Gasoline demand plummets

By Monday, oil and gasoline futures prices had given up all of the gain they’d experienced since Katrina. Today we learned that U.S. gasoline demand has plummeted. Both developments were pretty surprising, but are surely related.

I noted on Monday that the October futures price for both gasoline and crude oil were back to their values prior to Katrina’s landfall. Although both prices made significant moves back up today, I still find this quite remarkable. Fifty-six percent of Gulf oil production remains shut in, and the Energy Information Administration reported today that four refineries amounting to 5% of total U.S. capacity could be shut down for an extended period. Release of oil from the SPR, temporary rationalization of fuel standards, and assistance from foreign countries surely helped, but could not be expected to reverse completely the significant lasting damage wrought by the hurricane. So why didn’t the storm have a bigger effect on prices?

gas_demand.gif



Calculated Risk
caught today’s big story from the Energy Information Administration, which was that the drop in U.S. gasoline demand (discussed by Big Picture and William Polley last week) has turned into a freefall. As the above graph reveals, U.S. gasoline demand had been above the values of the previous year for all of June and July. But the August price hikes brought use back in line with the 2004 values. The post-Katrina price hikes and shortages sent it plummeting for the week ended September 9 to a value more than 6% below where it had been for the week ending September 10, 2004.

Is this tremendous drop in gasoline use just an anomaly? Perhaps. But the rising gasoline demand of the first two months of the summer in the face of rising prices seems a bit anomalous itself. A case could be made that U.S. consumers are finally responding in a significant way to price incentives.

In a third related development, the Chinese seem to have put on hold plans to fill their own strategic petroleum reserve. They may be Communists, but they have the same sensible desire to avoid buying high and selling low as anybody else.

Do demand curves slope down? I’ve been telling my students they do for years. And I imagine I’ll be telling them the same thing for a few more years to come.


Technorati Tags: , , , ,

42 thoughts on “Gasoline demand plummets

  1. Stuart Staniford

    I wonder if demand curves frequently have near-discontinuities at psychologically significant values, rather than being smooth. Professional traders are always speaking of the importance of the market passing through particular levels. Seems like $3 is enough to finally have some impact.

  2. Hal

    I guess we’ll know if $3 is a magic number soon. Gasoline this week is back below $3. It’s down about 10-15 cents around here from their highs. We’ll see if demand starts climbing back up.

  3. Tom Lord

    “So why didn’t the storm have a bigger effect on
    prices”
    I’m amateurish at this economics and oil analysis stuff but lemme try a speculation:
    1. Release of strategic reserves in the U.S.
    2. Slightly surprising release of strategic reserves on the behalf of the U.S. by European allies.
    3. While the productive capacity in the gulf took
    a bad hit, the refineries, import ports, and
    pipelines could have been a lot worse.
    4. Little bits of ability to ramp up production, here and there, elsewhere in the world were displayed.
    5. Lots of political will was apparent throughout the industry and many governments.
    The market is “thinking” that there is enough in
    reserve and production capacity to cover any
    initial shock and an event like Katrina is likely
    to help push drilling of known but previously
    off-limits fields. That doesn’t much change
    peak oil predictions and it certainly doesn’t
    help the environment but it explains why the
    futures markets look like they do.
    It’ll still be a rough winter and there’s a good
    chance the market is underestimating the risk of
    additional infrastructure loss and overestimating
    the will to dig everywhere that’s currently
    off-limits.
    -t

  4. CalculatedRisk

    I’m glad to see the drop in gasoline demand. This is definitely a time for caution and conservation – even if conservation is a result of high prices.
    Lets just hope Invest95L doesn’t become “Philippe” and hit the GOM late next week. We’ve had enough for this year.
    Best Regards.

  5. TI

    The destruction in Louisana and Mississippi must have some effects too. When roads are flooded there are not much driving and most of the evacuees still commute nowhere. Might this be a 1 – 2% share?

  6. Robert Schwartz

    The commentors on this blog had absolutely convinced me that the demand for gasoline was price inelastic. These numbers can’t be right;-)

  7. Houston's Clear Thinkers

    Markets at work

    A funny thing happened in response to the recent run-up in gasoline prices resulting from Hurricane Katrina — demand for gasoline dropped dramatically. Clear Thinkers favorite James Hamilton puts it all into perspective….

