Appendix D: Continuous Futures Contracts
posted under
by ceecabolos
With a clean database of "raw" commodity data, there are numerous types of contracts that can be gleaned from the raw data, such as: Nearest Contracts, Next Contracts, Gann Contracts, and Continuous Contracts. Following, are ideas for constructing these futures contracts derivatives. The symbols used are for illustration purposes only. These continuous contracts can be created through the Dial Data Service (56 Pine Street, New York, NY 10005, [212] 422-1600.)
$$ NEAREST CONTRACT $$
A nearest contract is primarily used by traders who just want a large file of continuous data made up of actual trading prices. They are content with the data going to expiration and then rolling over automatically.
It is quite probable that no one trades the nearest contract within 15 to 30 days of expiration. This is because the liquidity dries up very fast in the latter days of a contract. The number of days before expiration that an individual rolls over to the next contract is a function of the commodity that is being traded (the number of months till the next contract), and the individual's trading style. It is quite conceivable that the same individual will rollover at different times for different commodities.
When to rollover to the next contract will more than likely be based upon the current contract's volume. When it begins to erode, that is the time to roll forward.
Therefore, one should have available a choice as to when to rollover his Nearest Contract. Remember, Nearest Contracts are made up of actual data. Here are some examples: Portfolio Manager A is content to rollover at expiration; so all he wants is the "standard" Nearest Contract with symbol TRNE00 (Treasury Bonds). Manager A is probably managing money and needs equity calculations which he can derive from the data. Trader B feels that trading in the month of expiration is not liquid enough for him; so he wants his Nearest Contract to roll over 15 days prior to expiration—the symbol could be TRNE15. Analyst C would like to evaluate different roll-over dates, so he might like to download multiple Nearest Contracts, such as: TRNE00, TRNE05, TRNE12, and TRNE21 (which roll-over 5, 12, and 21 days before expiration.
Keep in mind that all of these contracts are Nearest Contracts and contain actual contract data. The only difference is which actual contract the data comes from.
$ NEXT CONTRACT $
A Next Contract is a unique offspring of the Nearest Contract. It is exactly the same as the Nearest Contract except that it is always
the contract that follows the Nearest Contract. In other words, if the Nearest Contract is using December data for T-Bonds (TR), then the Next Contract is using data from the March T-Bond contract. When the December contract expires, the Nearest rolls to the March and the Next rolls to the June contract. This is defined as the Next-1 contract.
From this concept, another Next Contract is available, called a Next-2. Here, the data is always coming from the contract that is two contracts away from the Nearest Contract. Keeping with the above example, if the Nearest is using data from the December contract, the Next-2 Contract is using data from the June contract. When the December contract expires, the Nearest begins to use data from the March contract and the Next-2 Contract uses data from the September contract and so on.
Ticker symbols for the Next contracts are: TRNXT1 and TRNXT2. Of course, the actual futures ticker will be used instead of the TR used in this example.
$ GANN CONTRACT $
Gann Contracts refer to the use of a specific contract month and rolling over only to the same contract in the next year. For example, July Wheat would be used until the July contract expires, then the Gann Contract would start using data from the July Wheat contract of the next year.
Examples of ticker symbols for Gann Contracts are: WO7GN, GC04GN, JY12GN, etc. (representing July Wheat, April gold, December Japanese yen).
CONTINUOUS CONTRACTS
Continuous Contracts were developed to help analysts overcome the problem of liquidity dry up and premium (or discount) gaps in futures data. This becomes a problem whenever an analyst is testing a trading model or system over many years of data. It allows for a continuous stream of data with compensation being made for rollover jumps in price trends.
CONSTANT FORWARD CONTINUOUS CONTRACTSA Constant Forward Continuous Contract looks a constant length of time into the future. It uses more than one contract to do this. A common method is to use the nearest two contracts and do a linear extrapolation of the data
One possibility is to give the futures trader (as with the Nearest Contracts) the ability to construct his own Constant Forward Continuous Contract. Three things are needed to do this: The commodity symbol, the number of contracts he wants used in the calculation, and the number of weeks into the futures he wants to look. For instance, if he wanted T-Bonds, using 3 of the
nearest contracts, and looking 14 weeks into the future, the symbol could be: TRCF314. TR is the symbol, CF is for Continuous (Forward Looking), 3 is the number of contracts used, and 14 is the number of weeks the price is projected.
