Australia has required all eligible electors to enrol (register) to vote now for one hundred years. In recent years, however, changing demographic patterns and citizen mobility have made it increasingly challenging to maintain the accuracy and currency of the federal electoral roll. These same social trends mean that strategies such as habitation reviews which have been successful in the past are no longer effective or feasible. Two of the state electoral commissions have responded to the declining enrolment rate by implementing systems in which administrative data from other agencies are used to add individuals to the roll and update their details without their intervention (direct enrolment and direct update). Recently passed Commonwealth legislation now allows the AEC to adopt similar processes. The current paper examines the processes by which the NSW and Victorian Electoral Commissions have implemented direct update and direct enrolment through previously published information and interviews with the commissions. The paper also reports on their progress so far, including the impact on the 2010 Victorian and 2011 NSW state elections. The implications of these reforms for electoral roll management in Australia generally are discussed.
Federal direct update and direct enrolment legislation was passed in June 2012 and given Royal Assent on 24 July 2012 in response to the first two recommendations of the inquiry of the Joint Standing Committee on Electoral Matters (JSCEM) into the 2010 election. The recommendations were that the Commonwealth Electoral Act 1918 (CEA) should be amended to allow the Australian Electoral Commission (AEC) to directly update and directly enrol eligible persons based on data or information that person had provided to an agency approved by the AEC. A similar recommendation was also contained in the report into the 2007 federal election (Recommendation 10).
The purpose of this report is to examine the existing approaches to direct update and direct enrolment implemented by the electoral commissions of New South Wales (NSW) and Victoria. The report will cover both publically released information about the operation of direct enrolment and direct update in NSW and Victoria, and report on consultations with the NSW and Victorian electoral commissions focusing on the operation of their programs. The aim of the current research is to allow the AEC to better anticipate the likely impact of recent legislative change for federal direct update and direct enrolment on its business including consequential impacts on turnout and informality. The research will also assist the AEC in developing mitigation strategies to address any adverse impacts should they arise.
The Commonwealth Electoral Act 1918 charges the AEC with the responsibility for maintaining the Commonwealth electoral roll. Each of the Australian states and territories has a Joint Roll Arrangement with the AEC, and the Commonwealth roll also forms the basis of each of the rolls used by the state and territory electoral commissions1. This has traditionally meant that Australians could complete one enrolment form and be enrolled for local, state and territory, and federal elections.
Prior to 1999 the AEC updated the roll using habitation reviews, which involved door-knocking of addresses across Australia's states and territories to confirm enrolments (AEC 2005). Habitation reviews were time consuming (taking up to six months to complete) and expensive around $16 million in 1999 dollars), and occurred only once every two years. Changing residential patterns and increasing population mobility made periodic habitation reviews a less effective roll maintenance strategy (AEC 1999).
In 1999 the AEC in conjunction with its joint roll partners, the state and territory electoral commissions, introduced a new process for updating the roll called Continuous Roll Update (CRU). CRU matches data obtained from various state and commonwealth government agencies against the roll to identify individuals who have moved or who might not be correctly enrolled. The AEC also has an address register of habitable addresses and can identify when there are no (or too many) electors enrolled at an address. When the data matching process identifies potential enrolment variations the individual is sent a letter and enrolment form asking them to confirm their new details (ECA 2005). The enrolment is not updated until the individual returns a signed enrolment form to the AEC or an objection action is commenced by the AEC (see below).
The first iteration of CRU in 1999 involved the receipt of 600 000 change of address notifications from Australia Post. The response rate to the initial mail out was 32.3 per cent (ECA 1999). A report by the Australian National Audit Office in 2001 into the integrity of the electoral roll found that CRU was an effective means of managing the roll and would provide a roll that was "highly accurate, complete and valid" (ANAO 2002, 13). Two years into the operation of CRU, ANAO found that while the process was beneficial, the data used for CRU was not consistent across states and territories.
In 2004–05 4.1 million change of address and new potential elector records were received, primarily from Australia Post, Centrelink and motor vehicle licensing authorities. The resulting 2.4 million letters from the AEC resulted in 1.4 million responses, a response rate of 55 per cent (ECA 2005). Following the recommendations of the 2002 ANAO report the AEC also undertook Sample Audit Fieldwork (SAF). The SAF exercise aimed to test the accuracy and completeness of the roll by doorknocking in 225 randomly selected Census Collection Districts (CCDs)2. The 2005 SAF found enrolment participation at 98.4 per cent, enrolment completeness at 96.3 per cent, and enrolment accuracy at 91.1 per cent (AEC 2005). The positive results from the SAF appeared to validate the effectiveness of CRU, and SAF exercises undertaken in 2006, 2007 and 2009 produced comparable results (AEC 2011a).
The CRU process has continued to be the primary means of updating the commonwealth roll, with 4.5 million letters sent in 2009–2010, and 33.7 per cent of all enrolments resulting from CRU letters (AEC 2011a). However response rates to CRU letters have declined over time. Current response rates to CRU letters are between 15 and 20 per cent (AEC 2012), resulting in CRU gradually becoming a less effective way of keeping an up-to-date roll.
In addition to generating enrolment letters, CRU can also result in individuals being removed from the roll. Section 114 of the CEA requires the AEC to begin objection action where there is reasonable grounds for believing that the person has not lived at the address for which they are enrolled for one month. If the response to the CRU letter indicates that the individual no longer lives at their enrolled address (such as a return to sender response) the AEC may send a notice indicating that if the individual does not respond they may be removed from the roll after 21 days (AEC 2012). As such, lack of response to the AEC's letters may result in the individual losing their franchise.
The CEA, as it stood until June 2012, required an individual to submit a signed enrolment form in order to enrol. CRU could identify individuals who were not enrolled, or who were enrolled at the wrong address, but the AEC could not do anything with this information other than notify the individual. Unless the individual took action to correct their enrolment, they risked being objected from the roll.
