Building on the country case studies presented in the previous chapter, the chapter draws lessons from international practice of homelessness measurement, examining how different data collection approaches are shaped by policy objectives, institutional arrangements and governance frameworks. The chapter identifies considerations for strengthening homelessness measurement systems and proposes a set of recommendations for Germany. This includes short-term recommendations to improve the clarity, usability and implementation of data collection tools, in addition to medium- to long-term recommendations to improve coverage and data linkages to build an integrated measurement system.
Measuring Unsheltered and Hidden Homelessness in Germany
4. Strengthening the measurement of unsheltered and hidden homelessness in Germany
Copy link to 4. Strengthening the measurement of unsheltered and hidden homelessness in GermanyAbstract
Lessons from international practice
Copy link to Lessons from international practiceSeveral lessons can be drawn from the country practices, along with a set of overarching trade‑offs to be considered in determining the appropriate data collection approach(es) to measure specific forms of homelessness, notably unsheltered and “hidden” homelessness.
Homelessness measurement is anchored in specific policy objectives
The country practices presented in Chapter 3 illustrate different approaches to measuring homelessness, which vary in their scope, design and relationship to policy. This is also reflected in the main lessons from the European Homelessness Counts (Box 4.1), where the data collection method has been designed to inform, evaluate and improve local policy responses, with clearly defined expectations regarding its use (Hermans et al., 2026[1]). However, the strength and directness of the connection between measurement and policy varies across contexts, both in terms of the stated purpose of data collection and the extent to which the findings are used to inform policy design and implementation.
Scotland represents a particularly direct case: homelessness measurement is embedded within the homelessness support system itself and captures data on its functioning, making a clear and immediate connection between measurement and policy action. The Nordic countries are similarly highlighted in the academic literature as examples where homelessness measurement serves as a central steering mechanism for the policy area, providing feedback on whether political initiatives and national strategies are effective in reducing or preventing homelessness (Benjaminsen et al., 2020[2]; Dyb et al., 2021[3]). Enumeration is also understood as a means of recognising the needs and claims of the affected population, identifying priority groups that can be targeted through policy measures, as well as legitimising political prioritisation and the allocation of state funding (Dyb et al., 2021[3]). The annual report in Finland illustrates the practical relevance of measurement as a steering tool: it includes dedicated sections allowing municipalities to explain the reasons for changes in their figures and to document the measures taken to reduce homelessness locally.
In Australia, homelessness data are intended to serve the targeting, planning, and funding of services to meet the needs of people experiencing homelessness. In Canada, the PiT Count framework makes explicit that the frequent annual enumeration is intended to measure progress towards ending homelessness, while the triennial Survey on Homelessness is designed to help government and community organisations better understand and serve individuals experiencing homelessness. Germany’s Homelessness Reporting Act likewise states that the reporting is intended to strengthen the evidence base for political action, though the Act does not elaborate on this objective in further detail.
The objectives of homelessness data collection in New Zealand and Czechia are more descriptively neutral: in New Zealand, the purpose is to provide information on the number, prevalence, and characteristics of people experiencing severe housing deprivation, while in Czechia the objective is to obtain qualified estimates of the number of people experiencing homelessness and housing exclusion across the national territory.
Box 4.1. Key findings from the European Homelessness Counts
Copy link to Box 4.1. Key findings from the European Homelessness CountsThe European Homelessness Count project developed and tested a common methodology to harmonise homelessness measurement across the EU (see Box 2.1). The final report, covering the second round of counts conducted in 2025, highlights several general key takeaways, as well as specific methodological insights (Hermans et al., 2026[1]).
Overarching takeaways
Policy purpose: Data collection serves as a tool to evaluate and improve local policies, with an expectation that findings will be acted upon.
Frequency: Annual surveys risk creating fatigue, while less frequent surveys may miss important trends.
Service paradox: A well-documented finding that a higher number of available services correlates with a higher recorded number of people experiencing homelessness.
