This document is at http://www.privacy.org.au/Papers/SubTFN9910.html
The Australian Privacy Foundation has a particularly strong interest in the issue of Tax File Number Administration, as the enhanced TFN system was introduced in the immediate aftermath of the identity card debate in the late 1980s. The enhancement of the TFN and introduction of the employment and investment income reporting were accompanied by very rigorous privacy safeguards.
The Foundation has been closely involved in seeking to defend the privacy safeguards in the TFN system through several major extensions in the early 1990s, with limited success. As predicted from the start, the Executive government has proceeded by a process of 'function creep' to massively extend the level of routine surveillance of individuals' regular transactions - particularly amongst income support recipients and students.
The early history of the TFN system and its privacy implications are well documented in a 1991 paper by one of our members Mr Roger Clarke. We commend this paper to you as an essential resource for your inquiry. It is available on the Web at http://www.anu.edu.au/people/Roger.Clarke/DV/PaperTFN.html. Mr Clarke also maintains a regularly updated paper on the history of privacy developments in Australia at http://www.anu.edu.au/people/Roger.Clarke/DV/OzHistory.html.
We are pleased to see that the Committee has recognised the "the significant privacy sensitivities surrounding this matter" and undertaken to give these appropriate attention.
While the ANAO report acknowledges privacy as an issue, it gives the impression of not really understanding the significance of the balance between state power and individual freedoms struck in the TFN system. This is as much, if not more a general civil liberties issue as it is about privacy.
The ATO, having lived through successive parliamentary debates about the TFN, and having to deal with the reality of compliance enforcement, understands these issues much better, and this explains why the ATO has previously considered but not acted on many of the changes now recommended by the ANAO.
It is important at the outset to recognise that perfect compliance with tax obligations is impossible to achieve for a very wide range of reasons. By international standards, we understand that Australia is already estimated to have high levels of compliance.
As tax administration seeks to close in on the relatively small areas of non-compliance, the measures and powers required become more and more intrusive and oppressive, often for severely diminishing returns.
Even on narrow cost-benefit grounds it is doubtful if some of the changes recommended by the ANAO are worthwhile. Once important non-financial considerations, such as privacy and freedom are taken into account, the case for many of the others is clearly not sustainable.
Our principled opposition to further major extensions of the TFN surveillance system is not an endorsement or defence of tax evasion. It is simply a recognition that in a free and democratic society, there are clearly limits on the powers that citizens are prepared to allow the authorities to exercise. The price of setting such thresholds is equally clearly some degree of inefficiency and anti-social behaviour, including some level of tax evasion.
We do not oppose some minor 'tidying up' changes to remove obvious inconsistencies and anomalies, or the many improvements in the administration of the TFN system within the ATO and reporting organisations that could no doubt be made without extending the level of intrusion or surveillance.
The ANAO suggests that one argument for extending TFN quotation obligations and making them mandatory is to achieve fairer and more consistent treatment of all taxpayers. The suggestion made in paragraph 4.64 is that because quotation is effectively already mandatory for income support recipients and students, it would be more equitable and consistent for everyone to be required to quote. To accept this argument would be to make a mockery of the concept of fairness. The effective compulsion for some categories of taxpayers was introduced in the early 1990s in direct contravention of previous government promises. It was this that created unfairness and inconsistency, which cannot be undone by subjecting all taxpayers to the same heavy-handed regime.
As indicated earlier, a level of surveillance has been imposed on income support recipients and students which at the time would have been unacceptable to the majority of taxpayers. This imposition took advantage of the vulnerability and lack of political voice of those groups, and of the apathy of the majority of the population who could see no immediate application to themselves. It is unconscionable to now use this precedent as a reason for further extension.
The specific concerns that follow are ones that have stood out to us. There is so much detail in the report that we have almost certainly missed some other important issues and we reserve the option of raising these at a later stage if the opportunity is presented.
We note that while the text of the report correctly refers to TFN quotation being effectively mandatory (because of the severe consequences for exercising the right not to quote), the Table in Appendix 5 simply uses the heading Mandatory, which is not correct. This may just be an oversight, but has the effect of contributing to a perception that TFN quotation is already legally mandatory - a significant symbolic threshold which has yet to be crossed, and which would require legislative change, including in the Privacy Act.
The ANAO has put forward a figure of $250 million as the potential additional revenue from reporting of real estate transactions (Summary paragraph 10 and footnote 2). It is only when one reads the ATOs response to this finding at page 68 of the report that one becomes aware that the ATO cautions against the reliability of this estimate which is based on a very small and localised sample (The ATO also goes on to explain why other measures to address rental income issue have been preferred).
