15th IAEE European Conference 2017
OPEC, Saudi Arabia, and the Shale Revolution:
Insights from Equilibrium Modelling and Oil Politics
Dawud Ansari, DIW Berlin & EADP
Vienna, 4th September, 2017
OPEC, Saudi Arabia, and the Shale Revolution: Insights from - - PowerPoint PPT Presentation
15 th IAEE European Conference 2017 OPEC, Saudi Arabia, and the Shale Revolution: Insights from Equilibrium Modelling and Oil Politics Dawud Ansari, DIW Berlin & EADP Vienna, 4 th September, 2017 Outline 2014 2016: Oil price crash
15th IAEE European Conference 2017
Dawud Ansari, DIW Berlin & EADP
Vienna, 4th September, 2017
2
shale elasticity
an important player
Dawud Ansari | 9 / 4 / 17
Content
Dawud Ansari | 9 / 4 / 17
3
Dawud Ansari | 9 / 4 / 17
4
Dawud Ansari | 9 / 4 / 17
5
WTI quarterly
Thomson Reuters Crude oil production capacities Data: IEA, own calculations
Price drop
Production cut agreed (incl. Russia, Iran) Nov 2014 Vienna talks: No OPEC cuts Jun 2016 OPEC deal talks fail Dec 2015: Vienna talks: Still no cuts Price collapse Feb 2016 Russia signals interest in cuts
Dawud Ansari | 9 / 4 / 17
6
Shale oil revolution (e.g. Aguilera and Radetzki, 2015) Financial speculation (e.g. Fantazzini, 2016, Tokic, 2015) Geopolitical stabilisation (e.g. Baffes et al., 2015) Dampened demand (e.g. Baumeister and Kilian, 2016) OPEC Behaviour New economics of oil Dale (2016)
Uncertainty
Behar and Ritz (2017) Coy (2015) Gause (2015) Mănescu and Nuño (2015) Baffes et al. (2015) Baumeister and Kilian (2016) Dale (2016) Kaletsky (2015) The Economist (2015)
Fattouh et al. (2016) Huppmann and Livingston (2015)
7
“[Ali al-Naimi’s] biggest move was the latest
Mohammad al-Sabban, June 2015
[…] It is not in the interest of OPEC producers to cut their production. […] Whether [the price] goes down to $20/B, $40/B, $50/B, $60/B, it is irrelevant. […] But if it goes down, others will be harmed greatly before we feel any pain.
Ali al-Naimi, November 2014
Dawud Ansari | 9 / 4 / 17
Dawud Ansari | 9 / 4 / 17
8
e.g. Kisswani (2016) Plaut (1981) e.g. Huppmann & Holz (2012) Almoguera et al. (2011) e.g. Kisswani (2014), Hochman and Zilberman (2015)
And even worse: How to model that? Fattouh and Mahadeva (2013): Changing OPEC objectives and behaviour over time make it impossible to have a single model explaining all OPEC history.
9
Dawud Ansari | 9 / 4 / 17
Dawud Ansari | 9 / 4 / 17
10
Bathtub market
function
Golombek production costs Linear demand
𝑟𝑗𝑢
𝑡
𝑢: 2011 Q4 – 2015 Q4, quarterly An extension of Huppmann (2013)
Dawud Ansari | 9 / 4 / 17
11
Source Type IEA (29 suppliers with 94.4% of global supply) Supply OPEC: IEA , non-OPEC: 97%-of-output rule and IEA (e.g. Behar & Ritz, 2017) Capacities DIW data set (e.g. Langer et al, 2016) Production costs Calculations based on US Dept. of Energy, EIA, Oil & Gas Journal Oil quality adjustment Survey-based: Javan & Zahran (2015), Caldara et al. (2016) Demand elasticity
Share in global crude production capacities Gini coefficient: 0.505 Data: IEA and own calculations
Solver Formulation Setup PATH MCP Cournot, Perfect Comp. Bonmin, Couenne MPEC MINLP Stackelberg
Dawud Ansari | 9 / 4 / 17
12 Actual and computed price trajectories
ARME in % KSA-FR PC KSA-CO Cournot UNI-CO Overall 23 27 35 52 120 First period 25 31 24 43 121 Second period 18 18 63 75 119
Goodness of fit
Dawud Ansari | 9 / 4 / 17
13 Computed profits (left) and production quantities (right) in the United OPEC setup in Q1 2015 by Saudi Arabia (KSA) and other OPEC members
Dawud Ansari | 9 / 4 / 17
14 Robustness of the perfect competition results to cost variations (overall cost reductions in %)
15
Dawud Ansari | 9 / 4 / 17
Dawud Ansari | 9 / 4 / 17
16
Fiscal breakeven prices in USD / bbl 2013 – 2015. Data: IMF
Dawud Ansari | 9 / 4 / 17
17 Month-to-month and quarter-to-quarter changes in US rigs (left) and quarter-to-quarter and year-to-year changes in US daily crude oil production (right). Data: EIA
18
Dawud Ansari | 9 / 4 / 17
Dawud Ansari | 9 / 4 / 17
19
Thank you for your attention.
