Ownership links, leverage and credit risk

1. Ownership links, leverage and credit risk Elisa LUCIANO Giovanna NICODANO University of Torino and Collegio Carlo Alberto International Financial Research Forum, Paris, March 27, 2008

2. Purpose Study the relationship between ownership and credit risk. Examine optimal capital structure and credit worthiness of parent-subsidiary links, such as business groups & multinationals, private equity funds, LBOs, MBOs, joint ventures & project financing, and compare with the ones of separately incorporated (independently managed) and merged activities. Key role of internal capital markets in exploiting tax savings-default costs trade off

3. Why credit and ownership? • Structural approaches to credit risk do not model ownership links: in doing so, • they do not model asset transfers and • They do not measure appropriately default dependence, both ex ante (pricing) and ex post (system stability). Usually default correlation is due only to asset correlation.

4. The holding may transfer funds to its subsidiary Khanna and Palepu (2000) document transfers in Indian groups Bertrand et al. (2002) document cash transfers in several forms - from asset sales to internal loans at subsidized rates Holding enjoys limited liability vis-à-vis the subsidiary's debt Hadden (1986): common characteristic across major jurisdictions Boot et al. (1993): holding writes comfort letters assuring subsidiaries' lenders - legally unenforceable - not honored when holding would be unable to survive What is the main feature of holding-subsidiary?

5. Among parent-subsidiary links, what is the main feature of private equity (private versus public groups)? Private equity represents 25% of global M&A activity & 33% of high yield market, with the following features (Jensen, 2007, Kaplan, 1996): • Leverage higher than in public firms, with debt (and equity) at the divisional level • Very few bankruptcies, even when leverage is at 95% or higher A theoretical model is still missing.

6. Outline • The model: stand alone versus parent-subsidiary • Optimality problem: endogenous versus exogenous quantities • Solution for symmetric, dependent firms • Asymmetric firms and constrained groups (limited leverage capacity) • Default correlation

7. The (double) stand alone model joint survival default firm 1 default firm 2 joint default DEFAULT, NO TAXES DEFAULT WITH TAXES NO DEFAULT

9. Main mechamism in groups: guarantee or contingent rescue • The holding can help its subsidiary out of default, if this does not endanger her survival = contingent rescue; • This distinguishes groups from M&A, in which rescue is uncontingent • The holding receives dividends from its subsidiary, when the latter is solvent.

11. The group model without dividends rescue joint survival without rescue selective default subsidiary selective joint default default holding DEFAULT, NO TAXES DEFAULT WITH TAXES NO DEFAULT

14. The group model with dividends joint survival via rescue joint survival without rescue and dividends selective default subsidiary joint survival via dividends joint default selective default holding

15. Optimality problem max value = debt+ equity Or, equivalently: min (default costs – tax savings) using debt policy (Ph,Ps). EXOGENOUS cash flow distributions (Xh, Xs), tax rates & default costs ENDOGENOUS • Optimal tax shield and default threshold (Xz,Xd), • Current values of optimal debt (Dh, Ds) and equity (Eh, Es) • Default and rescue probabilities, • Recovery rates, • Spreads • Value maximizing ownership structure

16. Main results for symmetric BBB firms (stand alone & merger of Leland, 2007) From two stand alone firms to group, as well as from merger to group Leverage (and value) up Joint default probability down How? 100% of debt is shifted onto the subsidiary The rescue feature dominates over dividends: non strategic rescue (provided by dividends) and ownership share do not affect results Figures resemble the private equity ones

17. Base case: the role of correlation

18. Base case: the role of correlation

19. Base case: the role of correlation

20. Proposition 1 • Assume no agency costs, positive bankruptcy costs, fiscal deductibility of interest and the ability of the parent company to commit to state contingent transfers to its subsidiary. Consider BBB stand alone firms. Then the leverage and the value of a group exceed the one of conglomerates and of the corresponding stand alone companies. The holding is unlevered. The recovery rate of the group, which coincides with the subsidiary one, is lower, while the spread is larger than in other parent-subsidiary links.

21. BBB case: group versus stand alone

22. BBB case: group versus conglomerate

23. Asymmetric BBB firms (costs, volatility & size) • Groups are value enhancing when the role of each firm is endogenously characterized • Selective default increases and credit worthiness of the subsidiary deteriorates with its size and risk • With asymmetric size the holding becomes optimally levered

24. Constrained leverage • Suppose that the subsidiary cannot raise more debt than a stand alone, since regulatory costraints – which protect the subsidiary shareholders or prevent thin capitalization – exist. • The group arrangement creates value, independently of correlation and ownership share. The optimal holding leverage is close to the subsidiary one • The holding leverage is increasing ownership share. • Our figures are close to the actual business group ones

25. Proposition 2 • Consider a subsidiary with face value of debt equal to that of a BBB stand alone company. Then the parent optimal leverage is positive and close to the subsidiary and stand alone one; the default probability of both the subsidiary and the parent, their recovery rate and spreads are lower than the stand alone ones.

26. Constrained leverage, infinitesimal ownership: optimal debt values and leverage as a function of correlation

27. Constrained leverage, the role of ownership

28. Impact on portfolio default correlation • Usually structural models permit to compute default correlation taking into account cash flow (asset) correlation, but without legal and financial ties • Our model includes both asset correlation and ownership links

29. Default correlation, unconstrained case

30. Default correlation, constrained case

31. Contributions • The model helps explaining why observed leverage and credit risk features, including joint default occurrence, differ according to ownership; • It explains why prediction of default frequencies improves when ownership links are included (van Hulle et al. (2006)); • It separates cases in which capital structure entails high leverage in a subsidiary and zero in the holding (private equity) from the ones in which the holding is levered too (business groups); • It rationalizes very high leverage, before agency costs or when no agency costs exist (private equity); • It rationalizes zero-leverage firms: does it solve the puzzle?

32. Constrained leverage, infinitesimal ownership

33. Constrained leverage, the role of ownership: