Mechanics of Future Oil Price Volatility (A Flubber Cobweb)
Posted by jeffvail on February 5, 2009
I previously examined the interface between peaking oil supplies and oil price volatility
as a predator-prey system. With the rapid drop in oil prices, it’s time to add another
wrinkle to that story: widespread acceptance (psychosis?) about the stability of high oil
prices acted as a damper on oil price volatility. Now that a collapse in oil prices is more
than a mere theory, oil markets are poised for a long-term increase in price volatility.
The fundamental problem facing oil markets at present it this: while present supplies are
sufficient to meet present weak demand, these sources of production face rapid decline.
The current low oil prices are not sufficient to support the long term investment in future
supplies, conservation, and consumption efficiency that will be necessary to mitigate the
impact of this decline. Because of the time-lag between a sufficient price signal and oil
reaching the market (or demand being reduced), and because of the impact of the recent
price collapse on producer psychology, volatility will rapidly incrase as the market's price
signal must make incrasingly exaggerated moves to bring supply and demand into
equillibrium.
Without ongoing investment to support present production levels, production
decline rates will accelerate
Figure 1: This graph, from Merril Lynch, shows that reduced capital expenditure
will have a sharp impact on field decline rates, but assumes this impact will diminish
as our economy recovers and capital expenditure picks back up…
There are two key issues here:
1) The market’s price signals react over much shorter time-spans than new supplies or
investments in conservation or efficiency can be brought to market. Depending on the
specifics, it can take anywhere from 2 to 10+ years to bring a new oil-field into full
production. Therefore, even when oil was at $147/barrel, oil producers couldn’t
immediately realize profits from oil that would only cost $100/barrel to produce. The
same is true with many efficiency and conservation measures—while the most elastic
demand (e.g. Summer driving vacations) can be reduced rapidly, other conservation and
efficiency efforts take much longer. Electrified rail takes years to fund and build out, and
the gradual upgrade of the fuel efficiency of our vehicle fleet (or the replacement with
electric vehicles) requires years of consistently high fuel prices or efficiency regulations
that generally don’t take effect until several years in the future.
2) Our recent market experience—a rapid crash in prices—undermines efforts at longterm investments in supplies. At $150/barrel, oil companies were willing to invest
billions in new projects that, several years in the future, would bring production on line at
a cost of $50, $60, or even $80/barrel. Because of the time-value of money, and because
our energy futures markets are incapable of economically hedging entire oil megaprojects
a decade or more into the future, oil companies needed to leave a large price cushion.
$150/barrel oil did not justify a ten-year lead time to produce oil that would cost
$140/barrel, even under the assumption that oil would remain at $150/barrel. This
cushion shrank some during the (relatively) steady increase in oil prices from 2001 –
2008, and oil companies’ willingness and ability to finance these projects increased. But
now that the price of oil has crashed from $150 to $40/barrel, the prospect of a sudden
drop in price is more than a mere possibility.
The result of this will be long-lasting: if oil prices again reach $150/barrel, there will be
much greater reluctance to invest in long-range projects to produce oil at even $50 or
$60/barrel. Similarly, the cost and availability of financing such projects has been
dramatically reduced by the credit crunch. The result is that the more aggressive
producers—those with the least cash and greatest incentive to take risks—are the least
able to finance such projects. Those with the greatest ability to undertake such projects—
select oil majors and many national oil companies—are also the most conservative and
least willing to risk embarking on long-term, expensive production projects.
Continually Increasing Price Volatility
The principle results of this time-lag and shift in producer-psychology is that price
signals must become increasingly over-exaggerated to create the desired market effect.
The level of interest and willingness to invest in oil production that was spurred by
2008’s $100+ oil prices will not be regained if oil again hits $100 or $150—it may be
necessary for oil prices to hit $200, $250, or more to provide an adequate incentive for oil
companies to invest in oil with a $50+ cost per barrel.
And, even when the price level necessary to spur sufficient investment is reached, the
resulting production will not come on line immediately. Instead, it will take several more
years to reach the market. During this time-lag period between sufficient price signal and
new oil reaching the market (or new demand being destroyed), prices will continue to
move and push this market signal. If oil at $100/barrel is sufficient to incentivise
investment in oil production to meet then-current demand, the market will continue to
push the price signal past $100/barrel, especially as geological depletion continues to
grind and the incentives for geopolitical disruption continue to rise.
