The Management of Financial Risks at German Nonfinancial Firms: The Case of Metallgesellschaft

I Boardof Governorsof theFederalReserveSystem InternationalFinanceDiscussionPapers Number560 August1996 THEMANAGEMENTOFFINANCIALRISKSATGERMANNONFINANCIALFIRMS: THECASEOFMETALLGESELLSCHAFT AllenB. FrankelandDavidE. Palmer NOTE: InternationalFinanceDiscussionPapersarepreliminarymaterialscirculatedto stimulate discussionandcriticalcomment. Referencesinpublicationsto InternationalFinanceDiscussion Papers(otherthanan acknowledgmentthatthewriterhashadaccessto unpublishedmaterial)should beclearedwiththeauthoror authors.

ABSTRACT late 1993andearly 1994,thewholly-ownedU.S.subsidiaryof a Germanconglomerate experiencedsubstantiallossesin connectionwiththeimplementationof a petroleummarketing t r s a a b a m

The Management of Financial Risks at German Nonfinancial Firms: The case of Metallgese]ls~haft Allen El.Frankel and David E. Palmer* I In Germany, banks are involved with client firms by advancing credit and providing financial services, similar to U.S. commercial banks. German banks also hold equity stakes in firms and exercise proxy rights for firm shares owned by bank clients. The combmed exercise of both direct and proxy voting rights can be refiected in the presence of German bank managers on client firms’ supervisory boards. It has been theorized in the corporate governance literature that the German model of bank relationships is well designed to minimize agency problems.l In turn, questions have been raised as to what private information is actually distributed within the German system of corporate governance, and how such information is transmitted.2 Our paper is a case study of the transmission of private information between Deutsche Bank and one of its large German corporate clients, Metallgesellschaft AG (MG). The paper * authorsareChief,InternationalBankingSection,andresearchassistant,respectively.inkhc Divisionof InternationalFinance,BoardofGovernorsoftheFederalReserveSystem. Theywishto thankcolleaguesattheBoardofGovernorsforcommentsandsuggestions.Thispaperrepresentstic viewsoftheauthorsandshouldnotbe interpretedas reflectingthe of oftheFederalReserveSystemorotherrnernbersofitsstaff. Thepaperwasoriginallypreparedfora June 1996conferenceoftheAmericanInstituteof ContemporaryGermanStudies. Wewishtothank attendeesofthatconferencefor[heircommentsandsuggestions.

-2draws on a special auditor’sreport on the near bankruptcy of MG and other materials that provide a rare, if not unique, opportunity to assess information issues associated with the German system of corporate governance. We focus on MG’s formulation and use of financial derivatives to implement a risk management strategy associated with a petroleum marketing initiative. The MG case attracted substantial public interest in the United States and in Germany. In the United States, the case was widely cited by proponents and opponents of various regulato~ proposals for derivatives markets in support of more and less regulation, respectively. In Germany, the case was responsible for a boomlet in interest in the perpetual debate on the role of the German Gro~banken in the management of German industrial concerns. In our view, the case raises questions about the efficacy of incentives offered to German banks with respect to their interest in monitoring and assessing the implementation of strategic initiatives by client firms, such as MG’s marketing of long-term, fixed price petroleum supply contracts. B In late 1993 and early 1994,the wholly-owned U.S. subsidiary of a German conglomerate, Metallgesellschaft Corp., experienced losses sufficiently large so as to trigger a DM3.4 billion emergency recapitalization of its German parent company, Metallgesellschaft AG.S The rescue of the MG group was overseen by the parent firm’s supervisory board, which was chaired by Deutsche Bank’snominee, Ronaldo Schmitz, a member of Deutsche Bank’smanagement board.4

-3- Schrnit.zhad succeeded IXesdner Bank’snominee, Wolfgang Roller, as chair of MG’s supervisory board early in 1993. The alternation of the two banks’nominees as chair of the supervisory board reflected shareholdings and control of proxy voting rights by the two ba~]ks that together summed to no less than 40 percent.5 The proxy rights assigned to the two banks not only included those of custodial customers but also those associated with the shares h~~ld in each bank’sinvestment funds ( K a p i t a l a n l a g e g e s el In late November 1993,MG’s management board disclosed plans to reshape the company and announced that it would pay no dividend for the fiscal year i992/93 (which had ended on September 30).6 Stock analysts concluded that MG had limited prospects for a near-term return to profitability. They attributed their pessimistic assessment of MG, which had grown rapidly through acquisitions in the late 1980s,to the depressed level of metal prices that had left MG very vulnerable to the German recession. The sensitivity of MG’s financial position to metal prices had persisted even though the company had become a conglomerate with sizeable trading activities. The analysts’assessments supply context to understand why MG had decided to hire Arthur Benson in November 1991to establish a new business for MG.7 The business was to consist of two elements: petroleum product sales and hedges of related contracts. Benson supposedly put in place a “text-book”hedging strategy to manage market risk associated with an aggressive marketing effort by MG to enter U.S. oil markets. The same basic market opportunity was also independently recognized by the managers of a trading subsidiary of Enron, a publicly-traded U.S. diversified natural gas company. Both MG and Enron offered long-term energy product supply contracts at fixed prices. The business plans