The mechanics of this are fairly simple. First, a fixed rollover date would need to be set for each commodity. A good one to start with could be 10 days prior to expiration. What is important is that there is a rollover sometime prior to actual expiration. Second, the number of contracts used will never be less than 2 and probably never greater than 4. The number of weeks used should probably always be greater that 3 and could go up to 40 in some cases.
Example: This is the method used by Commodity Systems, Inc. (See Perpetual Contract in Chapter 8.)
T-Bonds will be used again, because they have a uniform expiration cycle of every 3 months. Let's say a trader wants a Continuous Contract of T-Bonds using the 2 nearest months and looking 12 weeks into the future (symbol = TRCF212). Today's date is December 1. A graphical portrayal makes this easier to understand (see Figure D.1). The vertical axis is price and the horizontal axis is time. Today's date is marked on the horizontal axis and the expiration dates of the two nearest contracts (December and March), are also marked. He wants to look 12 weeks into the future so a mark is made 12 weeks from today which is about February 25. The close price of the December contract was 88.25 and the close of the March contract was 87.75 These points are then put above their expiration dates at the corresponding prices. Then a linear extrapolation is made by merely drawing a line between the two points. The slope of this line will vary up and down depending upon the outlook for long term interest rates (in this T-Bond example). In this particular example the outlook is for higher rates because the March futures price is lower than the December price.
To find the value of the TRCF212 close price for today, find the point on the horizontal axis that is 12 weeks from today (Feb 25th) and go up to the line drawn on the chart. Then from the line go to the right and that is the price of the close for this Constant Forward Continuous Contract (about 87.91). You can 510
$ Appendix D $
also visually see from the chart that the March contract is carrying more weight than the December contract because the point of interception is closer to March. This method can be done on the Open, High, Low, and Close in the exact manner. Of course, a computer does it mathematically; this is just a visual explanation of how a Perpetual Contract is constructed.
$$ NEAREST CONTRACT $$
A nearest contract is primarily used by traders who just want a large file of continuous data made up of actual trading prices. They are content with the data going to expiration and then rolling over automatically.
It is quite probable that no one trades the nearest contract within 15 to 30 days of expiration. This is because the liquidity dries up very fast in the latter days of a contract. The number of days before expiration that an individual rolls over to the next contract is a function of the commodity that is being traded (the number of months till the next contract), and the individual's trading style. It is quite conceivable that the same individual will rollover at different times for different commodities.
When to rollover to the next contract will more than likely be based upon the current contract's volume. When it begins to erode, that is the time to roll forward.
Therefore, one should have available a choice as to when to rollover his Nearest Contract. Remember, Nearest Contracts are made up of actual data. Here are some examples: Portfolio Manager A is content to rollover at expiration; so all he wants is the "standard" Nearest Contract with symbol TRNE00 (Treasury Bonds). Manager A is probably managing money and needs equity calculations which he can derive from the data. Trader B feels that trading in the month of expiration is not liquid enough for him; so he wants his Nearest Contract to roll over 15 days prior to expiration—the symbol could be TRNE15. Analyst C would like to evaluate different roll-over dates, so he might like to download multiple Nearest Contracts, such as: TRNE00, TRNE05, TRNE12, and TRNE21 (which roll-over 5, 12, and 21 days before expiration.
Keep in mind that all of these contracts are Nearest Contracts and contain actual contract data. The only difference is which actual contract the data comes from.
$ NEXT CONTRACT $
A Next Contract is a unique offspring of the Nearest Contract. It is exactly the same as the Nearest Contract except that it is always
the contract that follows the Nearest Contract. In other words, if the Nearest Contract is using December data for T-Bonds (TR), then the Next Contract is using data from the March T-Bond contract. When the December contract expires, the Nearest rolls to the March and the Next rolls to the June contract. This is defined as the Next-1 contract.
