The Effectiveness and CostEffectiveness of Homelessness Prevention and Assistance Programs: A Report

1
The Effectiveness and Cost-Effectiveness of
Homelessness Prevention and Assistance Programs
A Report of Preliminary Findings

• Australian Social Policy Conference
• Sydney 2007
• Paul Flatau
• Murdoch University

2
Conclusion

• Programs produce positive outcomes for clients
  and do so at low cost of delivery relative to the
  delivery of other services.
• If homelessness programs were able to reduce the
  utilisation of health and justice facilities by
  clients of homelessness programs down to
  population rates of utilisation the savings
  achieved would pay for the homelessness programs
  several times over.
• This suggests that there is potential for
  homelessness programs to be dramatically
  cost-effective.

3
Starting point

• AHURI 2005 Priority Research Question
• What are the whole of government costs and
  benefits of not preventing homelessness
  including, for example, in relation to health,
  crisis accommodation, policing, criminal justice,
  and housing assistance?

4
Our Key Research Questions

• What is the effectiveness of homelessness
  prevention and assistance programs in WA?
• Effectiveness Differential outcomes achieved by
  clients relative to needs and homelessness
  histories
• What is the cost-effectiveness of homelessness
  prevention and assistance programs in WA?
• Cost Effectiveness Differential outcomes
  achieved by clients relative to differential
  costs
• Taking into account the cost offsets of such
  programs whole-of-government budget savings
  achieved in health, crisis accommodation,
  policing, criminal justice, and housing assistance

5
Differential Costs and Outcomes

• Compare the outcomes and costs of providing
  homelessness-based programs (Actual)
• With
• The outcomes and costs incurred and outcomes
  achieved if these programs and services were not
  provided (Counterfactual)

6
Cost-Effectiveness Analysis
Cost Difference (difference between homelessness
program and no program)
Better outcomes from Homelessness Programs
Outcomes Difference
(difference between homelessness program and no
program)
Lower costs involved in Intervention
7
Cost-Effectiveness Outcomes
Cost Difference (difference between homelessness
program and no program)
Homelessness Programs more costly and less
effective
Homelessness Programs more costly but more
effective
Outcomes Difference
Homelessness Programs less costly and less
effective
Homelessness Programs less costly and more
effective
8
Cost-Effectiveness Outcomes
Cost Difference (difference between homelessness
program and no program)
Least Desirable Quadrant
Standard Quadrant
Outcomes Difference
Dominant Quadrant
9
Illustration
Consider a program to provide secure supported
accommodation for otherwise homeless people with
mental health conditions
Homelessness Program
Counterfactual No Program

• Potential Costs
• E.g., Significant costs associated with
  utilisation of acute psychiatric units, emergency
  departments, police and justice facilities
• Potential Outcomes
• E.g., Worsening of condition, no employment
  prospects, poor quality of life

• Potential Costs
• E.g., Significant capital investment in terms of
  dwellings and recurrent expenditure on staff and
  other resources
• Potential Outcomes
• E.g., Stabilisation of condition, employment
  options, improved quality of life

10
Scoping the project

• Where?
• WA Perth, South-West, Southern
• Programs?
• WA Homelessness Prevention Programs
• the Community Transitional Accommodation and
  Support Service (TASS) and the Re-entry Link
  program
• Designed to assist prisoners re-enter into the
  community on release
• the Supported Housing Assistance Program (SHAP)
  and Private Rental Support and Advocacy Program
• Designed to assist public and private tenants
  maintain their tenancies and,
• SAAP/CAP

11
Research Design

• Agency and Program Collaboration
• Working with agencies and programs
• Project Advisory Group
• Quantitative Analysis
• Background, needs and outcomes of clients
• Client Survey Wave 1 post-entry Waves 2/3
  3-month/exit survey Wave 4 12 month point
• Community Centre Survey One-off survey
• Administrative data sets
• Cost analysis
• Program funding information
• Agency Cost Survey
• Cost offset service utilisation outcome data
  from the Client Survey and the
• Qualitative Analysis

12
Needs Homelessness Histories

• Multi-dimensional approach to client needs
• Case worker assessed needs of clients
• Instability in early family environments
• Experiences of homelessness and unsafe living
  environments prior to the age of 18 after the
  age of 18 and in the year prior to support
• Self/caseworker assessed client experiences of
  mental and long-term physical health conditions,
  client concerns about own alcohol and drug use.
• Quality of life WHOQoL-BREF (Australian version)

