While the 2014 FATCA information transfer to the IRS was widely reported, since then we have had no idea how much data has been flowing from the ATO to the IRS. To get a better idea of the scope of the data exchange, Carl sent an FOI request to the ATO for a summary of the data sent to the IRS under FATCA for all three reporting years that have now been completed (2014, 2015, and 2016). The ATO complied with this request in a timely manner, sending us a pdf file of a printout of an excel worksheet that spans several pages both vertically and horizontally. 
FATCA requires Australian financial institutions (very broadly defined) to report account holder details as well as account balance, dividends, interest and other income paid, and gross proceeds from sale or redemption to the ATO for transmittal to the IRS. It is evident from the graphs below that the amount of data going to the IRS has exploded since the initial data transfer of 2014 data (transferred 30 Sept 2015).
Before I report the detail, a few caveats:
- The ATO have advised that the data is raw data. It includes all data that passed their validation and were sent to the IRS. However, some of these files were rejected by the IRS and resubmitted. This would result in the affected accounts being double counted. The ATO gave us no indication of the magnitude of this problem.
- It appears that not all accounts reported tax residence. The ATO only gave us the number of accounts reporting US or Australian tax residence and a list of other tax residence countries included in the report. The tax residence numbers and numbers of individual and entity accounts relate to all accounts, not just AUD denominated accounts.
- The ATO provided us with separate balance columns for each currency in which accounts were denominated. There are significant balances denominated in other than AUD, but the graphs above and the balances below include only AUD denominated accounts. USD denominated accounts were 38% of total balances in 2014, 27% in 2015 and 14% in 2016. In all years USD and AUD denominated accounts combined were more than 98% of total reported balances. There’s no indication of the number of accounts by currency.
Summary of Accounts Reported
|No. of Accounts||44,120||291,605||862,339|
|AU tax residence||10,501||23,510||124,702|
|US tax residence||11,988||17,679||72,007|
|AU tax residence||1,373||5,517||38,031|
|US tax residence||713||1,602||11,563|
From this we can see that the proportion of Australian resident accounts has been rising. If reporting tax residence is equally likely for Australian and US residents, then around 60% of individual accounts and 75% of entity accounts are Australian-resident accounts. As fewer than 30% of accounts have tax residence reported in 2016, drawing any firm conclusions on the proportion of Australian resident accounts is difficult.
The increasing proportion of entity accounts over time is expected due to the later deadlines for identifying entity accounts relative to individual accounts.
AUD Denominated Amounts Reported (A$ millions)
Yes, you read those numbers correctly – almost A$184billion of AUD denominated accounts were reported to the IRS for 2016! If you include USD and other currency accounts (translated to AUD), the total reported balances are more than A$216 billion – that’s 6% of the total financial assets of Australian businesses and households excluding superannuation!
Australia is not a large country. Imagine how much data is flowing into the IRS from the 113 jurisdictions that have signed FATCA IGAs. The IRS must be drowning in FATCA data. Given the IRS’ current budget, along with the additional demands place on the agency by the recently passed Tax Cuts and Jobs Act, this is not something the IRS will be able to easily match to 1040s and FBARs. Given the terms of the IGAs, many of the records sent will be false positives. And most non-resident US citizens live in high-tax jurisdictions where their FTC will offset most (if not all) of their US tax liability. For all but the largest accounts, it is simply not cost effective for the IRS to do anything with data from non-resident taxpayers.
Compiling all of this data and sending it to the IRS is not free. What does the ATO (and the Australian government) get for the added compliance costs imposed on banks, account holders, and the ATO? Professor William Byrnes of Texas A&M has written that, even for the US, FATCA costs more than it returns in tax revenue. In Australia, the CRS implementing legislation requires the Commissioner of Taxation to report to Parliament on the data sent out under CRS but not the data received! And there’s no report required at all for FATCA data. Without these figures, how will we ever know whether this global experiment in extreme financial transparency is paying any dividends?
Where do we go from here?
It would be ideal if we could get our banks or the ATO to automatically notify account holders when their data is sent overseas (either under FATCA or under CRS). Unfortunately, this will probably require legislation. In the meantime, anyone who believes their data has been sent to the IRS should consider asking the ATO to confirm the details that have been sent. Yes, we tried this before. And the ATO were not particularly cooperative. But there have been some developments since then. The Inspector General of Taxation published a Review into the Taxypayers’ Charter and Taxpayer Protections. Chapter 6 on on Cross-Border Information Exchanges contains the following:
6.53 Once taxpayers are made aware of the EOI, they must be given an opportunity to review, correct or appropriately contextualise any relevant information obtained.
Given that the ATO agreed with this chapter of the Review, they should be willing to provide individual taxpayers with any information sent under FATCA without prior approval from the IRS. Should the ATO respond in the same way as they did a year ago, we would be happy to assist in preparing an official complaint to the Inspector General of Taxation.
 A big thank you to the volunteers who re-constituted the spreadsheet and proofread all of the numbers.