-4of the two companies both called for the management of the financial risks of these contracts through exchange-traded and over-the-counter (OTC) derivatives contracts. The MG episode raised issues about how we]]German banks can perform as financial monitors of nonfinancial firms, particularly those with sophisticated financing activities. It has been presumed by commentators that the H relationaship, such as exisutsed between Deutsche Bank and MG, involves superior access on the part of the bank to information about the firm’s future strategies and financing plans. A highly visible part of the Deutsche Bank- MG relationship was the chairing of MG’s supervisory board by a Deutsche Bank representative.8 Materials released in response to public interest in the near-collapse of MG permitted us to assess the efficacy of the process through which information is transferred within the bank-dominated system of corporate governance. We also found it useful to compare the character of private information disclosures by MG’smanagement (to MG’s supervisory board and to Deutsche Bank, respectively) with the disclosures by Enron in its annual reports and other public documents in the early 1990s: Enron’sdisclosures were more forthcoming. The MG episode highlights the potential exposure of the Hausbank to moral hazard risks. Based on the now extensive public record, we conclude that Deutsche Bank did not closely monitor the nature and extent of MG’s use of Deutsche Bank’sfinancial backing. We appreciate that there is no basis to extrapolate from the MG episode with respect to the quality of monitoring by German banks in general. Nevertheless, the importance of Deutsche Bank’s involvement with MG raises questions about the effectiveness of incentives for bank monitoring in cases where smaller amounts could be at risk.9

-5- Our paper is organized in three sections. Section 1examines Enron’s andMG’s trading strategies and characterof financial risks inherent in those strategies. The section will consider what risks were foreseeable and how they were to be managed by the two firms. Section 2 examines what is on the public record concerning communication between (a) MG’s management and MG’s supervisory board and (b) MG’s management and Deutsche Bank. We review the private and public disclosures by the management board with respect to the management of risk exposures undertaken by MG’soil-trading activities and compare them with Enron’spublic disclosures, which we also examine. We find that MG’spublic disclosures were limited and less informative then Enron’spublic disclosures. We also discover that the passage of information from MG’smanagement board to its supervisory board was limited. The final section (Section 3) sets out a model of how banks are assumed to exercise control over the activities of affiliated nonfinancial firms. We then evaluate how well the model, which focuses on information issues, captures the MG case. Following that, we briefly comment on some of the proposed suggestions for reform in the German system of corporate governance, focusing on those suggestions involving communication of information to the supervisory board. The strategies of Enron and MG have both been characterized as involving the delive~ of financial engineering services to an industry clientele. In the case of Enron, this

-6involved creating a “gas bank.”10The bank is organized to make long-term, fixed-priced commitments of 15 years or more to both buyers and sellers of natural gas, by contracting for physical gas supplies and obligating itself under gas supply contracts, respectively. To coordinate its physical product offerings, Enron created a risk management unit. All financial contracting by Enron was the responsibility of this risk management unit. The unit managed Enron’s risk exposures by entering into short-term exchange-listed as well as “ OTC transactions, such as natural gas swaps, with financial firms’ commodities groups and traders of physical gas.ll The maturity of OTC contracts could extend up to 20 years. The unit employed dynamic hedging techniques in its management of the risk exposures of written options positions (caps and floors). These techniques replicate option positions through ongoing adjustments of forward exposures. These adjustments are reflected in variations of counterpart credit exposures. To manage counterpart risk, Enron created and maintained an independent credit department. Enron outlined its procedures in its annual report and 1O-Kfilings to the SEC. For example, Enron required less than investment-grade firms to post letters of credit or some other collateral to qualify themselves as Enron counterparties. Furthermore, Enron detailed how the firm’s accounting recognized its potential exposure to credit 1ossesthrough charges to current earnings. That is, Enron sought to inform readers of its public disclosures that it was conducting a disciplined trading business in which sources of credit risk were identified and managed. In Section 2, below, we will discuss the character and the specifics of public disclosures made by Enron. This is done to illustrate what a U.S. firm disclosed in response

-7to demands by a dispersed group of financial stakeholders for information about the implementation of a critical trading strategy. Enron’sdisclosure inits 1993 financial statements reads as if it learned what has come to be regarded as the important lesson of MG for other nonfinancial firms: the disclosure focused on how Enron planned to manage the potentially risky consequences involved in implementing a hedging strategy.12 By the late 1980s,MG had committed to become a player in U.S. oil markets.is At the end of 1991,it began to offer fixed-price contracts to supply heating oil and gasoline with terms of up to 10 years to independent wholesalers and retailers.14 Most of the long-term supply contracts were so-called “firm-fixed” and “firm-flexible” contracts, the remainder were guaranteed margin contracts. The guaranteed margin contracts were short-term arrangements that were extendable at MG’sdiscretion. The guaranteed margin contracts did not raise the same kind of risk management issues as the other contracts and thus are not discussed further in this paper. The firm-fixed contracts required that MG supply end users with a defined total volume for a fixed term at a firm price. As a rule, MG’scustomers were obligated to accept delivery of an agreed amount of product per month over a 5- or 10-yearperiod. MG has not publicly disclosed how it determined the creditworthiness of purchasers of supply contracts or if it allowed them to qualify as customers through the use of credit mitigants such as letters of credit or collateral.is

The contract price was to be calculated with reference to a basis that depended on futures contracts maturing within the next 12 months: a simple average of futures prices plus a fixed premium. The firm price quoted did not differ by maturity, that is, the same offered price applied for 5- and 10-year contracts. MG did not offer a rationale for why it chose not to ask for additional compensation for 10-yearas opposed to 5-year supply commitments.1[) Benson offered some commentary on the reason for identical pricing for the two maturities, but his rationale disregarded the consequences of time for valuation and for the computation of future credit exposures.17 From mid-1993 on, MG began to offer firm-flexible contracts. Under these contracts, MG entered into long-term delivery commitments for 5 or 10years at firm prices. But unlike firm-fixed contracts, counterparties had the option of indefinitely deferring (but to no later than the final month of the contract) acceptance of any month’sphysical delivery. In effect, the firm-flexible contract modified the firm-fixed contract by adding to it an option that permits the purchaser to increase the duration of its contractual obligation to purchase oil. A higher contract price was set by MG for oil delivered under these firm-flexible contracts. Many of the long-term supply contracts included “blow-out” options which allowed the buyer (MG’scustomer) to close out the contract if the spot price rose to or above the price specified in the supply contract. Under the original firm-fixed contracts, the customers would receive one-half of the difference between the nearest month futures price and the contract price, multiplied by the quantity of undelivered oil; MG would receive the other half. Under the firm-flexible contracts, the customer would receive the entire difference between the

