The old s1073B seems clear to me. It's saying once the earned, derived or received period has been determined (separate issue) then the income will be assessed over the relevant instalment periods in such a way that the daily income will be the same for a particular instalment period. Say somone earned $10 a day for 17 consecutive days, 1 day in intalment period 1, 14 days in instalment period 2, and 2 days in instalment period 3. s1073B says that the income will be assessed over the 42 days of the 3 instalment periods. For instalment period 1, the assessed amount will be $10/14 per day for each of the 14 days. In instalment period 2, the assessed amount would be $10/day for each of the 14 days. For instalment period 3, the assessed amount would be $20/14 days per day for each of the 14 days. The ombudsman's report says that this would be correct. What Centrelink was doing wrong was assessing $10 per day over 17 days and this was incorrect because the 17 days were not contained in one instalment period.
Hi Robert, yes; that is right. But this example is an fairly straightforward case, because when the income was 'earned, derived, or received' is known on these facts. This is also a case, perhaps, of 'disaligned' fortnights. But harder problems arise when the income records do not sufficiently clarify when the income was earned, derived or received.
For many employers who were made to cooperate with Centrelink through an investigation, all the records they produced would say was that Jack had an income of $3100 for the month of January. Centrelink would then input the $3100 into its calculator, the Earnings Apportionment Tool within the ADEX Statement, and the EAT would spread the income across the whole month ($10/day) and across multiple instalment periods.
In addition, the agencies had adopted a policy position -- despite the fact that s 1073B gave them three ways to attribute income (earned, derived or received) -- only as when 'earned.' This was explicitly to stop recipients taking advantage of deferred income arrangements, where they could render themselves entitled to payments in the first fortnights of a period by deferring the 'receipt' of their income until a later time. Before the amendment of 2020, the policy was expressed in the Department of Social Services’ Social Security Guide as follows:
As a matter of policy, income test assessment is generally based on whichever event occurs first, which is usually when people earn the money. This is because assessing earnings only when received would create inequities, as it would enable some people to defer receiving their earnings until the income would have less impact on their income support entitlement. This would place people who can defer receipt of income in a better financial position than those paid on a regular basis.
(1) had accepted that payslips showing when income was received (and not earnt) could only be sufficient evidence of when the income was received; and
(2) had not been so strict in their policies about anti-avoidance, which is not, after all, in the legislation -- and is in my view contrary to Parliamentary intention on a plain reading of the provision, because s 1073B gives the decision-maker the option to analyse on a 'received' basis for a reason (it is not surplassage!); and
(3) had, based on accepting the 'when received' option as a matter of evidentiary probity (Briginshaw, etc), then used s 1073B properly, by only dividing the 'received' sum by reference to the closest nearby instalment period
Hi Chris. Thanks for your thorough reply. I realise you've done a lot of research on this, and I'm being slack on mine. I think I understand your point. If I'm understanding you correctly, you're saying that if Centrelink only know when the money was received, then it was received on one day which is therefore in one instalment period and should be asssessed evenly among the 14 days of that instalment period according to s1073B. If they want the opportunity to calculate a different debt amount, over multiple full instament periods by assessing the money as earned or derived, then they need to get more info.
Yes, that is pretty much what I am suggesting. But I'm also venturing an explanation for the scale of the problem now too.
As you know, s 1073B gives the agencies options to calculate a person's income when it is 'earned, derived, or received.' But what I am noting is that Centrelink *always* elected to calculate it on the basis that is was *earned* in x instalment period.
If they were not so gung-ho on that policy, then maybe they would not have so easily misapplied s 1073B.
I am venturing the possibility that as soon as they adopted a policy of income being assessed when it was earnt (rather than derived or received), they always thenceforth had a tendency to 'spread' the income forwards in time. For example, if your bank records and payslip both said you received $3100 at the end of the month, Centrelink would spread/apportion the daily rate across the whole month. What I am saying is that this tendency was at least partly caused by this urge to claim that the income was 'earnt' earlier in time than when it was received.
As noted earlier, the reason they want to make this policy assumption was to avoid people delaying the actual 'earning'/reciving of their income (through asking the boss to defer payment) until later in time, and thereby obtaining the benefit of being entitled to Centrelink payments for all the fortnights during which they deferred payment.
