Restructuring Strategies and Post-Bankruptcy Performance, Thesis of Corporate Finance

Download Restructuring Strategies and Post-Bankruptcy Performance and more Thesis Corporate Finance in PDF only on Docsity! TECHNISCHE UNIVERSITÄT MÜNCHEN Lehrstuhl für Betriebswirtschaftslehre – Controlling Restructuring Strategies and Post-Bankruptcy Performance Marcel Bruno Naujoks Vollständiger Abdruck der von der Fakultät für Wirtschaftswissenschaften der Techni- schen Universität München zur Erlangung des akademischen Grades eines Doktors der Wirtschaftswissenschaften (Dr. rer. pol.) genehmigten Dissertation. Vorsitzende: Univ.-Prof. Dr. Dr. Ann-Kristin Achleitner Prüfer der Dissertation: 1. Univ.-Prof. Dr. Gunther Friedl 2. Univ.-Prof. Dr. Christoph Kaserer Die Dissertation wurde am 29.02.2012 bei der Technischen Universität München ein- gereicht und durch die Fakultät für Wirtschaftswissenschaften am 02.05.2012 ange- nommen. II Foreword A corporate insolvency normally has far-reaching consequences for employees, financiers, suppliers and customers alike. Both scientists and practitioners have there- fore paid very close attention to this phenomenon. To attenuate the negative repercus- sions, insolvency law often provides for mechanisms that can help to restructure firms, enabling them to survive and continue operating as a going concern. In the U.S., this kind of reorganization takes place under what is known as "Chapter 11 bankruptcy". It is therefore worth asking what companies can do to make a success of their reorgani- zation – and to stay successful when they emerge from Chapter 11 protection. Al- though the question is of great significance to modern economies, too little investiga- tive attention has so far been paid to it. This being the case, the objective of the dissertation that follows is to enrich ex- isting research into the effects of restructuring actions. The author analyzes an exten- sive range of actions taken by U.S. firms under Chapter 11 protection and examines their impact on the firms' performance. In several respects, he goes beyond the scope of past studies. First, he addresses not only those actions taken during insolvency, but also steps undertaken during the post-bankruptcy phase. This addition makes sense, as it is reasonable to assume that not only actions taken during Chapter 11 bankruptcy are likely to be crucial to the lasting success of a restructured company. Second, he ex- plores a more detailed array of actions than is the case in existing empirical literature. His analysis is based on a sample of large public U.S. firms that filed for bankruptcy between 1993 and 2005. The author's analysis shows that certain restructuring actions have a measurable impact on a company's long-term performance in the wake of in- solvency. V I thank Daniel Arand and Maximilian Doms for all the Wednesday morning re- search breakfasts, at which we shared the ups and downs of being a doctoral student. I owe special thanks to Dr. Kevin Aretz, who critically reviewed my thesis, challenged me by asking the right questions and was thus instrumental in making this a better the- sis. Finally, and most importantly, I would like to thank my parents, Alwine and Walter Naujoks, and my beloved wife, Annika Naujoks, for always and unconditional- ly supporting and motivating me during the good and the not-so-good times. Without their help and understanding, this project would not have been possible. I wish to dedi- cate my thesis to my parents. Munich, June 2012 Marcel Naujoks VI Table of Contents List of Figures ........................................................................................................... VIII List of Tables ............................................................................................................... IX List of Abbreviations .................................................................................................. XI 1 Introduction .............................................................................................................. 1 1.1 Motivation and Research Questions .................................................................... 1 1.2 Research Gap and Contribution ........................................................................... 4 2 Literature Review ..................................................................................................... 8 2.1 Financial Distress and the Efficiency of Chapter 11 ........................................... 8 2.2 Financial Distress and Corporate Restructuring ................................................ 13 2.3 Bankruptcy, Reorganization and Post-Bankruptcy Performance ...................... 16 3 Corporate Bankruptcy in the U.S. – Liquidation and Reorganization ............. 23 3.1 Principles of U.S. Bankruptcy Law ................................................................... 23 3.2 Liquidation under Chapter 7 .............................................................................. 25 3.3 Reorganization under Chapter 11 ...................................................................... 26 4 Research Model ...................................................................................................... 31 4.1 Definitions ......................................................................................................... 31 4.1.1 General Definitions .................................................................................. 31 4.1.2 Chapter 11 Outcomes and Post-Bankruptcy Outcomes ........................... 33 4.2 Research Model ................................................................................................. 35 4.2.1 Post-Bankruptcy Performance as Dependent Variable ............................ 37 4.2.2 Restructuring Actions as Independent Variables ..................................... 45 4.2.2.1 Operational Restructuring ........................................................... 47 4.2.2.2 Financial Restructuring ............................................................... 55 4.2.2.3 Managerial Restructuring ........................................................... 62 4.2.2.4 Portfolio Restructuring ............................................................... 65 4.2.3 Context Factors as Control Variables ....................................................... 69 VII 5 Methodology and Data ........................................................................................... 77 5.1 Statistical Methodology ..................................................................................... 77 5.1.1 Choice of Regression Model .................................................................... 77 5.1.2 Test for Sample Selection Bias ................................................................ 78 5.2 Data Sources ...................................................................................................... 81 5.3 Sample Selection ............................................................................................... 83 5.3.1 Sample Selection Process ........................................................................ 83 5.3.2 Sample Size Requirements and Treatment of Missing Data ................... 85 5.4 Descriptive Statistics ......................................................................................... 86 6 Analysis and Research Findings ........................................................................... 97 6.1 Univariate Analysis ........................................................................................... 97 6.2 Multivariate Analysis ....................................................................................... 100 6.2.1 Final Model ............................................................................................ 100 6.2.2 Final Model with Interaction Terms ...................................................... 108 6.2.3 Models by Restructuring Strategies ....................................................... 109 6.2.4 Effect of Discrete Changes .....................................................................111 6.3 Robustness of Findings and Regression Diagnostics ...................................... 115 6.3.1 Test for Sample Selection Bias .............................................................. 115 6.3.2 Test for Endogeneity Bias ...................................................................... 119 6.3.3 Collinearity among Independent Variables ............................................ 122 6.3.4 Normality of Residuals and Influential Cases ....................................... 124 6.3.5 Robustness of Cutoff Levels of Independent Variables ......................... 125 7 Conclusions ........................................................................................................... 127 7.1 Main Findings and Concluding Remarks ........................................................ 127 7.2 Limitations and Directions for Future Research ............................................. 131 References .................................................................................................................. 133 Appendix .................................................................................................................... 144 X Table 27: Collinearity Diagnostics.............................................................................. 123 Table 28: Emerged Public Sample Firms .................................................................... 144 Table 29: Multivariate Probit Regression of Post-Bankruptcy Performance - Final Model .......................................................................................................................... 145 Table 30: Multivariate Logistic Regression of Post-Bankruptcy Performance – Cumulative Models by Restructuring Strategies ........................................................ 146 Table 31: Correlation Matrix of Independent and Control Variables ......................... 147 XI List of Abbreviations AICPA American Institute of Certified Public Accountants APR Absolute priority rule BAPCPA Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 BRD UCLA-LoPucki Bankruptcy Research Database CAPEX Capital expenditures CEO Chief Executive Officer COGS Cost of goods sold CPI Consumer price index DBeta Pregibon’s Delta-Beta DIP Debtor in possession EBT Earnings before taxes EBIT Earnings before interest and taxes EBITD Earnings before interest, taxes and depreciation EBITDA Earnings before interest, taxes, depreciation and amortization EDGAR Electronic Data Gathering, Analysis, and Retrieval system FCF Free cash flow FDIC Federal Deposit Insurance Corporation LR Likelihood ratio M&A Mergers and acquisitions PBGC Pension Benefit Guaranty Corporation ROA Return on assets ROS Return on sales XII SDNY United States District Court for the Southern District of New York SEC Securities and Exchange Commission SG&A Selling, general and administrative expenses SIC Standard Industrial Classification SOP Statement of Position of the American Institute of Certified Public Accountants UCP Unanimous consent procedure UK United Kingdom U.S. United States U.S.C. United States Code USD U.S. dollar VIF Variance inflation factor 1.1 Motivation and Research Questions 3 tering viable and nonviable firms, and for being biased in favor of the survival of inef- ficient firms. Baird (1986) argues that the U.S. bankruptcy law is biased toward reor- ganization.10 Hotchkiss (1995) conjectures that failing to replace the incumbent man- agement during reorganization may be related to this bias.11 This is in line with Brad- ley and Rosenzweig (1992), who argue that provisions of the Bankruptcy Code give the incumbent management preferential treatment at the expense of stockholders and bondholders.12 Reconciling these two streams, White (1994) argues that the U.S. bankruptcy law faces a trade-off between letting inefficient firms reorganize under Chapter 11 (type I error) and liquidating efficient firms under Chapter 7 (type II error).13 Conse- quently, estimating the level of the type I error is an empirical question, as Hotchkiss (1995) suggests.14 However, the trade-off illustrated by White (1994) assumes that the Chapter 11 process is static and that firms are either viable or nonviable.15 In contrast, I regard reorganization under Chapter 11 as a highly dynamic process that allows im- portant stakeholders in a firm – namely the shareholders, the management and the creditors – to define and implement value-preserving and value-increasing restructur- ing actions that may contribute to (i) firm survival and (ii) post-bankruptcy success. This interpretation of the reorganization process under Chapter 11 relates to the dy- namic liquidation theory put forward by Kahl (2002). He models the decision of the creditors to liquidate a firm in financial distress as a dynamic process that gives a firm’s shareholders and management time to convince the creditors that continuing the firm may be the preferred option.16 Due to the creditors’ postponement of the liquida- tion decision, this time can be useful to implement value-increasing restructuring ac- tions. 10 See Baird (1986), pp. 133-134 and p. 145. 11 See Hotchkiss (1995), p. 4. 12 See Bradley and Rosenzweig (1992), pp. 1049-1050. 13 See White (1994), p. 269. 14 See Hotchkiss (1995), pp. 4-5. 15 See White (1994), p. 268. 16 See Kahl (2002), pp. 136-138. 1.2 Research Gap and Contribution 4 Accordingly, this study concerns itself with analyzing the restructuring strate- gies and the post-bankruptcy performance of large public U.S. companies and ad- dresses the following research questions: First, how effective are restructuring efforts of bankrupt U.S. firms in contributing to post-bankruptcy success? Second, which re- structuring strategies in general and which restructuring actions in particular signifi- cantly impact the probability of post-bankruptcy success? This focus is motivated as follows: First, the number of bankruptcies is still not receding despite decades of re- search into business failure, implying the need for further research into restructuring strategies.17 Sudarsanam and Lai (2001) argue that “[…] insolvency is the ultimate non-recovery and thus merits analysis as to [the] recovery strategies employed […]”18. Second, Chapter 11 of the U.S. Bankruptcy Code has served as a role model for many countries in recent decades.19 Third, since large bankruptcy cases such as those of American Airlines or General Motors catch public attention, it is mostly in respect of these cases that both academics and practitioners debate the efficiency of the bankruptcy system and firms’ post-bankruptcy performance.20 Besides, the challenges that small companies face during bankruptcy can differ significantly from those with which large companies are confronted.21 Fourth, given the highly developed capital markets in the U.S., data availability in terms of the number of bankrupt public com- panies places no material restrictions on the study. 1.2 Research Gap and Contribution This chapter details how this study contributes to and extends existing post- bankruptcy performance literature. The scope of the analyzed restructuring actions is enlarged compared to prior contributions to post-bankruptcy performance literature. I have relied on the seminal contribution to the restructuring literature by Lai and Sudar- 17 Similar arguments are put forward by Arogyaswamy, Barker, and Yasai-Ardekani (1995), p. 493, and Lai (1997), p. XVI. 18 Sudarsanam and Lai (2001), p. 190. 19 See Warren and Westbrook (2009), p. 604. 20 See Denis and Rodgers (2007), p. 113, and Baird (1993), p. 637. 21 See Lemmon, Ma, and Tashjian (2009), p. 6, and, more generally, Evans and Koch (2007). 1.2 Research Gap and Contribution 5 sanam (1997), who define four generic restructuring strategies that financially dis- tressed firms may choose.22 Thus, I examine restructuring actions of bankrupt firms falling into one of the four generic restructuring strategies operational, financial, ma- nagerial and portfolio restructuring.23 To make sure that the analysis of restructuring strategies and post-bankruptcy performance is sufficiently detailed with respect to the restructuring actions taken, I have combined both quantitative and mostly hand- collected qualitative data from different sources for my analysis. In doing so, I have sought to provide a sufficiently detailed, but also comprehensive analysis of which restructuring actions contribute to a higher probability of post-bankruptcy success. To analyze the impact of restructuring on post-bankruptcy performance, this study scrutinizes the restructuring actions undertaken during both the bankruptcy phase and the post-bankruptcy phase. This novel approach in post-bankruptcy litera- ture to date is supported by theoretical and conceptual models from restructuring lite- rature such as those of Robbins and Pearce (1992) and Arogyaswamy, Barker, and Ya- sai-Ardekani (1995). I have adapted these contributions from restructuring literature modeling turnaround as a process to the bankruptcy context.24 This appears to be a promising approach, since both process stages can be considered to be very different. During Chapter 11, the firm as a debtor in possession is subject to the rules and regula- tions of U.S. bankruptcy law and is supervised by the bankruptcy court. In recent years, firms in Chapter 11 have also increasingly been controlled by their creditors.25 Many of the rules and regulations that apply during bankruptcy, such as the automatic stay or the right to reject executory contracts and unexpired leases, are designed to give a bankrupt firm the opportunity of a fresh start when it emerges from bankrupt- cy.26 Once it has emerged from Chapter 11 and is no longer “[…] largely shielded from 22 See Lai and Sudarsanam (1997), p. 207 and p. 209. 