We have not been able to exhaustively check the other claims made by the ANAO, but this example of an exaggerated benefits estimate, devoid of an important qualification, make us suspicious of the overall integrity of the ANAO report. It is almost as though the ANAO at some stage formed a view as to its preferred recommendations and then selectively sought out facts and figures to support these.
Many assertions are made in the report about the incidence of non-quotation of TFNs in different contexts. These assertions are usually made to support a proposed extension of reporting requirements. However, there are few references in the text to figures or even estimates of the for the actual levels of non-quotation. It is only from paragraph 4.61 and Table 9 in Appendix 9 that we find that quotation levels for all categories are over 89%. These figures suggest to us two things. Firstly that the alleged non-quotation 'problem' is not massive. Secondly that such high levels of voluntary quotation may mean that individuals may misunderstand the implications of not quoting (see below).
We are concerned that the practice has grown up in the financial community of only partially informing investors of the consequences of non-quotation. Investors are typically told something along the lines of "You do not have to give your TFN but if you don't tax will be deducted at the highest marginal rate". They are rarely also told that any overpaid withholding tax can be reclaimed at the end of the tax year.
Most financial institutions use a form of words suggested by the Tax Office. In the early 1990s, the Privacy Commissioner sought to ensure that the ATO advice included a full notice of the implications, but this was often not done. It appears that practice has worsened - partly under pressure from forms designers seeking simplicity, but probably also because it is a 'nuisance' for institutions to have to deal with witholding tax.
The convenience of reporting institutions should not take precedence over clear and accurate information to customers. If the ATO itself was collecting the information, it would be required under the Privacy Act (Information Privacy Principle 1) to give individuals a full explanation. Reporting institutions effectively collecting information on behalf of the Tax Office should be required to do the same (they will be under the proposed private sector privacy legislation).
While quotation remains voluntary, a full and accurate explanation should be given to individuals about quotation and non-quotation and their implications.
As we feared when the enhanced TFN and associated income reporting were introduced, there has been a tendency for the authorities to regard non-quotation with suspicion, and this attitude has been passed on to reporting institutions and the general public.
The ANAO fall into the same trap when in paragraph 4.63 they state that "Non quotation of the TFN tends to be associated with situations where tax affairs are complex and where there are greater opportunities to evade, avoid or reduce the tax payable." There is no acknowledgement that some taxpayers may choose to exercise their right of non-quotation as a matter of principle - as a protest against the 'reversed onus of proof' inherent in the TFN system.
To the extent that there seems to be an issue about trusts (a high proportion of which do not quote all the relevant TFNs), we submit that this could be addressed by separate treatment of trusts. They are arguably more like corporate structures than individual taxpayers and an argument can be mounted for individuals waiving their privacy rights if they choose to take the advantages of trust formation. Certainly the privacy rights of all individuals should not be compromised because of specific concerns about the use of trusts for tax evasion.
We accept that the whole purpose of a high integrity TFN is to facilitate information matching. The use of the TFN by the ATO itself for matching against data from various other sources is controlled by the interaction of the tax laws and the Privacy Act and its TFN Guidelines.
The use of the TFN in matching programs for other Commonwealth government purposes is strictly regulated by the Data-matching Program (Assistance & Tax) Act 1990. The TFN Guidelines have been revised and re-issued by the Privacy Commissioner several times to accommodate the extended TFN functions authorised after vigorous parliamentary debate.
One very important factor in assessing the ANAO's recommendations is that the history of data-matching programs, both in Australia and overseas, has been one of massive overstatement of benefits in order to justify the investment in computer technology and the establishment of bureaucratic programs. The various programs falling under the 1990 Act are a good example. The ANAO itself has documented the savings achieved by these programs and shown them to be substantially less than the original estimates, which have been progressively downgraded.
In this report, the ANAO seems to have fallen into the same trap as agencies such as the ATO and DSS before, by making extravagant claims as to the likely benefits from extended TFN matching. As we have already noted, at least one major figure - the $250 million estimated from real estate reporting - is seriously questioned by the ATO.
It seems likely that compared with the matching of major data sets already taking place, the proposed 'mopping up' of residual and marginal data sets proposed in this report are even more likely to cost more, and yield fewer benefits than is optimistically claimed, with very little hard evidence.
We submit that in light of experience the Committee should significantly discount the estimated financial benefits from further information matching before weighing those benefits against the costs both administrative and in non-tangible detriment to privacy and other freedoms.
Experience of data-matching also shows that there are often major and often unanticipated problems with data quality, inconsistent definitions, different time periods etc. This can not only reduce the effectiveness of matching, but can also lead to many individual falling 'under suspicion' erroneously. This is one reason for the emphasis in both the statutory programs and the Privacy Commissioner's Guidelines on verifying matching results with individuals before any action is taken. Individuals are often able to provide satisfactory explanations for apparently negative discrepancies.