13
Outcomes Measurement

• Outcomes are assessed across a range of
  dimensions
• Changes in client status measurese.g., labour
  force, level of income, income source, education
  and training, and housing changes in the level
  of capability to manage circumstances and needs
• Changes in the utilisation of homelessness
  program services and non-homelessness program
  services
• Changes in self-assessed satisfaction with
  various dimensions of life, knowledge gained as a
  consequence of support, and quality of life
  outcomes.
• Program-specific client outcome indicators e.g.,
  in the case of tenant support programs, the
  reduction in debt levels to housing authorities.
• Outcomes are assessed over time Immediately
  following support three months exit and 12
  months

14
Funding Costs

• Recurrent and capital funding per client across
  different homelessness programs (data drawn from
  administrative sources)
• Better measure would be full-time equivalent
  load. Problems in obtaining information to derive
  estimates of full-time equivalent load.
• Cross-program comparisons of funding per client
  problematic
• Different methods for collecting and reporting on
  client numbers leading to potential differences
  in counts of the number of clients simply because
  of these different methods.
• Differences exist across programs with respect to
  the average duration of support, the rate of
  capacity utilisation, and client needs.

15
Funding Costs

• Agency cost analysis based on Agency Cost Survey
• The gross funds available for service delivery
  and the source of these funds
• In addition to government funding, providers of
  services raise income via other grants and
  donations and operating income from rent and
  other sources (e.g., vending machines).
• Ongoing costs involved in providing accommodation
  and support to clients
• The unit cost of providing accommodation and
  support to clients.

16
Cost Offsets

• Three approaches to estimating cost offsets
• Impute the cost of the client groups use of
  government services and compare this with the
  population in general. Differential is cost
  offset
• Within each client group, do the above analysis
  comparing cost offsets for those who experienced
  homelessness (or unsafe living) in the prior
  year with those who didnt
• Impute the cost of the clients group immediate
  past use of government services and compare this
  with the imputed cost of government services post
  the provision of support. Differential is the
  cost offset.
• To determine the value of cost offsets, the unit
  costs of delivering a range of health and justice
  services is estimated and applied in conjunction
  with prevalence indicators of service utilisation
  by the various client cohorts and for the
  population in general.

17
Where we have got to

• Evidence gathered to this point
• Program administrative data outcomes and funding
  information
• Wave 1 of Client Survey and most of Waves 2/3
  tabulated
• Community Centre Survey
• Partial analysis of Agency Cost Survey
• Population-based estimates of cost offsets
• 12 Month Follow-up Begun

18
Homelessness Histories
19
Client Needs

• Half of respondent clients currently experience a
  mental health condition
• Close to a third of all respondent clients
  experience a long-term physical health condition
  and around a fifth of all respondent clients
  express concerns about their own alcohol and drug
  use.
• Clients typically present to homelessness
  services with multiple needs requiring ongoing or
  intensive support.
• Clients self-esteem, own-assessed quality of
  life and satisfaction with various dimension of
  life is very low relative to Australian norms.

20
(No Transcript)
21
(No Transcript)
22
Accommodation

• SAAP National Collection Data - a high proportion
  of those who enter from a position of primary
  homelessness leave to secure permanent
  accommodation.
• In the Client Survey those who were in primary
  homelessness on entry or who were living in
  temporary accommodation at that time remained
  housed through the survey follow-up period.
• The majority of those assisted in homelessness
  prevention programs (SHAP PRSAP) retained their
  housing and partially or fully resolved the
  immediate housing problems that brought about the
  initial referral (e.g., tenant liabilities,
  rental arrears).

23
Re-entry to the Community

• Programs supporting those leaving prison provide
  housing to prisoners - Accommodation support
  provides a critical element of stability for
  clients and enables them more effectively to
  reintegrate into the community.
• The early evidence from the TASS and Re-entry
  Link programs is that these programs are proving
  beneficial in lowering rates of recidivism and a
  making positive contribution to the lives of
  individuals who have previously returned to
  prison.

