Business Management Assignment

11111//111

Illd

1/1

till

1/

/I III

II/ I\'

)( Ita Airlines and the Trainer Refinery Andrew Inkpen Michael H. Moffett

As creative as Delta might try to be with Trainer, it is hard to escape the feeling that a capital-intensive, highly cyclical commoditized business with a history of poor returns on capital would be buying ... more of the same.

"Delta Chases Fuel's Gold," by Liam Denning, The Wall Street Journal, April 5, 2012.

Reduction in refining activity in the Northeast, as reflected in recently announced plans to idle over 50 percent of the regional refining capacity, is likely to impact supplies of petro- leum products. The transition period as supply sources shift could be problematic for Ultra- Low Sulfur Diesel (ULSD), gasoline, andjetfuel supplies. Prolonged uncertainty over the coming months with regard to the disposition and operation of important logistical assets such as pipelines, ports, and storage would compound adjustment challenges. Reduced short-term product supply flexibility due to longer delivery times and potential transporta- tion bottlenecks for sources outside the region could also increase price volatility.

"Reductions in Northeast Refining Activity: Potential Implications for Petroleum Product Markets," Energy Information Administration, December 2011.

1 In late April 2012, Richard Anderson, chief executive officer of Delta Airlines (NYSE: DAL), and his acquisition team were evaluating one last time the proposed acquisition of the Trainer Refinery in Philadelphia. Delta had been seriously negotiating and pursuing the purchase of Trainer, owned by Phillips 66 (NYSE: PSX- WI), for nearly a year. Although the team believed that owning Trainer would allow Delta to manage its rising jet fuel costs- particularly in the Northeast Corridor-the analysts, markets, and press had categorically labeled it Delta's folly. The time to decide was now (Appendix 1 identifies the limited his- tory of vertical integration into energy by other large non-energy companies).

LTA'S DILEMMA

2 Delta's problem, the same problem suffered by all airlines, was fuel cost. Fuel costs aver- aged anywhere from 30 percent to 50 percent of total operating costs in the airline industry, and crude oil and jet fuel costs had been on the rise. Delta had been hit hard in 2011, as illus- trated in Exhibit I, as jet fuel prices had risen. Delta's total fuel expenses had risen by nearly $3 billion in 2011 (34 percent) over those of 2010, and 2012 was looking even grimmer.

3 Delta was an airline on the rebound. It had closed 2011 with $35 billion in revenue, up 10 percent from 2010, with profits up 40 percent to $854 million. It was driving profitabil- ity by flying fewer planes fewer miles with fuller seats. It had 80,000 employees worldwide and $3.6 billion in cash. It was the world's largest airline in terms of both fleet size and scheduled passenger traffic. And jet fuel costs were killing it.

THUNDERBIRD Source: Copyright © 2013, Helen Grassbaugh, Thunderbird School of Global Management, [email protected]. All rights reserved. This case was prepared by Professors Andrew Inkpen and Michael H. Moffett for 11'1 purpose of classroom discussion only, and not to indicate either effective or ill ·n" .rive 1l1111lllg '111 'Ill.HOO 01 I OUA

2009

2010

2011

3,853

3,823

3,856

$ 8,291 .$ 8,9G);11'

$11,783

$ 2.15

$'2.33

$ 306

I()I tilt' 111111,11111 tilt ill( h 1'.\1 Y

• DuPont's pur IIi.1 01 all 0, 1 !31. lhl WtI'> .I tually till' Il",lill 01 d I,\ldINll( lil'l(,II',(' "<1.1111',101 \1n III by Seagram's, but was close to a true vertical int gratlon a qui ilion, COIlO(Owould PIOVUit, (ilidl (II natural gas feedstocks for much of DuPont's chemical businesses. The two p rtod w~y ill 19(1H

• U.S:.3teel.'sacquisitiOq of Mara~hon refinin)g, 1982. liihliswas alii effort to diversify U.s. Sto I'~ 1)11,>"1(',, un The company holding company, USX, eventually divested the steel businesses in 2001, r "_lInlllq 1111 r I company Marathon Oil in 2002.

• The Kansas farm cooperative, Cooperative Refinery Association, purchased a refinery own d by IIH' N II Refinery Association in 1944 to provide diesel for its own consumption needs. The refinery, ori9ill"lIy blilit I 19p6, continues to operat~t9day. The refi,rnerywas purchased by Coffeyville Resources LIi.C in 2000,

• Private equity and investment banking interests have accumulated larger and larger interests in crud: II tion, storage, and transportation over the past decade. These financial firms are quickly becornlnq I Iqnlll owner of oil and gas assets in the United States.

EXHIBIT 1 Delta Airlines, Fuel Cost Summary

Year Gallons

(Millions) Cost

(Millions) Avg Price

(US$/gallon)

%

Source: Delta Airlines, 2011 Annual Report, p. 9.