-9second-neares~futures price and the contract price, multiplied by an amount (chosen by the buyer) not exceeding the volume of undelivered oil under the contract. In addition, from mid-1993 MG sought to modify existing firm-fixed contracts through the purchase of automatic closeout options from oil buyers. Under this modification, cash settlement would occur automatically once the nearest month futures contract reached an exit price, a strike price above the contract price. The firm-fixed contract customers who accepted this modification were granted confessional discounts on oil to be delivered. The Special Auditor’sReport concluded that MG introduced the firm-flexible contracts partly to improve the 1992/93fiscal year results of its U.S. operations: “As internal memoranda and descriptions by the management of MGR&M show, the contractual delivery obligations were expanded between July and September 1993,motivated in part to offset the losses—for accounting purposes, at least—that had already occurred.”lg the conclusion is supportable. However, it would be misleading to disregard other aspects of the initiative, namely, with respect to MG’srisk exposures. In this respect, it seems plausible to view motives as wanting to create an exposure where the highest payoff would occur if a rise in the price of oil coincided with a persistent reversion of the futures curve to backwardation (negative slope). That is, before the modification (to firm-flexible contracts), profits resulting from the exercise of the blow-out options would not have been dependent upon the slope of the oil futures curve.19(For an outline of MG’sbasic strategy and examples of possible outcomes under that strategy, see Appendix A.) An interesting corollary of our proposition that the contractual modifications were designed to rearrange MG’s market risk exposures can be deduced. It is that the managers of

- 10- MG’s oil trading position revealed no concern about a liquidity constraint. indeed, in the face of a sharp decline in oil prices and a persistence of contango, MG’s U.S. operations required more than $1 billion in funding over the fourth of 1993. More than $800 million was accounted for by direct and indirect loans from the German parent firm, with the remainder accounted for by direct bank lending to MG’sU.S. subsidiary. In its March 1995statement, MG revealed that the Group’s financial position had been . seriously impaired since September 1991as a result of a substantial depletion of undisclosed reserves.20 Nonetheless, MG was able not only to participate in a substantial volume of unexchange futures transactions involving large variation margin payments, but also transacted a very large volume of OTC swaps.2i An obvious question concerns how MG was able to retain its access to financial markets as a borrower and as a counterpart to OTC transactions over a period in which its financial position had become dubious. The limited evidence available suggests that MG’scontinued access to financial markets was based on its relationship to Deutsche Bank and Dresdner Bank. Throughout the 1980s, these two banks were consistently involved in arranging and financing MG’sprogram acquisitions. The two banks provided substantial credit (over DM700 million) to MG for these acquisitions. The collaborative nature of their involvement is illustrated by their assistance in MG’s takeover of Dynamit Nobel; the assistance included investments by the two banks in shares of the acquired company.22The two banks managed rights issues and equity-related public debt issues for MG in the late 1980s and early 1990s, totaling more than DM 1 billion.2s In May 1992,Dresdner Bank received a mandate to arrange a DM1 billion multicurrency credit facility for MG. Two months later, the mandate was revised to a DM1.5

-11billion facility’involving 46 German and non-German creditor banks. The stated public purpose of the facility was to provide liquidity in support of general corporate operations; however, the Special Auditor’sReport refers to an internal MG (parent company) newsletter piece on the facility that mentions its possible use for financing the liquidity needs of MG’s oil operations.24 MG’snew management refers to a December 6, 1993 article in the F ra A22gemeineZeitung (FAZ) as the first public revelation of its financial difficulties. With the publication of the article, MG’sfree access to market sources of finance and to transaction counterparties was closed down. (The circumstances and indicators of MG’s loss of access to market financing are outlined in Appendix B.) To avoid immediate closeout of its NYMEX positions, MG turned to Deutsche Bank on December 10for a bridge loan; by the end of December, Deutsche Bank and Dresdner Bank had advanced $900 million.25 In early 1994, the amount advanced by the two banks was increased to assist in the repayment of over DM800 million of maturing commercial paper. (See Section 3.3, below, for a discussion of the banks’role in MG’sfinancial rescue.) Coincidentally, it was concluded by MG’snew management board that the “material adverse change” clause of the 1992DM1.5 billion credit facility barred MG’sdrawing on the facility. The willingness of Deutsche Bank and Dresdner Bank jointly to take the initiative in arranging for emergency bridge lending to MG reflected the presence of their nominees on MG’ssupervisory board. In turn, the presence accorded a standing under German law that permits banks to act as organizers of financial rescues without undue concern about legal