  8. John

    Unfortuantely, at times of panic (and there was some gas panic following Katrina) consumers are often not the rational actors we need to think they are for a typical analysis of price elasticity. So I wonder whether demand was responding to factors other than just price.

  9. Bill Ellis

    From the DOE’s web site we note the following description of the methodology used in estimating consumption.
    “However, while petroleum analysts equate “Product Supplied” with consumption, there is a lag between petroleum delivered into the market and petroleum actually consumed. The product may sit in a tank belonging to a wholesaler, a retailer, or even a consumer before it is used. We cannot capture these small movements and therefore can be surprised by short-term volume fluctuations as these tanks are unexpectedly filled or emptied. Thus, the methodology overstates “demand” when the product moves into wholesaler or retailer storage and understates it during the period when it is actually consumed.”
    As I understand this, consumption is essentially measured as shipments from the refineries, or from import terminals. Thus, any draw down in retail inventories, or in consumer inventories (the amount of gas in the car), is not reflected in the reported consumption figures.
    Undoubtedly there were some that filled up thier tanks inn fear that there would be a limit to availability, and others that delayed refueling due to funds constraints. This would happen at both the retail and consumer level.
    While the inelasticity of petroleum demand is probably overstated, it is likewise probable that this spot report overstates the drop in consumption.
    Bill

  10. ed

    When talkinga bout “elasticity,” everyone should always remember that there may be a big difference between long run elasticity and short run elasticity. Furthermore, there may be a different response to changes in price that are expected to be temporary vs. those that are expected to be permanent.
    For example, perhaps some consumers thought the price spikes were only temporary, and so filled their tanks only part way…this could account for some of the drop in demand.

  11. MJB

    Maybe I’m missing something, but isn’t this obvious?
    How could demand for gasoline not fall when there is less gasoline available? How could demand have conceivably remained at 9400 when there wasn’t 9400 out there to be had?

  12. John

    MJB
    Except for changes in stocks (inventories).
    I don’t think there is any evidence the US is *short* of gasoline (in the sense that gasoline is just not available ie the pumps are empty). What there is is a market which is raising the price to reflect the costs of getting gasoline*
    * put it another way, people are not being turned away from the pumps, they are simply having to pay a price they don’t wish to pay. The price is what is making supply = demand.

  13. JDH

    Bill and MJB– Excellent points. All we really know is that the shortfall didn’t come out of refiners’ inventories, meaning that somebody at the wholesale or retail level wasn’t buying as much gasoline from refineries as before. One possibility is that wholesalers and/or retailers reduced their inventories. So what we observe is a combined demand from consumers, retailers and wholesalers for the output of refineries, and what we know for sure is that this combined demand went down.
    And, if you were a wholesaler or retailer looking at the Sept. 1 spot price and the October futures price, it certainly would have made sense for you to draw down your inventories. That response I think is part of what we think of as the elasticity of demand– if the price is higher in September than October, then the rational response is for demand to be lower in September than October. However, to the extent that’s part of what’s going on, it’s only helping us get around the short-run problem of the specific shortfall in September rather than the longer term problems.

  14. Bill Ellis

    Professor:
    I concur that there are definitional problems of “consumption.” However, my concern is that when the news trumpets a fall in consumption that may be based only on draw downs they are doing a great diservice. Course the news is good at that.
    Bill

