Firearm tracing, in its broadest sense, refers to the tracking of a firearm from ‘cradle to grave’—that is, from manufacture to (its eventual) deactivation, destruction or legal export. In this scenario, a firearm is traced from its source through ‘different points in its line of supply’ to its eventual removal from the registration record. The trace line shows the passage of transfer between manufacturer, importer(s), dealers, owners and (if it occurs) police possession. If methodically followed, this process improves the likelihood of identifying the site of diversion if the firearm is transferred into the illicit market. This represents the ideal for authorities engaged in firearm regulation and control but an ideal that has proved difficult to realise.
The preventative provisions outlined in the UN Protocol incorporate action to prevent the illicit transfer of firearms, their parts and components and ammunition across state borders, but some of the measures serve domestic agendas too. Of particular pertinence to Australia is the management and exchange of information on firearms that, along with the application of unique identification marks, is fundamental to firearm tracing. This section reviews the current status of firearms information management in Australia and where improvements in the collation and sharing of this information could occur.
The problem with data
Issues around quality, consistency and standardisation of data are certainly not unique to the collation of information on firearms. The quality of the data used for the analysis, which was dependent on the recording of consistent data from different agencies, and subsequent discussions with project stakeholders have shown this to be the case. The tracing of firearms in Australia has been compromised by two factors—a general absence of historical data and issues around current standard data collection procedures, information sharing and resources. These factors have restricted the scope of the analysis that could be achieved for this study.
There are numerous reservoirs of primary firearms data in Australia. These include:
- state and territory police firearm registers;
- material inventories, ballistic library inventories and record systems of firearms in police possession (ie firearms surrendered to, seized by or otherwise appropriated by police) administered by police services;
- the Integrated Cargo System, Firearms Tracking System and Detained Goods Management System administered by ACBPS;
- the Defence Export Control System administered by the Defence Export Control Office. This Office controls the export of firearms and ammunition through the issuance of permits and licences, and import of firearms under Regulation 13E of the Customs (Prohibited Exports) Regulations 1956 (Cth); and
- the Attorney-General’s Department Firearm Policy Unit, which manages applications for the importation of certain firearms prescribed under Regulation 4F of the Customs (Prohibited Import) Regulations 1956 (Cth).
At the most fundamental level is the data collected by state and territory police firearm registries. Firearm registers compile information on licensed firearm owners and dealers and the firearms registered to them. As discussed earlier, only handguns were subject to compulsory registration in Australia before the period of firearm reforms described earlier, and not every state and territory required registration of long-arms. The absence of a nationwide registration system for long-arms contributed to the phenomenon of the grey market, or the assemblage of long-arms that sit outside the legal pool that, and while not necessarily used or owned by persons involved in criminal activity, can and do flow into the illicit market. These firearms are effectively untraceable—records might exist on their place of manufacture and/or year of import and the circumstances of their seizure (if recovered by police), but documentation on the transfer of ownership between these two ‘life markers’ is often missing. It is for this reason that the pattern of long-arm diversion in Australia tends to be opaque, as shown in the analysis section where the majority of seized long-arms were denoted as originating from the grey market. Compulsory handgun registration does provide for better historical data and hence contributed to better delineation of sources regarding common points of diversion. The large amount of unknown or missing data, however, indicates underlying problems with data gathering and recording.
Among the resolutions from the National Firearms Agreement (1996) was the establishment of an integrated licence and firearm registration system in all jurisdictions. All states and territories adopted or modified their systems to incorporate a licensing scheme for persons to possess/use firearms and a registration scheme for firearms. Western Australia implemented a system different to other jurisdictions whereby the register is a record of firearm licences, permits and approvals, rather than a register of firearms per se. The licence details particulars about the licence owner and the types of firearms owned by the licence holder, which are then recorded in the register.
Together with the implementation of nationwide registration of all firearms was the recommendation that state and territory firearm registers be linked to enable the exchange of information. The original model for information exchange, still in operation, is the NFLRS, which is administered by CrimTrac. NFLRS stores data on registered, lost, stolen and destroyed firearms, licence holders and licensed firearm dealers and can be linked to other CrimTrac-administered police reference systems. The data on the system, however, is not complete and there are problems with misclassified and miscoded records originating from police registers (Project stakeholders personal communication 28 November 2011).
The ultimate goal is the implementation of an integrated national firearms licensing and registration (or national firearms management) system that would allow information on firearms to be electronically transferred between jurisdictions. The primary purpose of such a system is two-fold—to facilitate law enforcement agencies’ capability in tracing the movement of firearms throughout Australia and to streamline existing licensing processes for firearm owners. At its simplest, the system would allow state and territory firearm registers to ‘communicate’ (particularly important for reconciling inter-jurisdictional movements of firearms, which has been the cause of, or method for, diversion) but could incorporate links to other government firearm data resources. Two studies have already examined the logistics of establishing a NFMS but a final product has yet to materialise.
