COMMITTEE OF PUBLIC ACCOUNTS debate -
Thursday, 3 May 2012

Vote 4 - Central Statistics Office

Mr. Gerry O’Hanlon (Director General, Central Statistics Office) called and examined.

Chairman

I remind members, witnesses and those in the Visitors Gallery to turn off their mobile telephones as interference affects the sound quality of the transmission of the meeting. I advise witnesses that they are protected by absolute privilege in respect of the evidence they are to give to the committee. If they are directed by the committee to cease giving evidence on a particular matter and they continue to do so, they are entitled thereafter only to a qualified privilege in respect of their evidence. Witnesses are directed that only evidence connected with the subject matter of these proceedings is to be given and are asked to respect the parliamentary practice to the effect that, where possible, they should not criticise or make charges against a Member of either House, a person outside the House or an official by name or in such a way as to make him or her identifiable. Members are reminded of the provisions within Standing Order 158 that the committee shall refrain from inquiring into the merits of a policy or policies of the Government or a Minister of the Government or the merits or objectives of such policies.

I welcome Mr. Gerry O'Hanlon, director general, Central Statistics Office, and I ask him to introduce his officials.

Mr. Gerry O’Hanlon

Thank you, Chairman. I am accompanied by Mr. Aidan Punch, assistant director general for economic and environmental statistics, Ms Elaine Lucy, senior statistician in charge of our IT corporate systems and Mr. Joe Treacy, director of IT and corporate systems in the Central Statistics Office. We are joined by Mr. Dermot Quigley and Ms Ciara Morgan from the Department of Public Expenditure and Reform.

Mr. Seamus McCarthy

Vote 4 provides for the salaries and expenses of the Central Statistics Office which collects, compiles and analyses and disseminates a wide range of statistical information, relating to the economic and social life of the country. Its key outputs include the five yearly census of population, which is a critical reference point for economic and social planning purposes and the national accounts. The appropriation account shows the gross expenditure under Vote 4 in 2010 was €53.5 million. This represented an increase of 6% on the 2009 outturn, including costs associated with preparations for the taking of the 2011 census of population. Pay costs amounted to around €36 million or two thirds of the gross spend. Chapter 19 of the Comptroller and Auditor General's report reviews the development and implementation of a new data management system in the CSO. The system was the central component of the CSO's information technology strategic implementation plan, which was formulated in March 2002. Contractors were engaged to assist the CSO in designing, building and implementing the data management system. The examination looked at all stages of the project, from initial appraisal and planning in 2003 to the commissioning of the system in 2007 and its utilisation since then. The overall conclusion of the examination was that the project had delivered on its core technical objectives of modernising the CSO information technology environment, facilitating centralised data storage and providing a set of survey processing applications available for use on existing and new surveys. The CSO's strategic plan proposed a number of IT projects, including the data management system. Together, these were expected to deliver a set of business benefits. The plan included estimated costs for individual projects but the benefits from the individual projects were not separately identified. The examination found that no focused business case or appraisal was prepared in respect of the data management system project and that detailed consideration of potential alternative approaches was not documented.

The estimated cost for the project was just over €10 million, which did not include any internal CSO costs that would inevitably arise. Ultimately, payments related to the project amounted to approximately €13.2 million, including €9.2 million for the main development and support contract and software and hardware costs of approximately €3 million. Internal staff resources attributable to the project were subsequently costed at approximately €5 million, bringing the total project cost to approximately €18 million.

The outturn costs for the development and support contract was 55% higher than originally expected for two reasons. First, changes and enhancements agreed during pilot implementation added €2.4 million to the cost. This suggested that the original systems specifications were not as comprehensive as they could have been. In addition, there was an agreed extension of the post-implementation support and maintenance period from 12 months to 21 months at an extra cost of €880,000. The CSO complied with the relevant procurement rules in relation to the project.

The system went live in September 2007. This represented a 16 month overrun on the original schedule of 30 months. As the timing of the payments to the contractors was linked to delivery, several of the scheduled payments were deferred as a result of the delays. Maintenance and support provision for the contractors continued until May 2009.

The examination noted some issues relating to the use of the new system. As at 1 June 2011, only 30 of the 73 surveys falling within the scope of the project had migrated to the new system. Consequently, there was a continuing need for parallel running of the pre-existing systems which had overlapping functionality with the new system and consequent support costs. The Accounting Officer will be able to provide an update on the migration of surveys to the system.

The original concept for the project was for an end to end statistical processing system. The examination found that three of the ten survey processing applications that were developed had been discontinued because they were either not being used or were being used only to a limited extent. Overall, the examination found that the CSO applied the expected project management and implementation processes, including the carrying out of a thorough post-project review which documented the lessons learned. This is an important phase in the process which should help to ensure better project planning and management in future.

Chairman

Thank you. Mr. O'Hanlon, would you like to give your opening statement, please?

Mr. Gerry O’Hanlon

I thank the Chairman and members for affording me the opportunity to speak about the Central Statistics Office Appropriation Account for 2010 and related matters.

It is relevant to note that the activity programme of the CSO follows a cyclical pattern dictated by periodic statistical requirements. In the paper circulated to the committee, this cyclical pattern is illustrated in a graph showing the spending profile from 2005 to 2011. The CSO's budget peaks in census years – most recently 2006 and 2011, as is clear from the graph. In 2010, preparations for Census 2011 were well under way. This included building the IT processing system for scanning the census forms; printing of the forms themselves; recruitment and training of field supervisory staff; and other preparations for census collection and the subsequent processing operation.

The census took place in April last year. The preliminary results were published in June and processing of the forms was completed in December ahead of schedule. We are now seeing the fruits of this work in the current series of detailed census reports, which will continue through 2012. The year 2010 also saw the collection of the household budget survey and census of agriculture. The former is undertaken every five years and the latter every ten years. The results of both surveys were published in spring 2012.

With these cyclical activities in addition to our core statistical programme, spending by the CSO in 2010 was 6% higher than in 2009. The office's gross allocation in 2010 was slightly more than €60 million, while the outturn for the year was €53.538 million, a saving of approximately €6.5 million. These savings were achieved without cutting back on our core programme of statistical outputs and indicators. About half the amount saved related to the three major cyclical projects to which I referred.

In 2011, we implemented further savings of €3.5 million in the costs of collecting and processing the census. These were significant once off savings in the CSO's largest cyclical project. Our gross expenditure in 2011 was €74.7 million and, following the census operation, the CSO's allocation for 2012 is €45.4 million, which is close to our ongoing core requirement.

Like all Departments and offices we have to work within resource constraints. Our priority for 2012 is to deliver the core annual statistical programme, which is needed to inform national policy and meet European Union and other international obligations, while keeping a strong focus on continuous improvement and cost reduction. This is reflected in our action plans under the public service reform programme and Croke Park agreement.

In 2012, in addition to the census publication programme, the CSO will publish some 240 regular statistical releases and publications. All our outputs are available online, including the detailed tables on our StatBank service, and the StatCentral site for statistics published by other Departments and agencies.

Chapter 19 of the Comptroller and Auditor General's report for 2010 deals with the development of the Central Statistics Office's comprehensive data management system, DMS, from 2002 to 2009. The DMS project comprised an ambitious but necessary agenda to replace old IT systems, which had become dangerously obsolescent, with a unified and modern IT system for the collection and processing of statistics. By and large, we accept the findings of Chapter 19, which reflect the CSO's own closure report for the project in which we were frank about weaknesses as well as strengths. I emphasise that all stages of the project were managed with probity and integrity and with an emphasis on delivering a robust IT system which would be of value in the long run. I suggest also, perhaps contrary to inferences from the report, that it the project came in broadly in line with what I would term indicative budgetary estimates. I can come back to that issue.

The resultant data management system is now in everyday use in the CSO's Dublin and Cork offices. The system is maintained completely by the CSO's own IT staff, without ongoing consultancy costs and provides an integrated central survey processing system for CSO statistical areas. Importantly, the DMS makes it easier than before to implement new surveys and the system is designed to support more efficient organisation of statistical collection and processing.

In conclusion, I will mention the five high level goals which the Central Statistics Office has set in its strategy for 2012 to 2014. I referred earlier to two of these, namely, meeting statistical needs and increasing efficiency. In addition, we are committed to promoting greater awareness and use of statistics, developing greater use of data in the wider Irish statistical system and reducing the burden on data providers. I will be happy to expand on these points or on any other questions the committee wishes to raise.

