In a follow-up on my earlier post, I’d now like to discuss the second part of my paper, Causes and Consequences of the Oil Shock of 2007-08, which I presented today at a conference at the Brookings Institution. Here I’ll review the role that the oil price shock may have played in causing the economic recession that began in 2007:Q4.
My paper uses a number of different models that had been fit to earlier historical episodes to see what they imply about the contribution that the oil shock of 2007-08 might have made to real GDP growth over the last year. The approaches surveyed include Edelstein and Kilian (2007), who examined the detailed response of various components of consumer spending,
Blanchard and Gali (2007), who studied the extent to which the contribution of oil shocks has significantly decreased over time, my 2003 paper, which emphasized the role of nonlinearities, and a model-free data summary of the observed behavior of different economic magnitudes following this and previous oil shocks. Although the approaches are quite different, they all support a common conclusion: had there been no increase in oil prices between 2007:Q3 and 2008:Q2, the U.S. economy would not have been in a recession over the period 2007:Q4 through 2008:Q3.
One of the most interesting calculations for me was to look at the implications of my 2003 model. I used those historically estimated parameters to find the answer to the following conditional forecasting equation. Suppose you knew in 2007:Q3 what GDP had been doing up through that date and could know in advance what was about to happen to the price of oil. What path would you have then predicted the economy to follow for 2007:Q4 through 2008:Q4?
The answer is given in the diagram below. The green dotted line is the forecast if we ignored the information about oil prices, while the red dashed line is the forecast conditional on the huge run-up in oil prices that subsequently occurred. The black line is the actual observed path for real GDP. Somewhat astonishingly, that model would have predicted the course of GDP over 2008 pretty accurately and would attribute a substantial fraction of the significant drop in 2008:Q4 real GDP to the oil price increases.
The implication that almost all of the downturn of 2008 could be attributed to the oil shock is a stronger conclusion than emerged from any of the other models surveyed in my Brookings paper, and is a conclusion that I don’t fully believe myself. Unquestionably there were other very important shocks hitting the economy in 2007-08, first among which would be the problems in the housing sector. But housing had already been subtracting 0.94% from the average annual GDP growth rate over 2006:Q4-2007:Q3, when the economy did not appear to be in a recession. And housing subtracted only 0.89% over 2007:Q4-2008:Q3, when we now say that the economy was in recession. Something in addition to housing began to drag the economy down over the later period, and all the calculations in the paper support the conclusion that oil prices were an important factor in turning that slowdown into a recession.
It is interesting also that the observed dynamics over 2007:Q4-2008:Q4 are similar to those associated with earlier oil shocks and recessions. The biggest drops in GDP come significantly after the oil price shock itself. What we saw in earlier episodes was that the drops in spending caused by the oil price increases resulted in lost incomes and jobs in affected sectors, with those losses then magnifying other stresses on the economy and producing a multiplier dynamic that gathered force over subsequent quarters. The mortgage delinquencies and financial turmoil in the current episode are of course not the specific stresses that operated in earlier downturns, but the broad features of that multiplier process are surprisingly similar to the historical pattern.
My paper concludes:
Eventually, the declines in income and house prices set mortgage delinquency rates beyond a threshold at which the overall solvency of the financial system itself came to be questioned, and the modest recession of 2007:Q4-2008:Q3 turned into a ferocious downturn in 2008:Q4. Whether we would have avoided those events had the economy not gone into recession, or instead would have merely postponed them, is a matter of conjecture. Regardless of how we answer that question, the evidence to me is persuasive that, had there been no oil shock, we would have described the U.S. economy in 2007:Q4-2008:Q3 as growing slowly, but not in a recession.
thanks for this article. intuitively this makes sense. at a micro level, it makes sense that an overly indebted subprime mortgagee (living in a house he couldn’t afford, and living a fair distance from work) would tend to also be driving a car he couldn’t afford. an additional unexpected expense for gasoline would be the trigger for not being able to make debt payments.
If the downturn was largely due to high oil prices, should not have things rebounded once oil became cheap?
But it did not. At least not in the first 4 months of cheap oil.
So this thesis is invalid.
“If the downturn was largely due to high oil prices, should not have things rebounded once oil became cheap?
But it did not. At least not in the first 4 months of cheap oil.
So this thesis is invalid.”
Do you know what a vicious cycle is?