  15. the long view

    With respect to the question why didn’t the storm have a bigger effect on prices?, capacity expansion might be part of the answer. Rising price 1998-present appears to be calling forth a big capacity expansion, especially since around the fourth quarter 2002 when price began to exceed its 30-year average. Drilling initiated 2-3 years ago should be putting new product on the market right about now. Product and futures prices seem to have finally hit this ceiling on speculation.
    But how did price reach such heights is in the first place? Futures traders read the same things as everybody else and these things offer incompatible portraits of the market. Peak Oil, which seems to have morphed from theory to axiom, tells them price has barely begun to rise. The explosion of drilling activity since 2002 tells them new capacity is building so price should fall. Traders are in a kind of schizophrenia with respect to these ideas. This week, they believe in new capacity more than they believe in Peak Oil.
    Recall also the long string of events that have driven price ever upwards in the past 2-3 years; jailing Khodakovsky, attempted coup in Venezuela, civil violence in Nigeria, war in Iraq, bombing of foreign workers in Saudi Arabia and so on. Except for war in Iraq, the supply consequences of these disturbances, if there was any, have receded. Price, on the other hand, has not. So, again, the collapse of futures and the rapidity with which price has fallen may have more to do with an overbought, news-driven market before the storm than with some anomalous behavior after it. Or rather, rationality, which had become anomalous, has returned for the moment.
    As for the demand free fall, precautionary demand (hoarding) along the supply chain could help explain both the price rise before the storm and the apparent demand decline after it.
    With respect to higher 2005 demand despite higher price than 2004, the US consumer base is not static. Car and truck registration grows at about 1.6%, I believe. Also, new homes and commercial buildings that use oil heat are built every year. More than just these factors confound calculation of demand elasticity. Demand is also elastic with respect to income. We like to think of price and income elasticity as discrete measures, but surely they are not.

  16. TI

    There were of course physical supply distruptions that must show in the statistics. In disaster area there were refinery and wholesalers stocks and it was not possible for the wholesalers and retailers get gas in those conditions. It was not easy to get it from elsewhere, either. This did cause sharp price rises. What does this tell about price elasticity? If there are no gas there are none. The remaining gas will be allocated somehow, partly trough price mechanism, partly through overt or covert rationing (“We have closed. No gas”).
    The prices rose elsewhere, too. But the price hikes were seen as temporary which always causes some postponing of purchase. This is normal. We get the whole picture only afterwards.

  17. muhandis

    Oilcast #23 (www.oilcast.com)reports that the drop in gasoline consumption in the US is a result of a large number of cars and roads destroyed by Hurricane Katrina (not much traffic in and around in Louisiana and Mississippi).
    There is also a good report/warning at about 9min00 of Oilcast #23 from Barclay’s on future refinery capactiy and gasoline prices.

  18. Joseph Somsel

    I’d like to float the notion that consumers just stayed home more.
    When you see thousands of people without homes you begin to appreciate having a roof over your head. The disorders and damage were a shock to many people and I think a natural reaction to that shock is to “cocoon” a bit.
    I know I saw it in myself and my household.

  19. BK

    Quiz: Where would lower gasoline demand be first noticed in concrete terms? Answer: In the highest trafficked regions.
    Anecdotal evidence: #1 on my bad-traffic list is the DC area. I haven’t been there since Katrina, but maybe some of you could report in. #2 by my experience is my hometown of Alpharetta, Georgia ( a north Atlanta suburb). I remember in Los Angeles thinking “Hey, this traffic isn’t so bad.”
    Since Katrina, Alpharetta traffic must be 25-30% lighter. People are definitely staying home more – maybe working from home instead of driving the hour to downtown offices.
    Hook 2. ref pricing impacts
    Having worked in commodity sales/marketing for 15 years, you notice several lags in price impacts:
    1. Reaction time. People are busy. It takes time to A) get their attention, B) for them to digest what’s going on, and C) to make changes based on the new information. In the milk business, price impacted sales revenue around 90 days out. If that is the case, we’ll see Katrina’s true short-term impact in the next couple of months.
    2. Discretionary vs. Necessity. The amount that demand can quickly drop will be largely based on the percent of usage that is necessary (getting to work, delivering products, getting kids to school) versus the amount that is luxury (driving to ballgames or movies, trips to the beach or mountains). While both of these usage patterns will change with a surge in prices, the luxury usage will drop in the short term, while the necessity usage will drop over time (better MPG vehicles, moving closer work, etc.). The large impact of Katrina may be next Spring or Summer, if prices hold in the $60 range.

  20. peakguy

    But how elastic was demand really – the whole scale is not shown on the graph so it looks more dramatic than it actually is. Prices went from about 2.50 to 3.25 in a week – a 30% jump. Yet it only caused a 6% decline in demand. Add in the people in the NOLA affected areas that couldn’t consume (1-2%?) and the drop in demand is really pretty unimpressive.