-12risks. By contrast, in the United States, the courts have moved to deter banks from assuming responsibility for rescues of financially troubled corporate clients.2b A The business plans of both Enron and MG called for marketing programs of long-term contracts on energy products that depended on active management of market risk exposures through the use of exchange-traded and OTC financial instruments. The plans were both formulated in the period soon after the deregulation of energy prices. In both cases, the business plans were formulated by individuals who anticipated that the market would pay for the intermediation services embedded in long-term fixed-price private sector contracts. The available evidence suggests that Enron’s management has implemented a disciplined riskmanagement system that encompasses not only market risk but also credit and liquidity risk.27 On the other hand, there are considerable questions concerning MG’s approach to the monitoring, measurement and management of financial risks associated with its oil-trading activities. The Special Auditor’s Report describes a presentation made by MG’s management to MG’ssupervisory board in January 1993 in which the broad trading strategy was laid out without a detailed discussion of risk management practices.28 Reportedly, the discussion was limited to the role of refining capacity in the execution of that strategy as well as a general description of the market risk management strategy to hedge exposure related to supply contracts. In addition, the Special Auditor’sReport refers to meetings between Deutsche Bank staff and MG staff in which the hedging strategy was discussed. However, the Special Auditor’sReport does not indicate that Deutsche Bank advised MG on the plausibility of

-13assumptions embedded in MG’sstrategy. One conceivable topic of discussion might have been the assumed relationship between the level of the price of oil and the slope of the oil futures term structure. The Special Auditor’sReport also does not indicate that MG’s managers themselves conducted such evaluations, nor if either the supervisory board or Deutsche Bank had enquired if MG had conducted stress test analyses of the potential consequences of the assumption. The examination of the MG case by Edwards and Canter (1995) provides us with a summary of analyses discussing the sensitivity of MG’sexposures to particular risks.29 First, the authors suppose a counterfactual case in which MG had instead chosen to adjust dynamically the hedge ratio of financial contract amounts to physical delivery obligations, as opposed to maintaining the one-to-one hedge ratio (stack hedge). They found that such a minimum-variance hedge would have not allowed MG to benefit as much from rising energy prices or lose as much from falling prices. This analysis produces the assessment that MG assumed greater funding risk and greater rollover risk in exchange for the possibility that it could benefit from higher energy prices.so Edwards and Canter extend their analysis by observing that the magnitude of MG’s funding risk associated with a fall in prices would be sensitive to the credit quality of MG’s counterparties for supply contracts because of the non-performance risk. This analysis points to the conclusion that the MG management’schoices of contractual terms (such as the closeout options) and hedge ratio should have intensified interest among stakeholders in MG for direct information. In the U.S. system, there would be a demand for increased public information, with the stakeholders being responsible for their own evaluation of the

- 14information on the quality of MG’sability to manage counterpart credit risk. By contrast, it is often said that in the German system, the responsibility is shifted to the Hausbank to acquire and evaluate information on behalf of other stakeholders. In Section 2, below, we examine the evidence on private and public disclosures by MG and Enron. With respect to MG, no evidence was found of systematic private disclosure of counterpart risk information. By contrast, Enron discloses considerable quantitative and . qualitative information on its exposure to counterpart risk and how it manages those exposures with its choice having been, no doubt, influenced by critical public analyses that pointed to the crucial importance of how Enron managed counterpart risk.st 2: In recent years, economists have come to recognize the rudimentary state of the economic analysis of institutions.s2 In response, they have constructed frameworks for thinking about institutions, such as firms. They have concerned themselves with assessing how alternative forms of contractual relationships affect the flow of information between economic agents, the incentives created by such relationships for information flows, and the use of relationship-based information flows to influence or control firm behavior. One application of the approach concerns the analysis of downstream relationships of banks with nonfinancial firms in various national financial systems. It provides a line of reasoning that has been used to construct a theoretical justification for banks to “control” nonfinancial firms.sq

- 15- The case is based on three observations: (1) Concentration of financial claimholder relationships limits the scope for “free rider” problems. Free rider problems can act as a powerful disincentive to the costly gathering of information and hinder the effective exercise of control. That is, all c claims should result in more information c (2) The ability to ensure long-term relationships may be necessary ensure compensation for the costs incurred in control in contingencies that could not be fully set out in a contract. That a long-term relationship is a necessary condition for a bank to participate willingly in a rescue of a financially troubled customer. (3) The simultaneous holding of debt and equity can operate to cement long-term relationships and improve the information and control possibilities open to creditors. It should also operate to reduce the scope for conflict between equity and debt holders, especially in situations of financial distress. The three observationsjointly support the case for no constraints on links between banks and nonbank financial firms. They jointly suggest the presumption that Deutsche Bank as MG’sHausbank was well informed aboul MG’sconduct of its oil-trading activities. Conversely, the observationsjointly suggest the presumption thaf Enron’slarge number of claimholders would not be well informed compared to the standards set in the relationship of Deutsche Bank and MG. In normal circumstances such a comparison would not be possible. However, in the case of MG, the availability of the Special Auditor’sReport and other

I -16sources provided an opportunity for us to assess the process of information transfer that took place between Deutsche Bank and a large corporate customer. The assessment involves comparing the disclosures by Enron and MG of quantitative and qualitative information on the management of financial risks associated with their trading operations. A C In July 1993, the Group of Thirty published a highly influential report with recommendations for the management of derivatives activities both by derivatives dealers and by end-users.~4 The report identified basic information needs for monitoring the use of derivatives as: an explanation of the purpose of the transactions, the amounts placed at risk, the character of risks involved, and the accounting treatment. We constructed an information disclosure framework based on the Group of ‘l““h“lrty‘’s recommendation. In Table 1, we report on Enron’spublic disclosure of information regarding its risk management activities. Table 1 suggests a substantive enhancement took place in the 1993 financial statements, as compared with 1992. The 1993 statements introduce information on quantitative measures of credit exposures, separating exposures by counterpart types (such as independent power producers; gas and e]ectric utilities; oi] and gas producers; industrials; financial institutions; other) as well as into investment-grade and non-investment-grade. In addition, Enron disclosed the amount of reserves that had been set aside to absorb credit losses. In Table 2, we report on MG’spublic and private disclosures of information on its derivatives related activities. For public disclosures, we examined the annual reports of MG

