Canada’s draft National Strategy Respecting Environmental Racism and Environmental Justice 2026-2031

Working together to advance environmental equity

Introduction

Royal Assent of the National Strategy Respecting Environmental Racism and Environmental Justice Act (the Act) marked an important national commitment to develop a Strategy to assess, prevent and address environmental racism.

As Canada works to build a stronger, more resilient economy and tackle climate change, we must ensure that no community is left behind. We must build Canada strong- for all. While the federal government has responsibilities within this space, environmental justice (EJ) is rooted in community experience, often led at the local and grassroots level. Meaningful participation, community level knowledge, and engagement with those affected must inform this national strategy. This strategy provides a starting point to consider the roles of the Government of Canada within the context of EJ. At the same time, it recognizes and respects the jurisdiction, including provinces, territories and municipalities, and roles and leadership of other key partners in this space. Success will require collaboration, shared responsibility, and a commitment to building solutions together.

Everyone in Canada should have the opportunity to enjoy a healthy, safe and sustainable environment. The benefits of a growing clean economy, and strong environmental protection should be shared by all Canadians, regardless of who they are or where they live. By acknowledging and addressing environmental racism, we can build a stronger, more united Canada—one that protects people and the environment and moves forward together. This includes pursuing the necessary conditions to advance EJ:

• equitable decision‑making
• fair access to information and participation
• deliberate efforts to avoid disproportionate environmental impacts on marginalized and under-represented groups

Abbreviations

CEPA – Canadian Environmental Protection Act

CIRNAC – Crown-Indigenous Relations and Northern Affairs Canada

DA – dissemination area

ECCC – Environment and Climate Change Canada

EJ – environmental justice

EEPCI – Environmental Effects and Population Composite Index

FCSAP – Federal Contaminated Sites Action Plan

ISC – Indigenous Services Canada

LSES – low socioeconomic status

NPRI – National Pollutant Release Inventory

PEH – potential environmental hazard

UN Declaration – United Nations Declaration on the Rights of Indigenous Peoples

What the Strategy is

Canada’s National Strategy Respecting Environmental Racism and Environmental Justice (2026-2031), or “the Strategy”, will guide the federal government and partners in advancing EJ and addressing environmental racism across the country. As Canada works to build a stronger economy, and create healthier more resilient communities, the Strategy aims to support more equitable benefits for affected communities. The Strategy is informed by analysis of the links between race, socio-economic status, and environmental risks, and considers measures to advance EJ.

The Strategy recognizes the work already underway across the country. It reflects the contributions of individuals and communities who have been advancing EJ for years. Their insights, advocacy, and lived experience guide and strengthen the federal approach.

As reflected in the Act, the Strategy focuses on situations where environmental hazards intersect with systemic inequities. It concentrates on areas where the federal government has the authorities, leadership, responsibilities or capacity to advance change, including:

• guiding federal institutions in embedding principles of EJ, equity, and intersectionality into their policies, programs, regulatory activities, and decision-making processes
• seeking to strengthen collaboration and coordination across federal departments and agencies, other orders of government, industry, civil society, academia, and communities
• enhancing participation and representation of impacted communities, particularly Indigenous, racialized, and marginalized groups, in federal environmental decision-making processes
• improving federal capacity, tools, and knowledge systems to assess, anticipate and prevent environmental disparities, while supporting ongoing learning and adaptive approaches

The Strategy recognizes that broader social, historical and economic factors shape environmental racism and unequal environmental outcomes. These issues go beyond the responsibility of any one government or sector — it is a complex whole-of-society challenge. For this reason, the Strategy focuses on actions the federal government can take to best promote collaboration and complement shared responsibility across jurisdictions, sectors and partners.

How we listened and used evidence to shape the Strategy

This Strategy was informed by input from engagement across a variety of activities and regions. The feedback received provided rich insights and valuable considerations pertaining to the requirements outlined in the Act.

Canada’s Environmental Justice and Racism Symposium

Environment and Climate Change Canada (ECCC) hosted a Symposium on September 13-14, 2024.  It was the first national event of its kind in Canada and launched a national conversation on EJ and environmental racism. Over 300 participants − youth, people with lived experience of environmental racism, academics, advocates, and policy experts − joined the conversation in person and online. The Perspectives and Insights Report summarizes discussions during the two-day event. It highlights key themes, ideas, and resources from participants.

Engagement with Indigenous Peoples

Indigenous peoples in Canada have a special constitutional relationship with the Crown. Section 35 of the Constitution Act, 1982 recognizes and affirms this relationship, including existing Aboriginal and treaty rights. Indigenous Peoples also have a unique relationship with the natural environment, often intersecting with section 35 rights and rights outlined in the UN Declaration. As such, Indigenous perspectives on EJ may be considered in a separate context than other Canadians.

The department has emphasized engagement specific to contexts of Indigenous Peoples. These engagements highlighted challenges in awareness of federal initiatives and identified common findings:

• strong emphasis on distinctions-based approaches reflecting differences across Indigenous groups, regions, and circumstances, urban and those not represented within a single distinction.
• environmental racism is often understood through lived experience, including cumulative impacts over time and across jurisdictions, and is linked to broader social and historical factors.
• there is emphasis on respecting Indigenous rights, including earlier and clearer application of rights-based approaches.
• understanding specific impacts based on context and history requires direct engagement with affected Indigenous nations and groups.
• participants raised concerns about the effectiveness of consultation approaches versus strengthening Indigenous capacity to participate, and the need to enhance accessibility and simplify mechanisms.
• participants highlighted a need for clarity on constitutional responsibilities between levels of government and Indigenous governance. In the context of seeking greater assertion of Indigenous governments’ authorities (particularly for First Nations lands and territories), there is a general desire for clarity around:
  • the intersections of constitutional responsibilities and interactions with Indigenous rights
  • the increasing role of Indigenous-led governance.
• concerns were raised about data gaps, accuracy, and the importance of Indigenous data sovereignty and knowledge systems.

Forward considerations

This Strategy represents a starting point to continued dialogue and relationship building in advancing environmental justice. Ongoing efforts to strengthen engagement and consultation practices will be important, this is particularly the case where decisions intersect with section 35 and UN Declaration rights. While the importance of increasing communities’ internal capacity is understood, this challenge is broader than the scope of this Strategy.  Whole-of-government efforts, including those led by Crown-Indigenous Relations and Northern Affairs Canada (CIRNAC) and Indigenous Services Canada (ISC) are advancing improvements to consultation practices.

