During the last 20 years, many traders, from the smallest retail accounts to the largest hedge funds, have focused their trading interests on the development and execution of mechanical trading systems vs. wholly discretionary trading. This shift, combined with technological advances, has caused spread trading to fade from favor as an old-fashioned and cumbersome methodology.
However, most commercial firms still understand that spreading is an important part of a complete approach to trading. But today, a viable spread trading program requires significant resources in terms of research tools. Fortunately, modern software allows us to build powerful spread-based trading systems without the massive investment in a research department.
This series of articles will explain the mechanics of spread trading and show how backtesting can be used to develop mechanical trading strategies for spread markets.
WHAT ARE SPREADS?
A spread trade is the purchase of a futures position combined with the simultaneous sale of a related futures position; the industry routinely expresses spread prices as the premium or discount of the “buy side” of the spread. It is generally assumed that the positions, or legs, that make up the spread are economically related.
The markets that are traded in an exchange-recognized spread have a natural price relationship based on economic reality. This naturally leads to two questions that traders seek to answer when analyzing a spread:
• What normal relationships exist between the two markets in the spread?
• How do divergences in these relationships resolve themselves?
The answers to these two questions are important to spread trading and in trying to develop mechanical systems that can be used to analyze them.
One type of fundamental interrelationship in a spread is when two commodities have a substitution relationship where one commodity can be used to replace another in a nearly identical situation. Examples include short-term interest rate products or livestock feeds such as wheat or corn. The relationship between corn, soybeans and cotton is also a substitute relationship because these commodities compete for acreage at planting time.
Other relationship spreads such as the soybean crush spread and the crude oil crack spread are product based. These spreads emulate the internal workings of the oilseed or refining industries. This type of relationship also exists in livestock with spreads such as the hog/corn ratio and the cattle feeding spread.
Other spread relationships focus on the cost of carry or price differential between the same commodity with different delivery months or growing seasons. Spread traders should understand the economic relationship of a spread and recognize the conditions where a spread could fail. The most important aspect of any spread research is to develop a valid theory as to why a relationship exists. Without a valid theory to accompany a trading idea, a spread could be very dangerous to trade. In doing research, it is OK to reverse the process and test for patterns and then try to develop a theory to explain it.
WHY TRADE SPREADS?
In most situations, spread trading requires considerably less margin to maintain than outright positions. The spread trader’s main concern is the price differential between the long and short legs of the spread. Much less important are the markets’ actual price levels or the direction of the trend.
Because spread trades are considered hedged positions, they feature less volatility than outright futures positions. Exchange-recognized spreads typically have much lower margin requirements than outright positions even though many old-crop/new-crop spreads do not behave like hedged positions and can be very volatile. This is a critical point. Although spreads generally exhibit less outright volatility, they can be quite risky during times when fundamental relationships break down. That is one reason understanding the economic link is so important.
Spread trading also offers a new avenue for profits to traders. The spread between two markets generally trends more often than outright futures and often trends strongly when outright futures markets are flat. Because of their trending characteristic, different filters can be used to analyze spreading opportunities. The list of techniques includes seasonality and carrying charge studies plus many of the most popular technical indicators and chart patterns.
This means that classic technical trading methods that have worked on outright futures positions during strongly trending years can be employed. The analysis methods that worked during the 1970s and 1980s will work on today’s spreads because these trends have not been distorted by system-based trading, which gives the next generation of spread traders a big edge.
“One market, two ways to profit” (below) displays the July 2006 corn chart in the main panel and the July corn/July wheat spread as an indicator in the secondary panel. July is the first new-crop month for winter wheat as harvest begins in late May in the southern regions of the United States and moves north throughout the summer months. July usually finds wheat harvested in Kansas, Colorado and southern Nebraska.
By observing the chart, you can see that corn prices and the spread seem to move more or less independently of each other. You also can see that the spread is a lot less noisy and it holds a trend that recalls the commodity markets of 20 to 30 years ago.
SPREAD TYPES
If you determine that spreading may be for you, there are several types of spread positions to consider.
Seasonal spread tendencies in the agricultural markets do not change much when analyzed through a long period because the tendency is tied to the production/consumption cycle of the physical. As mentioned, it is important to recognize the market relationships and understand why a spread may or may not work in a given year.
More spreading opportunities are recognized by traders today because increases in computing power and market data provide the ability to perform more detailed testing to locate seasonal spreading opportunities. However, the seasonal spreads tend to be shorter and are much less obvious on the charts when performing analog comparisons.
Intramarket spreads. These involve the same commodity on the same exchange but using different delivery months. For agricultural crop contracts, there are two distinctions. Intra-seasonal uses two contracts that trade within the same crop year. Carrying charges are important in this type of spread trade. Contango is the normal alignment of carrying charges, and backwardation describes a reversed or abnormal carrying charge relationship. Interseasonal involves two contracts that are in two different crop years. Even though the contracts are the same commodity on the same exchange, the price dynamics between the contracts are quite different due to differences in crop size and demand between the crops.
It’s important to note that because of differences in growing seasons in various areas where a commodity is produced, the harvest may overlap several months within the crop year. An example would be the wheat market where harvest begins in south Texas in mid-May and moves north to finish in southern Canada in late August.