Obviously, the legislation gave the agencies options so that they could monitor any such attempts to avoid tapered or cancelled entitlements in instalments periods when money was earnt, derived, or recived (now, after all, the legislation just goes by 'received'). However, surely no legislative drafter ever expected Centrelink to *always* assume that it was earnt as early as possible (but that became the policy); and surely, further, it was expected that, even if Centrelink did want to 'spead' the income back in time to avoid precisely this issue, they would be constrained in doing so the the *single instalment period.*
But that's not what happened.
When we consider the issues in this way, it becomes harder in my mind to say that Centrelink made an honest mistake. In being, if I may, paranoid and hypervigilant about people deferring their incomes and taking advantage of the entitlements that would thereby accrue, Centrelink overcorrected; they imputed 'earnt' income to people further back in time than the provision allowed. They were, it seems, blinded by their policy position -- which was probably a combination of hypervigilance and ethical 'right-sideism' -- to the fact that they were actually acting ultra vires -- beyond their legal power.
Thanks Chris. I understand better now. Hopefully the mess can eventually be sorted and s1073A (i'm assuming this is new too?) will make it easier for people on benefits reporting their income (or their employers reporting it to Centrelink electronically) and for the Centrelink staff processing too.
Thanks for reading/viewing Len. In my view, yes. But not just because bank statements may have been subject to apportionment under s 1073B.
I have argued here (https://welfare.substack.com/p/centrelinks-net-to-gross-earnings) that they should be reviewed and cancelled because the net to gross income calculator uses annualised ATO PAYG data to identify the 'gross-up' ratio and applies that ratio to the net amounts in the bank statements. Given that the annual data is the source of the gross-up rate, they would suffer the same defect identified in the Amato reasons -- namely, that 'the conclusion that the Applicant had received social security benefits' to which they were not entitled 'was not open on the material before the decision-maker.'
Beyond this, however: yes, I agree any apportionment of the gross-up bank statement information across more than one instalment period would be reason to review and, preferably, cancel the debts. If the agencies wished to pursue the debts, then they must recalculate from the ground up using the ATO-issued tax tables and lawfully applying the apportionment rule in 1073B.
But there should also be an amnesty and an immunity on any debts that exceed the original calculations in recognition of the vice of the unlawful original debt decision.
Just thinking that in view of 1073B, there may be a strong(er) argument that all those "Rbodebters," effectively coerced into providing bank statements and thus missing out on having their debts "zeroed" following court cases & change of policy should now all be zeroed & refunded.
Thanks Len. Yes, I raised this with ACOSS; the s 1073B apportionment problem is likely to have affected a fair whack of those who co-operated both before and after the bank statement alternative proof point was imposed: ie, those in category 3a, 3b, and 4.
The recalculation of each of these robodebts in the three different circumstantial categories should not be set aside from the review because they were subject to 'resolution' via the robodebt settlement process.
That's because each of them (3a = overpaid false debt; 3b = underpaid false debt; 4 = possible recalculation on bank statements or income reports) may still be affected by s 1073B apportionment, no matter what source of data was used, especially since the debts were pre-single touch payroll and calculated before the 7 December 2020 repeal of s 1073B.
Yes. Some will be in the invidious and unsustainable position of having been wronged three ways: first by robodebt, then by the N2G residual ATO averaging (to coin a phrase: 'bank statement robodebt'), and thirdly by unlawful s 1073B apportionment.
The point about cooperation disfavouring those who acted in good faith is a difficult one legally; I am yet to consider the actions in detail (eg, action on the case for wilful injury?). But the paradox at the level of ethics is obvious. I agree there is probably an arguable remedy.
The old s1073B seems clear to me. It's saying once the earned, derived or received period has been determined (separate issue) then the income will be assessed over the relevant instalment periods in such a way that the daily income will be the same for a particular instalment period. Say somone earned $10 a day for 17 consecutive days, 1 day in intalment period 1, 14 days in instalment period 2, and 2 days in instalment period 3. s1073B says that the income will be assessed over the 42 days of the 3 instalment periods. For instalment period 1, the assessed amount will be $10/14 per day for each of the 14 days. In instalment period 2, the assessed amount would be $10/day for each of the 14 days. For instalment period 3, the assessed amount would be $20/14 days per day for each of the 14 days. The ombudsman's report says that this would be correct. What Centrelink was doing wrong was assessing $10 per day over 17 days and this was incorrect because the 17 days were not contained in one instalment period.
