• 沒有找到結果。

Mark-t se of this se

在文檔中 借屍還魂 : 基差交易重現 (頁 23-32)

Section 4 The purpos

4: Mark-t se of this se

9 Professor

4: Mark-t se of this se is package h

ues are calc value assoc e bond minu effects on th ove in isola different fac

noting that largely offs alue while a ng in sign th ple, when an d value and

market-to-m pective this e zero. As

:

r Shyan Yuan

to-Market ection is to a has been en culated by p

iated with t us the price he CDS leg ation, the ba ctors individ

because ba etting; i.e. a at the same t hey may not n issuers cre cash flows

market valu s means at in

s the CDS c

n Lee, “Cred

t Change analyze the ntered, as op product, and the changing e at which th

of the basis asis is explo

dually is im

asis trade is an increased

time bond v t be in magn edit rating is

on the basi

ue of a CDS nception va curve chang

it Derivatives 23

s in the V mark-to-ma pposed to be d not explici g market pr he bond was

s package. F oiting disloc mportant.

a fully hedg d default ris values are fa

nitude. Also s downgrad is trade, whi

S on inceptio alue of the d ges the mark

s and Structu

Value of a arket risks a efore enterin itly ‘netted’

rice of a bon s purchased Finally, whi cations betw

ged position sk portrayed

alling. Ho o, some cha ed, it trigge ile leaving t

on is always default leg m k-to-market

ured Finance

Basis Pac associated w ng the basis

’ by trading nd is simplis d. Thus, I on

ile it is unli ween market

n, Mark-to-m d in CDS ma owever, whi

anges are un ers step-up c the CDS leg

s zero. Fro minus the va

value of the

e” class note

ckage with changin s package. m activit. T stic, the cur nly analyze kely these f ts, thus unde

market chan arket, increa ile some cha nique to one covenants o g unaffected

om the prot alue of the p

e CDS is co

s, Fall Seme

ng factors to-The

mark-rrent traded the mark-factors will

erstand the

nges

tection premium ommonly

ester 2008.

Ann pt Dt

I expand th

Where: The first te market rate the expecte which one the contrac

Mark to m

Changes in First, they spread mat of the carry default we

nuity($) = v = surv = disco his formula

re = change erm is simpl

e CDS curv ed payoffs o

counterpart ct value for

market sens

n the CDS le affect the n tching the m y on the bas

change the

value of CD vival probab ount rate at

to calculate

e in carry current mark

ount rate at vival probab val probabi market CD ginning reco ent recovery ly the prese ve. The dif

on default.

ty would pa both parties

sitivity to p

eg of the tra net carry com maturity on

sis package e present val

S premium bility of defa

year t ed based on

ket CDS spr year t bility of defa

ility of defau DS curve

overy rate a y rate assum ent value of

fference bet This chan ay to “tear”

s is zero.

parallel shif

ade affect th ming from t

the bond (th changes.

lue of payof

24

cash paym ault implied

n the change

read - begin ault implied ult in year t assumption mption

the change tween next t nge in mark-up the cont

fts in CDS

he mark-to-the basis pa he 10yr spre

Second, by ffs received

ents and de d by CDS cu

e in fair valu

nning CDS s d by beginni t implied by

in expected two terms i -to-market v tract by brin

curve:

market valu ackage. By ead in this c y changing t d on default.

fault payme urve

ue of the CD

spread ing CDS cu y the current

d cash flows s the chang value is, in e ng it up to m

ue of basis p y raising or case) we cha the expected . While th

ents

DS as:

urve

t

s, implied b e in present effect, the f market value

package in t lowing the ange the pre d timing of he value rec

by the new t value of fair price

e, where

two ways.

CDS esent value

bond eived

never chan years disco

Figure 14: M

As the cred is gained(l contract pr out. Gain because up premium.

Despite the curve, it is asymmetri deterioratio slow and d

nges, par × ( ounted, and

Mark-to-mark

dit curve sh ost). The ga remium and ns from upw pward shifts

e nearly line important t c. This m on in the cre downward v

(1-R), chang thus the ma

ket sensitivity

hifts up(dow ain recorded d the current ward shifts i s increase th

ear relations to remembe means that su edit quality volatility sm

ging the exp ark to mark

y to parallel s

wn), toward d is equal to t market pre in the credit he probabili

ship betwee er that credi udden upwa of the refer mall.

