While perusing the ever-fascinating Federal Register, I came across this Comment Request from last month, asking for submissions from the public about collection of information under Notice 97-19 and 98-34.
What are those, you might ask?
Notice 97-19 and Notice 98-34 provide guidance regarding the federal tax consequences for certain individuals who lose U.S. citizenship, cease to be taxed as U.S. lawful permanent residents, or are otherwise subject to tax under Code section 877. The information required by these notices will be used to help make a determination as to whether these taxpayers expatriated with a principal purpose to avoid tax.
The Comment Request also helpfully gives the “estimated number of respondents”: 12,350. Another similar notice from December 2000, pertaining specifically to Form 8854, gives the “estimated number of respondents” as 11,000.
First, a bit of backstory. I’d been looking for these numbers on-and-off for a few months now. This is part of my ongoing quest to sanity-check the numbers derived from the Federal Register “name-and-shame” list — by finding anything to which those numbers could possibly be compared, and seeing whether or not it matches up. Under the Paperwork Reduction Act, whenever the IRS creates a form, it has to file a submission with the Office of Management and Budget giving an estimate of the number of responses, the time burden, and the cost burden for their latest piece of red tape. Every IRS form displays an OMB control number (“OCN”) with which you can look up the history of the submissions related to that form.
The OCN on the current version of Form 8854 is 1545-0074. However, if you go over to reginfo.gov, put that OCN in the search field, pick the “ICR” option (for “Information Collections Review”), and find the most recent submission, you’re greeted with an estimate of 152 million filers. Half the country renounced citizenship? Apparently not. As I soon figured out, the IRS now uses a single OCN for most forms relating to individual income tax returns, and lumps them all together in a single estimate of Form 1040 filers. Oddly enough, it was not always this way: old versions of Form 8854 have a dedicated OCN, 1545-1567, but it was discontinued in 2006.
I’d given up on finding these numbers for a while, until today when I came across the above Federal Register notice, almost by accident — I was idly browsing through instances of the word “expatriate” on .gov sites, and landed on the page for that notice. A conspiracy theorist might think the government made a deliberate effort to keep people from finding this number. But I’m a firm believer in Hanlon’s razor: “Never ascribe to malice that which is adequately explained by incompetence”. And indeed, the history of OMB submissions on Form 8854 demonstrates plenty of incompetence, as I’ll describe further below.
As we all remember, back in 1996 Congress had a two-minute hate against renunciants, which resulted not just in the creation of the Reed Amendment, but new taxes as well. After the IRS designed its new procedures for collecting those taxes, they submitted two estimates to OMB. The first was for the “Expatriation Information Statement”, giving an estimate of 11,000 respondents per year and a time burden of 2.5 hours per respondent. The second was for “Guidance For Expatriates”, giving an estimate of 12,300 respondents per year and a time burden of about half an hour per respondent. (I’m not sure how one “responds” to a guidance notice, but anyway …)
The exit tax regime as it stood in 1996 consisted of ten years of taxation on U.S. source income; however, as the old instructions make clear, people losing citizenship would file Form 8854 only once. If they were found to be Evil Unpatriotic Tax Avoiders, they would file other forms in subsequent years, but not Form 8854 itself. That means the IRS was expecting 11,000 people to turn in their passports or green cards every year. I have no idea how the IRS came up with that number. The Federal Register list did not yet exist at that point; the only basis the IRS would have had for their estimate was this Joint Committee on Taxation report listing all the people who relinquished or renounced in 1994. Maybe they had access to some data which is not available to members of the public. Or maybe they were just padding their estimates.
The subsequent IRS submissions to the OMB on the “Expatriation Information Statement” maintained the same estimate of 11,000 respondents per year (November 1998, December 2000, November 2003, February 2004, February 2005, May 2005). Submissions on “Guidance for Expatriates” actually increased the estimated number of respondents slightly to 12,350 (May 1998; April 2000, April 2003, April 2006, June 2009).
Accountants American Jobs Creation Act of 2004 drastically increased the number of 8854 filers. As the 2007 version of the 8854 instructions points out, “covered expatriates” after June 2004 have to file Form 8854 annually for 10 years. However, the IRS did not adjust their estimate of the number of filers. Similarly, after 2008, the HEART Act switched the system from a 10-year tax on U.S.-source income to a one-time exit tax on worldwide asset appreciation. So the IRS should have decreased their estimated number of filers, as Form 8854 returned to a one-time filing and the number of people filing under the old ten-year regime slowly decreased. It seems that the IRS simply couldn’t be bothered to update their estimates in response to changes in the law.
