October 28, 2007
$90 a barrel: Is it time to start worrying about the oil price shock of 2007?
Oil shocks in 1973, 1979, and 1990 were each followed by a recession. But we saw the price of oil climb from $20 a barrel in 2002 to $75 a year ago, and so far it has not resulted in a significant economic downturn. What's different now, and can we count on it to continue?
The question of why the economy may be less vulnerable to oil price shocks today has been examined by a number of recent academic studies. A new paper by Olivier Blanchard and Jordi Gali identified four factors as all making a contribution:
(a) good luck (i.e. lack of concurrent adverse shocks), (b) smaller share of oil in production, (c) more flexible labor markets, and (d) improvements in monetary policy.
All the new analyses agree that the more flexible economy that we have now, allows us to cope more easily with oil price shocks.... It is time for America to get over its inordinate fear of oil shocks. Soaring prices aren't pretty, but they are scarcely the existential threat to our economy posited by many.
Munechika Katayama, a Ph.D. candidate here at the University of California, San Diego, also has an interesting new study on this question. Katayama notes that regulation of the transportation sector during the 1970s resulted in excessive monopoly power and inefficiencies in the trucking industry, and suggests that deregulation may have been another factor that has helped the economy to respond more flexibly to oil price changes.
Katayama builds a dynamic model of the dependence of intermediate-good deliveries on the transportation sector, and simulates the response of aggregate output to a 10% increase in the price of oil. The blue curve in the figure below gives the prediction from his model for the economy as it was configured in the mid-1970s, and implies that an oil shock of this size might reduce aggregate value added by more than 1%. The fuchsia curve is the simulation rerun with just a single change, namely, a greater degree of competition in the transportation sector, as would be implied by the observation that markups in trucking have fallen from 35% in 1977 to near zero today. The red line is the implied path if transportation still exhibited as much regulatory-supported monopoly as it did in 1977, but the share of energy in total output had changed by the amount observed since 1977.
When the two effects are combined together along with the observed decreased serial correlation of oil price changes, Katayama concludes that today's economy would respond significantly less than the economy of 1977 to a 10% increase in the price of oil.
By the way, Katayama is currently seeking a job, and I'd encourage any departments who want someone with strong research and communication skills to take a look at him.
One idea that all of the studies mentioned above have in common is the attribution of a significant part of the declining importance of oil shocks to the increased efficiency in energy use, a theme emphasized again this weekend by Mark Perry, Greg Mankiw, and Mike Moffatt. In almost all of the models with which I am familiar, the key parameter that matters for this is the ratio of the dollar value of energy purchases to the dollar value of GDP. Here's a rough graph of the value share of crude oil over the last 40 years:
Before we become completely relaxed about this magnitude, it's worth emphasizing that its behavior is not a long-run downward trend, but instead a U-shape for which we are currently on the way back up. There is a very simple economic reason for this. The price elasticity of oil demand is substantially less than one, meaning that when the real price of oil goes down (as it chronically did between 1981 and 1997), the share of energy expenditures in total GDP will also decline. This fact, rather than the pure efficiency gains that many researchers often think of first, is a very important reason for the declining share of oil in GDP over this period. As oil prices go back up, however, this is starting once again to become a bigger factor in the budgets of consumers and firms-- a 10% increase in the price of oil is harder for any of us individually to shrug off today than it was in 2003.
I would also note that although the domestic auto sector is less important for the American economy today than it was in the 1970s, it is still quite significant, particularly with the housing downturn subtracting 1% from real GDP growth each year. We can avoid a recession only as long as consumers and firms remain as confident as Taylor and VanDoren above, and everything outside of housing continues to do well.
Should we count on that as oil surges to $90 and beyond? I am not so sure.
Posted by James Hamilton at October 28, 2007 03:13 PMdigg this | reddit
Listed below are links to weblogs that reference $90 a barrel: Is it time to start worrying about the oil price shock of 2007?:
» Opening Bell: 10.29.07 from DealBreaker.com
O'Neal Out as Merrill Reels From Loss (WSJ) No press release just yet from Merrill, but nobody seems to have any doubt that today is the day of O'Neal's departure. The only real unanswered question, it seems, is how... [Read More]
Tracked on October 29, 2007 04:38 AM
What we have witnessed over the past year is not an oil price shock.
What we have witnessed is a "price discovery game" being played with the consuming nations by OPEC+Russia wherein these producers are looking for the price level which will compel reduced consumption. They have not yet found the level ... not at $90.
What prevents that price in dollars from ever being found is the "debasement game" being played by US central bank wherein increasingly worthless paper is offered to the producers for the tangible products.