The idea of using the same information as the CRU to enrol an individual or update their details without requiring the individual to take any action is not a new one. In the JSCEM report on the 2007 federal election the committee recommended that information received by the AEC be used to directly update the roll where the elector had specifically consented to their information to be used for that purpose (Recommendation 10). Directly enrolling electors was specifically recommended against (JSCEM 2009). The Australian Government's 2009 Electoral Reform Green Paper also discussed "automatic enrolment" and "automatic update", and noted that the NSW Government was then examining these approaches as part of their SmartRoll project. The history of the NSW SmartRoll is outlined in more detail below.
In late 2009 the Special Minister of State asked JSCEM to examine the NSW Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Bill 2009, the legislation enabling NSW's SmartRoll process. In its submission to the inquiry the AEC stated that direct enrolment and update were "positive initiatives from the perspective of facilitating enrolment" (AEC 2010, 2). The AEC reiterated its support for direct enrolment and direct update in its submission to the JSCEM inquiry into the 2010 federal election (AEC 2011a). A majority of the committee clearly agreed, and the first two recommendations resulting from the majority report of the JSCEM inquiry into the 2010 election were that the CEA should be amended to allow the AEC to directly update and directly enrol eligible persons based on data or information provided to an agency approved by the AEC (JSCEM 2011).
On 23 June 2012 the Electoral and Referendum Amendment (Protecting Elector Participation) Bill 2012 and the Electoral and Referendum Amendment (Maintaining Address) Bill 2011 were passed by the Senate. These bills respectively allow the AEC to directly enrol and directly update eligible electors. At the time of writing the AEC was in the process of developing its direct update and enrolment systems.
This report discusses systems by which electoral management bodies (EMBs) can use administrative data from other agencies to enrol an individual or update the details of an already enrolled elector without the intervention of that individual. At present, such systems are in operation in Victoria and NSW at the state level. The NSW Electoral Commission has adopted the term "SmartRoll" as the branding of their system; the Victorian equivalent is not branded.
The best practical guide to what direct enrolment and direct update involve comes from existing programs in place in NSW and Victoria. The NSW and Victorian models of direct enrolment and direct update allow the Electoral Commissioner to enrol eligible electors and update their details when these details change at the Electoral Commissioner's initiative. In practice, the electoral management body receives data from other agencies, including vehicle registration and licencing authorities and other bodies such as utilities and higher education authorities. These data are subjected to predefined business rules to determine their validity, with suitably trustworthy and verified data then prompting processes to add eligible individuals to the roll, or change the enrolment information of those already on the roll, without the intervention of the individual. The individual is then notified of their enrolment or enrolment change, and given the opportunity to dispute that change. If no objection is received, the roll is updated with the new information by the commission and the elector is notified to that effect.
In much of the discussion of these systems they are referred to as "automatic enrolment" programs. For example, the NSW enabling legislation, the Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Bill 2009, explicitly labels the approach as automatic enrolment. While the Victorian legislation doesn't explicitly name the approach, the description of the legislation on the VEC web site refers to the amendment as implementing "automatic enrolment" (VEC nd), as does the Victorian Electoral Matters Committee (2012). Brent and Hoffman (2011) also collectively term these approaches, including some international approaches, as "automatic enrolment".
"Automatic enrolment" has become the default way of referring to these systems, despite the fact they are also used to update electors' enrolment details. It is a convenient, if somewhat misleading, shorthand to refer to what we will show is a collection of approaches that are more similar in terms of outcome than process. Notwithstanding this, the preferred terminology used by the AEC is "direct enrolment and direct update".
In its Second Supplementary Submission to the Joint Standing Committee on Electoral Matters on the Conduct of the 2010 Federal Election, the AEC (2011b) outlined its preferred definition:
In the following discussion, the term 'direct' is used to indicate actions initiated by the AEC rather than actions initiated by an elector. 'Enrolment' refers to transactions where a person is enrolling for the first time (and re-enrolling after a hiatus where there is no current record), whereas 'update' refers to updating an elector's address details for an existing enrolment. For these purposes a change in name is considered a new 'enrolment' and not an 'update'. (p. 3)
Use of the term "automatic" in relation to enrolment from third-party data may inaccurately suggest that the data does not undergo stringent quality assurance and are not checked by electoral officials before being used to add an individual to the roll. It conveys an undesirable blasé impression about the process. While unwieldy, the phrase "direct enrolment and direct update" (DE/U) will be used to refer collectively to these systems.
Another electoral reform which has in practice been tightly coupled with DE/U is the ability to enrol, or change an existing enrolment, on election day at the polling place. Again, the systems in place currently in NSW and Victoria are different in practice, but share the same putative objective. For convenience, this general approach will be referred to as "election day enrolment" (EDE). These systems are also discussed briefly in this paper.
The dissenting report of the JSCEM inquiry into the 2010 election argued against a direct update or direct enrolment system being implemented for the federal roll (JSCEM 2011). The argument made by Opposition Members and Senators3 in the dissenting report focused primarily on a concern for the integrity of the roll. The dissenting report argued that other government administrative data are often incorrect and that there cannot be sufficient confidence in the data for them to be used to populate the electoral roll. The dissenting report claimed a particular danger in using data from state governments, as they may be outside the scope of oversight by commonwealth bodies such as the Commonwealth Parliament or Auditor-General. The use of direct update and enrolment would also eliminate the existing "paper trail" provided by a signed physical enrolment form.
Opposition Members and Senators were also concerned that decisions around which data are trusted for the purposes of enrolment would inadvertently politicise the AEC, and stated that maintaining a correct enrolment is not an unduly onerous obligation for Australian citizens. Similar arguments were made in the dissenting report into the JSCEM (2010) inquiry into the NSW Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Bill 2009. Objections to direct enrolment and direct update are not along strictly partisan lines, however, with the Electoral Matters Committee of the Coalition Victorian government recommended that the VEC expand their direct enrolment and direct update program (EMC 2012).
These concerns highlight the importance of selecting reliable sources of data and scrupulously validating information before it is used to inform the roll. Operating direct enrolment and update in the most transparent manner possible will be essential for anticipating and managing claims of unfairness or partisanship.