THE STRUCTURE AND PROFITABILITY OF U.S. REFINING

4 The refining industry in the United States is defined regionally by Petroleum I\dllllill I tion for Defense Districts (PADD), a system put in place during the Second Wllrld The five districts, described in Exhibit 2, differ significantly over the three major' 1111'11 refining profitability: (1) source and cost of the crude oil input; (2) the scale, VilllllVI

EXHIBIT 2 U.S. Refining Capacity and Operations by District

Fuel Ca,pacity Percent P~ima~y Sour PADD District Refineries (bbl/day) '1 of Crade 011of Total

1: East Coast 12 2,153,000 12% North Africa, We t I\llIe

2: Midwest 25 3,579,000 19% Canada, US

3 Gulf Coast 44 8,802,100 48% Southern US, Gulf 01 M 4: Roc:kyMountains 15 614,750 3% Canada, US

5: West Coast 27 3,251,200 18% Alaska, Californi

123 18,400,050 100%

Source: "The U.S. Oil Refining Industry: Background in Changing Markets and Fuel Policies," Anthony Andrews, Robert Pirog. IIlId l'I'tl,ll Sherlock, Congressional Research Service, November 22, 20] 0, and author analysis. PADD-Petroil:llllilldillillistration rill' I)dllllNl' I )I~IIIII

I

)1 d

PADD 3: Gulf Coast

PADD-Petroleum Administration for Defense District.

complexity of the refinery facilities themselves; and (3) the product markets the refineries served. As a result, the different refineries across the districts experienced very different profitability over time. A map of the five districts can be found in Appendix 2.

CRUDE SOURCING

5 The five different refining districts had distinctly different sources of crude oil. The East Coast relied heavily on imported crude oil, much of it from West Africa and North Africa. The Midwest and Rocky Mountain districts historically relied on crude oil via pipeline from Canada, but the recent boom in shale oil in the northern Rocky Mountain states had provided the district with a new, larger, cheaper, and closer source of crude oil. The Gulf Coast district, the core of U.S. refining today with nearly half of the nation's refining capacity, sourced crude oil from the southern part of the U.S. (largely Texas, Oklahoma, and Louisiana), as well as the U.S. Continental Shelf in the Gulf of Mexico. The West Coast district purchased oil domestically (Alaska and California) and internationally (Asia).

6 The cost of crude oil acquisition was the cost driver in refining economics. Although there were a number of different benchmark crudes, the two most prevalent in the Western Hemisphere were Brent Blend (an index of North Sea crude oils) and West Texas Interme- diate (WTI, an index of oil traded via Cushing, Oklahoma). Brent Blend had grown to be a dominant price benchmark in recent years as the global crude oil market became more and more liquid, and was often characterized as water-borne as opposed to WTI's land-based pricing. Brent Blend and WTI had historically served different refining markets, but given competition and the growth in refining demands over time, their prices rarely differed sig- nificantly. As illustrated by Exhibit 3, however, this had changed. Beginning in the summer of 201 0, Brent Blend prices had traded at varying premiums over WTI. By March of 20 12, HI' .nt wa. trading at $12Slbbl to WTf's $99, a spr ad of $26/bbl.

L- ~ ~________ __ _

$140

$120 +----------------{----J:----·b-

$40~-------------------------------4~~~----------------------~ $20~--------------------~----~----~----------------------------~

___ West Texas Intermediate (WT!) -+- Brent Blend

Source: Calculated by authors. Spreads shown are the differences between the monthly average price per barrel of Brent III 'lid Iliid W II Gulf Coast Kerosene-Type Jet Fuel prices (FOB) as collected by the Energy Information Administration.

7 The spread between prices was driven by growing supplies of domestic oil ill 1111'11111 States and its inability to gain access to major refining centers like the U.. EUsl( '1111I Gulf Coast districts. The rapid development and production of shale oil froIII tllIlIl sources like that of the Bakken Field in the Dakotas was landlocked. Pip 'lilli' I capacity, and oil was stockpiling at crossroads like that of Cushing. Tran port.n hili 1111I tives like railroad were more costly. The result was a discounting of dom 'sl k I crude oil.

8 This flowed directly through to sourcing costs. As illustrated in Exhibit 4, II, 11'1III as January of 2010 the cost of crude oil acquisition across refining districts WII 1111 the same. But by the latter half of 2011, and decisively by the spring of 2012, Ilh' I II I crude were diverging dramatically. Crude acquisition costs in the Midw SI, WIII'II'I I tively cheap inland crude from the mid-continent and Canada had steadily I''plllli II II more expensive crude oil piped from imports and coastal sources, now enjoy xl II ,~Irt/I discount to the Gulf Coast. This was a dramatic change from the $lfbbl PI' mium II It paid as recently as the spring of 2010.