Gaps in available data are a recurring issue in local efforts to address environmental racism. Finding practical ways to improve data, while respecting Indigenous data sovereignty, remains a key challenge. These issues will continue to be important areas of discussion as Indigenous rights and perspectives are increasingly reflected in environmental decision-making.

Part 1: Understanding environmental racism and justice

Environmental racism

Environmental racism is a form of systemic racism that occurs when environmental policies and practices harm certain groups disproportionately based on their race. It has been linked to the disproportionate placement of environmentally hazardous sites (e.g., polluting industries) in or near Indigenous, racialized or marginalized communities. This is shaped by a history of exclusion or underrepresentation in environmental decision-making.

The impacts of environmental racism can include socio-economic, health, general well-being, culture and more.  Because environmental racism is systemic, EJ considers how overlapping forms of oppression compound harm. It focuses on how environmental benefits and burdens are distributed, how inequalities link to broader social and economic systems, and on centring community voices in decision-making.

Environmental justice

EJ promotes fair and meaningful inclusion in environmental decision-making and equitable distribution of benefits and burdens. EJ has long been advanced by grassroots organizations.  EJ is commonly understood through four principles: procedural, recognitional, distributive, and restorative justice.

Procedural justice

Procedural justice means including affected people in decisions that impact their lives and ensuring meaningful opportunities to participate. It also requires removing barriers to participation so decisions reflect lived experience and community needs. This includes elevating the voices of Indigenous, racialized, and marginalized communities and improving access to clear, reliable, and timely information.

Recognitional justice

Recognitional justice means respecting the lived experiences and values of affected communities. It acknowledges that many marginalized communities have long cared for the land while being excluded from decision-making. This aligns with Indigenous worldviews that recognize the inherent value of ecosystems. Recognitional justice seeks to improve how community knowledge is treated as valid evidence and challenges narrow definitions of expertise.

Distributive justice

Distributive justice focuses on the equitable sharing of environmental benefits and burdens. The goal is that no group bears disproportionate environmental risks based on race, income, or other characteristics, and for affected communities to enjoy a healthy and safe environment.

Restorative justice

Restorative justice means acknowledging past harms and focusing on those who experienced them. It prioritizes inclusion of affected people in creating solutions. In the context of EJ, this includes taking responsibility for environmental harms and risks. It also involves rebuilding trust and strengthening relationships with affected communities.

Part 2: Examining the link between race, socio-economic status and environmental risk

As a part of the National Strategy Respecting Environmental Racism and Environmental Justice Act (the Act) the Strategy must include a study that includes:

• an examination of the link between race, socio-economic status and environmental risk
• information and statistics relating to the location of environmental hazards

This study aims to provide an understanding of broad relationships between the location of environmental hazards, and particular community characteristics. To meet the specifications of the Act, it needed to be clarified whether the analysis should focus on:

• specific events
• routine exposure to environmental hazards or
• measured exposure to particular chemicals.

This in turn, had an impact on how to understand the concepts of “environmental risk” and “environmental hazard” in the context of the study.

The analytical approach taken in the study prioritized the use of simple terms. They are broad enough to apply across Canada, and its diversity of peoples, industries, and environments. The following interpretations of key terms in the Act were applied to support analysis:

• race: populations within the Census of Canada who self-identified as visible minorities (“racialized”) or as Indigenous peoples
• socioeconomic status: indexed value comprising information about income, education achievement levels, and unemployment rates that are often used to characterize lower socioeconomic status, regardless of race and Indigeneity, as found in the Census of Canada
• environmental risk/hazards: emphasis was placed on potential sources of industrial pollution, examined as sites across Canada rather than at levels of specific compounds or outputs. The hazards considered include sites on the National Pollutant Release Inventory, active or suspect federal contaminated sites, and federally-regulated wastewater effluent sources

A key consideration within the Act is the importance of understanding how different population groups are located in relation to environmental hazards. This was raised by participants throughout the legislative and engagement processes in the context of mapping. Therefore, analysis includes a map bringing together the key information described above. It offers a national picture of where potential environmental inequities may exist.  Communities will be able use the same baseline set of information to understand and compare circumstances.

Development of the Act also highlighted requests for information and statistics to understand possible patterns of communities being disproportionately affected by environmental hazards.  This was pursued using the same data resources behind the maps. This will help build a clearer national picture of how different environmental and social factors relate to one another across Canada. It will provide insights that can inform federal decision-making.

Based on these guiding considerations, ECCC developed an approach to examine how race, socioeconomic status, and environmental hazards overlap across Canada. This approach explores key questions such as:

• If there is a pattern between more racialized, more Indigenous or lower socioeconomic status communities and potential environmental hazards
• if racialized, Indigenous, or low socioeconomic status communities are more likely to be located near potential environmental hazards

The study used standard statistical methods to develop an approach that can be applied to different types of information. This helps ensure that similar analyses can be carried out consistently across Canada.

Results and lessons learned

The study’s findings are summarized below, focusing on key insights and lessons learned in support of developing a national Strategy. Additional methodological information and analytical results are available in Annex A.

Identifying patterns and disproportionality

Given that there is no single, universally acknowledged tool or approach to measure environmental racism or injustice, a variety of approaches were used. This provides more depth in how results are understood.

The analysis identifies linkages between certain population groups and environmental hazards, showing that some communities are more likely than others to be near certain environmental hazards. These patterns vary depending on population groups, geographic scale, and methods used. In particular, results vary according to the assumed spatial relationship. These understandings of “where” the relationships occur are important in being able to compare scenarios and outcomes.

Additional analyses examined disproportional co-location of hazards and populations to determine whether observed patterns differed from expected outcomes, reinforcing that findings are sensitive to both scale and methodology.

At a national level, results reveal patterns, not causes, and should be interpreted with recognition of what aggregate data do not capture:  local histories, contexts, and experience. Combining community experiences of environmental racism across Canada with national-scale data can better inform best practices in addressing EJ. Greater emphasis on localized and community-specific patterns allows a more accurate depiction of environmental racism.

Where potential environmental inequity is highest

Using the same data outlined above, ECCC developed an index to combine information on populations of concern and potential environmental hazards. The “Environmental Effects and Population Composite Index” (EEPCI) presents data in a single mappable measure, illustrating broader associations and enabling comparison across the country.

