CASE 5 The BP Amoco-ARCO Merger: Alaskan Crude Oil (2000) Jeremy Bulow and Carl Shapiro INTRODUCTION In March 1999 British Petroleum Amoco (BP) announced its intention to acquire the Atlantic Richfield Company (ARCO) for $25.6 billion in stock. As one of the largest oil mergers ever,the BP/ARCO deal was sure to attract intense public attention as well as antitrust scrutiny. Attention was further heightened because the deal was part of a more general consolidation in the unloved oil industry. In particular, the BP-ARCO deal came close on the heels of the massive 1997 Shell-Texaco joint venture,BP’s December 1998 acquisition of Amoco,and the then-pending Exxon-Mobil merger. At the heart of the BP-ARCO deal was the combination of the firms’ Alaska North Slope (ANS) crude oil reserves and related operations. The huge Prudhoe Bay oil field was the only one in the United States to have two operators. By 1999, with production having fallen by more than one- half since its 1988 peak,it had become far more efficient to have just one operator. Furthermore,the three primary owners of ANS—BP,ARCO,and Exxon—had disparate shares of oil and gas production. Exxon,for exam- ple,owned a larger share of the gas than the oil. This conflict made it more difficult for the partners to agree on an efficient development strategy. Bulow served as the Director of the Bureau of Economics at the Federal Trade Commission at the time that the Commission reviewed the BP-ARCO merger. Shapiro served as a consultant and ex- pert witness on behalf of BP and ARCO in the antitrust review and litigation of their merger. The opinions expressed here are an amalgam of the sometimes distinct views held by the two authors, and should not be attributed to the Federal Trade Commission,individual commissioners,or to BP or ARCO. We thank Simon Board,John Hayes,Paul Klemperer,and the editors for helpful com- ments on an earlier draft. 128 Case 5: The BP Amoco-ARCO Merger (2000) Overall,BP estimated it could save $100–200 million per year from reor- ganizing Prudhoe.1But the consolidation raised antitrust concerns. Exxon and some smaller investors were minority shareholders in the fields but did not operate in Alaska. Thus, the combined BP-ARCO would own 74 per- cent of ANS production and would operate every oil field in the state. BP entered the antitrust review process with considerable optimism. From its perspective, the deal was quite “clean” on antitrust grounds. Downstream, BP had no West Coast refining and marketing assets, so the merger would not affect concentration there. Upstream,the overlap was in the production of crude oil, arguably a world market where the combined share of BP and ARCO was quite small. But BP also recognized that there were various upstream overlaps related to the exploration,production,and transportation of ANS. The Federal Trade Commission (FTC), along with the states of Cali- fornia,Oregon,and Washington,was keenly interested in how the merger would affect the buyers of ANS, namely West Coast refineries, as well as final consumers,such as motorists. While the commission typically evalu- ates deals based on the effect on consumer welfare alone,as opposed to the sum of consumer and producer welfare, it presumed that an increase in prices charged to refineries would be largely passed along to final con- sumers. The state of Alaska had considerable interest in the deal,because of its strong financial interest in oil production (due to royalties,which domi- nate the state budget) and employment issues. The FTC staff and the state of Alaska originally divided responsibil- ity for the case so that the state would focus on the upstream (oil explora- tion and development, pipelines, and marine transportation) and the FTC mainly on the downstream (sales of ANS to West Coast refineries,impact on refined product prices). The theory behind this division of duties was that the state had more expertise in Alaska-specific issues,and that the in- terests of both the state and the Commission were to promote competition in exploration and development upstream. Downstream there was a diver- gence of interests,with the Commission preferring lower oil prices for con- sumers and the state preferring higher prices, which form the basis of its considerable royalties.2 While BP and ARCO dominated the North Slope,ARCO was also a major player downstream in California refining and marketing—businesses BP was not in. In fact,ARCO used all its own North Slope production and bought additional crude for its own refineries, raising questions regarding the treatment of captive capacity and the role of integrated firms in merger 1BP estimated the overall savings from its acquisition of ARCO at more than $1 billion per year, mostly from consolidating managerial and administrative operations. 