Homelessness data snapshot: An exploratory analysis of the outflow of shelter users

• Copyright

  © His Majesty the King in Right of Canada, as represented by the Minister of Housing, Infrastructure and Communities, 2025.

  T94-60/8-2025E-PDF
  ISBN 978-0-660-79859-2

On this page

• Introduction
• Data and methods
• Results
  • Annual outflow: Are a greater proportion of people flowing out of the shelter system?
  • Time to outflow: How long does it take to outflow?
  • Reasons for exit: What proportion have exits into housing?
• Key findings
• For more information

Introduction

This snapshot describes the outflow of emergency shelter users in Canada between 2018-19 and 2023-24. Outflow is the rate at which people exit the shelter system. This analysis can help measure progress in the goal of reducing homelessness and provide insights to better understand flow through the emergency shelter system.

The federal government tracks shelter use by measuring the annual number of shelter users, however, this is only one way to look at progress towards ending homelessness. It is important to also measure upstream prevention and downstream outcomes to determine whether the system is effectively reducing homelessness. The following questions aim to measure progress over time for each flow stage to present an overall picture of system health:

Upstream prevention: Are fewer people entering shelters for the first time? This topic was explored in the Analysis of the inflow of shelter users in Canada 2018 – 2022 data snapshot as well as The National Shelter Study 2024 Update.

Emergency shelter use: Are fewer people using shelters overall? Are their stays getting shorter or longer? This topic is explored in annual The National Shelter Study 2024 Update.

Downstream outcomes: Are a greater proportion of people flowing out of the shelter system? How long does it take to outflow? What proportion have exits into housing? These questions will be explored in the current report.

Data and methods

For this analysis, individuals who experienced homelessness in emergency shelters were tracked over time to observe outcomes.

The data were collected by service providers using the Homeless Individuals and Families Information System (HIFIS) and through data-sharing partnerships with jurisdictions that use alternative Homelessness Information Management systems.

Cohorts, or groups of people who accessed shelters from 2018-19 to 2023-24 were identified in communities with five years of complete and quality emergency shelter data (Table 1). These consisted of:

• three years of historic community data provided insight into past episodes of shelter use to determine whether a person was identified in the shelter system prior to their inclusion in the cohort;
• one year to select a cohort; and
• one follow-up year to determine each person’s outcome.

Table 1: Number of Individuals and communities represented in each cohort year.

Year	Number of communities	Number of individuals in cohort
2018-19	19	53,409
2019-20	13	38,160
2020-21	13	25,564
2021-22	10	28,804
2022-23	17	43,817
2023-24	21	47,961

This is the first study on emergency shelter outflow produced by the federal government. Longitudinal homeless studies like these are challenging to conduct as it is difficult to obtain quality and standard homelessness data over the long term. In each year, communities from British Columbia, Ontario and Nova Scotia are represented. Small, medium and large communities are included.^{Footnote 1} While this selection of communities may not be nationally representative, the results nonetheless offer important insights into how individuals flow through shelter systems over time, a subject that is notoriously difficult to explore. It also sheds light on areas to focus our data improvement efforts so that we can answer more questions in the future. As community-level efforts to improve data quality and completeness continue, we will gain an even better understanding of flow through the system with a more nationally-representative sample of communities.

Each person’s one-year outcome was determined based on the definitions in Table 2.

Table 2: 1-year outcomes and definitions

Outcome	Definition
Outflowed	A person was considered to have outflowed if their last exit date occurred in the cohort year or within 90 days of the end of the cohort year if their episode crossed over the end of the year. That is, they left the shelter system and did not return in the following year. It is important to consider that the analysis does not capture the full scope of housing instability (e.g., couch-surfing or hidden homelessness), and so some may continue to experience homelessness outside of the shelter system. Outflow in this analysis also does not capture people who may return to the shelter system after 365 days.
Retained	A person was retained if an episode of homelessness in the cohort year crossed over the end of the year beyond 90 days. That is, they did not leave the shelter system.
Returned	A person returned if they returned after a minimum period of 90 days inactivity within 365 days following their last known exit in the cohort year. That is, they left and came back to the shelter system in the following year.

Results

Annual outflow: Are a greater proportion of people flowing out of the shelter system?

The proportion of people that outflowed was compared for each cohort year to determine if there has been an increase in leaving the shelter system over time. An increase would suggest that homelessness is being resolved for more people.