As a result of the price rise during this time-lag, more production investment decisions
will be made, and more demand will be destroyed, than was actually necessary to reach
an equilibrium point. The result will be an inevitable overcorrection and price crash,
restarting the price cycle. With each successive boom and crash cycle, the market
psychology will become increasingly resistant to a given move in the price-signal, and
the degree to which the price signal must move to create an equivalent market effect will
increase. This will result in continually increasing price volatility as the market swings
wildly to reconcile the ever moving targets with production and demand.
A Flubber Cobweb
The time-lag between market and fundamentals is exacerbated by inadequate futures
markets and our own psychology. On a general level, this is a common process in
economics, described by the cobweb model:
Figure 2: A classical economic “Cobweb Model” of supply and demand
This model shows a supply/demand spiral characteristic of a market searching for
equilibrium in a steady-state environment—it is NOT characteristic of oil prices in a
post-peak environment. Rather, once oil production has peaked, the spiral will work in
the opposite direction with increasingly divergent P and Q brackets.
Under peak oil theory, Q must gerally decline, even if there is significant noise present.
From a civilizational-perspective, I think there will be a great deal of information carried
by what happens to P as Q approaches zero: If P1 and P3 diverge or generally trend
higher as Q approaches zero, this may show a sign of fundamental economic strength in
the face of peak oil—perhaps indicating that the realized EROEI of alternative energy
supplies is sufficient to maintain a global-industrial economy. However, if P1 and P3
converge or generally trend lower as Q approaches zero, then I see this as the dimming of
global-industrial system.
One important note: I don’t think that these values carry meaning when measured simply
in absolute dollar values—inflation, deflation, and other currency games introduce too
much distortion. Rather, I think a measure of the dollar price of oil as a percentage of
median national and global incomes (in dollars) will carry much more meaning when
measuring the divergence or convergence of P1 and P3 going forward.
Add in declining supply due to geological and geopolitical peak oil and increasing global
population and oil price volatility will rapidly accelerate over the next decade.
Discussion Questions:
1. How will this system interact with the global economy (itself a key driver of demand
and cyclic in nature)?
2. Will oil price volatility be primarily characterized by increasing spread between price
highs and lows, by a shortening of the period between highs and lows, or by some
combination of the two?
3. This article suggests that oil price volatility will continually increase. At the same time,
I generally criticize other theories that argue for perpetual increase (in population, GDP,
resource consumption, etc.). Clearly, at some point volatility must either (1) slow or
decrease, or (2) reach a functional maximum at which point market signals are reduced to
meaningless “trading noise” and the market function ceases to provide utility. Which
result, and why? In the spirit of Kurzweil (link) or Moore’s Law (link), does this process
lead inevitably to the end of markets? Half in jest, what if the causal mechanism of the
Maya 2012 hypothesis (or insert your favorite apocalypse meme here) is simply global
markets grinding to a halt as prices cease to carry meaning and we can’t find a way to
reverse the complexification that caused this?
**In light of these three concluding questions, I hope that it is clear that I am not
intending to argue some fundamental “truth” with regard to ever-increasing price
volatility, but rather that I’m using that argument to set up the three concluding questions
(which I consider much more interesting and important in the long-run).
15 comments on Mechanics of Future Oil Price Volatility (A Flubber
Cobweb)
You must log in to post a comment. If you don't have an account, create
one.
ShareThis
Shunyata on February 5, 2009 - 10:18am
The dynamics your are describing can be captured by making your equillibrium supplydemand graph a dynamic graph.
First of all, the demand curve is extremely convex, even L-shaped. This creates very
exciting dynamics when you are near the up-leg as we typically are.
Depletion means that the supply curve is drifting upwards to higher price levels. THE
DRIFT RATE IS A FUNCTION OF CURRENT PRICE LEVELS. Similarly, our
economic over-leverage means that the demand curve is drifting leftward to lower
consumption levels. THE DRIFT RATE IS A FUNCTION OF CURRENT PRICE
LEVELS.
When prices are high the SUPPLY drift rate is low, and equillibrium can form, and
volatility is relatively low. A few decades ago the drift could even go negative (creating
gluts) but that doesn't seem likely any more.
But high prices acclerate the DEMAND drift, destroy the equillibrium, drop prices and
rapidly restart the supply drift. This spikes prices and starts the whole cycle over again.