2As we shall see,only after Alaska settled with BP by negotiating the Alaska Charter did the Com- mission begin to focus seriously on upstream issues. 129 THE ANTITRUST REVOLUTION analysis.3From a legal perspective,should the deal be viewed as “horizon- tal,”since BP and ARCO were both major producersof ANS,or “vertical,” since ARCO was a net buyer of ANS on the West Coast while BP was a major supplier to ARCO competitors such as Chevron and Tosco?4 The analysis of the BP-ARCO merger can be divided into two major parts: upstream issues in Alaska, and downstream issues on the West Coast.5We organize our analysis along precisely these lines,starting with the upstream issues. As we shall see, however, an upstream divestiture of assets negotiated between BP and the state of Alaska would prove to have a major impact on the downstream analysis. THE UPSTREAM CASE: EXPLORATION AND BIDDING FOR OIL TRACT LEASES The state of Alaska and the federal government regularly auction off the rights to explore and drill for oil on new tracts of land on Alaska’s North Slope (both on-shore and off-shore).6 Under the terms of these auctions, bidders offer a price per acre,subject to a minimum. Winning bidders on a given tract of land obtain exclusive drilling and extraction rights to that tract,but must then pay rent on that tract as well as royalties on any oil that is extracted from it. As any other sellers would,the state and federal governments benefit from competition in these auctions. The basic upstream antitrust issue was whether the merger of BP and ARCO would substantially reduce competi- tion in these auctions,thus leading to a loss of revenue for the state and fed- eral governments and perhaps to a slower rate of development of North Slope oil tracts. There were good reasons for Alaska to fear that the merger would re- duce its revenues from auctions of oil exploration and production rights. BP and ARCO had historically been the largest bidders in auctions of oil leases 3The issue of captive capacity was a common one for the FTC. Generally,it preferred to ignore cap- tive capacity,as when it calculated the market share of Intel in microprocessors by ignoring IBM’s production for its own use. 4Plus, some of ARCO’s major competitors (Exxon) were integrated upstream while others (Chevron) had few or no assets in Alaska. 5There were also potential antitrust issues involving the Trans-Alaska Pipeline System (TAPS) and marine transportation of crude oil from Valdez,Alaska,to the U.S. West Coast. We do not explore those issues in this case. 6The Alaska Department of Natural Resources (ADNR) administers state leases. The Bureau of Land Management (BLM) administers lease sales for federal on-shore properties,and the Mineral Management Service (MMS) administers leases for offshore Outer Continental Shelf federal prop- erties. 130 Case 5: The BP Amoco-ARCO Merger (2000) in the State of Alaska.7In the decade prior to the merger,ARCO accounted for 38.4 percent of all successful bids and BP for 20.2 percent.8Other major bidders were Chevron, Phillips, Anadarko, and Petrofina. The FTC esti- mated that BP and ARCO had been the top twobidders on about 15 percent of all the leases that the state had sold. A loss of revenue equal to the differ- ence between the highest and second highest bids in those auctions would have cost the state and federal governments about $100 million in real terms over the bidding history of the North Slope. Recently,the two firms appeared to be in serious competition with one another in auctions on the western part of the North Slope,in the Alpine and the National Petroleum Reserve-Alaska (NPRA) fields. As usual in any merger involving a bidding market,one must look at the key assets necessary to be an effective bidder, as well as the actual shares of the merging firms in winning,or placing,bids. Here,the key as- sets include (1) control over processing facilities and feeder pipelines valu- able for oil production in new areas; (2) knowledge of the North Slope and experience in operating oil fields there; and (3) three-dimensional (3D) seismic data of the North Slope. In areas (1) and (2), BP and ARCO had advantages over other bidders and could be expected to have lower costs than their competitors for actually conducting North Slope operations. Anyone else who won an oil lease probably would need to negotiate with BP and/or ARCO to provide important services,such as processing facili- ties,pipelines,or operator services. Due to their possession of 3D seismic data, BP and ARCO also had some informational advantage over other bidders. The magnitude and dura- bility of this advantage was an issue in the merger review. The data were col- lected and initially processed by an independent firm,Western Geophysical, raising the possibility that other firms could also contract to obtain such data in the future (as well as obtain existing data as part of a divestiture package). Information is critical in these auctions; put simply, auctions for oil leases are an information-intensive business. There are good reasons to be- lieve that a bidder with superior information in an area will win the lion’s share of the tracts in that area and make almost all of the money.9 Unin- 7Some very interesting issues (beyond the scope of this case) arise in these auctions because joint bidding is common. Even the calculation of market shares is not straightforward in the presence of joint bidding. 