Just my humble opinion, correlation isn't quite the angle you're looking for. The ones you mention above are input/output and their values will float based upon price of feed, price of alternates, and season. If you're looking for purely correlation arbitrage, look at inter-month spreads of a single commodity. You can trade from a percent of carry basis, for example. The moves can be extraordinary (look at Dec wheat versus July wheat, for example, but that was an inter-crop spread). Eurodollars move similarly to treasuries, but just looking at the TED spread for the past few months will show that they're not really highly correlated lately. With Eurodollars, take a look at inter-month (or even inter-year) spreads. You're playing the yield curve then. Platinum/palladium is an interesting spread as they are largely substitutes industrially. The spread has gotten wild lately as rumors of a platinum ETF have swirled about. I hope those are good pointers to start with.
Wheat---Chicago versus Kansas City. They're both "Winter wheat". They correlate better than the "Spring wheat" Minneapolis contract. They're a little trickier to execute since they trade on separate exchanges.
The Commodity Systems Inc correlations page (pay per view) gave some unusual pairings: (Goldman Sachs Commodity Index) vs (Crude Oil) ....... Correl= +0.980 (Euro German Bund) vs (British Sterling Rate 3Mo) ....... Correl= +0.958 (Tokyo Rubber) vs (COMEX Gold) ....... Correl= +0.950 (Crude Oil) vs (Platinum) ....... Correl= +0.949 (London FTSE 100 Stockindex) vs (Eurex Dax German Stockindex) ....... Correl= +0.924 (Canola) vs (Soybean Oil) ....... Correl= +0.847 Some people just love to say "correlation is NOT causation" and perhaps the very high correlation between Tokyo Rubber and NYC Gold might be evidence for their view.
Correlation over too long a time period becomes useless i.e. significant moves are averaged out. Correlation over too short a time period is inaccurate. What time periods do you guys use to timebucket your correlations? Gas and power prices are significantly correlated with the spread being referred to as the spark spread. Coal and power prices are also significantly correlated with the spread being referred to as the dark spread. European emissions are also closely correlated to european power prices. Although coal, power and emissions are hardly liquid futures markets.
Certainly a correlation between crude oil and heating oil/gasoline which is the refining margin or crack spread is traded heavily. As stated in this thread, there is absolutely no correlation between crude and nat. gas. Nat. gas is mainly a domestic product (meaning almost no reliance on imports), and is closely correlated to heating oil. In fact the nat. gas to heating oil trade can be extremely lucrative for fundamental trading. When the price of nat. gas is above the price of heating oil, traders go long heating oil and short nat. gas. The fundamentals of the market will not allow this relationship to maintain these economics because heating oil will be substituted for the use of nat. gas. This trade takes DEEP pockets to maintain over time because you have to be willing to take the pain if the spread goes wider, but will ultimately prove to yield tremendous profits.
I've not heard of a correlation between nat gas and heating oil. They really aren't substitutable, their production and storage methods are different. Your logic makes this relationship somewhat suspicious. If Crude -> Heating Oil -AND- Heating Oil -> Natural Gas then why isn't Crude -> Natural Gas ... In any case, doing a correlation of prices between crude, heating oil, and natural gas reveals (going back to beginning of '01): Crude/Heating Oil: .98 Crude/Nat Gas: -.34 Heating Oil/Nat Gas: -.29 So, crude/nat gas is slightly more correlated than heating oil/nat gas, but crude/heating oil are extremely correlated. Do you have a reference which would show the correlation between HO and NG? If I'm wrong, I'd love to find out for sure.
My apologies, utilizing the term "absolutely no correlation" was from a fundamentalist point of view not in the definition of mathmatics. The point that I was trying to make was that many people trade the crude vs. nat. gas relationship when the only relationship between the two is that when you E&P natural resources you will extract both of these products. If you were to trade the relationship it is very dangerous. Putting on a crude oil vs. heating oil spread is synthetically creating a refining margin. Trading the nat. gas vs. heating oil spread makes sense from a fundamental point of view because when nat. gas is not available heating oil is substituted. Trading crude oil to natural gas doesn't make any sense because they aren't related to one another in the energy market unless you are a producer.
Don't forget the interest rate spreads. There is the NOB (Notes over Bonds), MOB (Munis over Bonds) and the TED spread.
bump thought i would bump up this thread to see which futures pairs are correlated in todays markets !?!!
Correlations are easy to find: http://www.mrci.com/special/correl.htm What is your purpose in asking? Do you feel there is some trading edge in correlations?