Gasoline Demand Is A Red Herring For The Oil Market

Image from

Gasoline demand is a red herring.

A red herring is something that ​takes ​attention away from a more ​important ​subject. Gasoline demand distracts from the more important subject that there is no fundamental reason for the current oil-price rally.

U.S. Gasoline Consumption Has Fallen 2 Million Barrels Per Day Since 2005

Those who believe that gasoline demand is the fire behind oil’s recent rally confuse production with consumption. They also don’t understand that Americans increased their driving when oil prices were $100 per barrel and continued to travel more miles throughout and despite oil-price highs and lows.

A recent Bloomberg article stated, “American gasoline consumption rose to 9.25 million barrels a day in March, an all-time high for the month.”

9.25 million barrels per day is product supplied, the measure of how much gasoline is produced in U.S. refineries. Total wholesale and retail sales is the measure of U.S. consumption and that amount is only 7.76 million barrels per day.

Consumption of gasoline in the U.S. has increased 802 thousand barrels per day (kbpd) since January 2014 but is 1,973 kbpd less than peak consumption in June 2005.

U.S. production of gasoline is 908 kbpd more than the post-Financial Collapse low in January 2012 but is 542 kbpd less than the peak in July 2007 (Figure 1).

Gasoline Product Supplied, Total Sales & Exports
Figure 1. Gasoline product supplied, total sales and net exports (12-month moving averages). Source: EIA and Labyrinth Consulting Services, Inc.

Meanwhile, net gasoline exports are at record high levels. Exports have increased 1,443 kbpd since June 2005.

So, consumption has increased but remains far below pre-2012 levels. Production is again approaching earlier peak levels but most of the increased volume is being exported. The belief that U.S. consumption is approaching record highs is simply not true.

Americans Are Driving More But Using A Lot Less Gasoline

Americans are driving more than ever before. Vehicle miles traveled (VMT) reached an all-time high of 3.15 trillion miles in February 2016 (Figure 2).

VMT have increased 97 billion miles per month (3%) since the beginning of 2015 and gasoline sales have increased 187 kbpd (2%). The rates of increase are not proportional.

Total U.S. Gasoline Sales & Vehicle Miles Traveled
Figure 2. Total U.S. gasoline (wholesale and retail) sales (12-month moving average) and vehicle miles traveled. Source: EIA, U.S. Federal Reserve Bank and Labyrinth Consulting Services, Inc.

For example, VMT was fairly flat from mid-2011 until oil prices collapsed in September 2014 yet gasoline sales fell more than 1 million barrels per day during the same period. Americans traveled the same number of miles but used a lot less gasoline. Even with the VMT increase since 2015, sales are still 539 kbpd less than in January 2009.

Vehicle Miles Traveled Independent of Gasoline Prices

Gasoline prices fell along with oil prices in late 2014 and consumption increased (Figure 3) but the relationship between prices and consumption is not straight-forward.

Gasoline Price and Sales
Figure 3. Gasoline retail prices and sales. Source: EIA and Labyrinth Consulting Services, Inc.

Prices reached a low in January 2015 and then increased to a peak in July 2015. Gasoline sales increased right along with prices, not exactly the relationship we would expect. Although gasoline usage is strongly seasonal, consumption is not simply a function of price.

Figure 4 shows that vehicle miles traveled increased from March 2013 through September 2014 when oil prices averaged almost $100 per barrel. It is unclear why Americans began driving more but it was not because gasoline prices were cheap.

Figure 4. Vehicle miles traveled and WTI crude oil prices. Source: EIA, U.S. Federal Reserve Bank & Labyrinth Consulting Services, Inc.

When oil prices fell to less than $43 per barrel by January 2015, there was no change in VMT growth. Furthermore, VMT continued to increase despite cycles of high and low prices in 2015 that first rose to $60 and then fell to less than $30 per barrel.

The factors that underlie driving behavior are complex. Gasoline price is among those factors but there is no clear correlation between price and consumption.

To put gasoline consumption in context, the 187 kbpd increase in gasoline sales since January 2015 is 1.2% of monthly U.S. oil consumption. The increase in product supplied is 1.7% of monthly consumption.

Crude Oil Imports Increased As U.S. Production Declined

As U.S. crude oil production began to decline in 2015, more oil was needed to produce gasoline and other refined products so imports increased.

From April 2015 to March 2016, oil production decreased 660 kbpd (-7%) but net crude oil imports increased 800 kbpd (+10%) (Figure 5).

Crude Oil Production and Net Imports
Figure 5. Crude oil production and net imports. Source: EIA April 2016 STEO and Labyrinth Consulting Services, Inc.

When analysts mistake gasoline production for consumption, they include gasoline made from imported oil in their celebration of increased U.S. “demand!”

Moreover, increased imports are also included in U.S. inventory volumes. If the whole situation seems complicated, that’s because it is and the lesson is to be careful about your sources of information.

