1 contingent capital: the case of coercs george pennacchi (university of illinois) theo vermaelen...

28
1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October 2011

Post on 19-Dec-2015

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

1

Contingent Capital:The Case of COERCSGeorge Pennacchi (University of Illinois)

Theo Vermaelen (INSEAD)

Christian Wolff (University of Luxembourg)

October 2011

Page 2: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

2

Contingent capital• Contingent convertibles (CoCos) are bonds that

mandatorily convert to equity after a triggering event.

• Motivation: providing discipline of debt (tax deductions?) in good times, avoiding COFD (bailouts!) in bad times.

• The Basel Committee continues to review CoCos and supports their use by national regulators.

• Swiss National Bank wants major banks to have capital/RWA ratios of 19 % with up to 9 % in CoCos.

Page 3: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

In the academic spotlight• Flannery (2005, 2009, 2010)• Hart and Zingales (2009)• Kashyap, Rajan, and Stein (2009)• Duffie (2010)• McDonald (2010)• Coffee (2010)• Albul, Jaffee, and Tchistyi (2010)• Sundaresan and Wang (2010)• Glasserman and Nouri (2010) • Bolton and Samama (2010)• Calomiris and Herring (2011)• Barucci and Del Viva (2011)• Berg and Kaserer (2011)• Hilscher and Raviv (2011)

Page 4: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

4

The ideal CoCo bond?• Avoids bankruptcy: going concern rather than gone concern.

• Based on market-based triggers as opposed to regulatory capital ratios

or regulatory discretion.

• Avoids economically unjustified conversions.

• Preserves pre-emptive rights of shareholders.

• Low risk-easy to value (facilitates ratings, large demand).

• Avoid risk-shifting and debt overhang problems

• No multiple equilibria.

• No negative signal around issue or conversion

Page 5: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

5

Why regulatory capital ratio triggers are ineffective

• Regulatory capital ratios could be “old” as they are calculated only once every quarter.

• Poorly functioning triggers invite regulatory intervention.

• This makes CoCos very risky and/or difficult to value

> low liquidity

> no credit rating

> bondholders will short the bank’s stock

Page 6: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

6

What financial crisis ?

0

2

4

6

8

10

March 31 June 30 Sept. 30 Dec. 31

Mean

Source : Y-9C Bank Holding Company Reports obtained from the Federal Reserve Bank of Chicago

Mean 2008 Tier 1 Common Ratios of 50 Major U.S. Banks in percent

Page 7: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

0

100

200

300

400

0

5

10

15

20

25

1Q04 1Q05 1Q06 1Q07 1Q08 1Q09

bps%

Figure 1: Market and Accounting Metrics for SCAP Firms

Market Value of Equity / Assets (left) Book Value of Equity / Assets (left) CDS Spread (right)

Notes: Market value and book value ratios are simple means for 18 FIs that participated in the SCAP, excluding GMAC. CDS spreads are simple means of available data.

Source: Kevin Stiroh, FRB-NY

7

Page 8: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

8

Market value based triggers

• Contingent Capital Certificates (Flannery (2005, 2009, 2010)).

• Now conversion takes place when equity market value hits a specific level.

• The corresponding stock price is also the conversion price.

Page 9: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

9

Numerical Example

Assets (A) Liabilities

1100 Senior Debt (D) 1000 CoCo (B) 30 Equity (E) 70

1100 1100

B/D = 3 % E/D = 7 % (E+B)/D = 10%

Page 10: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

10

Numerical Example (2)• Assume that there are 7 shares outstanding, i.e. stock

price of $10.

• Conversion takes place when equity falls from 70 to 35 (E/D = 3.5 %) and stock hits $ 5.

• This trigger price is also the conversion price so bondholders can convert $ 30 into 6 shares.

• Fully diluted stock price is now 65/13 = $ 5

• Bond holders get 6 shares worth $ 30

• Because of this, CoCos are risk-free.

Page 11: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

11

Problems

1. Unjustified dilution because of manipulation or panic.

2. Risk cannot be eliminated because of sudden jumps.

3. Multiple equilibria.

Page 12: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

12

Percentage of 100 Largest U.S. Banks witha Daily Stock Return less than -10%

0%

10%

20%

30%

40%

50%

60%

70%

1/3

/20

07

2/3

/20

07

3/3

/20

07

4/3

/20

07

5/3

/20

07

6/3

/20

07

7/3

/20

07

8/3

/20

07

9/3

/20

07

10

/3/2

00

7

11/3

/20

07

12

/3/2

00

7

1/3

/20

08

2/3

/20

08

3/3

/20

08

4/3

/20

08

5/3

/20

08

6/3

/20

08

7/3

/20

08

8/3

/20

08

9/3

/20

08

10

/3/2

00

8

11/3

/20

08

12

/3/2

00

82007 20081 12 3 4 5 6 87 9 10 11 12 2 3 4 5 6 87 9 10 11 12

Figure 1

Page 13: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

13

Unjustified dilution• Assume stock price falls to $ 5 because of manipulation

or panic

• CoCos convert into 6 shares • Because the true value of the capital is still 100, the fully

diluted stock price is now 100/13 = $ 7.69

• Bondholder gain: 6 x 7.69 – 30 = 16.1 (54 %)

• Bondholders have large incentives to manipulate stock downwards through false rumours or shorting (no short squeeze as bank delivers shares for covering!)

