Some time ago, a poster here (I believe it was Tim?) pointed us to the 1995 Joint Committee on Taxation report “Issues Presented By Proposals To Modify The Tax Treatment Of Expatriation”. This report was issued during one of the U.S. periodic witchhunts against the alleged flood of rich people who were “fleeing the country to avoid the estate tax”, and made public the names, birth dates, and dates of loss of citizenship of nearly a thousand people who had chosen to stop being Americans between 1 January 1994 and 26 April 1995.

I’m going to be doing a couple of posts over the next week or so based on this data, so I’ve decided to put it up here in machine-readable format; the original document contained only scanned pages images and not text, and thus was not easily searchable. I have not published the actual names or birthdates here, out of respect for the privacy of people who made individual decisions to give up U.S. citizenship more than a decade and a half ago, for whom no law at the time ever authorised that the data be made public, and who are no doubt going about their lives today with no desire to dredge up the distant past (and for those who have continued to pay attention to U.S. politics, a distinct sense of relief that they are not the targets of today’s IRS Offshore Crusades).

But first, a rant

Even a cursory eyeballing of the data shows that the premise for the 1995 Congressional witchhunt was utterly false: judging by names, the vast majority of people giving up citizenship were former immigrants who had elected to return to their home countries and resume their old citizenships, or accidentals and children of mixed-nationality couples who had just reached majority and chose to retain their other citizenship. And the accidentals were hardly wealthy heirs or heiresses; Google searches on some of the names find that they’ve gone on to become ordinary professionals. One of these alleged “fat cat traitors” is now a mid-level civil servant in a Southeast Asian government, another a law professor, a third an economist in South America. The minor issue of factuality, of course, did not slow down the demagogues one bit, and they quickly passed legislation to make the exit tax paperwork more onerous even if you owed no tax, and to give the Attorney-General the power to ban ex-citizens from ever re-entering the country, even for a visit.

In going through the data, I noticed some interesting things. There was one Japanese man whose relinquishing act occurred in 1941. There are toddlers who were recorded as relinquishing — when it is clear that a toddler cannot have the requisite mental intent to perform any such legal act. And finally, even in 1995 when the U.S. was still booming and the “Asian Tigers” like Taiwan and South Korea had still not reached “developed country” status, there were a lot of young people from those countries who gave up their U.S. citizenship, apparently concluding that opportunities would be better right where they were. I’ll explore these themes more in future posts.

Blatant errors in the data point to the lack of care with which these records were maintained. The U.S. Embassy in Seoul in particular appears to have been extremely sloppy about all this, not to mention ridiculously slow in processing the paperwork. Two people have their “application date” listed as being later than their “CLN issue date”, while numerous others have an “application date” earlier than their “loss date”. One man had his loss date recorded as his birth date. Others were recorded as renouncing on dates which do not exist such as the thirty-first of September. You can find an elderly man allegedly by the name of “Herr” and a young woman with the name “Frau”, their actual first names having gone unrecorded. There was a young man whose name is recorded as “Test”.

The government of a multi-trillion dollar economy made tax policy based on this garbage data set. I’ve seen two-person startups which exercise more care in maintaining the data collections which they use to make hundred-dollar decisions about web page design. But look on the bright side: I doubt the Congressional demagogues like Lloyd Doggett bothered to do the slightest statistical analysis of any of this data anyway. They just went based on stereotypes and gut feelings and started making their idiotic speeches claiming that renunciants had “discover[ed] one day that the Port Royal Golf Course in Bermuda is their hometown”.

Overview of the data

The list is contained in Appendix H; that’s the 305th page of the document, if you want to navigate to it in your PDF viewer. It contains 978 names. The report consists of a scanned PDF, which is neither searchable nor accessible to people who use screen-readers; I used the open-source OCRopus to convert the scanned image into a text file, and then went through and corrected it by hand. (OCRopus has lower accuracy than Tesseract, but in my experience it has better layout analysis, meaning it handles columnar data more accurately).

I’ve made the data into an HTML table with the following columns:

• Age: the age of the person on the day when the Certificate of Loss of Nationality was issued. The original table contained the birthdate; for privacy reasons, I substituted it with the age instead.
• Loss date: the date of the act which caused the person to lose U.S. citizenship. For renunciants, this is the date of the Oath of Renunciation; for relinquishers, this may be the date of naturalisation in a foreign country, or of swearing an oath for military or political service.
• Application date: the date the person went to a U.S. consulate to report their relinquishing act or swear the oath of renunciation
• CLN issue date: the date when the State Department issued a CLN to the person.
• Delay: the number of days between the “application date” and the “CLN issue date”. This was not in the original report, but I calculated it myself. The average is 109 days (71 days for renunciants, 127 days for relinquishers), but with an extremely high skew: many people got their CLNs in under a week, whille others were left waiting for more than a year, especially in Seoul where the embassy very clearly seems to have practised “batching”. The longest wait was for a renunciant whose CLN took nearly threee years to get issued.
• Classification: whether the person is a relinquisher or a renunciant. This was not in the original report; see below for the methodology I used to classify people as renunciants or relinquishers. The ratio of renunciants to relinquishers is about 5:4; this is slightly lower than Isaac Brock’s own renunciation data maintained by calgary411 & Pacifica, in which the ratio is closer to 6:4.

Classification as renunciant or relinquisher

I’ve counted as a renunciant anyone whose act causing loss of nationality was performed on the same day (the “loss date”) that the act was reported to the consulate (the “application date”). In almost all cases, the only way that could happen was if the act in question was execution of an oath of renunciation. There may be some rare cases where a person performed a relinquishing act in the morning and then ran over to the U.S. Consulate on the same day in the afternoon and got an appointment — that is to say, I may have miscounted some relinquishers as renunciants.

Unfortunately, a large number of the entries do not give either the “loss date” or the “application date”. The report states:

The State Department does not generally maintain on its computer system the date of application, because this date is not meaningful for State Department purposes … the State Department was asked to review source documents to ascertain this date if possible. The date of application was not included in the State department printout and was supplied separately, to the extent available, by the State Department’s Office of Consular Affairs for citizens who had performed an expatriating act other than formal renunciation before a Consular Officer.

The dates entered in Tables H-1 and H-2 reflect the dates contained in the State Department lists even if there was a question as to whether the date was correct. A blank in any column on the tables indicates that the relevant information was missing from the State Department computer printout. N/A and NR mean that the date could not be readily obtained from the State Department files when the State Department was asked to supply application dates.

On that basis, I’ve classified all the people with non-blank “loss date”, blank or missing “application date” as relinquishers; in most cases this seems to make sense given the other data available, for example, the “loss date” being several years before the “CLN issue date”, or the person having a name which indicates they probably come from a country where they would be able to relinquish rather than renounce if they resumed their former citizenship. The only exception I made here is for the married couples (same surname, same “loss date”) where one’s “application date” was recorded but the other’s was missing, in which case I assumed they both had the same application date.

Records in which both the “application date” and “loss date” are missing are listed as “unclassifiable”. I can’t think of any particular reason why the ratio of renunciants to relinquishers would be different among those records as compared to the complete records, so hopefully this doesn’t introduce another source of error into the analysis.

The data