More amusingly than that, you may have noticed that the 2009 extension was received by OMB one day before the expiration of the 2006 extension. This is part of a larger pattern during Geithner’s reign at Treasury: last-minute or late filings of mandatory paperwork, just like the “name-and-shame” list. Since the IRS apparently waited until the night before the deadline to get their paperwork in, I guess they simply didn’t have any time to update the estimated number of filers. It’s like a bad joke from university days that would fit right into Animal House — the professor changed the assignment, the paper is due tomorrow, but the frat brothers are still handing in some piece of junk they copied from the guys who took the class ten years ago, with the expectation that they’ll keep getting a “Gentleman’s C”.
The question we’re all wondering about: was that initial estimate correct? Are there really 11,000 filers of Form 8854 per year? The only way I can really think of to approach this problem is to look at the numbers which the IRS provides in its other OMB submissions, and compare those to the actual numbers of filers of the corresponding tax forms.
For Form 3520, the form on which you report your “foreign trusts”, the IRS’ notices give a very low estimated number of responses, even when compared to the IRS’ own statistics about the number of Form 3520 filers. For example, the most recent notice gives an estimate of 1,320 filers. However, the actual number of 3520 filers shown in the Statistics of Income Bulletin (SOI) — as we discussed earlier — is closer to 2,000, and the actual number of people who are supposed to be filing the form is a few orders of magnitude higher. Similarly for Form 3520-A, the most recent submission gives an estimate of 500 filers, while the IRS’ own statistics stated 3,819.
Statistics for other forms turn out to be harder to interpret. Consider the foreign-corporation-related Form 5471. The IRS’ July 2010 OMB submission estimates 86,130 responses (actually a decrease from the December 1982 estimate of 88,000), while the Winter 2006 SOI Bulletin states (at page 208) that 11,873 U.S. corporations controlled a total of 78,249 foreign corporations. One Form 5471 must be filed for each CFC; one entity may thus file multiple 5471s.
Does the OMB notice count each filing of the form as one respondent, or each filer? The time burden estimate in the OMB submission comes out to 50 hours per respondent, about the amount of time you’d expect for filling out a single 5471 — suggesting that they’re counting on the basis of forms filed. But there are other categories of Form 5471 filers too: U.S. corporations with minority stakes in foreign corporations, as well as natural persons (such as every U.S. Person entrepreneur who dares to incorporate where he lives instead of back in the Homeland). So the OMB number is probably an underestimate.
The data for Form 1065 (partnership return) is equally difficult to interpret. The 2008 OMB filing estimates 22 million responses, whereas a July 2010 Comment Request estimates 2.4 million responses. An SOI Bulletin from 2008 states that the IRS received returns from 3 million partnerships with a total of 19 million partners. It’s unclear whether these “estimated numbers of respondents” are referring to the partners or the partnerships. (Form 1065 is supposed to be filed by the partnership itself.)
Perhaps the basis for the number given is just chosen at the whim of whichever IRS employee is filling out the form. I’m sure they didn’t imagine that a member of the public would be so obsessed with the topic as to go around checking their numbers for consistency …
Form 990 and 990-EZ
Finally, a straightforward case, almost! In 2008 there were 192,968 filers of Form 990, and 238,914 filers of Form 990-EZ — for a total of 431,882. The OMB notice in effect for that period gave an estimate of 434,569 filers, almost exactly on the money (presuming that it includes both forms) — except that, oddly enough, the IRS filed an amendment to that notice that same year with a lower estimate of 403,068.
Another relatively straightforward case. The OMB submission estimates about four million filers, while SOI statistics show about three million. So the OMB submission had an overestimate of about 25%. “Close enough for government work”, as the phrase has it …
If I were as careless filling out my Forms 1040, 1116, 2555, 3520, and 5471 as the IRS very clearly is when they fill out these OMB submissions, I’d have been fined 300% of my assets by now. This lackadaisical attitude towards accuracy makes it hard to interpret the data with any confidence. As far as I can tell, the IRS at least got their estimated number of filers in the right ballpark for Forms 990, 1016, and Form 5471, but they may have drastically flubbed Form 1065, and it’s hard to say anything at all about Form 3520 since the actual number of filers is so much smaller than the number of people who are supposed to be filing. And of course, don’t forget that FinCEN estimated 400,000 FBAR filers back in 2011, but then ended up getting more than 600,000.
Anyway, the second quarter “name-and-shame” list should be appearing sometime in the next week. Whether you think that list is nearly complete or grossly understated, it’s a safe bet that the data will be full of errors — just like in previous lists which were full of repetitions and misspellings, and just like all the other data I’ve discussed in this post.