Inside Debt and Bank Performance During the Financial

Inside Debt and Bank Performance During the Financial Crisis DNB 2012 Sjoerd van Bekkum

Initiatives to reduce excessive risk taking incentives after the crisis §Compensation Fairness Act 2009 § Shareholders should approve pay § Shareholders appoint compensation directors §Dodd-Frank Wall-Street Reform 2010 § Shareholders should vote on pay § Shareholders can design their own pay proposals § G 20’s FSB Sound Compensation Practices 2009 §“Engage shareholders with compensation” § 50% of variable pay should be awarded in shares or share-linked instruments §All shareholder-based! 2

Why this focus on aligning shareholder interests within in banks to limit risk? §Academic evidence on shareholder alignment and bank risk-taking is mixed at best (e. g. , Fahlenbrach & Stulz ‘ 11) § Equity-based pay is convex, and increases risk…. §Holds for options and restricted stock §Equity-based pay encourages risk-shifting in banks: 1. Banks are highly levered (E/A<0. 10) losses are easily shared with debtholders (Bebchuk & Spamann 2010) 2. Bank losses are also shared with taxpayers (LLOR, deposit insurance, implicit government guarantees) if executive’s actions are unobserved, she will shift risk to taxpayers and other debtholders (Bolton et al. 2006) 3

This paper focuses on aligning debtholder incentives using inside debt § § 4 Inside debt consists of debt from firm to manager: - pensions - retirement plans - Other deferred compensation So if bank goes broke, the managers waits in line with other creditors (Sundaram & Yermack, 2007) Hence, ID aligns interests between managers and debtholders (and un-aligns them with equityholders) Indeed, inside debt leads to more conservative decision making for industrials (Cassell et al. , 2011)

Inside debt encourages conservatism which pays off during bad times: 1. Shareholder implications: More inside debt better returns during crisis 2. Debtholder implications: More inside debt less downside risk during crisis 3. Investment policy: More inside debt better quality asset portfolio 4. Financing policy: More inside debt less “outside” debt 5

Main Result: Returns over the crisis 6

Data § Inside debt is disclosed mandatorily from 2006 compensation § Crisis runs from July 2007 to March 2009 July 2007 – March 2009 risk/return measures § All lending institutions from largest 3000 firms § Fiscal year ending in December 2006 § “substantial” part in lending business § executives with nonzero inside debt (is relaxed later) 319 banks 7

Measuring enterprise-wide bank risk § Enterprise-wide bank risk: shareholders + debtholders § No bonds/CDS data for most banks § Risk measures based on equity returns distribution Volatility (T/S/I) ES 8 Co. Va. R

Measuring realized enterprise-wide bank risk § Enterprise-wide bank risk: shareholders + debtholders § No bonds/CDS data for most banks § Risk measures based on equity returns distribution § Va. R § Expected shortfall § Co. Va. R § Total / Systematic / Idiosyncratic volatility § Prob(equity loss > 80%) 9 § Risk-taking: § Asset quality: % nonperforming assets ‘ 08 §Jan 07 -Jun 07 repo growth

Measuring inside debt § Main idea (Jensen & Mecklin 1976): Optimally, manager’s D/E = firm’s D/E. § Measured in levels (Edmans & Liu 2011) and 1 st differences (Wei & Yermack 2011) 10

Empirical model where bank risk R is captured by, Va. R, ES, Co. Va. R, and Prob(equity loss > 80%) 11

Results: BHRs 12

Results: Volatility 13

Results: Enterprise. Wide risk 14

Results: Risk-taking 15

Endogeneity concerns These findings are also consistent with: § Inside debt being only awarded when expected future risk is low! Risk is measured around unanticipated shock Impact of 2006 debt on 2007 -2008 realized risks Accumulated pensions are stock, not flow § Inside debt is only awarded by banks operating in a stable environment 2006 volatility does not explain realized risks inside debt discourages pre-crisis repo growth § Re-estimation using IV: 16

Endogeneity concerns Re-estimation using IV: § Executive age (Fahlenbrach & Stulz; Sundaram & Yermack 2007) § Size (exluding TBTF) §Tax status (Sundaram & Yermack 2007) § Maximum state tax rate (Anantharaman et al. 2010) § Newly hired executive (Sundaram & Yermack 2007) § Industry median inside debt (Murphy 1999) § previous controls 17

Results: IV estimates 18

More alternative explanations: I also do the following robustness checks: § 10% of all banks disappear from sample §Adjust BHR for return after merger/delisting §Include executives without any inside debt § Create a 0/1 variable for whether sb has inside debt § Re-estimate model on expanded sample § Presented estimates are at the executive level §Results also hold at the CEO/CFO/bank level § Too-big-to-fail institutions might pay more inside debt AND are less risky § Results also hold without TBTF banks 19

Results: Robustness 20

Conclusion § Inside debt limits bank risk: • less negative returns • lower volatility and tail risk • discouragement of risk-taking § Power shift in governance from shareholders to debtholders limits risk Inside debt less risk better performance during crises § Trade-off between upside potential and downside risk is beyond the scope of this paper § But given that risk should be limited, inside debt might be a way to do it. 21

22