Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
 Download Current HansardDownload Current Hansard   

Previous Fragment    Next Fragment
Wednesday, 27 October 1993
Page: 2642

Senator PATTERSON (4.55 p.m.) —by leave—I move:

  That the Senate take note of the document.

I rise to speak on this important efficiency audit by the Australian National Audit Office in relation to the Department of Social Security's data-matching program. Honourable senators will well remember that the necessary legislation to facilitate the introduction of this program passed through the Senate in December 1990. Considerable concern was expressed by both the coalition and the Australian Democrats in the months leading up to the passing of this legislation about the privacy implications of this initiative and the assumptions underlying the estimated savings to be derived from the program.

  At that time the government claimed that the data matching would be a central element of its fraud control strategy. According to the Labor government, its introduction would conservatively result in savings of at least $300 million per year. In addition, the parliament was assured that the technology that had been earmarked by the bureaucracy to undertake this program would ensure that the intrinsically intrusive nature of this fraud control method would be minimised.

  We were also informed that the success of the program depended directly upon a significant extension of the tax file number. People will remember that when the tax file number was introduced the previous Treasurer, now the Prime Minister (Mr Keating), promised Australia that it would never need to be extended for any purpose other than taxation. But he broke another promise by extending it to the social security system.

  In the end, the Labor government secured the passage of this legislation by way of the not too subtle technique that it usually uses of a gun to the head of the coalition. The coalition was told that, if it opposed the bill, it would be branded as supporting people who deliberately fraud the social security system. When I suggested that we change a word, the then Prime Minister, Mr Hawke, wrote to Dr Hewson claiming that changing a single word in the bill would result in lost savings of up to $150 million a year.

  This audit report not only shows that the government's loaded gun was not loaded at all; it also clearly demonstrates that virtually every one of the concerns expressed by the coalition was entirely valid. As I said, the government claimed that the savings from this program would be $300 million a year. The auditor has concluded:

The downward variation in payments, which were the basis of the original estimates, represent only 17 million dollars of the 1992-93 savings.

This is a damning indictment of the assumptions that were used in the calculation of the original savings estimates. The minister and his bureaucrats were claiming that their savings estimates were conservative. Now we find that a mere six per cent of these savings are being realised. We were told that $150 million would be down the tube if we changed one word.

  A year ago, I highlighted to the Senate how the Department of Social Security was camouflaging this embarrassing saving shortfall. In its answers to questions that I put in Senate estimates committee hearings, the department had continued to claim that the annual savings from the data-matching program were in excess of $70 million during 1992-93. But a closer examination of these latest claims shows that some $40 million of these savings are a direct result of the so-called deterrent factor and not directly related to data matching whatsoever; that is, when people found out that there was data matching involved or that the tax file number would be taken, they decided to declare their real income, to withdraw from applying for social security or whatever else they decided to do.

  A further $13 million was attributable to identified overpayments which were also not included in the original savings estimates. In effect, this means that the department has changed the basis upon which it calculates its derived savings after the commencement of the program. Worst of all, it has quite deliberately tried to give the Australian parliament the false impression that the savings from this program are higher than they actually are. The Auditor-General's report states:

The DSS saving estimates available at June 1993 contained changes in methodology from the previous estimates, which should be clearly indicated. The ANAO considers that, when reporting to Parliament, the detailed basis on which DSS calculates estimates of savings from the program needs to be more transparent, with changes in methodology explained.

This matter is central to the concerns expressed by the non-government parties in 1990. At that time it was necessary for senators to make an informed decision as to whether the financial savings to be derived from data matching warranted the introduction of such an intrusive method of fraud control. In effect, we were asked to undertake a cost-benefit analysis that entailed an assessment not only of the financial costs and benefits but of the privacy repercussions as well.

  It is disturbing to see that the department has not seen fit to undertake its own cost-benefit analysis of the program, despite the fact that analyses are a requirement of the program protocol that was passed with the data-matching legislation. The Auditor-General found in his report:

The full range of costs incurred by the data-matching program are substantially in excess of those previously quantified by DSS in their reports on progress.