Hi Robert, yes; that is right. But this example is an fairly straightforward case, because when the income was 'earned, derived, or received' is known on these facts. This is also a case, perhaps, of 'disaligned' fortnights. But harder problems arise when the income records do not sufficiently clarify when the income was earned, derived or received.
For many employers who were made to cooperate with Centrelink through an investigation, all the records they produced would say was that Jack had an income of $3100 for the month of January. Centrelink would then input the $3100 into its calculator, the Earnings Apportionment Tool within the ADEX Statement, and the EAT would spread the income across the whole month ($10/day) and across multiple instalment periods.
In addition, the agencies had adopted a policy position -- despite the fact that s 1073B gave them three ways to attribute income (earned, derived or received) -- only as when 'earned.' This was explicitly to stop recipients taking advantage of deferred income arrangements, where they could render themselves entitled to payments in the first fortnights of a period by deferring the 'receipt' of their income until a later time. Before the amendment of 2020, the policy was expressed in the Department of Social Services’ Social Security Guide as follows:
As a matter of policy, income test assessment is generally based on whichever event occurs first, which is usually when people earn the money. This is because assessing earnings only when received would create inequities, as it would enable some people to defer receiving their earnings until the income would have less impact on their income support entitlement. This would place people who can defer receipt of income in a better financial position than those paid on a regular basis.
(see https://parlinfo.aph.gov.au/parlInfo/download/legislation/billsdgs/7203508/upload_binary/7203508.pdf)
If the agencies
(1) had accepted that payslips showing when income was received (and not earnt) could only be sufficient evidence of when the income was received; and
(2) had not been so strict in their policies about anti-avoidance, which is not, after all, in the legislation -- and is in my view contrary to Parliamentary intention on a plain reading of the provision, because s 1073B gives the decision-maker the option to analyse on a 'received' basis for a reason (it is not surplassage!); and
(3) had, based on accepting the 'when received' option as a matter of evidentiary probity (Briginshaw, etc), then used s 1073B properly, by only dividing the 'received' sum by reference to the closest nearby instalment period
then, arguably, none of this would have happened.
Hi Chris. Thanks for your thorough reply. I realise you've done a lot of research on this, and I'm being slack on mine. I think I understand your point. If I'm understanding you correctly, you're saying that if Centrelink only know when the money was received, then it was received on one day which is therefore in one instalment period and should be asssessed evenly among the 14 days of that instalment period according to s1073B. If they want the opportunity to calculate a different debt amount, over multiple full instament periods by assessing the money as earned or derived, then they need to get more info.
Hi Robert -- all good!
Yes, that is pretty much what I am suggesting. But I'm also venturing an explanation for the scale of the problem now too.
As you know, s 1073B gives the agencies options to calculate a person's income when it is 'earned, derived, or received.' But what I am noting is that Centrelink *always* elected to calculate it on the basis that is was *earned* in x instalment period.
If they were not so gung-ho on that policy, then maybe they would not have so easily misapplied s 1073B.
I am venturing the possibility that as soon as they adopted a policy of income being assessed when it was earnt (rather than derived or received), they always thenceforth had a tendency to 'spread' the income forwards in time. For example, if your bank records and payslip both said you received $3100 at the end of the month, Centrelink would spread/apportion the daily rate across the whole month. What I am saying is that this tendency was at least partly caused by this urge to claim that the income was 'earnt' earlier in time than when it was received.
As noted earlier, the reason they want to make this policy assumption was to avoid people delaying the actual 'earning'/reciving of their income (through asking the boss to defer payment) until later in time, and thereby obtaining the benefit of being entitled to Centrelink payments for all the fortnights during which they deferred payment.
Obviously, the legislation gave the agencies options so that they could monitor any such attempts to avoid tapered or cancelled entitlements in instalments periods when money was earnt, derived, or recived (now, after all, the legislation just goes by 'received'). However, surely no legislative drafter ever expected Centrelink to *always* assume that it was earnt as early as possible (but that became the policy); and surely, further, it was expected that, even if Centrelink did want to 'spead' the income back in time to avoid precisely this issue, they would be constrained in doing so the the *single instalment period.*
But that's not what happened.