Gasoline Demand Is A Red Herring

Americans are driving more than ever but using less gasoline than a decade ago. A lot of this is undoubtedly because of greater fuel efficiency. It seems unlikely that gasoline consumption will ever again reach levels before the Financial Collapse in 2008-2009.

The correlation between gasoline price and vehicle miles traveled is not very good. Americans started driving more when prices were very high and kept driving more despite large fluctuations in price. I cannot explain this except to observe that aspects of gasoline consumption must be somewhat inelastic. Price matters but not as much or in the same way that I imagined.

What is clear is that gasoline consumption is not a significant factor in the recent oil-price rally. The collective consciousness that drives the oil market is fed up with low oil prices. It looks for and sometimes invents excuses to raise prices and ignores compelling reasons to lower them.

Gasoline demand is the most recent invention and it is a genuine red herring.



  • hugh owens

    Thank god and perhaps Forbes for a semblance of independent analysis. Bloomberg early on was independent to an extent but unfortunately has gone the way of independent news organizations such as the NYT and has deteriorated into just one more garbage publication spewing opinion disguised as reporting. Bloomberg would probably explain the increase in VMT with low gasoline use caused by the increase in Tesla drivers. Bloomberg has published glowing articles on Tesla and driverless cars and says driverless cars will cut congestion in cities if they were adopted. Tell that to the average LA commuter. Driverless commuting is a reality in a good part of the world. It’s called the train and mass transit and the first thing Obama did was bail out GM and Chrysler and pour billions into crumbling highways instead of rebuilding a national rail network. The neat thing about steel rails is that they don’t crumble .

  • Ryan Cobb

    Hi Art,

    Personally don’t believe this current rally has much to do with gasoline demand or any fundamental influence what so ever. What I see is a purely financial move whereby the the futures curve was manipulated in a way that short money could no longer exert pressure on the front month contracts and/or short money took profits and were unable to reestablish the short trend once that interest was cleared. For the true hedgers, those with physical supply, there was never much difference between $30 oil and $40 because the back end contracts guaranteed $40’s on a forward sale. The nature of financial markets, the prevalence of algorithmic trading, underpins a rally like this even when the fundamental reasons are not there. Gasoline demand, weakening dollar, declining production are all inputs skimmed from headlines that drive the buying. Now that the futures curve has flattened and the momentum is up, it is much harder for traders ,who see the fundamental picture, to exert pressure on the near month contracts. Something will have to give as fundamentals usually win out. Either the back months will have to move higher or spot prices have to collapse. In a roundabout way, spot prices represent inventories and back months- demand. Since inventories do not appear to be declining, I expect the move to come at the expense of spot prices. Just my two cents.

    • Arthur Berman


      I agree with your comments. It is certainly a trader’s market and to some extent, it always is. With so much inventory, however, the effect of trading seems amplified since massive storage puts fundamentals into another dimension of abstraction.

      I wrote the post to provide the data behind the single largest use of crude oil and, of course, to give readers better knowledge of this red herring.

      Whenever I speak or am interviewed, I get questions about the often silly noise about the reasons for oil price fluctuations and lately, gasoline consumption is the main theme.

      All the best,


  • Daniel Pearson

    “What is clear is that gasoline consumption is not a significant factor in the recent oil-price rally.”

    “The collective consciousness that drives the oil market is fed up with low oil prices. It looks for and sometimes invents excuses to raise prices and ignores compelling reasons to lower them.”

    “Gasoline demand is the most recent invention and it is a genuine red herring.”

    Art – Your conclusions that gasoline demand does not and has not pushed crude prices higher is very well demonstrated by the data you presented. Thanks for the data to support your gasoline assertions. So, who is the ‘collective consciousness’? Is it the media, or backers/bankers of US shale oil which need crude prices to be higher as well as needing investment money to stay afloat?

    Now that inventories are still high (as demonstrated by your previous articles) I assume crude oil prices should be lower. I just read an article regarding how much more production is being ramped up in Iran, Iraq, Saudi, and other countries. The increases in volumes are collectively higher than the production drops from the US shale producers. Do you see inventory fundamentals any time soon used by the media to demonstrate that crude prices should be lower? Regards, DRP.

    • Arthur Berman


      The collective unconscious is all of us and our perception of fundamentals, current events and the global economy. Put your attention on your daily-weekly-monthly sense of how things are and are most likely to go. That is your part of this consciousness.

      As Ryan has insightfully explained, traders are moving this market. With so much oil in storage, fundamentals are somewhat of an abstraction. There are, however, interesting reductions in certain sectors of U.S. comparative inventories that are probably symptomatic of why prices are higher and higher price is the only thing that will reduce inventories. Yesterday, U.S. crude oil inventories increased but comparative inventory fell and price is higher so far today by $1.41.

      The short workers strike in Kuwait emphasized the relative fragility of supply despite the global surplus. The effect of lower price on investment is sobering and the collective unconscious of the market is taking that into account.