9 The Gulf Coast refinery district, the largest and most highly developed disll'i('1 III II US. complex, had traditionally enjoyed a number of crude oil acquisition cost Ut!VIIIIIII including a mixture of inland, offshore, and imported crude oil sourcing options. Till' ( III Coast had been usihg greater and greater volumes of imported oil, primarily WeslAl1 11111 which was priced on the basis of Brent Blend. As a result, crude acquisition costs W"II' II ing relative to the other refining districts.

10 But it was the East Coast district which was now suffering th highest crud' 1I'qld 1111111 costs. Although East Coast refineries had purchased crude oi I IIIIh ' same pri . 'S lIN III\' II

100~ ~~~-+~~ ~~~~~-1

90i---------------~------~~~~~--------~~~----~

70t- ------------------------~~~~~----~------~

60~----_.------_,------,_--~_,------._----_.------,_----~ ~1lJ1-10 Jarn-11iApr-l1 Ott-11Jan-10 .Aipr-10 Oct-H)

-- East Coast -- Gulf Coast -- Midwest

-- Rockies - - - - US Average

Source: "Diverging Trends in Regional Crude Acquisition Costs," This Week in Petroleum, EIA, January 25, 2012.

average and Midwest refineries as recently as the spring of 2010, by late 2011 they were paying a premium of more than $lO/bbl over the U.S. average, and more than $16lbbl over the Midwest, as noted previously. By the spring of 2012, the East Coast refinery complex was suffering the highest crude oil acquisition costs in the nation-by far.

Refining Complexity and Vintage 11 The second driver of refining economics and performance was the scale, vintage, and com-

plexity of the refineries themselves. Refining technology had made major advancements over time, with newer refineries typically being of greater size and efficiency.

12 One of the primary methods of categorizing refineries is complexity, the characterization of the plant and equipment it is operating. I The three categories-simple, complex, very complex-distinguish the capabilities a refinery possesses to both utilize a greater range of crude oil qualities and to produce a greater range of refined products from those crudes. Older refineries, or refineries possessing older technology, may be capable of processing only a specific type of crude, and may have little flexibility in the products they may pro- duce from that crude. Newer technology, which of course would also imply higher initial capital costs and potentially higher operating costs, were more flexible in both crude oil inputs and refined product outputs.

'One measure of refinery complexity is the Nelson Index. The e1son Index assigns numerical values to the multitude of refining technologies and equipment. A simple refinery would possess crude distillation, cat reforming, and hydrotreating distillates, for a value between 2 and 5 in the index. A complex refinery would have simple refinery capability plus a vacuum flasher, cat cracker, alky plant, and gas processing capability, a Nelson Index of between 5 and 12. A very complex refinery would be a complex refinery plu a coker, and a NIIIHOIIs 'or of" 13 and above ( ee Appendix 5 on page 19-9).

10111I

110111.1

H()%

10%

60%

50%

40%

30% 30.9

20%

10% 9.8

West Texas Ara Intermediate

I_ Gasoline Source: Based on average Gulf Coast winter yiel Handbook, 2007, Energy Intelligence, www.ener

premium product? A barrel of crude oil, d~ produce only 10 percent or 20 percent of r

16 These products are themselves cone gional, and international markets." BlL complexity, and reliability of suppliers product buyers come in. As seen in ~xt the five refining districts in the spnng worst in the Northeast, and had led to refineries which were at the center of r::

EXHIBIT 5 u.s. Refining Margins, 20: Week Northeast Midwe;

March 16 8.51 29.43

,March 23 10.91 29.57 Source: "U.S. Refinery Margins Gain 6 pct-Cree

"Ih'hllklilly, jet fuel is generically considered a /I I I, III'11Iljl11l1l-k.roscne (B). Jet A is available II I,d Iliidl'IIiINlIl1wdint rnationally. Jet-B is a m= 1/l1'IIIIIIIIIItIlIVI IIIth"OI'yU 00 I whose price i I' 1.11111111111111,'1wl111'11IIlrr~\'(;lIlialcil from th. 1111111Ihlll,1111111111I 1,'1 lIPPI'I,whiL'his 99% pu 11111I Ilidl lIiI, II III1II. 11111111111111'III1I11'1'l'l''nt orga

R fin ry Iy t I rl f t tu ExxonMobil Baytown Rcfin ry Baytown IX GlJl'f (Ocl I 3, 'I'Vo OPI'IIIIIII J

ExxonMobil Baton Rouge Baton LA Gulf Coast .I1Vo OPIIIIIIIII Refinery Rouge Hovic Refinery Kingshill, St. VI East Coast 500,000 2.7% (Io~(ld I t

Croix 201)