Unless the economic situation stabilizes or significant new supply becomes available, we
are locked in a downward spiral.
Log in or register to leave a comment
nelsone on February 5, 2009 - 10:59am
Your predator/prey model reminds me of the old Wa-Tor game, from Scientific
American, IIRC. Here's an applet modeling that predator/prey dynamic:
http://www.leinweb.com/snackbar/wator/
The features include No equilibrium. It's chaotic, irrespective of initial conditions.
Vulnerable to extinction - particularly when the "world" is small
Stabilized by addition of more dynamic variables.
A friend of mine coded up the equivalent three-phase game, something like
plants/fish/sharks, and found that it's much more stable in all respects. That stability
couldn't be had by merely tweaking the lifespan, response time, or other features of the
simple two-element predator/prey relationship. But the more separate, interwoven
dynamic elements were present, the smaller the excursion magnitudes in population
became.
I think it's a worthwhile paradigm for the energy markets, if for no other reason because it
does away with the whole notion of equilibrium - complexity theory seems more
applicable.
Log in or register to leave a comment
Gail the Actuary on February 5, 2009 - 11:59am
I found some of the explanation of the cobweb model from Wikipedia helpful.
Simplifying, the cobweb model can have two main types of outcomes:
(1)If the slope of the supply curve is greater than the slope of the demand curve (in
absolute value), then the fluctuations decrease in magnitude with each cycle, so a plot of
the prices and quantities over time would look like an inward spiral, as shown in the
diagram. This is called the stable or convergent case.
(2) If the slope of the supply curve is less than the slope of the demand curve (in absolute
value), then the fluctuations increase in magnitude with each cycle, so that prices and
quantities spiral outwards. This is called the unstable or divergent case.
Oil prices would seem to be the divergent case.
It seems to me though that the world's economy is in a sufficiently weak state that the
very first bounce up will cause massive bankruptcies, and possible collapse of financial
systems around the world. I am not sure that we will have to wait for multiple iterations
to happen.
Log in or register to leave a comment
jeffvail on February 5, 2009 - 12:35pm
I agree that, depending on the resiliency of the global economy, we may not sustain very
many wild swings in the dollar-price of oil. I do, however, think that we need to also
measure volatility in the ratio of $/barrel:$/year median income.
In theory, oil could hold steady at exactly $40/barrel, but this could represent massive
price volatility if there are wild swings in household income, inflation leading to deflation
or vice-versa, etc...
Log in or register to leave a comment
nelsone on February 5, 2009 - 12:45pm
As useful as it is for a snapshot, IMO the Cobweb Model cannot account for the system
dynamics. Take a look at almost any real-world system's time history and you'll see the
same quasiperiodic but unpredictable behavior that oil (or pork bellies) prices display.
Log in or register to leave a comment
Kevin Gardiner on February 5, 2009 - 12:53pm
With the elasticity of demand so low, and our diminished future ability to compensate for
changes in demand by adjusting supply, I predict the seasonal variation in gasoline and
crude demand will have a larger influence on prices than even three years ago. The
normal seasonal varition in gasoline demand in the US is about 9 to 10 percent between
its low in mid January and its high, usually in June. With an elasticity of demand around -
0.1, we can expect January to June price increases of about 150%, or conversely June to
January price drops of 60%.
These wild swings will be layered on top of the predator prey swings. Also, by mid
summer 2008 it was obvious that the typical 9-10% seasonal demand swings may
become more muted going forward. Thus actual price swings due to seasonality may be
more on the order of 100% Jan to Jun and -50% Jun to Jan.
Log in or register to leave a comment
dohboi on February 5, 2009 - 1:10pm
How does the model represent the moment when dollars (or ultimately any currencies)
are left out of the picture completely and countries just start trading oil for other tangibles?
Wheat has already been mentioned as something the US might be trading for oil with SA.
Doesn't "price" become irrelevant then? Or is a bushel the new currency at that point?
How close are we to that scenrio? I've heard that some countries are starting to directly
trade commodities for commodities already, but I haven't heard that this is being done
with oil yet.