      I continue to expect this current price cycle to move generally downward over the next month or so before the next cycle starts that will move prices closer to $50.

      All the best,


  • Paul B Smith

    Have you accounted for fact that wholesale gasoline sales and exports are neat gasoline, but retail gasoline sales usually include 10% ethanol?

    • Arthur Berman


      Ethanol is of course a factor but it is also part of the calculation of liquids supply. I am not trying to forecast long-term consumption behavior in this post but am merely showing that gasoline consumption is not behind this price rally.

      Thanks for your comment,


  • Flora

    Just want to put out a big thank you to Art who has done a great job producing insightful analysis of the oil situation. Since early last year I have been following your work on oil and gas and many times amazed by your thorough analysis, acute business sense and very good calls. I wish I had followed you earlier and more closely and I would be in much better situation than I am right now. Please keep up the awesome work!

  • Daniel

    “The 70 percent rise in crude oil prices from the lows of $27.1 per barrel to a high of above $46/b in a matter of three months is being driven by speculative activity—make no mistake about it. The speculators have latched on to every bit of rumour and news to bid prices higher, and this has nothing to do with the real fundamentals.”

    “However, speculation can boost prices only to a certain extent in the short-term. After this, the fundamentals take over. The extent of speculation is enormous, though the daily production of oil in the U.S. is around 9 million b/d, the WTI crude oil contract trades more than 100 times the produced quantity, as highlighted in this January 2016 post.”


    First, thanks for your reply. Second, I read the above statements from ‘OILPRICE” (May 3rd) which is inline with your current article regarding drivers of crude prices, and your previous comments regarding the biggest Casino in town. Thanks again for your reliable data driven articles. Regards, Daniel. (PS – I do see the significant jump in oil of $2.75 from Wednesdays low to today’s high.)

  • Alexandre Ganea

    “It is unclear why Americans began driving more but it was not because gasoline prices were cheap.”

    Can higher VMT be incidently related with low-wage jobs being lost or replaced? For example, this picture (can be verified on – shows manufacturing jobs being replaced by “food and drinking” services, at an equal rate. Can this mean people travelling more miles to go to work? (I assume manufacturing jobs to be more stable, meaning people working there for longer time, so potentially living close to the factory – as opposed to volatile jobs in the food industry) That would also explain the inelasticity between the VMT and price, you have to get to work no matter what.

    Jean-Marc Jancovici has also observed a similar trend in France, where low-wage workers tend to travel more and more miles to work in the past decades. Interestingly, this is also the slice of population that correlates with higher right-wing populist votes.

    It would interesting to correlate VMT and gas consumption with other indicators or statistics?

    Many thanks for all the insightful articles! All the best!

    • Arthur Berman


      That was exactly my guess! The U.S. has added lots of mostly low-paying jobs with limited hours since the Financial Collapse. People have to drive more to earn at best the same as before.

      Thanks for the interesting reference to Jancovici’s work in France.

      All the best,


  • Frank Cassidy

    I woud like to add a thing or two. All the western world is consuming less, the trend is crystal clear but it’s also clear that the rest of the world is consuming more, but it’s absolutely not crystal ! Infact, if production/consumption data from the statistical offices of western countries are more or less reliable, their equivalent in the rest of the world are imprecise, delayed or utterly opaque. This means that real data and forecasts are becoming more obfuscated than ever and we’re collectively paving the road to some spectacular surprise along the road.

  • Arthur

    Not sure where else to post these questions: 1) on the STGS May 10, 2016 Power Point presentation what do you think accounts for the difference between the EIA and IEA supply balance estimates on slides 6 and 8, and which do you suppose is more accurate? 2) Any update on your thinking about natural gas supply? Apart from the DUCS I am still hearing that here in the Marcellus/Utica there are over 1500 already completed wells (DACS?) waiting to be brought online depending on a lot of factors, mostly pipeline access. Although storage dipped to near the top of the 5 year average last week we still seem well stocked with natural gas with ample additional supply at the ready.

  • […] In beiden Fällen spielen die sich nur langsam einstellenden, aber ständigen Effizienzverbesserungen der Verbrennungsmotoren eine Rolle  siehe dazu z.B. den Aufsatz zu Benzin in den USA von Art Berman, hier. […]

  • “product supplied, the measure of how much gasoline is produced in U.S. refineries”

    With all due respect, but this is a gross misinterpretation. Product supplied represents consumption, as it is

    “field production, plus renewable fuels and oxygenate plant net production, plus refinery and blender net production, plus imports, plus net receipts, plus adjustments, minus stock change, minus refinery and blender net inputs, minus exports”

    readable in the EIA glossary.

    I stopped reading at that point.

    • Arthur Berman


      With all due respect to you, if “x” number of cars are produced, does that equal how many cars are sold (consumed)? I don’t think so. Just because EIA has an arbitrary definition, does not make is so.

      All the best,


Leave a Reply

This website uses cookies and asks your personal data to enhance your browsing experience.