Direct update and direct enrolment systems have been used by the NSW and Victorian electoral commissions, to some extent, over a period of time with each covering at least one state election. There is a small body of knowledge in the public domain about the operation of these systems so far, which is outlined below.
In 2009 the NSW Parliament passed the Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Act 2009, authorising the use of what the NSW Electoral Commission calls "SmartRoll". According to the NSWEC (2009), SmartRoll:
will minimise the need to complete and lodge enrolment forms by proposing that electors who have changed their address details and notified a NSW agency of that change are enrolled at their new address. The elector will be given the opportunity to object to that proposed enrolment but, if there is no objection, it is proposed that the elector is then notified of their formal enrolment.
Under SmartRoll, voters are added to the roll on the basis of change of address data sourced from the NSW Roads and Traffic Authority (RTA), the NSW Board of Studies (BoS) and the NSW First Home Owner Grant Scheme4 (Beeren 2011). Electors are notified in writing that they are to be added to the roll, including by SMS and email, and are given at least seven days to inform the NSWEC that the changes are incorrect. The amendment also allows for election day enrolment, where electors who discover they are not enrolled, or that they are enrolled at the wrong address, are able to enrol before 6pm on election day and still have their vote counted.
As at 22 June 2011, a total of 55 405 potential electors had been processed through SmartRoll, with 96.4 per cent of these persons subsequently being enrolled. Two thirds of these early transactions were based on RTA change of address data, with 20 per cent from the BoS and 11 per cent from the First Home Owner Grant Scheme (Beeren 2011). By April 2012, 326 768 SmartRoll notices had been sent by the NSWEC, with 86.5 per cent resulting in enrolments or enrolment changes. Of these notices, 93.4 per cent were triggered by RTA data, with 4.2 per cent from BoS data and 2.3 per cent from the First Home Owner Grant Scheme. The vast majority led to changes of enrolment address (82.3 per cent), with 17.4 per cent being new enrolments.
Prior to the 2011 NSW state election (26 March 2011) the largest age group of SmartRolled electors were those aged 25 years or under, comprising 37.8 per cent of the transactions, both because of new enrolments (31% of the total transactions) and because young people more frequently change address. At the time of the state election, around 40 000 electors had been SmartRolled. Of those, 19 965 had their address changed via SmartRoll, and the turnout rate for this group was 87.5 per cent. The turnout rate of the 18 996 first time enrolments was much lower, however, at 64.3 per cent (NSWEC 2011). This is consistent with research which indicates that lower levels of turnout amongst young people are common internationally (Quintelier, Hooge & Marien 2011).
The CEA does not currently allow the details of the electors on the federal roll to be updated on the basis of SmartRoll changes to the NSW roll. The NSWEC sends SmartRolled electors commonwealth enrolment forms for these individuals to sign and return, however only 12 per cent have so far elected to return them. One consequence of direct enrolment and direct update programs being run by the states, particularly NSW's SmartRoll program, is the divergence between the NSW and federal rolls. Election analyst Antony Green estimated that, if this trend continued, by the time the next federal election is called there will be 200 000 electors enrolled on the NSW roll who are not on the federal roll (Green 2011). Many of these NSW electors may have an expectation they are actually on the federal roll and thus may unsuccessfully attempt to vote at the next federal election.
The Victorian Electoral Amendment (Electoral Participation) Act 2010 came into operation on 20 August 2010, and had not been active for long before the 2010 Victorian state election, held on 27 November 2010. The amendment allows the VEC to automatically enrol eligible citizens when they turn 18, with these individuals then notified of their enrolment and given the opportunity to correct any mistakes. The system began implementing direct enrolment almost immediately, but did not start directly updating electors until March 2012.
Currently the system uses information from the Victorian Curriculum and Assessment Authority (VCAA) to enrol current and former students when they turn 18. These data are sent to Births, Deaths and Marriages (BDM) in order to identify those who were born in Victoria before being used for enrolment. The legislation also allows for enrolment at polling places on election day.
According to Mr Steve Tully (2011), Victorian Electoral Commissioner:
The package of reforms is integrally and intrinsically connected and, in short, means that every citizen over the age of 18 in Victoria who attends a voting centre in Victoria has the ability to vote.
If a person were on the roll for a particular address, they were given the vote for the corresponding electorate and if they were not on the roll at all, but claimed to be eligible, they signed a declaration envelope and either produced prescribed proof of their identity or expressly gave permission to have their identity checked with the prescribed authority.
The VEC directly enrolled approximately 1 800 18-year-olds prior to the 2010 state election (held on 27 November 2010). They plan on running the direct enrolment process for VCE students in November and December in non-election years (student residence records are at their most accurate when students are expecting notification of their year 12 results). However, because some students had not turned 18 prior to the end of the year, and the VEC was unable to send intent to enrol letter until 21 days after they become eligible. As a result, approximately 3 000 of the 2011 VCAA-based enrolments occurred in January 2012.
The initial roll-out of the Victorian program focused on enrolments, but in February 2012 VEC began using VicRoads change of address data to update existing enrolments. Approximately 10 000 new enrolments were added based on VicRoads data in each of February and May 2012, with another 6 000 in July 2012. In addition, approximately 2 000 enrolments were updated using VicRoads data each in February, May and July 2012 (VEC personal communication, 14 August 2012). As at July 2012 the VEC had processed around 50 000 transactions under the new system, approximately half of which were new enrolments and half enrolment updates.
The VEC also intends to explore additional data sources. Under their legislation the VEC is not able to use data from the AEC for direct enrolment, however the first recommendation from the Inquiry into the 2010 Victorian State Election (EMC 2012) recommended that the Electoral Act 2002 (Vic) be amended to allow AEC data to be used for the purposes of direct enrolment.
The VEC report that 86 per cent of the directly enrolled electors turned out at the 2010 state election. None of the 238 directly enrolled electors who did not vote claimed that they did not know they were enrolled (EMC 2012).