Historical data shortcomings are likely affected by ‘weaker’ firearm laws and past ambivalence to recording firearm movements, compounded by technological limitations in the documentation of large amounts of data. In more recent decades, or at least since the firearm reforms, impetus and (presumably) technological capacity have both been present to produce better records on firearms. A case in point was the absence in the ACBPS of a centralised electronic recording system for firearms data and the changes made since the early 2000s to improve the consolidation of this information. In 2002, the ACBPS first started recording the serial numbers (and other firearm markers) of all handguns released from ACBPS custody (as part of the Category H Handgun Certification Scheme). This was followed by the introduction in April 2006 of the Detained Goods Management System, which allowed for the capture of serial numbers from all detained firearms into the one centralised database. The firearms that were now being recorded in the Detained Goods Management System accounted for the vast majority of firearms legally imported into Australia, such as those contained in commercial shipments and firearms that required safety testing on entry into the country. In response to a 2008 resolution from the then Ministerial Council for Police and Emergency Management Police around the collection and dissemination of firearm data, the ACBPS commenced the third phase in its consolidation of firearm data by recording information (including serial numbers) on all legally imported firearms. Starting from 1 January 2009, the ACBPS began recording serial numbers from all legally imported firearms, including those firearms entering Australia that had not been previously detained due to importers not having the relevant documentation with them at the time of import because they were not subject to safety training testing requirements. These firearms were typically ‘accompanied firearms’ that were entering Australia through the passenger stream.
Yet to achieve a consolidated record of firearms data, such as envisaged for a NFMS, that would permit straightforward firearm tracing, some fundamental processes are still in need of mastering. Many of these are specific to the recording of firearm information at the registry level, but the fundamentals of technical expertise and improved data recording practices extend to the maintenance of firearms data in a number of the other listed data systems, particularly with regard to import and export, seizure/recovery, ballistics and firearm disposal records.
Achieving the fundamentals of firearm data recording
First capture recording of firearm identification and other features
A critical fundamental in producing data suitable for tracing is the accurate recording of a firearm’s identification marker—the serial number—and other classifying features (eg make, model). Previous audits of serial number data have returned high error rates (Project stakeholder personal communication 30 November 2011), including evidence for high duplication rates. Data on other classifying features of recorded firearms (eg make, model, calibre) have been similarly compromised, although this is not as much a problem as incorrect serial numbers.
Firearm identification is highly technical and requires considerable proficiency and knowledge. The technical nature of firearm identification creates the risk (and the reality) that personnel, such as staff in firearm registries, may not always have the knowledge or training to accurately record the features that are vital to identifying individual firearms. The quality of recorded serial number data is particularly affected—serial numbers, depending on the make of a firearm, can be located on different components of the firearm and their visibility is not always obvious. Some firearms may also have multiple stamps or have been poorly stamped, thus rendering the serial number difficult to distinguish, but these anomalies may not be (or cannot be) noted. A lack of expertise in identifying or locating the serial number may result in an incorrect serial number entry—an erroneous ‘nil visible’ record, the model number recorded instead, or a modified or truncated version of the full serial number. Additional information regarding the location of the serial number is additionally pertinent, particularly where a serial number is located on a firearm component that is not accountable under firearm legislation (eg a slide versus a frame).
Serial numbers, while the most important firearm identifying marker, are not 100 percent unique and hence it is equally important other identifying features—make, model, calibre, action—are also captured correctly. An accurate record of these features is particularly useful if a firearm returns to police attention and the serial number has been defaced. Again, technical expertise is often necessary to properly identify or recognise these additional markers.
A further complicating factor relates to the initial capture of firearm data. In firearm registries, this is often in handwritten format, on registration forms completed by licensed owners. This carries the additional risk of inaccurate data being recorded if there is a misinterpretation or misreading of handwritten entries, or the provision of misspelt or otherwise inaccurate information.
Comprehensive training in firearm identification is an obvious response to rectify inaccurate recording practices and ideally would extend to all personnel responsible for extracting identification material from firearms. Training is a resource issue and outside crucial roles, such as in ballistics, might not always be feasible. One method being used in firearm ballistics is to compile digital images of firearms for examination, but this is not a practicable option in other data recording contexts, not least because electronic filing would be unmanageable. Instead, other measures need to be applied that assist in self-correcting and/or standardising identification material, as discussed below.
A further, exacerbating factor in the collation of consistent, quality firearm data is a lack of data standardisation. Different systems are operated across the states and territories, a few of which have been upgraded or replaced in recent years. Resource issues do not always permit regular, methodical data cleaning (which systems may benefit from) and hence first-level data entry is a crucial step in maintaining accurate records. Data entry systems relying on free text fields and no autocorrect function can (and do) produce multiple variations of the same classifier item (eg calibre) and the structure by which serial numbers are entered. Some of this inaccuracy may originate in the technical competency of the original recorder, but it is also created by human error in data entry and a lack of consistency produced by multiple data entrants.
