It has been widely remarked that world conventional crude seems to have reached a plateau of production in 2005, at a time when oil prices were ascendant. Since one would expect producers to respond to higher prices with higher production, this has struck many as strange. Since Saudi Arabia is the ‘king’ of conventional oil producers, it is logical to examine their production record during this period. One finds the same story – a strange inelasticity in their response to higher prices. Theories for why abound – they wanted to keep oil off the market as a weapon to keep other OPEC members in line; the crude they were selling was not the crude needed by the world’s refiners/users; they could not increase production without damaging the North Ghawar field; or they simply were at the maximum flow rates for their existing fields at the time. Instead of speculating as to why they responded the way they did, this article simply provides a narrative, largely using figures, that shows there was indeed a change in the way that Saudi Oil production responded to price changes.
Econ 101 and Cartels
Generally speaking, supply is increasing in price. Saudi Arabia reputation is as the enforcer for OPEC – they are the largest producer, and one of the only members who credibly adhere to quotas. Given this role, the elasticity of supply has historically been quite high for the Saudis meaning that a sizeable change in price would be expected to elicit a dramatic change in production. Often, Saudi shifts in production would increase during geopolitical events (such as wars and strikes) that simultaneously drove oil prices higher on worries. Conversely, large Saudi supply cuts have often marked oil price bottoms. There appears to have been a paradigm shift in roughly 2005, where Saudi oil supply became markedly less elastic. It will be interesting to see if they respond to current unrest in the Middle East (and $100 world oil prices) with an announcement or more production – thus far we have radio silence.
Figure 1: Why did production decline during a period (2005-2008) when prices increased by more than 100%?
There are many reasons why Saudis may not have produced more oil during the period 2005-2008 but what is undeniable is that practically speaking, there was a change in the response between price and new marginal supply from the Saudis.
Scatter Plots tell an interesting story
The statistical way of determining the relationship between price and supply would be to aggregate the data, and use a linear regression to see if there is a response in supply to price changes. Scatter plots are a nice graphic way of seeing linear regressions. In a scatter plot you simply plot the data points (in this case Saudi oil supply and price of oil. In the following figures a story emerges of SA production versus price.
- There is a clear response to oil price spikes of 1990 (Iraq war), 2000, and 2003 (2nd Iraq war). Price elasticity appears to be 100-200 thousand barrels a day (kbd henceforth) per dollar during these events.
- It appears price elasticity declines between late 2003 and 2005.
- From 2005 to 2007, the coefficient is negative, implying production is in decline for some other reason!
- Supply elasticity returns to positive from 2007 into the price spike of 2008 and then also the crash of late 2008 – however the coefficient is dramatically smaller (inelastic).
- 2009-present, production has increased with price, but the coefficient (that is the multiplier on the response to price change) is smaller still. Price elasticity over the past 2 years is only 9 kbd per dollar.
A Couple of Notes
Before letting the scatter plots ‘tell the story’, please allow a couple clarifications: in this presentation ‘elasticity’ refers to nominal change in production per nominal change in price – not the standard formula for Price Elasticity of Supply (% change in supply/%change in price). Standard approximations are given for standard PES for each period in the addendum. Also note that Saudi oil production figures come from the IEA without including their portion of production from the neutral zone, so production figures will consistently be a few hundred thousand barrels a day lower than total SA production found in other material. Including neutral zone production does not materially affect the results.
Figure 2: Saudi Oil production (vertical axis) versus price (horizontal axis) January 1989 – February 1991: As one would expect, supply responds positively to price: supply elasticity is 214kb/d per $ during this period.
Figure 3: November 1999 – October 2001: Supply response to the price spike of 2000 was quite robust. Prices exceeded the price band of - for the first time since 1990. Supply elasticity is 133 kb/d per $ during this period. OPEC shifts its price band to -/barrel.
Figure 4: May 2002 – October 2003: Going into the second Iraq war, supply elasticity is quite high again (195 kb/day per $). Note that SA production is cut to only 8.6 mb/d at the end of the series and prices stabilize.
Figure 5: October 2003 – June 2005: Prices leave OPEC band (OPEC band gradually fades from memory). Notably, price elasticity declines from 195 kb/d per $ to 63 kb/d per $. Whether SA “won’t” or “can’t” produce more than 9.8 mb/day is debated.
Figure 6: June 2005 – February 2007: This isn’t supposed to happen. Supply elasticity is negative during this period. While prices moved to all-time nominal highs, SA gradually cut (or lost) production.
Figure 7: February 2007 – February 2009: “Normal” elasticity returns during this period. SA adds 1 mb/d to production during this time, but elasticity is dramatically lower than in previous periods (17 kb/d per $). Did SA implicitly reset their price bands and supply elasticity? Or were they supply constrained and producing as much as possible?
Figure 8: February 2009 – November 2010: Supply elasticity falls even more during this period to 9 kb/d per $. Ostensibly, oil quotas remained unchanged throughout this period, but SA clearly still responds to price – but with reduced elasticity.
Figure 9: Old Paradigm – November 1999 to October 2004. Supply responds powerfully to price as SA attempts to keep price within OPEC bands.
Figure 10: New Paradigm (October 2004- November 2010): Supply appears to be unresponsive to price - R^2= .0063. It appears that price is not a useful determinant for supply. However, after adding a time factor SA is clearly still responsive to price.
Figure 11: New Paradigm: Decreasing production over time. Price regains its explanatory power during this period when testing the hypothesis that production is decreasing each year by 0.24 mb/d. This decline is modeled by using a dummy variable that adds 20 kb/day of production for each period (months) past the starting data point. Supply is still much more inelastic than in previous periods.
Conclusion: Actions are louder than words
While there are many explanations for why it might be so, the data presented above implies that:
- Using a definition of spare capacity as “willing to produce” rather than “capable of producing” Saudi Arabia has less spare capacity than they claim. In fact using the elasticity response of the past 2 years implies that they only have approximately 0.7 mb/d of spare capacity at an oil price up to $125.
- Saudi Arabia production has decreased for a reason unrelated to price by approximately 0.24 mb/day each year since 2005.
Addendum
Approximate Price Elasticity of Supply for Saudi Arabia in different time periods:
- January 1989 – February 1991: 0.59
- November 1999 – October 2001: 0.41
- October 2001 – May 2002: 0.10
- May 2002 – October 2003: 0.59
- October 2003 – June 2005: 0.20
- June 2005 – February 2007: -0.08
- February 2007 – February 2009: 0.13
- February 2009 – November 2010: 0.04
