Professor Eoin O'Sullivan:

I have prepared a detailed statement that we can take as read but I will take the members through the document.

The homelessness policy statement was issued by the then Government in February 2013. There is a clear acknowledgement in that document that the extent of homelessness in Ireland must be quantified with confidence in order that realistic and practical solutions can be brought forward. Arising from that, three sets of data were generated by the local authorities on behalf of the Department, the first of which was known as performance reports, which talked to what Mr. Murphy mentioned about the dynamics of homelessness. This data, produced on a quarterly basis since Quarter 1 2014, look at the number of new and repeat presentations to homelessness, the numbers of adults in emergency accommodation for more than six months, and the number of exits from homelessness. I provide some data in the document to give members an illustration of the type of data provided via these quarterly performance reports up to quarter 2 2018.

Second, quarterly financial reports were required to be produced by the lead authorities. The expenditure on homeless services under five categories was identified. The concern at the time was that approximately 50% of expenditure was going to emergency accommodation and that expenditure should be diverted instead to preventative services. In practice, however, that has not happened. As I point out, in the first half of this year, 80% of all expenditure is on emergency services.

Third, these monthly reports extracted from the Pathway Accommodation & Support System, PASS, which had been established in Dublin in 2011 as a bed management and client support system, was rolled out nationally in 2013. From April 2014, data on the number of adults, the number of child dependants, the type of accommodation, their age and gender have been produced on a monthly basis.

I felt it was important to go into some detail on the origins of those reports. They came from the national homeless consultative committee, NHCC, and the cross departmental team on homelessness, CDT. In 2013 and 2014, there was a recognition that there was no national data on homelessness and in terms of the implementation of the national homeless strategy, some baseline data were required to measure progress on ending homelessness. There had been an exercise in Dublin known as Counted In, which had been taking place every three years from 1999 but had been discontinued after 2008. There were a range of difficulties with that but particularly the fact that it was only taking place every three years, which rendered it problematic for providing timely data on progress on ending homelessness.

A data subgroup of the NHCC and the CDT was then established to examine this issue. Following a number of meetings, it was agreed and recommended to the overall committee that with the national roll-out of the PASS, it was possible to extract timely data on the number of adults in emergency accommodation. Initially, we proposed that the data be produced every six months. The chair of the committee, an assistant secretary of the Department, on the basis of the Department reporting to the Cabinet sub-committee on social policy every quarter, suggested it be reduced to three months and then later to one month.

There was a clear acknowledgement, which comes up regularly, that the monthly figure is not comprehensive. It excludes four particular categories, which we examined in detail. Regarding non-section 10 funded services, there are a small number of services which, for various reasons, neither receive nor seek section 10 funding. We conducted research with the Housing Agency, which concluded that there were fewer than 200 beds nationally. In terms of establishing a baseline figure, therefore, we were aware of that limitation but the numbers were not that significant.

The monthly figures do not collect data on rough sleeping but there was an alternative source of data via the twice yearly count in Dublin. At the time, we contacted all the other local authorities, apart from Cork, Galway and Limerick, rough sleeping was not an issue in their areas. We were fairly confident, therefore, that we had an accurate minimum figure for Dublin and an estimate from outside Dublin.

The monthly data does not collate data on the hidden homeless but, again, we felt that the housing needs assessments collected that data, so there was an alternative source of data in place. We also did not recommend that long-term supported accommodation, which is funded by section 10 of the Housing Act 1988, be included in the monthly figures. That was reinforced in census 2016 when the homeless methodology liaison group also recommended to the CSO that this group be excluded from its overall count of homelessness. There was a clear acknowledgement and understanding of the limitations of the monthly data but, nonetheless, there was a consensus that it provided timely useful data on those in emergency accommodation.

I will go through the census, the past data and the social housing needs to suggest that the past data are a robust measure of homelessness.

It is worth noting that some of this confusion about the data relates to the fact that in some accounts people talk about households, some about adults, and some bout adults and children. For example, in September, based on the monthly report, there were there 5,202 households in emergency accommodation, or 5,869 adults or 9,698 adults and children, and often these different categories are used interchangeably. There have been two significant modifications to the monthly data, the first on 1 January 2015, when accommodation or refuges for those escaping from gender-based violence, funded via section 10, were removed. This followed a recommendation of the homelessness oversight group in 2013 that they should be removed and that these agencies would be funded by Tusla. That was broadly supported by the national homelessness consultative committee and cross-departmental team and the data sub-group. The second was that in March, April and July of this year, approximately 625 adults, with 981 accompanying child dependants, were excluded from the monthly reports. I have provided several statements from the Minister on that fact and there is a document before the committee from the Department. I have presented one chart which gives the monthly data, excluding the domestic violence shelters and the households excluded in March, April and June 2018, and a second chart which presents the data for each September from 2013 onward and which shows what the figures would look like in September 2018 if those two sets of data had not been removed.

Two modifications have been made to the data. The first, the removal of refuges, did not undermine confidence in the data because there was a clear rationale and logic for its removal. The second modification has created some confusion and undermined confidence in the data because it is unclear what the criteria are for removing these households. It is not clear whether it is the legal basis of the residence or the physical characteristic of the residence that is the determining factor. Hybrid accommodation situations have emerged in recent years that neither the Housing Act 1988 nor the data sub-group of the national homelessness sub-committee anticipated. It would be helpful to spell out in greater detail the criteria utilised and the rationale for the removal of 625 adults from the monthly reports in 2018 to ensure confidence in the reports.