In Victoria and NSW, election day enrolment (EDE) is seen as an essential part of the direct enrolment and update strategy. Election day enrolment provides an opportunity for individuals who have been directly enrolled or updated to update incorrect details. It also enables individuals who are not affected by direct enrolment and update who attend at a voting centre on election day to enrol and vote. This affects the concept of a "close of rolls" date. The close of rolls is still an essential part of Victorian electoral administration and communications. The VEC emphasizes the need to enrol or update enrolment by the close of rolls, and regards election day enrolment as a savings provision to cater for people who for whatever reason have failed to either enrol or update their enrolment details in time.
Despite having the common aim of enrolling electors without their intervention, the NSW and Victorian DE/U systems are fundamentally different when examined in terms of their evolution and processes. These differences have potential implications for the applicability of their experience to any federal DE/U system.
The Victorian DE/U system is arguably more similar to the current federal continual roll update (CRU) program and federal DE/U system being developed, at least in terms of the underlying mechanics of the system, than it is to the NSW SmartRoll system. Unlike the NSWEC, the VEC maintains its own roll, and undertakes its own CRU program. That is, the VEC has traditionally been able to add individuals to its own roll due to receiving enrolment forms, or object them off the roll if it believes they are no longer living at their enrolled address. The Victorian roll is kept synchronised with the Victorian electors on the federal roll through the existing joint roll arrangement process, minimising potential roll divergence.
The Victorian DE/U system is an extension of the Victorian roll and CRU process. When information becomes available indicating that a citizen is eligible to enrol, rather than sending out a notification of the requirement to enrol, as in the current CRU process, the citizen is sent notification of the commission's intention to enrol them. If no response is received within a certain timeframe (14 days) that the individual is ineligible to enrol, they are enrolled and are notified of their enrolment.
As the VEC already had a roll system, and the DE/U system was an update to that rather than a discrete system of its own, it was relatively inexpensive to develop. Initially the VEC's DE/U system only used VCAA data. It has since been expanded to also use VicRoads change of address data and data from the Rental Tenancies Bond Authority. Extending the DE/U system to use these additional data was relatively straightforward, and the VEC reported that they estimate this expansion cost the VEC under $45 000.
Although the end result looks the same to the elector, the NSW DE/U system, SmartRoll, works in a fundamentally different way to the Victorian system and the federal DE/U system being developed. SmartRoll was built from scratch as a dedicated DE/U system, and while roll products can be produced from the system, it is based more around a customer relationship management (CRM) framework than a roll. The system is based around two databases – one which records addresses and one which keeps track of individuals (electors and those not eligible to enrol). The two databases are matched in order to create roll products for particular purposes. The NSW system was never designed to supplant the AEC's roll, but rather act as another enrolment channel to supplement the AEC's roll. Unlike the Victorian system it was never intended to be able to stand alone as a complete roll.
In addition to recording enrolment details and status, the database of individuals also keeps track of contacts with these individuals. This facility is central to the NSWEC's tracking and management of what it refers to as its "platinum" electors – those who have only ever been SmartRolled, who are enrolled but who have never had an elector-initiated contact with the NSWEC (platinum electors are discussed in more detail below). It also allows an individual to be flagged as being ineligible for enrolment (due to mental state or lack of citizenship) without being added to the roll. The system also allows an analysis of what categories of potential electors require the most resources to be devoted to them.
In summary, the key difference between the systems is that the VEC maintains its own roll, and uses the federal roll, CRU and DE/U to keep its roll current. The NSWEC, in contrast, uses SmartRoll to keep track of changes to enrolment of NSW electors, and applies those changes to a current extract of the federal roll when it is required for an election. The only list of electors NSWEC maintains itself is those who have been affected by SmartRoll (because they have moved or turned 18 since SmartRoll began, for instance), whereas the VEC has its own complete roll. In neither case are changes or additions to the state DE/U elector databases currently able to be automatically applied to the commonwealth roll due to the requirement for a signed enrolment form5, hence leading to roll divergence.
A DE/U system was seen by the VEC as being a logical progression of the existing Victorian roll and CRU approach. The state government of the day was approached by the VEC about legislation to allow it to directly update and enrol individuals. The legislation, when passed, came into operation not long before the 2010 state election (27 November 2010). Due to the small window before the state election, the initial use of the new powers was relatively small scale, and only focused on new enrolments, rather than updates.
The developmental process of the NSW SmartRoll system was somewhat more structured. A description is provided in the NSW Electoral Commissions Report on the Conduct of the NSW State Election 2011. Briefly, however, in 2006 the NSW Joint Standing Commission on Electoral Matters was interested in ways of increasing enrolment, particularly amongst youth and disadvantaged citizens. In 2008 the NSWEC conducted a trial at two TAFE campuses to determine how accurately data could identify eligible but unenrolled students and the reactions of the students to the initiative. The system was then developed following positive outcomes of the trial and was ready for implementation in September 2010, when the enabling legislation was passed.
Where the initial run of the Victorian system used only one data source (the VCAA school enrolment data), with its limited shelf-life6, the NSW system used data from the Board of Studies, the Roads and Traffic Authority and the First Home Owners Grant. The NSW system started with a small sample of data (with records for only 152 individuals in the first two batches). With no problems encountered, the batch numbers were increased to the thousands. Typical batches generally now contain 5 000 change of address records, with occasional runs of 10 000 records (the largest containing 10 572 records, which resulted in a 96 per cent enrolment rate).
Both the Victorian and NSW DE/U systems are predicated on requiring no response from the correctly enrolled elector as a default position. When individuals are identified as requiring enrolment or a change to their enrolment they are contacted by the electoral commission and notified that the commission intends to enrol or update the individual. If the individual is happy that the commission's intended actions are correct, they do not have to take action. If the intended actions of the commission are incorrect (for example, they are not an Australian citizen, are not eligible to vote due to mental incapacity, or have not moved), the individual must contact the commission to inform them.
