THE ROLE OF INVENTORIES AND SPECULATIVE TRADING IN THE GLOBAL MARKET FOR CRUDE OIL

There is a consensus in the literature that the short-run price elasticity of oil supply is close to zero, if not effectively zero.

“There is a consensus in the literature that this shortrun price elasticity of oil supply is close to zero, if not effectively zero.”

• cites [14] 14. Davis LW, Kilian L. 2011. Estimating the effect of a gasoline tax on carbon emissions. Journal of Applied Econo- metrics 16: 1187-1214. sourceCould not verify

  Section 4.1 (Price Specification; page unavailable) discusses a flat short-run gasoline supply curve as a modeling assumption. The passages found when searching for short-run oil supply elasticity and a literature consensus do not address either claim about oil supply. (LLM, conf=0.77)

  Experimental — accuracy not yet measured

The paper's estimates rule out speculative trading as an explanation of the 2003-2008 oil price surge.

“Our estimates rule out explanations of the 2003-2008 oil price surge based on unexpectedly diminishing oil supplies and based on speculative trading.”

• Also stated in: § 1. INTRODUCTION · plain

  “our estimates rule out speculation as a cause of the surge in the real price of oil between 2003 and mid 2008.”

• Also stated in: § 1. INTRODUCTION · plain

  “all speculative trading in the spot market for crude oil combined had little effect on the real price of oil between 2003 and mid 2008”

Speculative trading played a central role in 1990, following Iraq's invasion of Kuwait.

“It also played a central role in 1990, following Iraq's invasion of Kuwait.”

Even after accounting for the role of inventories in smoothing oil consumption, the paper's estimate of the short-run price elasticity of oil demand is much higher than traditional estimates from dynamic models that do not account for the endogeneity of the price of oil.

“We also show that, even after accounting for the role of inventories in smoothing oil consumption, our estimate of the short-run price elasticity of oil demand is much higher than traditional estimates from dynamic models that do not account for for the endogeneity of the price of oil.”

Speculative demand shifts played an important role during earlier oil price shock episodes including 1979, 1986 and 1990.

“There is evidence, however, that speculative demand shifts played an important role during earlier oil price shock episodes including 1979, 1986 and 1990.”

• Also stated in: § 1. INTRODUCTION · plain

  “although speculative trading does not explain the recent surge in the real price of oil, we show that it played an important role in several earlier oil price shock episodes.”

Speculative trading helps explain the sharp decline in the real price of oil in early 1986 after the collapse of OPEC.

“it helps explain the sharp decline in the real price of oil in early 1986 after the collapse of OPEC.”

There is little probability mass on values close to zero for the short-run price elasticity of oil demand.

“Our posterior median estimate of the short-run price elasticity of oil demand of À0.26 is four times higher than standard estimates in the literature and there is little probability mass on values close to zero.”

The paper's estimates rule out explanations of the 2003-2008 oil price surge based on unexpectedly diminishing oil supplies.

“Our estimates rule out explanations of the 2003-2008 oil price surge based on unexpectedly diminishing oil supplies and based on speculative trading.”

The recovery of the real price of oil starting in 1999 was aided by coordinated supply cuts.

“there is evidence that the recovery of the real price of oil starting in 1999, following an all-time low in post-war history, was aided by coordinated supply cuts.”

Flow supply shocks played only a minor role in 1979.

“In contrast, flow supply shocks played only a minor role in 1979.”

Preliminary tests provided no evidence of cointegration between oil production and oil inventories.

“Preliminary tests provided no evidence of cointegration between oil production and oil inventories.”

The posterior median estimate of the short-run price elasticity of oil demand is four times higher than standard estimates in the literature.

“Our posterior median estimate of the short-run price elasticity of oil demand of À0.26 is four times higher than standard estimates in the literature and there is little probability mass on values close to zero.”

Even after controlling for the effect of other shocks on crude oil inventories, there is no evidence of rising speculative demand after 2003.

“Our model suggests that even after controlling for the effect of other shocks on crude oil inventories there is no evidence of rising speculative demand after 2003.”

The posterior median estimate of the short-run price elasticity of oil demand is -0.26.

“Our posterior median estimate of the short-run price elasticity of oil demand of À0.26 is four times higher than standard estimates in the literature and there is little probability mass on values close to zero.”

Speculative trading was a central feature of the oil price surge of 1979, following the Iranian Revolution.

“it was a central feature of the oil price surge of 1979, following the Iranian Revolution, consistent with the narrative evidence in Barsky and Kilian (2002)”

• cites Barsky and Kilian (2002)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

A flow supply disruption causes inventories to be drawn down in an effort to smooth production of refined products.

“A flow supply disruption causes inventories to be drawn down in an effort to smooth production of refined products.”

The increase in speculative demand in 1990 was consistent with a sharp increase in oil inventories in the months leading up to the invasion.

“This result is consistent with a sharp increase in oil inventories in the months leading up to the invasion.”

The small supply-side-driven increase in the real price of oil is dwarfed by the price increases associated with flow demand shocks.

“What evidence there is of a small supply-side-driven increase in the real price of oil is dwarfed by the price increases associated with flow demand shocks.”

A positive speculative demand shock causes a persistent increase in oil inventories.