In each year, the majority of people outflowed and did not return to the shelter system in the following year (Figure 1).

Figure 1. Proportion of shelter users that outflowed, were retained, or returned between 2018-19 and 2023-24

In each year between 2018-19 and 2023-24, the majority of individuals outflowed from the shelter system (62% to 69%). Between 22% and 24% returned within a year and 9% to 15% were retained in the shelter system. While the trends remained relatively stable over time, the 2021-22 cohort saw the lowest outflow rate (62.1%) and the highest rate of retention (15.2%). This suggests that there was more of a system bottleneck compared to other years. People were entering, but departing more slowly. The 2020-21 cohort saw the highest rate of returning clients (24.3%). Consistent with these findings, the annual The National Shelter Study 2024 Update shows that the number of shelter users experiencing chronic homelessness began to rise after 2020.

Outflow and indigenous status

All cohorts were combined to compare outflow results between Indigenous and non-indigenous persons.^{Footnote 2} There were some statistically significant differences. Indigenous Persons were significantly less likely to outflow (58.3%) compared to non-indigenous persons (67.6%). Indigenous Persons were slightly less likely to be retained (6.6%) compared to non-indigenous people (10.1%) but were significantly more likely to return within a year from inactive (35.2%) compared to non-indigenous people (22.3%). This suggests that Indigenous Persons had less stable outcomes compared to non-Indigenous people. Coupled with an overrepresentation in the shelter system compared to the general population, this indicates a systemic inequality (Figure 2).

Figure 2. Proportion of shelter users that outflowed, were retained, or returned between 2018-19 and 2023-24 by indigenous status

Outflow and gender

All cohorts were combined to compare outflow results between gender identities.^{Footnote 3} Those who identified as female were statistically more likely to outflow (70.7%) in a given year compared to those who identified as male (63.2%) as well as those who identified as gender diverse (66.1%). Those who identified as a gender diverse were statistically more likely to outflow compared to males.

Those who identified as male and those who reported another gender identity were more likely to return within a year compared to those who identified as female (Figure 3). This data suggests that females have slightly better outcomes compared to males and those who identify as gender diverse.

Both males and females were more likely than those who identified as gender diverse.

Figure 3. Proportion of shelter users that outflowed, were retained, or returned between 2018-19 and 2023-24 by gender identity

Outflow and age

All cohorts were combined to compare outflow results between age groups.^{Footnote 4}

• Accompanied children with families had the highest outflow (81.9%) followed by youth (70.4%), seniors (66.7%), then adults 25 to 49 (63.3%). Older adults had the lowest outflow (61.6%).
• Youth and adults aged 25 to 49 had the lowest rate of retention (9.7%). There was no significant difference between rates of retention among accompanied children, older adults and seniors.
• Accompanied children had the lowest proportion of returns within a year (2.1%) followed by youth (11.4%). Adults aged 24-49 had the highest proportion of return (14.7%)

This data suggests that accompanied children have the most stable outcomes compared to other age groups. This might be due to increased supports for families experiencing homelessness both in terms of prioritizing available housing, but also financial supports.

Figure 4. Proportion of shelter users that outflowed, were retained, or returned between 2018-19 and 2023-24 by age group

Time to outflow: How long does it take to outflow?

The time that it took people to outflow was compared for each year to determine if there has been a decrease. A decrease would suggest that homelessness is being resolved more quickly.

Among those who outflowed, the median length of the homelessness episode prior to outflow increased significantly between 2018-19 and 2023-24 (Figure 5).^{Footnote 5} In 2018-19, the median episode length was 41 days. In 2020-21, it was 70 days. In 2023-24, it was 66 days. This increase suggests that people are taking longer to exit shelters post-pandemic compared to pre-pandemic.

Figure 5. Median length of shelter episode prior to outflow between 2018-19 and 2023-24

Reasons for exit: What proportion have exits into housing?

The proportion of people who exited to housing was compared each year to determine if there has been an increase over time. A decrease would suggest that people are flowing to a more stable housing situation compared to a less desirable outcome like hidden homelessness.