Chairman

Thank you, Mr. O'Hanlon. May we publish your statement?

Mr. Gerry O’Hanlon

Yes.

Deputy Derek Nolan

I welcome Mr. O'Hanlon and his officials and thank him for his opening statement. As a public representative, the information I receive from the Central Statistics Office website, including statistics and reports, is invaluable and I compliment the office on its work in this regard. Unfortunately, the role of the committee is to examine problems and investigate how money was spent, rather than to praise and compliment.

I will concentrate mainly on chapter 19, which relates to the data management system. From what I understand, a strategic plan was introduced in 2003 which involved changing the office's hardware from an obsolete, silo based process where people worked on individual surveys to a more comprehensive, interlinked system, so that people could work on different projects at the same time. Part of that plan was to introduce this data management system. I understand from the report that it has achieved its major goal of replacing the hardware but the expectation in 2003 that it would provide an end to end statistical data system has not materialised.

There were significant cost overruns of approximately 30% from €10 million to €13 million for the data management system. Mr. O'Hanlon said in his opening statement that the budgets were broadly in line with the indicative figures. Can he explain the difference between the conclusion of the Comptroller and Auditor General and his own conclusion?

Mr. Gerry O’Hanlon

The advance work done by Cap Gemini Ernst & Young at the beginning of the project produced an indicative figure of €10.1 million for the design and build of the system. Essentially, it concentrated on the external costs of bringing in contractors to the operation. This figure was generated before any of the detailed specifications were measured or anything else, and we would have treated it as a broad indicative measure of the order of magnitude of what we were facing. Ultimately, when we went to tender for the project, three companies were considered at the final stage for putting forward a viable solution to our requirements. They were all very close to one another on technical specifications. The report shows that the firm we chose, Cognizant, came in at around €6 million. The two other tenderers at that time came in at between €3 million and €7 million higher, at €9 million and €12.5 million respectively. I mention this to put into context the €10 million initial estimate. The market was far from being something one could buy off the shelf. Effectively, the basic starting point of the other two tenders would already have been at the €10 million mark.

That being said, the Comptroller and Auditor General made a point that we came in ultimately with external costs of €13.2 million. However, that included some maintenance and support subsequent to the design and build. We spent approximately €11.1 million on the design and build of the system, which was really what CGEY stated earlier. We were 10% above that initial indicative figure, and that is the basis of my contention that we were broadly in line with that. However, as I said, the estimate was very crude as the market response to the tender would have indicated initially.

Deputy Derek Nolan

That is lucky rather than being an example of good project management. The Cognizant contract was for €6 million. It came in at €11 million in the end, so it overran by dramatic amounts. The CSO has the cover of a €10 million estimate by its consultants, but that is no justification for a project going so far over budget. The purpose of tenders is to get the best value for money. The CSO got that with the €6 million tender, but the fact that it then went over by 55% cannot be excused simply because it had an earlier report stating that it would cost €10 million.

Mr. Gerry O’Hanlon

When one is building a system rather than purchasing an off-the-shelf product, there is a variety of tendering mechanisms to try to reach a price for it. One can go for a fixed price which covers everything, a time and materials solution, or a compromise between the two. It is true that the contract as originally signed involved a fixed price for very specified requirements, but at all stages it also provided for a fixed price at specified rates for changed controls, where there would be a need to go beyond that. The intention in our strategy when going to tender on that was to get as much of the contract as we felt we could be definite about at the time for the fixed price, while leaving open an avenue to deal with some other elements where we would not have been precisely clear on exactly what we wanted. From that point of view, we had a very tight negotiation system drawn into the contract, which allowed for those additional elements to be factored in.

We sought additional services in two dimensions, both for the design and build itself and in looking for an increased period of maintenance. There were elements of uncertainty in these areas at the beginning and it would have been wrong for us to incorporate them into the fixed price because we just did not have the full definition of what was involved. The full implementation and delivery of this contract ran over a six year period, as the report states. The report also indicates that trying to get this is a fact of life.

We are satisfied with the process. Our approach towards the tendering system was a very cost effective way for a project such as this.

Deputy Derek Nolan

There were two figures given for some of the cost overrun. There was €2.4 million provided for agreed project changes, and €900,000 for an increase in the amount of time for which they will be willing to provide support to the project. I can understand the second figure as it is straightforward, but can Mr. O'Hanlon explain the €2.4 million for project changes?

Mr. Gerry O’Hanlon

I may call in my colleague to give more precise examples of what was involved, but as the project was being developed and as we saw exactly what was being delivered and what was required, we identified that what was precisely specified in the contracts did not exactly meet our requirements. The other element was that there were new business requirements coming in over that period. New surveys or new types of functionality were required. Perhaps my colleague can elaborate on some of those.

Ms Elaine Lucey

We were prototyping the system during the course of the project. As users were looking at the prototypes and evaluating them, the change requests came in. There was a fine level of detail and it was difficult to anticipate them beforehand. It was through the use of the prototypes and through testing that these changes came about.

Deputy Derek Nolan

Did the CSO not get a report from Ernst & Young on how to set up the system? Was there not a preliminary report on what kind of project and data systems would be needed? Had that preliminary scoping work not been paid for in the consultant report by Ernst & Young?

Mr. Gerry O’Hanlon

That is true to the extent that they could identify it. This is the factor that it was taking. This was a very ambitious project that we had undertaken. It is probably fair to say that what we were attempting to do was to develop a generic, all embracing system that would cover everything. In principle, that has been recognised as the right way to go but it was clear that no other office had implemented it - I am referring to our peers abroad - even though a number of them have moved in that direction since. So to some extent, even though we had the assistance of external contractors, they had nothing they could draw on that would enable them to be totally precise about what was required. There was an element of setting the broad direction, but the contractors and anybody else who assessed it would have discovered the full implications and the difficulties involved only as they progressed through the development and testing of the prototype.

Deputy Derek Nolan

What was the cost of the consultants' report?

Mr. Gerry O’Hanlon

The previous one? The Accenture report - that is, ITSIP A - which dealt with the detailed requirements, came in at about €600,000. It involved about 6,000 hours of work. The previous report, by Ernst & Young, may have cost less. Unfortunately, I do not have the figure, but I can provide it for the committee.

Deputy Derek Nolan

Does Mr. O'Hanlon understand where I am coming from? First he mentioned that he could not provide comparisons, but the CSO is not the only statistics office in the world. Surely there are lots of other national statistics offices that could have been peer-reviewed. I am questioning the fact that these two reports gave the scope of how the CSO wanted the project to be run, what needed to be done, and the systems that needed to be put in place. There are many statistics offices in the world that could have been brought into the discussion. I am not blaming Mr. O'Hanlon. These people were paid money to produce a plan for the CSO to update its systems. It subsequently turned out that the office needed to spend an extra €2.4 million on changes to that system to make sure it was fit for what was required. The CSO spent a lot of money on a report that did not tell it everything it needed to know, so that it could not operate a proper tender, and it was caught later on having to implement changes to the system that cost an extra €2.4 million.

Mr. Joe Treacy

With regard to Deputy Nolan's question about international benchmarking and so on, the first thing to say is that statistics offices around the world are not identical to each other. They face similar processing requirements, but they work in different national environments and cultures in terms of the balance between survey data and the use of administrative data. We are not all identical, but we do actually work together a lot to share approaches. There is much work done internationally, which is mainly to do with defining the business process for statistics and developing a common understanding of that. At the time, we brought in experts from Statistics Netherlands and Statistics Sweden and from New Zealand to discuss how to modernise the way in which the CSO was organised. Organisation and change management was involved; we needed to reorganise to become a more efficient, more aggregated statistical organisation. In addition, we examined what kind of core IT system we would need to establish to support that. That was the international context. There was no off-the-shelf international solution from another office. As Mr. O'Hanlon was explaining, we went through a process of business review that started with the Cap Gemini Ernst & Young report, setting out the kind of IT strategy and changes that we needed to establish. The data management system - the ITSIP stage A project, as we called it - set out a requirement specification. Even at that point, we could not know everything that would be needed. The documentation behind the DMS project is enormous, and the documentation that describes the different applications - from collection through processing to having clean data for the production of statistics - is very extensive.