“A positive speculative demand shock causes a persistent increase in oil inventories.”

The resurgence of the real price of oil starting in early 2009 reflected primarily increased flow demand.

“the resurgence of the real price of oil starting in early 2009 reflected primarily increased flow demand (see Figure 2).”

In the short run, 29% of the variation in crude oil inventories is driven by speculative demand shocks.

“It can be shown that in the short run 29% of the variation in crude oil inventories is driven by speculative demand shocks, followed by oil supply shocks with 26%.”

Reduced-form VAR models based on the paper's oil inventory proxy generate more accurate real-time out-of-sample forecasts of the real price of oil than other methods even during 2009-2011.

“Baumeister and Kilian (2012a) show that reduced-form VAR models based on our oil inventory proxy generate more accurate real-time out-of-sample forecasts of the real price of oil than other methods even during 2009-2011, further strengthening our claim that the oil inventory proxy is reasonably accurate.”

• cites Baumeister and Kilian (2012a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Figure 2 shows no evidence that speculative demand shocks explain the bulk of the surge in the real price of oil after 2003.

“Figure 2 shows that there is no evidence to support this view.”

In the short run, flow demand shocks have a negligible impact on crude oil inventories, explaining 2% of variation.

“Flow demand shocks have a negligible impact with 2%.”

Doubling the bound on the impact price elasticity of oil supply has little effect on the shape of the posterior distribution of the impulse responses.

“It can be shown that doubling this bound, while increasing the number of admissible models, has little effect on the shape of the posterior distribution of the impulse responses.”

There has been no systematic upward movement in the real price of oil after 2003 associated with speculative demand shocks.

“There has been no systematic upward movement in the real price of oil after 2003 associated with speculative demand shocks.”

In the long run, 87% of variation in the real price of oil is attributable to flow demand shocks.

“For example, in the long run, 87% of the variation in the real price of oil can be attributed to flow demand shocks, compared with 9% due to speculative demand shocks and 3% due to flow supply shocks.”

The estimates of the posterior median price elasticity of oil demand are remarkably robust to changing the bound on the impact price elasticity of oil supply.

“the estimates of the posterior median price elasticity of oil demand reported in Section 5 are remarkably robust to this change.”

There was a substantial increase in speculative demand already in the months leading up to the invasion of Kuwait.

“The bottom panel of Figure 2 reveals there was in fact a substantial increase in speculative demand already in the months leading up to the invasion.”

In the long run, 9% of variation in the real price of oil is attributable to speculative demand shocks.

“For example, in the long run, 87% of the variation in the real price of oil can be attributed to flow demand shocks, compared with 9% due to speculative demand shocks and 3% due to flow supply shocks.”

The real price of oil peaks after three months following a negative flow supply shock.

“It peaks after three months.”

The oil futures spread does not Granger cause the remaining model variables at conventional significance levels for maturities of 1, 3, 6, 9 and 12 months.

“We formally tested this proposition and were unable to reject the null of no Granger causality at conventional significance levels for maturities of 1, 3, 6, 9 and 12 months, consistent with the view that the inventory data are as accurate as the oil futures spread.”

• cites Giannone and Reichlin, 2006 sourceNot checked — no full text of the source

The real price of oil rises only temporarily after a negative flow supply shock.

“The real price of oil rises only temporarily.”

Figure 2 provides no indication that negative flow supply shocks were an important determinant of the real price of oil between 2003 and mid 2008.

“Figure 2 provides no indication that negative flow supply shocks were an important determinant of the real price of oil between 2003 and mid 2008.”

Oil production rises somewhat but only temporarily after a positive flow demand shock.

“Oil production also rises somewhat, but only temporarily.”

A positive flow demand shock causes a persistent hump-shaped increase in the real price of oil with a peak after one year.

“It causes a persistent hump-shaped increase in the real price of oil with a peak after one year.”

After one year, oil inventories reach a level in excess of their starting level following a positive flow demand shock.

“After one year, oil inventories reach a level in excess of their starting level.”

The decline in the real price of oil starting in November 1990 was almost entirely caused by a reduction in speculative demand rather than increased oil supplies.

“Figure 3 also highlights that the decline in the real price of oil starting in November of 1990, when the threat of Saudi oil fields being captured by Iraq had been removed by the presence of US troops, was almost entirely caused by a reduction in speculative demand rather than increased oil supplies.”

At long horizons, flow demand shocks explain 15% of variation in crude oil inventories.

“At long horizons, in contrast, the explanatory power of speculative demand shocks declines to 27% and that of flow supply shocks to 24%, while the explanatory power of flow demand shocks increases to 15%.”

A positive shock to the flow demand for crude oil is associated with a persistent increase in global real activity.

“A positive shock to the flow demand for crude oil, in contrast, is associated with a persistent increase in global real activity.”

At long horizons, speculative demand shocks explain 27% of variation in crude oil inventories.

“At long horizons, in contrast, the explanatory power of speculative demand shocks declines to 27% and that of flow supply shocks to 24%, while the explanatory power of flow demand shocks increases to 15%.”

In the short run, 26% of the variation in crude oil inventories is driven by oil supply shocks.