-17- -Table 1: Enron’s Public Risk Management Disclosure T o d i s yp 19c y l e 1 o ye ar su f s i t f n sa in an tate Q d e D d L i mi D D D E m Y a d d c e o l n i t rvae R fo c er i l s r is e on ed r ve t d c e o ler n i t rve ac E m sv o p t a o i a s r m i a lu ke ti te 1 E 1 f sn i d t th ty rno i a 99i usna tbo p sscn ne tn’ro mcei ru i tn th e ot m i w’ thx wo aa i rt h hitesm snosnu aedtr ti S E 1 oa p f uwi nth SnE (1u r 19i99 an r nurco Ob-l li epes E 1 1 1 a r n O- n9e9 99 ro ponu

- N Table 2: Nletallgesellschaft’s Public and Private Risk Management Disclosure P d iP ub d s rii cl sc T o d i s ypA cr l D nn ep to i Di susc t fo f ye s is buo Dpe B er 1 1 9 99 90 Qualitativedisclosures Generalexplanationof strategy Ye Y Y Discussionof derivatives D N N D N N D N N N E N N N p Q d d N N N N N N N N N N N N 1 I it a r fo th f yen 1 e MG mi 9 n bo apost t9 iscnw u ua 1/ag e x f c t h h it o ua i lbn dgedtl oc e iTeh- onwur tng de nlritg- at ad a th s o th h p b no a e itom iz p dThsoau w boar n meitgtui o t of

-19- Table 2 (cont.) OT s t I e a a wm h d Th 1hapvfda e re apdo9il n medgnntt uoyo90 io f c o s uoa a h d n t ew t e u dr vi ap reac gi 2 A d i tho S A c R ptu u ar c e dma eo ab hheein pmcoc ici nftlto nf wa p o b M m a b a t R Sn s c G’ o aooacMG s h g se nbaohum Oeami pe J 28 1 M am ga a a p n9 nor t baG’neua o t ris93mg areatst aenm tnt th c s o b u H m pth S ooApe R rpsuwa th narvde donepetecv hitea ity’ so s w p t M r o MGr f e em er as cl a thomS e t wn fu inn bceahgrmeog a te al d o th r m e s a a l dt nt su a tis ua o fra atath g t rat uitlienmed i th l c o th a o on bs n a gans lt c ri uk-Sri clw mtta hqauwch br ptrcdetr o o th b c o nal n c si e pt ~ I t c th S A R paua s hith th d eae c tagaaiseo abci pot nttorcc o fl i f n M m f b a ot D Bnr roi t Ge’ca maowah t la augac ana M t beum ts io p fo th p am or ri ma r po a r A a mt rnp kwi t D ao eneeeoBts geeduu L G oiN 23 1q oM sr o u thvri m 9u9i stoeaua adtls ontarimb it a si b s n a uR c Sc m kwa p o plwa thc m ta uarcenarni ahemes ddion autp b c o MG bu a a m oo thhD Ba eL e G ar Aiait Supmb A quro R ts ud d f o M s c c th oo o uo dta mltac rc o t1h9e cD ur B nmouepeas eu ts w g a c e oo xth di erc p m De ol Ba ppve aenl o t eontleth uitv c rast ra t M s w r t d e d t ohathic l eare t roisu c af t tanccloT Sp hetrta nt A R nu t M s cd e o th l oi d o opc hat sene notteoM epiolr li g-nt r’pa d an te D B p l ed a fo rea u r egp ett nk a m sc virt en 4 B o th i c n ia th Sfo A o Rn pthu rsie n et empdt ecM saivait it p th s r b u o D opBa wi qee ova iu r u o nacir v ri nfts i de tior so S S A o R pMu 1u a e dr e rM9 1 i pannne ccre 90t 99 or epsoiaor

- 2(.)for the following fiscal years: 1990/91, 1991/92, 1992/93. The last of these three reports was released in February, 1994 and included revised results for the fiscal year 1992/93 and a brief discussion by MG’s new management board acknowledging that losses had been incurred by MG’sU.S. oil operations.s5 We also reviewed the next two MG annual reports (not reflected in Table 2). The report for the fiscal year ending September 30, 1994 provided more qualitative information concerning MG’s use of derivatives. Finally, the annual report for the year ending ‘ September 30, 1995 introduces quantitative information in MG’spublic disclosures. The limited amount of public disclosure by MG concerning its use of derivatives to conduct risk management did not surprise us. What was surprising was that private disclosures (to MG’s supervisory board and to Deutsche Bank, as outlined in the Special Auditor’sReport) by the management board were only somewhat more informative than its public disclosures: the private disclosures did not provide useful information concerning MG’s risk management activities and fared poorly in comparison with Enron’scontemporary public disclosures.3b Indeed, Ronaldo Schmitz has confirmed our finding in a recent statement that the reporting by MG’s management board has been reformulated under the new management board to improve the quality and timeliness of information passed on to the supervisory board.s7 In summary, the information developed in this section supports the view that there was limited interest on the part of stakeholders in MG to encourage the management to be more forthcoming in both public and private disclosures. By contrast, in Enron’scase, the enhancement of disclosure coincided with greater public interest in firms’ use of derivatives, in general, and in Enron’s involvement,