Intermarket spreads. These are the similar commodities for the same delivery month on different exchanges. An example of this type of spread relationship would be wheat contracts traded at the Chicago Board of Trade, Kansas City Board of Trade and/or the Minneapolis Grain Exchange. The commodity is wheat at all exchanges, but the type of wheat, its use and delivery points are different at each exchange. This causes each market to have distinctly different fundamentals.
Intercommodity spreads. These are the same delivery month for different but related commodities and may or may not be traded on the same exchange. Examples of this type of spread relationship would be silver/gold, corn/wheat, live cattle/lean hogs. This type of spread also covers currency cross-rates and arbitrage techniques. In interest rates, spreads of different bond maturities have little to do with the overall direction of interest rates, but are a view of the yield curve.
Intercommodity spreads work for logical seasonal or economic reasons. Corn and soybeans maintain a spread relationship largely because of competition for acreage. Because of yield differences, the price of soybeans is typically 2-1/2 to 3 times the price of corn. This keeps the return per acre about the same. If the price of one or the other gets out of line, farmers shift acreage to the other market.
Source product spreads. These work as the result of the economic relationship(s) between a commodity and one or more of its related products. There are several examples of this type of spread relationship, including the soybean crush, which involves soybeans/oil and meal, and the feedlot spread, which involves live cattle/feeder cattle and corn.
This type of spreading actually emulates the actual business of processing a commodity such as crude oil. The spread compares the equity value of the source market with the equity values of its major end products, such as soybean meal and oil. This type of spread trading is more involved because there are at least three contracts in the position and the contract sizes are not as precise because of processing waste. Several contracts on each side of the trade may be required for a balanced position.
SPREADS & FUNDAMENTALS
Anyone associated with agriculture knows that prices move up and down with some degree of regularity. This movement occurs for sound economic reasons such as increased supply at harvest with farmer selling, subsequent dwindling supplies as the old crop year ends and a new harvest approaches. These price moves are seasonal in nature and add validity to the statement: “The cure for low prices is low prices and the cure for high prices is high prices.”
Based upon knowledge of the fundamentals, you could conclude that wheat should be worth more in March than in July after the new crop becomes available. The spread situation needs to be analyzed with its seasonal tendencies plus the overall technical and fundamental picture before concluding that a spread trade should be entered.
The relationship between different contract months varies with the market — livestock tendencies don’t necessarily match those for energies. For the grain spread trader, it’s necessary to understand that agricultural grain products have a well-defined carrying charge relationship within a crop year. Expected differences in supply and demand make delivery month patterns different. Prices rarely achieve full carry largely because of commercial participation.
SEARCHING FOR A PREMISE
For most U.S. farm commodities, the annual harvest falls within a fairly well defined period. Producers sell a substantial part of their crop at harvest time to meet production obligations, the costs of new machinery, the cost of land preparation for a new crop, fertilizer costs and other related production expenses. A portion of the crop may be held in on-farm or local elevator storage waiting for potentially higher market prices in the months ahead.
In any case, the concentrated selling during harvest pressures cash prices at harvest time. If the crop is large enough to satisfy normal demand requirements, buyers are not eager to accumulate supplies. The net result is usually a seasonal price decline that tends to coincide closely with harvest.
As cash prices weaken, they pressure near-month futures prices. Farmers and dealers who expect to sell the physical commodity within a few weeks hedge in the nearby delivery months. If prices of the nearby futures months do not decline to near the level of the cash markets, the owners of inventory will deliver their inventory against their short futures position.
Normally, the process of selling the near futures months continues until its price has declined to a level close to that of the cash market, which tends to depress the price of the nearby futures contracts relative to the prices of the more distant delivery months. This can be expected to continue as long as nearby cash market supplies are ample and plenty of storage space for the commodity is available.
Seasonal price tendencies are more consistent in some markets than in others. The timing of seasonal highs and lows may be distorted by several different fundamental factors such as extended adverse harvesting weather. This may result in a later-than-usual seasonal low price or it may diminish the impact of the harvest on prices.
An unusually early harvest can cause earlier-than-expected harvest movement and historically premature hedge selling in futures. Even if harvesting occurs on schedule, producers holding production off of the market expecting higher prices later can prevent seasonal downward pressure on prices. Under such circumstances, the seasonal low price may be higher than expected and seasonal gains from this level may be limited.
Many spreads in the agricultural markets have strong seasonal relationships. Some of these relationships have changed over time as production and consumption patterns have changed.
An example of this change would be the movement of corn and soybean production to areas outside the United States. Research is necessary to analyze these effects because changes in one market might not lead to a total breakdown of the seasonal relationship of the spread.
In short, we need to reexamine seasonal tendencies with an eye toward 100% objective analysis.
In next month’s article, we will show how to conduct objective spread analysis on the grain markets and begin developing mechanical systems for spread trading that can be backtested.
Murray A. Ruggiero Jr. is a consultant in East Haven, Conn. His firm, Ruggiero Associates develops market-timing systems. He is editor-in-chief of Inside Advantage Gold Club (www.iagoldclub.com) and author of Cybernetic Trading Strategies (John Wiley & Sons). E-mail: ruggieroassoc@aol.com.
