah McDermott (2020), “Report on the second comprehensive review of the Market Basket Measure”, /Catalogue no. 75F0002M.

Employment and Social Development Canada (2018), “Opportunity for All – Canada’s First Poverty Reduction Strategy”, Catalogue no. SSD-212-08-18E.

Gustajtis, Burton and Andrew Heisz (2022), “Market Basket Measure Technical Paper: The other necessities component”, Catalogue no. 75F0002M.

Heisz, Andrew (2019), “An update on the Market Basket Measure comprehensive review”, Catalogue no. 75F0002M.

McDermott, Sarah and Madeleine Steinmetz-Wood (2023), “Market Basket Measure research paper: Market Basket Measure thresholds for remote regions”, Catalogue no. 75F0002M.

Steinmetz-Wood, Madeleine (2023),“Market Basket Measure Research Paper: Delineating Remote Regions for the Market Basket Measure”, Catalogue no. 75F0002M.