It is estimated that the annual costs entailed in the data-matching program are about $14 million. This means that a program that entails matching the confidential financial and personal data of up to 11 million Australians on an ongoing basis is being legitimised on the basis of net savings that may be as low as $3 million per year. This is a long shot, and way out from the estimates trundled out by the government less than three years ago.

  The auditor has suggested that the department conduct detailed cost-benefit analyses to assess the cost-effectiveness of the program more fully. It must be noted that the department has moved quickly to respond to this recommendation and included an analysis in its recently tabled report, which unfortunately I did not have the chance to speak on because I dashed down to the chamber to find that it had already gone through. However, it is quite telling that it had not seen fit to do so until it received this, what I would see as, not so gentle a reminder from the Auditor.

  There can be no doubt that the data-matching program was and is technology driven. Three years ago a number of officers from the department went overseas to acquaint themselves with a new technology that would allow the automatic matching of millions of pieces of data in a relatively short period of time. The fact that this program is technology driven is recognised by the Auditor-General, who states in his report:

A key factor justifying the data-matching program was information technology developments which allowed millions of matches to be undertaken automatically resulting in the confirmation of voluntary compliance by clients in a great majority of cases.

However, the Auditor-General's report also reveals that the department has not been administering this technology driven program in the most efficient way possible. For example, it has been shown that a massive 85 per cent of the so-called matches that are identified through the data-matching process are not actually matches at all.

  In some ways I wish Senator Powell were still here to have seen this Auditor-General's report. In fact, I feel inclined to post it to her. We sat through that committee hearing and we heard the officers from the department telling us that there would be 100 per cent accuracy. They told us that when we got a `hit', when it was a match, it would be definitely a positive hit and it would be somebody who was not giving the correct information or who was giving information to Social Security about their income that was not the same as the information in the Taxation Office.

  We were told categorically that a match would be a match—that it would be a `hit' was the word they used. And what do we find? We find that 85 per cent of the reviews that are undertaken to resolve the data-matching results do not reveal an incorrect payment. They are wrong 85 per cent of the time. We match data on millions of people in Australia who are not defrauding the system. They get pulled out of the system, they get examined in much more detail and they have not done one single thing wrong. We were told categorically—Senator Richardson would remember; he was sitting there listening to it and he heard Mr Trevor Murphy say to us over and over again—`When we get a hit it will be right; we will be spot-on. Someone will be in receipt of an overpayment'.

  Mr Murphy was wrong. Somehow the department now is examining 85 per cent of those matches which do not reveal an incorrect payment. But not only that; these 85 per cent represent some 64,000 clients whose financial and personal arrangements are subjected to departmental review without good reason.

  The auditor has highlighted the fact that annual labour saving costs of $2.9 million would result if 46 cent or 30,000 of the unsuccessful matches could be eliminated. So we would have made savings if the department had got it right, if a hit was a hit, as it was called. But no, we find that 85 per cent of those people indicated as hits have not been in receipt of an overpayment. If we could correct 46 per cent of those, if we could eliminate those incorrect ones, then there would be approximately $2.9 million worth of savings.

  The unnecessary, inappropriate and costly intrusion entailed in the data-matching program does not end here. Fortunately, the data-matching act does not allow the department to stop the continuation of a benefit as a result of data matching. Can you imagine all these people having their social security reduced or cut out because we had found these `hits', that were going to be so right as Mr Murphy said? Benefit does not stop, because they have to receive a section 11 letter; it has to be sent to the client first. This means that the review of the so-called match can be undertaken before a person's benefit is put in jeopardy. So we get these hits and the department then checks them. When the department has checked the match and believes that there has been an overpayment it sends out a section 11 letter.

  Unfortunately, the auditor has found that a high proportion of section 11 letters are successfully disputed by clients. This means that the department is drawing inaccurate conclusions even after it has reviewed the so-called matches. So it does the matches, they come out of the computer, they are checked—these were the matches that were going to be 100 per cent right—and the department sends out section 11 letters, and still gets it wrong in 11 per cent of the cases. They have to provide evidence to the Department of Social Security.