23 Lai and Sudarsanam (1997) use the term asset restructuring instead of portfolio restructuring, which I employ in line with Bowman and Singh (1993) and Eichner (2010). 24 A similar approach has been recently applied by Eichner (2010) with regard to the restructuring of financially distressed (non-bankrupt) manufacturing firms. He analyzes which restructuring actions effectively contribute to a higher turnaround probability, relying on a two-phased process model involving early and late restructuring actions. 25 See Skeel (2003), p. 918, for example. 26 See White (1989), pp. 144-145, and Asquith, Gertner, and Scharfstein (1994), pp. 651-652. 2.1 Financial Distress and the Efficiency of Chapter 11 8 2 Literature Review The literature review provides an overview of the most relevant and influential contributions to research in the fields of financial distress, corporate restructuring and corporate bankruptcy. The literature mentioned here is far from being exhaustive. In- stead, I have limited myself to contributions that directly relate to my research topic of restructuring strategies and post-bankruptcy performance.33 My research thus inte- grates into the three research fields mentioned above. While the focus remains on con- tributions from financial economics, other related fields such as strategic management and the legal perspective on corporate bankruptcy have also been taken into account. 2.1 Financial Distress and the Efficiency of Chapter 11 According to Wruck (1990), financial distress is defined as “[…] a situation where cash flow is insufficient to cover current obligations.”34 Bankruptcy is explicitly included in this definition of financial distress.35 Much of the theoretical work on fi- nancial distress and corporate bankruptcy focuses on the efficiency of the bankruptcy process. In this context, efficiency may refer to two related yet distinct topics. First, efficiency relates to the screening or filtering process by which inefficient firms should be separated from efficient ones.36 Inefficient firms should be liquidated under Chap- ter 7 while efficient firms should be given the opportunity to reorganize under Chap- ter 11. As White (1989) shows, this is in line with basic economic theory which pre- dicts that competition will drive prices toward the equilibrium price and results in 33 See Hotchkiss et al. (2008) or Altman and Hotchkiss (2006) for a review of literature on financial distress and corporate bankruptcy and Eichner (2010) or Nothardt (2001) for reviews of literature on corporate restructuring and turnaround. Refer to chapter 3 for an over- view of U.S. bankruptcy law from a legal perspective. 34 Wruck (1990), p. 421. This definition includes, for instance, unpaid liabilities to suppliers or employees, liabilities (actual or potential) arising from any litigation (e.g. asbestos claims), and default on any principal or interest payments. Wruck (1990) defines financial dis- tress as flow-based insolvency in contrast to stock-based insolvency. Refer to Altman and Hotchkiss (2006), p. 5, for more on the stock- and flow-based definitions of financial distress. 35 See Wruck (1990), p. 422. 36 See White (1989), White (1994) and Mooradian (1994). 2.1 Financial Distress and the Efficiency of Chapter 11 9 firms being driven out of the market if their unit production costs are above the equili- brium price.37 However, White (1989) demonstrates that, under Chapter 11, there is a tendency to keep alive inefficient firms that should have been liquidated, or at least to unnecessarily delay the move of corporate resources to more favorable uses.38 One of the driving factors behind the continuation of inefficient firms is argued to be con- nected to the manifold subsidies that are granted to firms under Chapter 11.39 These subsidies include the retention of accrued tax loss carryforwards, exemption from tax on the gains from any forgiven debt, the right to terminate pension plans under certain conditions, the automatic stay of most interest and principal payments due by the deb- tor and the right to reject executory contracts and unexpired leases.40 White (1994) presents a game-theoretic model about the initial outcome of the bankruptcy process and interprets the process as a filter that may not work perfectly. It follows that some inefficient firms could reorganize under Chapter 11 (type I error) while some efficient firms might be liquidated under Chapter 7 (type II error).41 These results may materialize because efficient firms can benefit from appearing less effi- cient than they actually are (as, in this case, creditors might forgive more debt in order to rescue the firm) while inefficient firms can profit from appearing more efficient than they actually are (as this might result in reorganization rather than liquidation).42 White (1994) concludes that the bankruptcy policy practiced in the U.S. leads to a trade-off between type I and type II errors with a bias in favor of type I errors, i.e. letting ineffi- cient firms reorganize under Chapter 11.43 She explains this bias in favor of the type I error by pointing to the immediately visible cost of type II errors, namely job losses, 37 See White (1989), p. 129. Similar arguments are put forward by Baker and Kennedy (2002). 38 See White (1989), p. 130. 39 These subsidies are especially valuable compared to firms outside of Chapter 11 and also to firms that liquidate under Chapter 7, see White (1989), p. 144. 40 See White (1989), pp. 144-145, and White (2007), p. 1021. 41 See White (1994), p. 269. Bradley and Rosenzweig (1992) argue that the U.S. bankruptcy law introduces biases for incumbent manag- ers toward reorganization in Chapter 11, even when liquidation should be preferred. A similar position is adopted by Baird (1986). This can also be regarded as management entrenchment in the sense put forward by Shleifer and Vishny (1989). Bolton and Scharfstein (1996), p. 2 and p. 5, refer to the issue as a strategic filing by management intended to divert available cash to themselves. 42 See White (1994), p. 269. 43 See White (1994), p. 293. 2.1 Financial Distress and the Efficiency of Chapter 11 10 which policymakers try to avoid for obvious reasons.44 Validation of the scope of type I errors would be an empirical question analyzing the fate of firms emerging from bankruptcy.45 Second, as Gertner and Scharfstein (1991) show, efficiency may relate to in- vestment incentives of financially distressed firms. Inefficiencies in investment beha- vior may be triggered by coordination problems and conflicts of interest in the rene- gotiation (or bargaining) process among creditors and debtors.46 Building on earlier work by Bulow and Shoven (1978) and White (1983), they show that two inefficien- cies in the renegotiation process may occur: underinvestment and overinvestment.47 On the one hand, underinvestment may result since public debtholders of a distressed firm are likely to claim their share in any cash flows resulting from new investments. Given this circumstance, it may be difficult to convince some investors to provide fresh money to the firm. Consequently, some positive net present value projects could not be undertaken.48 On the other hand, overinvestment may occur because the share- holders of a distressed firm, finding themselves with their back to the wall, stand to receive much of the potential upside benefits of risky investments while bearing close to nothing of the cost of downside risks.49 Consequently, a distressed firm’s sharehold- ers and incumbent management might be incentivized to undertake high-risk projects and thereby effectively shift risk to the creditors.50 Additionally, Gertner and Scharfstein (1991) show in their model that it can be difficult to renegotiate with pub- lic debtholders in an exchange offer due to the holdout problem.51 Some debtholders 44 See White (1994), p. 293. 45 See Hotchkiss (1995), p. 5. 46 See Gertner and Scharfstein (1991), pp. 1190-1191. 47 See Gertner and Scharfstein (1991), p. 1191. 48 See Gertner and Scharfstein (1991), p. 1191 and p. 1195, and generally Myers (1977). 49 In most cases, this will also apply to the management as agent to the shareholders, especially when managers have stock holdings of the firm or when managers fear losing their job. In such a setting, managers might also be inclined to gamble in risky projects. 50 See Gertner and Scharfstein (1991), p. 1191 and p. 1195, and generally Jensen and Meckling (1976). 51 An example of the holdout problem is Amerco. Its efforts to restructure out of court were unsuccessful due to the differing interests of several creditor groups. By consequence, Amerco filed for bankruptcy as documented in Amerco’s 2003 annual report, p. 2. Another example can be found in Applied Magnetics’ 2001 annual report, p. 4, which states that “[t]he Company made a formal proposal to its trade creditors regarding a compromise and extension of the Company’s obligations. While certain of the Company's trade creditors accepted the Company's proposal, a substantial number of the trade creditors did not and in certain instances they commenced en- forcement actions against the Company.” 2.2 Financial Distress and Corporate Restructuring 13 timal investment activities by the firm, both of which can affect post-bankruptcy per- formance. On the other hand, the dynamic liquidation theory of Kahl (2002) allows financial distress to be linked to corporate restructuring.64 According to the dynamic liquidation theory, distressed firms (and bankrupt firms too) may win some time before creditors decide again whether to liquidate the firms or not. This time should allow the incumbent management to engage in comprehensive restructuring strategies such as restructuring operations or the firm’s portfolio, with a view to returning the company to profitability.65 These actions could help to convincing creditors that the firm is via- ble and that liquidation should be avoided. 2.2 Financial Distress and Corporate Restructuring For the purposes of this study, restructuring in the context of financial distress follows the definition given in Eichner (2010), who builds on the work of Altman and Hotchkiss (2006) and Bowman and Singh (1993).66 Accordingly, aimed at turning the firm around and overcoming financial distress, restructuring is defined as any material discretionary change in a firm’s assets, its capital structure, its operations, or its top management. While restructuring refers to discretionary changes or actions, turna- round refers to the outcome of the restructuring process: Either the firm managed to overcome financial distress or it did not. As Eichner (2010) points out, it is important to limit the definition of restructuring to material discretionary changes in contrast to any incremental continuous improvement programs.67 One pioneering contribution to restructuring literature comes from Lai and Su- darsanam (1997), who embed their empirical analysis in a theoretical framework that brings together the fields of financial economics and strategic management.68 Based on 64 In this case, corporate restructuring involves more than merely restructuring debt and should be understood in the sense used by Lai and Sudarsanam (1997) or Eichner (2010). 65 Note, however, that the potential issue of overinvestment might prevail unless the bankruptcy court or the creditors effectively supervise the decisions taken by the incumbent management during Chapter 11. 66 See Eichner (2010), p. 50, Altman and Hotchkiss (2006), p. 122, and Bowman and Singh (1993), p. 8. 67 See Eichner (2010), p. 50. 68 See Lai and Sudarsanam (1997), p. 198, which is based on Lai (1997). 2.2 Financial Distress and Corporate Restructuring 14 an agency model, they examine the influence of the ownership structure, corporate governance and lender monitoring on the chosen restructuring strategies for a sample of UK firms whose performance is in decline.69 They are thus concerned with potential conflicts of interest among the shareholders, managers and creditors of poorly per- forming firms and how these conflicts relate to the choice of restructuring strategies.70 They define four generic restructuring strategies: operational, financial, managerial and asset restructuring which are shown in Table 1 below.71 Lai and Sudarsanam (1997), and originally Lai (1997), deducted these generic restructuring strategies from prior turnaround research emanating from both the financial economics and the stra- tegic management perspectives.72 Accordingly, these restructuring strategies can be regarded as the attempt of Lai (1997) to establish a comprehensive and integrative re- structuring framework based on previous literature.73 These four generic restructuring strategies serve as the basis for my empirical analysis below. However, I follow the rationale elaborated by Eichner (2010) in replacing asset restructuring as defined by Lai and Sudarsanam (1997) by the term portfolio restructuring. This term seems better suited to distinguishing between significant changes to the business portfolio, such as divestments and acquisitions, and changes in capital expenditures that are regarded as an aspect of operational restructuring.74 Table 1: Generic Restructuring Strategies Defined by Lai and Sudarsanam (1997) Restructuring Strategy Individual Restructuring Actions Operational Cost reduction, improved financial control, closures and integration of production and other facilities Financial Equity-based (cash equity issue, dividend cuts or omission) and debt-based (debt refinancing, debt renegotia- tion) Managerial Replacement of CEO, chairman or managing director Asset New investments (e.g. acquisitions, capital expenditures in plant and machinery) and asset reductions (e.g. divestments, management buy-outs, spin-offs, sale-and-leaseback transactions) Source: Based on Lai and Sudarsanam (1997), pp. 207-209. 69 See Lai and Sudarsanam (1997), p. 198. 70 See Lai and Sudarsanam (1997), pp. 199-206, for a detailed discussion of the various conflicts of interest. 71 See Lai and Sudarsanam (1997), p. 207 and p. 209, and Lai (1997), pp. 72-75. 72 See the extensive review of prior turnaround literature in Lai (1997), pp. 72-82. 73 See Lai and Sudarsanam (1997), p. 198. 74 See Eichner (2010), p. 53, who draws on Bowman and Singh (1993), p. 8. 2.2 Financial Distress and Corporate Restructuring 15 Robbins and Pearce (1992) establish a seminal turnaround process framework that comprises two overlapping stages.75 The first stage of the turnaround process is called retrenchment and the second stage is called recovery stage. Cost-cutting and asset reductions are characteristic of the retrenchment stage whose aim is to ensure survival, reestablish positive cash flows and improve operational efficiency. The re- covery stage involves targeted investments and aims at establishing long-term profita- bility and conquering new markets. The process aspect had already been presented by Schendel, Patton, and Riggs (1976) in the context of corporate turnaround and Petti- grew (1987a) in the context of managing strategic change. According to the framework put forward by Pettigrew (1987a), the management of strategic change hinges on three important aspects to be considered successful: the content of the strategy, managing the process of change and taking the context into account.76 These aspects of success- ful strategic change lay the foundation for my research model, as described in more detail in chapter 4 below. Another important contribution to restructuring literature is provided by Arogyaswamy, Barker, and Yasai-Ardekani (1995). These authors present a two-stage contingency model for corporate turnaround. Initially, the model firm is faced with declining performance as defined by Schendel, Patton, and Riggs (1976). In other words, the decline is not just a temporary phenomenon. Moreover, the firm’s survival would be at stake if performance did not improve.77 It can thus be assumed that the model firm is in financial distress. The distressed firm responds by launching so-called decline-stemming strategies in the first stage whose aim is to increase effi- ciency.78 The scope of the decline-stemming strategy is a function of the severity of the performance decline and the available slack resources.79 In the second stage, the dis- tressed firm will implement recovery strategies subject to the initial causes of decline and the firm’s competitive position in the market.80 The stages need not be sequential 75 See Robbins and Pearce (1992), pp. 290-291. 76 See Pettigrew (1987a), p. 657. 