The existence of safeguards such as verification, while important, does not cancel out the privacy intrusion involved. Even if the individual concerned never becomes aware that they had fallen under suspicion, however briefly, their privacy has nevertheless been invaded. Privacy advocates have always had great difficulty in getting bureaucrats concerned to accept that it the act of matching itself is inherently objectionable, not just what is done with the results. Fortunately, parliamentarians have historically been more sensitive to this issue. Essentially it is about reversal of the onus of proof - matching is a speculative activity that puts everyone under suspicion and then 'clears' the majority. It conflicts fundamentally with our long established legal tradition of presumption of innocence.
Throughout the parliamentary debates about successive extensions of the TFN system, it has generally been accepted by all parties that one major principle should remain the quarantining of the TFN within the taxation and income support sector of government functions. While one can argue about how closely related some of the superannuation related uses are to taxation, there is at least an association between the administrative functions currently using the TFN.
The idea of allowing the TFN to be used to link records from completely separate government programs - most notably between the health programs and others, has rightly been strenuously opposed. Australians are understandably concerned that their health and medical information should be kept confidential, and many remember that the Health Insurance Commission that was a prime mover in the proposal to introduce an Australia Card in easy stages to allay public concern.
Unfortunately the government has managed to establish a link between health care and tax administration by enacting the private health insurance rebate and surcharge. This link must not however be allowed to be used as the basis of an extension of the TFN into the health care system.
The ANAO suggests that the ATO should be able to match its data with HIC data - this would require legislative changes to the secrecy provisions applying to the HIC and would be very controversial. We are disturbed that the ANAO has carried out test matching, and would like to know what statutory basis the ANAO has relied on to overcome the secrecy constraints.
The main benefit of ATO-HIC matching would appear to be for the ATO in gaining access to more regularly updated addresses, although there is a curious suggestion in paragraphs 4.38 and 4.39 that discrepancies might indicate improper receipt of benefits - the rationale for this is not explained. In relation to the address matching, we reject the assumption that the ATO needs to continually monitor taxpayers changes of address. As long as taxpayers provide the ATO with a valid address as and when contact is required (primarily when lodging tax returns), what business is it of the Tax Office to know what if any changes take place during the year, and when?
This proposal betrays an alarming mind-set of seeking to keep the population under continuous surveillance - paragraph 4.36 expresses it, almost gleefully, as an 'almost real time view'.
What would be even more alarming, and totally unacceptable, is the use of the TFN as the matching key for ATO-HIC matching. Although the ANAO report does not expressly recommend this, its discussion of the value of the TFN as a key under the heading of external matching, and immediately following the section on the HIC data, can be seen as an attempt to sow the seeds for this further function creep.
If the use of the TFN was ever extended from the tax/benefit area of government to health care administration, the bureaucracy would have substantially achieved by increments much of the power of the Australia Card surveillance system without the visible symbol of a photo-identity card which so galvanised public opinion in the 1980s.
We call on the Committee to firmly rule out any suggestion that the TFN should be used in the health care administration arena.
We hope that this submission has indicated our grave concerns about the overall direction of the ANAO report as well as much of its detail. Despite paying lip service to privacy issues, it is clear that the authors of the report fail to understand the depth and significance of the intrusion and surveillance that would result from adoption of their recommendations. The estimates of benefits are largely speculative and in some cases are heavily qualified by the ATO in its responses.
We would be pleased to elaborate on our concerns if the Committee so wished.
Other resources which the Committee may find useful are:
We note that this paper is a response to an Auditor-General report and subsequent Parliamentary Committee Inquiry that appears to parallel your committee's current inquiry. See also the Canadian Commissioner's view of information sharing at http://www.privcom.gc.ca/02_05_a_981126_e.htm and at http://www.privcom.gc.ca/02_05_a_990301_e.htm.
Nigel Waters
For
Australian Privacy Foundation
PO Box R507
Royal Exchange
SYDNEY NSW 2000
For immediate contact:
E-mail - nigelwaters@primus.com.au
Telephone: 02 4981 0828 or 0407 230342
For further information, contact also:
Tim Dixon, Director, (02) 9231 4949 or 0411 114411, TimDixon
Prof. Graham Greenleaf, (02) 9569 5310 or 9385-2233, Graham Greenleaf
Dr Roger Clarke, (02) 6288 1472 or 6288 6916, Roger.Clarke@anu.edu.au
Go to the APF Home Page.
Created: 22 November 1999
Last Amended: 22 November 1999
APF thanks its site-sponsor: