Japan futures, contract lifetime liquidity distribution...

General discussions about futures.
ecritt
Roundtable Knight
Roundtable Knight
Posts: 134
Joined: Sat Aug 28, 2004 3:44 am
Location: Phoenix, AZ
Contact:

Japan futures, contract lifetime liquidity distribution...

Post by ecritt » Fri Dec 24, 2010 6:34 pm

The distribution of liquidity (open interest and volume) over a contract's lifetime appears to very different for commodities in Japan, relative to commodities traded in other countries. Typically, an individual futures contract begins trading with very low levels of open interest and volume, which rise modestly over time, and then rise dramatically as the deliver month approaches, and then plummet as the delivery month is entered. However, in Japan commodity futures seem to start out with high levels of liquidity, open interest and volume reach their apex early in the contract's life and then decline slowly for many months, to almost nothing as the delivery month approaches.

In other words, the "liquidity bulge" for most futures contracts occurs during the 1 or 2 months just before the delivery month. But the "liquidity bulge" for commodities in Japan tends to occur 8 to 12 months before the delivery month.

Does anyone understand what causes this phenomenon and why it seems to be specific to commodities in Japan?
Attachments
Japanese_VS_USA_PalladiumFutures.jpg
Japanese_VS_USA_PalladiumFutures.jpg (188.74 KiB) Viewed 8366 times
Japanese_VS_USA_GoldFutures.jpg
Japanese_VS_USA_GoldFutures.jpg (180.77 KiB) Viewed 8366 times
Japanese_VS_USA_OilFutures.jpg
Japanese_VS_USA_OilFutures.jpg (171.09 KiB) Viewed 8366 times
Japanese_VS_USA_CornFutures.jpg
Japanese_VS_USA_CornFutures.jpg (183.18 KiB) Viewed 8366 times

RedRock
Roundtable Knight
Roundtable Knight
Posts: 939
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

Post by RedRock » Fri Dec 24, 2010 10:09 pm

Is it customary for the back months to be top step for speculators? Hmm. I'm curious too what the 'reason' might be.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Sat Dec 25, 2010 4:38 am

I had conversations with one of the Japanese exchanges after I had noticed this phenomena. Can't recall which though doubtless I recorded the details somewhere or another. The answer I got from this particular official at this particular exchange was that retail speculators are a very important part of the market in Japan and they like to trade the very far out contracts - they feel they get more bang for their buck that way.

I roll my contracts based on volume rather than open interest in Japan. Which may be a mistake - it means that back adjusted contracts are generated backwards not forwards. There is no option in CSI to generate contracts rolling on volume forwards and this may have unwanted consequences.

Yet another reason to develop software independent of CSI UA. All the more important to do as Sluggo does and to archive the entire folder of CSI back adjusted contracts each day.

Infuriatingly, rolling on OI can lead to a fall off in volume before you get an exit signal while rolling on volume can lead to the absurdity of switching from a later contract in to an earlier one (if you get a spike in volume around contract expiry for instance) which can ripple back in its effect over several years and alter your backtested results and make positions come and go in your theoretical portfolio (which thus ceases to match your actual portfolio).

I confess I have not yet looked at rolling Japanese contracts on fixed dates (although there may well be enough precedent in roll habits to do this). But it does make me yearn for more control.
Last edited by AFJ Garner on Sat Dec 25, 2010 12:14 pm, edited 1 time in total.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Sat Dec 25, 2010 11:28 am

I have also been looking at the interesting effect of increased backwardation in some cases, where leaving a roll until later in the day (along US lines) leaves room for the futures price to converge more closely with the spot. Against this of course you have the decrease in liquidity to worry about.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Sat Dec 25, 2010 12:32 pm

Have a look at the drop off in volume which usually occurs if you roll on OI as opposed to Vol. It can be a lot worse than in this example.
Attachments
Vol.JPG
Vol.JPG (113.21 KiB) Viewed 8315 times
OI.JPG
OI.JPG (145.04 KiB) Viewed 8315 times

RedRock
Roundtable Knight
Roundtable Knight
Posts: 939
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

Post by RedRock » Sat Dec 25, 2010 2:58 pm

Have you played with rolling on volume and OI? Just asking, I haven't ventured into the orient as yet...