24
Client Satisfaction

• Client Survey study participants reported a
  significant improvement in satisfaction with
  housing outcomes following the provision of
  support
• More than half of all Client Survey respondents
  reported that their housing position was much
  better than before assistance was forthcoming. A
  further quarter indicated that their housing
  position was somewhat better than before
  assistance was provided
• Agencies provide an environment, which improves
  clients perception of safety
• 61.1 per cent of clients in the Client Survey
  reported that assistance had resulted in improved
  feelings of safety. SAAP-DV and Single Women
  category of clients 87.1 per cent of clients
  indicated that assistance had resulted in an
  improvement in feelings of safety.
• One area where Client Survey study participants
  report much lower levels of positive change
  following the provision of support is that of
  employment opportunities.

25
Importance of Receiving Support and Client
Knowledge

• Importance of receiving support
• The vast majority of Client Survey study
  participants (86.1 per cent) indicated that it
  was very important to receive assistance and
  help from the service in meeting their needs. A
  further 12.6 per cent of study participants
  indicated that receiving support was important.
• Client understanding
• 55.6 per cent of study participants in the Client
  Survey indicated that support had resulted in
  them understanding the issues facing them and how
  to deal with them a lot more than before support
  was provided.
• A further 27.8 per cent of clients responded that
  they understood the issues facing them more than
  before (83.4 per cent in total).

26
(No Transcript)
27
Three Month Follow-up Evidence

• Follow-up evidence on client satisfaction
• Client satisfaction with housing, feeling part of
  the community, the neighbourhood and ability to
  cope with serious problems improved significantly
  from the Wave 1 survey point
• Client satisfaction with their financial
  situation, a feeling of safety and overall
  satisfaction improved marginally from the Wave 1
  survey.
• The two areas which did not exhibit an
  improvement in outcomes was employment
  opportunities and own health.
• There was however a small increase in the
  proportion of clients employed since the
  beginning of the support period.

28
Three Month Follow-up Evidence

• Follow-up evidence on quality of life outcomes
• Study participants displayed an improvement in
  the WHOQoL-BREF (Australian version) across all
  four quality of life domains (physical,
  psychological, social relationship and
  environment) from the point of the Wave 1 survey
  through to the follow-up survey.

29
Costs of running programs

• Significant differences between programs in
  recurrent funding per client
• However, estimates presented of government
  funding levels per client should be interpreted
  with some caution and judgments of the relative
  cost of delivering different types of
  homelessness programs made, on an unadjusted per
  client basis, are ill-advised.
• First, data collection methods exist between the
  various programs leading to potential differences
  in counts of the number of clients simply because
  of these different methods.
• Second, even if the same rules of defining and
  measuring clients and the same methods used to
  collect data were adopted across all homelessness
  programs, cross-program comparisons of funding
  per client are problematic because differences
  exist between programs with respect to the
  average duration of support, the rate of capacity
  utilisation, and client needs.

30
(No Transcript)
31
Costs of running programs

• Agency cost and expenditure structures
• In the case of SAAP Crisis/short-term services,
  on average, 74.3 per cent of total income is
  derived from government funding. The major cost
  component relates to staff costs, accounting for
  62.0 per cent of costs overall.
• Funding for SHAP services is predominantly
  program specific government funding, accounting
  for 98.4 per cent of all funding.

32
Cost offsets Community Centre Clients

• On average, the per annum cost of health services
  for a community centre client is 10,217/person
  greater than the population average, and higher
  for every service considered. The associated
  average life outcome is 250,544/person.
• The average cost of justice services for a
  community centre client is 3,810/person/year
  greater than the population average, with an
  associated average life outcome of
  93,414/person.

33
Cost offsets Client Survey

• For all programs, the average cost of both health
  and justice services used by clients exceed the
  population average. The total potential cost
  offset ranges between 7,647/person/year for
  PRSAP clients to 39,690/person/year for
  TASS/Re-entry Link clients.
• The associated average life outcomes range
  between 188,846/person for PRSAP clients to
  1,141,948/person for TASS/Re-entry Link clients.
  
• For all programs except TASS/Re-entry Link over
  two thirds of the cost difference relates to
  health services.

34
Cost offsets

• For all programs, the value of annual population
  offsets' is at least 2.7 times greater than the
  annual program cost, resulting in a significant
  potential net government cost savings from
  providing assistance.