Log in or register to leave a comment
jeffvail on February 5, 2009 - 2:03pm
I think you're exactly right: if it comes to trading wheat for oil, then the ratio of wheat:oil
becomes the new price. I'd imagine this would initially be much less volatile because it
would be much less fluid--rather than minute-by-minute trading between thousands of
independent parties, these ratios would be the results of protracted negotiations between
governments. However, I don't think this cause of lower volatility would be a good thing,
as some degree of volatility in a market is necessary to effectively bring supply and
demand into some form of equillibrium...
I don't think we're very close to that scenario, as nation-states everywhere simply have
too much invested in the currency-model. However, precisely because that system is
brittle, it could break down very, very quickly... I just don't think this will happen in the
next few years. I think we have the ability to keep up the smoke and mirrors for some
time still, though at the expense of exacerbating the eventual reconciliation with reality.
Log in or register to leave a comment
memmel on February 5, 2009 - 1:28pm
Nice post. Pretty much the concept I've been thinking of.
However I think we may have another factor thats entering the equation as this system
follows a model similar to what your describing. This is that the overall system efficiency
will decline and it gets disrupted. Its almost a law of thermodynamics that a complex
system disrupted from its normal pathways becomes less efficient. We are seeing this
situation unfold right now. A example is for shipping as profits fall it becomes more
profitable to make as many trips as possible. I think this is intrinsic in declining EROEI
or very closely related as your marginal profit declines you attempt to grow by doing
more marginally profitable transactions leading to higher and higher energy usage.
Disruption of the system can be seen as and artificial lowering of the overall EROEI. I.e
as the system gets disrupted it tends to become less efficient lowering the EROEI and
thus driving it to try and expand the marginally useful transactions making it overall even
less efficient.
Economically this is seen as the statement we will sell at a loss and make it up in volume.
Even though this is a nonsense statement its true. Whats happening is the original system
was designed to be efficient at a certain volume i.e economies of scale as the volume falls
then the survival of the overall system is just as dependent on maintaining volume even it
it results in a loss vs outright profitability. This attempt to maintain the volume
component or economies of scale at all costs results in a significant drop in overall
system efficiency.
I think if you add this one concept to your model then you will see that the system is
certain to collapse if it embarks on the path of attempting the maintain the current status
quo as it vainly tries to keep system tuned to profit marginally from economies of scale.
This can already be seen at the large scale as we move away from a market economy to a
socialist then probably command economy whit the government taking over more and
more sectors resulting in massive waste in and attempt to avoid having sectors fail as
economies of scale no longer work. We are playing the game of selling at a loss and
making it up in volume.
Log in or register to leave a comment
710 on February 5, 2009 - 2:13pm
First, do not measure the price of oil against median income. In the future, "median
income" measured in dollars becomes just as meaningless as oil measured in dollars,
because you don't know if $300K will buy you an island, a house, or a cup of coffee.
Instead, measure it against the relative prices of various inputs (human labor, other
energy sources, infrastructure), and relative prices of various outputs (pesticides, gasoline,
asphalt). The system contains feedback as pesticides contribute to labor (through food),
gasoline contributes to energy sources, asphalt contributes to infrastructure.
Second, now that we are approaching a dynamical systems perspective (including
periodic, complex, and chaotic effects), "Chaos: Making a New Science", by James
Gleick, is an excellent primer. The 24-part "Chaos" lecture series by The Teaching
Company is also very good.
To answer question #3, the system we are living in will go through two phase- or statespace changes. The system we are in right now is a complex system, meaning it has some
amount of differential feedback, functional cohesion, and regularity.
Biologically, this type of complex system encompasses nearly all of normal living
functions, like eating, breathing, drinking, playing, screwing, etc.
The system we pass into at the first change is a chaotic system, which has dysfunction,
disintegration, and irregularity. This is what we are beginning now, increasing volatility
of the parameters we measure. Feedback mechanisms are also strained with unexpected
or mistimed inputs.
In a living system, this chaos would be represented by fatal onset of cardiac arrhythmia,
respiratory failure, or severe blood loss.
The system we pass into at the second change is the steady-state system. This is when the
system as we know it stops permanently. Only the pieces are left over at this point. This is
when a living thing dies.
Now, within the bounds of organic tolerance on Earth and with positive (solar) energy
flow, the pieces can re-assemble into a different complex system. This would technically
be the third state-space change, the world engendered by the survivors post-collapse.
In a chaotic system, asking questions #1 and #2 are like proposing that we will head
straight for the hurricane in our fishing trawler, and we want to predict which parts of the
boat get ripped apart, and in what order people drown.