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Sat Dec 25, 2010 4:12 pm

Not sure that Volume "&" would help but at least it generates the data series forwards. Volume "or" would be ideal but it generates backwards.....

And generate backwards seems to ignore the "Confirmation Signals" - EG "Role on Second Consecutive Trigger".

Trouble is you only notice these occasional problems as they occur and when they do, they always take hours of work to investigate. Then further hours of head scratching to deal with.

JKE for instance (Kerosene) has just jumped back from June to January 2011 on a "roll on Volume" basis and the ripples extend back to May 2008. Back adjusted prices at the start date have gone from positive to negative.

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Sun Dec 26, 2010 2:10 pm

The gold plated Cadillac approach is the one designed and implemented by Roundtable member klatt_attack One of the advantages of this procedure is that, when complete, it gives you specific rollover instructions which can be used in standard, unmodified, CSI Unfair Advantage, with Generate Forward enabled.

However the effort required is, to put it delicately, nonzero.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Mon Dec 27, 2010 5:40 am

Thanks indeed for reminding me of Klatt Attack's approach which I have re-visited with interest. I have done the same sort of research myself in the past on a far less exalted scale - eyeballing the data for each and every contract rather than building the necessary algorithms within VBA (or whatever) to achieve Klatt Attack's vastly superior approach.

My problem with using fixed dates (which I did once upon a time) is that it seemed to me (on inspecting historical data) that trading patterns change over time. During one period (for instance) perhaps the November contract was popular and was heavily traded and then some years later people no longer used it. During one period (perhaps) a given fixed roll date worked well to keep you where the Vol or OI was and then the patterns changed.

The perceived advantage (to me at least) is that rolling on Vol or OI keeps me where I want to be and is self adjusting. The disadvantages are as indicated above with the JKE contract and as indicated in this thread:

viewtopic.php?t=7418&highlight=csi

The manual override in CSI can sometimes work nicely for you (it does with the JKE example) and sometimes it does not (as per the thread I quote above).

Of course none of this takes any account of points made frequently by ecritt and Roger Rines: different months of the same contract can trade very differently and there is profit to be had out of this. Some of the big CTAs I have been talking to recently DO profit out of this sometimes by trading the intra market spreads sometimes otherwise.

None of this takes account of the fact that how and into what you roll can have a very big effect on Contango/Backwardation.

However this takes me back to a point that I have made on several occasions: smaller operators have to be practical. Smaller operators KNOW they have made good money using bog standard CSI data and rolling mechanisms. They also know they may be able to make better money by devoting more time and research in the future to rolling mechanisms. They would doubtless benefit from building their own contract stitching software to include more satisfactory manual overrides.

So, as ever, there is no single answer. As ever there are different compromises to be made and advantages and disadvantages to each.

The longer you trade the more you realize that what you see in back testing is not necessarily what you get (or would have got) in real trading. So much depends on so many different real life problems. But again none of this should deter the beginner: despite whining and exasperations good money can still be made using a bog standard approach.

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Mon Dec 27, 2010 7:39 am

One advantage of building a fully automated rollover-procedure-chooser is that, once you've built it, you can run it and re-run it as often as you please.

If you observe (or think you observe, or overhear somebody else observe) that trading patterns have changed, possibly necessitating a change in your rollover procedure, you can simply turn on the machine and press the Go button. Then have a Boddington's, or a Cordon Rouge, or a Tanqueray and Schweppes, to relieve the exertion.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue Dec 28, 2010 6:08 am

Agreed!