  We were told categorically that a match was a match and that it would be 100 per cent right. I can remember the officers sitting there telling us until they were blue in the face. I did not believe one word of it, and I was vindicated. I am going to go back through that Hansard. I can say `I told you so' on about three or four issues. Senator Richardson can remember that. I think Senator Richardson would have the good grace to agree that many of those things I raised have now come to fruition.

  These are not the only data-matching related inefficiencies that have been highlighted by the auditor. He has recommended that the Department of Social Security distribute the costs incurred in data matching with the other government agencies that benefit from it. I am not sure but I think I also asked that question somewhere along the line. It is also recommended that greater emphasis be put on analysing the cost-effectiveness of each of the 70-odd projects that make up the data-matching program.

  The auditor also concludes that variations between Department of Social Security officers are wider than necessary and recommends that it, and I quote because I do not like to use this word, `prioritise a reduction in the variability of the performance between offices, building upon the national quality assurance results'. Significantly, the auditor has concluded that up to $12 million in savings could be achieved by investigating clients whose benefits are either cancelled or voluntarily surrendered as a result of not providing their tax file numbers.

  Finally, this audit report cannot be allowed to pass without making some comments about the extension of the use of the tax file number. I just said that the previous Treasurer, our now Prime Minister, promised us that the tax file number would not be extended. We were told that it was vital and necessary for this intrusive data-matching program that we should extend the tax file numbers. We were told that it was an integral part of the program. The distinct impression we were given was that, without the use of the tax file numbers, the data-matching program would be rendered largely unworkable. In stark contrast, the auditor has found that no identity frauds have been detected from data-matching projects like invalid or inconsistent tax file numbers that are used to identify personal identity data discrepancies.

  In fact, the department now appears to be of the opinion that tax file numbers may not be necessary at all in the data-matching program. Other identifiers may be able to be used for the purposes of income matching and alternative strategies are being examined for handling personal identity discrepancies.    We were told that we needed to extend the tax file number to make sure that we could catch people who were defrauding the social security system. We were told that the government would make $300 million of savings, and that was so far out that it is a joke. We were told that when a match was made, when data came across and was matched with the social security data, people plucked out of the computer would in fact be in receipt of an overpayment. The government was wrong, and it has been wrong on so many of these issues.

  There we were, the `unrepresentative swill' as we are referred to by the former Treasurer and now Prime Minister, in that Senate committee, without the resources of the department, asking questions about protecting the privacy of millions of Australians. We asked pertinent questions about protecting many people who are not even involved in the social security system, such as people's spouses whose tax information is matched—millions of people who are involved in this data-matching process. With hindsight, those questions were absolutely spot-on. The department knew less about it than we appeared to know from examining the situation very closely.

  I urge all senators, who have to make a decision within the next 24 hours or so as to whether the sunset clause that applies to the data-matching program should be removed, at least to read the summary of the Auditor-General's report. In particular, I ask the Green senators, the Democrat senator responsible for taking social security legislation through, and Senator Harradine, to read this summary and to decide for themselves whether we should let that sunset clause go. Should we say, `No; we still want to have a sunset clause to say that there is a period after which we may not accept this data-matching process'?

  I am sure that any honourable senator who takes the time to read the summary will have no doubt and will conclude that the sunset clause should be extended and that the agencies that are involved in this program should continue to be required to provide the parliament with detailed annual reports in relation to their implementation of the data-matching act. Those amendments went through the Senate to ensure that the Senate would be fully informed about what I believe is a very intrusive investigation into people's personal financial matters.

  We required that the agencies table their reports here. That has assisted us greatly in understanding exactly how that process is working, except, as we found out, the Auditor-General has implied that they have not been as transparent as they ought to have been in those reports. They have lifted their game in the recent report because of the not so gentle reminder from the Auditor-General. I call on honourable senators to read the summary of this report. I commend it to them, especially given the decision that we have to make in the chamber in the next day or two.

  Question resolved in the affirmative.