I am sorry about this lengthy answer. I will conclude by saying that having implemented this, we have been the subject of a lot of interest internationally. We have had study visits from other countries, including countries as far away as South Korea as well as many of our European Union colleagues. They wanted to see what we had done and learn from it - both the good and the bad aspects. Some countries used us as a benchmark for developing their own approaches. Thus, there were some positive outcomes despite the difficult path to having a system that is up and running and in use.

Mr. Gerry O’Hanlon

To add to that, the statistics business is one that is very much amenable to benefiting from wider developments in information technology. Part of our problem back in 2002 and 2003 was that we were still using our first major statistical system, which was a mainframe-based system that dated back to the mid-1980s in terms of technology. Since that period, particularly during the 1990s, there was a veritable revolution in the IT area, moving from mainframe computer systems to distributed client-server architectures.

Deputy Derek Nolan

As I said initially, there is no denying the CSO has changed its hardware, which needed to happen.

Mr. Gerry O’Hanlon

It is more than a simple matter of hardware. We did not just go out and buy machines. In fact, far more important is the development of software systems that could be applied. In the wider international statistical community, this whole issue of modernisation is an ongoing process. It often depends on where one is positioned. In 2002, unfortunately, we were not in a good position, because our technology dated back to the 1980s and was essentially obsolete, so we had to make a major leap. Hindsight is a great thing. At the time, we had to try to get ourselves into a new situation that would allow us to move into a modern environment over the next 15 to 20 years. We had to do this with a level of ambition; in other words, we could not just go out there and pick up what someone may have done in the mid-1990s. What we were embarking on was an attempt to identify and build a system that would be forward-looking and enable us to move forward in a modern environment. As I said, there was no off-the-shelf solution, so we had to try to get the best advice. As Mr. Treacy said, we got a combination of advice: we obtained consultancy advice from Ernst & Young, and we also brought in people from a number of statistical offices that would have been viewed at the time as being well advanced in their adoption of the new technologies. There was no perfect solution. To a certain extent, we would have seen it. We put a deliberate level of ambition into what we went looking for and, to address another point that was made, all of that ambition was not realised. Even for people who have come into the development subsequently, there is still an ongoing international attempt to fit the statistical requirements to emerging IT technologies. I am saying this to provide a context.

Deputy Derek Nolan

Mr. O'Hanlon is misinterpreting some of my questioning. Under no circumstances am I saying the CSO was embarking on a project that was unworthy or did not need to happen. Our job is to question budgets that were in place and how cost overruns occurred. That is the simple purpose of my questions. What I am saying is that although there were two consultants' reports guiding the process, the CSO carried out a tender process that turned out to be inadequate and had to make changes to the system midway through at a cost of €2.4 million. All I am trying to figure out is why that happened.

Mr. O'Hanlon mentioned that the project was extremely ambitious, which is evident from the report of the Comptroller and Auditor General. In fact, it may have been so ambitious that it tried to do more than it ought to have done. Some of the conclusions of the Comptroller and Auditor General have been acknowledged by the CSO action plan. These include the idea that instead of undertaking large scale projects, the organisation should move to smaller manageable projects. Although not spelt out in black and white it can be inferred from some of the commentary and it appears that at one point the project seemed to be detached from those who used the systems; the process was put in place by senior management but the input of business users, the people in the offices who used the systems, was not completely part of the process throughout; and at one point there was a significant reluctance from people within the organisation to use the system after several years of implementing change. This meant the project had to undergo several reviews to establish what was happening, to the point where, to this day, a great deal of work which, it had been anticipated, would be done using the system is not being done using it. Was the project over-ambitious? Does the organisation have the corporate governance and human resources ability to bring people along with it? If the organisation had to do it all again, would it undertake the project in the same manner?

Mr. Gerry O’Hanlon

I will address the last point first. If we had the benefit of hindsight we would have done certain things differently. I was a member of senior management when the project was embarked upon and I have reflected on it a good deal. I have asked myself given the information available to us then, what could we have achieved at the time and would we have done things differently. On the basis of the information available to us and the advice we were able to get at the time I do not believe we would have done anything substantially different. I have strong views in this regard and my colleagues in place at the time would have gone along with this view.

The first question posed dealt with the two preliminary contracts, under which we brought people in to advise us on what was to be done. They gave us good advice to the extent that such advice could be garnered at the time. I do not have a problem with it nor do I believe we received inadequate advice. It reflected what was asked of them and the extent to which the queries could have been answered with the completeness required. I do not accept that there was a serious overrun in the costs of the project. I have explained this already.

I was asked about the buy-in element in regard to the project. We brought it out together. Given the factors relating to its implementation, there were what may be seen as change requirements within the Central Statistics Office and there was possibly more resistance than we would have liked. To some extent, there is a view that the business was not properly involved in the project but I do not accept that. The project board included the directors in charge of each of the business areas. In addition, the two project managers were experienced CSO business area people with considerable experience in this area. There were teams of business analysts working closely on it. There may have been something, a layer in between. Had some people working in the business been involved, they might have had a different view but that is a different thing from claiming that the business was not actively consulted.

A question was asked about the overall take-up. As I stated initially, we were considering something with a long-term perspective. It is now 2012 and I envisage that the data management system, DMS, will be a core part of the information technology infrastructure in the coming decade at the least. This area is improving all the time. At this stage, it is most relevant to what we deem to be medium and large projects within the office. As the table we have provided demonstrates, some 80% of these projects, by far the most important, are now within the DMS. This situation is improving all the time, especially as new projects are being been embarked upon. One area kept out of the DMS initially was the national accounts area. We are currently pursuing a major upgrading of our information technology systems in this area. The DMS will be a central pillar of this requirement.

Many of the smaller projects are written or undertaken on simple systems using excel spreadsheets or the statistical analysis software, SAS, information technology system. The fact that we have not moved over is a deliberate choice at present, mainly on the part of users because they have simple requirements which are being met at the moment. We could put them into the DMS and the DMS could meet them but it is not necessary to put a priority on that the moment. To a certain extent, 80% of the real target areas within the office are currently within the DMS. We believe this figure will increase. Generally, our view of the information technology applications in such projects is that if one can get to an 80% adoption level, one is seen as having had a successful project.

Deputy Derek Nolan

I will finish presently. The report states:

The decision to discontinue the use of three of the applications and to focus primarily on the earlier stages in the survey processing lifestyle would appear to be inconsistent with the original vision ... Delays and technical problems during the implementation phase resulted in a degree of pessimism and doubt among users as to the achievability of the project ... The closure report acknowledges that the loss of engagement by business users during the migration phase was related to the quality of the new communication of corporate objectives ... As regards the prime contractor, although adequate change control procedures were in place the changes agreed added 40% to the original fixed price sum, which indicates that the original specification exercise was not comprehensive.

In his opening statement, Mr. O'Hanlon stated that by and large he accepted the report of the Comptroller and Auditor General. Despite everything I have asked him today I have seen no acceptance that there was any failure, any problem or any mis-communication on the part of the Central Statistics Office in the implementation of the project. Before coming here, I anticipated that this would be a straightforward report. However, I have seen nothing but defensiveness and a complete defence of everything that occurred on this project. There has been no acknowledgement that there were any communication or project errors. This is despite a clear report from the Comptroller and Auditor General stating that issues arose. I do not suggest the entire projects was a disaster. However, some problems were identified and I have tried to tease out the details. Basically, Mr. O'Hanlon has denied that there were any problems during the project. This contradicts what the Comptroller and Auditor General has stated in the report.

I have been part of this committee for one year and I have never before taken this attitude to a witness. I see nothing but a completely defensive attitude. It is a case of minding one's own patch rather than confessing to some of the contents of the report. These are little things that could be changed. We could have teased them out at the Committee of Public Accounts, where public moneys spent are scrutinised in a public way such that people can have confidence that when something happens in an organisation, the organisation can learn from it and put in place remedial procedures. I would rather that approach than coming here and suggesting everything is fine and that with the benefit of hindsight he would have done everything in exactly the same way.