RedRock
Roundtable Knight
Roundtable Knight
Posts: 939
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

Post by RedRock » Wed Dec 29, 2010 1:13 am

I suppose I will continue to be my own fully automated (single core) roll determination processor - to be awoken for adjustment as required. On demand.

longmemory
Roundtable Knight
Roundtable Knight
Posts: 128
Joined: Tue Apr 12, 2011 11:46 am

Post by longmemory » Mon May 16, 2011 4:40 pm

Hi Ecrit,

Looks like your question generated lots of enthusiasm, but few answers why Japanese contracts roll different. Allow me to offer a few observations and the answer will become clear.

Consider how currency carry trade works: the funding currency (low interest rate) borrows X units of cash and invests it in another currency where central bank short term rate (usually 30 to 90 day paper) has a higher yield. Say EuroYen effective yield is 0.25% (could be less in reality) and Brits pay 2.5%. (Making numbers up since Brits like a strong Sterling so pay more...) Simply said, you get 10 time the return with Stirling Yen carry then by liking your Japanese postage deposit book stamps. (The Japanese stamps actually pay ZERO.)

Carry risk is change in interest rate.
Way to offset risk is by becoming one of the boys.
AFC explanation hints at the right answer: small speculators wishing to collect premium. They cannot buy their way in the big boys club to offset risk, so they 'carry at home'.

Any commodity which is priced in Yens has zero exchange risk, but from the perspective of Japanese speculators (they are NOT system traders) they are there to collect the very inflated carry premium.

Japanese commodities roll EVERY month therefore, the mathematically about equal to currency carry trade with 30 day note, less exchange risk. If you are a preferred client and do NOT have to pay exchange fees, this can be a sweet deal.

Rather simple, once you see the world from their point of view.

How to roll:
Forget OI. Does not work.
Volume rolls are INDICATIVE of how you OUGHT to roll, but in real time, Volume rolls two days TOO LATE.

Go to exchange site, get the calendar.
Roll accordingly.

How to get in:
Place your LONG instructions the day before contract opens.
Be cautious about trading Short.

How to get out:
Liquidate two business day before new contract opens.

Slippage:
This is a long story, but to make it simple:
Study IFG contracts first & second day for Japanese commodities.
Manually adjust your buy price to the Highest(H, 2)
See if you made money during next 20 some days.
Do inverse for shorts.

Do not assume that published Open High Low Close and Volume have much to do with what preferred clients paid or received that day.

I've made my studies, tested them and traded Japanese commodities. Don't trade them any more.

L

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue May 17, 2011 4:40 am

longmemory wrote: Allow me to offer a few observations and the answer will become clear.
I confess that I am still somewhat mystified. Suziki San apparently sells short a few contracts of red azuki beans and goes long a few contracts of raw silk. He does not mind which way the price goes since in taking these positions he is in fact selling short the yen and investing in the higher yielding Australian dollar (or other higher yielding currency of his choice).

longmemory
Roundtable Knight
Roundtable Knight
Posts: 128
Joined: Tue Apr 12, 2011 11:46 am

Post by longmemory » Tue May 17, 2011 6:04 am

Hi AFJ

Your follow up question,

"I confess that I am still somewhat mystified. Suziki San apparently sells short a few contracts of red azuki beans and goes long a few contracts of raw silk. He does not mind which way the price goes since in taking these positions he is in fact selling short the yen and investing in the higher yielding Australian dollar (or other higher yielding currency of his choice)."

is not a question but an ANSWER.
Look at my brief outline of Carry Trade.
There must be some text out there which will go into details on Currency Carry, but a five sentence outline is good enough.

Suziki San is doing a both a Physical Carry and a Currency Carry. His returns ought to have very low correlation to price of Azuki Beans and high correlation to JPYAUD yield changes.

Suziki San is smart chap.

Do not assume that Suziki San just picks any old contract by picking numbers from a hat since the Physical Carry will pay off only when the contracts built in premium is distorted from real storage & interest rate charges.

I wish I were as smart as Suziki San.

L
p.s. The original question was about ROLLING.
The follow up question is about complex inter market swap & carry strategies so we are getting far off topic. :-(

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue May 17, 2011 7:39 am

I understand the (non-risk free) arbitrage possibilities inherent in currency carry and the potential from mis-pricing between cash and futures markets in commodities and financials.

