The Most Likely Futures for US Refining

London Houston Singapore Moscow Dubai New York The Most Likely Futures Portland Calgary for US Refining Santiago Rio de Janeiro Beijing Ed Arnold Shanghai May, 2017 Tokyo Sydney Astana Kiev Cape Town Riga San Francisco Washington

Market Reporting Consulting Events High level of uncertainty out 5-7 years

Oil prices - Shale oil production trends - OPEC behavior - Asian and Latin American refined products demand MARPOL VI effects - Light-Heavy spread outlook US refined products demand - Domestic demand - Export demand - G/D demand ratio RFS and CAFE regulations Border taxes New pipelines Alberta bitumen

World Oil Demand, Supply and Cumulative Oversupply (Data Source: IEA, March, 2017 Update. Calculations by Argus.)

9.0 1,400

1,200 8.8 1,000

8.6 800

8.4 600 Million bbls Million 400

Billionbbl/quarter 8.2 200

8.0 0 Q1- Q2- Q3- Q4- Q1- Q2- Q3- Q4- Q1- Q2- Q3- Q4 - 2014 2014 2014 2014 2015 2015 2015 2015 2016 2016 2016 2016 Global oil demand (left axis) Global oil supply (left axis) Cumulative oversupply since Q1, 2014 (right axis)

Early indications are that oil was in oversupply in Q1, 2017.

Copyright © 2017 Argus Media Ltd. All rights reserved. 4 Most likely scenario: Sticking in the shale band Light Crude OIL Price History ($2016 basis, monthy average prices, through March, 2017) $/bbl $140 Shale Band: $120 $35 - $65/bbl $100 $80 $60 $40 ? $20 $0

Brent/North Sea Dated LLS Month 1

High probability that monthly average prices will remain volatile, and will stay within the shale band, until shale oil production can’t keep up with demand growth, or until the long-term effects of lower E&P investment take hold.

Most likely scenario: Impact on US refining will be significant.

Distillate Component of Bunker Fuel Resid Component of Bunker Fuel World Resid Demand (kbpd)

Unless a ship-based option is widely adapted, a significant portion of the ~3mn bpd of high sulphur resid used in bunker fuel will need to find a new home or be converted to low sulphur material Demand for low sulfur distillate will, most likely, increase significantly

Regional delays in implementation, temporary exemptions, uneven enforcement Plan for multi-year widening of light-heavy spreads High-sulphur distillate will be an undesirable product ULSD distillate demand will increase ULSD prices will rise No significant pre-emptive conversion or desulfurization investment Eventually, look for some additional resid conversion investment Slowly growing shift to marine LNG

Wild cards: Longer term L-H spread differentials*, prolonged phase-in for the regulations, uneven enforcement

* if this occurs, could incentivize major conversion investment

USGC ULSD - 3%HSFO Differential (fob, through March, 2017)

$80 $/bbl $70 $60 $50 $40 $30 ? $20 $10 $0

Initial switch will disrupt normal price relationships. Our Creating a stronger incentive for investment in scrubbing. current, most Longer term, prices should return to pre-2020 basis once more vessels are scrubber ready, and as alternative fuel use grows. likely scenario No significant level of preemptive refinery investment, beyond the normal, but will see some additional desulfurization and conversion → capacity. (If long term L-H spread diff, expect conversion investment.)

Light-Heavy Crude Price Differential [Brent & North Sea Dated versus Maya (USGC). Yearly a verage prices, through 2016] $16 $14 $12 $10 $8 ? $6 $4 $2 $0

Brent & North Sea Dated - Maya (USGC) Differential

But – most likely – not long enough to incentivize significant new US resid conversion investment Margins for refiners making HSFO will shrink

Breakdown of Current US Low S Distillate Use Industrial Use Utilities Oil Firms 3.3% 0.9% 1.9% Military Off-Highway 0.2% 3.3% Bunkers 4.0% Commercial Use 4.3% Farms 5.3% Railroads 6.0% Trucks, Buses & Residential Automobiles 6.5% 64.2%

Different options available Total installed costs (“TIC”) probably more expensive than most current P50 estimates ($3m to $6m++ for new vessels, and $6m to $10m++ for retrofits) - High level of uncertainty with current TIC estimates. More likely on new ships versus retrofits Space requirement is significant, costing revenue Caustic storage (fresh and spent) and other waste will be a major technical issue. Caustic disposal may be the key implementation-cost issue Can’t install fast enough to fit all large, viable ships If resid conversion or resid desulfurization investment is substantial, scrubber investors could be “burned” 12 US gasoline demand finally above 2007 levels

US - Yearly Domestic Gasoline Component Demand 10,000 ’000 bpd 9,500 9,000 8,500 8,000 Petroleum- 7,500 derived gasoline 7,000 demand not yet up to 6,500 2004 levels 6,000 5,500 5,000 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 Ethanol used in gasoline Gasoline and blendstock imports US-produced oil derived gasoline for domestic use -Data from EIA. Calculations by Argus.