The initial method (Annex A, Figure 3) used Census Dissemination Areas (DAs) to group data. However, variation in DA size made visual comparison difficult and limited visibility at the national scale.

As an alternative, the data were transferred into a grid of equal-sized (500 square kilometre) hexagons (Annex A, Figure 4), improving consistency in mapping. This approach supports analysis of regional patterns rather than specific communities. The resulting map highlights areas where EJ considerations may be more pronounced, particularly regions with higher industrial activity and population density.

However, neither map should be used alone to determine which communities experience environmental racism or where inequity is most pronounced.  Results are influenced by underlying patterns and analytical assumptions.

Lessons learned and additional considerations

Environmental racism has various interpretations and is deeply tied to places, experiences, and perspectives. While statistics can support understanding, lived experience remains central to reflect the context, history and perspectives in affected communities. Similarly, weak statistical relationships do not necessarily indicate an absence of issues.

Since environmental racism and EJ have various interpretations, measuring these concepts can lead to:

• Results that can vary by method used, debate around which methods to use, and the need to apply the most reasonable and practical method for each question.
• National-scale results that reveal patterns, but cannot capture the full complexity of environmental racism, and shift with data, assumptions, and geography
• Proximity-based analysis without defined parameters and context that risks misrepresenting the communities it aims to describe
• The unit of measurement shaping results significantly. Larger units offer consistency but miss local experience, while smaller units capture specificity but complicate visualization; both scales are needed to better capture community experiences of environmental racism and EJ.

This highlights the importance of transparency in analytical methods and their limitations to support appropriate interpretation. Transparency supports credibility and relevance. This analysis can be considered as one of many inputs to inform federal decisions at a national scale and how community perspectives are understood.

When the Act was developed, there was frequent discussion of data shortages for measuring relationships or challenges. Discussions often raised the need for more data collection. This analysis did not confirm a universal lack of data. If anything, there is not necessarily an overarching shortage of data for all locations and media of concern. Rather, there may be more difficulties in:

• awareness of the data resources
• practical ability to use different data sets to bring together different variables
• capability to organize and use the data in ways that meet user needs.

Actual data gaps may still exist due to specificity, cost, or privacy constraints. Before further data collection, analysts should reflect deeper consideration of existing data, and the context and policy intent of the analysis.

There is also interest in understanding health outcomes. However, health data are often limited or aggregated due to privacy considerations. Assessing unique, community-level health risks requires detailed, context-specific information and is better suited to targeted analyses rather than national-level studies.

Many people have also expressed interest in including more variables that reflect socioeconomic conditions and intersectionality. However, adding more variables can make it harder to identify specific relationships, like those related to race or Indigeneity. Expanding the set of variables requires a clear understanding of the purpose of the analysis and the questions it aims to answer.

If the goal is examining broader social conditions, then existing tools such as the Canadian Index of Multiple Deprivation or the Canadian Social Environment Typology may be more appropriate for assessing socioeconomic status. These tools, however, measure different factors than those outlined in the Act.

This leads to final considerations regarding analytical approaches. There may be opportunities to establish a baseline and shared methodology to support others in their analysis. Focusing on clear and precise use of terms like “risk” and “disproportionate” can help place concerns in context.  All of the above reflect the importance of balancing national comparability with local relevance.

Further opportunities

A central objective of the study was to develop methods that can be applied to analyses of specific communities, places or environmental hazards. They use the same principles as the national analysis while recognizing data limits.

Although there has been public interest in more detailed analyses, the federal government’s objective is to support informed dialogue rather than identify or categorize specific communities.  Detailed case studies are more effective when shaped by the communities involved, which are best positioned to provide context and identify relevant factors. 

The analysis also highlights the importance of understanding what data may already exist. It is important to understand how best to use it before considering collecting new data. For example, focus on opportunities for existing data resources to make specific, targeted improvements where appropriate and improve linkages between datasets (e.g., common identifiers for facilities, inventories of federal monitoring sites) to improve integration and reduce duplication

Further opportunities remain an ongoing area for input and discussion.  Advancing the work requires participation from partners to increase efforts and provide on-the-ground perspectives. It is not solely the responsibility of the federal government. Partners’ involvement strengthens the work, reflecting the lived experience of a specific situation. To achieve the best results, it may be useful to develop mechanisms to receive feedback and discuss methodologies, including data access, use and measurement. 

Additional analyses

The text of the Act included a series of possible broad-scope measures in section 3 (3) (b) that could be envisioned for a Strategy. The broad scoping reflected a need to better understand the current context and relationship between impacted communities, industry /polluters and the environment before suggesting strategic approaches. This motivated inward-looking analysis in several areas that could support Strategy development.  

The focus of federal regulatory intersects was on analysis with respect to public perspectives’ suggestions that the federal legislative and regulatory system may be responsible for environmental inequities, but also offer potential opportunity to address environmental inequities. Analysis found that:

• EJ can help to inform and understand environmental decision-making processes and outcomes and how they are considered to cause ER
• regulatory outcomes seek to balance many factors impacting people in Canada, but are not experienced or perceived/accepted the same by everyone – there is not always a clear “good” or “bad” outcome
• complex regulatory spaces involving shared jurisdiction make it difficult to understand how environmental protection decisions, including assessment outcomes and compliance decisions, are made. Communications on processes and outcomes could improve understanding
• federal regulatory approaches have the “levers” to support environmentally just outcomes, but there is potential for challenges in understanding practical implementation and coordination, particularly for cumulative impacts and regulatory approaches
• environmental jurisdiction is complex across orders of government, and can change how EJ issues are perceived and addressed by communities
• more explicit consideration of environmental racism may exist in the context of regulatory “gaps” associated with First Nation reserves

The focus of representation in policy making was on “the involvement of community groups in environmental policy-making”.  Analysis considered practice and representation in federal environmental decision-making approaches from existing program/initiative/policy audit and evaluation processes. The results showed:

