Traders' Roundtable

Hi,

I check the JB closing price data from several sources. It seems that the data are not very clean in 1999, particularly at the end of the year. For example prices are constant from the end of sept until the end of nov 1999. Did you experience the same? Like to have clean data before testing!!!

Alexandre

A sad reality of life is that nobody runs a worldwide clearinghouse for the unambiguous assignment and dissemination of securities symbols. Each individual exchange is more or less free to make up its own symbols as it pleases, without regard to overlap with symbols used by other exchanges. From time to time, securities are de-listed (removed from the exchange) and, at the whim of the exchanges, those symbols may be reassigned to other securities.

Additionally, data vendors have a tradition of assigning their own symbols to various securities. The Reuters symbol is often different than the Bloomberg symbol is often different than the Metastock symbol, for example.

Which brings me to JB. Some people might feel that JB refers to (a chain of hifi stores in Australia), while others prefer to think it's (either 5-year or 10-year government bond futures traded in Tokyo), while others immediately think of (a trucking company in Arkansas that ships goods for Wal-Mart), and still others think of (10 Year Japanese government bond futures traded in Chicago). "JB" means different things to different people.

Citibank changed from "CCI" to "C" when Chrysler (formerly "C") got bought by Daimler-Benz. Sun Microsystems recently changed from "SUNW" to "JAVA". Orange Juice futures are "OJ" on some quote machines and "JO" on others. The NASDAQ exchange is rolling out a multi phase process for (changing the symbols) of its listed stocks. And on it goes.

Sorry.
I meant JB = Japanese 10y Govt Bond Future (Exchange Tokyo).

Alexandre

Shown below is the data supplied by CSI for 10 year Japanese Government Bonds traded on TSE. ("commodity number 158" in CSI parlance). Its behavior in October through December 1999 doesn't appear unusual, at least to me.

This is a backadjusted continuous contract; rollover parameters are shown in the second image.

The data is correct.

Due mainly to the liquidity problem of the Cheapest-to-Deliver cash bonds and due partly to the Y2K issues,
market participants unprecedentedly skipped DEC99 ( look at the volume ! ) and rolled to MAR00.

Sluggo, JGB10, thanks for your answers.

Sluggo, on you chart, the quality of the data in mid september doesn t seem very good. On my side, prices are constant through november. For both of us it must have to do with the issue described by JGB10. You roll on the 4th day, and I roll on the 2nd day, which is very similar. I checked the open interest as suggested by JGB10, and I can confirm the problem with the DEC99 contract. We would certainly be better by rolling on open interest and volume. I guess that if you change your roll trigger to OI, the september month will be better on your chart.
For trading do you generally use a fixed day of the month for the roll as in your mail, or the OI and Volume?

JGB10 (great name to answer this question!), what was the reason for this liquidity problem with the cheapest-to-deliver at this time? Is there any other similar problem at other points in the past?
Also, I am a bit surprised that Y2K was such a big issue. It should have been an issue for other contracts too. Was it so?

JGB10, I re-read and saw that you wrote "mainly" for the Cheapest, and "partly" for Y2K. So I conclude that Y2K was not the big issue.

JGB10 wrote:Due mainly to the liquidity problem of the Cheapest-to-Deliver cash bonds and due partly to the Y2K issues,
market participants unprecedentedly skipped DEC99 ( look at the volume ! ) and rolled to MAR00.

Thanks for the explanation! I added a couple new features to the "Futures Data Investigator" (link to downloadable Blox source code) that specifically look for this sort of volume and/or open interest crash.