BP Texas City Refinery Texas City TX Gulf Coast 475,000 2.6% Operannq Citgo Lake Charles Refinery Lake LA Gulf Coast 440,000 2.4% Op rc1IIIHI

Charles Marathon Garyville Refinery Garyville LA Gulf Coast 436,000 2.4% Op rllI.lnq

ExxonMobil Beaumont Refinery Beaumont TX Gulf Coast 345,000 1.9% Oper linq

Sunoco Philadelphia Refinery Philadelphia PA East Coast 340,000 1.8% To be 10.,( d J 2012

Chevron Pascagoula Refinery Pascagoula MS Gulf Coast 330,000 1.8% Operating

Valero Corpus Christi E&W Corpus TX Gulf Coast 315.000 1.7% operatlno Complex Christi Total, Top 10 4,261,000 23.2%

Total U.S. Refining 18,400,050 100.0%

Source: "The U.S. Oil Refining Industry: Background in Changing Markets and Fuel Policies," Anthony Andrews, Robert Pirog, and MIIlh I Sherlock, Congressional Research Service, November 22,2010; updated by authors.

13 The Gulf Coast district dominated the U.S. refining industry, with relatively higher 1'11111 plexity and scale. It was home to 48 percent of the total capacity of the U.S. industr , 1I11t1 eight of the ten largest refineries. Only two of the largest refineries in the nation \ \'1\ outside the Gulf Coast, and one of those, the St. Croix refinery, had closed in 20 I I (, 'I Appendix 3). The top ten facilities alone made up more than 23 percent of total caplli'il \ These larger and generally newer facilities enjoyed a number of operating benefits, ill 'I"d ing lower production costs and greater product mix flexibility (see Appendix 3.)2

Refined Products 14 The third and final driver of profitability in refining is the nexus of the products gcn '1'1111'11

by any individual refinery and the market for those products. The more valuable the pnu! ucts a refinery could make, the more it could pay for its crude oil.

15 Different crude oils generate different relative proportions of derivative products-the ga« ,IllIt•• kerosene Get fuel), diesel, gas oil, lube oil, and other residuals. But the differing product yi Ids III also a function of the refinery's individual technology and complexity, and the setup of the I' 'filllli \' by operating management (see Appendix 3 for a numerical example). Jet fuel was one of th ' 111111111 derivative products resulting from the refining of crude oil, and was considered somethin 1 (II I

'Although there have been no new refineries built in the U.S. in more rhan 40 '11I'S,Ihls 1)1'1.n-norc I obx 'rvIIIIIIII is somewhat misleading. The major oil refining companies have und 'rlllkl'\111IIItIIIlit1111'III'mnssivc r 'invl',_IIII\'1i1 efforts in the old facilities over time, making ad finiiion of' U! ill 1\11111111'II111'111IIIIPlIIINlhlIIISk,

HO' ,

60%

50%

40% 24.7

30% 30.9 39.641.620%

10% 23.7 9.8

Arab t.iqht Niggrian Bonny Light

DFueloil1

West Texas Intermediate

I_ Gasoline IIIKerosene/Jet 0 Diesel Source: Based on average Gulf Coast winter yields. Data drawn from The InternationaL Crude Oil Refining Handbook, 2007, Energy Intelligence, www.energyintel.com.

premium product? A barrel of crude oil, depending on what kind of crude and which refinery, may produce only 10 percent or 20 percent of its output as jet fuel (see Appendix 4).

16 These products are themselves considered commodities, which are sold into local, re- gional, and international markets." But the actual buyers are still confronted with costs, complexity, and reliability of suppliers. That is where Delta Airlines and other derivative product buyers come in. As seen in Exhibit 5, refining margins differed dramatically across the five refining districts in the spring of 2012. The margins per barrel were clearly the worst in the Northeast, and had led to the refinery closures. It was these same East Coast refineries which were at the center of Delta's problem.

EXHIBIT 5 u.s. Refining Margins, 2012 (US$/bb\) Gulf Coast RockiesNortheast MidwestWeek West Coast

March 16

March 23

28.98

29.70 31.79

33.82

21.31

20.00

8.51

10.91

29.43

29.57 Source: "U.S. Refinery Margins Gain 6 pet-Credit Suisse," Reuters, March 26, 2012.

"Technicaliy, jet fuel is generically considered a kerosene, and then classified as an unleaded kerosene (A or A-I), or naptha-kerosene (B). Jet A is available and used only in the United States, while A-I is the standard used and consumed internationally. Jet-B is a mixture used only in very cold climates. 'A commodity is in theory a good whose price is set on an open market, possessing no characteristics in the 'y s of' the market which differentiate it from the same product produced anywhere else by anyone else. By this 1I"lillilioll, II III ial like copper, which is 99% pure copper by definition and specification, is indeed a commodity, 111111"11111'oil, with ils multirudc ofdiffor nt organic 'oI11POIl()IlIS ilnd obviously dilf 1"11111111,'1( '\ pri 'e,s,ls IIOt.