The only event capable of trumping these games is an event which produces a sudden price discontinuity.
A US attack on Iran will NOT be such an event.
A FULL Iranian response to a US attack (use it or lose it) could be such an event.
Contemplate, if you will, CNN, MSNBC and FOX (and perhaps even CNBC) running continuous coverage on Iranian rocket strikes on ships in the gulf and industrial facilities up and down the western shore of the gulf.
$109 one day, $156 the next and $197 on the third day ... that would be an "oil price shock."
Let's see the Fed print its way out of that.
Posted by: esb at October 28, 2007 04:47 PM
Yes the US economy is less dependent on oil. Apparently.
In reality it is just as dependent on oil.
All that has happened is that manufacturing jobs have been shipped overseas, and now the oil is being consumed on our behalf by the nations we import finished goods from.
A true oil accounting would include the implied oil consumption of the imports we consume. As well as the extra oil consumed in transportation by the "global economy" that supports our total consumption.
Posted by: Daniel Dare at October 28, 2007 06:14 PM
I think that lack of regulation is the answer.
It's not only trucking. Airlines too. LOTS of big users of oil have a lot more flexibility to change their operations in the face of rising oil prices. That flexibility was lacking in the bad old days.
Also, what about price controls? They caused economic havoc back in the '70s. Gas lines were worse than higher prices. That was the lesson of the 1970s. I'd have to say that our experience with $90 oil bears that out.
Posted by: Buzzcut at October 29, 2007 07:32 AM
These factors contributed, but the other side of the coin is that in the 1970s oil was just one example of generalized inflation -- for example, in the 1970s unit labor costs were soaring at double digit levels. As a consequence of this real personal income and real profits growth turned sharply negative before and into the 1970s recessions. But now real income growth is running at 4% to 5% rates and outside of financial sector write-offs real profits growth is still positive. It is hard to see the US economy entering a recession as long as real income growth is this strong.
Posted by: spencer at October 29, 2007 08:31 AM
What continues to stun me in discussions by economists of this issue is the lack of recognition that the percentage price increase is not the proper metric of the size of an oil shock. Because oil shocks occur for geopolitical reasons (particular countries have wars, revolutions, etc), the natural size of oil shocks is set as a percentage of supply. Past oil shocks (eg in the 1970s) have been of O(5%) reduction in global supply. That's a serious reduction in the flow of energy through the economy which is bound to result in reduction in economic activity somewhere (for lack of energy with which to physically perform the activity in question).
By contrast, the current "oil shock" is instead a plateau in supply of oil, but not yet an actual reduction (at least not to any significant degree - see latest data here). Thus although oil is almost as expensive in inflation adjusted terms as during the 1980 shock , this shock is in fact significantly milder than that one (at present). That price is almost as high during a much milder shock reflects the fact that demand has become more inelastic since the 1970s.
If we were now, for some reason, to lose 5% of supply, oil would shoot up to somewhere in the general vicinity of $200. Then we would discover that the economy is no more resilient to oil shocks than it has ever been.
Posted by: Stuart Staniford at October 29, 2007 08:32 AM
The main reason there is minimal economic distress caused by the increase in prices this time is because the price increase has been caused by demand growth, not supply reduction.
Much in the same way that buying a new house because you want to live in a bigger one causes less distress than having to buy a new house because your last one burned down. Similar outcomes, different causes and, hence, different reactions.
Posted by: bartman at October 29, 2007 08:45 AM
Echoing what Stuart Staniford said.
I wanted to comment that I would also like to look at energy flows into the economy rather than dollar flows around the economy--though both are relevant considerations. He beat me to it. What I'd like to look at is the amount of energy used per unit of GDP to see how that has changed over time.
In addition, I have this concern that whatever efficiency gains we've accomplished since the late 1970s will make energy shortfalls all the more difficult. If we really are more energy efficient, then I'm concerned that energy shortfalls mean MORE impact on GDP, not less.
That said, I agree with the general argument of this post: we're more flexible than we were in the 1970s. But I'd also argue that we've not seen a steep drop in supply--yet. And that we have more disposable income than the 1970s, and a larger percentage of the population who are--at least right now--economically immunized to price increases because they have much more disposable income.
Posted by: General Specific at October 29, 2007 09:04 AM
My idea : it was not the increase in the oil price that produced the historic recessions, but the reduction in supply.
If true, this idea seems to be over-looked.
Posted by: George Barwood at October 29, 2007 10:26 AM
George Barwood wrote: "My idea : it was not the increase in the oil price that produced the historic recessions, but the reduction in supply."