Both systems allow contact with individuals via means other than the traditional paper postal letter – specifically SMS (Short Message System) and email. Currently only NSW is using these alternative forms of contact, while the VEC is sending paper letters. The electronic communications means are used in a fall-back manner. If a mobile phone number is present, an SMS is sent. If the SMS bounces, or a mobile number is not supplied, an email message is sent. If the email isn't able to be delivered, a traditional paper letter is sent. Due to the "opt-in" nature of the system, there is no requirement for the individual to post back a response. This approach has the added advantage of removing costs for return postage. An individual who believes they are being enrolled or updated in error can email or phone the commission to correct their details. Each of the commissions has a small call centre dedicated to DE/U queries. Importantly, the individual is only required to respond if their address is incorrect or they are not entitled to be enrolled.
The NSW experience so far is that 19.5 per cent of initial communications have been by SMS, with 24.7 per cent by email and the remainder (55.8 per cent) by post, with only a small number of the SMS and email messages bounced back. The email addresses and mobile phone numbers are obtained from the external data sources (RTA and Board of Studies).
The use of electronic communication is viewed as an important component in keeping the costs of the system down, postage being a major expense for electoral commissions. Both states have outsourced their communications (electronic and paper-based) to third-party providers. The currency of the electronic contact information is a potential issue. Email and mobile phone use and physical mobility vary between different demographic groups, and are likely to be particularly changeable amongst young people (ACMA 2011).
In both Victoria and NSW the public communication about the changes to enrolment were deliberately low-key and minimal. Both state electoral commissions made public announcements about the changes in the form of media releases, which received little media interest. The commissions have not made any concerted efforts to tell the general public about the changes, their communications strategy being primarily targeted at those who have been affected.
In Victoria there was effectively no dedicated communication strategy around DE/U and election day enrolment (EDE). The reforms were put in place only shortly before the election; only a small number of individuals were affected by the DE/U program, and these had been individually notified through the standard post-DE/U notifications.
Both commissions primarily use their DE/U notifications in order to inform individuals about the system. The NSWEC directs individuals to their web site at http://smartroll.info through their voter communications, whereas the VEC has a more detailed letter regarding DE/U, including where the data came from and contact details for enrolment objections.
The NSWEC has deliberately adopted a policy that the less contact that is required from SmartRolled electors the better. Under this policy SmartRolled electors only have to contact the NSWEC if their details are incorrect – no contact is required if the elector is correctly enrolled (as is also the case in the Victorian system). They have a deliberately small call centre devoted to SmartRoll (four staff) and are therefore sensitive to any increase in enquiries which may indicate increasing levels of data matching errors. In practice, the NSWEC has fielded enquiries about election events from SmartRolled voters to their SmartRoll call centre (as opposed to the regular NSWEC call centres). This suggests that SmartRolled electors have noted the specific SmartRoll contact information from their enrolment materials. Both Victoria and NSW also have the capacity to respond to enquiries electronically, and emphasise these electronic options to electors.
In both NSW and Victoria, identified potential electors are sent a letter through the postal service confirming their enrolment after the grace period for objection has expired and they have been enrolled. In NSW this letter is explicitly seen as being part of their elector engagement process. There are also efforts made to contact and engage with SmartRolled electors and to gauge the attitudes and experiences of the electors, although this is constrained by the available resources.
Both electoral commissions viewed election day enrolment (EDE) as a "safety net" for unenrolled and incorrectly enrolled individuals. If an individual had been incorrectly enrolled (either through the DE/U system or otherwise), they are still able to correct their enrolment at the polling place, and hence not be disenfranchised. While EDE is arguably a valuable addition to an electoral system's repertoire in itself, it was introduced in both Victoria and NSW in conjunction with DE/U legislation and will therefore also be discussed in this report.
The VEC emphasises that any eligible voter who attends a polling place on election day is able to cast a vote that would be entered into the scrutiny. While they do not place quite the same emphasis, the NSWEC also agreed that is also true of their approach. There are, however, subtle differences between how the states carry this out.
The most significant difference is that, in Victoria, a voter who is enrolled at an address other than their current address will have their vote cast for the division in which they are enrolled, and is then given an opportunity to change their enrolled address (ie., if they have moved to a different division, their vote will be cast for their enrolled division, not their new division). In NSW a voter attending a polling place on election day would be able to change their enrolled address and have their vote cast for the division associated with their new address. In practice, however, it is likely that some NSW electors have had their vote recorded at their old address (and previous division) rather than their current address (and current division).
Due to the federal election being held only three months before the Victorian state election, many voters were already correctly enrolled and the potential scale of the added workload associated with EDE was not apparent. The VEC expects that around 60 000 EDE transactions would occur in state elections, but only had around 30 000 at the 2010 state election. The experience of the NSWEC, however, suggests that EDE is resource intensive at polling places.
The core data sources for both the Victorian and NSW DE/U systems are the databases of year 12 students from the relevant state educational assessment authorities (VCAA in Victoria and the Board of Studies in NSW) and motor vehicle licencing authorities (VicRoads in Victoria and the RTA in NSW). The year 12 students databases are particularly useful in terms of new enrolments of eligible citizens aged 18 years, and motor vehicle licence database are particularly attractive in terms of changes of address. Both state commissions also use data from the Registry of Births, Deaths and Marriages (BDM) in order to determine citizenship and to also identify people who are deceased to remove from the roll. While the VEC are not currently using the VicRoads data for DE/U purposes, the data is being used for their CRU program, and it will be used for DE/U in the near future.
A small number of other supplementary data sources are also being used or considered. These include the Rental Tenancies Bond Authority (RTBA) and Victorian Tertiary Admissions Centre (VTAC) in Victoria and the First Home Owners Grant Scheme in NSW. There is interest amongst the commissions for other data, such as changes of address from utility companies, although no avenues currently exist for including these data.
Both state electoral commissions have a set of "business rules" that are used in order to determine the consistency of the data. While developed independently by the NSWEC and VEC, there have been recent moves toward increasing the consistency of these rules. The VEC emphasises that their legislation specified that they use trusted processes for filtering the data, rather than trusted data or trusted agencies.
The NSWEC receives drivers' license change of address data nightly and imports it into the system daily, although the data is not used to update enrolments for 30 days after it is received. The system process batches of cases. It could, however, continually import the data and update elector details. Processing the data in suitable batches is an administrative decision, rather than an inherent feature of the system. By April 2012 most (93 per cent) of the SmartRoll enrolments came from RTA data, with the BoS data comprising 4.3 per cent and the First Home Owner's Grant 2.3 per cent of enrolments.