The question to be asking is, "is it a good idea to go out on the water?"
"Yes, but if I don't fish, I can't eat!"
"If you go out in the storm, it will kill you today. If you don't go, you'll starve to death in
three to five weeks. You can use that time to figure out where else to get food."
Log in or register to leave a comment
jeffvail on February 5, 2009 - 3:06pm
I agree with you that measuring median income, in isolation, in dollars is meaningless
going forward. That wasn't my suggestion. What I did suggest (though it's probably my
fault for not being clear), the measure of oil price (in dollars) as a percentage of median
income (in dollars) is not meaningless, because it uses a common denominator to
establish how much oil an average person can buy. This will be a highly meaningful ratio
in our economy going forward, regardless of whether $300k will buy you a house, an
island, or a cup of coffee, precisely because it communicates relative buying power of
energy, which represents the fundamental ability to do work. As such, it will effectively
represent whether our development of alternative sources of energy is increasing or
decreasing our individual ability to do work.
I agree with you that chaos theory provides some insight into oil markets, but I'm not sure
there's any value in asking "is it a good idea to go out on the water?" What is the practical
equivalent of this question? "Is it a good idea to continue to use oil?" The answer to that
question is meaningless in the sense that it does not produce a realistic, implementable
course of action. It's more akin to realizing that we're already in the middle of the
hurricane, and asking "would it have been wiser to never have left shore last week in the
first place?"
I think the real lesson for us from chaos theory is that, given the incredible complexity
and non-linearity of our civilziational system, and the extreme dependence of any such
system on initial conditions, we have no realistic hope of understanding the long-term
outcome of present inputs. It's like asking the Venetians to correctly recognize the longterm impacts of the corporation (which, in a manner of speaking, they invented), and to
then decide whether it is a wise course of action. Our time-frame from action to result has
shrunk from centuries to years, but we're no more capable of guessing the correct
outcome. I do think, however, that by recognizing our inability we can arrive at a rational
course of action: consciously reduce the size and complexity of our societal and
economic networks. That, however, is a discussion for another time...
Log in or register to leave a comment
Gail the Actuary on February 5, 2009 - 2:39pm
Another issue is whether supply and demand curves are kinked because of large sunk
costs. This is an image from a presentation by Daniel Ahn.
Log in or register to leave a comment
Robert.Smart on February 5, 2009 - 2:43pm
Us computer folk can see that we need a buffer. Let's rename the SPR the "Tactical
Petroleum Reserve" and create a real Strategic Petroleum Reserve that is 10 times bigger
and that the government does actually fill when the price is low and empty when it is
high. By printing money to buy the oil, then unprinting money when it is sold the
government will also act against deflation/inflation respectively.
Log in or register to leave a comment
JonFreise on February 5, 2009 - 4:39pm
Hi Jeff,
Short term, I think volatility must decrease because the credit supply has been reduced.
Here is the mechanism:
A company, say airline, has a profit margin.
If the price of oil rises high enough, the profit margin is consumed.
A. Without credit, the airline would need to reduce purchases of fuel because it is
focusing on a smaller customer base that can afford the higher price to keep the company
in the black. The company's elasticity of demand has just changed from a very low oil
based elasticity of demand to a much higher airline customer based elasticity of demand.
B. With credit, the airline borrows money and maintains market share while operating in
the red. In this case the company's elasticity of demand is clearly much higher than in the
first case. It continues to pay ever higher prices while taking on more debt.
Companies have a strong incentive to take on debt because any that survive the oil price
spike will have reduced competition on the far side and hope to make up the earlier losses.
This is the same issue as a close stand of trees growing too tall for their root systems as
they race upwards competing for sunlight. They don't want to become top heavy, but the
other choice is stay short and starve.
Eventually, credit is exhausted. When that happens, the oil price debt driven bubble
bursts and all strategy B companies will be forced to become strategy A companies.
(Notice that no speculation was needed to drive up a price bubble. It would appear as all
commercial traders which was what we were seeing last fall.)
I feel credit acts as a volatility amplifier, but not in the speculative fashion most assume.
So I think that until there is a lot more credit available or a lot of debt is retired, that we
are unlikely to see such levels of volatility.
Does anyone else feel this mechanism has merit?
Log in or register to leave a comment