Transtrend: trading pork bellies

[fib21]

Up untill this year I had been trading pork bellies but because of the recent decline in volume I have decided to exclude it from my portfolio at least for now.

But looking at Trandtrend's latesdt disclosure document (updated 2010) it seems that they are trading this market, so I'm wondering if any of you have any idea how a large CTA managing billions could actually be trading this market?

[svquant]

Don't believe everything you read. I've noticed pork bellies on the list of tradeables for quite a few large CTAs and it does not compute given their size and the market liquidity. Most of these CTAs have been around for a while and I just think it is left over from when PB was more of a market one could trade.

Remember all d-docs I've seen also carry wording to the effect that the manager may change this list at anytime without notification and it is a representative list etc.

[absret111]

Does anyone know the reason behind the decline in volume in the pork bellies contrct?

[M20J]

A recent copy of the CME Daily Livestock Report gave a good explanation:

http://www.dailylivestockreport.com/doc ... 0-2010.pdf

[AFJ Garner]

Talking of liquidity, for what it is worth I had a conference call with a systematic CTA this morning who run $1bn plus. They say they have positions in 200 to 400 instruments at any one time and trade..........amongst other illiquids..............milk.

[absret111]

That's interesting
But how about the performance of this CTA? If he's diversified so broad, the equity should be realetiv smooth, if it's worth for him to trade such illiquid markets...

[AFJ Garner]

Over almost a decade they have recorded the following performance:

Annualised RoR 19.6%
Annualised St Dev: 22.56%
Worst DD -24.66%

10 year correlation to the usual suspects is around 65%

[sluggo]

How to trade a very large portfolio, which includes illiquid markets, in a huge account?

Try the following little homework assignment. First select a mechanical trading system + parameters which give a looooong average trade hold time on the very large portfolio. Something like "C" or "X" or "Z" on the attached diagrams (which I snagged from other posts on this website).

Then modify the trading system's exit code. Have the system exit 20% of its position on the day it gets the exit signal. Exit another 20% of the position the next day, exit another 20% of the position the day after that, then 20% more the following day, and exit the final 20% a day later.

For those who prefer fractions rather than percentages, exit (1/5th) on the first day, (1/5) the second day, (1/5) the third day, (1/5) the fourth day, and (1/5) the fifth day.

Set your simulated entry/exit slippage to "high", lots of slippage, lots more than you think you'll actually experience in real trading.

Now simulate the modified system on the very large portfolio. How do you like its performance? Does it make you think it might be possible to trade illiquid markets with a big account, if you're willing to adapt the way you trade?

Part II: modify the system again. Exit (1/10th) of the position at market-on-open of the first day. Exit (1/10th) of the position at market-on-close of the first day. Exit (1/10th) of the position at market-on-open of the second day. You get the idea. Simulate. Analyze results. Formulate conclusions.

Part III: modify the system again, scale-in entries to smear them across a number of days, just as you have previously smeared exits across a number of days. Simulate, analyze, etc.

Part IV: Partition the tradeables of your portfolio into two groups, the liquids and the illiquids. Modify your system's position sizing code so that when you get an entry signal in a market belonging to the "liquids" group, you take that trade at 100% of full size. But when you get an entry signal for a market in the "illiquids" group, you take that trade at X% of full size. Simulate for several values of X% -- be sure to include X=100%, X=50%, and X=5% in your trials. Analyze results. Formulate conclusions.

I am confident that an energetic and innovative trader could think up a dozen more schemes and strategems and contrivances, to code and to test.

[fib21]

Another thing that I'm sure many of the larger firms do is assign a trader to a market or a sector and when there's a signal they spend the entire trading day (or days) buying or selling as many contracts as needed and this could at times equal 100% or more of the average daily liquidity for that market.

For MILK if I have a four hour window to fill my trade and I need to buy 200 contracts I may be buying 10-15 contracts every 15 minutes this way I'm not making a splash as I would be if I placed a trade for 200 all at once.

So to me this means that these large firms have to spend a lot of time, thought and effort in the execution part of this business. A good strategy here for executing trades is essential and possibly unique to each market/sector.