DIW Berlin — Deutsches Institut für Wirtschaftsforschung e.V. Mohrenstraße 58, 10117 Berlin www.diw.de Editor Dawud Ansari dansari@diw.de
cooperative collusion or just non-cooperative? Oxford Review of Economic Policy 27, 144-168.
consequences, and policy responses. Available at SSRN: http://dx.doi.org/10.2139/ssrn.2624398.
Journal of the Association of Environmental and Resource Economists 3, 131-158.
Economics 63, 185-198.
Mimeo, Federal Reserve Board.
Energy Policy 96, 383-396.
trade-off: Saudi Arabia’s oil policy in the 2014–15 price fall. Oxford Review of Economic Policy 32, 223-240.
427-443.
OPEC Energy Review 39, 53-76.
Dawud Ansari | 9 / 4 / 17
21
Dawud Ansari | 9 / 4 / 17
22
Fringe.
33, 1.
Current Upheaval in the Global Crude Oil Market. International Association for Energy Economics Energy Forum Index Third Quarter 2015.
Economics and Finance 38, 96-118.
Energy Policy 97, 171-180.
Economics: Benchmarks, Breakeven Points, and Inelasticities. MIT CEEPR Working Paper.
equilibrium analysis. Energy Policy 97, 258-266.
855-866.
Dawud Ansari | 9 / 4 / 17 23