GK: Please consult my 2003 paper and New Palgrave entry for discussion of whether oil price decreases have a parallel positive effect to the contractionary influence of an oil price increase. In particular, note that the equation from which the above figure is generated does not have the implication that you assert.
GK: there is likely path dependence, as there probably is with credit conditions.
JDH: really great post!
James Rubin, then Chief Economist of CIBC, highlighted the importance of oil in the current recession back in late 2008. His analysis can be found on the Oil Drum.
In any event, if we believe this analysis, then the #1 goal of US (and global) energy policy must be to prevent a return to the oil pricing environment of mid-2008. That, in turn, implies a managed price regime.
Is there similar analysis of the early 80’s recession and the 79 oil shock?
(I learned, I think, that there was consensus that the 80’s recession was purposely created by the Fed to bring inflation under control.)
Thanks.
The Professor is right, the oil bubble blew up the economy, which was fueled with money borrowed from our homes. Be thankful, it was better to burst early then to have gone another lap. The reason the economy will not be back in gear, even with the lower fuel costs, is due in large part to the American middle class getting the message! We are currently paying down consumer debt, which is driving the savings rate up and the consumer debt bubble down. With, the coincidental tightening of credit by the banks making it hard to borrow, as well, additional money into these institutions via the “new” savings will build and eventually the credit will have to loosen up. The nice part of the almost forced debt repayment/savings bubble being created is that we (U.S.) may not need to borrow so much capital from the world for all the reckless spending going on. While Washington and the Fed are flooding the market place with cash, people are using it to pay down “personal debt” as fast they can. For those that do not like Obama, as I do not, it is going to be hard because we (U.S) will roar back in late spring, 2010. Unemployment will be high into 2012, but coming down enough to re-elect Obama. Whether we like it or not, we are being forced to accept a new, socialized world economy.
As an aside, go UNC! LOL
I agree that the gas price bubble was the wake-up call for the American consumer. He/she finally realized that he/she was in way over his/her head, and cut back substantially on spending, starting with travel. The rest, as they say, is history.
I love reading evidence that supports my own previously held conclusions. Without a recession, I think the mortgage problem would have gone unnoticed a little bit longer, of course. The underlying credit risk would still be there but, it took a recession and consequent slide in home values to expose the amount of risk in the system.
Thanks Chris and Barney. I also think the current amount of money being thrown at the affirmative action problem is absolute craziness. Thanks to mark to market the current values of those assets are now vastly understated where before they were overstated by not accounting for the level of risk. Didn’t I just hear that the rule was suspended? I would like to see how that plays out before throwing another trillion or two to the banking system. But then, imo, most of the crap being done to solve the problem has little to no basis in economic facts and very much to do with political agendas.
enjoyed your piece, even through a quick read, but i think the tie of gdp to oil may be a bit spurious. the cause of the recession begins with global capital flows. when the fed eased beginning late 07 it was a monetary boost to china and all the other dollar-pegged nations that were already growing quite rapidly. no surprise oil prices accelerated higher. well i think you can fill in the rest.
JDH wrote:
“…the evidence to me is persuasive that, had there been no oil shock, we would have described the U.S. economy in 2007:Q4-2008:Q3 as growing slowly, but not in a recession.”
This actually begs the question. Why was there an oil shock?
It would be perfectly reasonable for a politician to take this statement to support his solution of price controls on oil.
This actually begs the question. Why was there an oil shock?
It would be perfectly reasonable for a politician to take this statement to support his solution of price controls on oil.
Perhaps with the huge losses Wall Street took on speculating with long term securities in short term funds, they had to create a buble to try and recoup losses! Of course, I’m only speculating!
I am not familiar with these models. Is it a real shock we are talking about? That is, do the models control for the fact that a depreciation of the dollar due to an increase in the money supply would result in an increase in oil prices? Just trying to figure out whether this rules out the fact that part of this shock could be due to monetary policy shocks.
Steve said: “…The reason the economy will not be back in gear, even with the lower fuel costs, is due in large part to the American middle class getting the message!…”
I think Steve has it right. A rational actor (in the economic sense) would respond “symmetrically”, i.e., cutting back on spending with higher oil prices and increasing spending with lower oil prices. However, the oil/gasoline price spikes last year put the fear of God into everyone.:)
Sebastian
p.s. Go UNC.:)
couldn’t a monetarist substitute “the price of money” for oil? eventually the MR gained with oil at some level is no longer larger than the MC.
seems like this happens in every economic bubble and that the price of oil was a result of too much leverage everywhere
GK, the price of oil rose from $30 in 2003 to $50 by late 2004 and $60 by 2005, but the economy did not start slowing until late 2007. So yes, the price of oil is now down, but it will take time for the positive effects to impact the economy.