  21. reno

    Perhaps all those people that topped off their tanks during and after Katrina hit haven’t purchased any gasoline lately. I have spoken with a few people who haven’t bought any gas lately as they “wait it out” for lower prices.. I for one am waiting for the lst possible moment to fill up again.. And yes, I have curtailed my driving..

  22. The Big Picture

    Gas Returns to Pre-Katrina Levels

    Last Summer, we looked at the question as to why Gasoline prices shoot up so quickly, but take much longer to drop. As the chart nearby shows, Gasoline futures have fallen to the pre-Hurricane Katrina levels. Oil is back to early August levels. Yet man…

  23. donna

    We tanked up all our cars the day before Katrina hit when it was obvious the GOM was screwed already. I’ve driven much less since, only for necessary errands (I have eliminated a 60 mile commute in the last couple years as well). Husband drives only to work and son only to school since Katrina. It’s not that hard to cut back when price is high.

  24. cruder

    Demand by consumers may not have gone down. The amount supplied by refineries to wholesalers/retailers went down by about 400kbbls/day for 7 days = about 2.8 mil bbls for the week.
    there are 210 mil cars in the US, avg gas tank 20 gals = 4.2 bil gals = 100 mil bbls.
    I dont know what the capacity of tanks at retailers and wholesalers is (if anyone knows please respond) – but i would guess at least 100 mil bbls ( enough for every car to get one fillup seems in the ballpark?).
    It is quite possible that consumer demand did not go down at all ( other than those disrupted /distracted) and what happened was that 2.8 mil gals out of , say, 200 mil bbls of ‘inventory” residing in car gas tanks+ Retailer/wholesaler tanks – was drawn down. I am unable to draw ANY conclusions re demand elasticity from this data.

  25. cruder

    Sorry – that last paragraph should have read ” 2.8 mil bbls ” NOT ” 2.8 mil gals. ( I guess I had gals on my mind!!)

  26. cruder

    To continue:
    Why might retailer inventory be drawn down? Transport problems/weather prevented some deliveries to retailers?
    Why might consumers drawn down the inventory of gas in their cars?
    Weather, gas lines or having topped up the car in the previous week etc might have kept consumers from making a trip to buy gas?
    In any event , a draw down of ,say, 1% of the inventory in car gas tanks and retailer tanks is hardly big and quite plausible.
    I suspect end consumer demand for gas will be hardly affected by $3/gal or even higher. Europe/Asia has been paying $4+/gal for a long time ( due to high taxes). For someone who uses 100 gals a month – even a $2/gal increase is only $200/mo more which is relatively small compared to other costs – housing etc.

  27. dali lama

    It is quite possible that _some_ people are indeed driving less. This means that less will be spent on non-essential items, causing a softening in aggregate demand for September.

  28. Stuart Staniford

    Here in the SF Bay Area, there has been no tangible impact on traffic from Katrina whatsoever. We are all still ploughing up and down 280 as fast as the guy in the Porsche in front will let us. Traffic has been steadily increasing for over a year as the local economy continues to recover from the tech crash.
    On the question of elasticity. I was reflecting on it last night, and it seems it’s really something of a polite fiction to view the demand/price relationship as having a functional form at all. The relationship of demand and price varies from time to time and place to place. A lot of change in price can produce very little change in behavior, then a bit more price can change behavior much more, and if the price goes back down, there’s no guarantee demand will follow the same path it took on the way up. The hysteresis effects are not negligible.
    It’s qualitatively useful to think about the demand curve, and clearly it’s qualitatively true that demand for gas is inelastic, but I don’t think there is a quantitative theory there.

  29. cruder

    The whole concept of demand curves/quantification etc. is just a lot of physics envy from social science academeic types. Social sciences including economics are about where medicine was in the middle ages.
    Does the US have enough Dollars to buy all the petroleum it needs? Silly question – we can create as much US Dollars as we want and we are.
    Does Asia have enough US Dollars? Umm- not unlimited as in the US case but a couple Trillion or so aint pocket change.
    So – thats about as much as we need to know to assess how countries will react when faced with a supply constraint in energy.

