FDIC/JFSR Conference FDIC/JFSR Conference Discussion of: Discussion - - PowerPoint PPT Presentation

FDIC/JFSR Conference FDIC/JFSR Conference

Discussion of: Discussion of:

“ “Bailouts, contagion, and Bank risk Bailouts, contagion, and Bank risk‐ ‐taking taking” ” by Dell by Dell ‘ ‘Ariccia and Ratnovski Ariccia and Ratnovski “ “Regulatory capture and banking supervision reform Regulatory capture and banking supervision reform” ” by Boyer & Ponce by Boyer & Ponce “ “Capital Regulation and tail risk Capital Regulation and tail risk” ” by Perotti, Ratnovski and Vlahu by Perotti, Ratnovski and Vlahu

By: Anja n V. T ha ko r Jo hn E . Simo n Pro fe sso r o f F ina nc e a nd Dire c to r o f the PhD Pro g ra m

1

September, 2011

Overview Overview

• T

he se a re thre e pa pe rs tha t ha ve the fo llo wing c o mmo n q ue stio n uniting the m:

• Wha t fo rms o f mic ro prude ntia l a nd ma c ro prude ntia l re g ula tio n

(inc luding re g ula to ry struc ture ) a re mo st e ffe c tive in c o ntro lling b a nk risk ta king ?

Perotti‐ Ratnovski‐ Vlahu (PRV)

Boyer‐Ponce (BP) Dell ‘Ariccia‐ Ratnovski (DR)

Microprudential regulation in the form of capital requirements. Macroprudential regulation in the form of bailouts. Organization of regulatory agencies.

2

• Wha t is a lso c o mmo n to a ll thre e pa pe rs is tha t

e a c h se e ks to o ve rturn c o nve ntio na l wisdo m:

• PRV: hig he r c a pita l re q uire me nts ma y induc e b a nks to ta ke mo re

risk.

• DR: b a ilo uts ma y induc e b a nks to ta ke le ss risk.
• BP: ha ving multiple re g ula to rs ma y le a d to b e tte r re g ula tio n a nd

le ss risk.

3

Capital Regulation and Tail Risk Capital Regulation and Tail Risk

by: Perotti, Ratnovski, and Vlahu by: Perotti, Ratnovski, and Vlahu

Q ue stio n: Do e s hig he r c apital always le ad to lo we r risk

taking by the bank?

• Co nve ntio na l wisdo m sa ys “ye s”. T

he insig ht tha t hig he r c a pita l induc e s ba nks to ta ke lo we r risk g o e s ba c k a t le a st to Me rto n’ s 1977 JBFpa pe r: iso mo rphic c o rre spo nde nc e be twe e n c o mmo n sto c k put o ptio ns a nd de po sit insura nc e .

• Ba nks c a n inc re a se the va lue o f the de po sit insura nc e put b y

inc re a sing risk a nd le ve ra g e .

• T
• c o unte r the se inc e ntive s a nd re in in the risk-ta king

pro pe nsity o f b a nks, we ne e d re g ula to ry c a pita l re q uire me nts.

4

• T

his pa pe r a rg ue s, ho we ve r, tha t whe n b a nks c a n inve st in “ta il risk” pro je c ts (whe re a suffic ie ntly a dve rse re a liza tio n c a n wipe o ut all o f the ba nk’ s e q uity c a pita l, re g a rdle ss o f the le ve l o f c a pita l), the n a ba nk with hig he r c a pita l ma y ta ke mo re risk.

• Why?

Higher Capital

a. Lower probability of breaching minimum capital requirement in low‐return state and having to make costly capital adjustment b. Payoff in tail‐risk state is 0 regardless of level of capital.

5

• A ba nk – whic h is a ve rse to sta te a – c a n “a ffo rd” to

ta ke g re a te r risk with mo re c a pita l sinc e hig he r c a pita l me a ns lo we r pro ba bility o f re a c hing sta te a , whe re a s c a pita l do e s no t a ffe c t sta te b.