-21- I D A number of information-based activities have been identified as possible sources of competitive advantage for banks to serve as financial monitors of other firms. These include: (a) collecting information on firms’ongoing operations; (b) offering informed advice to firms’ managements concerning the formulation and execution of business strategies; (c) making informedjudgments on managerial performance; and (d) communicating well-informed signals to other potential creditors of the firm, M Under German law, superviso~ board members must treat information received from a firm’smanagement as confidential. A supervisory board member who is a bank’snominee is also not permitted to pass on bank-confidential information to a firm’smanagement, including confidential information on that same firm.q8 In connection with supervisory board service, the member may tap into the bank’sareas of technical expertise. The information that we reviewed certainly does not suggest that Deutsche Bank was well informed about MG’soil-trading operations simply because its nominee served as chair of MG’ssupervisory board. We also did not find evidence that the supervisory board requested information from MG’smanagement based on well-informed questions passed by Deutsche Bank to its supervisory board nominee. For example, MG’s increased use of credit facilities for the hedging’strategy was not cited by the Special Auditor’sReport as a topic discussed by the supervisory board. However, based on the available information, we cannot choose between two

-LL hypotheses concerning such behavior, namely, that German legal prohibitions were effective or that bank nominees on supervisory boards have little incentive to be actively involved in the oversight of the ongoing operations of firms. Evidence in a recent paper by Steven Kaplan supports the hypothesis of limited involvement by supervisory boards in the management of German firms.s9 Kaplan finds that the replacement of German managers is influenced by stock market outcomes in much the same . manner as in the United States. This suggests that German supervisory boards use the same sort of open-market price signals as available to U.S. boards of directors, rather than process private information.40 On the other hand, it is informative to review the supervisory board’sresponse to the receipt of a special audit of MG by KPMG in mid-December 1993, in the wake of the appearance of press reports that raised concerns about MG’sNYMEX trading positions.41 The superviso~ board responded to the loss revelations by replacing MG’s management board. The new management board then asked Deutsche Bank and Dresdner Bank, both with representatives on MG’s supervisory board, to assist in organizing a creditors’coordinating committee, which Deutsche Bank chaired. This creditors’committee, in turn, considered a rescue plan put forward by MG’s new management.42 Theodor Baums has drawn a picture of active involvement by bank nominees on supervisory boards.43 He suggests that they provide banks’client firms with specialized advice, financial knowledge, and information. He buttresses his case by noting that: “Large banks have departments specialized in corporate finance, analyzing financial markets as well as the financial needs of their clients.”% However, in our review, we found no evidence of an information flow

7J the s r o In s its ex p os on four issues: (1) the creditworthiness of MG’scustomers, (2) the management of MG’s liquidity position, (3) the exit options, and (4) the character of German accounting methods. With respect to the creditworthiness of MG’scustomers, Schmitz commented that 200 of MG’scustomers were middle-sized firms whose likely default had not been considered by noted that the default risk had been taken into account upon the reorganization of MG’soiltrading business in late December 1993. We take this comment as an indicator of Schmitz’s heightened awareness of MG’srisk management activities in his role as MG’ssupervisory board chairman; in addition, Schmitz’scomments could be representative of a new sensitivity on the that MG should not draw on its back-up a would have violated the facility’sclause that drawing should not take place if there had been a material deterioration in business health. He acknowledged“thatthis decision triggered the need for Deutsche Bank and Dresdner Bank to set up the emergency credit facility in December 1993. On the other hand, he was not forthcoming about the possible legal risks associated with a violation of the negative pledge clauses contained in the 1992back-up credit facility through the collateralization of the emergency credit facility.

1 _ Schmitz alluded to the technical problems involved in hedging the exit options. In particular, he focused on the technical problems raised by firm-flexible contracts. The Special Auditor’sReport contains no evidence that the supervisory board routinely reviewed the specifics of contracts offered by MG prior to December 1993. Finally, Schmitz argued that German accounting methods had not misled the supervisory board.47 Schmitz mentioned his concurrence with KPMG’sdecision to apply value adjustments that had resulted in oil derivatives-related losses of $800 million. Schmitz’s statement can be taken as an indication of his post-crisis command of the MG situation. In summary, Schmitz’scomments about the obvious deficiencies of the MG operation, if they are taken at face value, point to an absence of his active involvement as a financial advisor of MG during 1993. That is, Schmitz’s involvement with IMGis not of the character described by Baums. I has been argued that a firm can benefit from being identified as a client of a bank, in the sense that i a ” identification is enhanced if the bank is directly involved in the governance of the firm. In the case of MG, a number of parties may have chosen to act as if MG were backed by Deutsche Bank and that the bank actively monitored MG’soperations.48 One possible instance where such an assumption played a role is NYMEX’Sauthorization of higher limits for MG at the end of 1992,even though NYMEX had concluded as early as September 1992 that MG’s NYMEX member was insufficiently capitalized.”9

-25- Based on the evidence presented in this paper, financial market participants might have reason for concern about the performance of German banks as financial advisers in normal circumstances. But it is not clear that market participants have been as much concerned with this performance asopposed to the performance of German banks with respect to financially troubled customers. It is widely accepted that German banks will assume responsibility for the refinancing of financially troubled customers. In the MG case the major German banks did play this role; nevertheless, there were expressions of dissatisfaction among other creditors. This suggests that it might be revealing to compare the structure of the proposed financial rescue with the structure that was finally negotiated. 3.3 In the aftermath of the public disclosure about MG’slosses from its oil-trading operations, Deutsche Bank and Dresdner Bank participated in a restructuring of MG. Table 3 outlines the structure of the DM3.4 billion financial rescue originally proposed on January 6, 1994and also the structure accepted on January 15, 1994,including subsequent amendments (added through March 1995). The sizeable discrepancy between the structure of the proposed rescue and of the agreed rescue indicates that the German shareholder banks did not impose a solution on other creditors.sOThe agreed solution required the German bank shareholders to purchase convertible shares from other creditor banks and to increase their percentage of MG’sshares outstanding to over 27 percent as compared with less than 23 percent before the rescue.