8These bidding data are publicly available. See ADNR 1999. 9To illustrate,suppose that an oil field has a true value that is equally likely to be any amount be- tween 0 and 100. One informed bidder knows the exact value,while others are literally clueless, other than knowing the distribution. Each firm submits a sealed bid,with the highest bid winning. Then the (Bayesian Nash) equilibrium is that the informed bidder will bid half the true value and the uninformed bidders will randomize in such a way that the highest bid among them will be equally likely to be any amount between 0 and 50. In this equilibrium,the informed bidder bids v/2 where v is the true value,and each uninformed bidder makes zero in expectation regardless of how much it bids:contingent on winning it knows that the true value must be between 0 and twice its 131 THE ANTITRUST REVOLUTION formed bidders do hold down the profits of the informed bidders—if the un- informed bidders did not participate, then a single informed bidder could win all the licenses for next to nothing—but the seller’s revenue is consid- erably lower than when there is competition between twoinformed bidders. The state of Alaska and the federal government had a strong interest in ensuring that they would receive full value for their property by hav- ing competition between equally well-informed bidders. If the two best- informed bidders were to merge,it might be necessary for the state and the federal government to protect themselves in other ways,such as by raising the minimum price or royalty rate at which they would lease fields. But in- creased minimum prices might cause some leases that otherwise would have been purchased to go unsold. This is hardly a phantom concern: Leases had been awarded on only about 40 percent of the acreage available in state auctions prior to the proposed merger. During November 1999 Alaska negotiated an agreement with BP that was intended to preserve upstream competition in the bidding for leases and more generally to preserve competition on the North Slope. This agreement known as the “Alaska Charter”,was unveiled on December 2,1999. Under the terms of the Alaska Charter,BP would sell 175 thousand barrels per day (MBD) of ARCO’s production to two other production companies10along with seismic data and other upstream assets that would make these compa- nies stronger bidders on the North Slope.11 BP felt that the Alaska Charter fully addressed the upstream issues raised by Alaska and the FTC. It also believed that the state was the natural party with which to negotiate upstream issues since the FTC seemed fo- cused on the downstream issues,and anyway the state had much greater ex- pertise than the FTC in Alaskan production. Once the Alaska Charter was negotiated,Alaska and BP became allies, at least to the extent of arguing that the Alaska Charter dealt adequately with upstream competition issues. Nonetheless,the FTC later challenged the BP-ARCO merger in court pri- marily because of upstream issues. It is doubtful that the FTC knew better than the state about competition within Alaska, but the FTC might have had a better sense of its own bar- gaining position. For example,the state may have been concerned that if it went to court the government would lose the case and the merger would go through as announced. The FTC probably recognized that BP would regard going to court as very costly. Furthermore,once the state had negotiated the bid,while if the high uninformed bid is equally likely to be any amount between 0 and 50 the in- formed bidder will maximize expected profits by bidding v/2. See,for example,Klemperer (1999) for a primer on basic auction theory. 10BP was willing to sell all 175 MBD to one buyer if that was the FTC’s preference. 