“It can be shown that in the short run 29% of the variation in crude oil inventories is driven by speculative demand shocks, followed by oil supply shocks with 26%.”

About one third of the price increase from July to August 1990 was caused by speculative demand shocks and two thirds by flow supply shocks.

“It shows that about one third of the price increase from July to August of 1990 was caused by speculative demand shocks and two thirds by flow supply shocks.”

The effects of a positive speculative demand shock on global real activity and global oil production are largely negative but small.

“The effects on global real activity and global oil production are largely negative, but small.”

In the long run, 3% of variation in the real price of oil is attributable to flow supply shocks.

“For example, in the long run, 87% of the variation in the real price of oil can be attributed to flow demand shocks, compared with 9% due to speculative demand shocks and 3% due to flow supply shocks.”

At long horizons, flow supply shocks explain 24% of variation in crude oil inventories.

“At long horizons, in contrast, the explanatory power of speculative demand shocks declines to 27% and that of flow supply shocks to 24%, while the explanatory power of flow demand shocks increases to 15%.”

Speculative demand played an important role in 1979.

“Speculative demand also played an important role in 1979.”

Despite additional loss of Iraqi output in early 2003, global oil production unexpectedly increased.

“Despite the additional loss of Iraqi output in early 2003, global oil production unexpectedly increased.”

The short-run price elasticity of gasoline demand is approximately of the same magnitude as the short-run price elasticity of oil demand in use.

“As shown in the online Appendix (supporting information), the short-run price elasticity of gasoline demand is approximately of the same magnitude as the shortrun price elasticity of oil demand in use.”

There is evidence of a small resurgence of speculative demand following the outbreak of the Iran-Iraq War, reflected in rising inventories and a higher real price of oil.

“There is also evidence of a small resurgence of speculative demand following the outbreak of the war, reflected in rising inventories and a higher real price of oil.”

The production shortfalls in Iraq and Venezuela were more than offset at the global level by increased oil production elsewhere.

“The production shortfalls in Iraq and Venezuela were more than offset at the global level by increased oil production elsewhere.”

Much of the observed increase in the real price of oil in 1979 only occurred later in 1979 after Iranian oil production had resumed.

“Much of the observed increase in the real price of oil, however, only occurred later in 1979 after Iranian oil production had resumed.”

The increase in the real price of oil in response to the 1980 Iran-Iraq War flow supply disruption was slightly larger than the price response to the 1979 supply disruption.

“As Figure 5 shows, the increase in the real price of oil in response to this flow supply disruption was slightly larger than the price response to the 1979 supply disruption.”

There is little evidence of flow supply shocks being responsible for the oil price surge of 1979.

“In contrast, there is little evidence of flow supply shocks being responsible for the oil price surge of 1979, consistent with the fact that overall global oil production increased in 1979, reflecting additional oil production outside of Iran.”

The posterior median estimate of the short-run price elasticity of oil demand in production is -0.44.

“Our posterior median estimate of this elasticity, as shown in the first column of the upper panel of Table II, is À0.44.”

There is little probability mass on oil demand elasticity values close to zero in the model.

“Moreover, the second column of Table II again shows that there is little probability mass on elasticity values close to zero, casting doubt on models of speculation that do not involve a change in oil inventories.24”

The paper's structural model implies a larger role for flow supply shocks in explaining fluctuations in the real price of oil than previous estimates.

“For example, our structural model implies a larger role for flow supply shocks in explaining fluctuations in the real price of oil than previous estimates.”

• AssumptionRests on an assumption: Standard structural VAR models of the market for crude oil implicitly equate oil production with oil consumption and ignore the role of oil inventories.

Shifts in speculative demand played a more important role during several earlier oil price shock episodes, notably in 1979, 1986, 1990 and 2002.

“Shifts in speculative demand played a more important role during several earlier oil price shock episodes, however, notably in 1979, 1986, 1990 and 2002.”

Similar relationships between price and inventory responses hold during other key historical episodes.

“Similar relationships were shown to hold during other key historical episodes.”

• AssumptionThe authors state this as an assumption: The structural shocks in the model are mutually uncorrelated.

Oil supply shocks played only a small role in the observed oil price increase in 1978-1981.

“Oil supply shocks played a small role only, violating the premise of Hamilton's calculations.”

• AssumptionThe authors state this as an assumption: The structural shocks in the model are mutually uncorrelated.
• AssumptionThe authors state this as an assumption: The real price of oil is covariance stationary for the historical decompositions.

In the context of a model that nests all leading explanations of how the oil price surge came about, the paper provides direct evidence against the popular view that the increase in the real price of oil from 2003 until mid 2008 was driven by speculators.

“In the context of a model that nests all leading explanations of how this oil price surge came about, we were able to provide direct evidence against the popular view that this increase was driven by speculators.”

Oil demand shocks were the main cause of the observed oil price increase in 1978-1981.

“We showed that oil demand shocks were the main cause of the observed oil price increase in 1978-1981.”

• AssumptionThe authors state this as an assumption: The structural shocks in the model are mutually uncorrelated.
• AssumptionThe authors state this as an assumption: The real price of oil is covariance stationary for the historical decompositions.