Resource and time constraints: Counts place a substantial additional burden on service providers, suggesting they should be remunerated for this effort.
Methodological lessons
The project also yielded concrete methodological lessons, including:
Legal grounds for data collection: This is highlighted as an often‑overlooked dimension in homelessness data collection. Despite the EU framework under the GDPR, governments interpret it differently: some rely on a public-interest basis without informed consent (e.g. the Nordic countries), while others apply stricter, consent-based requirements (e.g. Germany).
Role of independent research institutions: In service-based methods, an independent research institution backed by a data processing agreement can facilitate data-sharing where service providers would otherwise be reluctant to share client data.
Limitations of administrative data: Useful for longitudinal insight, but working with it is time‑consuming, as it typically requires extensive cleaning and restructuring; differing variable definitions across providers also limit comparability.
Conditions for local ownership: The report identifies three key preconditions for local buy-in of the data collection exercise: i) political will and policy value at city level; ii) collaboration across a wide range of actors; iii) active involvement of outreach teams and social workers.
Length-of-questionnaire dilemma: Cities favour short questionnaires to ensure feasibility, yet the resulting core variables offer only a partial picture.
Source: (Hermans et al., 2026[1]), European Homelessness Counts Final Report Year 2, https://www.kuleuven.be/lucas/en/images/eu-homelessness-counts-final-report-2026.pdf.
The frequency of data collection can also have an impact on its timeliness and usefulness for informing and steering policy (Dyb et al., 2021[3]). For example, the comparatively long intervals of 6 years between surveys in Sweden (Table 2.6) has been linked in the literature to lower levels of political interest at the national level; where counts are used less intensively as instruments to measure policy impacts, the frequency of data collection tends to be lower (Dyb et al., 2021[3]). Additional considerations about data collection frequency are discussed below.
Different institutional and governance arrangements underly the data collection approaches
The case studies reveal considerable diversity in the institutional and governance arrangements underpinning data collection approaches, which in turn are likely to influence the results and their reliability.
One dimension of this diversity is the degree of historical continuity. The Nordic countries are characterised by long-standing, stable data collection systems. Finland began collecting homelessness data roughly four decades ago, and Denmark has conducted its homelessness mapping consistently since 2007. Other countries have introduced their approaches more recently. Canada launched its first Everyone Counts survey in 2016, and Germany’s Homelessness Reporting Act dates from 2020. Long-standing systems benefit from institutional continuity, established processes, and accumulated expertise, which may support higher response rates and enables more robust monitoring of trends over time. However, longevity can also produce path dependency, making it difficult to adapt the approach in response to structural changes, as illustrated by the challenges associated with the shift in register ownership in Finland.
Dedicated legal frameworks mandating homelessness data collection are the exception rather than the rule among the country practices examined. Germany’s Homelessness Reporting Act is a notable outlier in this regard. In Scotland, the statutory regulation of local authority homelessness services entails mandatory reporting through the case management system. Where a population census is used to measure homelessness, the relevant legal mandate applies to the census as a whole rather than to homelessness measurement specifically. In Canada, Denmark, Finland and Czechia, participation in data collection is voluntary, though it is typically incentivised through linkages to funding for grants relating to homelessness policy. The Danish case is instructive in this sense: following legislative changes to the municipal funding mechanism for homelessness services, reporting among some units increased, suggesting that such linkages can serve as an effective lever for improving coverage.
The case studies also highlight considerable diversity in the actors responsible for homelessness measurement. In some countries – including Germany, Denmark and Czechia – data collection is commissioned by ministries and outsourced to research institutes. While this model offers flexibility and may facilitate data sharing and processing through clear agreements (see Box 4.1), it carries risks of discontinuity and knowledge loss in the event of a change in provider. In other contexts, data collection is more institutionally integrated: in Finland, it sits within a public agency under the Ministry of Environment, with a high degree of required municipal participation. In Canada, data collection also relies heavily on local community entities. In such contexts, perceived relevance at the local level is a key determinant of participation and ultimately quality of estimates. As the Finnish example illustrates, this relevance can be reinforced where clear links are established between the data and local policy developments.