I found the paper on causes quite implausible (as I said in a comment there), but the 2003 black box generated consequences that seem very plausible to me. Energy is a big chunk of family budgets and when it doubles from 2.5 to 5 % of GDP something has to give–and pretty quickly.
Using the Nov endpoint of the graph the percent decline from trend in real GDP is about 0.63%. In 4th Q 2008 GDP was falling at about 3.8% annual rate. So, what was elasticity of GDP with respect to oli price change and oil share GDP?
“Do you know what a vicious cycle is?”
Do you know what a virtuous cycle is?
“So yes, the price of oil is now down, but it will take time for the positive effects to impact the economy.”
The impact to consumer spending is almost immediate (given that tanks have to be filled once a week, and shipments to retail stores are weekly, daily in the case of perishable produce).
“Energy is a big chunk of family budgets and when it doubles from 2.5 to 5 % of GDP something has to give–and pretty quickly. ”
And when it shrinks back to the lower size relative to GDP, the reverse happens quickly.
JDH,
I scanned over your paper. I would argue that oil was above $100 for only a few months, and it has been below $60 for a longer period than that already. Certain things are not linearly corelated as the demand curve is not linear, but general inflation/deflation effect is equal in both directions.
If one examines the pyschological impact, it is even more in favor of a drop being stimulative. Until 2004, $52/barrel was considered very high. But today, no one considers that to be troublesome, simply because $147 was briefly hit before.
James, I calculate a projected GDP of 12,210B, and a counter-factual GDP of 11,533. This gives an oil price impact of -5.9%.
First – is this correct?
2nd – what does the model project for 1 year from now, say?
GK: I agree consumer spending on non-energy goods should pick up quickly in response to retail price declines. A large majority of people have been spending every dime of income. Two possible reasons for a slow consumer response to declining oil prices: high oil prices were not the only important contributor to the recession (I would certainly agree with that) and/or non-linear effects–one of which I will suggest.
If I want my stores fully stocked for the holiday selling season, the goods need to ship from China no later than October 15, and I must have ordered them weeks earlier, usually no later than August. That means the latest numbers I will have on which to project demand will be for July. If the goods aren’t in the stores, they won’t be sold and that’s reflected in lower GDP.
The U.S. economy is operating in a world in which we are competing with Germany and Japan and a few other nations for a way to participate in world trade with a high pay labor force. We must sell some goods on the international market.
Our major competitor – Germany – has already adjusted its economy to high priced gasoline.
WE must figure out someway to move toward equal trade AND move toward an economy that can operate with high priced gasoline.
That is the problem that Obama will be forced to solve – unless his dream of some alternative to gasoline can be found.
I realize Germany GDP is down more than U.S. because their exports are down. Nevertheless, Germany is better positioned to compete in the global economy.
If the relationship you find between oil and gdp is correct, then should we expect a series of severe boom and busts until we solve the energy problem?
It seems with all the production cutbacks in oil producing states that it will be even easier for quantity demanded to quickly exceed quantity supplied as the global economy recovers.
Expectations of higher oil prices are already beginning to cause the price of oil to bubble higher in the futures market even though we are ‘swimming in oil’ right now in the spot market.
Expectations of higher oil prices shift the demand curve to the right and the supply curve to the left, no? I think it will take a prolonged L shaped recession to purge the expectation of a quick rebound in oil prices from the minds of producers and consumers.
JDH, congratulations on teasing apart the impact of oil prices and mortgage stress inside a complex feedback system that we call the economy. Very nice. I particularly like your Chart 3 in the paper which shows how consumer oil expenses as a percent of GDP “came apart” from the trend channel in 2005 and beyond. Maybe we should call this the SUV/Hummer recession.
GK: One possible explanation for low oil prices not being reflected yet in demand or GDP in my view would be the statements of our dear leader that he wants to bankrupt coal using electricity generators and see electricity rates triple and his view on the need for cap and trade to curb those noxious gases polluting our planet. What will be the impact of that? Inquiring minds want to know and we will not spend until we do.
GK: I think politics is what is preventing a quick recovery. If one attempts to simplify (a very difficult proposition) the voters into the producers and the dependents, I would say the producers are scared to death and the dependents that mostly contributed to our current political regime (and in a state of euphoria) don’t have any money to spend.