The added explanatory power of flow supply shocks for the real price of oil, especially in 1990, comes at the expense of the explanatory power of speculative demand shocks.

“The added explanatory power of flow supply shocks for the real price of oil, especially in 1990, comes at the expense of the explanatory power of speculative demand shocks.”

• AssumptionRests on an assumption: Standard structural VAR models of the market for crude oil implicitly equate oil production with oil consumption and ignore the role of oil inventories.

The collapse of the real price of oil in late 2008 and its partial recovery since early 2009 have been primarily driven by flow demand.

“Likewise, the collapse of the real price of oil in late 2008 and its partial recovery since early 2009 have been primarily driven by flow demand.”

The paper's short-run elasticity estimates are substantially higher than standard estimates cited in the literature.

“Our short-run elasticity estimates are substantially higher than standard estimates cited in the literature, casting doubt on models of speculative trading based on perfectly price-inelastic short-run demand for oil and for gasoline.”

The conclusion that the 2003 to mid-2008 oil price increase was not driven by speculators holds even for the 2007/08 period.

“This conclusion holds even for the 2007/08 period.”

The sustained run-up in the real price of oil between 2003 and mid 2008 was caused primarily by shifts in the global flow demand for oil.

“Rather, our results support recent findings in the literature that the sustained run-up in the real price of oil between 2003 and mid 2008 was caused primarily by shifts in the global flow demand for oil.”

The price and inventory data during the Persian Gulf War episode can be explained only by a combination of a positive speculative demand shock and a negative oil supply shock.

“Our analysis showed that the price and inventory data can be explained only based on a combination of these two shocks.”

• AssumptionThe authors state this as an assumption: The structural shocks in the model are mutually uncorrelated.

The paper's structural VAR model incorporates changes in oil inventories in computing the price elasticity of oil demand for the first time.

“it also incorporates for the first time changes in oil inventories in computing the price elasticity of oil demand.”

The accumulation of inventories is accomplished by a reduction in oil consumption and an increase in oil production as the real price increases.

“The accumulation of inventories is accomplished by a reduction in oil consumption (reflected in lower global real activity) and an increase in oil production, as the real price increases.”

The paper's structural model of the global market for crude oil explicitly allows for shocks to the speculative demand for oil as well as shocks to flow demand and flow supply.

“We develop a structural model of the global market for crude oil that for the first time explicitly allows for shocks to the speculative demand for oil as well as shocks to flow demand and flow supply.”

The paper's structural VAR model allows the construction of a direct estimate of the elasticity parameter based on exogenous shifts of the oil supply curve along the oil demand curve.

“Not only does this model allow the construction of a direct estimate of the elasticity parameter based on exogenous shifts of the oil supply curve along the oil demand curve”

Data on the oil futures spread are redundant in the paper's structural VAR model, given that changes in above-ground oil inventories are included.

“the analysis in Alquist and Kilian implies that data on the oil futures spread are redundant in our structural VAR model, given that we have already included changes in above-ground oil inventories.”

• cites Alquist and KilianNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• AssumptionThe authors state this as an assumption: Arbitrage between the spot and futures markets for crude oil holds.

A positive speculative demand shock shifts the demand for above-ground oil inventories, causing in equilibrium the level of inventories and the real price of oil to increase on impact.

“A positive speculative demand shock will shift the demand for above-ground oil inventories, causing in equilibrium the level of inventories and the real price of oil to increase on impact.”

With the exception of 1990, the major oil price shocks were driven primarily by oil demand shocks.

“we conclude that, with the exception of 1990, the major oil price shocks were driven primarily by oil demand shocks.”

Additional regulation of oil markets would not have prevented the 2003-2008 oil price surge.

“Our analysis implies that additional regulation of oil markets would not have prevented the 2003-2008 oil price surge.”

• Also stated in: § 1. INTRODUCTION · plain

  “additional regulation of oil futures markets would not have prevented the increase in the real price of oil in the physical market for oil”

Under reasonable assumptions about the oil refining industry, the short-run price elasticity of gasoline demand is about as high as the short-run price elasticity of oil demand.

“We provide a theoretical model that shows that under reasonable assumptions about the oil refining industry the short-run price elasticity of gasoline demand is about as high as the short-run price elasticity of oil demand.”

• AssumptionThe authors state this as an assumption: The oil refining industry satisfies assumptions such that the short-run price elasticity of gasoline demand is about as high as the short-run price elasticity of oil demand.

Under the maintained assumption of arbitrage between spot and futures markets, the absence of speculative pressures in the spot market implies that there cannot have been speculative pressures in the oil futures market either.

“under the maintained assumption of arbitrage between spot and futures markets, the absence of speculative pressures in the spot market implies that there cannot have been speculative pressures in the oil futures market either.”

• AssumptionThe authors state this as an assumption: Arbitrage between the spot and futures markets for crude oil holds.

The speculative component of the real price of oil is identified with the help of data on oil inventories.

“The speculative component of the real price of oil is identified with the help of data on oil inventories.”

The 2003-2008 oil price surge was caused by unexpected increases in world oil consumption driven by the global business cycle.

“Instead, this surge was caused by unexpected increases in world oil consumption driven by the global business cycle.”

