The results are in – see how the Australian Business Register (ABR) has been performing in "Survey 21".
Earlier this year, we asked you to complete an online questionnaire to confirm that your business details are correctly recorded in the Register.
We've published the results and commentary in Measuring ABR Integrity 2016.
You can also download Integrity of the Australian Business Register 2016 (PDF, 1.2MB)
On this page:
The ABR is Australia’s source of registered business information. A whole-of-government resource and valuable public asset, the data in the ABR is regularly accessed and used by both the private and government sectors to facilitate business and government interactions.
It provides a unique trusted identifier for Australian enterprises (the Australian business number – ABN) and continues to be developed as the nation’s accurate and complete source of business information.
Measuring ABR Integrity is an annual program to determine accuracy of ABR data (ABN in use and key fields) by sampling ABN holders across all entity types with an even mix of characteristics (eg recent and aged registrations) to answer questions regarding the data we hold.
The survey was conducted from 6 April 2016 (initial email dispatch) to 31 May 2016 (conclusion of phone follow-up) using data extracted on 1 April 2016.
As in 2015, ABN holders with an email address were sent an email with a link to complete the questionnaire online. Those without an email address were sent a letter with information to access the questionnaire online. ABN holders were given two weeks to complete it. Those who didn't respond were followed up via phone. The results were then sent to the ABS for statistical analysis.
The methodology is the same used as last year.
Between 2015 and 2016, the overall ABN in use rate increased, with the majority of key fields also showing improvement. The concept of “minor/insignificant error” was first introduced in 2015 and can now be compared in conjunction with “completely correct”. While what might be classed as insignificant or minor is subjective, the interpretation is the data is useable from the ABN holder’s perspective (or the data would have been marked as “incorrect”). The results are summarised below.
* Useable = Completely correct + Minor/insignificant error
Completely correct 2015
Completely correct 2016
Change to Completely correct
Change to Useable
With our ongoing compliance program (including bulk cancellation), overall ABN in use rate has improved by 5.4% (4.3 percentage point increase). Comparing the completely correct rate across the key fields (excluding Legal Name which is already over 95%), there is an average improvement of 3.2% (1.7 to 4.6 percentage point increase in 5 out of 7 key fields).
We can now compare data useability by combining “completely correct” and “minor/ insignificant error”, which shows improvement across all key fields (except for Legal Name which moved from 99.6% to 99.4%).
For key fields, the data integrity program in 2015–16 focused on completeness (eg missing address elements) and validity (eg invalid email format). Accuracy (currency) of data, which is the primary purpose of the survey, is still reliant on ABN holders to update the ABR (or other tax systems where synchronisation with the ABR is in place).
The compliance program, in addition to improving overall ABN in use rate, also contributed to the improvement in key field accuracy (as non-operating business tended not to keep their information up to date). We have ongoing Client Information Review (CIR) campaigns for ABN holders with no update for a set period to prompt them to review and update where necessary. In 2015–16, we also piloted a targeted CIR campaign for incorrect number of associates (eg partnerships with 0 or 1 associate).
The findings from Measuring ABR Integrity 2016, in conjunction with other intelligence (eg Agency Data Quality Survey), have shaped our data integrity program for 2016–17. For example, we are piloting “reconciliation to source” initiatives for contacts and associates as they have the lowest accuracy rate. We will continue with targeted CIR campaigns for other data categories. Email address is another key area for improvement.
The following sections will present accuracy measures for ABN in use and each key field in more detail with breakdowns by entity type as well as other findings from the survey.
ABN status by entity type
Marked improvement in the ABN in use rates for companies, partnerships and trusts can be attributed to our bulk cancellation program, using Australian Securities and Investments Commission (ASIC) deregistration data and new criteria for partnerships and trusts. Superannuation funds and government entities remain stable as expected.
For the first time, we have seen the ABN in use rate peaking for individuals despite ongoing bulk cancellation program and desk-based compliance activities. In order to gain insight into this finding, we have analysed the raw data from the samples to see if there is any emerging pattern.
In the survey, 480 individuals were sampled with 255 responding. 194 out of 255 (76%) indicated their ABN was in use. For the remaining 61, we asked why and from when the ABN was not in use. The results (with 1 who did not answer) are profiled below.
ABN not in use (individuals)
Within 12 months
1 – 2 years ago
3 – 5 years ago
More than 5 years ago
Used the ABN for a while but no longer operating
Got the ABN to start an enterprise but did not go ahead
Changed entity/business structure
Applied in error – discovered not required for the activity
Other (please specify)
The standout result is that 41 out of 60 (68%) indicated the ABN was not in use within two years. Within this timeframe, we can't cancel the ABN using our current bulk cancellation criteria and will have to rely on other compliance activities or voluntary cancellation. Also worth noting is that “enterprise not going ahead” is the 2nd most common reason (excluding “Other”). Lastly, analysing the text specified in the “Other” category revealed that most could have chosen one of the set answers instead (highlighting there is always some degree of user error in a “survey” environment).