25

pected time et value.

shifts in CDS

the right(lef o the presen

emium, plus t curve are s ty of defaul

en mark-to-m t risk is a on ard shifts of

rence entity

of receivin

S curve

ft) side of th nt value of th s the increas smaller than lt, lowering

market valu ne-sided jum f the credit c y. Whereas d

ng par chang

he horizonta he differenc

sed probabi n losses for the expecte

ues and para mp risk, thu curve are lik downward s

ges the num

al axis, cont ce between t ility of a def downward ed value of

allel shifts in us volatility kely, given a

shifts are ge mber of

tract value the original fault

pay-shifts the credit

n the CDS is

a enerally

l

Mark-to-m

Like parall CDS contr outs. Unl changes. N the same ti The loss on payments.

Figure 15: M

Mark to m Making ma recovery ra implied de

market sen

lel shifts, ch ract, it chang

like parallel Negative cha ime the redu n the reduct

Positive c

Mark-to-mark

market sens ark-to-mark ate is not a q

fault probab

sitivity to c

hanging the ges the pres l shifts, the anges in slo uced premiu tion in prem changes hav

ket sensitivity

sitivity to c ket changes

quoted rate.

bility, and t

changes in t

slope of the sent value o

effects are ope increase ums and inc miums and c ve equal and

y to changes

hanges in t in value ba . As I show thus has an e

26

the Slope o

e credit curv of the credit

offsetting b e the chance creased cum cumulative s d opposite e

s in the steep

the recover ased on chan

ed before re effect on th

of the CDS

ve has the t premium an based on the e of early de mulative surv

survival dom effects.

pness of the

ry rate:

nging recov ecovery rate

e fair value

curve:

wo effects o nd the prob e structure o efault payme

vival rates r minates the

CDS curve:

very rates is es have a sig

of the CDS

on the fair v bability of de of our steepn ents, a gain reduce cont gain on def

tricky beca gnificant ef S contract.

value of a efault pay-ness

, while at tract value.

fault

ause the ffect on the

The recove volatility is entity issue Recovery r lower reco recovery ra single nam

Figure 16: M

Mark-to-m basis packa may fall si the scenari prepared fo outs becom

ery rate on a s relatively es more deb rates are als very rates, w ates, in gene me CDS is la

Mark-to-mark

market chang age. How gnificantly io in which or significan me more like

a bond is les symmetrica bt (reducing so related to while perio eral, are ver arge.

ket sensitivity

ges in value ever, these as credit ris a basis hold nt mark-to-m ely.

ss volatile t al. Larger g recovery ra o the default

ds with low ry idiosyncr

y to changes

e tend to mo changes are sk increases der receives market vola

27

than the cred changes in ates) or reti t cycle. Peri w default rat ratic and the

s in the recov

ove in the op e not necess s; however, s the largest atility, and e

dit premium n recovery ra res debt (in iods with hi tes have hig e variance o

very rate

pposite dire sarily bad fo

this means t pay-out.

expect volat

m, however ate are likel ncreasing rec

igh default r gh recovery

on realized r

ection expec or the basis companies Thus basis tility to incr

unlike CDS ly when a re covery rates rates tend to rates. How recovery rat

cted returns investor. B

are closer t holder mus rease as def

S premiums eference

s).

o have wever, tes for

on the ond prices to default, st be fault

pay-s

Section 5

One possib contract to entering in in which th intervening effective d exchanged mathemati

The CDS n (value to a slightly to

Where:

5: Basis p

ble modifica buy protec nto a forwar he CDS con g period. T date is later t d until after t cally below

no-arbitrage protection get the no a

n = the co k = the co pt = the cum Φt = cumul

= [the su from 0

=(probab = [pt – p

ackage us

ation to the ction in the f d CDS cont ntract is effe The forward than a vanil

the effectiv w:

e condition sell must eq arbitrage co

ontract matu ontract effec mulative de

ative probab um margina

to k] divide bility of def pk)] / (1 - pk)

sing forw

basis packa future at a p tract today, ective, but a d CDS cont lla CDS wh ve date, even

above, state qual the val ondition for

urity date ctive date efault probab

bility of def al default pr

ed by proba fault from k )] (by Baye

28

ward CDS

age is using pre-specified

the CDS tra assumes the

tract is also ile having t n if the refer

es that the p lue to a prot a k to n yea

bility fault condit robabilities ability of sur k to t) / cum

s’ theorem)

forward CD d premium ader locks i risk of CD of a shorter the same ma rence entity

premium leg tection buye ar forward C

tional on sur from k to t a rvival from mulative surv

)

DS. A For for set leng in a lower p S spread ch r duration, a aturity date.

y defaults. I

g must equa er). I mod CDS:

rvival until and the prob 0 to k vival probab

rward CDS gth of time.

premium for hanges in the

as the contra . No prem

show this r

al the defaul dify this con

t=k

bability of s bility at k

is a By r the period

e act’s miums are

relationship

lt leg ndition

survival

The net eff premium.

credit prem pictured m

Figure 17: C

A basis pac the forward loss in the the BRD fo which is th defaults be increased c basis packa

fect is that t Figure 17 mium falls, a mature in 10

Changes in t

ckage holde d CDS. Th

event of an or different he same as a efore the eff carry) if the age quickly

the orward C shows that assuming th

years).

he CDS prem

er, however hus while th n un-hedged

forward CD a vanilla CD fective date e bond matu y decreases b

CDS contra t as k, the ef he final mat

mium for forw

r, must assu he net carry d default dom

DS contract DS. In eac of the forw ures or defau

bellow zero

29

acts with late ffective date turity date o

ward CDS co

me the cred y on a basis minates the t options.

ch case, the ward CDS, b ults after the o as the effe

er effective e of the con of the contra

ontracts

dit risk on th package wi

gain off the The base ca returns are but positive e effective d

ctive date is

dates have tract, becom act is unchan

he bond unt ith forward e carry. Fi ase is a 0x1

highly nega and slightly date. The s extended.

lower the c mes further nged (all co

til the effect CDS is incr igure 18 bel 0 year forw ative if the b y higher (du expected re

credit distant, the ontracts

tive date of reased, the llow shows ward CDS,

bond ue to the eturn on the

Figure 18: B

While usin basis packa negative sl are based o can be com estimated r quickly the will be trem

Basis pack the credit q effective d More impo

BRD for CWL

ng forward C ages are the lope, forwar on the condi mparatively recovery rat e loss will b mendous.

kages utilizin quality of th date of the F ortant, if the

LN with Forw

CDS on the e same. In rd CDS redu itional marg

small. Gi tes, the astu be minimal,

ng forward he reference FCDS, using e investor ha

ward CDS

Cable & W n cases of ex

uce the carr ginal defaul iven that bo ute trader co

while if it d

CDS can al e entity. If g an FCDS g

as some con

30

Wireless bas xtreme cred ry on the tra lt probabilit nd prices on ould purchas defaults afte

lso be usefu f an investor greatly incr ntrol over th

is package m it risk, wher ade significa ties, which,

n extreme ri se a bond w er the effect

ul if the inve r is sure the reases the ca

he timing of

may not be re the credit antly. For

in the case isk names c with the expe

tive date of

estor has in e firm will n ash flow on f the default

a good idea t curve has rward CDS of steep cre can hover at ectation if it the FCDS,

side inform not default b

the basis p t of the refe

a, not all an extreme premiums edit curves t or around t defaults

the gain

mation on before the

ackage.

erence

entity, grea The next se

Section 6

The pay-ou finance lite refinancing credit defa this section scenarios a predatory b

At the mos of lenders.

or liquidati CDS contr raising the contract on loan issuan

The key fa refinanced CDS pricin

at profits ca ection deals

6: How CD

uts and strat erature. A g or liquidat ault swaps ar n I briefly d as illustratio

behavior an

st fundamen When ch ing a loan, a racts lower’

liquidation n an ex ante nce on an ex

actor in this , or liquidat ng and impl

an be achiev s directly w

DS Chan

tegies for le A creditors o tion and his re fundame discuss how

ons of this c nd the affect

ntal level, cr hoosing the

and make th s lenders pa n payoff if th e basis, and

x post basis.

equation is ted. Every lied surviva

ved by ‘guid ith this kind

ge the Ru

enders and b optimal cont

s beliefs abo ntally chang

CDS affect change. I foc

t of changin

redit default optimal deb heir contrac ayoffs if it c he firm defa

changes the .

that it is wi ything discu al probabilit

31

ding’ the de d of manipu

ules of the

bond holder tract is base out the lend ging the rel t this relatio cus primaril ng capital str

t swaps are bt contract, ct decision b choose to re aults. By d e lenders de

ithin the len ussed in the ies, has assu

fault past th ulations.

e Game

rs are relativ ed on his inf der’s reactio

lationship b onship broad

ly on the ba ructures on

changing th lenders esti based on the efinance a po

definition th ecision if CD

nder’s powe previous se umed that d

he effective

vely well de formation se n set. As etween cred dest sense. I asis trade, in basis holde

he optimal c imate their p ese factors.

oorly perfor his changes DS contract

er to decide ections up to default timin

date on the

efined in rec et, payouts a hedging in ditor and de

I use severa ncluding hed ers.

contracting payoffs on r The prem rming firm, the lender’

ts are entere

if the firm o this point, ng is indepe

e FCDS.

cent on

nstrument, ebtor. In

al different dging,

decisions refinancing mium on

, while s optimal ed into after

is

, including endent and r

beyond the literature.

Scenario 1

In this scen

“senior sec forms of de the loan is not have su capital mar

I analyze th renegotiati

e influence o

在文檔中 借屍還魂 : 基差交易重現 (頁 23-32)

相關文件