Page 14: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

14

Solution: COERC

• Whenever stock hits $ 5, bondholders have to convert at $ 1.

• But shareholders have the right to buy back shares from the bondholders at $ 1.

• We baptize this as a COERC: call option enhanced reverse convertible.

Page 15: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

15

COERC and manipulation • Suppose stock falls to $5 through manipulation.

• Bonds with par of $ 30 convert at $1, receiving 30 new shares.

• Shares outstanding: 37 • Bondholders’ proportion of ownership: α = 30/37

• New fair value per share: $ 100/37 = $ 2.70

• However, when trigger is hit, a rights issue is announced for 30 shares at $ 1 per share.

• As $ 2.70 > $ 1, the rights are exercised and bonds are repaid.

Page 16: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

• Avoids death spiral or other incentives to short the stock• Preserves pre-emptive rights of stockholders• Setting trigger price >>> conversion price keeps bond

risk low• Low risk means no incentive to engage in risk-shifting• Under no condition tax payers have to bail out

bondholders• Multiple equilibria avoided• No negative signal as trigger is based on observed

variables• The COERC is a credible commitment to refinance

because it is a COERCive tool

Benefits of COERC

Page 17: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Multiple equilibria

• Sundaresan and Wang (2010) show that if the conversion event transfers value between CoCo investors and shareholders, then the individual values of the bank’s shares and CoCos are not unique.

• Thus, a trigger based solely on the market value of the bank’s shares can lead to multiple equilibria.

• However, the sum of the values of CoCos and bank shareholders’ equity is unique.

• Thus, we avoid multiple equilibria by basing the COERC (CoCo) trigger on the market value of total capital ratio, defined as

• If senior debt CDS spreads depend on this market capital ratio (as it does in our valuation model), then a senior debt CDS spread trigger also avoids multiple equilibria (c.f., Hart and Zingales, 2009).

Market value of equity + Market value of COERCValue of senior debt (deposits)

Page 18: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Valuation and risk analysis: a formal model

• We use a structural credit risk model of a bank1 to compare COERCs, standard CoCos, and non-convertible (subordinated) debt in terms of:

1. New issue yields (credit spreads)

2. The issuing bank’s risk-taking incentives

1 Pennacchi, G. (2010) “A Structural Model of Contingent Bank Capital,” Federal Reserve Bank of Cleveland Working Paper 10-04.

Page 19: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Assumptions: assets and deposits1) To model potential sudden, extreme losses as might

occur during a financial crisis, the value of a bank’s assets follows a jump – diffusion process.

2) The term structure of interest rates follows the model of Cox, Ingersoll, and Ross (1985).

3) Bank deposits have short (instantaneous) maturities, are default-risky, and are paid a fair credit spread.

4) The bank targets a 10% total capital-to-deposits ratio by adjusting deposit growth (mean-reverting leverage).

Page 20: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Assumptions: bonds and equity5) Fixed- or floating-coupon bonds are issued at their par

value of 3% of deposits and have a five-year maturity.

6) The types of bonds considered are:a) COERC with conversion triggered when total capital = 6.5% of

deposits (3.5% equity value). b) Standard CoCos with conversion triggered when total capital =

6.5% of deposits. Conversion price = trigger price.c) Non-convertible subordinated debt.

7) A bank is closed by regulators (and equity holders are wiped out) if the value of bank asset falls below the par value of deposits plus any non-convertible bonds.

Page 21: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: Fair new issue yields on COERCs decline as the proportion of shares they receive, , increases.

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5

5.1

5.2

7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12 12.5 13 13.5 14 14.5 15

Percent Capital to Deposits

Co

up

on

Rate

New Issue Yields on Fixed-Coupon COERCs For Different Numbers of Shares Issued

Five-Year Maturity, Initial COERC Value = 3% of Deposits,Conversion Triggered when Total Capital = 6.5% of Deposits

Dashed Line is Five-Year Default Free Treasury Yield

COERC shares to total shares ratio = 20/27

COERC shares to total shares ratio = 30/37

Figure 3

Co

up

on

Ra

te (

%)

Capital to Deposits (%)

Page 22: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: Delaying conversion of COERCs to when capital is lower increases their new issue yields.