7,000

6,500 US-produced oil derived gasoline for 6,000 domestic use 5,500

5,000

- Data from EIA. Calculations by Argus. Gasoline export growth has been accelerating the last 4 years Most likely export demand scenario: Demand growth rate will gradually shrink Most likely US demand scenario: RFS continues as is. Only moderate changes to CAFE policy. (“54.5 in 2025”). Demand growth shrinks and eventually disappears.

Million bbls 260 250 240 230 220 210 200 190 2001 - 2015 inventory range 180 2016 inventory levels 170 2017 inventory levels 160 1 3 5 7 9 11131517 1921232527293133353739 414345474951 Week Number

This places downward pressure on gasoline prices (and oil prices)

US- Yearly Domestic Distillate Demand 5,000 ’000 bpd 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

Biodiesel Distillate imports US-produced distillate for domestic use -Data from EIA. Calculations by Argus.

4,000

3,500

3,000

2,500

2,000 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015

Distillate exports US-produced distillate for domestic use

-Data from EIA. Calculations by Argus. 2016 growth in distillate exports could compensate for loss in domestic demand. Most likely scenario: Flat to slowly declining demand, until MARPOL VI hits, then look for tight market to develop.

2016 and 2017 US Distillate Inventory Levels

Million bbls

190

170

150

130

110

90 2001 - 2015 inventory range 70 2016 inventory levels 2017 inventory levels 50 1 3 5 7 9 111315171921232527293133353739414345474951 Week Number

This places downward pressure on distillate prices (and oil prices)

US - Yearly US G/D Demand Ratios

2.5

2.3

2.1

1.9

1.7

1.5 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 US G/D Domestic Demand Ratio US G/D Domestic + Exports Demand Ratio -Data from EIA. Calculations by Argus.

Most likely scenario: This ratio will head back to lower levels once MARPOL VI takes effect

US - Major Refined Products Exports

3,500 000 bpd

Unfinished Oils 3,000 Jet fuel & other kero.

2,500 Resid fuel oil

Pet. coke 2,000 Gasoline & Blendstocks 1,500 Distillate

1,000

500

0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Most likely scenario: This will continue, but expect decelerating growth

US- Export Growth Trends for Distillate and Gasoline are On Different Tracks (Rolling CAGRs ) 50% 40% 30% 20% 10% 0% -10% -20% Distillate NGLs & LRGs Gasoline & Pet. coke Resid fuel oil blendstocks 2012 - 2014 2013 - 2015 2014 - 2016

RFS - RFS stays at current levels - RINS policy does not change

CAFE Standards - No significant change - CARB has “clout.” - Car manufacturers do not want to build multiple vehicle types.

Most likely scenario: Minor, measured changes

Trump administration appears to be listening to US industry leaders

Regardless, high level of uncertainty

Both Dakota Access and Enbridge’s Keystone XL have new life Govt. of Canada has approved KM’s Trans Mountain pipeline twinning Govt. of Canada has approved Enbridge’s Line 3 “replacement” (enlargement) Probability of completion, for all 4 pipelines, within 2 - 4 years, higher than in 2016 But numerous lawsuits and hearings are obstacles on the paths to completion

Still face hurdles - Hearings at Nebraska Pubic Services Commission - Likely lawsuits: Sierra Club, NRDC and others Start of construction for KXL could be months or years off KXL less critical than when first envisioned - Other pipelines and rail transport have filled the gap Alberta Dilbit and Synbit will get to the USGC, regardless Supply of Alberta bitumen will continue to grow, slowly Most Alberta bitumen majors are in the business for long term - SAGD operations cannot easily be turned on and off - Despite the flight of foreign capital, ~$14billion in oil sands spending is forecast for 2017 - Recent permitting requests for multiple new SAGD projects

Canadian Oil Imports to US versus US Oil Production and Mexican and Venezuelan Imports 10,000 000 bpd 8,000 US Oil Production Canadian Crude Imports 6,000 Mexican Oil Imports Venezuelan Oil Imports 4,000

2,000

0 Jan-09 Jan-11 Jan-13 Jan-15 Jan-17

Majority of Canadian import growth is Alberta bitumen Alberta bitumen import growth will compensate for the loss of Mexican and Venezuelan imports

2,000 kbpd

289 PADD1 1,500 293 PADD5

1,000 PADD4

1417 1436 PADD3 500 PADD2

0 2009 2010 2011 2012 2013 2014 2015 2016

- Data from EIA. Calculations by Argus.

Total US bitumen demand dropped in 2016. PADD 2 demand grew slightly. PADD 3 demand dropped.