When we consider the issues in this way, it becomes harder in my mind to say that Centrelink made an honest mistake. In being, if I may, paranoid and hypervigilant about people deferring their incomes and taking advantage of the entitlements that would thereby accrue, Centrelink overcorrected; they imputed 'earnt' income to people further back in time than the provision allowed. They were, it seems, blinded by their policy position -- which was probably a combination of hypervigilance and ethical 'right-sideism' -- to the fact that they were actually acting ultra vires -- beyond their legal power.
Thanks Chris. I understand better now. Hopefully the mess can eventually be sorted and s1073A (i'm assuming this is new too?) will make it easier for people on benefits reporting their income (or their employers reporting it to Centrelink electronically) and for the Centrelink staff processing too.
Thanks for reading Robert. Yes, the new s 1073A is part of the new regime where, crucially, income is always calculated when received. This sidesteps the whole issue. An additional provision was also added to stop the deferral concern: s 1073BB: http://classic.austlii.edu.au/au/legis/cth/consol_act/ssa1991186/s1073bb.html. FYI I go over the old sections in (too much) detail here: https://youtu.be/RhFHWes8qBQ
Asking for a friend - does this strengthen arguments that all Robodebts based on bank statements should be reviewed and cancelled?
Thanks for reading/viewing Len. In my view, yes. But not just because bank statements may have been subject to apportionment under s 1073B.
I have argued here (https://welfare.substack.com/p/centrelinks-net-to-gross-earnings) that they should be reviewed and cancelled because the net to gross income calculator uses annualised ATO PAYG data to identify the 'gross-up' ratio and applies that ratio to the net amounts in the bank statements. Given that the annual data is the source of the gross-up rate, they would suffer the same defect identified in the Amato reasons -- namely, that 'the conclusion that the Applicant had received social security benefits' to which they were not entitled 'was not open on the material before the decision-maker.'
Beyond this, however: yes, I agree any apportionment of the gross-up bank statement information across more than one instalment period would be reason to review and, preferably, cancel the debts. If the agencies wished to pursue the debts, then they must recalculate from the ground up using the ATO-issued tax tables and lawfully applying the apportionment rule in 1073B.
But there should also be an amnesty and an immunity on any debts that exceed the original calculations in recognition of the vice of the unlawful original debt decision.
Hard agree.
Just thinking that in view of 1073B, there may be a strong(er) argument that all those "Rbodebters," effectively coerced into providing bank statements and thus missing out on having their debts "zeroed" following court cases & change of policy should now all be zeroed & refunded.
Thanks Len. Yes, I raised this with ACOSS; the s 1073B apportionment problem is likely to have affected a fair whack of those who co-operated both before and after the bank statement alternative proof point was imposed: ie, those in category 3a, 3b, and 4.
The recalculation of each of these robodebts in the three different circumstantial categories should not be set aside from the review because they were subject to 'resolution' via the robodebt settlement process.
That's because each of them (3a = overpaid false debt; 3b = underpaid false debt; 4 = possible recalculation on bank statements or income reports) may still be affected by s 1073B apportionment, no matter what source of data was used, especially since the debts were pre-single touch payroll and calculated before the 7 December 2020 repeal of s 1073B.
Sure, but fwiw I think Robodebts calculated/recalculated due to the provision of bank statements (arguably coerced?) are especially egregious.
All of these alleged debts would have ultimately been zeroed had the alleged debtors done nothing.
ie those that tried to comply with Centrelink's (unlawful?) demands were ultimately worse off than those that didn't.
Yes. Some will be in the invidious and unsustainable position of having been wronged three ways: first by robodebt, then by the N2G residual ATO averaging (to coin a phrase: 'bank statement robodebt'), and thirdly by unlawful s 1073B apportionment.
The point about cooperation disfavouring those who acted in good faith is a difficult one legally; I am yet to consider the actions in detail (eg, action on the case for wilful injury?). But the paradox at the level of ethics is obvious. I agree there is probably an arguable remedy.