77 See Arogyaswamy, Barker, and Yasai-Ardekani (1995), p. 497, relying on Hofer (1980). 78 See Arogyaswamy, Barker, and Yasai-Ardekani (1995), p. 498. 79 See Arogyaswamy, Barker, and Yasai-Ardekani (1995), pp. 498-499. 80 See Arogyaswamy, Barker, and Yasai-Ardekani (1995), p. 499. 2.3 Bankruptcy, Reorganization and Post-Bankruptcy Performance 18 gence.92 The same holds true for the pre-filing size of the firm.93 Additionally, compa- nies with higher pre-bankruptcy industry-adjusted operating margins exhibit superior post-bankruptcy performance.94 Denis and Rodgers (2007) document a positive rela- tionship between improvements to the operating margin during Chapter 11 and post- bankruptcy performance.95 Conversely, companies that do not manage to improve their operating margin in Chapter 11 are more likely to experience further financial distress later on.96 Both the pre-filing firm and industry operating margin show a negative rela- tion to the time spent in Chapter 11.97 With regard to the initial outcome of the Chapter 11 filings, Denis and Rodgers (2007) conclude that the pre-bankruptcy industry- adjusted operating margin of emerging firms is significantly greater than that for liqui- dated or acquired firms.98 Furthermore, larger firms show a higher likelihood of reor- ganizing instead of being liquidated or acquired.99 The work of Denis and Rodgers (2007) can be criticized, as they compare changes in firm characteristics such as firm size or the leverage ratio from the last 10-K before filing (denoted as F-1) to the last 10-K before the resolution (denoted as R-1).100 They therefore do not include the pre- sumably beneficial effects of the reorganization in their analysis of post-bankruptcy performance. By contrast, I compare the last available 10-K before the filing to the first 10-K after emergence to ensure that the effects of the reorganization are reflected therein.101 From a legal perspective, LoPucki and Whitford (1993b) document mixed re- sults for a sample of bankrupt firms between 1979 and 1988. Some firms are consi- dered successful because they survive Chapter 11, emerge with their core business in- 92 See Denis and Rodgers (2007), p. 101 and p. 116. This holds true for the industry-adjusted operating margin as performance metric. 93 See Denis and Rodgers (2007), p. 116. Size shows a significant impact on the likelihood of positive operating margin in at least two years after emergence and on the likelihood to survive through three years after emergence. 94 See Denis and Rodgers (2007), p. 104 and p. 118. The operating margin is defined as operating income before depreciation and the liabilities ratio is defined as total liabilities over assets. 95 See Denis and Rodgers (2007), p. 116. 96 See Denis and Rodgers (2007), p. 116 and p. 118. 97 See Denis and Rodgers (2007), p. 118. 98 See Denis and Rodgers (2007), p. 109. 99 See Denis and Rodgers (2007), p. 103. 100 See Denis and Rodgers (2007), p. 116. 101 These effects may, for example, include reductions in leverage resulting from debt renegotiation with debtholders. 2.3 Bankruptcy, Reorganization and Post-Bankruptcy Performance 19 tact and show financial success.102 Yet they also document high leverage ratios with 76% of the sample firms emerging with a leverage ratio above industry benchmarks. Furthermore, they also report a high refiling rate of almost one third of the emerging firms.103 Gilson (1997) finds similar patterns for firms emerging from Chapter 11. He notes that 70% of the emerging firms maintain leverage ratios that are above the re- spective industry median.104 More than 25% of the firms need to refile for bankruptcy or require a second debt restructuring.105 Gilson (1997) argues that firms might choose to keep leverage high upon emergence to make it easier for creditors to monitor the firm’s managers.106 Accordingly, he does not blame these findings on inefficiencies in the Chapter 11 process which is in line with the dynamic liquidation theory of Kahl (2002). Analyzing firms that adopted fresh-start reporting between 1990 and 2003, Heron, Lie, and Rodgers (2009) find that post-bankruptcy debt ratios correlate posi- tively to the pre-filing debt ratios.107 They also document that most firms emerge with debt ratios significantly above industry levels.108 Heron, Lie, and Rodgers (2009) ar- gue that these findings are generally consistent with potential inefficiencies in Chap- ter 11 that can hamper firms’ ability to reset their capital structures to optimal levels.109 Kalay, Singhal, and Tashjian (2007) examine the changes in firms’ operating performance during bankruptcy from the last fiscal year before filing until the first fis- cal year following emergence.110 The sample period spans the years 1991 through 1998. Their key finding is a significant improvement in average operating performance during Chapter 11, which they interpret as net benefits of the Chapter 11 proceed- ings.111 Firms with higher pre-filing debt-to-asset ratios seem to benefit more from Chapter 11, while the complexity of debt renegotiation – measured by the number of 102 See LoPucki and Whitford (1993b), p. 611. 103 See LoPucki and Whitford (1993b), pp. 607-608 and p. 611. 104 See Gilson (1997), pp. 165-166, based on book values of long-term debt to assets. 105 See Gilson (1997), pp. 166-167. 106 See Gilson (1997), p. 190. 107 See Heron, Lie, and Rodgers (2009), p. 742. 108 See Heron, Lie, and Rodgers (2009), p. 742. 109 See Heron, Lie, and Rodgers (2009), p. 742. 110 See Kalay, Singhal, and Tashjian (2007), p. 782. 111 See Kalay, Singhal, and Tashjian (2007), p. 789 and pp. 794-795. 2.3 Bankruptcy, Reorganization and Post-Bankruptcy Performance 20 classes in the reorganization plan – is significantly negatively related to the improve- ment in operating performance.112 For their sample from 1991 through 2004, Lemmon, Ma, and Tashjian (2009) find that the post-bankruptcy performance of financially distressed firms is stronger compared to that of economically distressed firms.113 This evidence seems to lend sup- port to the efficiency of the filtering process in Chapter 11.114 Some scholars have ex- amined factors that could potentially lead to a subsequent refiling for bankruptcy, the so-called “Chapter 22”.115 Bandopadhyaya and Jaggia (2001) find that firms that spend more time under Chapter 11, reduce their leverage and retain more business lines exhibit a lower probability of reentering bankruptcy.116 Several restructuring strategies before and during bankruptcy are analyzed by Datta and Iskandar-Datta (1995). For a sample of bankrupt U.S. firms from 1980-1989, they document that, dur- ing Chapter 11, almost 50% of the firms replace the top management, 19% lay off per- sonnel and 66% engage in asset restructurings. However, they do not explore the per- formance effects of the different restructuring strategies.117 Alderson and Betker (1999) take a different approach to measuring post- bankruptcy performance. They rely on cash flow figures instead of accruals for their sample of firms that emerged from Chapter 11 between 1983 and 1993.118 They regard their cash flow performance metric as superior to accrual-based operating margins – which are frequently used as proxies for operating cash flow – since these may signifi- cantly deviate from cash flows.119 Their key finding is that the performance of reorga- nized companies is comparable on average to the performance of a benchmark portfo- 112 See Kalay, Singhal, and Tashjian (2007), pp. 794-795. 113 See Lemmon, Ma, and Tashjian (2009), p. 4. 114 See Lemmon, Ma, and Tashjian (2009), p. 1. 115 Altman and Hotchkiss (2006), p. 12. Surprisingly, TransTexas Gas filed for bankruptcy protection four times, as shown by Altman and Hotchkiss (2006), p. 90. 116 See Bandopadhyaya and Jaggia (2001), p. 217. 117 The restructuring strategies analyzed are financial restructuring, asset restructuring, governance restructuring and labor recontracting. See Datta and Iskandar-Datta (1995), p. 19. 118 See Alderson and Betker (1999), pp. 69-70. 119 See Alderson and Betker (1999), p. 79. They specifically name asset sales and other transactions as factors that cause operating margins to deviate from cash flows. Other factors include, for instance, capital expenditures and changes in net working capital. 3.1 Principles of U.S. Bankruptcy Law 23 3 Corporate Bankruptcy in the U.S. – Liquidation and Reorganiza- tion Reorganization under Chapter 11 is one way of resolving corporate bankruptcy in the U.S. To put the reorganization under Chapter 11 into a broader legal perspective, this chapter explores some of the general provisions of U.S. bankruptcy law and presents the basic characteristics and governing rules of both liquidation under Chapter 7 and reorganization under Chapter 11.127 3.1 Principles of U.S. Bankruptcy Law Aghion, Hart, and Moore (1994) name two goals of any bankruptcy law. Be- sides resolving insolvency and financial distress in a timely manner, the assets of in- solvent and financially distressed firms shall be disposed of in a socially efficient way.128 According to White (2007), corporate bankruptcy can be characterized as the legal process in which firms in financial distress resolve their debts.129 Bankruptcy law prescribes rules as to which debtor’s assets are to be used to repay debts (“size of the pie”130) and how the proceeds are divided among the creditors (“division of the pie”131). The current U.S. bankruptcy law was enacted by U.S. Congress with the Bank- ruptcy Reform Act of 1978, which took effect on October 1, 1979. The Bankruptcy Reform Act of 1978 is commonly referred to as the Bankruptcy Code, which can be found in Title 11 of the United States Code (U.S.C.).132 Since 1979, the Code has been amended several times, one recent example being the Bankruptcy Abuse Prevention 127 This chapter provides an overview of the most relevant aspects of U.S. bankruptcy law with particular regard to reorganization under Chapter 11. For a detailed account of U.S. bankruptcy law, see e.g. Baird (2006) or Epstein and Nickles (2007). 128 See Aghion, Hart, and Moore (1994), p. 215. 129 See White (2007), p. 1016. 130 White (2007), p. 1016. 131 White (2007), p. 1016. 132 See Epstein and Nickles (2007), p. 2. 3.1 Principles of U.S. Bankruptcy Law 24 Consumer Protection Act (BAPCPA) in 2005.133 The most important changes intro- duced by the BAPCPA include limiting the exclusivity period for the management to produce a reorganization plan to 18 months and granting the debtor only a one-time extension of 90 days to assume or reject executory contracts and unexpired leases.134 Two important principles lay the foundations of U.S. bankruptcy law. First, the automatic stay relates to the legal consequences of a bankruptcy filing.135 Once a deb- tor has filed for bankruptcy, the automatic stay provides for instantaneous protection of the debtor against individual creditors trying to enforce their pre-petition claims stante pede.136 Returning to the analogy of the pie, the automatic stay thus guarantees the pure existence of the pie that can be orderly divided among the creditors. Second, the absolute priority rule (APR) provides guidelines on how the pie is to be distributed among the creditors.137 The APR establishes a hierarchy according to which claims are to be settled. First in line are secured senior lenders, followed by unsecured junior lenders and, lastly, the equityholders as residual claimants.138 In Chapter 7 liquidations, the APR is used to distribute the proceeds of asset sales to the creditors. According to the APR, a higher priority class needs to be paid in full before a lower priority class receives anything.139 Generally, the APR applies also to Chapter 11 reorganizations.140 Deviations from the APR are, however, permitted by the Bankruptcy Code in Chap- ter 11 reorganizations and do occur, as documented by prior research.141 The APR can 133 See Altman and Hotchkiss (2006), pp. 26-28 and pp. 47-55, Epstein and Nickles (2007), p. 2 or Baird (2006), p. 6. According to Altman and Hotchkiss (2006), p. 49, and Bharath, Panchapagesan, and Werner (2010), p. 13, most provisions of the BAPCPA took effect on Oc- tober 17, 2005. None of the firms in my final sample filed for bankruptcy on or after October 17, 2005. It seems therefore reasonable to conclude that my analyses are not subject to any bias linked to the BAPCPA. 134 See Altman and Hotchkiss (2006), p. 48. 135 Refer to 11 U.S.C. § 362. 136 See e.g. Epstein and Nickles (2007), p. 15 or Baird (2006), p. 207. 137 See e.g. White (2007), p. 1016 or Baird (2006), pp. 66-67. 138 See generally Bris (2008). 139 See White (2007), p. 1019. However, some scholars, such as Bebchuk and Fried (1996), p. 934, argue that strict adherence to the APR also has its costs. These costs include excessive use of security interests, reducing the firm’s incentives to carefully select investment projects and, according to Bebchuk and Fried (1996), p. 934, introducing distortions to the monitoring mechanism between creditors and borrowers. 140 See Baird (2006), pp. 81-86. 141 See Weiss (1990), Eberhart, Moore, and Roenfeldt (1990), Franks and Torous (1989) and the discussion in Eberhart and Weiss (1998). A more recent study is Carapeto (2000). The notion of frequent deviations from APR (often documented in the 1980s) has changed in re- cent years, as e.g. Ayotte and Morrison (2009) find in their study for petitions filed in 2001. They argue that deviations in favor of equi- tyholders are largely replaced by creditor control. Similar arguments can be found in Adler, Capkun, and Weiss (2006) and Baird and Rasmussen (2003). 3.2 Liquidation under Chapter 7 25 thus be considered as a starting point for the bargaining about the distribution between the debtor in possession (management and equityholders) and the creditors in Chap- ter 11 reorganizations.142 A debtor may file for bankruptcy protection without being insolvent, i.e. the debtor is still able to satisfy its current obligations and the debtor’s total liabilities do not exceed its total assets.143 Apart from a voluntary filing by the debtor, involuntary filings submitted by creditors are also possible. U.S. firms seeking bankruptcy protec- tion can file for either Chapter 7 or Chapter 11 both of which are explored in more de- tail below. 3.2 Liquidation under Chapter 7 Chapter 7 of the Bankruptcy Code formulates rules on how to liquidate the as- sets of a bankrupt firm. It can be regarded as the benchmark or baseline to which Chapter 11 needs to be compared. The objective of Chapter 7 is to dissolve the compa- ny by selling its assets. The bankruptcy court names a trustee in charge of the dissolu- tion.144 Once the assets have been sold, the proceeds are distributed to the creditors in accordance with the APR.145 Typically, when a firm files for Chapter 7 the value of the assets has already decreased considerably. This mostly results in the equityholders re- ceiving nothing in the final distribution.146 Large public firms usually file for reorganization under Chapter 11. Only in rare cases do they file for liquidation under Chapter 7.147 Sometimes business prospects deteriorate heavily during Chapter 11 leading firms to convert their Chapter 11 cases to Chapter 7 liquidations. Furthermore, some reorganization plans confirmed under Chapter 11 are actually so-called liquidating plans.148 There are two forms of liquidat- 142 See Altman and Hotchkiss (2006), p. 34, and White (1989), p. 139. 143 See Bradley and Rosenzweig (1992), p. 1044, Warren and Westbrook (2000), p. 47, or Baird (2006), p. 9. 144 See White (2007), p. 1019. 145 See White (2007), p. 1019, and generally Bris (2008). 146 This is in line with the absolute priority rule. 147 See Bebchuk (1998), p. 1, White (2007), p. 1022, and Lemmon, Ma, and Tashjian (2009), p. 6. 148 See Warren and Westbrook (2009), p. 611 and 11 U.S.C. § 1123(b)(4). 3.3 Reorganization under Chapter 11 28 court must check that any subsequent reorganization or liquidation is unlikely.167 The company formally emerges from Chapter 11 on the effective date of the reorganization plan, which is usually shortly after confirmation by the bankruptcy court.168 Apart from the standard Chapter 11 procedure described above, firms can file prepackaged bankruptcies. In a prepackaged bankruptcy (alternatively called prepack or prearranged bankruptcy), the reorganization plan is typically filed in conjunction with the actual filing.169 Tashjian, Lease, and McConnell (1996) distinguish between “pre-voted”170 and “post-voted prepacks”170. The difference between the two being that, with the first, the voting procedure for acceptance of the reorganization plan takes place before filing, while, in the latter, the voting takes place after filing. The advan- tage of prepackaged bankruptcies, especially the “pre-voted”170 ones, is that they usually take less time under court supervision and as a result can be less costly. Pre- packs are sometimes also described as a hybrid form between Chapter 11 reorganiza- tion and out-of-court restructurings.171 According to Hotchkiss et al. (2008) prepacks began to replace some out-of-court restructurings as of the 1990s.172 In recent years, more and more bankruptcies under Chapter 11 have resulted in the going-concern sale of the bankrupt firm’s assets.173 According to Jensen (1991), acquiring assets through an auction process can improve the efficiency of the bank- ruptcy process.174 Hotchkiss and Mooradian (1998) document a 45% discount for the acquisition of bankrupt targets relative to non-bankrupt targets.175 They argue that an 167 See 11 U.S.C. § 1129 (a)(11) and LoPucki and Whitford (1993b), pp. 608-609. 