• Pa pe r a rg ue s tha t this is wha t ha ppe ne d in the re c e nt

c risis – ba nks with mo re c a pita l to o k mo re ta il risks.

• T

wo ke y assumptio ns:

i. Pa yo ff to b a nk sha re ho lde rs whe n ta il risk e ve nt is re a lize d is inde pe nde nt o f c a pita l le ve l. ii. Ra ising c a pita l in lo w-re turn sta te in whic h minimum c apital re q uire me nt is b re a c he d is c o stly fo r the b a nk b e c a use “…e q uity issue s a re vie we d b y ne w inve sto rs a s ne g a tive sig na ls”.

• Ba se d o n this, the pa pe r pre sc ribe s tha t re g ula tio n

sho uld fo c us o n mo nito ring ta il risk in a dditio n to c a pita l re g ula tio n.

6

Comme nts:

• I

nte re sting pa pe r: Ana lysis is pla usib le . Pa pe r is c a re ful no t to o ve r-re a c h in re g ula to ry pre sc riptio ns (e .g . do e s no t pre sc rib e uppe r b o und fo r c a pita l re q uire me nts).

• But, nume ro us issue s to de a l with:
• 1. Pa pe r sho uld e xpla in diffe re nc e s b e twe e n its a na lysis a nd Cale m

and R

• bb’s JFI pape r

(“Impac t of Capital-Base d Re gulation on Bank Risk T aking: A Dynamic Mode l”)

Ma in ide a he re se e ms q uite simila r!

Capital Risk

Standard effect Capital is so high that probability of bankruptcy becomes low enough to make risk taking attractive.

7

• 2. T

he c la im tha t the re is a ne c do ta l e vide nc e tha t b a nks with hig he r c a pita l to o k mo re risks prio r to the c risis se e ms to be que stio na ble . I n fa c t, e ve n tho ug h ba nks did ha ve hig h c a pita l in the e a rly 2000s, ma ny spe nt hundre ds o f millio ns o f do lla rs in re purc ha sing sto c k to re duc e c a pita l…Prio r to the c risis, c a pita l ra tio s in inve stme nt ba nks we re de c lining . (Re me mbe r Be a r-Ste a rns ha d the lo we st c a pita l ra tio a mo ng the ma jo r inve stme nt ba nks.)

• T

he re is a lso e vide nc e tha t ba nks with hig he r c a pita l a c tua lly did be tte r in te rms o f surviving the c risis a nd a lso in g a ining ma rke t sha re a nd fina nc ia l pe rfo rma nc e (se e

Be r ge r and Bouwman, “How Doe s Capital Affe c t Bank Capital Dur ing Cr ise s?”, Whar ton F

• in. Inst. Ce nte r

WP, 2011).

8

• 3. Simila rly, the a utho rs ne e d to re c o nc ile the ir a rg ume nt

with the e vide nc e in Me hr

an and T hakor , “Bank Capital and Value in the Cr

• ss-Se c tion”, RFS, Apr

il 2011-- bank

c a pita l a nd va lue a re po sitive ly re la te d in the c ro ss-se c tio n!

• 4. T

he ke y a ssumptio n tha t raising a dditio na l c a pita l wo uld be c o stly fo r the ba nk due to a dve rse -se le c tio n-induc e d dilutio n is inc o nsiste nt with the e mpiric a l e vide nc e . Cor

ne tt and T e hr anian, “An E xamination of Voluntar y ve r sus Involuntar y Se c ur ity Issuanc e s by Comme r c ial Banks: T he Impac t of Capital Re gulations on Common Stoc k Re tur ns”,

JFE 1994, do c ume nts tha t whe n ba nks e ng a g e in

“invo lunta ry e quity issue s” (to me e t c a pita l re quire me nts), the pric e re a c tio n (a nno unc e me nt e ffe c t with a 2-da y windo w) is statistic ally insig nific ant.

9

• 5. Pa pe r ma y a lso b e ne fit fro m a dire c t c o mpa risio n

to Ac har

ya, Me hr an and T hakor , “Caught be twe e n Sc ylla and Char ybdis: Re gulating Bank L e ve r age Whe n the r e is Re nt-Se e king and Risk Shifting” whic h

hig hlig hts be ne fits a nd c o sts o f ha ving ba nk c a pita l.

Appropriate range of leverage

Bank leverage is so low that creditors lack sufficient skin in the game to threaten credible liquidation to discipline rent seeking by bank’s manager. Bank Leverage is so high that it invites risk shifting that increases bank shareholder value but is socially inefficient.

10

• 6. No t a t a ll c le a r tha t ba nks de libe ra te ly to o k ta il risks.

Ra the r, like e ve ryb o dy e lse , the y we re lulle d into a fa lse se nse o f se c urity b y a ve ry lo ng string o f g o o d

• utc o me s.

E vide nc e in F

ahle nbr ac h and Stulz (JFE, 2011) milita te s

a g a inst the no tio n o f de lib e rate ta king o f ta il risks.

11

Bailouts, Contagion, and Bank Risk Bailouts, Contagion, and Bank Risk

by: Giovanni Dell by: Giovanni Dell ‘ ‘Arricia and Lev Ratnovski Arricia and Lev Ratnovski

• Co mmo n wisdo m is tha t b a ilo uts – e spe c ia lly tho se

tha t do no t impo se a ny ha irc uts o n c re dito rs – inc re a se inc e ntive s o f b a nks to ta ke mo re risks b e c a use the y c re a te mo ra l ha za rd.

• T

his pa pe r a rg ue s the o ppo site .

12

• I

de a:

• Ba ilo uts insure a g a inst sta te a a nd thus induc e the b a nk to

inve st mo re prude ntly b e c a use the ma rg ina l re turn to b e ing prude nt is e nha nc e d b y insura nc e a g a inst sta te a .

• I

ntuitio n so me wha t simila r to princ ipa l-ag e nt mo de l in whic h insuring the a g e nt a g a inst risks b e yo nd his c o ntro l (i.e . b y no t pre dic a ting his pa yo ff o n sig na ls tha t a re uninfo rma tive in the

Holmstr

• m (1979) se nse ) he lps to impro ve inc e ntive s.
• Ba ilo uts a re no t a b a d re g ula to ry to o l.
• No te tha t hig he r c a pita l re q uire me nts c a nno t pro vide the

sa me a tte nua tio n tha t b a ilo uts c a n.

Risk a. Systematic risk of entire banking system failing (beyond bank’s control). b. Idiosyncratic risk of failure for a bank (within bank’s control).

13

Comme nts:

• Ve ry pro mising pa pe r. Simple b ut ne a t ide a .
• I

ts the first fo rma l justific a tio n I ha ve se e n fo r b a ilo uts b a se d o n e x ante e ffic ie nc y g ro unds. He re a re so me sug g e stio ns fo r impro ve me nts:

• 1. A b ig imple me nta tio n pro b le m fo r re g ula to rs is

disting uishing b e twe e n syste mic and idio sync ratic failure s. I n re a l time , syste mic fa ilure s do no t a ll o c c ur simulta ne o usly – the y o c c ur se q ue ntia lly.

So …whe n o ne o r a fe w b a nks fa il, ho w do yo u kno w if it is idio sync ra tic o r syste mic ?

14

• 2. A b ig g e r issue (tha n the dire c t e ffe c t o n b a nk risk

ta king ) in b a ilo uts is the a dve rse e ffe c t the se b a ilo uts ha ve o n c re dito r disc iplining inc e ntive s, e spe c ia lly whe n c re dito rs a re spa re d ha irc uts in b a ilo uts. T his is no t c o nside re d he re . Se e Ac har

ya and T hakor , “T he Dar k Side of L iquidity Cr e ation: L e ve r age and Syste mic Risk”, 2011, whe re

this issue is c o nside re d. …What wo uld happe n if this c o nside ratio n was intro duc e d in the mo de l?