• more active approaches to engagement (e.g., community focused) are often seen as providing better outcomes and better opportunities for monitoring. However, these approaches may not be appropriate for all scenarios, with mixed perspectives on their success and more substantial capacity and resource requirements better aligned with longer-term representation and participation. Using those approaches needs to be considered in terms of clear objectives.
• there is no universally accepted or practical approach to engagement and consultation that yields environmentally-just representation, reflecting challenges in:
  • identifying who could or should be engaged to participate, which will be further affected by
  • the timing and scale of the potential outcomes of the decision or approach, relative to the timing of the engagement or consultation process, which can drive
  • differing levels of interest in participating, trust in process, or burden on participants, which can often impact
  • the actual “representativeness” of those who come forward, in terms of if or how they are able to speak on behalf of their communities (or themselves).
• efforts to develop clear and consistent guidance may help to improve overall understanding of needs and tools. However, there will be an ongoing need for flexibility to design approaches that best suit operational considerations, although early emphasis could consider:
  • guidance around how to identify partners for engagement to clarify representation in action
  • guidance to improve timing of engagement to align with obtaining necessary perspectives and laying out intended outcomes

Part 3: National Strategy Respecting Environmental Racism and Environmental Justice 2026-2031

The National Strategy sets a path to build a stronger Canada, one that protects communities, advances EJ, and addresses environmental racism. It focuses on reducing barriers, strengthening accountability, and working together so all communities, especially those most affected, can benefit from a healthier, safer, more secure environment.

The Strategy recognizes that environmental racism and unequal outcomes stem from broad social, historical, and economic forces that no single government or sector can address alone. Meeting this challenge means coming together, uniting efforts across governments, sectors, and communities—to build lasting solutions.  The Strategy focuses on the federal role, demonstrating how we can move forward together to build a stronger, a more inclusive and resilient future.

Vision

To advance environmental justice supporting equal access to healthy, safe and more secure environments for communities.

Guiding principles

The Strategy recognizes the importance of actions that are developed with the participation of rights holders, where appropriate, in a manner consistent with the Government of Canada’s commitment to take effective measures to achieve the objectives of the United Nations Declaration on the Rights of Indigenous Peoples. Other key international frameworks, like the United Nations Convention on the Rights of Persons with Disabilities, can help to drive inclusive actions. There are four guiding principles envisioned here.

Protecting the environment in environmentally hazardous areas 

Efforts to protect the environment should be grounded in a clear understanding of the harms disproportionately experienced by certain communities. Priority should be given to preventing environmental damage and strengthening ecological safety for affected communities. Environmental protection should be guided by best-available science and Indigenous Knowledge Systems.

Acknowledging lived experiences and realities, and respecting differences

Policies and actions should be informed by a clear understanding of what causes inequities and how environmental, social, and economic well-being are connected. Particular attention should be given to Indigenous, racialized, and other marginalized communities, and those whose livelihoods depend on natural resources. Learning more about each community’s local realities helps us see how environmental pollution makes existing inequalities worse.

Encouraging environmental education

Communities should have access to meaningful learning opportunities about the environment, including environmental hazards and their impacts. Timely, accurate, and relevant information is essential for informed decision-making. Strengthening access to information enables communities to better understand and respond to EJ challenges.

Building meaningful relationships and ensuring inclusive participation

Inclusive decision-making requires creating meaningful opportunities for affected communities of different backgrounds and experiences to shape decisions that impact their lives. Efforts should focus on identifying and removing barriers to participation, including considerations for fostering safe, respectful spaces for participation. Valuing the knowledge and experience within communities, unlocks better, more practical solutions that reflect local realities and strengthen our collective future.

Priorities and actions

Working toward the Government of Canada’s vision to advance environmental justice so communities can benefit from a healthier, safer and more secure environment is an ongoing process. This Strategy outlines initial proposed federal actions, grounded in four priority areas identified through engagement and consultation.

Priority area one: Promoting environmental justice across federal government institutions

This priority area supports embedding EJ principles across federal policies, programs, and priorities. By integrating EJ and intersectionality into decision-making, incremental systemic changes could:

• reduce environmental disparities
• improve trust and relationships between the federal government and Indigenous, racialized and otherwise marginalized communities
• support a fairer distribution of environmental outcomes for all Canadians

Potential priority actions:

• task a dedicated federal team to lead Canada’s efforts to advance EJ
• lead a federal community of practice on intersectionality, equity, and inclusion in environmental policies, programs and initiatives, including through tools like Gender-based Analysis (GBA) Plus
• identify opportunities to reflect EJ as a priority across federal policies, frameworks or strategies

Priority area two: Strengthening collaboration and coordination on assessing, preventing and addressing environmental racism for a stronger Canada

ER is a complex, systemic, and shaped by overlapping challenges. Many of these issues fall outside federal jurisdiction. However, the Government of Canada can play a key role to assess and address ER in a way that is unified and contributes to effective solutions that better serve communities unduly impacted by environmental harms.

Potential priority actions:

• develop and share a transparent methodology to assess environmental burden and social vulnerability nationally
• lead policy development on restorative justice approaches in the context of environmental policy
• establish a mechanism/platform to share consolidated information with affected communities
• identify opportunities to work with granting agencies and federal funding initiatives to address gaps in research, knowledge, and resources

Priority area three: enhancing impacted communities’ participation in federal environmental policy and decision-making

A key component of advancing EJ is recognizing and reflecting the perspectives of diverse communities, and integrating those perspectives into policies, programs and initiatives that impact them. This priority area aims to improve access and participation of Indigenous, racialized and marginalized communities in engagement and consultation, and to improve understanding of these processes. This includes clarifying whose input was reflected, and how input was reflected in the final outcomes in federal environmental policy decisions.

In the context of Indigenous peoples, this includes continuing conversations to understand and recognize the variety of Indigenous perspectives on ER and EJ. It involves seeking ways to clarify implications of environmental decision-making with respect to Indigenous rights as a way to advance EJ.

This work supports more inclusive and transparent environmental policies and engagement practices, contributing to increased trust and more equitable outcomes for communities disproportionately impacted by environmental harm.