168 See Zhang (2010), p. 1722. However, as the case of American Banknote Corporation has shown, the time between confirmation of the reorganization plan and consummation of the plan, i.e. the effective date, can be quite extensive. After the bankruptcy court confirmed the third amended reorganization plan in November 2000, it took another amendment to the plan and almost two more years until American Banknote Corporation emerged from bankruptcy in October 2002. Refer to the 2002 10-K of American Banknote Corpora- tion for more information on the causes of the delay. 169 See Tashjian, Lease, and McConnell (1996), p. 138 or Betker (1995), p. 3. 170 Tashjian, Lease, and McConnell (1996), p. 138. Post-voted prepacks are sometimes referred to as prenegotiated prepacks. See, for instance, Baird and Rasmussen (2003), p. 674. 171 See Tashjian, Lease, and McConnell (1996), p. 135. 172 See Hotchkiss et al. (2008), p. 16, and, for prepacks in general, Tashjian, Lease, and McConnell (1996). 173 See Baird, Bris, and Zhu (2007), p. 4, Skeel (2003), p. 921, and Baird and Rasmussen (2002), pp. 35-36. 174 See Jensen (1991), p. 32. Estimation of the value of the bankrupt firm would thus be left to the market for corporate control. 175 See Hotchkiss and Mooradian (1998), p. 243. 3.3 Reorganization under Chapter 11 29 “[…] acquisition is a substitute for a reorganization in Chapter 11 […]”176. In Chap- ter 11, a business can be sold as a going concern in one of two ways: either through a § 363 sale or as part of a confirmed plan of reorganization.177 One reason why many cases have resulted in § 363 sales over the last years is that the assets can be bought free and clear of all claims without the approval of a reorganization plan.178 The differ- ence between the legal entity and the business entity thus becomes evident.179 The le- gal entity, deprived of most of its assets, remains in Chapter 11, while the business ent- ity (i.e. most of the assets) has been sold off. Consequently, from a legal perspective, the question about the going-concern value of the bankrupt firm (the size of the pie) is separated from the question of the distribution to the creditors and equityholders (who gets how large a slice of the pie). The bankruptcy court is in charge of the distribution only.180 Selling a business as part of a confirmed reorganization plan typically takes more time than a § 363 sale, since creditors have to approve the reorganization plan.181 During reorganization under Chapter 11, the debtor in possession will in many cases be confronted with financing problems. Upon approval by the bankruptcy court, a debtor in possession may obtain new financing means, know as DIP financing or post-petition financing.182 This financing is intended to pay professionals such as law- yers, consultants and accountants during the bankruptcy proceedings, to satisfy work- ing capital requirements and to fund necessary capital expenditures to keep the busi- ness running.183 DIP financing typically enjoys higher seniority and increased securi- ty.184 The terms of the DIP loan are specific to each contract. Nonetheless, they are mostly floating rate notes provided as a short- to medium-term revolving credit line.185 Some legal scholars argue that, through the terms of the DIP financing such as debt 176 Hotchkiss and Mooradian (1998), p. 241. 177 See Hotchkiss and Mooradian (1998), p. 251. § 363 sales refer to 11 U.S.C. § 363. 178 See Baird and Rasmussen (2002), p. 35. 179 Compare to LoPucki and Whitford (1993b), pp. 601-602, on the separation of “business survival” from “entity survival”. 180 Compare to Jensen (1991), p. 32, or Baird and Rasmussen (2002), p. 36. 181 See Hotchkiss and Mooradian (1998), p. 251. 182 See White (2007), p. 1021. 183 See Altman and Hotchkiss (2006), p. 44. 184 See Altman and Hotchkiss (2006), p. 44 and 11 U.S.C. § 364. 185 See Altman and Hotchkiss (2006), pp. 44-45. 3.3 Reorganization under Chapter 11 30 covenants, creditors have de facto taken control of the debtor in possession in recent years, contrary to the widely held view that Chapter 11 would be debtor-friendly.186 Finally, U.S. bankruptcy law provides the debtor in possession with many op- tions to facilitate a successful fresh start after emergence. As shown in chapter 2.1, these include the right to terminate pension plans under certain conditions. Pension plans can be stopped under Chapter 11 and handed over to the public Pension Benefit Guaranty Corporation (PBGC).187 Furthermore, the debtor in possession has the right to reject executory contracts and unexpired leases.188 186 See Muro (2008), pp. 3-4, Baird and Rasmussen (2002) and Skeel (2003), p. 918. Baird and Rasmussen (2009), p. 30, provide a list of covenants and provisions approved by the bankruptcy court in DIP financing orders. 187 See White (2007), p. 1021, or Datta and Iskandar-Datta (1995), p. 28. 188 See White (1989), pp. 144-145, and White (2007), p. 1021, who states that penalties for breach of contract are assigned to the class of general unsecured claims that will only be satisfied at the class’ pro-rata distribution under the plan of reorganization. 4.1 Definitions 33 4.1.2 Chapter 11 Outcomes and Post-Bankruptcy Outcomes Table 2 provides an overview of the Chapter 11 outcome definitions used in this study. Firms emerging from Chapter 11 can be classified as reorganized (public or pri- vate), merged, liquidated or dismissed. Table 2: Chapter 11 Outcomes Outcome Definition Reorganized public Firms that left Chapter 11 as going concerns having had their reorganization plans confirmed (or all or substantial- ly all assets were sold in a § 363 sale) and continued to file documents with the SEC (i.e. remained public enti- ties). In addition, firms that were acquired, but that continued to operate independently and filed documents with the SEC Reorganized private Firms that left Chapter 11 as going concerns having had their reorganization plans confirmed (or all or substantial- ly all assets were sold in a § 363 sale), but that no longer filed documents with the SEC (i.e. firms that went private). In addition, firms that were acquired, continued to operate independently, but that no longer filed docu- ments with the SEC Merged Firms that were acquired and were merged into the acquiring firm, i.e. that lost their status as independent entities (including § 363 sales and sales as part of the reorganization plan) Liquidated Firms that either had a liquidating plan confirmed or were converted to Chapter 7. Assets were sold piecemeal Dismissed Firms that filed for Chapter 11, but the bankruptcy court dismissed the case Source: Author’s own illustration, influenced by Hotchkiss (1993) and Denis and Rodgers (2007). Firms that emerge as reorganized entities have either had their reorganization plans confirmed or all or substantially all assets were sold in a § 363 sale. It is crucial that these firms continue to operate as independent entities to be classified as reorga- nized.198 The legal entity of a reorganized company can either remain unchanged or be newly established. In some cases, the reorganized company changes its name before, upon or shortly after emergence. Continuing to file documents with the SEC distin- guishes firms that reorganized as public entities (reorganized public) and those that reorganized as private entities (reorganized private). In Chapter 11, companies can be acquired either through a § 363 sale or through a confirmed reorganization plan.199 Compared to prior literature, the distinction be- tween acquired and merged is crucial, since performance improvements of firms that 198 See Hotchkiss (1993), p. 11 of the second essay. 199 See Hotchkiss and Mooradian (1998), p. 251, who explain in detail what differentiates the two methods from each other. Compare also to chapter 3.3. 4.1 Definitions 34 were merged into the acquiring firm cannot be accurately disentangled from perfor- mance improvements induced by the acquiring firm.200 Accordingly, I classify these firms as merged. Conversely, I assign firms that have been acquired but that continue to operate as independent public entities to the reorganized public group. This is in line with Hotchkiss (1995) and Bandopadhyaya and Jaggia (2001).201 Firms that are liquidated see their assets sold piecemeal either pursuant to a li- quidating plan under Chapter 11 or through conversion to Chapter 7. Most of the time, the reason for liquidation is that the estimated going concern value falls below the li- quidation value.202 In a few cases, petitions for bankruptcy protection are dismissed by the bankruptcy court. These cases are excluded from the analyses.203 Some cases may be dismissed because the debtor and its creditors have been able to reach an agreement which they had not been able to work out outside the courtroom before the filing.204 To summarize, firms are only considered for the analysis of post-bankruptcy performance if they emerged from Chapter 11 as independently operating public entities.205 Since I am interested in the ultimate fate of firms that emerge from bankruptcy, it is important to define possible post-bankruptcy outcomes. These can be public, pri- vate, merged, refiled or liquidated. These are summarized and defined in Table 3. Firms are tracked from the year of emergence from Chapter 11 (E) until three years after emergence (E+3).206 200 Hotchkiss and Mooradian (1998) analyze acquisitions in Chapter 11, but from a transactional point of view. They conclude that acquir- ing (parts of) an insolvent firm can create value for the acquiring firm. Including firms in the reorganized public outcome group that were acquired during Chapter 11 and remained independently operating public entities reflects the prevailing reality in Chapter 11 as documented by e.g. Baird and Rasmussen (2003), p. 691. They stipulate that the "dominant feature of the large corporate Chapter 11 today is the asset sale." 201 See Hotchkiss (1995) who builds on Hotchkiss (1993), p. 11 of the second essay, and Bandopadhyaya and Jaggia (2001), p. 203. 202 See Hotchkiss (1993), pp. 11-12 of the second essay. 203 Of the initial sample, 7 cases are dismissed. 204 See Warren and Westbrook (2009), p. 611. 205 Compare to Hotchkiss (1993), p. 11 of the second essay. The same definition is applied in Hotchkiss (1995). Denis and Rodgers (2007) use a similar definition, but exclude firms that were acquired in Chapter 11 from their post-bankruptcy performance analyses. 206 This choice conforms to time horizons used by prior research such as Denis and Rodgers (2007). Refer also to chapter 4.2.1. 4.2 Research Model 35 Table 3: Post-Bankruptcy Outcomes Outcome Definition Public Firms that continued to operate as independent public entities through E+3 Private Firms that went private through E+3 Merged Firms that were acquired and merged into the acquiring public firm through E+3 Refiled Firms that refiled for Chapter 11 or Chapter 7 after their initial emergence through E+3 Liquidated Firms that were liquidated after their initial emergence through E+3 Source: Author’s own illustration, influenced by Hotchkiss (1993) and Denis and Rodgers (2007). 4.2 Research Model The aim of this study is to identify restructuring strategies and actions that are associated with a higher probability of post-bankruptcy success in a multivariate set- ting. The research model is grounded in the conceptual work of Pettigrew (1987b), who introduced the triangle of content, process and context recommended for any meaningful strategic management research and especially the management of strategic change.207 Restructuring in a bankruptcy setting can also be understood as the man- agement of strategic change in a broader sense. In the context of this study, content builds on the four generic restructuring strategies introduced by Lai and Sudarsanam (1997) that are intended to turn around the bankrupt firm and eventually facilitate post-bankruptcy success. The process aspect is operationalized by two distinct stages referring to (i) the bankruptcy phase, begin- ning when the firm files for bankruptcy (F) and ending when it emerges from Chap- ter 11 (E), and (ii) the post-bankruptcy phase, which is defined as three full fiscal years following emergence (E+1 through E+3).208 The process aspect seems especially im- portant in the bankruptcy context, since the institutional framework during bankruptcy is substantially different to the post-bankruptcy phase. This becomes evident when one considers that the firm acts as debtor in possession during bankruptcy supervised by the bankruptcy court. The automatic stay provides relief from honoring debt payment 207 See Pettigrew (1987b), pp. 4-6, and Pettigrew (1987a), p. 657. 208 This process approach resembles the approach taken by Eichner (2010) for analyzing the effectiveness of restructuring actions of non- bankrupt manufacturing firms. 4.2 Research Model 38 Table 4: Performance Metrics of Selected Post-Bankruptcy Performance Studies Author(s) Performance Metric Accounting-Based Performance Metrics Lemmon, Ma, and Tashjian (2009) EBITDA scaled by total assets (also as industry-adjusted value) Denis and Rodgers (2007) Operating income before depreciation scaled by total assets (also as industry-adjusted value) Kalay, Singhal, and Tashjian (2007) EBITDA scaled by total assets (also as industry-adjusted value and as normalized value scaled by the industry standard deviation) Dawley, Hoffman, and Brockman (2003) Return on assets (also as industry-adjusted value) Alderson and Betker (1999) • Net cash flows to claimholders, defined as net cash flow from operations + net cash flow from investment + cash interest paid – change in cash – other cash flows from financing • EBITDA scaled by sales (as industry-adjusted value) Maksimovic and Phillips (1998) • Plant-level productivity • Operating cash flows (also as industry-adjusted value) Hotchkiss and Mooradian (1997) Operating income, i.e. net sales - COGS - SG&A before depreciation and amortization (also scaled by total assets or sales, or as industry-adjusted value) Hotchkiss (1995) Operating income, i.e. net sales - COGS - SG&A before depreciation and amortization (also scaled by total assets or sales, or as industry-adjusted value) Market-Based Performance Metrics Jory and Madura (2010) Stock price performance Eberhart, Altman, and Aggarwal (1999) Stock price performance Source: Author’s own illustration. Within the class of accounting-based performance metrics two major groups emerge from prior research. While Hotchkiss (1995), Denis and Rodgers (2007) and Kalay, Singhal, and Tashjian (2007) use operating performance metrics such as operat- ing income or EBITDA, others, such as Asquith, Gertner, and Scharfstein (1994), Jos- tarndt and Sautner (2010) or Eichner (2010) use performance metrics that additionally take financial or investment activities into account.212 These metrics include, for in- stance, EBITDA less interest expenses or EBITDA less interest expenses less capital expenditures. These metrics typically serve as free cash flow proxies.213 Hotchkiss (1995) uses a similar free cash flow proxy for her descriptive statistics to compare her results to the operating performance metric.214 Similar cash flow proxies have been 212 Another distinction could be made between accrual-based accounting metrics and cash flow-based accounting metrics. However, in most of the related prior studies cash flow-based metrics are approximated using accrual-based metrics which presumably contain some non-cash items. For a more detailed discussion of the differences between accruals and cash flows, refer to Dechow (1994). 213 See Jostarndt and Sautner (2010), pp. 16-17 and Asquith, Gertner, and Scharfstein (1994), p. 628. These metrics are often interpreted as interest coverage ratios. 214 See Hotchkiss (1995), p. 9. 