Set Indices 𝑗 ∈ 𝐽 Crude oil producing countries 𝑘 ∈ 𝐾 ⊆ 𝐽 Stackelberg leaders 𝑙 ∈ 𝐿 ⊆ 𝐽 Stackelberg followers 𝑢 ∈ 𝑈 Time periods in quarterly steps from 4th quarter 2011 onwards Parameters 𝛾1𝑢, 𝛾2𝑢 Demand parameters 𝜁 Price elasticity 𝜒𝑢 Observed actual price 𝜓𝑢 Observed actual quantity 𝛿1𝑗, 𝛿2𝑗, 𝛿3𝑗 Cost parameters 𝜆𝑗𝑢 Production capacity 𝜃𝑗 Quality of oil index Variables 𝑞𝑢 ∈ ℝ0
+
Market price in period 𝑢 𝑟𝑗𝑢 ∈ ℝ0
+
Quantity supplied by producer 𝑗 in period 𝑢
Dawud Ansari | 9 / 4 / 17
24
Dawud Ansari | 9 / 4 / 17
25
2 − 𝛿3𝑗 𝑟𝑗𝑢 − 𝜆𝑗𝑢
𝑗∈𝐽
𝑟𝑗𝑢 β1t = φt 1 − ε−1 β2t = φt χtε −1 𝑟𝑗𝑢 ≤ 𝜆𝑗𝑢
Dawud Ansari | 9 / 4 / 17
26
0 ≤ 𝑞𝑢 − 𝜃𝑗𝑢𝑁𝐷𝑗𝑢 ⊥ 𝜆𝑗𝑢 − 𝑟𝑗𝑢 ≥ 0 ∀𝑗 ∈ 𝐽 ∀𝑢 ∈ 𝑈 𝑁𝐷𝑗𝑢 = 𝛿1𝑗 + 2𝛿2𝑗𝑟𝑗𝑢 − 𝛿3𝑗 ln 1 − 𝑟𝑗𝑢 𝜆𝑗𝑢 ∀𝑗 ∈ 𝐽 ∀𝑢 ∈ 𝑈 𝑞𝑢 = 𝛾1𝑢 + 𝛾2𝑢 𝑟𝑗𝑢
𝑗∈𝐽
∀𝑢 ∈ 𝑈
Dawud Ansari | 9 / 4 / 17
27
0 ≤ 𝑞𝑢 − 𝜃𝑗𝑢𝑁𝐷𝑗𝑢 − 𝜐𝑗 ⊥ 𝜆𝑗𝑢 − 𝑟𝑗𝑢 ≥ 0 ∀𝑗 ∈ 𝐽 ∀𝑢 ∈ 𝑈 𝑁𝐷𝑗𝑢 = 𝛿1𝑗 + 2𝛿2𝑗𝑟𝑗𝑢 − 𝛿3𝑗 ln 1 − 𝑟𝑗𝑢 𝜆𝑗𝑢 ∀𝑗 ∈ 𝐽 ∀𝑢 ∈ 𝑈 𝑞𝑢 = 𝛾1𝑢 + 𝛾2𝑢 𝑟𝑗𝑢
𝑗∈𝐽
∀𝑢 ∈ 𝑈
Dawud Ansari | 9 / 4 / 17
28
max
𝑟𝑘𝑢 ∀𝑘∈𝐾
𝑞𝑢 ∗ 𝑟𝑘𝑢
𝑘∈𝐾
− 𝜃𝑘𝑢 𝐷
𝑘𝑢 + +𝜐𝑘𝑟𝑘𝑢 𝑘∈𝐾
∀𝑢 ∈ 𝑈 𝐷
𝑘𝑢 = 𝛿1𝑗𝑟𝑗𝑢 + 𝛿2𝑗𝑟𝑗𝑢 2 − 𝛿3𝑗 𝑟𝑗𝑢 − 𝜆𝑗𝑢 ln 1 − 𝑟𝑗𝑢
𝜆𝑗𝑢 − 1 ∀𝑘 ∈ 𝐾 ∀𝑢 ∈ 𝑈 0 ≤ 𝑞𝑢 + 1 − 𝑔 𝛾2𝑢𝑟𝑙𝑢 − 𝜃𝑙𝑢𝑁𝐷𝑙𝑢 ∀𝑙 ∈ 𝐿 ∀𝑢 ∈ 𝑈 𝑁𝐷𝑙𝑢 = 𝛿1𝑙 + 2𝛿2𝑙𝑟𝑙𝑢 − 𝛿3𝑙 ln 1 − 𝑟𝑙𝑢 𝜆𝑙𝑢 ∀𝑙 ∈ 𝐿 ∀𝑢 ∈ 𝑈 0 ≤ 𝜆𝑗𝑢 − 𝑟𝑗𝑢 ∀𝑗 ∈ 𝐽 ∀𝑢 ∈ 𝑈 𝑞𝑢 = 𝛾1𝑢 + 𝛾2𝑢 𝑟𝑗𝑢
𝑗∈𝐽
∀𝑢 ∈ 𝑈 𝑞𝑢 + 1 − 𝑔 𝛾2𝑢𝑟𝑙𝑢 − 𝜃𝑙𝑢𝑁𝐷𝑙𝑢 ≤ 𝑠𝑙𝑢𝐶𝐽𝐻 ∀𝑙 ∈ 𝐿 ∀𝑢 ∈ 𝑈 𝜆𝑗𝑢 − 𝑟𝑗𝑢 ≤ (1 − 𝑠𝑙𝑢)𝐶𝐽𝐻 ∀𝑙 ∈ 𝐿 ∀𝑢 ∈ 𝑈