Although the VEC had not (at the time of the consultations) started using the VicRoads change of address data for direct enrolment, they had undertaken initial trials to determine the scope of the exercise. Of approximately 35 000 cases in the data, 18 700 were matched by the business rules as suitable for direct enrolment (a "hit rate" of around 50 per cent). The VEC started using VicRoads change of address data for updates in February 2012 and record that 61 per cent in July 2012 resulted in enrolments. Interestingly 12 per cent of those contacted with an intent letter are recorded as "subsequently enrolled". In most cases this means they contacted the AEC to update their enrolment after having received the letter but before they could be processed by the VEC's direct enrolment and direct update program.
Both Victoria and NSW use the same protocol with silent electors. Any elector residing at the same address as a currently silently enrolled elector is also considered a silent, and is not directly enrolled or updated. The NSWEC and the AEC jointly reached an agreement that this approach would be appropriate, and the VEC came to the same conclusion independently.
For persons ineligible to enrol due to intellectual disability or other mental conditions the situation is less straightforward. On the whole, these individuals cannot be identified on the basis of the data received by the electoral commissions. It is possible that excluding records from schools which specifically cater to intellectually disabled students from direct enrolment might go some way toward addressing this issue, but the range of disabilities catered to by such schools will likely vary, and will not necessarily align with the expectations of the CEA. It is also likely that students who are unable to vote due to intellectual disability cannot be identified in this way. In practice, the NSW and Victorian electoral commissions rely on parents of intellectually disabled persons contacting the commission in response to DE/U notifications.
Both the NSW and Victorian electoral commissions have access to state Births, Deaths and Marriages (BDM) data and have in the past used that data to determine citizenship. A potential elector identified as having been born in the state was used as a proxy for citizenship.
The method of access of the data varied depending on the state. In NSW an extract of all births in the state for the past hundred years has been obtained from BDM, along with regular extracts of deaths data. In Victoria the list of potentially eligible electors was sent to BDM to filter out those who were not born in Victoria (or who have since died).
Both states faced the issue of directly enrolling individuals who were born in Australia and are still in Australia at the age of 18, but who are not Australian citizens. The largest group of these is likely to be children born to New Zealand citizens who live in Australia, although the actual number of these is not known (and is not expected to be large). The VEC also checks the birthplace of the individual's parents in order to identify and exclude such people.
It was anticipated by the respective commissions that those individuals who are notified they will be directly enrolled who are not Australian citizens will contact the commission to report they are not eligible to be enrolled. Since 2012, however, the Department of Immigration and Citizenship (DIAC) has agreed for the AEC to supply DIAC data on citizenship status to the state commissions. As DIAC will not necessarily have information on Australian-born Australian citizens, BDM data is still an important means of verifying citizenship.
Analysis conducted by the NSWEC on the SmartRolled electors prior to the Clarence by-election (19 November 2011) indicated that approximately 4 per cent of those added to the roll were new enrolments ("new news", as opposed to reinstated enrolments). At this point, the number of these electors who voted is not known. Overall, 17.4 per cent of electors who have been SmartRolled since the inception of the system are new enrolments.
Electors not on the roll are essentially absent for either passive or active reasons. The passive missing are those who have not enrolled due to disinterest, apathy, or simply not having got around to enrolling or updating their details. The passive missing are unlikely to object to DE/U enrolment. The active missing, in contrast, are those who have a philosophical or ideological objection to enrolment and/or voting, and refuse to be enrolled. With the exception of a trial carried out recently by the VEC to prosecute non-enrollers, Australian electoral authorities have traditionally not vigorously pursued the active missing. The details of at least some of the active missing, however, will be captured in DE/U programs (some of them will manage to stay "off the radar" sufficiently to avoid being included in DE/U data).
The NSWEC estimated during the consultations it will take four to five years of SmartRoll operation to make serious inroads into the number of missing electors in NSW. This estimate, based on the estimated number of missing electors and the number of transactions undertaken by SmartRoll, may prove to be overly-optimistic. Due to the design of SmartRoll, electors who have limited contact with the utilised data sources (e.g., non-year 12 completers, electors without drivers licenses) are unlikely to be picked up by SmartRoll.
Due to the short implementation lifetime of the Victorian DE/U system (at the time of investigation) it is too early to draw any firm conclusions on how successfully it identified and enrolled missing electors. The VEC reports that 31 per cent of persons identified from the Victorian Curriculum and Assessment Authority data were not enrolled. It is likely that at least some of these people would have enrolled at the next enrolment trigger such as an election event. Analysis of the results of more datasets being processed through the DE/U system will provide a better indication of the number of missing being enrolled.
The implementation of DE/U has so far been limited to only a small proportion of electors and only one state election in each of Victoria and NSW. The full extent of any "knock-on" effects of these programs may not be apparent for some time. It is too soon to say whether electorally disengaged electors who are directly enrolled are more likely to vote informally, or to not attend at a polling place, for example. Certain knock-on effects are, however, being anticipated and planned for.
One significant example is the NSW future commitment to SmartRolled electors. The NSWEC believes, not unreasonably, that electors who have been enrolled or had their details updated by SmartRoll will have little or no motivation to manually update their details in the future, regardless of their legal obligations. The electoral commission is therefore obligated to continue to monitor these electors and update their details (and communicate the update to the elector) when necessary. In practical terms, these electors have been freed of the obligation to maintain their own enrolment.
The NSWEC divides its SmartRolled electors into the "gold" and "platinum" groups. The gold group will have filled out an enrolment form at some point in their lives, but are now having their details updated automatically by the electoral commission. The platinum group have been directly enrolled, and can expect to be directly updated in the future. These are the electors who have never filled out an enrolment form. It is expected that the experience of being SmartRolled will change their future enrolment behaviour, leaving them unlikely to manually update their enrolment.