4.2 Research Model 39 used in other research fields such as post-merger performance (Healy, Palepu, and Ru- back (1992)) or buyout performance (Guo, Hotchkiss, and Song (2011)).215 Table 5: Performance Metrics of Selected Restructuring Studies Author(s) Performance Metric Accounting-Based Performance Metrics Eichner (2010) Interest coverage, i.e. EBITDA – capital expenditures – net interest expenses Jostarndt and Sautner (2010) Interest coverage, i.e. EBIT < interest expenses Jostarndt and Sautner (2008) Interest coverage, i.e. EBIT < interest expenses Buschmann (2006) Return on investment, i.e. EBT/(fixed assets + working capital) Kahl (2001) EBITD scaled by total assets or sales (also as industry-adjusted value) Asquith, Gertner, and Scharfstein (1994) Interest coverage, i.e. EBITDA < interest expenses (also scaled by total assets, or as industry- adjusted value) Market-Based Performance Metrics Lai and Sudarsanam (1997) Stock price performance Ofek (1993) Stock price performance Source: Author’s own illustration. Often, these metrics are scaled either by total assets or net sales to produce a re- turn on assets (ROA) or return on sales (ROS) which can be better compared over time and across firms.216 In cross-industry studies, performance metrics are often scaled by subtracting the respective industry median matched by SIC codes. Industry medians are defined as the contemporaneous values for a specific item (e.g. total assets) of all firms appearing in a given 3-digit SIC group. In line with Denis and Rodgers (2007) it is required that at least five distinct firms form the basis for calculating the industry median.217 Where this criterion is not met, I have moved from 3-digit SIC groups to 2- digit SIC groups, and ultimately to 1-digit SIC groups. As Hotchkiss (1993) points out, the comparison of emerged firms’ performance with the median industry performance might overstate both positive results and negative results for the bankrupt firms in rela- 215 See Healy, Palepu, and Ruback (1992), p. 139, using pretax cash flow return on assets to measure operating performance improvements. The pretax cash flow is defined as sales less cost of goods sold less selling, general and administrative expenses before depreciation and amortization. Guo, Hotchkiss, and Song (2011), p. 515, use net cash flow defined as EBITDA less capital expenditures to measure firm performance. 216 See Healy, Palepu, and Ruback (1992), p. 139. 217 See Denis and Rodgers (2007), p. 104. 4.2 Research Model 40 tion to the respective industry.218 This is due to the fact that many emerging companies adopt fresh-start reporting, recording assets at fair market values which are presumably below the book values recorded at historic costs which probably apply to the respec- tive industry peers.219 This problem is overcome by using sales instead of total assets as a scaling factor in line with Hotchkiss (1995).220 In contrast to prior post-bankruptcy performance studies, which focus on firm and industry characteristics and only selectively examine specific restructuring actions such as replacing the CEO or reducing leverage, this study scrutinizes the restructuring actions employed by bankrupt firms in a comprehensive manner.221 It therefore seems appropriate to alter the performance metric, since the variety of potential restructuring actions is not restricted to improving a firm’s operating performance, but also aims to improve a firm’s financial position and investments.222 Accordingly, I use a proxy for pretax free cash flow as a performance metric in this study. The pretax free cash flow proxy is defined as shown in Table 6.223 My definition draws on prior literature by As- quith, Gertner, and Scharfstein (1994) and the extension employed by Eichner (2010), who additionally takes capital expenditures into account.224 My choice is corroborated by Alderson and Betker (1999), who do not use operating margins for their analysis of 218 See Hotchkiss (1993), pp. 14-15 of the first essay. 219 Fresh-start reporting in accordance with Statement of Position (SOP) 90-7 of the American Institute of Certified Public Accountants (AICPA) refers to financial reporting by entities in reorganization under the Bankruptcy Code. The provisions of SOP 90-7 can be ap- plied subject to two conditions: (i) the reorganization value of the assets of the emerging entity before confirmation is less than the sum of all post-petition liabilities and allowed claims (i.e. the firm is insolvent in the stock-based definition) and (ii) holders of voting shares before confirmation receive less than 50% of the voting shares of the emerging company (i.e. a substantial change in ownership takes place). If these conditions are met, the firm should apply fresh-start reporting upon emergence from Chapter 11. The application of fresh-start reporting entails allocating the reorganization value of the entity to the assets of the emerging firm which will result in many if not all assets being written down to their fair market values. Liabilities are also set to their fair values (calculated as discounted ex- pected cash flows). For further details refer to e.g. Lehavy (2002) or Heron, Lie, and Rodgers (2009). 220 See Hotchkiss (1995), p. 8. 221 Refer to the literature review in Hotchkiss et al. (2008), pp. 31-35. 222 A simple example serves to illustrate this: Assume a firm manages to reduce its debt significantly over the course of the three years after emergence. This is likely to reduce its interest payments in absolute terms and should, ceteris paribus, contribute to better performance. This, in turn, is not directly reflected in any operating performance metric (e.g. operating income). Only indirect effects could be cap- tured by an operating performance metric, such as spending the saved interest payments for marketing purposes, possibly resulting in higher sales. Applying a broader performance metric that takes interest payments into account would directly measure this effect. 223 For the remainder of this study, I refer to the proxy for pretax free cash flow as free cash flow. 224 Typically, free cash flow includes changes in net working capital. I do not uphold this due to limitations in data availability. Kaplan (1989), p. 224, provides similar reasons for not including changes in net working capital in his cash flow measure. The limitations in data availability mainly relate to firms whose post-bankruptcy data was not available in Worldscope and which had to be manually ex- tracted from company filings which would have been outside the scope of this study. This is consistent with prior literature in the re- structuring and turnaround field. Eichner (2010), for instance, leaves changes in net working capital out too. 4.2 Research Model 43 Table 8: Accounting-Based Definitions of Turnaround Success of Selected Studies Author(s) Definition of Turnaround Success Eichner (2010) (EBITDA – capital expenditures – net interest expenses) > 0 three years after the onset of distress Jostarndt and Sautner (2010) Firm in financial distress avoids bankruptcy filing and completes debt restructuring Buschmann (2006) EBT/(fixed assets + working capital) above 9% for at least two years and long-term average of at least 5% Kahl (2001) No explicit success definition. Firm performance measured as industry-adjusted EBITD scaled by total assets for up to five years following distress resolution Source: Author’s own illustration. Based on the above mentioned contributions from post-bankruptcy literature and restructuring literature, I define post-bankruptcy success as returning to a non- negative free cash flow in year E+3, as defined in Table 6. This means that the compa- ny generates enough cash to cover its operating costs, interest payments and capital expenditures. Post-bankruptcy success is coded as a dichotomous variable as opposed to a continuous variable, following the rationale put forward by Hotchkiss (1993).230 To allow for an easier comparison among firms and industries, free cash flow is either scaled by sales or the respective industry median is subtracted.231 The treatment of the post-bankruptcy outcomes refiled, liquidated, merged and private is explained in more detail below.232 In addition to the performance-based defi- nition of success using the free cash flow metric, these four outcomes determine whether or not a firm should be counted as being successful after its emergence. Clas- sifying companies that refiled or liquidated as failed seems straightforward and is sup- ported by Denis and Rodgers (2007) and Hotchkiss (1995).233 Bandopadhyaya and Jaggia (2001), however, do not regard a second bankruptcy filing as failure a priori. Taking the point of view of creditors, they argue that a second reorganization can be deemed successful if the creditors receive at least as much as they would have received 230 See Hotchkiss (1993), p. 21 of the first essay. The potential influence of missing observations is mitigated by employing performance groups in the logistic model as opposed to continuous accounting variables. Refer to chapter 5 for a description of the treatment of missing data in this study. 231 Scaling free cash flow by total sales yields a return on sales performance metric. In line with Hotchkiss (1995), scaling by sales is preferred to scaling by total assets since some firms adopt fresh-start reporting which could bias any performance metric scaled by total assets. 232 Refer also to chapter 5.3.1 for the treatment of firms that refiled, liquidated, merged or went private in the sample selection process. 233 See Denis and Rodgers (2007), p. 116 and Hotchkiss (1995), p. 17. 4.2 Research Model 44 in a liquidation.234 While this is true from the creditors’ perspective, a second filing violates one pivotal condition on which the bankruptcy court based its confirmation of the reorganization plan of the initial bankruptcy. This condition requires that “[c]onfirmation of the plan is not likely to be followed by the liquidation, or the need for further financial reorganization, of the debtor or any successor to the debtor under the plan, unless such liquidation or reorganization is proposed in the plan.”235 Conse- quently, I treat refiling and liquidation as failure. Denis and Rodgers (2007) and Hot- chkiss (1995) even regard a subsequent out-of-court restructuring as evidence of fail- ure.236 I refrain from following this example, however, as my definition of success is primarily based on the performance metric. If out-of-court restructuring is mirrored in negative firm performance, this will in any case lead my performance metric to indi- cate failure. According to the model by Kahl (2002), one should observe more liquida- tions (i.e. real failures in my terminology) when firms reenter financial distress.237 As- signment to the successful or unsuccessful group is less clear-cut for firms that merged or went private after emerging. While Hotchkiss (1993) assumes that merged firms are to be treated as successful, I do not automatically treat merged and private firms as successes.238 Instead, these firms are categorized according to their performance one year before merging or going private, which seems to be more plausible.239 Measurement to determine whether post-bankruptcy performance can be re- garded as a success takes place at the end of year E+3. This corresponds to a typical time horizon in prior post-bankruptcy performance literature since both Denis and Rodgers (2007) and Hotchkiss (1995) track their accounting data for up to three 234 See Bandopadhyaya and Jaggia (2001), p. 202. 235 11 U.S.C. § 1129(a)(11). Refer also to LoPucki and Whitford (1993b), pp. 608-609. 236 See Denis and Rodgers (2007), p. 116 and Hotchkiss (1995), p. 17. 237 See Kahl (2002), p. 157. 238 See Hotchkiss (1993), p. 21 of the first essay. 239 This entails measuring the restructuring actions over a shorter period of time to avoid any endogeneity by construction, as explained in chapter 4.2.2. 4.2 Research Model 45 years.240 Eichner (2010) tracks his distressed firms for three years from the onset of financial distress.241 4.2.2 Restructuring Actions as Independent Variables This section examines which restructuring actions tend to influence post- bankruptcy performance, how these are operationalized and which hypotheses are de- ducted from prior theoretical and empirical research. In contrast to prior contributions such as Denis and Rodgers (2007) and Hotchkiss (1995), I explicitly model restructur- ing actions depending on the phase of the restructuring process in which they are im- plemented. Furthermore, relying on results from prior restructuring research, I extend the scope of analyzed restructuring actions in the bankruptcy context.242 The categori- zation into operational, financial, managerial and portfolio restructuring strategies fol- lows Lai and Sudarsanam (1997) and Eichner (2010).243 Table 9 below summarizes the definitions of each independent variable which will be detailed in the following. The reference period is always the last available fiscal year before filing (F-1) for the bank- ruptcy period and the fiscal year of emergence (E) for the post-bankruptcy phase. To avoid any endogeneity by construction in the post-bankruptcy phase, the independent variables are measured from E through E+2, while the dependent variable is measured with a lag of one year in E+3.244 The lagged measurement to avoid or mitigate any en- dogeneity is standard in other research areas such as asset pricing.245 Nevertheless, I explicitly test the exogeneity assumption for the restructuring actions in the post- bankruptcy phase in chapter 6.3.2. 240 See Hotchkiss (1995), p. 17 and Denis and Rodgers (2007), p. 116. Hotchkiss (1995), p. 17, tracks her sample firms for up to five years to check whether they liquidate, refile or need a private workout during this time. Her main regression models take a time horizon of three years into account. As can be seen in Table 7 and Table 8, some authors use shorter time periods (e.g. Lemmon, Ma, and Tashjian (2009) with two years) while some use longer time periods (e.g. Kahl (2001)). 241 See Eichner (2010), p. 71. 242 Refer to chapter 2.3 for more details of post-bankruptcy performance literature. 243 Refer to chapter 2.2 for a more detailed motivation of this approach. 244 In line with Eichner (2010), p. 173, I assume that restructuring actions during bankruptcy can be deemed exogenous, since they differ by at least three years from the measurement of post-bankruptcy success in E+3. 245 See, for instance, Welch and Goyal (2008). In another context, notably the influence of security class actions on takeovers and CEO disciplinary events, Humphery-Jenner (2012), p. 159, applies the same technique. 4.2 Research Model 48 phase, where the firm is given the chance of a fresh start to recover and possibly grow its market share again. 4.2.2.1.1 Increasing Sales Increasing sales seems to be a natural way to overcome distress or bankruptcy and has been suggested by contributions from the restructuring literature such as the theoretical contribution of Hofer (1980).254 Revenue generation is one of the four ge- neric operating turnaround strategies put forward by Hofer (1980).255 Notwithstanding, the positive effect of sales-increasing actions may be difficult to realize in many bank- ruptcy situations, especially when taking into account that e.g. increased marketing efforts do only translate into higher sales after a time lag. It therefore seems appropri- ate to distinguish the potential effects of sales-increasing actions in terms of the time when they are employed (bankruptcy phase vs. post-bankruptcy phase) and in terms of the underlying reason for distress (economic vs. financial distress). During the bankruptcy phase it may be difficult to implement sales-increasing actions successfully as management will likely be distracted by other, more urgent top- ics. These might include convincing suppliers to continue supplying, retaining key employees, receiving DIP financing to keep the business running, and preventing im- portant customers from turning their back on the bankrupt firm. Once the business emerges from bankruptcy as a reorganized going concern, the perspective is likely to change. In this situation, management might define sales-increasing actions in an at- tempt to regain lost market share, as shown by Hofer (1980), or to conquer new mar- kets. In the restructuring literature, these different notions of sales-increasing actions are highlighted by Buschmann (2006).256 It is important to distinguish between economically and financially distressed firms. While increasing sales for economically distressed firms could even be counter- 254 See Hofer (1980), p. 26. Similarly, Hambrick and Schecter (1983), p. 233. Sudarsanam and Lai (2001), p. 185, support the suitability of sales-increasing strategies to overcome financial distress. 255 See Hofer (1980), p. 20. 256 See Buschmann (2006), pp. 54-56. 4.2 Research Model 49 productive if some of the products sold return negative contribution margins, firms that went bankrupt mainly for financial reasons are not a priori susceptible of selling unprofitable products with negative contribution margins.257 Sales-increasing actions should therefore always be a function of the primary underlying distress cause. Dis- continuing unprofitable product lines (and reducing sales accordingly) could be bene- ficial to economically distressed firms, whereas financially distressed firms might ben- efit if sales increase due to better capacity utilization and greater economies of scale.258 Empirically, Nothardt (2001) finds that sales-increasing actions exhibit a signif- icant positive influence on the turnaround probability.259 Both Eichner (2010) and Buschmann (2006) find no significant relation between sales-increasing actions and the turnaround probability.260 Empirical evidence for bankrupt firms is sparse. Kalay, Singhal, and Tashjian (2007) document for their U.S. sample that firms reduce sales on average by 14% while in Chapter 11. This is interpreted as focusing on the core of the business.261 Taking the ambivalent notion of sales-increasing actions as a whole, I for- mulate the following hypothesis: H1: Pursuing sales-increasing actions in Chapter 11 (out of Chapter 11) is not (posi- tively) related to the probability of post-bankruptcy success.262 Sales increase is operationalized as an increase in net sales or revenues by at least 10% compared to the reference period. Like Eichner (2010), I use a 10% thre- shold to grasp significant changes in the sales level only.263 Additionally, to control for any inorganic sales growth (through acquisitions), I include an interaction term of both sales growth and acquisitions in my analysis.264 257 See Buschmann (2006), p. 55. 