L._~ ~ -~---

UI'lhll'II,1II III 1I:/IIIIlIg 11I'1/1'//1' III IIII' NII/IIt"II,I/, //,\ 11'/It'I'II't! III 11'1'1'11111'1IIIIItli/II

III kll« III'I'/' 50 111'1'1'1'111'!!"II' I'I'glllllllln:/illlllg 1'llfIlIl'III', i,\' fIA,'/I' III 1111111/1'11/11'1I troloum pmducts. 'llu! transition period II,\' ,\'111111/.1',1'111/1'1'1',1'",hil/mllltllll' fllllhl III Ultra-Low Sulfur Diese! (Ul.Str), gnsotlne, utuljetjuo! ,I'IIIIIIIII'.\'.

"Reductions in Northeast Refining Activity: Pot .ntial lmpli 'aliollN 1111 I' 1111 Product Markets," U.S. Energy Information Administration, I "'lnlll'l 1111 I

17 Although the 2004-2007 period was, in the word of one analyst, th ' .. 'oldl'lI I ing, the years since 2007 had been less golden. Due to the run-up ill 'I'lIdl' 01111 first half of 2008, with the collapse of commodity prices and the ens ,t of till' 1111111 in the second half of 2008, refining profitability fell.

18 For the refineries operating on the East Coast of the U.S., much of 111'11 I depended upon gasoline. But the increase in gasoline prices in 200H, 1'0110 I nancial crisis-induced recession, had caused a substantial drop in gasoline 'Oil the United States. This fall in the demand for the primary product of East ('011 I many of which are the smaller and older facilities with higher operati ng '(I~I 111\II uct mix flexibility, yielded a series of facility closures. Exhibit 6 provide II III the operating-and idled/closed-refineries in the PADOI East Coast district 0111

EXHIBIT 6 u.s. East Coast PADD District 1 Refineries Capacity Percent

PADD East Coast Facilities City State (bbls/day) of Total 1-01 Hovic Refinery Kingshill, St. Croix VI 500,000 23%

1-02 SunocOPhih'ldelphia Refinery Philadelphia PA 340,000 16%

1-03 ConocoPhillips Bayway Refinery Linden NJ n8,000 11% -04 P>etropll!JsDelaware CitY' Refinery Delaware. City DE 2QO,'00O 10%

1-05 Valero Paulsboro Refinery Paulsboro NJ 185,000 9%

1-06 ConocoPhillips Trainer Refinery Trainer PA 185,000 9%

1-07 Sunoco Marcus Hook Refinery Marcus Hook PA 175,000 8%

1-08 Chevron Perth Amboy Perth Amboy NJ 80,000 4% Refineryfiferminal;

1-09 Amerada Hess Port Reading Refinery Port Reading NJ 70,000 3%

1-10 United Refinery Warren PA 70,000 3%

1-11 Western Yorktown Refinery Yorktown VA 70,000 3%

1-12 Ergon Newell Refinery ,. Newell WV 20,000 1%

1-13 Bradford Refinery Bradford PA 10,000 Q..%. 2,153,000 100%

Source; "The U.S. Oil Refining Industry: Background in Changing Markets and Fuel Poli 'j 's:' Anthony Andrews, Roil 'rl I'IIIIP, 111111 F. Sherlock, Congressional Research Service, November 22,20 I 0; updated by authors.

~20+---------------------------------------~-------t~----. $15+-------------------~------~~~_=----~--~~ $10~r_--------------------~------~~--_+--r, $5 $0

?>t;)?>t;)Pi" ?>" ?>'ij,?>q,?>""?>~ ?>~ ~~'?>~?>~ ?>tQ?>f¢?>'" ?>'" ?>'b?>'b ?>OJ'!>JOJs:P r:P10" s»"s:;,"'s:;,"'s:;,'"s:;,"',h::h:~~ !&s»~~Cos5' s:;," s:;,'bs:;,"I;>s:;,OJ~OJ "t;) "t;),," ,,"~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~v~~~~~~~~~~~~~~~~~~~O~O~O~O~O~O~&~O~O~O~O~&~O~O~O~o~&~&~if~&~o~o

Source: Calculated by authors. Spread shown is the difference between the monthly average price per barrel of Brent Blend and the Gulf Const Kerosene-Type Jet Fuel prices (FOB) as collected by the Energy Information Administration.

States. The combination of closure and idling (planning for closure if not resold) was radi- cally changing the East Coast product markets.

• In September 2011, Phillips idled the 185,000 bpd Trainer Refinery in Philadelphia. The primary explanation was the facility's lack of profitability.