Mr. Gerry O’Hanlon

I am sorry if Deputy Nolan considers my reaction to his questions defensive. In my opening statement I remarked that we accepted there were weaknesses and strengths in the project. I stand over that totally. With regard to the adoption stage, what was written in the report reflected what happened at the time. Let us consider one specific point, that is, the reaction of the business when we were developing the system. There was a reaction from the business to the adaptation. This happens in change projects everywhere and we had it rather heavily. It is interesting however that the level of resistance is declining as we move along and this is consistent with change projects. Naturally, the operation was not perfect but nor was it a failure. To some extent my reactions today suggest that given the realities of trying to implement a project such as this, one cannot achieve everything. One must set reasonable and acceptable objectives.

Undoubtedly, we did not achieve everything and difficulties arose during the course of the project. As we mentioned in our submission, our strong conclusion is that we have a new and modern IT system that compares favourably with those of international peers. This is unlike the situation that obtained in 2002 and 2003 when we had a near obsolete IT system. Our system had a positive effect and played an important role during our recent organisational changes to, for example, the structure of our business, the processing of surveys and so on.

There were problems in the implementation and we did not achieve all of our objectives but the project's underlying objective was delivered broadly in line with overall cost budgets following a deliberate process on the part of the CSO's management.

Chairman

Mr. O'Hanlon stated that 80% of projects were on the data management system, DMS.

Mr. Gerry O’Hanlon

Of medium and large projects.

Chairman

Is the system being used to its full potential on each project?

Mr. Gerry O’Hanlon

No, but its adaptation is constantly increasing.

Chairman

Was it overspecified? Is Mr. O'Hanlon happy that the CSO will use the system to its full potential?

Mr. Gerry O’Hanlon

If the Chairman means overspecified in the sense that we did not use everything we set out to get, the answer is "Yes". More to the point, we were trying for a generic, all-embracing system that would cover everything. Unfortunately, that did not occur. The nuances are better and more flexibly handled outside the system. Had the Chairman asked me two to three years ago whether we were satisfied with the adaptation of the DMS, I would have referred to severe limitations. At this stage, I am far more sanguine. The system has delivered in line with our expectations.

Our IT system has other elements and the DMS is not the sole part. There have been other developments, for example, in our use of the statistical analysis system, SAS, which is proprietary international software. We have gone beyond what was seen as the core DMS specification, but DMS remains part of our solution.

Chairman

If the CSO is not using it to its full potential, does that mean the CSO overspent on the specifications? Was the €18 million value for money? Had the CSO set out a better plan, could it have acquired a product it is currently using at a cheaper price? Reverting to Deputy Nolan's questions, has the CSO examined this issue so that lessons can be learned from the project, how it was embarked upon and the end purchase? How much has the State lost in terms of the sum spent versus the end result?

Mr. Gerry O’Hanlon

It is difficult to tell. One would need a yardstick for comparison. Compared with ten years ago, €18 million, including internal CSO staff costs, is a reasonable figure. Our dedicated expenditure on IT, including for IT specialists, during the ten-year period was in the order of €70 million to €80 million. My broad answer to the question of whether the system is value for money is "Yes". Consider the costs of the alternatives. The basic cost of what was slightly the best technical option suggested to us was initially tendered at more than €12 million. This did not include CSO costs or so on. Adding the CSO's staffing costs of €5 million would amount to €17 million. Given the existing market, the offers and what we were seeking, €18 million was broadly what we were going to spend.

Chairman

Mr. O'Hanlon is saying that in spite of the fact that the CSO is not using the system to its full potential, the cost of €18 million represents value for money.

Mr. Gerry O’Hanlon

We could have gone about it differently. With the benefit of hindsight, we might have received a more focused response in some respects. Deputy Nolan asked whether we learned anything from the process. We learned that a more modular, focused and limited approach might have been successful, but others argue that we might not have covered as much ground in the same period had we taken such an approach, given the higher overheads in specifications, the delays and so on. In a project like this one, a trade-off is necessary. From everything we have seen, we may have overspecified but the cost level is broadly of the magnitude expected to be spent to achieve the end result.

Mr. Joe Treacy

I will add one or two points. Some of the Chairman's questions touch on the issue of the long-term return on investment. I will enumerate some of the benefits achieved. One measure would be to determine what IT costs the CSO today. We are free of maintenance charges on the DMS because we maintain it using our staff. Our A5 subhead, which covers IT expenditure, has dropped dramatically. I have been examining it for the past ten years or so. While IT maintenance spend used to cost between €5 million and €7 million per year, including cyclical projects, it is now approximately €3 million. This is a significant reduction and the €3 million is the amount required to keep our environment in place and updated as regards hardware and so on.

Another significant benefit is the ability to implement new surveys without additional or higher IT development costs, which would have been the case had we continued with our one-by-one silo approach to surveys. There are concrete examples of this benefit in terms of research and development, innovation, business energy utilisation and so on. The costs avoided are concrete benefits. The CSO is doing more with less because it has a DMS.

We have implemented a new organisational structure for business statistics. It is more aggregated and is more efficient because it is a functional organisational model rather than a survey-by-survey mode. This structure has been enabled thanks to our central IT system, the DMS.

Chairman

I have a final question. Given our hearing late last year on the general Government debt account and the €3.6 billion, is there a chance of such an error recurring with this new system? How did the error occur? We are still awaiting a response from the Government. What changes have been made to the systems to ensure this does not happen again? Was the €18 million accounted for?

Mr. Joe Treacy

If I may clarify, there is not a connection between investing in the DMS and the particular errors. I wish to be clear about that.

Deputy Kieran O’Donnell

I think the Chairman is fully aware-----

Mr. Gerry O’Hanlon

My colleague is getting his defence in first, that IT is not responsible for everything.

Deputy Kieran O’Donnell

No defence is required.

Mr. Gerry O’Hanlon

The CSO and the Department of Finance have a shared responsibility in regard to the gross Government debt statistics. I am also aware that the Minister commissioned two reports and they have still to be put in the public domain. The CSO has co-operated with the finalisation of those reports. I do not believe it would be appropriate to discuss the areas which the reports will deal with.

Solely from a statistical perspective, which is a perspective we tried to put forward at the time, the error - it is very much to be regretted that it occurred - has been portrayed very much in an accounting context, whereas, in fact, it was an error in statistical consolidation, in a process. In the minds of many people, there is a view that statistics and accounting are the same and that they have the same requirements but there is a fundamental difference between the two. Accounting is focused on accounting for everything and auditing and ensuring that all the bits and pieces are properly accounted for and relate to one another. Statistics, on the other hand, is in the business of using every piece of information available, sometimes from accounting sources or from Estimates or from a wide variety of different sources, in order to provide overall aggregate estimates of what is happening in the economy, for example. Our judgment of the seriousness of the €3.5 billion problem is very much viewed in that context. We regret very much that it happened because it should not have happened but its seriousness is, frankly, in a statistical sense much less than perhaps has been portrayed.

I emphasise that the key indicator we were looking at is the gross debt to GDP ratio. In 2007, the Irish figure, as we have indicated in our note, was around 25%. It jumped, unfortunately, to 45% in 2008, to 65%, approximately, in 2009. The first estimate for 2010, the erroneous one, was 95% and the revision took it back to 93%. It was unfortunate that an error occurred but in the statistical sense it did not in any way misrepresent Ireland's position either at the level of the debt-to-GDP ratio or, indeed, the trend taking place at the time. As I am director general and head of the CSO and responsible for official statistics, I wish to make that point firmly. I heard a question being asked recently that these figures are reported to EUROSTAT and what was the EUROSTAT reaction to this. It was very much as I have explained to the committee. It was a case of thanking us for the notification of the revision which it would incorporate in its figures. These types of errors, because of the nature of statistics, can happen from time to time but the Irish position has not been seriously misrepresented in any respect by that error.

As ever, the CSO is willing to learn from the exercise. Mr. Punch will comment on the procedures in more detail and on what has happened in the meantime. It was a most unusual circumstance and in that regard, it is highly unlikely that it will ever be repeated. In any case, we will be prepared for it in our systems so that it will not get through again. It was a most unusual error. As my colleague, Mr. Treacy said, it has nothing whatsoever to do with the DMS.