11The Alaska Charter also required BP to divest the necessary pipeline and tanker capacity to bring this crude oil to the West Coast. We discuss the impact of the Alaska Charter on downstream mar- kets below. The Alaska Charter is available at http://www.bp.com/alaska/ARCO/charter.htm. 132 Case 5: The BP Amoco-ARCO Merger (2000) Alaska Charter,its terms would have been binding on BP even if the FTC lost a bid to block the merger. In that sense,the Charter reduced the riskiness to the merger’s biggest skeptics of going to federal court to block the merger. One concern at the FTC was that while ARCO Alaska (the unit within ARCO conducting ARCO’s Alaskan operations) was a going business,the smaller companies created or enhanced under the terms of the Alaska Char- ter might not be viable competitors. These concerns were partially based upon a study of divestitures conducted by the FTC’s Bureau of Competition (FTC 1999). This study measured the success of a divestiture by whether the divested assets were later “operated viably” in the same industry they had operated in prior to the divestiture. Of twenty-two divestitures of whole businesses, nineteen were deemed successful by this measure. Of the fif- teen divestitures of something other than whole businesses, only six were successful.12In addition,information economics implied that the commis- sion should put a thumb on the scale in favor of divesting complete busi- nesses rather than a set of assets cobbled together from two or more sepa- rate enterprises.13 This “clean sweep” policy of selling whole businesses intact made three commissioners lean heavily in the direction of requiring a complete divestiture of ARCO Alaska,or at least something close to it. The Alaska Charter was negotiated well before the FTC challenged the BP-ARCO deal in court. Therefore,it was natural and sensible to evaluate the impact of the proposed merger given the Alaska Charter. Though de- signed to deal with the upstream issues in the case, the Alaska Charter in fact eliminated the rationale for the commission’s downstreamcase,which is precisely where the FTC’s pre-Charter efforts had been concentrated. This realization gradually led to a change in the FTC’s approach to the case, and affected the subsequent litigation in federal court. WEST COAST CRUDE OIL SURPLUS, DEFICIT, AND ARBITRAGE CONDITIONS The remainder of this case study focuses on the downstream impact of the merger. The basic downstream antitrust issue in the BP-ARCO merger was whether the acquisition of ARCO would allow BP to elevate the price of ANS crude oil to West Coast refineries. Ultimately, higher ANS crude oil prices might lead to higher prices of refined products,especially gasoline, on the West Coast. Certainly this concern was salient to politicians in Cali- 12There were scientific concerns about the study within the Commission; also,a proper study of the success of the Commission’s divestiture policy should evaluate whether consumers were ultimately helped or hurt by the Commission’s orders. The Bureau of Competition report ignored this impor- tant factor. 13A good example of the “sell a whole business”concept was the FTC’s decision to ask Exxon Mobil to sell the Exxon jet oil business,which operated on a stand-alone basis,instead of the Mobil business,which did not. 133 THE ANTITRUST REVOLUTION fornia, Oregon, and Washington. To address these issues, we begin with some background information on the supply and demand of crude oil on the West Coast. Quantities, Imports and Exports In the mid-1970s,West Coast refineries relied largely on California crude oil and imported crude oils. Roughly 45 percent of the crude oil used in PADD V14 was from California, 45 percent from imports, and 10 percent from Cook Inlet in Alaska. The West Coast was “in deficit”; that is,it was a large net importer of crude. Total use of crude oil was roughly 2.5 million barrels per day,or 2500 MBD. These conditions were changed dramatically by Alaskan North Slope production of crude oil. ANS production started in 1977, peaked around 1990 at about 2000 MBD, and has now declined to about 1000 MBD, as shown in Figure 5-1. When ANS production was high,the West Coast was “in surplus”as a net exporter of crude. But by 1999 the West Coast was again deeply in deficit,importing more than 600 MBD of crude oil,as shown in Figure 5-2. By 1999, some 42 percent of crude oil used on the West Coast was from Alaska,33 percent from California,and 25 percent from imports; see Fig- ure 5-3.Of the Alaskan oil,three-fifths was sold on the merchant market (this includes all of BP’s ANS),and the rest was transferred internally (primarily ARCO and Exxon ANS crude oil used in their own West Coast refineries). Crude Oil Prices As a general principle,the price of ANS crude oil closely tracks other crude oil prices over time. Figure 5-4 shows the price of ANS and a number of other crude oils between 1989 and 1999. In this sense,crude oil prices on the West Coast are governed by conditions in the world crude oil market. The spike in prices in 1991,for example,reflects the Gulf War. The corre- lations among these different crude oil price series are very high,typically in the 0.97 to 0.99 range. However, price differentials between different grades of crude oil do vary somewhat over time. We will be examining these differentials closely. In particular,we look closely at the time series of the difference between the price of ANS crude oil and the price of the benchmark West Texas Intermediate crude oil (WTI).15 The price differentials between ANS and WTI crude oils can largely be explained by import and export arbitrage conditions. In the late 1980s and 14Much of the data in this industry actually covers Petroleum Area Defense District (PADD) V, which encompasses not only California,Oregon,and Washington,but also Alaska,Hawaii,and Arizona. 