However, PADD 6,000 ’000b/d 3 represents 5,000 the largest 4,000 3,000 opportunity 2,000 1,000 0 Note the relative 2009 2010 2011 2012 2013 2014 2015 2016 stability of PADD 3 heavy sweet light sour light sweet medium sweet & sour heavy sour import heavy sour levels, since 2009 PADD 2 Oil Import Quality Trend by Oil Type

PADD 2 demand 6,000 ’000b/d growth is probably 5,000 close to its peak 4,000 demand level 3,000

2,000

1,000

0 2009 2010 2011 2012 2013 2014 2015 2016 Copyright © 2017 Argus Media Ltd. All rights reserved. heavy sweet light sour light sweet medium sweet & sour heavy sour There is room for additional Canadian bitumen in PADD 3

US PADD 3 Heavy Sour Crude Source Breakdown

kbpd Other 100% Saudi Arabia 80% 10% Ecuador 14% 60% Brazil 27% 40% Colombia

Canada 20% 32% Mexico 0% 2009 2010 2011 2012 2013 2014 2015 2016 Venezuela

Much of the Venezuelan heavy sour crude is high TAN and is a obvious candidate for “replacement.”

US Average Refinery Crude Oil Input Density and Sulfur Content API wt% S 32.5 1.6 1.5 32 1.4 Average Crude Oil API (left axis) 31.5 1.3 1.2 Average Crude Oil S level, wt% 31 1.1 (right axis) 1 30.5 0.9

30 0.8

...... as US oil logistics systems become less constrained, and as US light oil finds it way to additional export markets.

Oil Prices: Most Likely in shale band for multiple years Expect volatility Wild cards: Supply disruption, demand surprises, or OPEC successfully managing long-term production cuts MARPOL VI: Plan for multi-year widening of light-heavy spreads High-sulphur distillate will be an undesirable product ULSD distillate demand will increase ULSD prices will rise No significant preemptive conversion of desulfurization investment Wild cards: Long term L-H spread differentials (if they materialize, could incentivize major conversion investment), prolonged phase in for the regulations

US Gasoline Demand - Most likely that export demand growth rate will gradually shrink - Most likely that US domestic demand growth shrinks and eventually disappears - Wild cards: CAFE 54.5 disappears, RFS diminished, RFS increases

US Distillate Demand - 2016 growth in distillate exports was not enough to compensate for the loss in domestic demand - Most likely scenario: Flat to slowly declining domestic + export demand, until MARPOL VI hits, then look for tight market to develop - Wild Cards: Marpol timing delays, major jump in GDP growth rates in developing or developed countries

Pipelines & Alberta Bitumen - DAPL, KXL, Trans-Mount. twinning & Enbridge Line 3 all have new life - Most likely years of lawsuits and hearing for some of these - Additional Alberta bitumen will come into PADD 3 regardless (and a small additional amount into PADD 2) - Although the game has changed, Alberta majors are in the game for the long run - Wild cards: Successful law suits, major spills, technology breakthroughs

RFS and CAFE Standards - Most likely, RFS stays at current levels - Most likely, RINS policy does not change - Most likely, CAFE policy sees no significant change - Wild cards : Near-term political-control shifts in DC, Trump administration policy shifts

Border Taxes - Most likely scenario: Minor, measured changes - Trump administration appears to be listening to US industry - Regardless, high level of uncertainty

Ed Arnold, Senior Consultant [email protected] 630 544 6492 Houston www.argusmedia.com http://blog.argusmedia.com

Stay Connected @ArgusMedia Argus-media +Argusmediaplus argusmediavideo

Copyright notice Copyright © 2017 Argus Media Ltd. All rights reserved. All intellectual property rights in this presentation and the information herein are the exclusive property of Argus and and/or its licensors and may only be used under licence from Argus. Without limiting the foregoing, by reading this presentation you agree that you will not copy or reproduce any part of its contents (including, but not limited to, single prices or any other individual items of data) in any form or for any purpose whatsoever without the prior written consent of Argus.

Trademark notice ARGUS, the ARGUS logo, ARGUS MEDIA, ARGUS DIRECT, ARGUS OPEN MARKETS, AOM, FMB, DEWITT, JIM JORDAN & ASSOCIATES, JJ&A, FUNDALYTICS, METAL-PAGES, METALPRICES.COM, Argus publication titles and Argus index names are trademarks of Argus Media group.

Disclaimer All data and other information presented (the “Data”) are provided on an “as is” basis. Argus makes no warranties, express or implied, as to the accuracy, adequacy, timeliness, or completeness of the Data or fitness for any particular purpose. Argus shall not be liable for any loss or damage arising from any party’s reliance on the Data and disclaims any and all liability related to or arising out of use of the Data to the full extent permissible by law.