• Also stated in: § 1. INTRODUCTION · plain

  “our model implies that both spot and futures prices during 2003-2008 were driven by unexpected increases in world oil consumption.”

• Also stated in: § 1. INTRODUCTION · plain

  “unexpected fluctuations in global real activity explain most of the surge in the real price of oil between 2003 and mid 2008”

The limiting case of a zero price elasticity of demand is an unlikely explanation of the 2003-2008 surge in the real price of oil.

“Thus the limiting case of a zero price elasticity of demand discussed by Hamilton (2009a) is an unlikely explanation of the 2003-2008 surge in the real price of oil.”

• cites Hamilton (2009a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

An unexpected increase in the uncertainty about future oil supply shortfalls would have much the same effect as news about future oil supplies and demand.

“An unexpected increase in the uncertainty about future oil supply shortfalls would have much the same effect.This point has been demonstrated formally in a general equilibrium model by Alquist and Kilian (2010).”

• cites Alquist and Kilian (2010) sourceNot checked — no full text of the source

Existing estimates of the short-run price elasticity of oil demand are based on dynamic reduced-form regressions that ignore the endogeneity of the real price of oil.

“These existing elasticity estimates, however, are based on dynamic reduced-form regressions that ignore the endogeneity of the real price of oil.”

Given that crude oil is storable, the real price of oil also depends on the demand for oil stocks or inventories.

“Given that crude oil is storable, the real price of oil also depends on the demand for oil stocks or inventories.”

Both spot and futures prices during 2003-2008 were driven by unexpected increases in world oil consumption.

“our model implies that both spot and futures prices during 2003-2008 were driven by unexpected increases in world oil consumption.”

• AssumptionThe authors state this as an assumption: Arbitrage between the spot and futures markets for crude oil holds.

The estimate of Hausman and Newey (1995) suggests a bound of -0.8 ≤ Use_t ≤ 0.

“Their estimate suggests a bound of À0:8≤ Use t ≤0.”

• cites Hausman and Newey (1995) sourceNot checked — no full text of the source

• AssumptionThe authors state this as an assumption: The impact price elasticity of oil demand is lower in magnitude than the corresponding long-run price elasticity of oil demand.

Professional forecasters persistently underestimated global growth during 2003-2008, especially growth in emerging Asia.

“Even professional forecasters persistently underestimated global growth during 2003-2008, especially growth in emerging Asia, lending credence to our model results.”

• cites Kilian and Hicks (2013) sourceNot checked — no full text of the source

Even dramatic increases in US oil production would not lower the real price of oil substantially at the global level.

“Our structural model also implies that even dramatic increases in US oil production would not lower the real price of oil substantially at the global level, while a full recovery of the global economy would raise the real price of oil by as much as $50 in real terms (see Baumeister and Kilian, 2012b).”

• cites Baumeister and Kilian, 2012bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Further regulation of oil markets would have done nothing to stem the increase in the real price of oil.

“This finding is important because it implies that further regulation of oil markets would have done nothing to stem the increase in the real price of oil.”

The surge in the real price of oil between 2003 and mid 2008 was mainly caused by shifts in the flow demand for crude oil associated with the global business cycle.

“It supports the substantive conclusion in Kilian (2009a,b) that the surge in the real price of oil between 2003 and mid 2008 was mainly caused by shifts in the flow demand for crude oil associated with the global business cycle.”

• cites Kilian (2009a,b)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The inventory data are as accurate as the oil futures spread.

“consistent with the view that the inventory data are as accurate as the oil futures spread.”

Inventory data allow the paper to impose identifying information about the price elasticity of oil demand, which could not be imposed when using the oil futures spread.

“inventory data allow us to impose identifying information about the price elasticity of oil demand, which could not be imposed when using the oil futures spread.”

The paper's inventory data, while less than perfect, are still informative.

“Thus there is reason to believe that our inventory data, while less than perfect, are still informative.”

A benchmark for the long-run price elasticity of oil demand is provided by studies of nonparametric gasoline demand functions based on US household survey data such as Hausman and Newey (1995), which have consistently produced long-run price elasticity estimates near -0.8.

“A benchmark for that long-run elasticity is provided by studies of nonparametric gasoline demand functions based on US household survey data such as Hausman and Newey (1995), which have consistently produced long-run price elasticity estimates near À0.8.”

• cites Hausman and Newey (1995) sourceNot checked — no full text of the source

Impact sign restrictions alone are typically too weak to be informative about the effects of oil demand and oil supply shocks.

“Impact sign restrictions alone are typically too weak to be informative about the effects of oil demand and oil supply shocks.”

The model has the ability to detect speculative demand shocks when they exist.

“Although speculative motives played no important role after 2003, there are other oil price shock episodes when they did, suggesting that our model has the ability to detect speculative demand shocks when they exist.”

The run-up in the real price of oil occurred because of the cumulative effects of many positive flow demand shocks over the course of several years.

“Our model shows that the run-up in the real price of oil occurred because of the cumulative effects of many positive flow demand shocks over the course of several years.”

The main results are robust to the choice of admissible structural model.

“We also conducted the same analysis with every other admissible structural model and verified that our main results are robust to the choice of admissible model.”