Mr. Aidan Punch

We are not passing the parcel between us. As head of macro-economic statistics, I concur completely with Mr. O'Hanlon. To answer the Chairman's question as to whether the error would have occurred if, for instance, national accounts IT had used the DMS, my straight answer is that the probability of it occurring would be far less. My biggest priority as head of macro-economic statistics is to ensure that the compilation of all our macro-economic statistics - not just the debt and deficit material which is very important clearly - is put on a proper IT footing. This is not the case at the moment and I am being totally up front in this regard. This system has evolved over 20 or 25 years and it is not the best system. A priority for us is to move our IT processing system in order to utilise the DMS. We are in contact with Ms Elaine Lacey and her colleagues in an effort to move this forward. We have difficulties and these have to do with resourcing; we are getting no extra staff and neither are we looking for any extra staff. If I am to use the DMS it means I will have to put a number of my officials working on a two-year IT transformation project which is what I have in mind. One of the officials who accompanied me on the last occasion, Mr. Mick Lucey, who was a very valuable member of staff, has not been replaced. Luckily, the Department of Public Expenditure and Reform has agreed to a replacement because it is a key job. Once that job is replaced, we will be moving towards transforming our IT and national accounts and in particular, in the GGD area which is a very important component. I hope that answers the Chairman's question.

Chairman

Before I call Deputy O'Donnell, while Mr. Punch's explanation may be quite valid and this is the way he views it in his world, it was viewed quite differently at the time. From a systems point of view, to put it mildly, it should not happen. Errors will be made but when an error is made, particularly of that kind - I refer to Deputy Nolan's comments - it is much better if it is explained promptly and that we all understand how it happened and what actions are being taken to correct the system so that errors do not happen in the future, rather than having to wait from late last year until this morning to receive some sort of explanation as to where the problem arose and why it arose. That simple explanation could have been provided without much of the fuss that has accompanied it since then, between this committee and the Government. It is a lesson for all of us that big issues such as this, involving in this instance a sum of €3.6 billion, should be explained immediately and promptly.

Deputy Kieran O’Donnell

I welcome Mr. O'Hanlon and his colleagues. I wish to deal with the issue of the sum of €3.6 billion. Mr. O'Hanlon describes the difference between a statistician and an accountant. I will be the accountant now. With regard to the €3.6 billion it is not as simple as saying it is to do with the general Government debt. Our job is to examine how it impacts on people's lives. We all have our separate jobs. The interest cost on €3.6 billion affects the everyday lives of ordinary people. We wanted to know whether the interest element reflected in the current account for the budget. Everything has an implication. I do not agree it is simply a case of saying the figure was wrong. EUROSTAT contacted the CSO two years after the event and the bird had flown. In my view, it is very simple and I refer to what Mr. Punch said. There should be a way of having checks in an operating system. The CSO is examining the level of annual national debt and a quick calculation based on the interest charge should reconcile to the amount that is physically showing as being an interest charge. One can do that on the back of a page. It is basic accounting, nothing to do with computer systems.

The delegates are saying the discrepancy had no impact but it is more accurate to say that while it apparently had no impact, we are still awaiting the report as to whether there was an understatement or overstatement in respect of the amount of interest charge that went into the annual accounts for the State. We are talking about an environment, back in 2010, where there was pressure on the taxpayer from austerity measures in terms of the budgets that were brought through. In that context there are huge implications arising from any discrepancy of this type.

Will Mr. O'Hanlon indicate what precisely happened in terms of the compilation of data? The discrepancy was based on information compiled by the CSO which was relied upon, in turn, by the Comptroller and Auditor General. How did it transpire that a reconciliation was not done at the time as between the amount of interest that went in on the basis of the national debt and the actual national debt figure? I am not convinced it is as simple as the delegates saying they spoke to EUROSTAT and there was no issue. It may not have been an issue statistically in terms of the debt to GDP figure, but if someone put in an incorrect interest charge, either by understating or overstating it, that could potentially have had a huge impact by way of further cuts being imposed on taxpayers. Every euro that is taken out in expenditure reductions has an effect on people's lives, people whose disposable income is already under enormous pressure. It is not an abstract concept. Will Mr. O'Hanlon outline the reconciliation procedure and how this could have happened? Setting aside his claim that it had no major impact, which I do not accept, how was an understatement of €3.6 billion allowed to happen?

Mr. Gerry O’Hanlon

What happened is fairly well known.

Deputy Kieran O’Donnell

I would like Mr. O'Hanlon to tell me. I have not heard it from the CSO, which is the body responsible for the error.

Mr. Gerry O’Hanlon

I will give the Deputy my take on it, although I am not the expert in this area.

Deputy Kieran O’Donnell

Mr. O'Hanlon is the boss of his organisation.

Mr. Gerry O’Hanlon

I will tell the Deputy how it happened. The Housing Finance Agency had previously used the National Treasury Management Agency as its agent for raising money. In 2010, however, in the wake of the major turmoil in the financial markets, instead of acting as agent, the NTMA met the HFA's funding requirement from its own resources. Even though nothing essentially changed from the perspective of the HFA and perhaps even from the business perspective of the NTMA, this did have an implication for the statistical measure of gross Government debt. Unfortunately, in the forms we used at the time or the understanding of the people involved, the nuance of that change was not recognised. This will be the subject of the report from the Department. I do not wish get into it and it is as well it is left for the reports, in fairness to the people concerned and so on.

On the other point the Deputy made, my understanding is that all the accounting figures of the NTMA and the HFA, all in their separate rights, were totally above board. There was no question over them. It is also my understanding - this is a matter for confirmation by the Department of Finance - that the statistical measure of gross Government debt has no direct role in the calculation of the State's interest payments as such. There would not have been a method for independently cross-checking the figure. Unfortunately, what occurred in this instance was a classification error, which was made by the statisticians and accountants in the NTMA, which balanced itself out. In other words, as it moved through the system, everything appeared correct because the initial misclassification was not picked up.

Deputy Kieran O’Donnell

It strikes me as logical that one would take the gross national debt and reconcile the interest charge with what physically appears on the Government's accounts on an annual basis. One would assume that to be normal practice, but Mr. O'Hanlon is saying it is not being done.

Mr. Gerry O’Hanlon

My understanding is that what would be seen as the interest payments would not have been calculated by reference to that figure.

Deputy Kieran O’Donnell

I accept that.

Mr. Gerry O’Hanlon

It would have been calculated perhaps via other figures that were related to it.

Deputy Kieran O’Donnell

The point I am making is a technical one. If a reconciliation was done between the gross national debt and the actual interest figure itself, that €3.6 billion would have automatically stood out.

Mr. Aidan Punch

It may or may not have done. The Deputy asked for our explanation of what happened. We have issued two press statements to that end. As soon as the issue was discovered, we issued press releases indicating the extent of the debt and communicated the matter to EUROSTAT. There was no delay in any of this and we took the initiative, as an independent body, to take that approach. I submitted a three-page note to this committee on the morning we appeared before it. That note still stands and is available on the committee's website. It sets out as cogently as possible, in plain English, what happened. Our explanations have not changed in the meantime.

I assume, once the committee has received the two reports, that it will consider them in detail, in which case we will be available again, presumably along with our colleagues from the Department of Finance and the NTMA to answer any questions members have. We are not fudging the matter. The error was made and it was most unfortunate. It was a very simple, consolidation error. To come back to information technology and make the link with that, when one puts in place a good, solid information technology system, as we have done through the new data management system, DMS, and so on, one can build in counterparting information. In other words, it is possible to compile the data from the various sources and confront one with the other. We did not have that counterparting capability n place previously and that is how the error occurred.

Deputy Kieran O’Donnell

I have two further questions for the delegates. Mr. Punch referred to the CSO's new DMS system. For people in the business sector, the amount of form-filling for various purposes is a cause for complaint and something which places undue demands on time and resources. Is the new DMS system capable of doing a comparison, say, of the common information coming in through all forms? Does the CSO undertake an annual review of forms with a view to identifying the potential for reducing the number of forms businesses have to complete?

In terms of discrepancies between the live register and household survey figures, what level of detail does the DMS system have the capacity to produce? Can it, for instance, break down the profile in detailed terms as to who is unemployed, where they are from and so on? That level of detail would assist Departments and State agencies in, for example, linking the provision of training to the profile of people who are unemployed and on the live register.