JDH has touched upon that aspect in past posts so I'm not sure why he didn't mention it in this one.
Posted by: General Specific at October 29, 2007 10:36 AM
Thanks, GS. Post was getting a little long so I left out a number of relevant issues, which several of you have kindly been bringing up.
Posted by: JDH at October 29, 2007 10:50 AM
In addition, I have this concern that whatever efficiency gains we've accomplished since the late 1970s will make energy shortfalls all the more difficult.
Isn't that just the law of diminishing returns restated?
Posted by: Buzzcut at October 29, 2007 11:53 AM
Googling for charts of energy per unit GDP, I found:
It's not a primary source but it's got a picture anyway. U.S. energy intensity has fallen steadily since the 1970s.
I wonder why, if energy got cheaper 1980-2000, energy intensity fell? You'd think if energy is cheaper then people would use more of it and it would become a larger input into GDP.
Posted by: Hal at October 29, 2007 01:47 PM
As far as this idea that quantity of energy determines "shockiness" and production levels, I don't buy it. People make decisions on the basis of observable local information like price, not abstract macroeconomic aggregates like total quantity. Faced with a price increase for gasoline it doesn't matter whether it is because demand is up or supply is down. You've got to pay the same thing either way and you'll make the same decisions regarding tradeoffs.
Seeing energy as determinative of production levels is the same kind of mistake as the Marxist labor theory of value. Focusing on one factor of production ignores the larger picture.
Posted by: Hal at October 29, 2007 01:48 PM
Spencer, wrong. Real income growth always look positive to its not. Nor are real profits rising, they are falling.
Posted by: Sandman at October 29, 2007 07:20 PM
In addition, I have this concern that whatever efficiency gains we've accomplished since the late 1970s will make energy shortfalls all the more difficult.
Isn't that just the law of diminishing returns restated?
Yes, of course, but the actual facts matter.
Specifically, in the 1970's a non-trivial fraction of electrical generation was oil-powered, and a substantial amount of home-heating similar.
Today, virtually no electrical generation (by fraction of total produced) in developed countries
is oil powered, because of price.
This is because gas and coal generation offered their technical and economic advantages (coal, in much cheaper fuel with less flexibility, and gas, with somewhat cheaper fuel and greater flexibility and lower emissions).
The remaining petroleum consumption is primarily transportation, for which there are almost no alternatives--for ground transportation--and absolutely no alternatives---for air transportation.
Seeing energy as determinative of production levels is the same kind of mistake as the Marxist labor theory of value. Focusing on one factor of production ignores the larger picture
And suppose those laborers were literally starving to death, as in not obtaining sufficient joules of combustible hydrocarbons through their eosophagal fill pipe?
Certainly there is much more than the Marxist labor theory, if all the workers are well fed.
But suppose they aren't?
That is the equivalent problem with petroleum.
Posted by: Matthew Kennel at October 29, 2007 08:27 PM
Hal: Nice find.
I haven't thought this issue through very well, energy per GDP that is. It appears that the US--and China as well-are using less oil per unit of GDP (though China shows an upward trend in the last couple years). And the comment above regarding natural gas and coal fired power plants must be taken into account--we're acquiring our energy from other sources.
I just don't know the answer. If an economy uses energy more efficiently, and in particular oil energy, does that imply a larger quantity of GDP that is contingent or dependent on that energy use? In other words, do we consider energy use just another part of the economy, or the base of the economy? I tend to think the latter.
Anyway, nice find.
Posted by: General Specific at October 29, 2007 09:12 PM
The sectoral share of oil consumption seems important here. In the '70s a substantial fraction of U.S. oil was used in creating industrial process heat and electricity -- affecting a substantial fraction of the PPI in non-trivial ways. Only tiny amounts of petroleum are used for those purposes now -- most of this fraction has been replaced by natural gas inputs. Almost all petroleum in the U.S. is now used for transportation sector purposes (rather than residential, commercial or industrial sector). I am aware that the official numbers are in the vicinity of 2/3rds of oil going to the transportation sector, but this is a misaccounting which does not reflect the fact that most "industrial" usage is refinery usage, and that 85% of asphalt is used for roads.
Posted by: benamery21 at October 30, 2007 01:25 AM
From my comments on the Blanchard-Gali paper over at Economist's view: Seems to me that the sectoral shift of oil share of the economy is even more important than the decline in absolute share. That is, that the decline in oil price impact on PPI as compared to CPI is the big story. Yes, on a second order this would have to be due to decline real wage rigidity. But on a first order, fuel switching in all non-transportation sectors from the early 70's to mid 80's has significantly lowered the impact of oil price shocks on PPI. Not so?