258 See Buschmann (2006), p. 55. 259 See Nothardt (2001), p. 271. 260 See Eichner (2010), p. 215 and Buschmann (2006), p. 187. 261 See Kalay, Singhal, and Tashjian (2007), pp. 789-790. 262 To make reading easier, the presented hypotheses are always formulated as alternative hypotheses. 263 See Eichner (2010), p. 136. 264 Eichner (2010), p. 138, states that he corrected sales growth for any acquisitions. However, he provides no details of how this correction was operationalized. 4.2 Research Model 50 4.2.2.1.2 Reducing Costs Reducing costs as part of the retrenchment stage of a turnaround is widely ac- knowledged, for instance in the two-stage turnaround model of Robbins and Pearce (1992).265 They argue that firms in distress “should activate the turnaround strategy by sharply reducing operational costs through an aggressive retrenchment response.”266 However, as Arogyaswamy, Barker, and Yasai-Ardekani (1995) point out, cost cutting alone may not be sufficient to turn the firm around.267 The results of cost-cutting ac- tions usually materialize faster than sales-increasing actions, as argued by Hofer (1980).268 During the recovery phase, cost cutting should become less important com- pared to increasing sales and carrying out investments to achieve long-term profitabili- ty and growth.269 Nevertheless, continued cost controls are recommended by e.g. Rob- bins and Pearce (1992).270 Empirical results concerning cost-cutting actions are mixed. Buschmann (2006) finds that almost all distressed German firms in his sample undertake cost-cutting ac- tions. At the same time, he finds no significant difference between recovery and non- recovery firms.271 Sudarsanam and Lai (2001) find that non-recovery firms employ operational restructuring actions (including cost-cutting actions) more often which are associated with a lower chance of recovery.272 Nothardt (2001) finds only insignificant contributions of cost-cutting actions other than personnel-related costs to the turna- round probability.273 Slatter (1984) shows that both recovery and non-recovery firms engage in cost-cutting actions although the frequency is higher for non-recovery firms.274 In the bankruptcy context, Kalay, Singhal, and Tashjian (2007) and Datta and 265 See Robbins and Pearce (1992), p. 291. Other contributions to strategic management literature include Schendel, Patton, and Riggs (1976), Hofer (1980), Arogyaswamy, Barker, and Yasai-Ardekani (1995) or, more recently, Filatotchev and Toms (2006). 266 Robbins and Pearce (1992), p. 304. 267 See Arogyaswamy, Barker, and Yasai-Ardekani (1995), p. 495. Similarly, Robbins and Pearce (1992), p. 303. 268 See Hofer (1980), p. 26. 269 See Balgobin and Pandit (2001), p. 305. 270 See Robbins and Pearce (1992), p. 291, and Slatter (1984), p. 120. 271 See Buschmann (2006), pp. 186-189. 272 See Sudarsanam and Lai (2001), p. 197. 273 See Nothardt (2001), p. 259. 274 See Slatter (1984), p. 120. 4.2 Research Model 53 cline.286 John, Lang, and Netter (1992) find that firms in performance decline quickly reduce their personnel costs ratio as part of adjusting operations and the organiza- tion.287 The extent to which reducing the number of employees impacts performance, especially during Chapter 11, remains largely unexplored. Datta and Iskandar-Datta (1995) show that around 19% of their sample firms implement layoffs during bank- ruptcy.288 Kalay, Singhal, and Tashjian (2007) report that the number of employees declines on average by 23% during Chapter 11.289 Khanna and Poulsen (1995) docu- ment positive announcement effects of layoffs in the three years before filing for bank- ruptcy.290 The bankruptcy phase, on the one hand, should provide for a realignment of the number of employees that is needed to return the firm to adequate performance levels. On the other hand, reducing the number of employees in the post-bankruptcy phase could be negatively related to post-bankruptcy success when the firm should be on a growth or recovery path. Bringing the aforementioned together, I formulate the following hypothesis: H3: Reducing personnel in Chapter 11 (out of Chapter 11) is positively (negatively) related to the probability of post-bankruptcy success. Personnel reductions are operationalized by a reduction in the number of em- ployees by at least 10% compared to the reference period, as found in Worldscope. This follows the operationalization by Atanassov and Kim (2009).291 The number of employees thus serves as a proxy for personnel-related costs. 4.2.2.1.4 Changing Capital Expenditures Reducing capital expenditures in financial distress fits into the retrenchment stage of Robbins and Pearce (1992) as it aims to improve cash flow in the short term 286 See Ofek (1993), p. 27. 287 See John, Lang, and Netter (1992), p. 907. 288 See Datta and Iskandar-Datta (1995), p. 28, who do not report any performance effects from layoffs. 289 See Kalay, Singhal, and Tashjian (2007), p. 789. 290 See Khanna and Poulsen (1995), pp. 927-928. Note that these effects are measured before filing for bankruptcy. 291 See Atanassov and Kim (2009), p. 349, who use a threshold of 20% as opposed to my threshold of 10%. 4.2 Research Model 54 and, thereby, to contribute to firm survival.292 Conversely, during the recovery phase, capital expenditures should increase as a result of focusing on investment and growth to return the company to long-term profitability, as put forward by Robbins and Pearce (1992).293 Capital expenditures can be changed relatively easily at the discretion of management.294 Compared to layoffs, the cash flow effect can also be significant de- pending on the capital intensity of the industry. Implementing capital expenditure changes seems also fairly uncomplicated. With regard to capital expenditures during financial distress and bankruptcy, the problem of overinvestment induced by the risk- shifting incentives for the incumbent management and shareholders must be consi- dered, as described in chapter 2.1. However, it should be emphasized that overinvest- ment can only occur if firms have sufficient internal capital left to invest, or if they can raise external capital such as DIP financing, which does not impose restrictive cove- nants on the firm’s investment behavior. Accordingly, if overinvestment prevails, it is reasonable to expect capital expenditures to increase during Chapter 11. Empirically, Buschmann (2006) finds that distressed firms in his German sam- ple reduce their investments during the crisis. However, he cannot trace any significant difference between successful and unsuccessful firms.295 Sudarsanam and Lai (2001) analyze the contribution of increased capital expenditures on the likelihood of recov- ery.296 They find no significant impact of an increase in capital expenditures. Bergauer (2001) shows that distressed firms in her German sample reduce their investments dur- ing the crisis and moderately increase investments after turnaround has been achieved.297 For a sample of junk bond issuers, Asquith, Gertner, and Scharfstein (1994) show that capital expenditures plummet by 66% during financial distress.298 This corresponds to findings by Andrade and Kaplan (1998) for highly leveraged 292 See Robbins and Pearce (1992), p. 291. 293 See Robbins and Pearce (1992), p. 291. Similarly, Balgobin and Pandit (2001), p. 305. 294 See Eichner (2010), p. 85. 295 See Buschmann (2006), p. 193. 296 See Sudarsanam and Lai (2001), p. 189. 297 See Bergauer (2001), pp. 214-215. 298 See Asquith, Gertner, and Scharfstein (1994), p. 650. 4.2 Research Model 55 transactions that subsequently became distressed.299 Eichner (2010) analyzes both an increase and a reduction in capital expenditures for the retrenchment and the recovery phase for his distressed sample. However, he cannot substantiate any significant im- pact of changes in capital expenditures on the turnaround probability.300 For bankrupt firms empirical evidence is sparse. In her descriptive statistics, Hotchkiss (1995) documents that, in the years leading to bankruptcy, firms reduce their median capital expenditures from 6% to 4% of total assets. The median remains at 3% in the post-bankruptcy phase.301 Due to the automatic stay and the possibility of DIP financing it can be assumed that the pressure to reduce capital expenditures during Chapter 11 merely to generate cash is not as pronounced as for an out-of-court restruc- turing. To shed more light on the impact of changes in capital expenditures in the bankruptcy context, I formulate the following hypotheses: H4: Increasing capital expenditures in Chapter 11 (out of Chapter 11) is negatively (positively) related to the probability of post-bankruptcy success. H5: Reducing capital expenditures in Chapter 11 (out of Chapter 11) is positively (negatively) related to the probability of post-bankruptcy success. The change in capital expenditures is measured as a change of at least 10% with respect to the reference period. Capital expenditures are operationalized as capital ex- penditures scaled by total assets as defined in Worldscope. Expenditures associated with acquisitions are not included. Accordingly, only additions to property, plant and equipment and investments in machinery and equipment are included in capital ex- penditures. 4.2.2.2 Financial Restructuring Financial restructuring concerns itself with the right-hand side of the balance sheet, i.e. with changes to a firm’s liabilities and stockholders’ equity. 299 See Andrade and Kaplan (1998), p. 1464. 300 See Eichner (2010), p. 215. 301 See Hotchkiss (1995), p. 9. 4.2 Research Model 58 to the automatic stay during Chapter 11 that allows bankrupt firms to suspend most of their debt payments. This is obviously more beneficial to highly leveraged firms.317 Denis and Rodgers (2007) find that firms that reduce their leverage ratio during Chap- ter 11 exhibit a higher probability of emerging.318 Likewise, firms that reduce their leverage ratio during Chapter 11 are more likely to show positive post-bankruptcy per- formance.319 I therefore formulate the following hypothesis: H6: Reducing the leverage ratio in Chapter 11 (out of Chapter 11) is positively re- lated to the probability of post-bankruptcy success. The reduction in leverage is measured as the change in total liabilities divided by total assets of at least 10% compared to the reference period. Total liabilities are taken rather than total debt in accordance with Denis and Rodgers (2007) and Kalay, Singhal, and Tashjian (2007).320 My choice is also substantiated by Kalay, Singhal, and Tashjian (2007) who do not find any significant difference between employing a leve- rage ratio based on total debt or total liabilities.321 4.2.2.2.2 Issuing New Equity Issuing new equity for cash might seem as an easy way to resolve financing problems in financial distress and bankruptcy. A number of important issues must nev- ertheless be addressed in relation to this view.322 First, a firm in bankruptcy typically lacks a successful track record needed for a convincing equity story. Second, the debt overhang problem introduced by Myers (1977) and extended by Gertner and Scharfstein (1991) discourages new equity investors due to likely wealth transfers at the expense of the new shareholders. However, as Jostarndt (2007) argues, investors may provide new equity to the firm if the net present value of the going concern is 317 See Kalay, Singhal, and Tashjian (2007), p. 792. The typical exception for debt payments which are continued even under Chapter 11 relates to trade creditors as documented by Kalay, Singhal, and Tashjian (2007), p. 791. 318 See Denis and Rodgers (2007), p. 113. The leverage ratio is measured as liabilities scaled by total assets. 319 See Denis and Rodgers (2007), p. 117. 320 See Denis and Rodgers (2007), p. 104, and Kalay, Singhal, and Tashjian (2007), p. 790. 321 See Kalay, Singhal, and Tashjian (2007), p. 779. 322 For a general literature review regarding security offerings, see Eckbo, Masulis, and Norli (2007). For more on the buyers of already issued stock of bankrupt firms, refer to Coelho, Taffler, and John (2010). Clark and Weinstein (1983) analyze the behavior of the com- mon stock of bankrupt firms, albeit under the old U.S. bankruptcy regime before 1979 (the Chandler Act). 4.2 Research Model 59 greater than any wealth transfers to the creditors.323 Alternatively, new cash equity in- vestments during Chapter 11 are more likely if creditors make concessions to the new shareholders that restrict any wealth transfers at the expense of new shareholders.324 Third, from a transactional and organizational view, it seems less complicated to se- cure fresh money through a DIP financing arrangement. Nonetheless, new equity is- sues do occur during Chapter 11, but they more frequently take place in the post- bankruptcy phase to finance recovery or growth.325 During Chapter 11, new equity issues typically take place shortly before or at the emergence from bankruptcy.326 New shareholders (so-called vulture investors)327 will invest even during bankruptcy, pro- vided their investment will be reflected in the reorganization plan and their equity rights also apply to the reorganized entity.328 Hotchkiss and Mooradian (1997) show that the presence of a vulture investor has a positive impact on the post-restructuring performance of financially distressed firms.329 Consequently, one might argue that is- suing new equity during Chapter 11 may serve as a signal that the equity investors trust in the viability of the firm. Accordingly, one might conjecture that firms receiving new equity during Chapter 11 will likely show a better post-bankruptcy performance. Once the company emerges from bankruptcy, issuing new equity to finance future growth is likely less burdensome than in the bankruptcy phase, since the relative im- portance of the debt overhang and wealth transfer problems should now diminish.330 Buschmann (2006) finds a significant positive correlation between the issue of new equity and turnaround for his sample of distressed German firms, whereas Eichner 323 See Jostarndt (2007), p. 131. 324 See Jostarndt (2007), p. 131. 325 Refer to the descriptive statistics on the frequency of restructuring actions in Table 15 below. 326 Based on the findings in my sample. 327 See e.g. Hotchkiss and Mooradian (1997). 328 The investment motives of vulture investors are presented in Altman and Hotchkiss (2006), chapter 8. These include gaining active control of the target, as in the case of the investor W. L. Ross in the U.S. steel industry, who bought and controlled firms such as LTV, Acme Steel, Bethlehem Steel, Weirton and Georgetown Steel. In addition to buying old debt and already issued shares of the bankrupt firm, vulture investors also buy new shares in some cases. 329 See Hotchkiss and Mooradian (1997), p. 401. 330 See Eichner (2010), p. 110. 4.2 Research Model 60 (2010) and Sudarsanam and Lai (2001) find only insignificant relations.331 In sum, I formulate the following hypothesis: H7: Issuing new equity for cash in Chapter 11 (out of Chapter 11) is positively re- lated to the probability of post-bankruptcy success. The issue of equity is operationalized as a dummy variable. Information con- cerning equity issues is hand-collected from company filings, as described in more detail in chapter 5.2 below. It is crucial that the new equity issue is completed rather than only announced and it must be conducted in return for cash in line with Jostarndt (2007).332 Furthermore, equity issues can include private placements and public offer- ings as well as rights offerings for common or preferred stock.333 4.2.2.2.3 DIP Financing The drawbacks and benefits of DIP financing for bankrupt firms are discussed controversially in the literature. Some authors argue that DIP financing increases the problem of overinvestment as modeled by Gertner and Scharfstein (1991).334 Others, such as Gilson, John, and Lang (1990), regard DIP financing as part of the solution to potential underinvestment due to the debt overhang problem.335 Stulz and Johnson (1985) show in their model that secured debt (which DIP financing typically is) can help firms to invest in profitable projects in which they would not have been able to invest using equity or unsecured debt.336 However, providing security to new lenders may result in wealth transfers in favor of the new secured lenders and to the detriment of the existing unsecured lenders.337 Additionally, DIP financing may prevent bankrupt 331 See Buschmann (2006), p. 190, Eichner (2010), p. 218, Sudarsanam and Lai (2001), p. 196. 