. 26 - Table 3: The Financial Rescue of MG T f r pli th wha p n e o J r6 a19 c eM sacnoonpth omn p ci (A a c o Do b w n o dei vinM1c efpboro ll r crsnv of er (B a b c i D i an Ml 1c pi re li ( c l o m r f Ci) in eDdM l ]i P rescue (Janruary 6, 1994) o Agreed rescupe (January 15, 199o4)s and subsequent amendments S h (A se tac e r ba deo e x ( agh t asni ao a Dslt veddmi ede b i c ac o e n r fr nvtotntk c ebia p frer oisc ti ( B D e u an I Ma t 19 D sa Dr eup chb an D r e Ds 1 m tdb o00 t poil neo 1 B a f nck ba or re (B se t p i nae s r ( ag tt i hap i nac(Dc 1rt ip o p ftr M s co f p p o n ashoe hao er rDr rpiif1f ti p o h ( 1 pD ee n elsh0. o ru f cets fe D 1 p r e e r s c dnen (C n p ai D mr t( ag tiipM5 Dc 1 mi illero f lpa o l o “ m c roc in r net DMe rmeine ( ai( diu “ c b r an ed Other creditor (A se t c e ba deo x ( so c ni ba ( re nstovein b banks4 i c c o e n r n nvt t a eei rt re frt rgop ic tianair ba de c Do a Drnv eua t ma u t d ( 1i mi ff I Ma 19 1 fo ba s t po t D a D e f D 10 re muit (B t p 3 p ou ne s e( rs p rccaihanma s r pu ha ti o f f e r in (C se t p e D r ( nsc ba opM5 ai crti lvir f D ti m i ne c i r m l t DM med il(li mo il Non-bank (B m s (h tha i aj ( mra ns e hag c t hor a p luols re shareholders I O n o K Av f su e l hgi ts chfwiala t matm r t ira eh an D a se t p i am sh o Ml r c ev at ri ris apic -B i p t Mr s c ho I M pa19 haM H o el (pi rt sol Aui T w D o an D e g 3 r sutit. ie Me 5 [s tshd -p t m s wh a h a aj r ssoue es h ol 6 p o th c i e an r c pi ce re i R o D e B o e if a j w t a uwnli Al n t noekli e ru sc clrits

-27- Table 3 (cont.) 2 Th tw b n t t ha j e a oD ob x t ha hae s nkitibr tMeh1rl i ec en nltl D 1 e c 9 e m93 be ~ A o 1 D f B ha ie i9M n MuGt 13 p caan sDt 95 Br sterr1csrce eata p e r c e nt 4 T w 12 c b i hr I G n baere a we ca Fere dB annk lSw Drue itri rmud C I aan A t b n H m o aM ac a eo pa old d ta r pmldnikiiat idmicean r e bu w sr tp oef wir p puo n ecrr oc et ai rou wre r te cei -m ng S B oo F r A rsu Ze la F nr Tnle- ai otZckgp ie n imfe eimes turtauinr un

-28- It a to the MG rescue. But the fact that there were disputes about the financing structure should be of no surprise in light of the diversification of financial relationships of Germany companies. This outcome is in line with a basic point in a recent paper by Bolton and Scharfstein which is that the structure of debt affects the negotiations that follow recognition of financial problems.si They demonstrate that negotiation of outcomes can be affected by the number of creditors, the distribution of security interests in specific assets, and by the character of voting rules. In summary, the MG case and its aftermath raise questions about the sustainability of traditional German financial restructuring techniques in the presence of passive monitoring by supervisory boards. The questions arise due to the increasing interest of managements of major German firms to establish and maintain more complex financial relationships, and to have these relationships with a larger and more diverse group of creditors. 3.4 The MG case has raised different issues on the two sides of the Atlantic. In the United States, MG was viewed as another instance of sizeable losses associated with derivatives trading. In Germany, the MG case led to renewed interest in German corporate governance and calls for reform. Our paper was primarily motivated by the second set of issues and we comment on two proposed sets of reforms. One proposed set of reforms focuses on improving information flows to the supervisory board, for example, to enhance responsibility and independence of firm

-29auditors.52The auditors would be charged to keep the supervisory board sufficiently well informed so that it can provide ongoing oversight of management. Related proposals call for the creation of audit committees by supervisory boards. The committees would presumably force supervisory board members,to spend more time examining a firm’s risk management strategies and controls. A second set of reforms aims at improving accountability to shareholders by reducing or removing the ability of banks to vote the proxies of nonbank shareholders. Advocates of such reforms, e.g.. the German Shareholders Association, view them as a means of encouraging supervisory board accountability to shareholders. The two sets of reforms differ in how they would “fix”the German system of corporate governance. The first emphasizes enhancing resources available to the supervisory board. For example, advocates of this approach could well argue that independent auditors with broadened mandates would have scrutinized the MG trading strategy; such scrutiny might well have led the supervisory board to examine critically MG’sproposed trading strategy. On the other hand, the second set of reforms places particular emphasis on correcting conflicts of interest that supposedly permeate German bank relationships with their corporate customers. C The objective of this paper is not to add to the voluminous amount already written on MG’soil trading activities. Instead, we have viewed the incident as providing us with source material on how the German corporate governance system functions. We are sensitive to the fact that the case presented might not be representative. However, the size of MG and the

sizeof bank stakes in MGsuggestthat the relationship was sufficiently important to merit attention. Theevidencereferencedinthispaperpointstotheconclusionthatalthoughtwo majorGermanbankshadsufficientfinancialinterestsin theywerenotwellinformed abouttheimplementatioonf anoil-tradingstrategythatalmostbankruptetdhefirm. This aspectof theMGcasedemonstratewshytheremightwellbereservationasbout overwhelmingrelianceonthetransmissionof privateinformation:asupposedlykey componentof theGermanmodelof corporategovernance.