Estimates which show only 2/3 of oil consumption in the transportation sector fail to account for the fact that most 'industrial' use is in the refining industry itself (thus included directly in transportation fuel cost), and that another substantial fraction is road asphalt (government). This means that most industry exposure to oil shocks is solely via transportation expenses (not electricity or process fuel). The airline industry was bailed out during this timeframe, and the trucking industry was largely able to pass thru costs but these were minimal for most industries. Consumers got shafted and extracted home equity debt to pay their gas cards off but had no market power to exact wage increases and little to reduce gas consumption.
Posted by: benamery21 at October 30, 2007 01:34 AM
I think everyone should take a look at graph on Mark Perry's blog titled: Energy Consumption per Dollar GDP. It helps explain why high oil prices have yet to effect the economy as much as one would think.
Posted by: macquechoux at October 30, 2007 07:12 AM
The point is that the causes of oil shocks (turmoil in specific countries) are denominated in supply terms. A revolution in Saudi Arabia would be a 8.5 million barrel/day event. An invasion of southern Iraq by Iran would be a 6mbd event. A civil war in Venezuela would be a 3mbd event. And so on. From the perspective of economics these are exogenous events - not caused by the normal smooth operation of markets, but rather by a failure of political/economic arrangments to adjust smoothly.
There is no natural "size" of an oil shock in price terms. How much price adjustment would a hypothetical revolution in Saudi Arabia cause? There is no direct answer. The only thing that is fairly fixed is the amount of supply that comes from Saudi Arabia.
How much price movement each of those things cause in the parts of the market that are operating normally is then basically set by the elasticity of demand at the time. The reason a very mild shock in supply has caused a large price movement in the last few years is because demand has gotten more inelastic since the 1970s. That should not inspire confidence about what will happen if we do actually experience a real oil shock at some point - rather it suggests that the price movement will be enormous.
Posted by: Stuart Staniford at October 30, 2007 08:52 AM
We also had the oil use per dollar of GDP graph here.
Posted by: JDH at October 30, 2007 09:17 AM
Mr. Benamery21 makes a critical point - petroleum is today mostly consumed in transport and much less is process and electrical uses. For example, in 1970, 35% of US electric production was oil-fired and today it is less than 2%.
While petroleum is a vital factor of production it is not the only one. Since the 70's several hundred million new workers have been added to the free world's labor supply with China and India joining the global economy. Low energy intensity labor can serve as a partial substitute for plateaued petroleum production.
Posted by: Joseph Somsel at October 30, 2007 09:24 AM
Stuart - You can define oil shocks as reductions in supply and look at how different economies respond to a given percent reduction. I question whether the recent historical experience (past 7 years) constitutes an oil shock by this definition, or frankly by any definition. The name "oil shock" should be reserved for sudden changes because that is what a "shock" implies.
I don't agree with the implicit premise that the most important thing about an oil shock is the percent reduction in supply, or the perhaps further implicit premise that an X% reduction in supply will cause an X% reduction (on some time scale) in GDP.
By definition, delta price and delta supply are in proportion to demand elasticity, so given a certain value of elasticity, one is as good as the other as the basis for any analysis. But elasticity changes over time and is different from country to country. Given global markets, during an oil shock all countries face the same price, so if they have different elasticities they will experience different supply reductions. This is another reason why price rather than supply would seem to be the more useful variable.
Posted by: Hal at October 30, 2007 10:51 AM
I don't think demand has gotten less elastic, I think it has taken a little while for people to decide this isn't temporary, and a lot of people still aren't sure. Look at UPS: unwilling to commit to the large numbers of hybrid trucks necessary to bring their price down (they bought 100 and stopped buying, complaining that their price was too high).
I think OPEC has found the price level at which substitution will take place, they just don't know it yet, because of this lag, as well as capex lag for PHEV's.
PHEV's will start to arrive in 2009 and 2010 in volume, and then we'll start to see real substitution.
Stuart, where would supply be if production hadn't plateaued lately? 3-4% higher? That's the magnitude of the "oil shock".
Posted by: Nick at October 30, 2007 11:41 AM
PHEV's will start to arrive in 2009 and 2010 in volume,
Really? Reference? My understanding is that the only company to make hybrids in any volume is Toyota (~150k Prius sold so far this year, simply an amazing number) and they're not doing a PHEV.
Because oil use is transportation related, it takes time for the fleet to turn over. If high gas prices are causing substitution, you should be able to see it when people trade in their SUVs for cars, and their big cars for smaller cars.