332 See Jostarndt (2007), p. 174. The firm only receives fresh money in the case of cash equity issues as opposed to debt-to-equity swaps which help to decrease leverage ratio but fail to provide the firm with fresh liquidity. 333 Not included are warrants, convertible bonds, debt-to-equity swaps and executive compensation. 334 See Gertner and Scharfstein (1991), p. 1191, and Dahiya et al. (2003), p. 260. 335 See Gilson, John, and Lang (1990), p. 320. 336 See Stulz and Johnson (1985), p. 501. 337 See Bebchuk and Fried (1996), p. 870. 4.2 Research Model 63 man) has a positive effect on post-restructuring performance.352 Eichner (2010) docu- ments a significant negative impact on the turnaround probability due to a disciplinary CEO change in the late phase of the restructuring process.353 Buschmann (2006) finds no significant correlation between changing the management and turnaround suc- cess.354 Prior contributions from bankruptcy literature, such as Baird (1986), argue that U.S. bankruptcy law favors reorganization over liquidation.355 Accordingly, the bank- ruptcy law might be biased toward the survival of inefficient firms in Chapter 11 that should have been liquidated. Hotchkiss (1995) suggests that the role of incumbent management could play a role in this bias.356 This would be in line with Bradley and Rosenzweig (1992), who argue that incumbent management benefits from provisions of the Bankruptcy Code at the expense of stockholders and bondholders.357 Datta and Iskandar-Datta (1995) support the hypothesis of Bradley and Rosenzweig (1992) that Chapter 11 indulges incumbent management.358 One could also argue that incumbent managers tend to overinvest in risky projects during bankruptcy, hoping to turn the firm around. This was done in the well-known case of Eastern Airlines, where credi- tors’ cash collateral was used to finance unprofitable business with court approval.359 Hotchkiss (1995) observes that retaining the pre-filing management is associated with poor post-bankruptcy performance among her sample firms, which filed for bankrupt- cy between 1979 and 1988.360 Assuming that managers prefer to control larger firms, incumbent managers are likely to refrain from selling too many assets or subsidiaries during Chapter 11 as this 352 See Hotchkiss and Mooradian (1997), p. 401. The involvement of a vulture investor is a reference to the vulture investor becoming either CEO or chairman or acquiring the majority of the voting stock. 353 See Eichner (2010), p. 229. In agreement with Hotchkiss (1995), I do not distinguish between disciplinary (or forced) CEO turnover events and those not related to a disciplinary event (unforced). 354 See Buschmann (2006), p. 195. 355 See Baird (1986), p. 134. 356 See Hotchkiss (1995), p. 4. 357 See Bradley and Rosenzweig (1992), pp. 1049-1050. These provisions include the exclusive right to propose the plan of reorganization, see Hotchkiss (1995), p. 3. 358 See Datta and Iskandar-Datta (1995), p. 15 and p. 27. 359 See Weiss and Wruck (1998). 360 See Hotchkiss (1995), p. 4. 4.2 Research Model 64 might hamper an effective restructuring of the firm.361 Somewhat surprisingly, Khanna and Poulsen (1995) do not find any significant differences between the management actions of firms filing for bankruptcy and those of firms in a control group that did not file for bankruptcy in the three years before filing for bankruptcy.362 They conclude that bankruptcy should not be attributed to incompetent or self-serving managers.363 Managers who have led firms into bankruptcy may have acted in a similar way to their counterparts whose firms did not go bankrupt. Accordingly, these managers did not necessarily do anything bad – but nor did they do anything good to prevent the firm from having to file for bankruptcy. More recently, Skeel (2003) has argued that Chapter 11 has changed from being pro-debtor to being pro-creditor.364 The reasons for this change in Chapter 11 practice, according to Skeel (2003), are (i) the changed terms of DIP financing arrangements that now allow creditors to effectively control the debtor in possession, and (ii) reten- tion bonuses and performance-based compensation schemes during Chapter 11 that tempt key managers to stay onboard and honor the rapid resolution of Chapter 11 in accordance with the interest of the creditors.365 To summarize the above, I propose the following hypothesis: H9: Replacing the top executive who was in office at the time of filing in Chapter 11 (out of Chapter 11) is positively related to the probability of post-bankruptcy success. The change in the top executive (CEO or president) is defined as the initial re- placement of the person who was in office at the time of filing and is coded as a dum- 361 See Hotchkiss et al. (2008), p. 33. 362 See Khanna and Poulsen (1995), pp. 920-921. 363 See e.g. Bolton and Scharfstein (1996), p. 5, on the ability of managers to divert cash flow from the firm. 364 See Skeel (2003), p. 919. Similar arguments are put forward by Bharath, Panchapagesan, and Werner (2010), Adler, Capkun, and Weiss (2006) and Baird and Rasmussen (2003). 365 While the terms of the performance-based compensation schemes vary, Skeel (2003) names the most common strategy, namely to honor rapid closing of the Chapter 11 process. Alternatively, if the business is to be sold, managers are paid in relation to the sale price. See Skeel (2003), p. 919 and pp. 926-928. 4.2 Research Model 65 my variable.366 Unlike Hotchkiss (1995), I have set the cutoff date for identifying the pre-filing top executive immediately before filing, instead of setting it at two years before filing.367 This choice was made to ensure that the management turnover variable captures the change of the top executive who actually filed for bankruptcy. Setting the cutoff date at two years before filing does not consistently grasp this effect since the top executive may have changed before the filing.368 4.2.2.4 Portfolio Restructuring Portfolio restructuring is concerned with any significant divestment or acquisi- tion that changes the asset portfolio of the firm as defined by Eichner (2010).369 4.2.2.4.1 Divestments Divestments in a distress situation are typically undertaken to generate the cash needed to repay debts or keep the business running, to focus on core business or to divest unprofitable business lines in accordance with the targets of the retrenchment stage described by Robbins and Pearce (1992).370 Divestments during financial distress may be a substitute for new debt or equity issues, both of which might be difficult to realize in times of distress.371 However, as Asquith, Gertner, and Scharfstein (1994) argue, certain problems can arise while a firm is in financial distress. First, conflicts between shareholders and creditors may hinder divestments. According to Jensen and Meckling (1976), equity in a distressed firm is an option on the firm’s assets that is out of the money. Shareholders thus benefit from riskier assets and are likely to lose some of the value of their option 366 Two special cases in the sample stand out and require some further explanation. Sterling Chemicals temporarily had two co-CEOs during bankruptcy. In this case, I coded the appointment of the co-CEO as a change in top management, as it can be assumed that the old CEO would relinquish at least some of his responsibilities. After successful emergence, Calton re-hired the CEO who had filed for bankruptcy and then been dismissed during bankruptcy. 367 See Hotchkiss (1995), p. 16, who states that her calculations are not dependent on the cutoff date, be it three years, two years or one year before filing. 368 This means that a change in the top executive might be flagged as having occurred during Chapter 11 while the change in fact took place e.g. one and a half years before filing. 369 See Eichner (2010), p. 53, who relies on Bowman and Singh (1993). 370 See Robbins and Pearce (1992), p. 291. John, Lang, and Netter (1992), p. 892, show that firms in distress respond by sharpening their business focus. For a detailed literature review with regard to divestitures, see Eckbo and Thorburn (2008). 371 See Hotchkiss (1993), p. 3 of the third essay. 4.2 Research Model 68 4.2.2.4.2 Acquisitions While divestments and asset sales in the bankruptcy context have been analyzed by prior empirical research, acquisitions both during Chapter 11 and after emergence and their potential impact on post-bankruptcy performance have been omitted. Yet in- vesting in profitable growth by acquiring other (parts of) companies, would appear to be a promising option for firms that emerge with an adjusted capital structure and po- tentially improved operations. Many contributions to restructuring literature, such as Schendel, Patton, and Riggs (1976), Slatter (1984), Robbins and Pearce (1992) and implicitly Arogyaswamy, Barker, and Yasai-Ardekani (1995), support the notion that acquisitions can add value, especially in the recovery phase of the turnaround process.387 Acquisitions during the retrenchment phase or Chapter 11 do not seem to fit into the two-stage turnaround model of Robbins and Pearce (1992) who instead advo- cate divestments in the retrenchment phase.388 Accordingly, I assume a negative rela- tion between acquisitions effected during Chapter 11 and post-bankruptcy perfor- mance.389 Empirical studies on the impact of acquisitions on turnaround probability report insignificant results only, such as Eichner (2010), Buschmann (2006) and Sudarsanam and Lai (2001).390 However, Sudarsanam and Lai (2001) note that recovery firms and non-recovery firms behave differently over time, with recovery firms focusing more on investments and acquisitions, while non-recovery firms are shown to be more preoccupied with operational and financial restructuring.391 To summarize, I formulate the following hypothesis: H11: Significant acquisitions in Chapter 11 (out of Chapter 11) are negatively (posi- tively) related to the probability of post-bankruptcy success. 387 See Schendel, Patton, and Riggs (1976), p. 8, on acquisitions in the upturn phase, Slatter (1984), pp. 120-121 on growth through acqui- sition, Robbins and Pearce (1992), p. 291, on acquisitions as strategies in the recovery stage, and Arogyaswamy, Barker, and Yasai- Ardekani (1995), p. 510, on strategic reorientation during the recovery phase. 388 See Robbins and Pearce (1992), p. 291. 389 One possibility in which an acquisition during Chapter 11 might be positively related to the likelihood of post-bankruptcy success occurs if the bankrupt firm is not restricted in its liquidity (i.e. it is not insolvent in the flow-based sense) and if the target improves the firm’s profitability. 390 See Eichner (2010), p. 229, Buschmann (2006), p. 181, and Sudarsanam and Lai (2001), p. 196. 391 See Sudarsanam and Lai (2001), p. 197. 4.2 Research Model 69 In line with Eichner (2010), an acquisition is operationalized as the mentioning of any completed majority-owned acquisition of another company or business unit through either an asset deal or a share deal.392 The closing of the deal (rather than the signing or the mere announcement) is crucial to assign the acquisition to a specific year.393 Only majority-owned transactions are considered, since these transactions can potentially alter the portfolio of the company. This is not typically the case with minor- ity-owned transactions, which are primarily considered to be financial investments. Significant transactions are defined following the same criteria as for divestments. 4.2.3 Context Factors as Control Variables This chapter briefly introduces the control variables used and explains their use by referring to prior research. Table 10 below summarizes the definitions and opera- tionalization of the control variables, which represent context factors in the sense of Pettigrew (1987b). Table 10: Definitions of Control Variables Category Variable Definition Bankruptcy Proceedings Prepack Mentioning that the bankruptcy filing was a prepack in the UCLA-LoPucki BRD Duration Time from filing to emergence from Chapter 11, in months. In the event of a § 363 sale, emergence is defined as the date of consummation of the sale (closing) Main Distress Source Economic Distress Industry-adjusted operating margin [Operating income/sales scaled by industry median] below zero in F-1 Financial Distress • Operating income less capital expenditures insufficient to cover interest ex- penses in F-1 • Leverage ratio [Total liabilities/total assets] above one in F-1 • Filing related to asbestos claims • Not economically distressed Firm and Industry Characteristics Firm Size Firm size in F-1 measured as ln(total assets) or ln(sales) Leverage Ratio Leverage ratio in F-1 measured as (total liabilities/total assets) Median Industry Performance Measured as the median return on assets or return on sales in F-1 for all firms within the same SIC group Source: Author’s own illustration. 392 See Eichner (2010), p. 138. 393 The same rationale applies as for the divestments shown above. 4.2 Research Model 70 Two control variables are used in relation to the bankruptcy proceedings: pre- packs and duration. Prepacks have been analyzed by Heron, Lie, and Rodgers (2009), Tashjian, Lease, and McConnell (1996), Chatterjee, Dhillon, and Ramirez (1996) and Betker (1995), for example. Not surprisingly, prepacks spend less time on average in Chapter 11.394 Chatterjee, Dhillon, and Ramirez (1996) show that economically viable firms with liquidity problems file prepackaged bankruptcies more often, while eco- nomically distressed firms file for traditional Chapter 11, and economically viable firms with no liquidity problems tend to employ an out-of-court restructuring.395 Nev- ertheless, empirical evidence concerning the post-bankruptcy performance of prepacks is relatively sparse. Lubben (2008) finds that prepacks exhibit a higher probability of refiling.396 Conversely, Alderson and Betker (1995b) argue that prepacks should exhi- bit better post-bankruptcy performance than usual Chapter 11 reorganizations, since they resemble out-of-court debt restructurings which tend to be more frequently cho- sen by efficient firms.397 Their empirical analysis supports this view, as the average excess return for prepacks is significantly higher.398 Despite the inconclusive evidence, I expect a positive relation between filing a prepackaged bankruptcy and post- bankruptcy performance because prepacks typically leave Chapter 11 much more quickly. This should reduce the costs associated with bankruptcy, such as direct costs incurred by employing lawyers, consultants and accountants and indirect costs as a result of any potential business disruption costs. To summarize the above, I formulate the following hypothesis: H12: Filing a prepackaged bankruptcy is positively related to the probability of post- bankruptcy success. 394 See Heron, Lie, and Rodgers (2009), p. 742, and Tashjian, Lease, and McConnell (1996), p. 142. 395 See Chatterjee, Dhillon, and Ramirez (1996), p. 5 396 See Lubben (2008), p. 268 and p. 281. 397 See Alderson and Betker (1995b), pp. 13-14. 398 See Alderson and Betker (1995b), p. 15. 4.2 Research Model 73 probability of post-bankruptcy success.413 Heron, Lie, and Rodgers (2009) conclude that pure financial distress can be resolved more quickly than economic distress.414 As Asquith, Gertner, and Scharfstein (1994) point out, firms may be both economically and financially distressed at the same time.415 To avoid any inconsistencies, firms that are simultaneously in economic and financial distress are categorized as economically distressed.416 Asquith, Gertner, and Scharfstein (1994) study different sources of finan- cial distress and define three of them: leverage, firm operating performance and indus- try operating performance.417 I have refrained from using their convincing definition of the sources of financial distress for the following reason: Asquith, Gertner, and Scharfstein (1994) base their definition of financial distress on an interest coverage ratio only. Accordingly, they focus solely on the flow-based definition of financial dis- tress.418 For the bankruptcy context, this definition alone is not sufficient, since some firms that enter bankruptcy are not insolvent in a flow-based sense.419 This definition would fail to include filings that relate to the stock-based insolvency definition or fil- ings related to asbestos claims. The latter firms filed for bankruptcy protection to re- solve their actual and potential future liabilities due to asbestos litigation claims.420 413 See Denis and Rodgers (2007), p. 116. 414 See Heron, Lie, and Rodgers (2009), p. 727 and p. 742, who regard firms with positive operating performance before filing for bank- ruptcy as financially distressed. 415 See Asquith, Gertner, and Scharfstein (1994), p. 628. 416 This follows Asquith, Gertner, and Scharfstein (1994), p. 632, who define economic distress (either poor firm or poor industry operating performance) as a source of financial distress. 417 See Asquith, Gertner, and Scharfstein (1994), p. 632. 418 Excessive leverage affects the interest coverage ratio only indirectly (through higher interest payments), but not directly. 419 Instead, these firms might exhibit excessive leverage or a negative net worth. See also the discussion in Wruck (1990), p. 422. 