-31 - Notes Th Ya La R ]0 ev 2 Se 3 o a 4 a s ( A i u f s i ch s r a m ( V or r i s “ W i r t s c h a f t sw A d a S ( s S s s q a a B (

-32andcash f t a A b a a 1995andtheBerichttiberdieS on n $ /4 A 1 A k tac ibs e W n g es In ( c Me t 15 a c a d a 9 m a C a ” 21 co t c p e a i F a a s K Be

24 a d d r a d 25 s 26 Th c o e so q u u U.bnu o a i rc nd dta iepa ab na a a a a b I P B P r o E Ac~ cN 1 cm ip on tiki R is d p Z9 E F Rd an M r S Cw ( ai a ann1 cM r kl 99 hads U J o F Maou ut a s I a F o ” rb p d b p os u F C an F o i S i n t rmn t r anrauc d in e I a a A D eP anr P r i r a v i a ct nc ti 2 d d S d f

% W A u i D f eVserlag iuct Das‘JVertpa~ierp,. 66. C a a a i s a a a s J o L E o an cO ur o agw i ra l n niaom za 40 , S a p d m p r m G Th J Ma B a S pa y a nk st ne a d c the a a c a B o a r s e n - Z e i t un c - C a Me D Qe ru i a va rt - “M ” Jo o A C F p o i r p np l oa ie ra nc - Ri - D Qe ru i a v r a t t e iv rl

47 a d - “ Me D Qe ru i a v r a te ti - C “ J o A C o F p o u i r p rnnp li oarn - Ri pp.49-50. 48. In variousanalysesofMG,a numberof authors(mostprominentlyMertonMiller)haveasserted thatfinancingofMGshould a d B o r s e n - Z e it A a s Kl a a n B o r s e n -Z u m a a C J O P Eo o c ul o i rn ntoi S S W a c A a me f A Un Di A k t i e n g e s e l ls $ , ---- , ...=. >

References A G Th J Ma B a S pa y an nk st ” C B ( and C J O P Eo o c ul o i rn ntio i a a W B i di S o n nj 1i4 Ab d 1 A e k Ferbruarbtye6a.ep i r e t ngif C&LTreuarbeitDeutscheRevisionandWollert-ElmendorffIndustrieTreuhand(1995b), “StellungnahmezurErwiderungvonFrau C a Me D Qe ru i a va rt “ J o A C F p o i r p np l oa ie ra nc Ri D Qe ru i a v r a t t er iv Sc W A u i D f eu si W

“ Me D Qe ru i a v r a te ti M . o F lM ut ouar B F a i ian i Gne na ve U R October,pp.28-M. is B P r o E Ao a cNo 1 co o p tk n ii er omvnig . Un Di A k t i e n g e s e l ls F M in an D eP anr P r i r a v i a ct nc ti F C an F o i S i n t rmn t r anra uc M a M E H M Jo o L E an Oc r o gaw na n o i zmi at O M K Z f eB uni Ba at nn s ti k kc w ihrre rt L “ F May o rb

a Me J o C oF o i urr napo Ri pp.49-50. a a D Qe ru i a v r at te Me a Me N C # ( Th Y La R 10 e al v ie ”

-39- . k .c- 2 ~ 0 :Gj .F-4 .m L0 0 n + + m c o .+ 0 o C u + cm G0 $ 0 + + + 1 -1 t&o o c’s 3 2 g I + z I .. 4

-40m c o + I $ , - 2 I I I 2 m I I I IP I c 0 . .. <I iiiEllB I I 1 ,’ – - >k . o 0 0 —+--A--+.- +-- Oooooooog N comwmmJ - . * * \

-41- B Beginning of December 1993: NiG started having extreme difficulty meeting margin l calls on NYMEX. MG required liquidity support of $1 billion to finance futures and swaps positions. December 6: Revelations of losses appeared in the German press. MG share price l fell 13percent. December 10: Deutsche Bank and Dresdner Bank agreed to provide a DM1.5 l billion collateralized bridge loan. QDecember 10: NYMEX notified MGR&M that it could not expand its futures positions and that higher margin calls would be assessed. QDecember 17: MG’s supervisory board fired four members of MG’s management board, including CEO Schimmelbusch and CFO Forster. The day before, Siegfried Hodapp, CEO at MG Corp, had resigned. Mid to late December: Some counterparties in OTC swap transactions refused to l roll over positions with MG Corp without significant collateral; some banks showed hesitance to lend money for financing MG’sposition, others cancelled credit lines. s December 28: NYMEX notified MGR&M that it had to reduce positions on the exchange, and that in the future, positions could not exceed limits imposed by NYMEX. December 30: By this date, approximately 80 percent of the firm-flexible supply l contracts had been terminated. c End December: Share price had fallen an additional 20 percent since December 6. January 5, 1994: MG’s new management announced revised losses of DM1.8 l billion, stating that the concern would require a sizable capital injection to avoid bankruptcy.

-42- InternationalFinanceDiscussionPapers IFDP Number 560 N M 559 L 558 S 557 P I a 556 A 555 S 554 S 553 M 552 551 A 550 A A I 549 548 E S I I 24, W

-43- InternationalFinanceDiscussionPapers IFDP Number 547 U M 546 I 545 Ba A 544 Ba 543 D a 542 P a P 541 A 540 E Kamin S 539 A S 538 E I 537 E a 536 N 535 I 534