Anecdotal evidence shows that this is happening, however some of it is just the baby boomers moving out of their childrearing years and getting rid of SUVs and minivans for sedans.
Posted by: Buzzcut at October 30, 2007 12:19 PM
Chevy Volt. Now in GM's production pipeline. I want one SO BAD ...
Posted by: Avo at October 30, 2007 02:28 PM
US hybrid sales are on track for 350,000 in 2007, for about 2% of the 16M new light vehicles.
GM plans a Saturn PHEV in 2009, and the Volt in 2010, with initial production of 60,000/year and fast expansion.
Toyota has vowed that GM won't take the lead away from them. They were hoping for a li-ion Prius in model year 2008 or 2009, but now it looks like model year 2010 due to their getting cold feet relative to their outdated li-ion chemistry.
Posted by: Nick at October 30, 2007 03:12 PM
Hal - I somewhat agree on the semantics of "oil shock" but was accepting JDH's use of the term for discussion purposes.
Nick - I find this paper quite good evidence that short-run price elasticity of demand is a lot lower than it used to be.
I would agree that long term efficiency responses have started to get into the pipeline, but they have had no discernible impact yet.
Posted by: Stuart Staniford at October 30, 2007 03:38 PM
Read why at the link. It is time to take the bull by the horns, and force the petroleum exporters to burn the candle at both ends. The US economy can win this battle, and end the scourge once and for all.
Posted by: GK at October 30, 2007 04:58 PM
For what it's worth, from the financial dictionary online:
"Economic shock Events that impact the economy which originate from outside it. They are unexpected and unpredictable (e.g., Hurricane Andrew in 1991, the rise in oil prices by OPEC)."
Not very quantitative. But the price of oil has basically quadrupled in the past five years. One could consider it a price shock.
Going to the reference JDH provided above in which he discusses or debates Stuarts energy theory of economics (mentioning that we don't have water or labor theories of economics), I agree with his point that an economic theory, in general, should not take into account the external inputs because an economy can adapt around those external inputs in a way that makes it misleading to include them. An equine theory of economics that takes into account the needs of a society that uses horses for transporation, pulling plows, and the likes might have made sense at one time, but would look very silly now.
It seems that economics really is about measuring what is circulating around the economy--e.g. exchange value.
But. I still think one can, at a particular time, consider whether there are significant inputs into the economy--e.g. energy--that are are so fundamental that, even if not reflected by price as a large percentage of the economy--because the input is plentiful and cheap, as oil has been--still is considered fundamental. Describing that economy in terms of energy as an input--as has been attempted by some in the past--might provide a better predictor of future economic performance when energy supplies are less plentiful.
The traditional economist points at the energy percentage of GDP and says "see, it's not a problem if energy supplies tighten up" whereas the economist with energy as an input can point out the contribution of energy to the economy irrespective of price--while still taking into account efficiencies that might take place as prices rise.
I would work out the details on this theory but (a) Stuart Staniford would beat me to it (b) his theory would be right while mine would be riddled with errors and (c) I'm still busy working on my equine economic theory. Better late than never.
Posted by: General Specific at October 30, 2007 11:40 PM
Do you have specific references for the dates you cite? That website didn't say anything about when the Volt is going into production.
I don't think that the Volt is slated for production. My understanding is that GM is waiting for improved batteries.
I've never even heard of a Saturn hybrid other than the 2 that they have in production (and which are a stretch to even call a hybrid).
Posted by: Buzzcut at October 31, 2007 05:40 AM
All this discourse around energy intensity feels like "look! my building is growing taller ever year so it must means that I don't need the foundation anymore"
Posted by: Khebab at November 2, 2007 07:11 PM
None of you bloggers has mentioned "peak oil"! One dosen't have to be very observent to realize that cheap oil is getting more scarce every day and that eventually the demand is going exceed the supply. When that happens the market is going to cause the price of crude to sky rocket. This situation is going to cause hoarding, chaos, economic depression and war.
There is an easy answer to the question "why hasen't $90 oil affecetd the economy" is simple, refiners are still using oil that they bought at $65. The economy is going to choke big time when the refiners start processing $90 oil and passing on the extra cost on to consumers.
Posted by: Gale Whitaker at November 4, 2007 07:57 AM
Gale: The only reason "peak oil" hasn't been mentioned is that (a) people who participate on this site are familiar with the concept and (b) the point of the post wasn't "why is oil getting expensive" but "why hasn't the economy choked on $90 oil."
Posted by: General Specific at November 4, 2007 08:21 AM