Mr. Joe Treacy

The Deputy's first question is essentially about the response burden and the means of collecting data. The DMS is what we call multi-channel in that it allows us to type in a form, scan a form, receive data online and from other administrative sources, and so on. We are increasingly making use of administrative data in order to reduce burden. For example, in the census of agriculture 2010 we used a great deal of the agricultural administrative data and have continued to do so in the June and December surveys. In the case of annual services inquires, in the past year we have substantially reduced the burden on small businesses employing less than ten staff through a combination of administrative data, a shorter questionnaire and a smaller sample. These types of initiatives are reducing the burden and we are measuring that effect in an annual publication on response burden. We have a burden barometer which shows that between 2008 and 2010, we reduced the burden on business by 18%. Given that the Government strategy is to achieve a 25% reduction by 2014, we are well on the road to meeting that target.

Mr. Gerry O’Hanlon

As I indicated in my opening statement, it is one of our strategic objectives to try to reduce the burden at all stages. The perception of the burden placed on businesses in the context of their supplying statistics and so on is much greater than the actual burden itself. We have carried out studies in this regard

Deputy Kieran O’Donnell

I have dealt with businesses and I have burned the midnight oil while trying to complete forms for them.

Mr. Gerry O’Hanlon

I would be quite happy to forward to the Deputy our report in which we try to analyse the position. As Mr. Treacy indicated, we are trying to make use of directly administrative data as an alternative to direct survey collection. It is a key strategic aim of the National Statistics Board and the Central Statistics Board to progress this. The Government has supported the strategy that was recently communicated to it.

The Deputy inquired about the live register, which is one of the key administrative sources of information and is based on the numbers of people availing of jobseeker's benefit and allowance. In the past we used to be obliged to obtain summaries from each of the local offices. Now, however, the figures for the live register are actually based on each individual who is registered. In so far as the information relating to registrants is picked up, we can carry out a full analysis if we can obtain their addresses, etc., and we can break it down to quite a detailed level.

Deputy Kieran O’Donnell

Does the CSO currently receive detailed information from social welfare offices in respect of people's names, dates of birth and previous occupation? Does it produce reports in which this information is reflected?

Mr. Gerry O’Hanlon

We certainly produce age and occupational analysis. The location analysis is down to county level. Part of the difficulty with addresses is, as the Deputy is aware, that they are just literal strings. We are strongly pushing for the adoption of post codes throughout the country. If these codes are adopted, then it will be possible - at the flick of a switch - to carry out a detailed analysis on the impact of unemployment at quite a localised level. As statisticians, this is a matter which we are pursuing strongly.

Chairman

Mr. O'Hanlon and I have discussed this matter on previous occasions. In that context, those in the SME sector would still consider the forms to be a major burden. I refer, in particular, to smaller companies. I often wonder about the accuracy of the information supplied to the CSO on these forms. Many small companies merely tick the boxes and return the forms in order to ensure that they are in compliance. As a result, the accuracy the CSO is seeking is lost. At times, the forms that are sent out are not suitable for some of the companies which receive them. As a result, they either cannot complete them accurately or they do not have employees who are suitably qualified to complete them. This means that they simply engage in an exercise of ticking boxes. The more of this the CSO can remove from the system and rely instead on more definite and accurate information, the better it will be for everyone involved. This will serve the interests of the CSO and will come as a relief to those businesses which Mr. O'Hanlon states are happy to shoulder the burden of bureaucracy placed upon them by the forms in question but which I contend are deeply unhappy with that burden.

Mr. Joe Treacy

One of the more difficult questionnaires for businesses to complete is that which relates to quarterly payroll or earnings data. The CSO has worked with payroll software providers in the interests of developing a solution in order that small or medium-sized businesses will be able to extract the relevant information in the format required for the CSO questionnaire. We have informed such providers that the way they can help small businesses to extract the relevant data is by building the necessary software into their systems.

Chairman

That is fine but there are small businesses which cannot afford the burden that is being placed on them and which do not have the requisite systems in place. Some of the forms could be simplified in order that they might suit the needs or whatever of these businesses and perhaps that is the matter at which the CSO should be looking.

Deputy Eoghan Murphy

As someone who was involved in filling out forms and interacting with the relevant personnel, I thought that last year's census was extremely well conducted. The information emanating from the census is very interesting. How much did it cost to conduct?

Mr. Gerry O’Hanlon

In broad terms, my recollection is that over a five-year period the project would have cost - obviously most of the costs would have related to 2011 - in the region of €50 million to €55 million.

Deputy Eoghan Murphy

Was that cost incurred over a five-year period?

Mr. Gerry O’Hanlon

Predominantly, it would have been incurred in 2011. In my introductory remarks I stated that total expenditure for the office in 2011 was approximately €74 million. In 2012, it will be €45 million. That means that the additional expenditure for 2011, mainly relating to the census, was approximately €30 million.

Mr. Joe Treacy

I have some details in that regard. In 2011 we spent €34.5 million on the census. In 2010 our spend on the census was €8.9 million, while in 2009 it was €3.2 million.

Deputy Eoghan Murphy

I thank Mr. Treacy. Is he in a position to provide a breakdown in respect of the way in which that money was spent? On what, for example, was the €8.9 million for 2010 spent?

Mr. Joe Treacy

The largest item within that would have been IT development costs in respect of the processing system for the census. I do not have an exact figure for 2011 in that regard. However, it certainly was the largest item.

Deputy Eoghan Murphy

What that an entirely separate IT system?

Mr. Gerry O’Hanlon

Absolutely.

Deputy Eoghan Murphy

Was the work in that regard outsourced?

Mr. Gerry O’Hanlon

Yes.

Deputy Eoghan Murphy

Was the company involved paid to build a new IT system or-----

Mr. Joe Treacy

We have gone to market with an RFT on each occasion but it was the same company which installed and delivered the systems in 2002, 2006 and 2011. Essentially, it is a large-scale processing system which scanned approximately 1.6 million census forms - which add up to many millions of pages - and processed them in a short period, namely, from April to December 2011. It is quite a specialised, once-off type of IT system.

Deputy Eoghan Murphy

Did the money spent in 2010 cover all elements of the use of the system? Was the same system used on the three occasions on which censuses have been conducted since 2000?

Mr. Joe Treacy

The technological requirements and opportunities move on. The same company has done the work on each occasion. As far as I am aware, we have minimised the additional costs by trying to reuse as many elements as possible.

Deputy Eoghan Murphy

So the system is not necessarily built from scratch on each occasion.

Mr. Joe Treacy

That is correct.

Deputy Eoghan Murphy

So the cost in respect of the system for the years 2002, 2006 and 2011 should have decreased.

Mr. Joe Treacy

In fact the amount we paid in 2011 was less because it related to the utilisation rather than to the build of the system.

Deputy Eoghan Murphy

Was that for the build of the system on this occasion compared to the build on previous occasions?

Mr. Gerry O’Hanlon

Mr. Punch was previously responsible for the census so he may wish to comment.

Mr. Aidan Punch

In the first instance, competition rules dictate that one must go to the market on each occasion and spec one's system. This is one system that was delivered on time and within budget. I accept that the committee is not scrutinising the position in that regard today but this is an example of a good IT system. In saying that I am not trying to denigrate the DMS in any way. My point is that, as Mr. Treacy indicated, the same company was successful on three occasions. Effectively, however, it was implementing a new solution on each occasion. It rebuilt and reused what it utilised on the previous occasions and it charged the going rate. We went to tender and the company in question offered the best price. It went on to deliver the system on time and with no fuss. The system processed all of the census forms in the period to which Mr. Treacy referred.

Deputy Eoghan Murphy

This was CACI UK.

Mr. Aidan Punch

Yes.

Deputy Eoghan Murphy

I appreciate the costs will not necessarily come down that much and technology must change. The company was successful twice previously and a third time. To what extent is there actual competition when it applies?

Mr. Aidan Punch

They were successful in 2002. The 2001 census had to be cancelled because of the outbreak of foot and mouth disease. At the time, that contract contained a seven-year period which enabled us to go back to the company for the 2006 census on price. We negotiated very hard on the price and it was really down to the wire. We had to get this over the line as quickly as possible or otherwise we would have to go to the market again. CACI UK was subsequently used in 2006.

In 2011, we went again to the open market and it was the successful bidder again. There is not a significant number of companies in this area. It must be remembered at the time the US and New Zealand were also doing censuses. There are few enough players in this market.

Deputy Eoghan Murphy

In 2011, what did the cost figure of €34.5 million cover?

Mr. Aidan Punch

The cost of the field operation is the big thing in 2011.