The involvement of multiple actors across different levels also makes data sharing, integration, and processing and the respective frameworks a recurring implementation challenge. Access to administrative data is frequently constrained by legal, technical, or institutional barriers; even in countries with relatively advanced register systems, such as Finland. A country’s broader digital infrastructure also appears to influence both the quality of estimates and the scope for deeper analysis. Systems built around personal identification numbers, as in Finland and Denmark, can reduce the risk of double counting. In Denmark specifically, strict but enabling data linkage rules allow for the integration of homelessness data with administrative records, supporting more detailed analysis regarding the experiences of homelessness.
No method can generate an exact count of the extent of homelessness – but, when consistently implemented, it can help policymakers monitor trends
No single data collection approach can provide a complete or exact picture of the full extent of homelessness. Both structural limitations inherent to a given method and specific implementation challenges may contribute to an undercount or inaccurate estimate.
Most approaches reviewed in this study are point-in-time in nature, capturing homelessness at a single moment rather than over a longer period. Denmark, for example, has estimated that the number of people experiencing homelessness over the course of a year would be at least double the point-in-time figure recorded during the survey week.
Methods that rely on service contact – including service-based surveys (Denmark, Germany), administrative data from service providers or local authorities (Finland), or case management systems (Scotland) – are by definition limited to individuals who are in contact with relevant service providers or local authorities at the time of enumeration. All countries using such approaches acknowledge in their reporting that figures should be interpreted as lower-bound estimates. The accuracy of these estimates is further affected by the willingness of service providers and local authorities to respond to and/or engage with the surveys, which appears to be constrained by growing administrative reporting burdens. None of the countries quantify the extent of undercounting in terms of people experiencing homelessness who are not captured in the resulting estimates.
Street counts present analogous limitations. In Canada, it is explicitly acknowledged that the PiT Count cannot provide a reliable estimate of the total extent of homelessness; the Count is framed as a snapshot with clearly defined boundaries, which are also further affected by local flexibility in implementing the count. In New Zealand, the identification of only 330 rough sleepers through Census field operations is likely to also represent a substantial undercount (Table 2.4).
Approaches that derive hidden homelessness figures from population censuses or household surveys, as instruments not originally designed to capture homelessness, rely on proxy variables and assumptions that may distort the resulting estimates. The high level of uncertainty is illustrated by the plus/minus 50% confidence interval around the estimate of 11 000 people identified as ETHOS Light 6 through extrapolation from EU-SILC data in Czechia.
Nonetheless, the consistent presence of uncertainty across approaches does not necessarily diminish the value of the data produced. Particularly where collection methods are consistent over time, they can support the analysis of longer-term trends and provide an evidence base for policy objectives, including the monitoring of policy effectiveness, the identification of diverse profiles among people experiencing homelessness, and the recognition of their lived realities.
Box 4.2 outlines a set of methodological trade‑offs that are important consideration for any government reviewing its approach to measuring unsheltered homelessness and temporary stays with family or friends.
Box 4.2. Measuring unsheltered and hidden homelessness involves important methodological trade‑offs
Copy link to Box 4.2. Measuring unsheltered and hidden homelessness involves important methodological trade‑offsThe case studies and the patterns emerging across methodological approaches highlighted in this report imply a set of cross-cutting trade‑offs when deciding how to measure unsheltered homelessness (ETHOS Light 1) and temporary stays with family or friends (ETHOS Light 6). These include:
Direct vs. indirect measurement: Methods range from direct enumeration – such as street counts or service‑provider surveys – to indirect estimation based on assumptions or proxy variables applied to existing data sources that were not originally designed to assess homelessness. This distinction carries corresponding trade‑offs in precision and coverage. Direct methods can only capture what is visible within a defined window (the night of the count or the week of the survey) and while likely accurate for the individuals recorded, they will rarely achieve full coverage. Indirect approaches offer the additional potential to identify a broader range of people experiencing homelessness through existing sources, beyond direct contact with those who are visible or accessible during the specific enumeration window. However, the assumptions required as part of the estimation process may distort the results.