From this concept, another Next Contract is available, called a Next-2. Here, the data is always coming from the contract that is two contracts away from the Nearest Contract. Keeping with the above example, if the Nearest is using data from the December contract, the Next-2 Contract is using data from the June contract. When the December contract expires, the Nearest begins to use data from the March contract and the Next-2 Contract uses data from the September contract and so on.
Ticker symbols for the Next contracts are: TRNXT1 and TRNXT2. Of course, the actual futures ticker will be used instead of the TR used in this example.
$ GANN CONTRACT $
Gann Contracts refer to the use of a specific contract month and rolling over only to the same contract in the next year. For example, July Wheat would be used until the July contract expires, then the Gann Contract would start using data from the July Wheat contract of the next year.
Examples of ticker symbols for Gann Contracts are: WO7GN, GC04GN, JY12GN, etc. (representing July Wheat, April gold, December Japanese yen).
CONTINUOUS CONTRACTS
Continuous Contracts were developed to help analysts overcome the problem of liquidity dry up and premium (or discount) gaps in futures data. This becomes a problem whenever an analyst is testing a trading model or system over many years of data. It allows for a continuous stream of data with compensation being made for rollover jumps in price trends.
CONSTANT FORWARD CONTINUOUS CONTRACTSA Constant Forward Continuous Contract looks a constant length of time into the future. It uses more than one contract to do this. A common method is to use the nearest two contracts and do a linear extrapolation of the data
One possibility is to give the futures trader (as with the Nearest Contracts) the ability to construct his own Constant Forward Continuous Contract. Three things are needed to do this: The commodity symbol, the number of contracts he wants used in the calculation, and the number of weeks into the futures he wants to look. For instance, if he wanted T-Bonds, using 3 of the
nearest contracts, and looking 14 weeks into the future, the symbol could be: TRCF314. TR is the symbol, CF is for Continuous (Forward Looking), 3 is the number of contracts used, and 14 is the number of weeks the price is projected.
The mechanics of this are fairly simple. First, a fixed rollover date would need to be set for each commodity. A good one to start with could be 10 days prior to expiration. What is important is that there is a rollover sometime prior to actual expiration. Second, the number of contracts used will never be less than 2 and probably never greater than 4. The number of weeks used should probably always be greater that 3 and could go up to 40 in some cases.
Example: This is the method used by Commodity Systems, Inc. (See Perpetual Contract in Chapter 8.)
T-Bonds will be used again, because they have a uniform expiration cycle of every 3 months. Let's say a trader wants a Continuous Contract of T-Bonds using the 2 nearest months and looking 12 weeks into the future (symbol = TRCF212). Today's date is December 1. A graphical portrayal makes this easier to understand (see Figure D.1). The vertical axis is price and the horizontal axis is time. Today's date is marked on the horizontal axis and the expiration dates of the two nearest contracts (December and March), are also marked. He wants to look 12 weeks into the future so a mark is made 12 weeks from today which is about February 25. The close price of the December contract was 88.25 and the close of the March contract was 87.75 These points are then put above their expiration dates at the corresponding prices. Then a linear extrapolation is made by merely drawing a line between the two points. The slope of this line will vary up and down depending upon the outlook for long term interest rates (in this T-Bond example). In this particular example the outlook is for higher rates because the March futures price is lower than the December price.
To find the value of the TRCF212 close price for today, find the point on the horizontal axis that is 12 weeks from today (Feb 25th) and go up to the line drawn on the chart. Then from the line go to the right and that is the price of the close for this Constant Forward Continuous Contract (about 87.91). You can 510
$ Appendix D $
also visually see from the chart that the March contract is carrying more weight than the December contract because the point of interception is closer to March. This method can be done on the Open, High, Low, and Close in the exact manner. Of course, a computer does it mathematically; this is just a visual explanation of how a Perpetual Contract is constructed.
