Intermarket analysis gets a bad rap because if you pick the correct predictive markets it works well most of the time except for these times it decouples. Even with these problems the winning percentages are very high. Let me give you an example I used in my April 1998 article. I want to make a note because I am sure someone will that I am not showing a portfiolo here. I wanted to show what I did in my old article and since TradeStation could not do portfiolo analysis back then or even today, I did not do that much of it until I designed TradersStudio. We used Silver to predict Thirty Year bond. We used my classic intermarket system with the parameters 14 for bonds and 26 for silver, shown in earlier post. This system did well since 1998 except in 2003. During this time both Silver and bonds rallied, they decoupled. I have attached a year by year table ,which is referenced here. You might say that the average of about 42700 per year since 1998, data used in article was though Dec 30, 1997 is not impressive and you would be correct. The problem is this decoupling. Another point is during the first five years out of sample it average over 45200.00 a year, which is not bad. [ See 2003 problems in spreadsheet , plus chart attached] These three trades took up over eight months and produced disaster, these trades could have been filtered out. If you look at the attached screen shot you see two trades ,which failed when the market was trending. Intermarket analysis , the way I defined it is a counter trend methodology, it does bad when markets trend. We could have used correlation analysis or simple a trend following methodolgy to turn intermarket off when bonds trend. The winning percentages are huge, almost 70% overall. The fact I am using a almost 8 year old set of parameters and that after a bad year with bonds are back in a trading range this relationship is producing historical normal returns, over 9K the first 8 months and is did make over 4K in 2004, you can see that intermarket analysis has promise, if you understand it.
Thanks for that Murray. I'm curious - why Silver vs Long Term rates and not Gold? Would seem to be a more logical choice?
Gold has had other issues. The reason is that for the period from the mid 1990's until 2002 the central banks , sold there stock piles to support the dollar and give the picture that inflation was nonexistent. Silver is a industrial metal and has not been manipulated like gold. I wrote about this in one of my futures articles.
Let's first spend a little time discussing inflation and interest rates. You might ask why the negatively correlated market system works. You can think of it as a form of arbitrage. My model for interest rates is as follows: T=A+B+(C*A) where A is current inflation B is perceived future change in inflation C is risk factor. This model states that interest rate for time T are equal to current inflation plus perceived future change in inflation for time "T". The final term is a risk factor times current inflation. During periods of steady growth and low inflation the ratio between long bond rates varies between 220% and 250%. This would indicate an inflation rate of 2.4% would lead to a 30-year bond rate of 5.75% to 6.25%. You might be saying that rates are not this high currently and ask why. This leads to several questions? 1) Are things different this time; we heard that before the Internet bubble. 2) Is the yield curve predicting another recession? The fed keep raising rates with $3.00 gas to protect the dollar from being devalued due to budget deficits. Will the fed overdo it and create another recession? 3) Do we have a perceived lower cost of goods as US companies need to compete with foreign countries for labor. I think numbers 2 or 3 are possibilities, because I donât believe it is different this time.
I have showed an example of using Silver to trade bonds a few replies ago. I am now going to share with you the fact that many stock indexes, ETF's and even some mutual funds are useful in predicting the prices of the underlying related futures. Stock prices lead commodity prices in many cases. One reason I think is that much more money is spent in the stock analyst world, so there research contains future prediction of the underlying prices of the related commodity or futures. Here are some classic examples Gold stocks and gold futures US Chemical stocks and bond prices, these are negatively correlated because the price of chemicals increasing which would be good for these stocks indicates future inflation. There are many others types of stocks, which are negatively correlated to bond prices. Steel stocks, Metal stocks, and Food producer stocks. Intermarket analysis is just not for predicting bonds: Gold stocks and gold futures Oil producers and Crude oil Copper company stocks, as an index and Copper futures. You get the idea. We will cover this more in a few days.
When markets move it's not just because of new money. When dominant market players are allocating, changing allocations and adapting, this is the time to pay attention. Especially, in these narrower range days... I keep exhaustive spreadsheets on "logical" and "not logical" relationships. Their purpose is to discover "movement" and thus track/discover the inverse or correlation to come. One day I accidentally discovered this and made some money, WHILE IT WORKED. Since then it is an ongoing study to find what combination of instruments is working. I only know of one trader that specializes with this with OPM...steve something or other...What he does is very secretive and I do not believe he posts here on ET anymore, and when I stumbled onto this he pm'd me one or two years ago. Frankly, if you could enter an instrument with a very high probability of it going in your direction and it was just to learn "how much", that would not be something shared freely. What I like about Mr. Ruggiero is that he stimulates thought with his pieces. The relationships discovered may indicate movement, which predicts the movement to come in its correlating/non correlating instrument, whether its with the trend or not. There is an uncanny relationship and I believe that this thread is discussing the only TA to be able to predict or forecast. Otherwise TA is merely a tool to quickly visualize price/time/volume without any predictive value whatsoever in my trading and in my opinion. good thread.... Michael B.
The price of Crude oil is a hot topic right now. Another issue is the budget deficit and the dollar. I am going to show you how both of these interact. Let's use the Intermarket system that we have used for bonds, my classic negatively correlated systems. We will take a look at trading crude oil using the dollar index. The reason that this is a valid relationship is that crude oil is priced in dollars so, if the dollar declines, oil prices need to rise just to stay constant. The dollar index was around 125.00 in 2001 while oil was stable around 25.00. The dollar index has declined over 30% while oil has increased about 250%. Let's talk a look at a relationship, which I originally discussed on CNBC in 1994. We will trade crude oil using the dollar index as our Intermarket as a negatively correlated market. We will use the same system we discussed earlier in this thread. This relationship goes though period of decoupling because geopolitical issues as well as disruptions which can overpower the effect of the dollar to crude like the Hurricane. Using a 5-day moving average for crude and a 40-day moving average for the dollar index produced the following results, over the years. See attached spreadsheets. We can see in this sheet that in 1999 and in 2000 this relationship totally disconnected. This is not a perfect relationship and in fact I would not use as a stand-alone system. My point is to show that if we don't worry about the value of the dollar we can create other problem, which we might not be obvious if we don't understand how markets interrelate.
I am going to do some work in this thread using some of the SP500 industry groups. I would like people to be able to follow along. I use both CSI and Pinnacle for my futures data. The problem is this SP500 industry group data is hard to find. Can the members help me find a free source for this type of data, so we can all follow along using the same source. We need as much history as possible, end of day, at least 10 years. If we find industry groups which are not the SP500 groups that also ok, we can try them. If we can find free sources with links and symbols, then we can share working together on this project. I have put one of my intermarket systems up in TraderStudio already , I am sure TradeStation user can port it into TradeStation.