An exogenous influx of financial speculators cannot have driven up the oil futures price, because the absence of speculation in the spot market rules out speculation in oil futures markets under the standard assumption of arbitrage.

“Second, this result tells us that an exogenous influx of financial speculators cannot have driven up the oil futures price, becauseunder the standard assumption of arbitrage between the futures and spot markets for oil maintained by proponents of the financial speculation hypothesisthe absence of speculation in the spot market also rules speculation in oil futures markets.”

• AssumptionThe authors state this as an assumption: Under the standard assumption of arbitrage between the futures and spot markets for oil, the absence of speculation in the spot market rules out speculation in oil futures markets.

Given one of the largest unexpected oil supply disruptions in history, one would have expected a much larger decline in oil inventories given the impulse response estimates in Figure 1.

“Given one of the largest unexpected oil supply disruptions in history, one would have expected a much larger decline in oil inventories given the impulse response estimates in Figure 1.”

The observed large increase in the real price of oil in 1990 was caused by both shocks working in the same direction.

“At the same time, the observed large increase in the real price of oil was caused by both shocks working in the same direction.”

The paper's structural VAR model is set-identified based on a combination of sign restrictions and bounds on the implied price elasticities of oil demand and oil supply.

“Our structural VAR model is set-identified based on a combination of sign restrictions and bounds on the implied price elasticities of oil demand and oil supply.”

A full recovery of the global economy would raise the real price of oil by as much as $50 in real terms.

“Our structural model also implies that even dramatic increases in US oil production would not lower the real price of oil substantially at the global level, while a full recovery of the global economy would raise the real price of oil by as much as $50 in real terms (see Baumeister and Kilian, 2012b).”

• cites Baumeister and Kilian, 2012bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

In the presence of changes in oil inventories, the standard definition of oil demand elasticity that equates production with consumption is inappropriate.

“In the presence of changes in oil inventories that definition is inappropriate.”

The relevant quantity measure for oil demand elasticity is the sum of the flow of oil production and the depletion of oil inventories triggered by an oil supply shock.

“The relevant quantity measure is instead the sum of the flow of oil production and the depletion of oil inventories triggered by an oil supply shock.”

The speculative supply shock hypothesis can be rejected.

“Hence we can reject the speculative supply shock hypothesis.”

• AssumptionThe authors state this as an assumption: In the model, OPEC holding back oil production in anticipation of rising oil prices is observationally equivalent to a negative flow supply shock.

There is no indication that flow supply shocks played an important role in the oil price surge of 1979.

“Thus there is no indication that flow supply shocks played an important role in the oil price surge of 1979.”

The evidence implies a much larger role for speculative demand in 1990/91 than the data for the month of August alone would suggest.

“Taken in conjunction our evidence implies a much larger role for speculative demand in 1990/91 than the data for the month of August alone would suggest.”

Whether the limiting case discussed in Hamilton (2009a) is empirically relevant depends on the magnitude of the short-run price elasticity of oil demand.

“Whether the limiting case discussed in Hamilton (2009a) is empirically relevant thus depends on the magnitude of the short-run price elasticity of oil demand.”

• cites Hamilton (2009a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• AssumptionRests on an assumption: The price elasticity of oil supply is near zero in the short run.

The combined cutback in oil production in 2002-2003 was of a magnitude similar to the oil supply disruptions of the 1970s.

“The combined cutback in oil production was of a magnitude similar to the oil supply disruptions of the 1970s (see Kilian, 2008a).”

• cites Kilian, 2008aNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The short-run price elasticity of oil demand is a key parameter in the theoretical models of speculation of Caballero et al. (2008) and Hamilton (2009a).

“For example, it is a key parameter in the theoretical models of speculation L. KILIAN AND D. P. MURPHY 472 Caballero et al. (2008) and Hamilton (2009a).”

• cites Caballero et al. (2008) sourceNot checked — no full text of the source
• cites Hamilton (2009a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The posterior median estimate of the oil demand elasticity in production is misleading in that it ignores the role of inventories.

“The posterior median estimate of O, Production in the first column of Table II, while instructive when compared to conventional estimates, is misleading in that it ignores the role of inventories.”

• AssumptionRests on an assumption: Typical estimates of the oil demand elasticity in the literature have been based on models that equate the percent change in quantity with the percent change in the production of crude oil, ignoring the existence of oil inventories.

Relatively high short-run price elasticities of oil demand are fully compatible with the view that economic fundamentals are responsible for the surge in the real price of oil after 2003.

“Indeed, this result illustrates that relatively high short-run price elasticities of oil demand are fully compatible with the view that economic fundamentals are responsible for the surge in the real price of oil after 2003.23”

Allowing for inventory responses will tend to lower the magnitude of the price elasticity of oil demand.

“By construction, allowing for inventory responses will tend to lower the magnitude of the price elasticity of oil demand.”

• AssumptionRests on an assumption: Typical estimates of the oil demand elasticity in the literature have been based on models that equate the percent change in quantity with the percent change in the production of crude oil, ignoring the existence of oil inventories.

The low elasticity estimates typically reported in the literature can be explained by the bias toward zero in reduced-form estimates.