Timeliness vs. comprehensiveness: More frequent data collection tends to sacrifice depth, while more comprehensive approaches are typically conducted at longer intervals. For instance, Finland’s annual report is more concise than Germany’s or Denmark’s biennial reports but also contains much less detailed socio-demographic information on people experiencing homelessness. Both patterns can also be observed within individual countries. Canada’s Everyone Counts data collection approach combines an annual enumeration focussed solely on enumeration, with a triennial survey designed to generate more detailed information about individual profiles.
Complexity vs. operational feasibility: Increased methodological complexity can improve coverage – for example, by complementing service-based methods with administrative or register data, or by applying multi-step algorithmic processes to identify people in more hidden forms of homelessness as part of census operations. More extensive surveys can also yield richer profiling information. However, complexity risks lower participation among reporting entities, higher costs, and implementation challenges that may affect reliability, such as the need to make calibration decisions in highly sophisticated identification systems.
Standardisation vs. flexibility: Approaches that permit local flexibility – such as Canada, where communities independently determine the geographic scope and method for the street count most appropriate to their context, or Finland, where municipalities determine which administrative data sources to draw on – are likely to increase participation and may be essential for achieving high response rates in approaches that rely on full national coverage. However, such flexibility comes at a cost to comparability and consistency across reporting units. Stricter standardisation, as in Germany’s three‑stage sampling procedure, limits this variation but may reduce a more precise representation of local conditions.
Continuity vs. development: As outlined above, long-standing systems benefit from institutional continuity, established processes, and accumulated expertise, which can support higher response rates, data reliability, and more robust trend monitoring over longer periods of time. At the same time, the capacity to adapt methodological approaches is important, particularly as methods to measure more hidden forms of homelessness continue to evolve. Continuity can also produce a certain path dependency, making it difficult to revise well-established approaches, even when improvements are warranted.
Recommendations to improve Germany’s measurement approach
Copy link to Recommendations to improve Germany’s measurement approachIn recent years, Germany has made significant progress to strengthen homelessness measurement. The 2020 Homelessness Reporting Act established a solid legal foundation for systematic data collection and created a framework for expanding coverage beyond sheltered homelessness. The introduction of biennial supplementary reporting on unsheltered homelessness and people staying temporarily with family or friends reflects a clear commitment to improving the evidence base on less visible forms of homelessness. Overall, Germany’s data collection approach can be considered a good international practice. Its broad service-based design, which mobilises institutions well beyond the homelessness sector, increases the likelihood of measuring ETHOS Light 6 in particular, as well as the heterogeneity of living arrangements of people experiencing homelessness.
At the same time, the review of international practices (Chapter 3) highlights opportunities for the German Government to improve clarity, coverage, implementation and long-term sustainability. This chapter proposes a set of recommendations to build on Germany’s existing strengths. The following recommendations distinguish between short-term improvements to the existing data collection process and medium- to long-term reforms aimed at developing a more integrated measurement system.
Short-term recommendations: Improve clarity, usability and implementation of data collection tools
Clarify the policy purpose of each data collection instrument to determine whether they suffice to meet policy objectives
Each homelessness data collection instrument should be linked to a clearly defined policy purpose. While the Homelessness Reporting Act identifies the objective of strengthening the evidence base for political action, the specific role of data collection could be more clearly articulated.