Legal name by entity type
Legal name accuracy remains stable and high across all entity types. For those who indicated “minor/insignificant” and “incorrect”, we asked for more information. Analysing the information provided, the residual inaccuracy has a large degree of “user error”:
Postal address by entity type
Postal address accuracy continues to improve across most entity types mainly due to removal of non-operating business. This is a key field that primarily relies on ABN holders to inform us of changes. We can't automatically use other data sources as ABN holders can use different postal addresses in their dealings with the ATO and other government agencies.
Unsurprisingly “address changed” is the most common reason when the ABN holders indicated the address is incorrect.
Business address by entity type
Business address accuracy continues to improve across most entity types mainly due to removal of non-operating business. This is another key field that primarily relies on ABN holders to inform us of changes. There is potential in reconciling with the business address provided in income tax returns to detect change.
As email address is an optional field, ABN holders were given different choices when responding to whether the recorded email (or lack of) was an accurate reflection of their circumstances. The results are presented below.
In compiling a “completely correct” figure, “recorded email correct” and “still don’t have email” are interpreted as such (reflecting the ABN holder’s view that what we showed was correct). This is consistent with how we presented the figure in 2015 to allow comparison.
Email address by entity type
If we focus just on those with e-mail address recorded (see table below), the overall accuracy is higher (87.3%) which compared favourably with other key fields (eg postal address which is at 86.1%). The high accuracy is attributed to the email validation and cleansing program conducted in 2015–16. All new registrations since January 2015 already have email addressed validated. In 2015–16, all stock email addresses were re-validated with blatantly invalid ones (eg firstname.lastname@example.org) removed from the system.
Breakdown by entity type (using those with recorded email as base)
Recorded email correct 2016
Recorded email incorrect 2016
For those with no email address recorded (see table below), instead of being an accuracy measure, we can interpret the results as that over 60% of ABN holders expected their email address to be recorded. While the email address might not have been provided initially, this is a reflection of their current mindset.
Breakdown by entity type (using those without recorded email as base)
Still don’t have email 2016
Have email but not recorded 2016
In 2016–17, email validation and cleansing program will continue (including correction to common domains, eg @homail to @hotmail) to further improve email accuracy.
Industry activity by entity type
This measure refers to the accuracy of the industry activity (using the text description of the 4-digit ANZSIC code). As background, ABN applicants provide their main business activity as free text.
Prior to December 2013, the system coded the ANZSIC using backend logic (and the applicant did not see what we coded at the time of registration). Since December 2013, we introduced a point of capture coder (POCC) where a “smart” dropdown is presented with choices (based on the free text entered) for the ABN applicant to select from to improve accuracy.
As evident from the results, accuracy continues to improve across all entity types due to the continual use of the POCC. We also have ongoing data integrity program to fix invalid ANZSIC ranges (9999x) in aged registrations. Superannuation funds are outliers as the ANZSIC code is predetermined (minor inaccuracy is attributed to the ABN holders’ perception that they should be described differently).
Contact name(s) by entity type
* Useable = Completely correct + Minor / insignificant error
The results are somewhat mixed mainly due to the volatile nature of the data element (contact more likely to have changed than address) and the lack of proactive update in general. In particular, if there are multiple contacts recorded, we showed them all. Hence, overall completely correct accuracy is the lowest across all key fields (and there is a 1% decline between 2015 and 2016). On a positive note, the usability rate has increased by 4.6%, suggesting at least one contact is correct.
Associates by entity type
* Useable = Completely correct + Minor/ insignificant error
Associates accuracy remains stable across all applicable entity types (individuals and government entities by default have no associate). Partnerships and superannuation funds are more prone to membership changes and experienced minor deterioration in completely correct accuracy (also due to the lack of proactive update in general).
Also worth noting is that for companies and trusts, business rules only enforce minimum number of associates at the time of registration. There is an additional response choice “all listed associates correct and there are more not listed”, which is counted as “completely correct” for consistency with previous year.
Measuring ABR Integrity is run as a review by which information to be completed online was requested under Section 15 of the ABN Act. Subsequent phone follow-up also emphasised the non-voluntary nature of the review. We also added to the script to advise ABN holders that non-compliance may be subjected to entitlement review.
Measuring ABR Integrity
The final completion rate is comparable to previous year. Feedback from phone follow-up indicated non-completion was mainly due to exhausted lists of contactable phone numbers and e-mail addresses, rather than stalling or downright refusal to comply.
Analysis of de-identified responses (for those who had indicated information shown was incorrect) indicated there is still a lack of understanding of obligation to update the ABR when circumstances changed (36.6%). Top four responses are as follows:
Reason for not updating the ABR 2016
Didn't know you needed to
Didn’t think about it
Provided the details to the ATO for another reason and thought they would flow through to update the ABR
We also asked about ease of update and reason if the response was “difficult” or “very difficult” (based on free text analysis showing the top two responses).
Ease of update 2016
Neither easy nor difficult
Reason (collated from free text) for indicating difficult / very difficult to update 2016
Could not update desired information online
Difficult/complicated / confusing process
There are still over 25% of ABN holders experiencing difficulties to update the ABR whilst the reasons point to lack of understanding and possibly issues with obtaining or using AUSkey.