15

• E

xisting re se a rc h (Ac har

ya and T hakor (2011))

sho ws tha t this…

• I

nduc e s b a nks to de lib e ra te ly se e k mo re c o rre la te d a sse t c ho ic e s

• Syste mic risk is no t e xo g e no us, b ut is e ndo g e no usly inc re ase d b y

b a ilo uts AND…

• F

ar hi and T ir

• le (AER, for

thc oming) a lso sho w tha t

b a ilo uts

• I

nduc e b a nks to b e c o me mo re hig hly le ve re d, furthe r inc re a sing syste m fra g ility. Ba se d o n the se pa pe rs…

• Diffic ult to c o nc lude tha t b a ilo uts a re g o o d fo r e x ante e ffic ie nc y.

T he ke y is tha t inte rc o nne c te dne ss/ syste mic risk is no t e xo g e no us, b ut is e ndo g e no usly a ffe c te d b y b a ilo uts.

• 3. I

t wo uld b e g o o d if the a utho rs we re to c a re fully disc uss the implic a tio ns o f the se impo rta nt issue s.

16

Regulatory Capture and Regulatory Capture and Supervision Reform Supervision Reform

by: Pierre Boyer and Jorge Ponce by: Pierre Boyer and Jorge Ponce

Que stion: Sho uld the re b e c o nso lida tio n o f re g ula to rs

into o ne sing le re g ula to r o r sho uld we ha ve multiple re g ula to rs? Ke y Re sult: T wo re g ula to rs a re b e tte r tha n o ne !

KE Y I DE A:

• Se lf-inte re ste d b a nk re g ula to rs who hide fro m fina nc ia l sta b ility

c o mmitte e supe rviso ry info rma tio n g a the re d via a udits, in o rde r to e xto rt b rib e s fro m b a nks.

17

18

Comme nts:

• 1. Cle ve r ide a to mo de l this e sse ntia lly a s a ho ld-up

pro b le m o f so rts (b a se d o n pro prie ta ry a c c e ss to supe rviso ry info rma tio n), so tha t the so lutio n invo lve s diminishing the re g ula to r’ s info rma tio n mo no po ly.

• 2. One thing the a utho rs sho uld do is to c ha ng e the

spe c ific a tio n fro m b rib e s (unre a so na b le in ma ny c o untrie s) to re putatio nal c o nc e rns. T he y c ite Boot

and T hakor (AER, 1993) whe re the mo ra l ha za rd is tha t

a re g ula to r ma y de la y c lo sing a n e c o no mic a lly inso lve nt b a nk b e c a use do ing so wo uld b e a n a dmissio n o f pre vio us mo nito ring e rro rs. …A simila r re puta tio na l a rg ume nt c a n a pply to splitting re g ula to rs.

19

I f o ne re g ula to r is re spo nsib le fo r mo nito ring b a nks a nd a no the r fo r de c iding whe the r to c lo se the m, the n wha t ke e ps the mo nito ring re g ula to r fro m c lo sing the b a nk in a time ly ma nne r wo uld no t a pply to the c lo sure re g ula to r.

 E

ffic ie nt c lo sure s!

20

Conclusion Conclusion

Paper Common Wisdom Paper’s Explanation Plausible Alternative in Reality

1) Capital Regulation and Tail Risk Higher bank capital  less risk Higher bank capital  more risk in the presence of tail risk When times are good, banks experience an increase in capital and see risks as being low. Thus, increase risk and sometimes also decrease capital to increase ROE (see Stulz’s JACF paper on ERM and LTCM).  Little protection when times get tough 2) Bailouts, Contagion, and Bank Risk‐Taking Bailouts increase moral hazard and risk Bailouts induce banks to invest more prudently because they are protected against (exogenous) systemic risk Bailouts endogenously increase systemic risk through correlated asset choices and higher leverage 3) Regulatory Capture and Banking Supervision Reform Multiple bank regulators are inefficient Multiple bank regulators are better than single regulator Paper’s story is probably true, but for more subtle reasons that may have more complex regulatory implications.

21