• In December 2011, Sunoco shuttered its 178,000 bpd Marcus Hook refinery in Pennsylvania. Again, the official explanation by Sunoco was lack of profitability.

• The Hovic Refinery on St. Croix in the U.S. Virgin Islands, a 500,000 bpd joint ven- ture between Hess Corporation and Petroleos de Venezuela (PDVSA), technically also serving the PADD 1 East Coast district, was shut down in February 2012. It will be converted to use as a crude oil storage terminal only.

• Sunoco's 340,000 bpd Philadelphia refinery, the largest operating refinery in the East Coast district after the closure of the Hovic St. Croix facility in February 2012, would be shut down in July 2012 if not sold. A number of private equity firms, including the Carlyle Group, were rumored to be interested. Again, the reason for closure was lack of profitability.

19 Three of the four closures were refineries operating in the Philadelphia area, and the three were also the major providers of gasoline and jet fuel in the Northeast, specifically the Philadelphia and New York areas. The four closures totaled a full 50 percent of the refining capacity within the East Coast district. The closures had been the subject of much debate within the region, cutting hundreds of jobs, reducing state tax revenue bases, and ul- timately requiring businesses across the region to search out new sources of the three crude oil derivatives produced at the plants-Ultra-Low Sulfur Diesel (ULSD), gasoline, and jet fuel. Gasoline imports as a result of refinery closures in the East Coast district are shown in Appendix 6.

I_ USVls _ Canada 0 W. Europe 0 Other I 2011

Source: "Potential Impacts of Reductions in Refinery Activity on Northeast Petroleum Product Mark 'IN," Energy Information Administration, February 27, 2012.

20 Delta's issues were typical of the airline industry. More than 80 percent of airline opt'11I1 II costs were fixed or semivariable. The only costs that varied per passenger were res 'I'vill I III system commissions, food costs, and ticketing fees. The operating costs of an airlin ' Iliphl depended primarily on the distance traveled, not the number of passengers on board, hll example, the crew and ground staff sizes were determined by the type of aircraft, 1101 1111 passenger load. Therefore, once an airline established its route structure, most of its OPtl1 II ing costs were fixed.

21 The industry had one of the lowest profit margins when measured against other uulu tries. Airlines were far outpaced in profitability by industries such as banks, health 'III,', consumer products, and energy. In recent years, a la carte revenues such as bagga 1 ' I~I, and change fees had become increasingly important for most of the airlines.

22 Although reducing operating costs was a high priority for the airlines, the nature I II the cost structure limited cost reduction opportunities. The large airlines' restrictiv ' UIIIIIII agreements usually limited labor flexibility. The airports controlled gates and landing ri "II and had local monopoly power. Maintenance was largely determined by the equipment pill viders, such as Boeing and Airbus. Finally, jet fuel, the largest single cost, averaged about 30 percent of total operating costs across the major airlines.

23 Delta had struggled with profitability for a decade. As shown by Exhibit 7, Della 111111 suffered losses continuously from 2002 to 2009, finally turning a profit in 2010 and 20 II Although total operating revenues had risen from $14 billion in 2002 to over $35 billiun in 2011, operating and financing costs had prevented profitability until 2010.5 The airl ill I'

DELTA, JET FUEL, AND THE TRAINER REFINERY

25

26

H 27 MI

IWII

'Delta acquired Northwest Airlines in October 2008; financial results III" '\1111111111111 1'111' 1111 'ms h 'gilll1illll 00 I

$20,000

$15,000

$10,000

$5,000

-$5,000

-$10,000-t------,---,-------,---,-------"j'----,-------,--,------,-""''''---l 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

I_ Operating Revenue _ Fuel Expense Net Income 1 Source: Delta Airlines. Financial results include Northwest Airlines acquisition beginning in 2008.

largest operating expense was salary and related costs every year leading up to 2005, when for the first time fuel expenses surpassed salary costs. Exhibit 8 shows how fuel expense had grown in size and significance since that time.

24 Hedging jet fuel costs had long been a subject of considerable debate in the airline in- dustry. Southwest Airlines had long been famous for reaping regular benefits with its fuel hedging program, and Delta and others had periodically hedged their fuel cost exposures with varying level of success.

25 In recent years, particularly the oil and jet fuel price run-up in 2007 and 2008, many airlines had instituted fuel surcharges on ticket sales, passing on the higher costs directly to the customer. Historically, this had been considered a non-starter strategy, but the rapid rise in prices of the time had made it a common practice across much of the airline industry. It was still an option open to Delta, but given the relatively quieter times in crude oil prices (although at $100lbbl, crude oil was still not back to historical levels), leadership was afraid that customers would not react positively to a return to cost pass-through.

26 Not all airlines had been equally successful at passing along those higher fuel prices in 2008. The 2008 jet fuel price peak had also seen the failure of five different major U.S. air- lines." In addition to the higher fuel prices, all five airlines were carrying very large debt service burdens, and with a decline in air travel and an increase in fuel costs, they failed quickly.

Hedging Jet Fuel Expense 27 Much of the debate over fuel hedging had been an inability to agree upon the objective of

hedging. There were two streams of thought on the objective: (1) to lock in jet fuel costs

"Four of the five airlines filed for Chapter 11 bankruptcy within three weeks of the price of crude oil reaching $1001 bbl: Aloha Airlines, March 20, 2008; ATA, April 2, 2008; Skybus, AprilS, 2008; Frontier Airlines,

]111111,2008; and harnpion Air, May 31,2008.

L- ~

40%

35%

30%

25%

20% $1.16

15% $ $0.820.65 $0.69 $0.67

10%

5%

0% 2000 2001

Av r

$1.79 $2.12

Prl

$2.24

$3,16

Source: Delta Airlines.

$2,1 $

over time which were cheaper than the costs of buying the fuel regularly on th ' spill III ket; or (2) to lock in a known cost of fuel 12, 18, or even 24 months into the future III 11111 to gain increasing predictability and control over corporate cash flows.

28 Jet fuel hedging instruments (derivatives) were only available on the over-the ('IHIIII (OTe) market, where an individual counterparty would write the contract and a ,t Il~ 11111 terparty to the airline. Although these contracts and derivatives could be custom-dcxi II to the buyer's need, they were more expensive and less liquid than exchange-trad 'd Ihllil cial derivatives. Exchange-traded derivatives, like those traded on the New York M 'n' 111111 Exchange (NYMEX), were highly liquid, but unavailable for jet fuel. They were avniluhl however, for crude oil (WTI) and heating oil. In some ways, jet fuel hedgers thercf 1I (' hll I to trade off hedging effectiveness for hedge liquidity." As illustrated by Exhibit 9, althuu I jet fuel and crude oil prices were highly correlated, they were not the same,

29 The primary derivatives available were swaps and options. Fuel swaps were cont I11 1 which may extend over a one- or two-year period in which the two counterpartics wouhl exchange payments, one fixed and one floating (spot). Under the most basic of these 'ilil tracts, the plain vanilla swap, the airline would typically be the fixed-price paying pnuy, allowing it to pay a fixed price for fuel monthly or quarterly for the duration of the Ugll'I' ment. In return, it would receive a floating or spot price payment in return, therefore SWllp ping payments. The obvious benefit to the airline was that it had known payments for 1'111'1 expenses for the duration of the agreement. The risk, however, was that it was locked 111111 this payment rate regardless of whether jet fuel spot prices were higher, the same, OJ' 'VI'II lower than the fixed price. The typical swap agreement required no up-front payment, IIl1d was settled on a real-time basis over the life of the contract.

'This liffcrcntial between a jet fuel exposure and a crude oil hedge is the risk rhat the exposure's value will nul chang exactly in correlation with the value of the hedge, Termed basis risk, il IIIIIy Ill' ~111,,1I or large dcpcndillf' on the dl'/"'" of' .orr '/"Iion h 'Iw' 'n.i 'I fuel price Hnd .rudc oil pri 'C' d""ll'c'~

~1 I I

,,, ~140 +--_._-

~120+------- ~100~------~------------------~--~~--~--------------------~4-----------~

$40+-~----------------------~----~----__~~~--~----~~--------------~

ource: Calculated by authors. Spreads shown are the differences between the monthly average price per barrel of Brent Blend and WTI and the Gulf Coast Kerosene-Type Jet Fuel prices (FOB) as collected by the Energy Information Administration.

30 Options are derivative agreements which allow the holder to either buy an asset at a set price over a period of time (a call option), or sell an asset at a set price over the term (a put option). As opposed to a swap agreement, an option is just that-the right but not the obligation to exercise the rights provided by the option contract Delta could purchase a call option or series of call options on jet fuel which would effectively assure it of the price it would have to pay over the contract term. This would assure Delta of a cap on fuel purchase costs. If, however, fuel prices were to fall precipitously, Delta could purchase its fuel on the cheaper spot market and allow its call option derivatives to expire.

31 Options, however, have an up-front purchase price, the option premium, which the swap agreement does not Option premium values are directly impacted by market volatility. Therefore, during periods of perceived crude oil and jet fuel price volatility, option premi- ums can become quite expensive. One method for financing these call option purchases is the simultaneous sale of put options on the same fuel amounts, but at a somewhat higher price than that of the call option. This combined position, an option collar, can be self- funding (a net zero premium payment up front, as the call option premium expense is equaled by the put option premium earned). The final hedge position effectively bounds fuel price costs between a maximum cost set by the call option strike rate, the cap, and a minimum ~et by the put option strike rate, the floor.

