Sunday, April 24, 2011

Seasonality of Crude: Buy in March, Sell in May

Overview:  Crude Oil & Gasoline Seasonal Tendencies
As we start this new year, a number of events are likely to occur along with the normal changes in the weather. January gasoline is typically the lowest in any year and, despite the common mythology, gasoline consumption does not normally fall steeply after Labor Day and then recover miraculously after Memorial Day.
We do see an element of driving disappear after Labor Day, as drivers in the 16 to 25 year-old age bracket tend to drive less, or at least more predictably. Family vacations are also over by that point, as a general rule. But, there are pockets of demand during foliage sighting season and Thanksgiving Weekend is always the best four-day driving period in any year in which July 4th does not fall on a Tuesday or Thursday.
There is usually good driving through the month of December into New Year’s Eve, but it traditionally falls off a cliff right after the champagne glasses touch to ring in a new year. People park their cars and drive to work and school and to appointments. But it is not until March or April that more discretionary driving normally returns.
Refineries know this and they typically plan maintenance turnarounds from January through April or early May. During this period, there is a definite tendency for gasoline inventories to be drawn down; even though demand starts the year at its lowest levels, the maintenance usually goes on long after demand has started to mount a comeback.
Market Reaction
The market reaction is not what most people might believe. Regardless of the overarching trend, prices have a long history of advancing from early March through the middle of May. Despite the fact that the so-called “Driving Season” starts with Memorial Day and peaks on Independence Day, most of the buying has come into the market long before the end of May.
Gasoline Seasonal Trade
We are now trading “RBOB,” but it is effectively “unleaded gasoline,” and we include its history in all its various incarnations, from leaded regular to unleaded to oxygenated to RBOB. The tendency for prices is to advance from early March. Following the loose guideline of buying in the first two weeks of March and selling sometime before May 15th, the seasonal tendency has worked in 25 out of 26 years - or 96.15% of the time. It is the strongest and most reliable seasonal tendency of any commodity futures contract.
Over the years, various specification changes seem to have helped. In 2003, prices rallied through March - in the leadup to the invasion of Iraq on March 19th of that year - and they then dropped into May. We advised against following this seasonal that year, as we saw prices advancing in January and February as it became clear that war with Iraq was coming.
Despite 2003, buying in March has yielded profits consistently. The buying part is a great deal easier than the selling part, but profits are there to be realized if one uses money management techniques and starts to take profits as soon as prices get overbought in May.
The following table shows whether one could have bought June gasoline futures on March 1st and then have been able to sell it at a profit at some point in the first two weeks of May. As one can see, there were profits available in May every year except 2003.

Tactical Variations   
One of the best variations is to buy on March 1st, and sell 50% on May 1st and 50% on May 15th.  One can use any system that buys in early March and sells later.  One can start selling at any time after April 1st.  The table below shows the profit and loss figures for the March 1st to May 15th period, taking half profits on May 1st and half on May 15th.  Four of the last five years have been spectacular, with the four largest gains coming in the four years between 2006 and 2009.

The following table shows the result of buying on March 1st and selling half on May 1st and half on May 15th, or on the days closest to these dates.

Using that system rigidly, there would have been profits in 23 out of 26 years for a net gain of 10.59 cents a gallon or $4.447 a year.  If one had placed this trade every year, one would have made or $115,626 per contract before commissions and fees, including the 20 cent loss in 2003.  If we remove that year, when we did not recommend this trade, there would have been average yearly profits over 25 years of 11.83 cents, or $4,969.60 per year – on one contract.  The accumulated profits over 25 years would have reached $124,240.20 before commissions and fees.

Six of the last eight years have yielded profits of more than 20 cents a gallon.  In 2005, it would have lost two-hundredths of a cent.  We recommended NOT doing this in 2003, because prices had rallied strongly leading up to the March 19th invasion of Iraq. 

Every year, our biggest fear is that too many people know about this trade for it to keep working, but this trade’s track record is second to none in commodities trading.  We are not sure what to do this year, with the current unrest in the Middle East.  The unrest could help or it could give us an early top.  We still have a couple days to decide whether to recommend the trade or to suggest traders avoid it. 

If we do recommend it, from a money-management perspective, one absolutely must pick a point at which one would admit to being wrong.  This needs to be viewed within the broader framework of being a trade one would commit to making each year for the next five or ten years.  If this year does not work, for any of a number of reasons, we must be in a position to try again next year and in the years that follow, to bolster our chances to make this strong seasonal tendency work for us over time. 

Viewing this as a one-time trade to be made just this one year, in a way that could leave one’s business vulnerable and unable to try again in following years would be a dangerous mistake.  The fact that it has worked in the past is no excuse to bet the ranch on it now.  It is simply a tool to be used – on a continuing basis – to increase one’s chances of long-term success in a business that has become more volatile and complicated.  Regardless of the results in 2011, this cannot be judged by the results of a single year or even two or three years, but must be viewed over an extended period.  One must trade this accordingly.

Because of the risks that such a trade entails, some traders may be better served by buying call options, in which the risk can be identified and limited ahead of time. 

Beyond May
While the trades outlined above take advantage of the seasonality from March to May, there is a general tendency for oil prices to advance through the summer as well.  This is more prevalent in crude oil than in refined products, and gasoline prices tend to see their highest prices in mid-May.  They do often move higher, though, from Independence Day into autumn, although here, as well, the trend is more firmly established in crude oil. 

In the case of both refined products, we tend to see yearly highs ahead of the season during which their demand reaches its peak.  As we move through summer, part of the strength in crude oil is lent to it by heating oil, which most often peaks in mid October.

Crude Oil Summer Seasonal Tendency
Recent years have confused a tendency that used to exist for prices to weaken from the end of May into early July.  But the tendency of prices to rally from early July to some point before the end of summer does still work.  Using the table below, we are defining success in those cases where one could have purchased October crude before July 15th and been able to sell it at a profit at a point before the end of summer, defined here as September 15th.  Check-marks in the left-hand part of the column suggest it was best to sell on August 1st.  Check-marks in the middle of the column suggest one could have held out until September.

The check marks that are aligned to the left signify years that one could have made profits, but would have sold on August 1st.  The center-aligned check marks are for years in which one would have made profits by selling in September, although in some cases one could have also made them earlier.  Based on our very loose rules, one could have made a profit by skillfully buying before July 15th and selling somewhere between August 1st and the end of summer in 24 of the last 28 years.  This shows a general trend higher through the summer.

One can see in the chart above a marked tendency for prices to advance through the end of the first quarter into the second quarter.  The year 2003 needs to be viewed differently because of the lead-in to the war in Iraq.  In a number of these years, it might have been better to buy in December or even January, but over the full history of trading in the gasoline contract, buying in the first two weeks of March and then taking profits before the middle of May has been the most consistent system seen.

Conclusions
While there is a general tendency for gasoline prices to rise from some point in December, January, February or March to a point in May and then possibly from July to September, the most reliable pattern is the one that sees prices advance from early March to the middle of May. 

The worst months to be long gasoline would seem to be June, October and November, with the first half of December added into November. 

Retailers should integrate a policy of buying caps or calls on product in early March, to be held into May, and may want to buy, less aggressively, around Independence Day, with a view to holding a long bias position through the end of summer.  The only way to approach this trade is as a multi-year campaign, rather than as a single-season “bet” that will either make or break one’s business.  One needs to be prepared to take this approach year in, year out, over a period of five, 10 or even 15 or 20 years.