It is difficult to say what constitutes a good turnout when it comes to direct enrolments, as direct enrolment mainly affects young people, who historically have low levels of turnout. Some proportion of those directly enrolled may also be electorally disengaged, and therefore less likely to both enrol and vote. With these provisos in mind, both commissions report relatively high levels of DE/U electors voting at subsequent elections. In NSW 77.2 per cent of SmartRolled electors voted at the 2011 state election, with 64.3 of the newly enrolled electors voting. In Victoria 79.8 per cent of directly enrolled electors voted at the 2010 state election. Although difficult to quantify, again because of the traditionally low turnout of young people, it also suggests that the DE/U programs are resulting in correct enrolments. The results are on the basis of small, but not necessarily unrepresentative, samples of the entire electorate, and are a promising indication of the potential success of DE/U systems.
While Australian election management bodies are determinedly apolitical, they operate in a highly politicised environment. It is interesting, therefore, to note that in both NSW and Victoria the DE/U legislation was passed by Labor state governments, both of whom then lost power to Liberal/National Coalitions in the first election for which the DE/U systems operated. In both cases the electoral commissions had received no indication that the incoming government was unhappy with the DE/U system or wished to shut them down. Under its new government, for example, the Victorian Electoral Matters Committee made recommendations7 consistent with the continuing operation of direct enrolment and direct update in Victoria (EMC 2012). This is no guarantee that a new government will not, at some point, require the system to be changed or repealed. The dissenting reports from several commonwealth JSCEM hearings regarding direct enrolment and direct update indicate some reservations about the process from several federal Coalition MPs and Senators. However the state experience indicates at least some level of comfort with DE/U systems by both major political parties at a state level.
The respective DE/U systems required substantially different financial outlays, largely because the NSW system was developed from the ground up as a new system whereas the Victorian system was a relatively minor enhancement to an existing system. The NSWEC reports that it received $1.2 million from the state government to implement the system and enrol the first batch of new electors. The initial seed funding was rolled over into the continuing program. Ongoing costs of the NSW system mainly consist of the six personnel who continue to maintain the system, including four call centre staff. The actual cost of the Victorian program is not known, but is reported to be considerably lower.
A significant cost-saving of the DE/U systems is the reduced requirement to contact electors through the postal service. Even in Victoria, where the communication is still paper-based, there is a significantly reduced postage cost because most electors do not need to send back a form (thus avoiding return postage costs). In the NSW system this has been further reduced through the use of non-paper based communications. It is not known exactly what the postage savings from these initiatives are at this point, however postage costs have always been a significant expense for EMBs.
In addition to the system establishment costs associated with DE/U, there is also likely to be an additional cost in terms of adequately resourcing EDE. At this point it is not possible to gauge how much widespread use of EDE by electors would cost. In theory EDE should only need to be used for a small proportion of eligible persons for whom DE/U has not resulted in an accurate enrolment, but that theory has not been adequately tested. The quantum of that extra funding is yet to be determined.
Roll integrity is not a simple concept to quantify. It has been defined by the AEC as consisting of five elements (AEC 2012, 3):
The five-element definition of roll integrity adopted by the AEC follows the four-element definition of roll integrity proposed by the Australian National Audit Office for its audit of the integrity of the roll (ANAO 2002). The main differences between the ANAO and AEC definitions is that the AEC has replaced validity ("the electoral roll includes all individuals who are eligible to enrol", ANAO 2002, 10) with the more inclusive element of entitlement. The AEC also added "processing correctness", reflecting the AEC's awareness of the importance of the data that forms part of the system, particularly through the CRU process. A similar multi-factorial approach is taken internationally. The International Foundation for Electoral Systems has a more expansive global principles for accurate electoral rolls which incorporate the AEC's five elements (Yard 2011).
In additional to composing multiple different elements, in practical terms roll integrity is a relative measure. That is, a given process may lead to greater or lesser roll integrity in terms of any or each of the five elements.
A full discussion of assessment of roll integrity is beyond the scope of the current paper, however the expressed opinions of the NSW and Victorian commissions was that DE/U inevitably leads to greater roll integrity. The process increases the accuracy and completeness of the information contained in the electoral roll well beyond what CRU is able to achieve. It uses fundamentally the same data as CRU, however, so entitlement, processing correctness and security are not compromised. Additionally, the states believe their DE/U processes contain appropriate opportunities for any elector who is incorrectly updated to correct those updates, and the existence of EDE mitigates the risk of electors being disenfranchised through errors made by the electoral commission.
Neither of the state commissions reported that they had encountered any significant privacy issues, either from privacy authorities or electors, as part of their DE/U programs. In practice, the commissions effectively bypass the need for individuals to opt-into the DE/U systems, or be notified of the possibility of their information being passed on to the electoral commissions, by using the demand powers in their respective legislation to obtain the data. In Victoria the contact letter sent to the enrolled elector notes the source of the data and that "a number of organisations are required to provide the VEC with information to assist in maintaining the electoral roll". Clients who engage in transactions with the agencies that supply data to the commissions (especially educational assessment authorities and motor vehicle license issuing authorities) may not be aware that their information may be passed on to the electoral commission.
A move to direct enrolment and direct update of the federal roll by the AEC will have a substantial impact on AEC operations, and particularly on election, enrolment and education functions. Direct enrolment and update will likely be a powerful tool for maintaining an accurate and current roll, as part of a broader suite of enrolment strategies. The exact scope of the changes to the work of the AEC is difficult to comprehensively predict, but may include:
Direct enrolment and direct update are likely to have substantial impacts on the key indicators of the voting franchise: participation rate, turnout and formality. The reforms primarily target the participation rate, through ensuring as many eligible persons as possible are currently enrolled. As such, successfully implementing direct update and direct enrolment should improve the participation rate.
The likely impacts on turnout and formality are more complex, and it is possible that the turnout and formality rate may both decrease. Some electors who are directly enrolled and updated may have lower levels of electoral engagement, and may either not vote on election day or may vote informally (either intentionally or inadvertently). There is a risk that negative movement in these measures may be employed as "evidence" that the direct enrolment and update process has failed.