Mr. Joe Treacy

The largest item was €22.2 million for field pay costs and €2.2 million for field non-pay costs, a total of €24.4 million.

Deputy Eoghan Murphy

How many people were employed in the field operations?

Mr. Joe Treacy

Approximately 5,500 people.

Deputy Eoghan Murphy

Were they on short-term contracts?

Mr. Joe Treacy

Yes.

Deputy Eoghan Murphy

Were they Irish?

Mr. Joe Treacy

They did not have to be Irish. They had to be resident here and available to work.

Deputy Eoghan Murphy

Were any of them taken off the live register? Was it used to assist people to get back to work?

Mr. Joe Treacy

It is not permissible to give preference to people on the live register. We advertised the posts through FÁS services but preference cannot be given to live register status.

Deputy Eoghan Murphy

Why not?

Mr. Joe Treacy

It would be discriminatory.

Deputy Eoghan Murphy

Companies are given tax breaks for taking a person off the live register if he or she has been on it for longer than six months. Is there any consistency?

Mr. Joe Treacy

If there is, then it is a policy question.

Mr. Gerry O’Hanlon

This is a long-standing question. Of the 5,500 people employed, 5,000 were enumerators employed part time for ten weeks. The issue went back to a finding at the Labour Court many years ago that it was discriminatory to give preference to people on the live register. For example, a woman working from the home may not be on the live register and must be given a fair opportunity to get this employment.

Instead, we gave a premium for people's availability for work. That would favour people who were unemployed and on the live register. As we were also operating on a 20 to 25-hour week model, it would not be attractive to people on the live register if they lost a complete payment for doing this work. We had to make an arrangement with the Department of Social Protection as to how these cases would be treated. As this was a piece-rate working system, we also had arrangements that there would be payment on account over the period to ensure the income flow was interrupted over it.

Getting one of these field jobs was not automatic for those on the live register. However, those on the live register interested in getting one of these posts did have a starting advantage from the point of view of availability.

Deputy Eoghan Murphy

Perhaps it is policy issue but we do give tax breaks to employers who take people off the live register which is somewhat discriminatory. I find it interesting. How much were these people paid an hour?

Mr. Gerry O’Hanlon

It was €1 over the minimum wage, working out at between €9.50 and €10 per hour. It was not specified in that way. Enumerators were paid on the number of households dealt with and the amount of time spent on training. It was a piece-rate working arrangement.

Deputy Eoghan Murphy

What made up the rest of the spend, €10 million, in 2011?

Mr. Joe Treacy

As I said, field collection costs came to €24.4 million. The headquarters salary costs, mainly in the processing operation in Swords, came to €4.7 million.

Deputy Eoghan Murphy

What do the staff do in headquarters?

Mr. Joe Treacy

Up to April, they were administering the census, such as running the field operation. From April, they were dealing with receipt of the forms and the implementation of the processing system. All the 1.6 million forms had been scanned, processed and were ready to analyse by December last year. They completed this very fast.

Deputy Eoghan Murphy

Was that CACI UK?

Mr. Joe Treacy

No. The IT system had been installed by CACI UK but it was all Central Statistics Office, CSO, staff who processed the forms.

Mr. Gerry O’Hanlon

All handling of forms was done by permanent CSO staff, additional staff recruited on temporary six-month or nine-month contracts or staff from other Departments. These were CSO-controlled staff. At the peak six months during 2011, we would have had up to 150 additional staff working in our Swords office in the processing of the census.

As a side note on the impact of technology, in 1996 before we moved to the automatic scanning operation and the population was smaller, the census would have absorbed about 700 man years of labour in the CSO. With the use of new technology, that figure has been reduced to 200 to 300 man years of labour. We are producing far more outputs and far quicker too. This is directly linked to the adoption of the new technology.

Deputy Eoghan Murphy

CACI UK provides the technology and it is then utilised by the CSO. Is the information stored offline or is it online? Are the servers on which this data is stored owned by the State?

Mr. Joe Treacy

The data is stored on CSO premises and computer equipment.

Deputy Eoghan Murphy

Is that connected to the Internet?

Mr. Joe Treacy

No.

Deputy Eoghan Murphy

At no point is this data held by CACI UK?

Mr. Treacy

No, it is not held by CACI UK.

Deputy Eoghan Murphy

There were concerns at the time that the State would have to pay CACI UK or its parent company for access to the census information.

Mr. Joe Treacy

No, it is not technically or legally possible for that scenario to happen.

Mr. Aidan Punch

The census system was separate from the overall CSO system and not even linked to it. It was completely locked down.

Deputy Eoghan Murphy

How do we know the data we get from people is accurate? For example, in this or the previous census, there was a thing going around online about ticking "Other" for religion and writing in one is a Jedi Knight. If enough people did so, we would have Jedi Knight classed as an official religion in Ireland. Is that possible?

Mr. Aidan Punch

It is possible on our census form.

Deputy Derek Nolan

How many Jedi Knights have we in Ireland?

Mr. Aidan Punch

I mean on how we know the information is accurate. If one compares the information from one census to another at the detailed level of geography, age group and so forth, one will get a remarkable consistency. Why do we hold a census of population every five years? The reason is that the population is volatile in terms of immigrants and emigrants. That is one of the main reasons in a small area dimension but the data are robust from census to census. That comes back to the Chairman's earlier point about businesses. I am not convinced that people just tick boxes. Looking at our business statistics over time we get remarkable stability in the data which points to the fact that people are taking this process seriously.

Deputy Eoghan Murphy

Do we have any Jedi Knights in Ireland?

Mr. Aidan Punch

No. I do not think anybody did that. If someone did not tick one of the boxes on the religion question, which is what the Deputy is talking about, they could write down their particular religion if it was not one of the three or four specified. I do not think we came up against any of those on this occasion.

Deputy Eoghan Murphy

They might have left.

Mr. Gerry O’Hanlon

If somebody wishes, under religious denomination, to consider themselves a Jedi Knight, the CSO is prepared to accept the entry as Jedi Knight because it is self-declaration. Whether such a thing exists is immaterial. We can probably tell the Deputy the number who have recorded themselves as such but we do not publish it.

Deputy Paschal Donohoe

If the witnesses come back to us with a follow-up note on the number of Jedi Knights in Ireland it would be most unusual piece of correspondence the Committee of Public Accounts has ever received.

I want to focus on three areas that touch on some themes colleagues have mentioned. On the discussion about the general Government debt, and my questions arise from the briefing notes the witnesses were good enough to send to us, which are helpful, a sentence in the last paragraph refers to having already implemented a number of key steps to improve procedures in relation to data on Government debt. Could we have an update us on those steps?

Mr. Aidan Punch

I will try to respond to that. We have already tried to learn the lessons. For instance, since we last met in committee here we have signed off our 2012 return, which would have given the GGD for 2011, 2010, 2009 and 2008, and that was published last month. I chaired the Government Finance Statistics Liaison Group and we got all the people around the table. That did not just happen on a day; they would have been circulated with the data beforehand. I am talking about the Central Bank, our colleagues in the Department of Finance and ourselves and we signed off on that particular statement. Everybody had to be happy about the accuracy of the statement. That was getting a type of common ownership around the processes and so on.

We have also put in place better templates, although not on the IT footing I would like to see achieved. That will be a longer-term project, perhaps over two years, but we have put in place better templates with counterparting information on it to ensure that the type of error that occurred will not occur again. There are other examples also. It is mainly to do with our collaboration with the other organisations including the National Treasury Management Agency, the National Housing Agency and so on.

We have a draft memorandum of understanding with our colleagues in the Central Bank specifically on this point because we already have a general memorandum of understanding with the bank, and we will roll that out to all the organisations. There is a more formal process in place rather than the previous informal system. There is a more formal set-up in terms of our dealings with this very important Government measure.

Deputy Paschal Donohoe

Within that area, Mr. Punch might help me to understand his full role. I understand the role he plays in regard to reporting general Government debt. Does he play a similar role regarding the reporting of deficit figures for the State?

Mr. Aidan Punch

That release covers two areas, the deficit and also the debt. It is called Maastricht debt, which means it is reported in gross terms, and we do not let out the assets on the other side of the account. We compile those as well. In fact, for the first time in our forthcoming national income and expenditure report in respect of 2011, which we publish in June, we will set out a set of tables which will chart the deficit and debt situation. Our plans are to introduce our own release on that and therefore it will be a statistical release setting out, in a specific definitional context, the debt and deficit situation on an annual and quarterly basis.