Different policy objectives require different types of data collection. If the primary objective is to estimate the number of people experiencing homelessness, assess and monitor changes over time, regular enumeration using a more focussed, yet consistent methodology may be sufficient. By contrast, understanding the profiles of people experiencing homelessness, how people enter and exit homelessness, and how their needs and circumstances differ requires more detailed information on individual profiles, experiences and pathways. These analytical objectives may therefore call for richer survey instruments or separate thematic studies, which can be conducted less frequently. Some countries separate these functions. Canada, for example, combines annual enumeration focussed on measuring the scale of homelessness and monitoring trends with a more detailed survey conducted every three years to inform policy development and service planning.
A clearer articulation of policy objectives would not only help guide methodological choices, but could also strengthen the perceived value of participation among involved stakeholders, such as service providers and authorities. Linking findings more systematically to policy discussions and follow-up government actions leverages the potential of homelessness measurement as an active steering tool for the policy area, as illustrated by the Nordic countries’ practices.
Simplify the biennial survey
The biennial supplementary survey – which is conducted by a research institute to report on a broad range of data and information on people experiencing homelessness – is methodologically ambitious and generates rich information on people experiencing homelessness. At the same time, its complexity creates reporting and processing burdens for the commissioned research institutes, as well as participating service providers.
The German authorities could assess whether the current questionnaire can be streamlined, while preserving the information most relevant to its policy objectives (as discussed above). An option worth exploring is whether thematic modules on specific topics, such as pathways into homelessness, are best integrated into the core survey or could instead be addressed through separate thematic studies (e.g. additional study on trajectories of people experiencing homelessness).
International experience suggests that simpler survey instruments can generate higher participation rates without necessarily reducing analytical value. Denmark, for example, collects a wide range of information through a comparatively concise reporting form (2 pages). A more focussed questionnaire could help reduce respondent burden, improve completion rates and simplify data processing.
Improve the reporting process for service providers, including further digitalisation
Given the central role of service providers in the German approach, reducing reporting burden should remain a priority. Participation in the supplementary survey requires additional efforts from organisations that are often already operating under considerable resource and capacity constraints.
The German authorities could review whether elements of the current reporting process can be simplified and further digitalised. In particular, moving progressively towards digital reporting would reduce administrative workload and facilitate data processing. The Danish model demonstrates how a digital reporting platform can support data collection, simplifying data submission for participating service providers and authorities.
Germany could also consider allowing trained staff to complete questionnaires on behalf of people experiencing homelessness where appropriate. This approach is used in Denmark and is also possible as part of the method promoted through the European Homelessness Counts project (see Box 2.1). This may not only simplify survey completion for staff, but also reduce selection bias associated with voluntary self-completion and improve coverage of people who face barriers to participating directly.
Review the timing and the reference period of the biennial survey
The current one‑week reference period provides a clear point-in-time estimate. However, it may contribute to undercounting, particularly among people staying with family or friends who are less likely to be in contact with services during a specific week.
The German authorities could review whether adjustments to the timing, duration and/or reporting arrangements could improve coverage. This could include assessing whether the current survey week is optimal from an operational perspective and whether contextual factors, such as monthly income cycles (including social assistance benefits) or school holidays, affect service use and reporting patterns.
The experience of Denmark and the European Homelessness Counts pilot project suggests that allowing services to report on behalf of individuals they have recent knowledge of, even if no direct contact occurs during the reference week, may improve coverage of less visible forms of homelessness. Any such adjustment would require clear rules regarding the recency of contact and verification of housing circumstances.
In addition, allowing a longer submission window following the reference period – as done in Denmark, where the submission window is around three weeks – may help reduce pressure on reporting institutions and improve response rates.
Medium-term to long-term recommendations: Improve coverage and data linkages to build an integrated measurement system
Evaluate the pilot use of administrative data to complement service-based reporting
The pilot use of administrative data in the 2026 supplementary reporting represents a promising development, particularly as a means to improve the measurement of people staying temporarily with family or friends.
Administrative data likely has to be viewed as a complement rather than a replacement for service-based reporting. International experience suggests that no single administrative source is likely to provide a comprehensive estimate of homelessness. Finland’s long-standing approach demonstrates both the potential and limitations of administrative data: useful results can be achieved, but only through the combination of multiple sources.