15ANS is more “sour”and thus cheaper than WTI,so the ANS-WTI price differentials are negative numbers. 134 Case 5: The BP Amoco-ARCO Merger (2000) FIGURE 5-1 Alaska North Slope Crude Oil Production, 1977–2005 2500 2000 1500 D B M 1000 500 0 19771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999200020012002200320042005 ANS Crude Oil plus Natural Gas Liquids Production Source:State of Alaska Department of Natural Resources. Note:2000–2005 data are projected. early 1990s, the West Coast was in surplus, and foreign exports of ANS crude oil were prohibited. Therefore,some ANS crude oil had to be trans- ported beyond the West Coast to the Virgin Islands and the Gulf of Mexico, where it would compete with WTI and other crudes. Competitive market ar- bitrage implied that the West Coast price should be the price in the alterna- tive markets,minusthe incremental transit costs. In the late 1990s,with the region in deficit,transit was going the other way. A competitive West Coast price for ANS crude oil should have reflected the price of crude in other markets,plusany incremental cost of shipping that crude to the West Coast. The price of ANS crude oil rose relative to WTI crude oil by about $1.50 from 1993 to 1995,as the market moved from surplus to balance and later into deficit (see Figure 5-2). Ironically,the move from surplus to deficit both raised prices and re- duced the chance that the merger would elevate ANS crude prices. Once the West Coast was in deficit,BP,ARCO,and Exxon were able to sell their oil at the cost of imports plus transit costs from a competitive world market. In- creasing the shortage by exporting out of the region would not raise prices very much,as the supply of imports was highly elastic,and therefore would only be a viable strategy if transport costs were very low. By contrast,in the early 1990s it was theoretically possible that an increase in exports could have raised prices significantly by moving total supply from surplus to shortage, potentially making exports profitable for a large supplier, even one with high transit costs. 135 THE ANTITRUST REVOLUTION FIGURE 5-2 PADD V Imports and Exports, 1989–1999 800 600 400 D) 200 B M Oil ( 0 e 89 90 91 92 93 94 95 96 97 98 99 d 9 9 9 9 9 9 9 9 9 9 9 Cru−200 1 1 1 1 1 1 1 1 1 1 1 −400 −600 −800 Imports Total Exports Net Exports Source:Department of Energy (Energy Information Administration) and company data. FIGURE 5-3 Usage of Crude Oil in PADD V, 1999 Imports Merchant Market ANS 25% 25% Arco and Exxon Captive ANS 16% California Crudes Cook 33% 1% 1999 PADD V Runs 2420 MBD Source:Company data. 136 Case 5: The BP Amoco-ARCO Merger (2000) FIGURE 5-4 Monthly Prices (per barrel) for Select Market Crudes $40.00 $35.00 $30.00 $25.00 $20.00 $15.00 $10.00 $5.00 $0.00 Jan-8M9ay-8S9ep-8J9an-9M0ay-9S0ep-9J0an-9M1ay-9S1ep-9J1an-9M2ay-9S2ep-9J2an-9M3ay-9S3ep-9J3an-9M4ay-9S4ep-9J4an-9M5ay-9S5ep-9J5an-9M6ay-9S6ep-9J6an-9M7ay-9S7ep-9J7an-9M8ay-9S8ep-9J8an-99 ANS Brent Dubai WTI Source:Reuters (ANS),Nymex (WTI),Platts (Brent and Oman). Demand for ANS since 1995 Since 1995,as shown in Figure 5-2,PADD V has increasingly relied upon imports to meet its crude oil needs. Data from this period provide strong ev- idence that West Coast refineries were capable of replacing ANS crude oil with foreign crude oils without incurring substantial incremental costs as a result of this substitution. In other words, the intermediate to long-term elasticity of demand for ANS crude oil on the West Coast is very high. Im- ported crude oils are very close substitutes for ANS crude oils. The experience of California refineries is illustrative. From 1995 to 2001,ANS crude sold in California declined by 342 MBD,from 725 MBD to 383 MBD; balancing this,imports rose by 370 MBD,from 156 MBD to 526 MBD.16Despite this tremendous decline in ANS crude oil availability, the price of ANS crude did not rise at all relative to the price of WTI. In 1995,ANS crude sold for an average of $5.91 per barrel less than WTI; in 2001,the differential was actually higher,at $6.44.17 Declining ANS crude oil production is a wonderful natural experiment that reveals a great deal about the demand for ANS crude oil on the West 16Source:http://www.energy.ca.gov/fuels/oil/_crude_oil_receipts.html. 17These data are taken from the Energy Information Administration website. See http://www. eia.doe.gov/pub/oil_gas/petroleum/data_publications/petroleum_marketing_monthly/current/pdf/ pmmtab22.pdf. In real terms the ANS discount declined slightly. 137