Deputy Paschal Donohoe

To return to the discussions on the IT system on which my colleague, Deputy Nolan, focused, I want to focus on the overall cost and the figure of €18.2 million that has been quoted. Page 8 of the briefing document gives a breakdown of the different components of the €18.2 million figure and then at the top of page 9 it states that the company was retained to provide maintenance support for DMS for 21 months - the initial 12 months and an additional period of nine months for knowledge transfer costing in total €2.1 million. Is that €2.1 million captured in the €18.2 million?

Mr. Gerry O’Hanlon

If the Deputy goes back to table 2, which is the breakdown on page 8, the second last column refers to maintenance and knowledge transfer and he will see there is a figure of €2,135,000. That is the €2.1 million we are talking about.

Deputy Paschal Donohoe

That €2.1 million, therefore, is not the €2.1 million referenced in the paragraph below it, out of a total of €7.1 million.

Mr. Gerry O’Hanlon

No.

Deputy Paschal Donohoe

They are separate.

Mr. Gerry O’Hanlon

They are separate, yes. The figures are just the same.

Deputy Paschal Donohoe

They just happen to be the same.

Mr. Gerry O’Hanlon

Yes.

Deputy Paschal Donohoe

That is fine. That €18.2 million includes the maintenance costs in the aftermath of the project going live.

Mr. Gerry O’Hanlon

By way of explanation, essentially, the €18.2 million is the capitalised cost of the whole project from the point of view of our balance sheets, asset registers and so on.

Deputy Paschal Donohoe

On the €5 million figure for CSO staff costs, I am surprised by the size of that figure. From my understanding from reading the notes it appeared to be omitted at the early stage of evaluating the project. To put that €5 million figure in context, it refers to the internal staff costs within the CSO, and is roughly equivalent to the initial estimate of the total contract of €5.9 million. Why was such a large figure not taken into account at the early stage of putting the project together? It is just under one third of the total cost of the project.

Mr. Gerry O’Hanlon

We did not take on any additional staff to-----

Deputy Paschal Donohoe

I am not suggesting that was done.

Mr. Gerry O’Hanlon

To a certain extent, even in our old legacy systems there would have been dedicated staff and IT systems working along and therefore no matter what we did, whether we stayed with the old system or tried another approach by doing in-house development, we had a complement of staff available for IT projects. Effectively, a large element of that €5 million would reflect our core existing capacity for development work, and therefore it was not explicitly identified.

The other part of the €5 million which brings it up was the adaptation costs for the business areas in changing their systems because the business staff would have to spend time in changing and updating their systems, and all of that was also costed and included as part of the cost of the project.

Deputy Paschal Donohoe

The point Mr. O'Hanlon is making is that the €5 million does not represent an additional cost for the CSO. It represents a diversion of its existing resources to the project at that time, but nonetheless I am sure the diversion of that €5 million to this project had an opportunity cost elsewhere in Mr. O'Hanlon's organisation in terms of work he could have done at the time. That brings me back to the observation that it is a large, internal cost which I am surprised was not included in the early stages of evaluating the project as opposed to the latter stages of it because that €5 million is just under one third of the total cost of the project.

Mr. Gerry O’Hanlon

It was probably an omission from the point of view of putting together an overall comprehensive project cost. The Deputy is right that had we been putting together an overall comprehensive project cost, it should certainly have been included. We were aware of the large element of it. Even before the arrival on site of the contractor, Cognizant, we had put together a fairly large and skilled, dedicated team from within the office. Perhaps we had ten to 12 staff dedicated immediately to the project. One should bear in mind the context, namely, that the cost accumulated over a period of six to seven years.

Deputy Paschal Donohoe

We have had a lot of discussion about the cost of the project. Mr. Treacy, in his testimony, referred to the benefits of the project that have accrued. Has an exercise been done to monetise those benefits and evaluate whether the expenditure paid off for the taxpayer?

Mr. Gerry O’Hanlon

Not in any comprehensive sense.

Deputy Paschal Donohoe

Okay. It always strikes me that in discussions on expenditure I would evaluate cost in two ways. First, I would ask whether the final cost was equal to that envisaged at the start of the project. We have had a discussion on that. Second, I would ask whether the revised cost was worth it. Mr. Treacy outlined a number of benefits that would allow one to conclude the expenditure was worth it. I am sure Mr. O'Hanlon believes this. Why has there been no process to determine whether the expenditure of the original €18 million was justified?

Mr. Gerry O’Hanlon

As we said, our existing IT system faced obsolescence. To be blunt about it, if that obsolescence had been realised, the CSO's business would have stopped dead in its tracks. The fact is that over the long period, there was no disruption to any CSO outputs and we greatly expanded them. Our putting in place an alternative IT system - fundamentally the data management system - was a major contributor to that.

The global answer is that the CSO's business was facilitated by the system continuing over the period in question. Categorising what precisely is relevant to the data management system as opposed to other elements of the IT system or office would be the problem. To some extent, we are reduced to looking at other milestones. Mr. Treacy said that, in our business statistics area in 2010, we engaged in a major reorganisation of our system for processing, analysing and disseminating our data. That has resulted in a very significant saving in terms of staff costs. In the order of ten to 20 staff have been released from that area on foot of the reorganisation. That is very much enabled by the availability of the data management system. As to how much precisely should have been attributed to it, there is a large element of subjectivity. We can make indirect references rather than engage in a formal costing exercise.

Deputy Paschal Donohoe

I take that point. As one who did many payout calculations and who analysed whether money was worth spending, I believe that when one sits down to engage in such exercises, one always ends up dealing with the intangible qualities. It does not matter whether one is dealing with soap powder, machinery or otherwise; the exercise is always riddled with uncertainties and intangible qualities on which one must make a value judgment in order to establish a monetary value. While I fully understand Mr. O'Hanlon's answer, I contend it is one that most people give when asked whether an investment has paid off. I urge that expenditure on infrastructure of an IT nature or otherwise be subjected to a retrospective payout analysis, for the simple reason that if one does so problems become less likely in the future. In this regard, consider the discussion we had on internal costs, amounting to €5 million. The session we just had would have proceeded differently had Mr. O'Hanlon stated the cost and then outlined why the expenditure paid off. I urge Mr. O'Hanlon to consider for the future the question of payout and whether the use of money is justified, even if there are intangibles that are difficult to value.

Mr. Gerry O’Hanlon

I take the Deputy's point. I will turn to my colleagues to determine whether they can, at some stage in the future, devise-----

Deputy Paschal Donohoe

The CSO will be engaging in this process again in respect of a different issue-----

Mr. Joe Treacy

Certainly.

Deputy Paschal Donohoe

-----but perhaps not on the same scale.

Some of the discussions we have had and the questions that arose on the allocation of costs might be handled differently if we considered the adoption of a payout model after work such as the work in question.

Mr. Seamus McCarthy

I would like to comment on the comments made by Deputy Donohoe. Many of the reports of the Comptroller and Auditor General over the years have emphasised the requirement for a good business case to be established for an investment project. It would inform decision-making and should identify the full life-cycle costs. It should also identify the expected business benefits, perhaps in terms of efficiencies. Even if they are not monetised, efficiencies may be describable. In the CSO's case, the impact on the cost of individual surveys could also be identified. Reference could also be made to the burden on business and the impact thereof.

The overall conclusion of the Comptroller and Auditor General was that this was a successful IT project by comparison with others we have reported on. The look-back exercise that was undertaken was very thorough. It has the potential to be a learning exercise for the CSO but also for other organisations. That is why the Comptroller and Auditor General made the recommendation that the Department of Public Expenditure and Reform examine this project and the look-back exercise and learn the lessons to be learned and perhaps disseminate them more widely.

Chairman

Does the committee agree to note Vote 4, pertaining to the CSO, and dispose of Chapter 19, pertaining to the data management system? Agreed.

Mr. O'Hanlon is to retire shortly. On behalf of the committee, I wish him well in the future. I thank all the witnesses for attending this morning's session.

The witnesses withdrew.

Sitting suspended at 12.20 p.m. and resumed at 12.35 p.m..