  30. Knowledge Problem

    WHY YES, THERE IS ELASTICITY IN THE DEMAND FOR GASOLINE!

    Lynne Kiesling James Hamilton has done the back of the envelope calculation to show that in the pasat two weeks, gasoline demand has plummeted, both relative to trend and relative to last year. His graph showing the effect is striking….

  31. T.R. Elliott

    The idea that a demand curve is not possible because of spatial and temporal variations is like saying an aggregate, statistical theory of nature (e.g. thermodynamics) is not possible because the motions of particles vary spatially and temporally. Or that there exist hystersis effects in the real world that are modeled.
    Paul Ormerod discusses hysteresis effects as they apply to economics.
    My feeling is that the demand curve right now can easily be described as elastic, because it’s the people at the margins–of society in terms of income–who are making the hard decisions of a beer after work or enough gas to drive home. The guy in the Porsche still isn’t thinking about it. And the reduction would not show up in commuting behavior to and from work. It would show up elsewhere.
    With significantly higher prices, e.g. $6/gallon, we will see a long-term hysteris effect in that people will forgo SUVs for smaller vehicles.
    The fact that currencies are fiat does not have any implications on supply/demand curves as far a I’m concerned.

  32. BK

    Stuart,
    Demand in the Bay area may not have dropped, but in my part of Georgia the price went from $1.89/gallon to $3.19, and the effect was noticeable.
    Maybe prices there didn’t go up 60%?

  33. Joseph Somsel

    “Physics envy”
    You nailed it, Cruder!
    And from that derives “social engineering”.
    Still, economics has provided insights and some predictive utility but as I tried to explain in my overly long, unreadable post above, no field of study is big enough to encompass the whole role of energy in human society.
    Maybe we need a “Unified Energy Theory”?

  34. Tim

    Ok, enough of the physics arrogance here.
    As a former engineer, and a current physician who also dabbles in computational fluid flow and computer programming in the spare time, I can assure you that the main focus of my day, human nature, is much more unpredictable than just about anything that is thrown at you in physics. Don’t forget that most of physics that people like to engage in is linearized models of whatever (pendulums, electronics, friction, celestial bodies in motion) allowing easier formation of governing laws which are then simplified into linear (or mildly nonlinear) equations. Human nature is much more highly unpredictable and nonlinear and not ammendable to derivation of laws governing its effect. Hence, more qualitative descriptions as opposed predictive equations.

  35. T.R. Elliott

    I agree Tim. As someone with a physics and engineering background, my primary concern with economics is that some economists (or people who like to play them on TV like Stossel and Stein) are unable and/or unwilling to accept the limitations. They speak with a determinism that then leads many to use them as straw men for the field of economics as a whole.
    Speaking of inelastic behavior, I’m looking at Reynolds paper “Determine the power of prices to change oil discover and production using a non-time dependent hubbert model.” His conclusion is that the supply of production and the rate of discover are inelastic with respect to price. He comes to this conclusion based upon an analysis of the lower 48 state US production, but my concern with analysis of this region is that it was–possibly–operating in an environment in which OPEC had sufficient cheap (to produce) excess capacity to increase investment risks for us producers. Just a thought. Not sure whether it is a valid idea.

  36. Coyote Blog

    Supply and Demand in Gasoline

    Via Lynne Keisling of the Knowledge Problem comes two good articles on supply and demand in the gasoline markets. The first is from James Hamilton, who analyzes the effect of gasoline price increase on demand and finds, amazingly to some

  37. The Glittering Eye

    Catching my eye: morning A through Z

    Here’s what’s caught my eye this morning:
    Argghhh! has a reflection on the whys and wherefores of rebuilding New Orleans. I continue to be confused by the public discussion. I certainly believe that if the people of New Orleans (where…

  38. Different River

    Elasticity of Demand for Gasoline

    (Note to non-economists: The “price elasticity of demand” is a measure of how much more or less of something people buy in response to a change in price.)
    Most people believe that the demand for gasoline is not very elastic — that is, if the pri…

Comments are closed.