Another possibility is that in the end they may not be profiting much if at all from these illiquid markets but just like short trades still gives the system a boost by improving risk adjusted returns etc.

That’s my 2cents worth

[AFJ Garner]

Managers I have spoken to tend to restrict shorter term trading to liquid contracts so that slippage does not destroy the potential profit. Illiquid contracts like milk would be reserved for longer term positions.

Big operators tend to trade all contract months not just the front month. This can help to extend their reach and enable them to take on greater AUM without reaching capacity problems.

Typically they would be looking at restricting themselves to a small percentage of total open interest rather than just (by way of example) a small percentage of 5 days' average volume on just the front month.

A lot of research and development goes into automating the dealing process and making sure that orders are dealt when the liquidity is there and so forth and so on.

[absret111]

I suspect that rollovers are handled in the same way as entries and exits.
Or any other idea?

[AFJ Garner]

A snapshot of one big fund manager's positions on 30th September this year shows:

Long 208 Milk Class III CME 201012
Long 322 Milk Class III CME 201011

Long 100 Oats CBOT 201012

Short 204 Rough Rice CBOT 201011

and a few other positions in illiquid markets. All grist to the mill it seems. Interesting to compare these positions with total open interest at the time and the average daily volumes in the relevant contracts. Scaling in and scaling out no doubts helps.

[fib21]

Just got this from Transtrend in response to my email asking about PB...

Dear Dario,

I tried calling you earlier without success.

The disclosure document contains a list of markets that Transtrend may trade. Pork bellies are currently not traded.

I hope this information is useful and sufficient. Please do not hesitate to contact us should you require any further information.

Kind regards,
Joris Tolenaar

+31 10 453 6510

[AFJ Garner]

Ask for a copy of the Transtrend Fund Alliance Annual Report 2009. It gives a list of positions held as at the year end. Interesting stuff but of course it is only a snapshot in time.

[AFJ Garner]

I recall Sluggo saying that he has gained much over the years from talking to other CTAs at trade fairs and the like and that while one should not expect too much disclosure in terms of the systems traded, they are often willing to discuss other areas of trading openly.

In an interesting conference call today with a well know US CTA, I learnt that they too trade milk. I think it brings into focus the incredulity shown on this forum from time to time as to a large fund manager trading illiquid contracts. They can and they do. Trading milk may not add much to a $1bn portfolio but every little bit of diversification helps.

[svquant]

Came across this quote from a Transtrend Q2 2010 performance update letter, "Another constantly changing factor that requires adaptation is liquidity. As markets grow and decline and liquidity shifts from one exchange to another, we aim to follow liquidity and add new markets where possible; while dropping others. For instance, DTP recently stopped trading Azuki red beans due to declining liquidity and kerosene and gas traded on the Central Japan Commodity Exchange due to the fact that they will be delisted. This is an endless process and our search for new markets will continue."

[Eventhorizon]

fib21:

Another thing that I'm sure many of the larger firms do is assign a trader to a market or a sector and when there's a signal they spend the entire trading day (or days) buying or selling as many contracts as needed and this could at times equal 100% or more of the average daily liquidity for that market.

This article reflects the ideas expressed by fib21 above:

http://www.automatedtrader.net/articles ... ct-capital

It talks about how Aspect Capital gives its traders the choice regarding whether to handle a trade themselves or allow an algorithm to handle execution.

As you might expect, this leads to the traders taking a hands-on approach in the less liquid markets while computers handle the liquid ones.

[cryder]

Thanks for posting the article. Found it interesting, in these days of pure electronic trading and direct order routing, the company is interested in 'minimizing or hiding' the trading footprint to prevent predation on their order flows was a consideration. Guess there must be some level of spying, electronic monitoring going on.

[Moto moto]

good article.
I also enjoyed the talk about hiding the footprint, as the other day on talking to someone at a social function, they were telling me a lot of the retail cfd providers look for clients that show traits of loosing all their money, and then they dont trade against them as such, but they dont hedge up/offset their trades in house. they just take them on.
This was working on the simple basis that most retail traders loose their money in the markets.
slightly different to hiding the footprint, but none the less interesting, and confirmation brokers do track what individual clients are doing in terms of style.