Reason for shelter exit was recorded for each stay whenever available. This information can give some insight into a person’s immediate destination after leaving a shelter. Response options were rolled up into six categories:

• Housed: This category includes people who found independent or private market housing, subsidized housing, supportive housing, or housing through an agency placement. This category also includes people who have returned to their previous address or partner, or have found housing in a residential care facility.
• Institutionalized: This includes people that were admitted to the hospital for medical or psychiatric reasons, people who were admitted into detoxification, a recovery program, or a treatment program for medical, psychiatric, or substance-use reasons. This also includes people exited to a correctional facility or jail.
• Temporarily accommodated: This category includes transitional housing.
• Hidden homeless: This includes people staying temporarily with friends or family members.
• Deceased: this is likely an undercount as some deaths are underreported.
• Unknown: Some exit response options do not provide insight into a person’s outflow path. For example, they may have left the area or returned to the street. Disagreement with the shelter rules, being expelled, completing the shelter program, or reaching maximum time allotted in the shelter are also reasons for exit, but it does not tell us of their outcome. Therefore, their outflow path is left unknown. Missing values are also rolled up into this category.

While most Homelessness Management Information Systems have the capacity to collect information about exit destinations, it is often difficult to record people’s trajectories out of the homeless shelter system using administrative datasets. Between 53% to 63% of people who outflowed in each cohort had an unknown reason for exit. Unless a person has been assisted in their housing search or has participated in a support program, information about their whereabouts is limited. The ability to get in touch with people after they exit the shelter system is also challenging due to housing instability, lack of access to communication technologies, and lack of time and resources to conduct thorough follow ups.

Approximately 34% of the individuals (n=46,654) who outflowed between 2018-19 and 2023-24 were recorded as going into housing. However, this number is likely an underestimate as the majority of people exited to unknown destinations (Figure 6).

Figure 6. Year-over-year comparison of reason for exit among those who outflowed between 2018-19 and 2023-24

Those who were chronic shelter users were less likely to outflow (55.6%) compared to non-chronic individuals (70.2%). However, those who experienced chronic shelter use and outflowed were more likely to have favourable outcomes. For example, they were more likely to report leaving for housing or temporary accommodation. Those who were chronic were also more likely to be reported as deceased compared to non-chronic individuals.

Figure 7. Median length of shelter episode prior to outflow between 2018-19 and 2023-24

Key findings

Overall trends

The majority of people (>62%) outflowed.

• In each year, the majority of people (>62%) outflowed from the shelter system.
• In each year, around 9% to 15% of shelter users were retained into the following year and about 22% to 24% of shelter users left and then returned within a year.

A smaller proportion of people outflowed in 2021-22 in comparison to previous years.

• Overall, the shelter system in 2021-22 experienced high pressure due to longer episodes of shelter use prior to outflow, combined with lower outflow rates and higher retention rates compared to pre-pandemic years. This suggests that there is more of a system bottleneck compared to previous years. Likely this is due to housing availability and affordability.

It is taking longer to outflow over time.

• Among those who outflowed, the median length of the shelter use episode prior to outflow increased between 2018-19 (41 days) and 2023-24 (66 days).

Demographic analysis

Children, youth and women are more likely to outflow. Indigenous people and adult men are more likely to return.

• Accompanied children had the highest outflow (81.9%) followed by youth (70.4%), seniors (66.7%), then adults 25 to 49 (63.3%). Older adults had the lowest outflow (61.6%).
• Youth and adults aged 25 to 49 had the lowest rate of retention (9.7%). There was no significant difference between rates of retention among accompanied children, older adults and seniors.
• Accompanied children had the lowest proportion of returns within a year (2.1%) followed by youth (11.4%). Adults aged 24-49 had the highest proportion of return (14.7%)
• Compared to non-Indigenous people, a lower proportion of Indigenous Persons outflowed (58.3% vs 67.6%) and a higher proportion returned to shelter (35.2% vs 22.3%). Coupled with an overrepresentation in the shelter system compared to the general population, this indicates a systematic inequity both in terms of overall shelter use and outcomes.
• A higher proportion of those who identified as female outflowed (70.7%) compared to those who identified as male (63.2%).

Exit analysis

Information about the exit destination of those who outflowed is limited.

• More than half of exits were to unknown destinations.
• Quality data on exits from shelter are important for understanding outcomes for those who have passed through the shelter system.
• Shelter users experiencing chronic homelessness are less likely to outflow. However, they are more likely to report exiting homelessness to housing upon outflow (51.5%) than those not experiencing chronic homelessness (28.6%).

For more information

Find out more about homelessness research on Reaching Home: Canada's Homelessness Strategy.

If you have any questions about this report, contact us.