Potential priority actions:

• support enhanced ways to track and communicate:
  • information about engagement opportunities
  • how input was incorporated
  • data related to representation, outcomes and participation
• lead the development of a federal framework on inclusive engagement in environmental policy
• investigate coordinated departmental approaches to better identify potential impacts to Indigenous rights, and engage with impacted groups
• develop and publish guidance on addressing common barriers experienced by communities in environmental policy engagement and consultation activities
• lead research and recommendations to strengthen recognitional justice in Canadian environmental policy development

Priority area four: Learning from experience and avoiding recurring challenges

This priority area aims to work towards better equipping the federal government with the knowledge, tools and capacity to:

• assess and anticipate complex risk factors
• prevent recurring challenges
• support continuous improvement in environmental outcomes for Canadians

Potential priority actions:

• identify and share lessons learned through local case studies assessing and addressing ER, where appropriate
• lead the development of federal policy approaches to better integrate social, economic, and environmental analysis in federal environmental decision-making processes
• work with federal institutions to develop an improved data framework that better supports an integrated assessment of environmental regulatory and enforcement data

Roles and responsibilities

The Strategy focuses on federal action. However, environmental protection in Canada is a shared responsibility across all levels of government, through legislation, regulation, policy, planning, and resource management.  Communities also play a vital role in providing perspectives that help inform and refine efforts to advance environmental justice.  This reality could be achieved by:

• leveraging EJ principles (procedural, recognitional, distributive and restorative) to help ensure that decision-making is equitable and inclusive
• considering diverse experiences of impacted people
• collaborating between governments, and other actors like industry, to facilitate accessible access to information and data to help the public understand potential impacts on communities

Some critical areas where Canada already has obligations or responsibilities that could contribute to   advancing EJ principles include:

• responsibilities on federal Crown lands, where the federal government holds primary responsibility for the stewardship and environmental management, including First Nation reserve lands
• Federal Contaminated Sites Action Plan (FCSAP), where the Government of Canada identifies, assesses, and remediates contaminated sites, some of which are on or near Indigenous or racialized communities affected by long-term pollution
• a Right to a Healthy Environment under the Canadian Environmental Protection Act, 1999  (CEPA), which includes consideration of EJ in exercising federal authorities to take action related to managing toxic substances, protecting air and water quality and assessing environmental and health risks
• Canada’s Anti-Racism Strategy, which commits the federal government to addressing systemic racism across institutions, policies, and initiatives. This could include environmental decision-making, by removing barriers and making systems more inclusive.
• longstanding commitments to intersectional analysis, through the application of Gender-Based Analysis Plus, to better understand the impacts of policies, programs and initiatives on diverse groups of people

Moving forward with implementation

Putting these strategic efforts into action creates opportunities to work together to assess, prevent and address environmental racism. It also helps define how we measure success, through shared progress and practical results in communities. Crucially, advancing EJ is something we do together, when we work together.

Putting the Canada’s National Strategy Respecting Environmental Racism and Environmental Justice (2026-2031) into motion requires a strategic and coordinated action plan. The action plan will draw on:

• an enhanced federal role
• ongoing collaboration with partners and communities
• leveraging the processes, tools and opportunities available across the Government of Canada to drive impactful change.

We welcome feedback on how the Government of Canada can best implement the actions outlined in the priority areas. Perspectives from communities, organizations, experts, and individuals will help ensure that the proposed measures are practical, effective, and responsive to diverse needs and experiences to build a stronger, secure, united Canada.

Annex A: Methodology and results of examination of links between race, socioeconomic status and environmental risk 

This annex outlines the methodology used to assess links between race, socioeconomic status, and environmental risk. It also includes the numerical results of the statistical analysis.

Datasets and variables

The datasets were selected based on a series of guiding principles including using nationally consistent data that were released by the federal government. This approach ensured nation-wide consistency and ability to access and manage data. Emphasis was also placed on public availability and routine updates to ensure repeatability and accessibility of outcomes.

The following population variables and data resources were used to complete the analysis:

• racialized people (“race”) as a variable was based on Census 2021 population statistics for Dissemination Areas (DAs) from Beyond 20/20, using line 1684 divided by line 1683 to create a racialized population rate
• Indigeneity (“Indigenous”) as a variable was based on Census 2021 DA-level population statistics from Beyond 20/20, using line 1403 divided by line 1402 to create an Indigenous population rate
• Low Socioeconomic Status (LSES) Index was a newly constructed variable using Census 2021 DA-level population statistics from Beyond 20/20. It was built by creating the following:
  • low income, after tax rate (line 340 over line 335)
  • the lower education rate based on people without at least high school diploma (line 1993 over line 1992)
  • unemployment rates (line 2226 over line 2224)
  • Where all three rates were available (i.e., not suppressed), they were averaged together to create an LSES Index value that could show relative highs and lows across these three social indicators combined

Population rates were used so that measures were standardized for differing population sizes. Census data were prioritized, given ability to access a broad range of relatively localized data. However, Census data considers home locations (where people live). Concerns have been raised about work locations as a potential source of environmental exposure and injustice, but data on workplace populations was not readily available for this assessment. As well, some Census data reflects sampling estimates and may reflect errors or inaccuracies associated with sampling processes. However, these are potential sources of error generally acknowledged and mitigated to the extent possible by Statistics Canada.

A metric of potential environmental hazard (PEH) was built to estimate environmental risk. This considers the possibility of interactions between industry/pollution and community populations in defined spaces. The PEH considers current environmental hazards, rather than the past (i.e., potential decades of potential hazards). This avoids drawing connections that may not currently exist. This approach included counts of:

• sites that had reported into the National Pollutant Release Inventory (NPRI) within the last five calendar years (i.e., 2021 to current)
• active or suspected sites in the Federal Contaminated Sites Inventory
• federally regulated wastewater sites (i.e., reporting under the Wastewater System Effluent Regulations) that are not located within 3 km of an NPRI site, to reduce potential duplicate reporting between the datasets

To transform the hazard counts to PEH, the number of hazards in each spatial unit was multiplied by the hundred-weight total population of that unit (based on line 1 from Census 2021 in Beyond 20/20, divided by 100). This gave a sense of a total population size that could be affected.

The resulting figure was then divided by the land area (in square kilometers) of the same spatial unit. This reflects the potential spread of the population and hazard counts but assumes that people and hazards are distributed equally over spaces. The resulting PEH value for each spatial unit gives a generalized sense of the possibility that people and hazards will interact. 

There are many other sources of environmental hazard that are not captured in the selected datasets. Furthermore, counts of sites will not consider the outputs in terms of:

• pollution volume
• specific compounds and their toxicity or
• timing and frequency of releases

However, it is not plausible to address all potential sources of data, compounds, or potential environmental racism/EJ concerns in a single approach. The range and type of hazards considered in this analysis will limit the applicability of the results at a local level. Selected hazard data sources reflect a broad cross-section of available federal data, focused on industrial activities and pollution sources. However, the expectation is that the methodologies developed here may be used to examine more specific relationships.