32 Another strategy which had periodically been employed by airlines was a form of market-timing. This was a selective hedging strategy, requiring the company to time it hedging activities in regard to the current spot price of crude oil/jet fuel relative to what it h licved was the long-term average-a fundamental mean.

$7 WTI (US$/bbl)

$6 1/

$5

$4

$3

$2

$1

$0

Source: Energy Information Administration and BP Statistical Review.

Trainer's Potential 33 Crude oil prices had risen over the past several yea.rs, and many market analysts h'lI

higher prices were here to stay. The spread between jet fuel prices and crude oil pll was rising. Hedging the differential was difficult. If Delta owned its own jet fuel slIppll whatever it lost in higher global jet fuel prices it gained from the increasing prohlllill ity of that same jet fuel sold by its own fuel subsidiary. Many analysts, however, III II that the margins potentially earned in the refining segment were driven by a multiuul I factors-as seen in the earlier sections, many of which are different from simply 11ll' pi I of the specific crude oil inputs. As illustrated by Exhibit 10, U.S. refining margins wet 11I1 consistently correlated with WTI price movements.

34 How critical was Trainer itself? The Trainer Refinery could produce more Ihlll 23,000 bpd of kerosene/jet fuel, roughly 20 percent of all the jet fuel production on IIH'I 'II I Coast. If Trainer was allowed to close (permanently), the East Coast market would he '0111 that much more isolated, and Delta's jet fuel costs were likely to suffer as a result. MOl and more of their fuel sourcing would have to be done from either the United Kingdom 01 the U.S. Gulf Coast-both of which would be more expensive.

35 And it isn't clear yet how those deliveries would take place. Pipeline capacity is sruu between the Gulf Coast and Northeast (the primary pipeline, the Colonial, is current! IIp erating at near capacity), as well as the availability of intercoastal shipping required L1II1II'1 the Jones Act." Even if those barriers could be overcome, and after a period of pos ibl 'ill ruption in supply, these sources would still likely suffer higher transportation and d 'Iiwi v charges, as well as a less-secure ongoing supply of the critically important fuel.

'The Jones Act, the Merchant Marine Act of 1920, requires that all shipments OCCIII'l'il'I' between U.. plH'IH, termed cabotage, be performed by U.S. flagships, using ships constructed ill 1111'II,S .• IIII'll 'd by U.S. 'itlNII , and crewed by U.S. citizens and permanent residents. The result was III'Ilstly 1I1i1t1II III 1'1vice sc ·tor rill' intercoastal shipping.

-----~-----T----------~--~~y __ ~~~~·~I '~ ••- __ ~_~~~ •• II lip I 1111"I III,' ('111111I111.hilt 1111'1,III "1,,i·IIII'·I'1III11\11'I1It 1llIIltl'd tho l' NIIl'l'lll'l ,lllhl'

1111111'1III' tilt, IIIIIld flllllll'('lts l'I'lilil' 1111(111'1111''lit 'I' PI'OPOI'lioll) 1'1'0111low 'I' 'oN11i1t '1'1Il!- I \' Iltll I I IlIIh' II IN pl'i ' 'd Oil (II ' husis of' W ·S(T 'x as lilt '1'111xliatc like the Bakken rude 1111111NIlIIIi I) 11,"111,Train 'r's cost competitiveness would increase dramatically.

7 Phillip (Ill's IIsl inn I rice was $180 million, little more than the price of a single new IIlIl'i "I' III wid 'body, The state of Pennsylvania had offered an additional incentive of $ I() million to reopen the refinery and put hundreds of workers back to work. But Trainer was an old refinery and would require new investment, maybe $100 million, to gain more efficient production as well as increase the yield on jet fuel--which was definitely desired. Delta's refinery consultants estimated that the investment in retooling the refinery could increase the jet fuel yield from 14 percent to about 32 percent of total output; others argued that a high of 26 percent or 28 percent was more likely. The consultants had also recom- mended that Delta arrange to exchange or swap the other refinery outputs from Trainer for jet fuel at other key locations in the Delta grid, reducing Delta's need to also market the refinery's products. That assumed, of course, that Delta could put together its owner management team for the refinery and rehire most of the workers who had been laid off the previous fall.

DECISION TIME

38 Richard Anderson decided what was needed was a full-blown final debate before mak- ing the final decision. He divided the analysis team into two groups, the For's and the Against's. Each team would have 20 minutes to make their case.

"African crude oil had become significantly more expensive relative to U.S. crude oil production in 2011 and ~()12as a result of African prices being based on Brent Blend, the NOl1hSea oil benchmark, which suw rapid pi lvi' 111.rcascsas a result or rising tensions with Iran.

I