I think perhaps I am merely a little confused as to what you are trying to say. I think maybe you are saying:

"My theory as to why the participants in the Japanese futures markets prefer long dated contracts is that the majority of them are expecting to profit from arbitrage opportunities rather than outright price moves and therefore like to lock in their trades for as long as possible so as to continue to collect the premiums they are expecting from such strategies. Open interest remains high in near term contracts long after volume has moved to further out contracts since many contracts are settled physically in pursuance of arbitrage trades".

Or am I still missing the point ?

longmemory
Roundtable Knight
Roundtable Knight
Posts: 128
Joined: Tue Apr 12, 2011 11:46 am

Post by longmemory » Tue May 17, 2011 10:14 am

Hi AFJ

Yes, you are missing the point.

Rolling: if this is what you are really after, roll on VOLUME 1st trigger and check published schedule.

Carry trade and Spreads: The topic is way to complex to cover in few lines.
If you'd really like to know about these topics, locate a book specializing in Agricultural Swaps. Find another book which covers currency trading from an econometric perspective. Such books are dull reading and leave out important details (on purpose, i suspect). Case and point: Sarno & Taylor: The Economics of Exchange Rates; Cambridge University Press, 2002.
Sarno & Taylor cover the math assuming you already know what to do.

Books leave out structural issues. Having made a long study of the issues and after developing and back testing strategies to exploit the opportunities, I found that structural impediments became a show stopper.

Examples:

Can I deliver a Short FUTURE MARKET Position against a Long CASH MARKET position?

What is my cost?
Depending on registration status, taking delivery or total cost of round turn can be as low as two dollars per contract or as high as several hundred dollars per contract. The cost depends on whom you are!

How much margin must be posted?
Registered market makers & Commercials may post less then 1% margin, others may have to come up with 10 or more of face value.

Computing Legs:
Are you eligible for Exchange Spread exposure margin, Broker Margin or the lesser of the two?

Do you have to pay the warehousing cost, or can you submit a letter of understanding that you have warehousing at your disposal?

And the real list is a lot longer...

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue May 17, 2011 10:34 am

longmemory wrote:Hi AFJ

Yes, you are missing the point.

Rolling: if this is what you are really after, roll on VOLUME 1st trigger and check published schedule.
Thanks. I did my investigation long ago on roll mechanisms in the Japanese futures markets and roll on volume with an early warning system on expiry and delivery just in case. All of this spews out on my daily order run.

What I was interested in was your theory as to WHY the Japanese futures markets roll on a different basis to US and other world futures markets. After all, let us face it - the US and European traders engage in much the same trades on their own markets as the Japanese do. I had thought your reasoning was as I stated above. It seems that I was wrong.

But in any event as a simple trend follower I am content to follow price in the Japanese and every other market and feel no overwhelming urge to change my spots and engage in arbitrage. I will leave that to Suzuki and his chums....and other masters of that particular universe!

ecritt
Roundtable Knight
Roundtable Knight
Posts: 134
Joined: Sat Aug 28, 2004 3:44 am
Location: Phoenix, AZ
Contact:

Post by ecritt » Tue May 17, 2011 2:49 pm

hi longmemory,

If I'm understanding you correctly, you are suggesting that Japanese traders are somehow participating in the interest rate differential between the Yen and some other currency by holding positions in Japanese futures contracts, that are priced in Yen? And it has something to do with the term structure of these futures contracts not being priced at "fair value"?

Thank you,

Eric Crittenden

ecritt
Roundtable Knight
Roundtable Knight
Posts: 134
Joined: Sat Aug 28, 2004 3:44 am
Location: Phoenix, AZ
Contact:

Post by ecritt » Tue May 17, 2011 2:51 pm

It looks like the exchange forces people to overweight their participation in the deferred contracts via position limits. Any idea why?

http://www.tge.or.jp/english/contract/c ... zuki.shtml

Post Reply