Spatial relationships

To better understand spatial interactions between populations and pollution, a clearly defined spatial approach that aligns people and potential hazards is critical. Acknowledging the different ways this could be examined, three different spatial approaches were examined. They were based on technical feedback from experts and efforts to better understand the implications of changing spatial relationships, and included:

Simple co-location

DA population data linked with PEH based on summary of hazard counts within the DA Census boundaries. This approach was simple. However, it did not account for “boundary” effects, where hazards near a boundary line may still affect other DAs nearby. The sizes of DAs were highly variable, creating the requirement to standardize hazard measures for differing land area.

Tessellated co-location

Canada was re-envisioned as a series of equal-sized (500 square kilometer) hexagons for consistency in spatial comparisons. Populations were assumed to be equally-distributed across DAs, and re-allocated into hexagons using ArcGIS tools. Hazard counts were allocated within the new hexagons as well, “washing out” the PEH standardization for unit size (i.e., all units were 500 square kilometers). Hexagons aggregate total population sizes to give a sense of potential hazard over larger spaces/regions.

Hybrid co-location/proximity

This used the same underlying approach as simple co-location. However, hazards were assumed to impact any DA within a specified distance. To do this, radiuses (0.25 km, 0.5 km, 1 km, 5 km, 10 km, 25 km) were drawn around all potential hazards. The number of hazards whose radiuses intersected the DA at a given distance was used as the count of hazards to create the PEH for each DA.

This considers that hazards may affect multiple DA spatial units, but also makes analysis subject to the relevance of the radius being considered. Selecting an appropriate scale for national context across different pollution source types and outputs can be challenging, and not aligned with local concerns, Selecting a specific radius also favours inclusion of populations from more population dense areas, since the same set radius will be able to “affect” more small spatial units than large.  Notably, the hybrid proximity approach was not used for all analyses. This is due to the possibility of duplicate-counting of hazards.

Community

Communities are based on the population within a spatial unit; within a DA, or within a hexagon. Considering different spatial relationships outlined above allows analysts to better understand nuances in how environmental racism/EJ relationships might be understood. Population and PEH values were joined into the same spatial units with the aid of ArcGIS software, and published Census boundary data provided by Statistics Canada.

Methodological approaches

The methodologies used for analysis fall into two broad categories: statistical analysis and visualization.

Statistical analysis

Statistical approaches  were selected to address the research questions being posed, with priority for methods that are rigorous and accessible, ensuring they can be interpreted by broader audiences. This resulted in the selection of five core statistical methodologies that emphasize relationships between population and hazard variables. These are outlined below.

Odds ratios examined the difference in odds of a community having a specified population group (e.g., racialized or Indigenous) depending on the presence of a hazard. This uses binary (e.g., yes or no) exposure and outcome selection. Population presence is the “exposure” and the existence of a hazard was treated as the outcome. Odds ratios provide a distinct consideration of proportionality based on comparison, but have a tendency to increase quickly.

Relative rates were used to examine difference in likelihood of a community having a hazard based on whether certain populations were present. This used the same underlying data as the odds ratios. Relative rates give a sense of disproportionality, but need to be carefully understood when underlying data do not have substantial controls for other factors.

Spearman ranked correlations that correlate the population values and PEH across all spatial units. Ranked correlation addresses strong skew in population rates, allowing directional consideration of very unique outcomes. Correlation values are between 0 (no correlation at all) and 1 or-1 (perfect positive or negative correlation, respectively). Correlation gives a general sense of whether the values moved in similar directions, but not of causality between variables. Interpretation of correlation requires an assumption that there is a stable relationship between variables, which may not be reasonable or logical.

Chi squared tests of independence examine whether having a specific population rate and having a hazard (or hazards) were independent across spatial units. If those variables were not independent, the proportionate deviation from an independent relationship can be measured. These tests are primarily visual (beyond basic independence test statistics) but provide context around odds ratios and relative rates.  Ranges of population and PEH were built to separate “0” hazard or population groupings from relatively equal, increasing quartile ranges of non-zero values.

Moran’s I measures clustering within variables across spatial units (e.g., whether similarly scaled values are found together). It provides context to support interpretation of other statistical results by examining values between 0 (no clustering) to 1 or-1 (perfect clustering, or perfect dispersion, respectively).

Visualization analysis

Visualization brings together population information and PEH into a single visual output. This task was done by creating an “Environmental Effects and Population Composite Index” (EEPCI). It included the following:

• a “population” variable for each spatial unit that averages the scaled quartile values for visible minority rates, Indigenous population rates, and LSES Index values for each spatial unit
  • single value providing a sense of “who lives here” in context of potential marginalization
  • reflects that having a substantial visible minority, Indigenous population alone may not be entirely “unique”. Communities that have one or both of those populations and relatively poor socioeconomic status may be of increasing concern from an equity perspective
• the quartile range of the PEH value for each spatial unit

The PEH and population quartile values were multiplied together to yield a value between 0 and 16. A value of 0 would indicate either, or both of, no hazard or no population of concern in a unit. A value of 16 would indicate that the unit was in the highest quartile of all variables (PEH, visible minority rate, Indigenous rate, and LSES Index).

Mapping of the EEPCI was completed for all communities where all required data were available. This was completed for both DAs and tessellated hexagons.

Consideration was given to providing separate mapped “layers” of the underlying data. However, this could lead to interpretive challenges for users. The visualized outcome should be clearly aligned with the broader analytical intent of this work. As a result, the maps have focused on providing a single visual outcome considering the same variables across the country to understand where there may be unique relationships or intersections between populations-of-interest and environmental hazards.

Results and initial interpretations

Odds ratios and relative risk

Odds ratios and relative risk ratios were calculated for all spatial relationships, using binary indicators of exposure (population presence, yes or no) and outcome (hazard presence, yes or no). The tables below provide the separate results for odds ratios and relative risk ratios.

Table 1: Odds ratios of hazards for communities with different spatial relationships

Population Group	Co-Location	0.25 km Hybrid	0.5 km Hybrid	1 km Hybrid	5 km Hybrid	10 km Hybrid	25 km Hybrid	Tess. Co-Location
Visible Minority	0.26	0.37	0.55	0.9	5.61	7	7.38	6.68
Indigenous	2.01	1.76	1.49	1.26	0.65	0.64	0.5	4.7
LSES*	1.78	1.72	1.66	1.51	0.6	0.31	0.11	30.56

* LSES results split exposure at above or below median LSES rate, due to a lack of communities without LSES values, other than for tessellation results.