4.2

4.4

4.6

4.8

5

5.2

5.4

5.6

5.8

6

6.2

6.4

6.6

Percent Capital to Deposits

Co

up

on

Rate

New Issue Yields on Fixed-Coupon COERCs For Different Equity Trigger Thresholds

Five-Year Maturity, Initial COERC Value = 3% of Deposits,COERC Shares to Total Shares Ratio = 30/37

Dashed Line is Five-Year Default Free Treasury Yield

Conversion Triggered when Total Capital = 5.0 % of Deposits

Conversion Triggered when Total Capital = 6.5 % of Deposits

Figure 4

Co

up

on

Ra

te (

%)

Capital to Deposits (%)

Page 23: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: New issue yields are higher for CoCos without a call option (minimum = 6/13).

New Issue Yields on Fixed-Coupon COERCs versus Contingent Capital without Call Option

Figure 5

4

4.5

5

5.5

6

6.5

7

7.5

8

8.5

9

9.5

7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12 12.5 13 13.5 14 14.5 15

Percent Capital to Deposits

Co

up

on

Rate

Five-Year Maturity, Initial Bond Value = 3% of Deposits,Conversion Triggered when Total Capital = 6.5% of Deposits

= COERC Shares to Total Shares Ratio Dashed Line is Five-Year Default Free Treasury Yield

COERC = 30/37

COERC = 20/27

COERC = 10/17

Contingent Capital without Call Option

Capital to Deposits (%)

Co

up

on

Ra

te (

%)

Page 24: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: Credit spreads on COERCs are lower than those of both standard CC and non-convertible debt.

0

50

100

150

200

250

300

350

400

450

500

550

600

3.5 4

4.5 5

5.5 6

6.5 7

7.5 8

8.5 9

9.5 10

10.5

11

11.5

12

12.5

13

13.5

14

14.5

15

Percent Capital to Deposits

New Issue Credit Spreads on Floating-Coupon COERCs, Contingent Capital, and Subordinated Debt

Cre

dit

Sp

read

(b

p)

COERC = 30/37

Contingent Capital without Call Option

Non-convertible Subordinated Debt

Five-Year Maturity, Initial Bond Value = 3% of Deposits,Conversion Triggered when Total Capital = 6.5% of Deposits

= COERC Shares to Total Shares Ratio

Figure 6

Capital to Deposits (%)

Page 25: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: A bank that issues COERCs has less incentive to invest in assets having a high probability of jumps.

Figure 7

(E

/D)/

Change in the Value of Shareholders’ Equity per DepositFor a 25% Increase in Frequency of Jumps ()

Five-Year Maturity, Initial Bond Value = 3% of Deposits,Conversion Triggered when Total Capital = 6.5% of Deposits

= COERC Shares to Total Shares Ratio

Capital to Deposits (%)

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

0.002

7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12 12.5 13 13.5 14 14.5 15

Contingent Capital without Call Option

Non-convertible Subordinated Debt

COERC = 30/37

Page 26: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Result: A bank that issues COERCs has less dis-incentive to issue new equity (debt overhang).

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

6.75 7 7.25 7.5 7.75 8 8.25 8.5 8.75 9 9.25 9.5 9.75 10

Capital to Deposits (%)

Figure 11 Change in the Value of Existing Shareholders’ EquityFollowing an Increase in New Equity of 0.125% of Deposits

Five-Year Maturity, Initial Bond Value = 3% of Deposits,Conversion Triggered when Total Capital = 6.5% of Deposits

= COERC Shares to Total Shares Ratio

Capital to Deposits (%)

COERC = 30/37

Contingent Capital without Call OptionNon-convertible Subordinated Debt

E/

A-1

Page 27: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

Intuition for results• Converting COERCs to a large number of shares

coerces shareholders to repay them at par, making COERCs less default-risky.

• Shareholders find it difficult to transfer value from COERCs to themselves by increasing risk.

• Shareholders’ risk-taking incentives become closer to those under unlimited liability, reducing moral hazard and debt overhang, thereby enhancing bank stability.

Page 28: 1 Contingent Capital: The Case of COERCS George Pennacchi (University of Illinois) Theo Vermaelen (INSEAD) Christian Wolff (University of Luxembourg) October

28

Summary and conclusions• A COERC mitigates debt overhang and should be attractive to

all firms wishing to reduce costs of financial distress.

• It is an early stage CoCo designed to prevent bankruptcy.

• It has a forward-looking, market-based trigger but is free from manipulation or death spirals.

• With a trigger value above its conversion value, its has low credit risk (high credit rating, easy to value, marketable).

• It preserves pre-emptive rights of shareholders by protecting them against dilution of control by bondholders.