Housing and transportation have a moderate correlation in the consumer’s mind. The home being an inventory center and the car the chief freight handler. The demand for housing, at the time being mainly tract housing, would have accelerated the use of the car for freight. Oil, being the least elastic, shot up the fastest, faster than housing. In the suburbia model, tract housing and automobile go together.
What was bubbling housing? Mainly technology, and secondarily some further constraint. Internet technology, in the house, removes 1/3 of the need for automobile freight handling in terms of locating cheap goods. About 10-15% of the housing bubble was justified by the broadband internet, probably another 10% was justified in waiting for technology to solve the next constraint in the household.
So, JDH would normally say oil shocks result in a period of innovation in transportation. But this time the consumer killed the car and kept the house, like a divorce. The consumer did that because his house had just enough technology to displace about 1/3 of his monthly fixed car expenses.
Hamilton is actually a genius. “It’s the oil stupid”. He has been saying this for decades, and people keep forgetting.
I even lost my faith a bit when oil prices were high and there wasn’t a ‘recession’. It turns out there was.
I was at a lecture in Calgary titled “Oil Price Uncertainty” by Dr. Apostolos Serletis. In it, he examined the effects of an oil price shock. Both a positive and negative shock. I believe his conclusion about a negative oil price shock was that GDP fell as well.
” would say the producers are scared to death and the dependents that mostly contributed to our current political regime (and in a state of euphoria) don’t have any money to spend.”
That is certainly true.
US GDP per capita is currently $45,000. But it has not risen in real terms in almost a decade.
I think that once a society reaches a certain level of prosperity, leftists become a large enough percentage of the population that they enact policies that prevent further economic growth. The ratio of leftists to productive people becomes high enough to paralyze any prospects of economic growth.
The 60-year-old women at Code Pink, in yesteryears, would have had to work in the fields, factories, or as housekeepers. They would have no luxury of causing trouble while on the taxpayer dole.
It might not be possible for any large country (more than 10m people) to ever cross $60,000 a year in per-capita GDP in 2009 dollars. That might be a self-inflicted ceiling that human societies impose on themselves.
GK – Could be that consumers do not need/want all the stuff that would be available at a higher income/GDP level.
More stuff is not the only value available. Personally, I would give up some of my current stuff for an economy operating smoothly, without stress -(reliant on oil, large trade deficit, high debt, reliant upon extensive credit).
The market price of oil does not impact the consumer as much as the price at the pump.
If one looks at US oil consumption, it dropped by over 20% after the implementation of the Carter energy policies. Reagan stalled efficiency efforts and consumption returned to 1978 levels by 2000. In 2000, pump prices first went over $2 in the Midwest. Fuel efficiency needs to get back on track.
Families have a “Transportation” budget, not a “fuel” budget. Increases in fuel costs translates into fewer car purchases. We have seen US car sales go from over 17 million to under 10 million. Some of that is people finding out how to live with fewer cars and liking the savings they realize. A typical car costs about $7,000 per year. A wage worker at $8 per hour earns only $16,000 per year. GO figure.
http://www.eia.doe.gov/emeu/aer/pdf/pages/sec5_20.pdf
Mark: The paper has an extensive analysis of four recessions that followed four earlier oil shocks prior to the most recent experience. To that list could also be added the Suez Crisis of 1956-57, and in fact there is some upward move in oil prices prior to every postwar recession with the single exception of 1960. I’ve been writing about this for many, many years.
wogie: Note that the graph is already in units of 100 times the log, so the actual change (not the percentage change) is already in units of approximate 100 basis points. That is, the correct calculation is 940.98 – 935.91 = 5.07, meaning that the oil shock is estimated to have reduced the 2008:Q4 level of real GDP by about 5.07% relative to what it would have been without the oil shock. Note also that the green dotted line is not a “trend line”, but instead is the forecast from a univariate fourth-order autoregression estimated over 1949:Q2 to 2001:Q3. The last two quarters of GDP before this forecast begins (2007:Q2 and 2007:Q3) were both above-average growth rates (4.8% at an annual rate), and the nature of an autoregression is that, in the absence of other disturbances, you would have expected the next few quarters (2007:Q4 through 2008:Q2) to have also come in a bit above average. For this reason, the green line predicts that real GDP would have increased by a total of 4.88% between 2007:Q3 and 2008:Q4 in the absence of an oil shock, whereas the red line predicts that real GDP would have decreased by a total of -0.19% with the oil shock, for a net difference of 5.07%. Actual real GDP fell by -0.87% between 2007:Q3 and 2008:Q4. All percentage changes reported here represent 100 times the log difference.
