Follow-up on RED ALERT - Don't Trust TIC data

Tuesday, June 07, 2005

The US Treasury has indirectly reacted to stories (see post " RED ALERT - Don't trust the TIC data" and the links there) pointing out that the Treasury Capital Inflow (TIC) data got altered retroactively to an extent that cannot be explained with statistical adjustments only. Rob Kirby posted the following news on that story at Financial Sense.
Rob Kirby: I wrote a piece a couple of weeks ago about the Pirates - no - not the ones who work in Pittsburg (a great steel town). The other ones who steal in the Caribbean in an article titled Pirates Reprise. The long and short of the story outlines how the U.S. Treasury's numbers don't add up. I received assistance with this story from a Dutch journalist named Willem Middelkoop - who was first to recognize the inconsistency of the Treasury's reporting based on an earlier article I had written. It's my understanding that Mr. Middelkoop primarily works in television. The discrepancy in the numbers prompted him to contact the US Treasury, citing my article, and query them on the numbers. He received a reply from none other than Mr. Tony Fratto , a fellow who apparently works in a capacity of public relations for the US Treasury. He forwarded Mr. Fratto's comments about my article to me as a FYI (for your information) - or journalistic pleasantry. Mr. Fratto suggested that the author (that would be me) "succumbed to a bit of hyperbole" in suggesting that anything was amiss at the good ole' US Treasury and their reporting of numbers. In fact, he said that there was nothing unusual or brazen at all with the reporting of the subject numbers.
Well, I must admit that I'm very flattered that Mr. Fratto has taken the time to read my article. It's perhaps the nearest thing to the truth the man has seen in quite some time. Since I've got his attention I would like to point out that, thanks to alert readers, we have found even more that appears to be egregiously wrong with the same set of numbers in the Pirates Reprise article. In the first data set, Norway is omitted. In the second data set, Norway is retroactively inserted to the tune of 35 billion dollars alone in the subject month of Jan. 05. With a show of hands please, does this appear normal to any of you?
Mr. Fratto was not the only person who should have an informed opinion who passed judgment on this article. I heard from a former high ranking official (who shall remain nameless for now) of the New York Fed who had this to say:
Dear Rob
Re - Pirates Reprise
I enjoyed your article on the Fed data. I was once chief of a division at the NY Fed that complies a good deal of this data for treasury.
A few points:
I noticed these discrepancies the minute they were out since I track this data set and must update it by hand. I agree that the revisions are large. Several facts may explain the changes although I have not placed calls to find out why. For now, it's a survey break...
But rest assured that having 'financial reporters' blame the problems and the risks of the world on HEDGE FUNDS gets us no closer to understanding the facts. It just gives the devil a name- and maybe the wrong one. In this case I think they were stabbing in the dark...
Some of the revisions may be pure error. Here I agree that I'd be shocked to know that the US treasury could not get straight numbers from other foreign governments. Or that the registration system has failed that badly. But there are myriad technical issues as well. Its not as simple as you might think...Maybe it is a result of getting the date of ownership wrong since treasuries are traded widely. It might also reflect some confusion over who owns a security for the purpose of the survey if it is OUT on repo...
However, there may be some discrepancy between book and market value if someone has changed their method of evaluating holdings - (unlikely but possible).
There is also the possibility of instrument mishandling. For example in the US FNMA securities are not treasuries but to foreign governments they are often treated with almost the same respect as investments in US governments (implicit government guarantee and all...). Were they improperly tabulated for a time as 'treasury holdings'? It's possible.
As for Caribbean hedge funds, well the new investor of de-merit is the hedge fund. We no longer have nameless speculators we can call them hedge funds and SEEM more authoritative.
Another problem with the treasury holder by location data is just that these data are by location HELD. These data DO NOT trace holdings back to the ultimate beneficial owner. For example the Japanese securities firms (...and I was once chief economist at the US offices of one the Big Four firms) sometimes have bought and sold US treasuries in a subsidiary in the Caribbean (NOTE - NOT a hedge fund) the purchase in this case is attributable to the Caribbean not to the Japanese. When Merrill Lynch UK buys bonds it is a holding by a UK resident, etc...
So for data on US treasury holdings by nonofficial holders the geographical reporting is not very clear. Large institutional and investment fund buying occurs in Switzerland Luxembourg, the UK, the Netherlands and Germany, as well. We have no idea who the ultimate beneficial owner of these investments are or where they really reside.
So look at the data and think about them. I am not opposed to your ire in this matter - indeed I share it but for the simpler reason that these huge revisions were released hidden - if you will - without comment... oh they called a 'series break'... never mind. It's the sort of thing US authorities are not supposed to do. They are supposed to be above all this.
Guess they aren't.
Notice that the official holdings data were revised but holdings were made higher there.
Perhaps Mr. Fratto should instead busy himself reading the Treasury's numbers prior to publishing rather than commenting on my hyperbolic efforts, ehhh? Hyperbolic efforts, last I heard, just happens to play third base.
The Prudent Investor notes with high interest that the Treasury spokesman did not give any reason for the altered data.


Wikinvest Wire