Table 2: Relative risk of hazards for communities with different spatial relationships

Population Group	Co-location	0.25 km hybrid	0.5 km hybrid	1 km Hybrid	5 km hybrid	10 km hybrid	25 km hybrid	Tess. co-location
Visible Minority	0.3	0.44	0.64	0.94	1.27	1.11	1.01	3.57
Indigenous	1.91	1.64	1.37	1.15	0.96	0.98	1	3.51
LSES*	1.71	1.6	1.49	1.29	0.95	0.96	0.99	20.32

* LSES results split exposure at above or below median LSES rate, due to a lack of communities without LSES values, other than for tessellation results.

Note: Shaded cell denotes that this value is not significantly different than "1" at 95% confidence intervals

Co-location and hybrid co-location odds ratios show opposite trends between visible minority communities and the Indigenous and LSES communities. As the radius of effects from a hazard increases, so do the odds of affected communities having visible minorities (moving from lower than 1 odds, or less likely, to nearly 7.5 times the odds). The opposite is true for Indigenous and LSES populations. Of particular interest, this includes a rough “inversion” of the trends between the 1 km and 5 km radius outcomes. In comparison, the tessellated approach indicates that all three population groups, when considered within a broader spatial unit, demonstrate higher odds of being located with hazards. This could reflect that it is more unlikely to find a 500 square kilometer portion of the country with no hazard sites for comparison.

Considering relative risk, the outcomes are substantially less variable than odds ratios. This is likely due to a shift away from a “ratio of ratios” to a ratio of rates. There are similar trends to the odds ratio for many of the results. Notably, with relative risk, the “inversion” of likelihood still happens at 5 km. However, the relative risk trends back towards “1”, where the risk is no different between populations. This could reflect that with proximities of that size or greater, most (if not all) of the DAs and population in Canada would end up being included. As a result, the rates (rather than odds) would start to look similar to “Canadian average”. As with odds ratios, tessellated results indicate higher rates of co-location between hazards and communities/areas with a visible minority or Indigenous population, or higher LSES population.

The “inversion” effect for the proximity results is expected to (at least partially) reflect that a radius will favour the addition of smaller, more population dense DAs into the “affected” population. This may increase visible minority populations included, given substantial visible minority populations living in densely populated areas. This could also partially explain why radiuses may not substantially increase Indigenous populations, given emphasis on more rural or remote areas with larger-sized DAs.

Overall, the results demonstrate that there can be disproportionality in basic relationships. However, the specifics of that disproportionality can change in terms of both magnitude and direction depending on how spatial relationships are characterized. This makes it challenging to draw a single, generalized conclusion around “disproportionality”.

Spearman ranked correlation

Correlation analysis was completed for both DAs and tessellated hexagons. It was not completed for the varying levels of hybrid co-location.

Table 3: Correlation results for population rates and PEH values

Population Group	DA Co-Location PEH	Tessellation PEH
Visible Minority	-0.19	0.45
Indigenous	0.13	-0.01
LSES	0.09	0.05

The results of the correlation analysis show varied correlation values and direction depending on spatial scale. In most cases examined, the correlation values are relatively weak (e.g., less than 0.2). This may reflect low likelihood of broad, simple variables aligning closely over entire scales of population and hazards. An exception to weak values is the moderate positive correlation between visible minority population rates and PEH for tessellated communities. With /moderate negative correlation at smaller scales, this may suggest that regional scales are better positioned to demonstrate a potential relationship unique to visible minority populations. That said, moderate results also merit caution against concluding any specific or stable relationship among these variables. 

Chi squared test of independence

The chi squared test of independence considers differences between observed and expected outcomes across scaled categories of exposure (i.e., population rates) and outcome (i.e. PEH). The scaled population rates were different between DA spatial units and hexagons due to population aggregation. The differences are reported in Table 4. In general, combining populations into hexagons reduced the within-quartile range for visible minority populations. An exception was the highest quartile. Combining populations in hexagons increased the range within quartiles for Indigenous populations (except the highest quartile), and ranges were relatively stable for LSES.

Table 4: Population rate characteristics by quartile group, based on spatial unit

Quartile	Visible minority % DA	Visible minority % Hexagon	Indigenous identity % DA	Indigenous identity % Hexagon	LSES index x 100 DA	LSES index x 100 Hexagon
Q1 Range	0.4 - 7.7	0.0 - 1.2	0.1 - 2.7	0.1 - 6.6	0.0 - 9.4	2.0 - 13.7
Q2 Range	7.7 - 18.9	1.2 - 2.6	2.7 - 4.9	6.6 - 16.7	9.4 - 12.4	13.7 - 16.7
Q3 Range	19.0 - 40.6	2.6 - 5.3	4.9 - 9.3	16.7 - 75.0	12.4 - 16.6	16.7 - 21.4
Q4 Range	40.6 - 100.0	5.3 - 78.5	9.3 - 100.0	75.0 - 100.0	16.6 - 64.3	21.4 - 56.7

For chi squared test estimates, the expected outcomes are intended to reflect an “independent” scenario (i.e., where the exposure and outcome are not related). The initial outcome of a chi squared test of independence is a test statistic summarizing the difference of observed values from expected values. In all cases tested, it was determined that population and PEH variables were not independent at the 95% confidence level. This is not surprising, given the relatively large number of communities in the analysis and the ranges of the underlying variables considered. This provides context for previous correlations. While relationships may not be strong, “stable” or consistent, the variables are not entirely independent. These findings, while limited, reveal a discernible relationship that future work could substantiate, when considering the underlying variables are not completely independent.

Potential relationships were considered in terms of percent difference between observed and expected outcomes (based on number of spatial units). The included analysis compared scenarios where the hazard outcome is binary (i.e., yes/no to having a hazard) with population rates scaled (i.e., increasing rates in non-zero quartiles). 

There are different results for the co-location and tessellated co-location outcomes (see Figures 1 and 2, respectively). The results are consistent with the findings of correlations, odds ratios, and relative rates. Specifically, direct co-location (Figure 1) tends to indicate inverse effects for visible minority populations and positive increasing effects for Indigenous and LSES populations. For example, inverse effects could be less units than expected actually having hazards in relation to increasing population rates. Positive increasing effects could be more units than expected actually having hazards as population increases.