420 Six firms in the emerged public sample filed mainly to resolve litigation claims arising from asbestos. 4.2 Research Model 74 Figure 4: Economic vs. Financial Distress of Emerged Public Sample Firms in F-1 The figure shows the number of firms from the emerged public sample that are categorized as either economically or financially distressed or not distressed one year before filing (F-1). Source: Author’s own illustration. Finally, firm and industry characteristics prevailing one year before filing for bankruptcy are also controlled for. One context factor that is frequently controlled for is the leverage ratio one year before filing. The leverage ratio thus reflects the preva- lent capital structure of the firm. As Zingales (1998) shows, the level of the leverage ratio can have a negative impact on a firm’s chances of survival.421 This could be attri- buted to the debt overhang problem introduced by Myers (1977). Harris and Raviv (1990) show in their model that more levered firms exhibit a lower chance of reorgani- zation after default.422 Conversely, high leverage may also have benefits as argued by Jensen (1986) and Jensen (1989). High leverage may serve as an effective monitoring and disciplining tool that gives managers an incentive to run the firm efficiently and to restructure faster if the orderly payment of contractual debt obligations is at risk.423 Furthermore, high leverage may serve as a catalyst for a timely bankruptcy filing. This can preserve relatively more firm value compared to a less levered firm whose firm value will have fallen more dramatically when filing for bankruptcy.424 421 See Zingales (1998), p. 906. 422 See Harris and Raviv (1990), p. 323. 423 See Jensen (1986), p. 324. 424 See Jensen (1989), pp. 41-42. 143 41 (29%) Emerged Public No Distress Economic Distress Financial Distress 12 (8%) 90 (63%) 4.2 Research Model 75 In the context of bankruptcy, several empirical papers analyze the impact of le- verage. Kalay, Singhal, and Tashjian (2007) find that leverage has a significant posi- tive effect on improvements in operating income during Chapter 11, which they ex- plain with the automatic stay.425 Denis and Rodgers (2007) find only inconclusive re- sults concerning the influence of pre-filing leverage on post-bankruptcy perfor- mance.426 Accordingly, I expect that the pre-filing leverage ratio will not affect the probability of post-bankruptcy success: H14: The pre-filing leverage ratio is unrelated to the probability of post-bankruptcy success. In line with Denis and Rodgers (2007) and Hotchkiss (1995), I define the pre- filing leverage ratio as total liabilities over total assets prevailing one year before fil- ing.427 The pre-filing size of the company is also typically used as a control variable. The relation between size and post-bankruptcy performance has mostly been found to be positive as, for instance, in Denis and Rodgers (2007), Dawley, Hoffman, and Brockman (2003) and Hotchkiss (1995).428 Franks and Torous (1989) posit that larger companies are better able to cope with the complexities of the reorganization process.429 Moreover, the larger the firm, the more slack resources can generally be drawn on.430 This has been emphasized by Dawley, Hoffman, and Lamont (2002) and Hannan and Freeman (1984), for example.431 LoPucki and Whitford (1993b) document high confirmation rates among large public U.S. firms in bankruptcy and conclude that there seems to be a relation between size (i.e. large public firms) and confirmation of 425 See Kalay, Singhal, and Tashjian (2007), pp. 790-791. 426 See Denis and Rodgers (2007), p. 116. 427 See Denis and Rodgers (2007), p. 109, and Hotchkiss (1995), p. 7. 428 See Denis and Rodgers (2007), p. 115, Dawley, Hoffman, and Brockman (2003) p. 417, and Hotchkiss (1995), p. 17. 429 See Franks and Torous (1989), p. 749. 430 In the sense put forward by Arogyaswamy, Barker, and Yasai-Ardekani (1995), pp. 498-499. 431 See Dawley, Hoffman, and Lamont (2002), p. 700, and Hannan and Freeman (1984), p. 159. 5.1 Statistical Methodology 78 the problem of missing data.443 This problem can be mitigated by using performance groups modeled as a binary dependent variable.444 5.1.2 Test for Sample Selection Bias In the following, I test the analysis of post-bankruptcy performance for a poten- tial sample selection bias.445 This bias could arise if those firms that emerge from Chapter 11 are not randomly drawn from the underlying population of bankrupt firms.446 In this case, the firms forming the basis for the final sample are determined by a selection process. The potential bias follows from the research design, which focuses on the behavior of those firms that emerge as public firms from Chapter 11 and, accor- dingly, whose post-bankruptcy data is observed. Data for those firms that do not emerge from Chapter 11 is, by definition, not observed. This can also be understood as a kind of sample attrition.447 If a sample selection bias is not controlled for, this can lead to inconsistent estimates of the factors influencing post-bankruptcy perfor- mance.448 As a test for sample selection bias, I use an extension of the classic Heckman two-stage model called the probit model with sample selection.449 Unlike the classic two-stage Heckman model, which uses a probit model in the first stage and an ordi- nary least squares model in the second stage, the probit model with sample selection uses probit models in both stages.450 Since the logistic and probit models typically yield similar results and since this will be demonstrated for my final model below, ap- 443 This may be due to refiling, liquidating, going private or merging. 444 See Hotchkiss (1993), p. 21 of the first essay. Additionally, she states that differences in accounting practices are also mitigated. These differences frequently occur after emergence in connection with the write-down of assets to their fair market values when fresh-start re- porting is adopted in accordance with SOP 90-7. 445 The sample selection bias has also been called sample selectivity, see Davidson and MacKinnon (1993), p. 542. In this study, selectivity bias, sample selection bias or short-hand selection bias are used synonymously. Seminal contributions with regard to sample selection bias in economics are Heckman (1974), Heckman (1976), and Heckman (1979). 446 See Wooldridge (2010), p. 790. 447 See Wooldridge (2010), p. 813. 448 See Wooldridge (2010), p. 805. 449 See Wooldridge (2010), pp. 813-814. Among the first contributions with respect to probit models with sample selection are van de Ven and van Praag (1981) and Dubin and Rivers (1989). 450 See Wooldridge (2010), p. 814. 5.1 Statistical Methodology 79 plying the probit model with sample selection instead of a logistic model should not make much difference to the results.451 To make it easier to understand how the probit model with sample selection is applied, I will introduce it in the context of the analysis of post-bankruptcy perfor- mance.452 The first stage is a selection equation as in (1). The selection equation de- termines which firms survive the Chapter 11 process and emerge from bankruptcy. The second stage in (2) is the outcome equation that models post-bankruptcy performance.

and 
are vectors of observations of independent variables for firm .  and  are parameter vectors. Finally, 
and 
are error terms for firm , which are assumed to be independent of  and  and exhibit a bivariate normal distribution, as shown in (3). 
=

 + 
(1)
= 
 + 
(2) 

~ 0 0 , 1   1  (3) Post-bankruptcy performance for firm  is observable only if the firm emerges from Chapter 11. Formally,
=     
> 0  
≤ 0 I am interested in the expected value of
conditional on selection (
> 0), which is formally expressed as 
| 
> 0]. (4) Replacing 
with

 + 
from equation (1) and rearranging yields 
| 
> −

]. (5) Taking the expected value in (5) results in 
 + 
| 
> −

]. (6) 451 Using a probit model to test for sample selection bias instead of a logistic model can be considered standard, as can be inferred from Wooldridge (2010), for example. 452 This paragraph draws heavily on Greene (2012), pp. 912-916, and Wooldridge (2010), pp. 813-814. 5.1 Statistical Methodology 80 Equation (6) can be rewritten as 
 + 
(7) where
= !(

) Ф(−

) = !(

) 1 − Ф(

) is the inverse Mills ratio, with ! being the normal probability density function and Ф being the normal cumulative distribution function.453 It is generally recom- mended to have at least one variable driving selection in the first stage equation that does not appear in the outcome equation of the second stage.454 As a result, it is sug- gested to use an instrumental variable as an exclusion restriction.455 Prior post-bankruptcy performance literature mostly ignored a potential sample selection bias. Only recently, Kalay, Singhal, and Tashjian (2007) have discussed a potential sample selection bias which they eventually dismiss based on insignificant differences in risk-adjusted returns between reorganized firms, on the one hand, and liquidated and acquired firms, on the other hand.456 Alderson and Betker (1999) con- cern themselves with a potential selection bias in their study of post-bankruptcy per- formance, albeit without explicitly testing for it.457 In the context of bankruptcy costs that materialize in different Chapters of the Bankruptcy Code, Bris, Welch, and Zhu (2006) present a two-stage model accounting for a potential selection bias.458 453 See Davidson and MacKinnon (1993), p. 544, and Heckman (1979), p. 156. 454 See Wooldridge (2010), p. 814. 455 Refer generally to Li and Prabhala (2007) or Vella (1998) for a discussion of the use of exclusion restrictions in models testing for or correcting selection bias. 456 See Kalay, Singhal, and Tashjian (2007), p. 775 and p. 794. A similar procedure is applied by Lemmon, Ma, and Tashjian (2009), pp. 57-58. 457 See Alderson and Betker (1999), p. 77. 458 See Bris, Welch, and Zhu (2006), p. 1260. 5.3 Sample Selection 83 More qualitative information about new equity issues, DIP financing arrangements or changes in top management was gathered by hand from company filings from the same data sources as above. To make sure that all DIP financing arrangements were considered, I also used a systematic keyword search to check LexisNexis for any company information or news stories indicating DIP financing.469 Information about mergers, acquisitions and divestitures was gathered from the Mergerstat M&A database, accessed through LexisNexis to obtain consistent and reli- able information on these transactions.470 The Mergerstat M&A database was syste- matically searched for acquisitions and divestitures. This database covers all publicly announced transactions in which the equity value is greater than USD 1.0 million and the interest amounts to at least 10%.471 This ensures that only transactions of relevant size are considered. Data on individual SIC codes for the bankrupt firms was taken from Worldscope and amended by year-specific data from company filings or EDGAR. Chapter 11 outcomes were taken from the UCLA-LoPucki BRD and cross-checked with data from company filings, the Mergerstat M&A database and the Public Company Bankruptcy Filing Information database in LexisNexis. 5.3 Sample Selection 5.3.1 Sample Selection Process The sample was taken from the UCLA-LoPucki Bankruptcy Research Database, which covers all large bankruptcy filings of public U.S. firms since 1979.472 The se- lected period is 1993-2005, with the year of filing determining whether a firm is in- cluded in the sample or not. The initial year was chosen as 1993, as this is the first year 469 See chapter 4.2.2.2 for a detailed description. 470 Despite the fact that firms usually report material transactions in the notes to the 10-K and in 8-K filings, I rely on the consistent cover- age of firm transactions in the Mergerstat M&A database. 471 See source description for Mergerstat M&A database in LexisNexis. 472 Refer to chapter 5.2 for a detailed description of the requirements to be included in the UCLA-LoPucki BRD. 5.3 Sample Selection 84 for which company filings are available in EDGAR.473 Prior to fiscal 1993, data avail- ability is weaker, although some filings are available in EDGAR online and SEC on- line. In addition, the Mergerstat M&A database started full coverage of M&A deals in 1993. The last year was 2005 to allow for sufficient post-bankruptcy data in the three years following emergence.474 This left me with 529 firms. In line with previous re- search, 46 companies belonging to the financial service sector (SIC codes between 6000 and 6999) were excluded from the sample.475 The reason for exclusion is that financial service firms have their own particular bankruptcy regulations (such as the FDIC), as stated by Kalay, Singhal, and Tashjian (2007) and Dawley, Hoffman, and Brockman (2003), for example.476 Moreover, these firms’ balance sheets and state- ments of operations differ significantly from other companies, which makes compari- son difficult. In addition, I discarded 7 firms whose cases were dismissed. This re- duced the sample by 53 firms to 476. Next, the firms needed to emerge from Chap- ter 11 if their post-bankruptcy performance was to be analyzed. 172 firms did not emerge reducing the sample to 304 firms.477 Of those that did emerge, 161 had to be discarded due to limitations on data availability in either the year of emergence (E) or the first full post-bankruptcy year (E+1).478 This resulted in a sample of 143 firms that emerged as independent public firms and for which data was available for at least two post-bankruptcy years. A further 15 firms had to be excluded, yielding the final sample of 128 firms.479 The detailed sample selection process is shown in Figure 5. 473 Assuming that calendar year equals the fiscal year, filings for the fiscal year 1993 were the first filings to be included in EDGAR during 1994. 474 Adding on average two years in Chapter 11 to the year of filing plus the three post-bankruptcy years, results in 2005 being the last possible year for my sample. The average duration of two years builds on findings from Denis and Rodgers (2007) and Hotchkiss (1995) who find a median duration of 18 and 17 months in Chapter 11. 475 See Kalay, Singhal, and Tashjian (2007), p. 775, Dahiya et al. (2003), p. 264, and Dawley, Hoffman, and Brockman (2003), p. 418. 476 See Kalay, Singhal, and Tashjian (2007), p. 775, and Dawley, Hoffman, and Brockman (2003), p. 418. 477 Of these 172 firms, 111 were liquidated, 60 merged and 1 case is still pending (as of fiscal 2010 W.R. Grace was still in Chapter 11). 478 Data availability is deemed insufficient if data for less than two years is available as of the year of emergence, in line with Hotchkiss (1993), p. 14 of the first essay. Of these 161 firms, 105 firms went private, 20 were merged, 9 refiled, 6 liquidated and 21 lacked suffi- cient data for other reasons. Some of these firms did indeed emerge as public entities. However, they went private or merged before fil- ing the first 10-K after the year of emergence (see also Lemmon, Ma, and Tashjian (2009), p. 57). 479 Of these 15 firms, 13 were merged or went private in E+2 and 2 firms (Calpine and Solutia) had been out of Chapter 11 for less than three full fiscal years in fiscal 2010. This is in line with Hotchkiss (1995), p. 9. In principle, the 13 firms referred to above that merged or went private in E+2 have sufficient post-bankruptcy data (at least two years). However, since I prefer to judge the post-bankruptcy performance of firms that merged or went private based on their performance in the last available year before leaving the sample, this would only leave data for a single year. For consistency, these 13 firms were therefore excluded from further analysis. 5.3 Sample Selection 85 Figure 5: Sample Selection Process Source: Author’s own illustration. In line with Hotchkiss (1995), I also included cases that filed for bankruptcy more than once during the selected time period.480 This practice contrasts with Denis and Rodgers (2007), who do not include firms that filed twice during the same time period.481 To check the robustness of my findings, I excluded the five second filings (the repeated bankruptcy cases) in an additional analysis reported in chapter 6.2.1 be- low. 5.3.2 Sample Size Requirements and Treatment of Missing Data The final sample consists of 128 firms that emerged as reorganized public enti- ties from Chapter 11 and had sufficient post-bankruptcy data available. In terms of the minimum sample size for a logistic regression, Backhaus et al. (2006) state that at least 25 observations per category of the dependent variable are required.482 This criterion is met. Long (1997) concludes that a sample size below 100 observations could be prob- 480 See Hotchkiss (1995), p. 15. In total, five firms in my sample occur twice since they refiled for bankruptcy and again qualified for inclusion in the sample. 481 See Denis and Rodgers (2007), p. 104. 482 See Backhaus et al. (2006), p. 480. 128143 161 304 172 47653 529 Final Sample Insufficient Data Insufficient Data until E+1 15 Emerged Public Not emerged Initial Sample EmergedBRD Sample 1993-2005 Financial Services & Dismissed