The creation of standardised templates (at least for important classifier data items), and filtering functions that validate item combinations and force prompts that constrain the length, type and format of alphanumeric entries would have two practical effects. First, it can prevent and correct inaccurate or incorrect items produced by typical data entry mistakes and potentially prevent the entry of misidentified items. Actual autocorrection in the latter situation is really only viable for classifiers such as make, model and action but potentially could flag, where other information is correct, problems with the configuration of the serial number. A number of firearm register systems already employ these functions, such as filtering, and the use of drop-down or standardised templates. Victoria Police have developed a series of standardised templates—the Weapon Identification System, or WIDS—that are available on their website to assist firearm owners to correctly identify their firearm. Verification searches can be undertaken if the owner knows the make or model of their firearm and retrieve information on other firearm characteristics. For example, if the owner knows their firearm has a model name of ‘700 Special’ the system retrieves the related template, which determines the firearm type (handgun), firearm category (‘H’), firearm make (‘Astra’), action (‘semi-automatic’) and calibre variants (.32AUTO) associated with that model type.
The current standard of jurisdictional firearm data, however, creates difficulty in trying to link a particular firearm record to the data stored within any firearm identification system. Until the existing data is subject to manual preliminary cleansing, the correlation of existing records to preferred identification standards will remain a problem. Once a preliminary cleansing has taken place then the preferred data standards may be further applied.
Removing the data gaps
The problem of the quality, consistency and standardisation of data collected for registration and evidentiary purposes is not a new one; nor are the remedies proposed to improve the quality of firearms data. These remedies, however, are not options that can be achieved quickly or without considerable financial investment, and jurisdictions and bodies such as the Firearm and Weapons Policy Working Group continue to make and consider methods of improvement.
An evident and important outcome of improving data quality is the elimination of data gaps, the next step in achieving the goal of tracing the life course of a firearm. Data gaps are created and sustained by dirty data and incomplete or unconnected systems of information exchange. The NFLRS was established as a national database to support information exchange on all registered firearms and licensed owners. It is used by firearm registries to upload new records and is available to registry staff and operational police to conduct record searches. An alternative or replacement to NFLRS has received consideration in past years, based on the creation of a single shared ‘authoritative identity record’ for each individual firearm, and onto which updates in its movement between custodians is documented. Key entry-into-the-system or transfer flags would include import or export, sale, inter-jurisdictional transfer, theft or loss, surrender, seizure, recovery, deactivation and destruction events. This deceptively simple premise, however, was determined, in the model proposed, to require substantial investment. The status quo was hence retained, albeit with incremental improvements in data quality assurances at the jurisdictional level and alternative approaches to better document and alert incoming jurisdictions of the transfer of firearms, an event known to be associated with an increased risk of firearms being lost to the system.
Stakeholders for this project suggested that an information-sharing scheme founded on linked records still represented the ideal solution to safeguard accurate data and minimise the emergence of data gaps. One option would be a simplified version of previously recommended products comprising a distributed database with single records for firearm and licensed owners. Jurisdictions would retain custodianship of their data but maintain communication with each other through a data linkage system based on firearm and licensed owner records. The specifics of a linkage system require further exploration that cannot be accommodated in this report but would necessitate adaptation to a common ontology for the classification of firearms. The importance of such a common ontology has featured in broader discussions by the Firearm and Weapons Policy Working Group on the development and instalment of a National Firearms Identification Database.
To resolve the suitability of a distributed database, stakeholders in the project further suggested the possibility of conducting a limited-scale study involving two jurisdictions (1 large, 1 small) to estimate the cost of integrating to a data linkage system. An important component of the study would be for participating jurisdictions to determine what is being lost (in time, resources and efficiency) with their current system and what might be gained through integration. If integration does prove to be the more efficient approach, it would help promote the creation of better, more consistent data. Efficiency in data collation and dissemination would also assist in freeing up crucial resources for additional compliance monitoring and auditing work.
Past practices, as evidenced by the data collated in the NFTD and discussion with project stakeholders, have resulted in certain data useful or critical to the tracing of firearms being captured only recently (such as serial number data on import and registered long-arms), being captured inconsistently or not being captured at all. Incomplete or incompatible data hamper (or potentially render impossible) the back-capture of information. At a minimum, it prevents more sophisticated analyses of firearm markets and adds qualifiers to the strength of the findings discussed earlier. At a more critical level it potentially impedes law enforcement agencies to reconcile firearms data during different stages of a firearm’s history. Yet when done well, it can help prevent or at least flag where firearms have been lost to the system and disrupt the flow of firearms into the illicit market.
Important steps have been made in the collection of firearm data, compelled by inter-governmental and domestic policies, enabled by technological capabilities and encouraged by genuine need to trace firearms. Further steps in standardising and harmonising data on a national level are still needed, although these steps are still being explored, with the dual purpose of ensuring that the logistics of application are achievable. If complemented with a system that supports cross-jurisdictional and cross-agency data transmission or access, and training of personnel in the accurate recording of firearms information, the compilation of Australian firearm data will be of a quality that promotes the efficient tracing of firearms and, consequently, a targeted enforcement response.