Nick G: See details given to wogie above. For what they’re worth, the forecasts from here out from equation (3.8) call for GDP growth rates (annual rate) as follows: 2009:Q1 -2.5%; 2009:Q2 -0.8%; 2009:Q3 +3.9%.
tj: You could argue that’s part of what did happen in the 1970s.
A telling thing on that graph is how the “actual” line starts diving below the “expected” trend around alte August. This is around the time all the crap started hitting the fan on how screwed up the financial world really was, and a minor recession became something muchn worse.
While the oil shock led to a change in behavior that slowed and eventually retracted the economy (more than real GDP figures showed at the time, since real people had inflation well above the 4% that was being used as a deflator at the time), it was the realization of the financial meltdown and the collapse in asset prices that forced the real changes and cutbacks that we’re still suffering from.
Regardless, it’s good info, Dr. Hamilton.
While the high energy prices may have helped trigger the economic contraction a readjustment of economic activity was inevitable. No matter how we look at it, we had a phony economy that was based on borrowing and spending. That had to come to a bad end no matter what happened with oil prices.
Looking forward we should be able to note some vulnerabilities that seem to be ignored by most analysts and commentators. The bottom line is that the price declines that come with inventory liquidations are bound to end sooner than most people expect because the supply side response has been swift as factories have closed down operations and commodity producers have acted just as rapidly to shut down marginal operations. In the case of oil that means that supply is in rapid decline and that without new capital investment even a depression will be unable to keep prices from rising eventually.
At the same time the risk to savers is very large as central banks have threatened the purchasing power of savings by creating money out of thin air. This increases the risks of an inflationary depression in the United States and Western Europe that is similar to national depressions seen in many Latin American countries during the past few decades.
Finally, I have the opportunity to provide a reminder to the world that oil (and its price) can not be taken for granted.
I haven’t read the paper yet, but the price shock seems to have been the result of 3 factors.
Fundamentals, where world crude oil supply fell in 2006 and 2007 as demand skyrocketed
Declines in the value of the dollar over much the period 2005-2008:Q2
A speculative premium after 2007:Q4
If oil prices had followed the trend since 2003, I figured the price at about $110-115/barrel. So, I chalk up prices higher than that to #2 and #3.
To the extent that the financial crisis itself caused some (but not nearly all) of the price shock, it is hard to separate causes and effects.
There will be no rebound of the economy stemming from low oil prices. Oil production is now declining worldwide, the recession just covers that up because currently the consumption declines even faster.
But look at what happened after the G-20 meeting. Stock markets went up about 5% in a bit of optimism but that immediately triggered a 10% rise in oil prices.
If the economy recovers enough to actually try some growth again, demand will outrun supply again very fast and the oil price, inelastic as it is as soon as you get a sellers market, will shoot through the roof once more, triggering the next recession cycle.
Past experience is the only guide we have to the future. At the same time, I am with the “look-at-what-is-different-today” school.
“the price declines that come with inventory liquidations are bound to end sooner than most people expect because the supply side response has been swift as factories have closed down operations and commodity producers have acted just as rapidly to shut down marginal operations. In the case of oil that means that supply is in rapid decline and that without new capital investment even a depression will be unable to keep prices from rising eventually”.
The rapid foward looking response of many corporations (to reduce production and employment) suggests to me that we will see an upward growth in corporate profits soon and an uptick in the stock market before corporate profits begin to increase.
This change will be short lived because the low employment will not sustain economic growth.
To compete in the global economy, the U.S. economy must be restructured to be able to produce profits when gasoline prices are high – which all parties refuse to acknowledge.
I have now had a chance to look at this paper.
It needs pointing out at this juncture that I can find nothing in Obama’s energy policies–I have looked–that shows any level of concern about the low long-run price elasticity of demand for oil. Magical efficiency improvements from revised CAFE standards seem to be all that is required. JDH has been tracking the terrible car sales data, and we might expect that to continue for some years. Obviously this will weaken the effectiveness of CAFE improvements.
I have written two articles on all this.
Steven Chu’s Energy Miscalculations
The Secretary of Synthetic Biology