However, when considering tessellated relationships, the inverse relationship for visible minority populations switches to a positive effect (Figure 2). In the same figure, the Indigenous result shows a slight “inverse” outcome in the trendline. This is largely due to a unique pattern. For example, fewer areas with hazards than expected at lowest and highest population rate, but more than expected in moderate population rate groups.

Figure 1: Percent variance in number of DAs with hazards compared to expected, by population type and quartile

Figure 2: Percent variance in number of hexagons with hazards compared to expected, by population type and quartile

The same analytical approach was taken using quartile scales for PEH as well. The resulting visualizations are not included here but largely demonstrate the same trends outlined in Figures 1 and 2. They have more “nuance” in specific levels of population and hazard. However, they are substantially more complicated to explain with similar overarching interpretation of results.

Moran’s I

The Moran’s I analysis was run on DAs rather than hexagons, given the larger number of communities to test. Considering population rates, rather than counts, avoids clustering effects related to total population size effects, rather than population composition. Hazards were examined counts and as PEH, to consider whether the adjustment to PEH could substantially change understandings of potential hazard. The Moran’s I test results (see Table 5) were significant to at least the 95% confidence level.

The results of this test indicate moderate levels of clustering in visible minority and Indigenous populations. There was less clustering in combined LSES. This may reflect the strong skew in visible minority and Indigenous rates, often because people choose to be co-located with “like populations”.

In comparison, underlying LSES data sources can be more “normally” distributed, giving a sense they would be less likely to cluster together. That said, some areas of increased values can lead to a smaller clustering effect. As well, LSES results may reflect “washing out” since LSES is constructed of three different sub-variables.

Importantly, there is little evidence of clustering in included hazards. This may reflect that hazards are genuinely dispersed, or that combining several hazard types into a single measure obscures individual clustering patterns, or both.

Visualized EEPCI Index

Visualization of the EEPCI was completed for both DAs (Figure 3) and for tessellated hexagons (Figure 4) across the country.

In Figure 3, many of the higher index values are difficult to see on a national/large scale due to relatively small community area. This could be addressed through “zooming in” on a map. However, this does create questions about the intent of mapping efforts as an approach to support comparison across the country. DAs can show very localized outcomes. However, the zoomed in results are focused within set boundaries.

Comparatively, Figure 4 provides more consistent units of comparison. Figure 4 is also perceptibly more effective at highlighting potential areas of interest in Northern communities that are less visible in DA-based maps. However, when “zooming in” it is unlikely that local communities would find the information as appealing. 

A notable trend from Figure 4 is the emphasis on coastal regions, likely reflecting access to transport and resources (e.g., water) for industry. Similarly, it highlights known regions of population density and industrial activity. In that sense, hexagonal approach may better approximate “known” or assumed intersections at higher levels. DA-based approaches may be more likely to identify unexpected local outcomes (both positive and negative) that require more explanation based on experience.

Figure 3: EEPCI Values Across Canada, by DA

Figure 4: EEPCI Values Across Canada, by Hexagon

Considerations and lessons learned

The results show that relationships between communities and environmental hazards vary across the country. However, the specific findings depend on the data used and the assumptions built into the analysis.  Patterns may reflect inequality, spatial distribution, or both. Because of this, it can be misleading to use simple national level results to draw broad conclusions or to ‘label’ communities. At these large scales, it is difficult to determine what would count as ‘typical’, ‘proportionate’, or ‘equitable’ outcomes.

Using proximity could be problematic without carefully defined logic models, parameters and variables.  Impacts vary depending on pollutant type, dispersion, and exposure. It could also be more difficult to identify a specific population to compare against to estimate disproportionality. A strong emphasis on proximity to pollution, or neighbourhood-specific concerns, may be more relevant to a community. However, it may also be driven by other underlying differences that are unique in context.

Emphasis on proximity-based statistics alone could even result in misleading conclusions. As an example, multicultural populations are a defining feature of Canada, and there are likely to be clusters of specific populations in certain neighbourhoods. However, if EJ is assessed solely based on proximity of a unique community without appropriate context, these patterns could be misinterpreted.

All of the above points reflect that metrics alone cannot fully capture community experience. These measures do not reflect people, context, or history and are sensitive to analytical choices. They require context from community narratives and histories. Using statistics alone misses several key factors, such as:

• how people in “affected” communities actually feel, and their perspectives
• how their perception is influenced by the perspectives of external players (e.g., Governments, industry, ENGOs, academia, etc.)

As noted, results are also influenced by the unit of measurement used (i.e., the size of the area being studied). Selected units may not align with specific community contexts. There is a balance between consistency (larger units) and specificity (smaller units). Both are needed to understand relationships and derive value from analytical outcomes. Larger scales may address some concerns around proximity but may miss local experiences. This also reflects the challenge of defining what constitutes a “community” for analysis.

This is particularly evident in mapping and comparing the differences between DAs and standardized hexagons across the country. There are also challenges in visualizing results at national scales. Increasing detail (e.g., percentiles) can make visualization more difficult, while aggregation helps preserve meaningful differences.

Additional practical lessons learned became apparent during analysis. Firstly, despite the intent and emphasis on public accessibility and repeatability, the analysis completed required access to GIS software. It also required access to trained officials that can join data and plot information for visualization. While the data itself may be accessible, the software (and ability to use that software) may be a limitation. This is particularly the case if data “cleaning” or adjustments are required.

Secondly, the results are based on selected variables aligned with the text of the Act and practical decisions of analysts. Results will change when variables are adjusted, reflecting the complexity of environmental racism and EJ that cannot be captured in a single definition or approach universally understood and agreed-upon across all communities’ experiences. Emphasis can be placed on developing methodologies that can be used, assessed, and refined over time through more specific analyses. This can support broad-scale policy dialogue about environmental equity and informing policy decisions.

This work provides a national-level foundation and points towards areas for further consideration, which will need to consider their own  specific context and logic models (e.g., examining specific, logical relationships).The work completed here contributes to national-level, policy-focused knowledge, and recognizes that community-level experience may require different scales and approaches that build off of national considerations. This community perspective often falls outside the role of the federal government, pointing to a potential role for governments as providers of information and tools (like methodologies). Community-level analysis is most powerful when community-led.

Those seeking more detailed information about the approaches and results outlined above are encouraged to contact ej-je@ec.gc.ca. Reach out to continue the discussion and gain access to more detailed materials as they become available.