Worryingly, internal AEC research indicates that a proportion of electors already believe they are being automatically enrolled onto the federal roll. As the Victorian and NSW (and potentially Queensland) direct enrolment and update processes affect more people—these states containing well over half of the Australian voting population—this proportion will inevitably grow. Maintaining all possible consistency between state and federal enrolment processes will reduce elector confusion and may increase confidence in the electoral process.
The convergent evolution of direct enrolment and direct update in Victoria and NSW reflects different approaches to addressing a common problem: the decline in the enrolment rate. DE/U was viewed by the electoral commissions as a logical extension to the common approach of the data-driven CRU process, removing only the need for the elector to actively respond in most cases. With the increasing desire of citizens to transact with government services online (AGIMO 2011), it removes what is seen by electors and potential electors as an anachronistic step in the process of enrolment.
Due to the long history of CRU, a system which was developed jointly between the state and territory electoral commissions and the AEC, the data used for DE/U is well understood. As such, from the perspective of the commissions, the risk of DE/U to the integrity of the roll is seen as being low. Concerns from parliamentary stakeholders regarding roll integrity are important for the commissions to address, and as such an evidence base should be developed by the commissions to assuage these fears. While large-scale habitation reviews are infeasible, processes along the lines of the Sample Audit Fieldwork can be used to verify the accuracy of enrolments resulting from DE/U.
In an environment of compulsory enrolment, the state has an obligation to make enrolment as easy as possible for electors, while still ensuring that those who are enrolled meet the requirements for enrolment. If the state, via the electoral commissions, is able to accurately enrol individuals without the individual's intervention, then it appears self-evident that it should do so.
Australian Communications and Media Authority (2011). Numbering: Implications of research into consumer issues (Consultation Paper 4).
Australian Electoral Commission (1999). Supplementary submission to the Joint Standing Committee on Electoral Matters: AEC response to hearings of 19 October 1999.
Australian Electoral Commission (2005). Measuring the electoral roll: A report on the results from the Sample Audit Fieldwork.
Australian Electoral Commission (2010). AEC submission to the Joint Standing Committee on Electoral Matters Inquiry into the NSW Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Bill 2009.
Australian Electoral Commission (2011a). Submission to the Joint Standing Committee on Electoral Matters on the conduct of the 2010 federal election.
Australian Electoral Commission (2011b). Second supplementary submission to the Joint Standing Committee on Electoral Matters on the conduct of the 2010 federal election.
Australian Electoral Commission (2012). Submission to the Inquiry by the Joint Standing Committee on Electoral Matters into the Electoral and Referendum Amendment (Maintaining Address) Bill 2011.
Australian Government (2009). Electoral Reform Green Paper: Strengthening Australia's democracy. Canberra: Author.
Australian Government Information Management Office (2011). Interacting with government: Australians' use and satisfaction with e-government services.
Australian National Audit Office (2002). Integrity of the electoral roll. Audit Report no. 42 2001–2002.
Beeren, P. (2011). Is there a case for automatic enrolment? Paper presented at Challenges of Electoral Democracy Workshop, Melbourne, 14–15 July.
Brent, P & Hoffman R (2011). Electoral enrolment in Australia: Freedom, equality and integrity (pp 20–35). In J-C Tham, B Costar & G Orr (Eds) Electoral democracy: Australian prospects. Carlton: Melbourne University Press.
Electoral Commission New South Wales (2011). Report on the conduct of the NSW state election 2011.
Electoral Council of Australia (1999). Report of the 1999 Continuous Roll Update activities to update the electoral roll for the Commonwealth, states, territories and local government.
Electoral Council of Australia (2005). 2004–2005 Continuous Roll Update: A report of the Enrolment Steering Committee to the Electoral Council of Australia.
Electoral Matters Committee (2012). Inquiry into the conduct of the 2010 Victorian state election and matters. Parliament of Victoria.
Green, A. (2011). NSW Automatic enrolment and its challenge for the Commonwealth.
Joint Standing Committee on Electoral Matters (2009). Report on the conduct of the 2007 federal election and matters related thereto.
Joint Standing Committee on Electoral Matters (2010). Inquiry into the implications of the Parliamentary Electorates and Elections Amendment (Automatic Enrolment) Act 2009 (NSW) for the conduct of Commonwealth elections.
Joint Standing Committee on Electoral Matters (2011). The 2010 federal election: Report on the conduct of the election and related matters.
Tully, S. (2011). Joint Standing Committee on Electoral Matters. Conduct of the 2010 federal election and matters related thereto. Hansard, 13 April 2011.
New South Wales Electoral Commission (2009). Smart-roll legislation to streamline roll maintenance.
New South Wales Election Commission (2011). Report into the conduct of the NSW state election 2011.
Quintelier, E., Hooghe, M, & Marien, S. (2011). The Effect of Compulsory Voting on Turnout Stratification Patterns: A Cross-National Analysis. International Political Science Review, 32(4), 396–416.
Victorian Electoral Commission (nd). Changes in legislation.
Yard, M (2011). Civil and voter registries: Lessons learned from global experiences. Washington DC: International Foundation for Electoral Systems.
1 In some cases the states will use the Commonwealth roll for their state as is, whereas in others the Commonwealth roll is one component of the state roll, amongst other information.
2 At the time the CCD was the smallest geographical area that the Australian Bureau of Statistics used for collecting statistics, and averages around 225 dwellings (less in rural areas).
3 The Hon Bronwyn Bishop MP, the Hon Alex Somlyay MP, Senator Scott Ryan and Senator Simon Birmingham.
4 The NSW Smartroll is maintained as a separate system to the federal roll, and does not attempt to capture enrolment changes that are already reflected on the federal roll. This is discussed further.
5 In practice they are added to the roll as "state-only electors" for the respective state and mailed a commonwealth enrolment form to complete and return to be added to or updated on the commonwealth roll.
6 That is, the address detail in the VCAA data may be much less accurate once the school year has finished.
7 Recommendation 3.1 recommended that the Electoral Act 2002 (Vic) be amended to allow enrolments from the AEC that have not be witnessed be allowed to be used for the purposes of direct enrolment and direct update in Victoria.