The German authorities could continue to explore how administrative data from job centres, social welfare offices and other public institutions can complement existing survey-based estimates. Particular attention should be given to identifying simple and transparent indicators of housing exclusion and homelessness that could improve the detection of homelessness within existing administrative systems.
Establish clearer data-sharing arrangements between involved actors
Effective homelessness measurement increasingly depends on the ability to combine data and information across different institutions. As Germany expands the use of administrative data, clear frameworks for data sharing, processing and governance will become even more important.
The German authorities should develop clearer guidance on when and how municipalities, job centres, social welfare offices and service providers can share data for homelessness data collection purposes. Greater clarity regarding legal bases, responsibilities and data protection requirements could reduce institutional uncertainty and improve participation.
International experience highlights that standards of the common GDPR framework are interpreted differently, and that most frequently challenges arise from uncertainty and inconsistency about how rules should be interpreted and applied in practice across institutions.
Assess whether survey frequency and content should be adjusted
Germany currently conducts supplementary reporting every two years, making it one of the more frequent service-based homelessness data collections internationally, compared with an international average of around three years. Over time, Germany could review whether this frequency continues to provide sufficient added value relative to the reporting burden on participating services and authorities.
This assessment could distinguish between different data collection purposes. Regular counting exercises may need to be conducted relatively frequently to monitor changes in the scale and profile of homelessness, while more detailed thematic surveys may be carried out at longer intervals. As proposed in the previous recommendation, this could involve a lighter recurring survey covering headline figures and core demographic information, complemented by less frequent thematic studies examining specific policy questions in greater depth.
Develop a long-term data strategy for homelessness measurement, including indicators to monitor progress
Germany has already demonstrated a strong institutional commitment to homelessness measurement through the Homelessness Reporting Act and the expansion of reporting beyond sheltered homelessness – this approach could provide inspiration to many other OECD and EU countries. The next step for Germany is to translate these efforts into a long-term measurement strategy.
Such a strategy could define measurement priorities, clarify the respective roles of different data sources and opportunities for their integration, establish principles for methodological development, and set out arrangements for follow-up and periodic review.
International experience, particularly from the Nordic countries, suggests that measurement continuity is an important asset. Consistent measurement approaches support institutional learning, strengthen participation and enable meaningful analysis of trends over time. The challenge will be to preserve sufficient methodological continuity while continuing to improve coverage, data quality and the relevance of the information collected.
In addition, alongside headline estimates of the number of people experiencing homelessness, Germany could consider developing a set of indicators to monitor progress in specific policy areas. These indicators could help explain the mechanisms and dynamics underlying homelessness, assess how policies and services are functioning, and identify where further action is needed.
Relevant indicators could cover areas such as prevention, access to services, housing outcomes, health, employment and sustained exits from homelessness. This work could draw on the forthcoming OECD Blueprint for Integrated Action Plans for Homelessness and Housing Exclusion.
References
[2] Benjaminsen, L. et al. (2020), “Measurement of Homelessness in the Nordic Countries”, European Journal of Homelessness, Vol. 14/3, pp. 149-169, https://www.feantsa.org/files/Observatory/Journals/Volume-14/V14-3/EJH_14-3_A7_web2.pdf.
[3] Dyb, E. et al. (2021), “Governing Through Definitions and Numbers: Analysis of the Nordic Homeless Registrations as Instruments of Governing Homelessness”, European Journal of Homelessness, Vol. 15/3, pp. 187-201, https://www.feantsa.org/files/Observatory/Journals/Volume-15/V15-3/EJH_15-3_A13.pdf.
[1] Hermans, K. et al. (2026), European Homelessness Counts Final Report Year 2, https://www.kuleuven.be/lucas/en/images/eu-homelessness-counts-final-report-2026.pdf.