“This fact helps explain the low elasticity estimates typically reported in the literature.21”

Much of the existing literature has attempted to estimate the oil demand elasticity from dynamic reduced-form models that do not distinguish between oil demand and oil supply shocks.

“First, much of the existing literature has attempted to estimate the oil demand elasticity from dynamic reduced-form models that do not distinguish between oil demand and oil supply shocks (see, for example, Dahl, 1993; Cooper, 2003).”

• cites Dahl, 1993 sourceNot checked — no full text of the source
• cites Cooper, 2003 sourceNot checked — no full text of the source

The difference between the paper's estimate and conventional estimates can be attributed to the difference between estimating a structural and a reduced-form model.

“The difference in results can be attributed to the difference between estimating a structural and a reduced-form model.22”

Traders in anticipation of an expected oil shortage will buy and store crude oil now with the expectation of selling later at a profit.

“The reason is that traders in anticipation of the expected oil shortage will buy and store crude oil now with the expectation of selling later at a profit.”

The posterior median estimate of the oil demand elasticity in production is much higher in magnitude than conventional reduced-form regression estimates of this elasticity.

“It is also much higher in magnitude than conventional reduced-form regression estimates of this elasticity.”

Identification of the oil demand elasticity requires an exogenous shift of the contemporaneous oil supply curve along the contemporaneous oil demand curve within a structural model.

“This is not possible because the identification of the demand elasticity requires an exogenous shift of the contemporaneous oil supply curve along the contemporaneous oil demand curve within the context of a structural model.”

Reduced-form estimates of the oil demand elasticity fail to account for the endogeneity of the price of crude oil and hence are biased toward zero.

“Reduced-form estimates of the oil demand elasticity fail to account for the endogeneity of the price of crude oil and hence are biased toward zero.”

The revised median estimate of the oil demand elasticity in use is still four times larger than conventional elasticity estimates.

“While this large reduction in the magnitude of the elasticity highlights the importance of accounting for changes in oil inventories, the revised median estimate is still four times larger than conventional elasticity estimates.”

• AssumptionRests on an assumption: Typical estimates of the oil demand elasticity in the literature have been based on models that equate the percent change in quantity with the percent change in the production of crude oil, ignoring the existence of oil inventories.

The large reduction in the magnitude of the elasticity from production to use highlights the importance of accounting for changes in oil inventories.

“While this large reduction in the magnitude of the elasticity highlights the importance of accounting for changes in oil inventories, the revised median estimate is still four times larger than conventional elasticity estimates.”

• AssumptionRests on an assumption: Typical estimates of the oil demand elasticity in the literature have been based on models that equate the percent change in quantity with the percent change in the production of crude oil, ignoring the existence of oil inventories.

The inclusion of oil inventories matters.

“The inclusion of oil inventories matters.”

• AssumptionRests on an assumption: Standard structural VAR models of the market for crude oil implicitly equate oil production with oil consumption and ignore the role of oil inventories.

The posterior median estimate of the oil demand elasticity in production is seven times higher than typical conjectures in the recent literature.

“This estimate is seven times higher than typical conjectures in the recent literature.”

Models of speculation that do not involve a change in oil inventories are cast into doubt.

“Moreover, the second column of Table II again shows that there is little probability mass on elasticity values close to zero, casting doubt on models of speculation that do not involve a change in oil inventories.24”

Other recent studies relying on alternative structural models have obtained similarly large oil demand elasticity estimates ranging from -0.35 to -0.41.

“Another way of putting these results into perspective is to observe that other recent studies relying on alternative structural models have obtained similarly large oil demand elasticity estimates ranging from À0.35 to À0.41 (see, for example, Serletis et al., 2010; Bodenstein and Guerrieri, 2012; Baumeister and Peersman, 2013).”

• cites Serletis et al., 2010 sourceNot checked — no full text of the source
• cites Bodenstein and Guerrieri, 2012Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Baumeister and Peersman, 2013 sourceNot checked — no full text of the source

Speculative demand shocks must be represented as shifts of the contemporaneous oil demand curve rather than the contemporaneous oil supply curve, even if the shift in expectations is about a cut in future oil supplies rather than an increase in future oil demand.

“Such speculative demand shocks must be represented as shifts of the contemporaneous oil demand curve rather than the contemporaneous oil supply curve, even if the shift in expectations is about a cut in future oil supplies rather than an increase in future oil demand.”

Without accounting for shifts in the speculative demand for oil, it is not possible to understand the evolution of the real price of oil during the earlier episodes of 1979, 1986, 1990 and 2002.

“We showed that, without accounting for shifts in the speculative demand for oil, it is not possible to understand the evolution of the real price of oil during these episodes.”

The speculative supply shock from oil producers leaving oil below ground would be associated with a negative flow supply shock in the paper's framework rather than the building of above-ground inventories.

“The latter form of 'speculative supply shock' would be associated with a negative flow supply shock in our framework rather than the building of above-ground inventories.”

Additional regulation of oil derivatives markets is not likely to lower the real price of oil.

“By contrast, additional regulation of oil derivatives markets is not likely to lower the real price of oil, nor can increased domestic oil production in the USA be expected to have much of an effect on the global real price of oil (also see Baumeister and Kilian, 2012b).”

• cites Baumeister and Kilian, 2012bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The paper's result breaks down only if demand for oil is completely price inelastic.

“This result breaks down only if demand for oil is completely price inelastic, a case that we discuss in more detail in section 5.”

The paper's conclusion regarding the determinants of the spot price of oil would remain equally valid if there were limits to arbitrage between spot and futures markets.

“It is important to note that our conclusion regarding the determinants of the spot price of oil would remain equally valid if there were limits to arbitrage between spot and futures markets.”

There is no evidence that oil producers have responded systematically to news triggering positive speculative demand shocks by lowering oil production in anticipation of predictable real price increases.

“Although oil producers could conceivably react to the same news that triggers a positive speculative demand shock by lowering oil production in anticipation of predictable increases of the real price, there is no evidence that oil producers have responded systematically in this way.”

The paper's model allows estimation of both the traditional oil demand elasticity in production and the more relevant oil demand elasticity in use, which incorporates changes in oil inventories.

“We provided a model that allows the estimation of both the traditional oil demand elasticity in production and of the more relevant oil demand elasticity in use which incorporates changes in oil inventories.”

Traditional estimates of the short-run price elasticity of oil demand are not credible.

“We also observed that traditional estimates of the short-run price elasticity of oil demand are not credible.”

The higher short-run elasticity estimates cast doubt on models of speculative trading based on perfectly price-inelastic short-run demand for oil and for gasoline.

“Our short-run elasticity estimates are substantially higher than standard estimates cited in the literature, casting doubt on models of speculative trading based on perfectly price-inelastic short-run demand for oil and for gasoline.”

Conventional estimates of the short-run price elasticity of oil demand from dynamic reduced-form regressions, as in Dahl (1993) and Cooper (2003), ignored the endogeneity of the real price of oil, causing the elasticity estimate to be downward biased.

“One problem is that conventional estimates of this elasticity from dynamic reduced-form regressions, as in Dahl (1993) and Cooper (2003), have ignored the endogeneity of the real price of oil, causing the elasticity estimate to be downward biased.”

• cites Dahl (1993) sourceNot checked — no full text of the source
• cites Cooper (2003) sourceNot checked — no full text of the source

Increased domestic oil production in the USA cannot be expected to have much of an effect on the global real price of oil.

“By contrast, additional regulation of oil derivatives markets is not likely to lower the real price of oil, nor can increased domestic oil production in the USA be expected to have much of an effect on the global real price of oil (also see Baumeister and Kilian, 2012b).”

• cites Baumeister and Kilian, 2012bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Existing estimates of the short-run price elasticity of oil demand, including the structural estimates of Baumeister and Peersman (2013), ignored the role of inventories in smoothing oil consumption in response to oil supply shocks.

“Moreover, existing estimates, including the structural estimates recently provided by Baumeister and Peersman (2013), have ignored the role of inventories in smoothing oil consumption in response to oil supply shocks.”

• cites Baumeister and Peersman (2013) sourceNot checked — no full text of the source

The absence of a sharp decline in oil inventories in August 1990 is inconsistent with the view that the price increase reflected a negative oil supply shock.

“At the same time, the absence of a sharp decline in oil inventories in August of 1990 is inconsistent with the view that the price increase reflected a negative oil supply shock.”

Using monthly well-level oil production data from Texas, there is essentially no response of oil production to either the spot price or the oil futures price.

“Kellogg (2011) using monthly well-level oil production data from Texas finds essentially no response of oil production to either the spot price or the oil futures price.”

• cites Kellogg (2011)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

A situation in which arbitrage breaks down is not consistent with the scenario envisioned by researchers who attribute rising spot prices to speculation by financial investors, because in that case speculation-driven increases in the oil futures price could not possibly be transmitted to the spot price of oil.

“Clearly, a situation in which arbitrage breaks down is not consistent with the scenario envisioned by researchers who attribute rising spot prices to speculation by financial investors, because in that case speculationdriven increases in the oil futures price could not possibly be transmitted to the spot price of oil.”

In the absence of significant excess production capacity, the short-run price elasticity of oil supply is very low.

“Hamilton (2009b, p. 25) observes that 'in the absence of significant excess production capacity, the short-run price elasticity of oil supply is very low'.”

• cites Hamilton (2009b)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

In the limiting case of no arbitrage, we could still infer from our model whether there is speculation taking place in the spot market.

“although we could still infer whether there is speculation taking place in the spot market.”

In the limiting case of no arbitrage, we would be unable to infer from our model whether there is speculation in oil futures markets.

“In the limiting case of no arbitrage we would be unable to infer from our model whether there is speculation in oil futures markets”

Our analysis of the spot market would remain valid if arbitrage were impeded or broke down completely.

“Our analysis of the spot market would remain valid if this arbitrage were impeded or broke down completely”

Even in the presence of spare capacity, the response of oil supply within the month to price signals will be negligible because changing oil production is costly.

“Kilian (2009a) makes the case that, even in the presence of spare capacity, the response of oil supply within the month to price signals will be negligible because changing oil production is costly.”

• cites Kilian (2009a)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

If arbitrage were impeded or broke down completely, the oil futures price would become disconnected from the spot price.

“but the oil futures price would become disconnected from the spot price in that case.”