The Credit Ratings Game - kif.re.kr Freixas PPT.pdf · The Credit Ratings Game Bolton, Freixas, and...
Transcript of The Credit Ratings Game - kif.re.kr Freixas PPT.pdf · The Credit Ratings Game Bolton, Freixas, and...
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
The Credit Ratings Game
Patrick Bolton, Xavier Freixas, and Joel Shapiro
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
The Subprime Crisis
Dollar value of originations of subprime mortgages rosefrom $65 billion in 1995 to $600 billion in 2006
Moody�s pro�ts tripled between 2002 and 2006, withpro�t margins of 50%
Income from rating structured deals amounted to 50% ofCRA total income
Extreme complexity of ABS, CDOs (Gorton, 2009)
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Key Elements of the Credit Rating Industry
Issuer pays, and payments may in�uence ratings:
Fees are renegotiated with regular clients.
Analytical managers participated in fee discussions withissuers and sta¤ discussed fees and market share (SEC)
CRAs o¤er related consulting services, such as pre-ratingassessments
Possible con�icts of interest
Issuers shop for ratings:
rating agencies are only paid if the credit rating is issued
"What the market doesn�t know is who�s seen certaintransactions but wasn�t hired to rate those deals" - BrianClarkson, president and COO (until July 2008), Moody�s
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Agencies�models are not precise:
Errors have been detected (and abused as in ABACUS)
agencies do not perform due dilligence on issuer data
Large barriers to entry exist:
SEC prohibited entry by creating NRSRO (NationallyRecognized Statiscally Rating Organization) designation
Reputation plays a role in decisionmaking :
agencies state that their business is dependent on marketcon�dence
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Motivation
Understand the con�icts of interest characteristic of creditrating agencies
Consider the e¢ ciency issues
Derive the pros and cons of three di¤erent plans to reformthe industry
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
This paper
We incorporate these elements in a model of how CreditRatings Agencies (CRAs) manage con�icts of interest andwe examine proposed regulations of the industry.
Our �ndings:
For the same information revelation strategy duopoly isless e¢ cient than monopoly in terms of both total ex-antesurplus and consumer surplusMore precise information increases current payo¤s, but alsoincreases the probability of getting caught.The most e¢ cient solution is to change to a system ofupfront fees, automatic disclosure of ratings, and oversighton analytical standards
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Roadmap
Literature Review
Monopoly
Duopoly
Welfare
Regulation
Conclusion
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Related Literature
Certi�cation intermediaries: Biglaiser (1993)
Sobel (1985), Benabou and Laroque (1992)communication strategies
Certi�cation intermediaries with endogenous lying costs:Strausz (2005) Mathis, McAndrews and Rochet (2008)
Certi�cation intermediaries with exogenous lying costs:Bolton, Freixas, and Shapiro (2007)
Transparency of a screening process: Farhi, Lerner, andTirole (2008), Pagano and Volpin (2009)
Shopping: Skreta and Veldkamp (2008), Sangiorgi etal.(2010)
Empirical results: Becker and Milbourn (2008)
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
The Model
Three players: Issuer, Credit Rating Agency, InvestorsAn Investment: ex-ante is of type g with probability 1
2 , botherwiseProb (g defaults)=0, Prob (b defaults)=pReturns R when not defaulting, r when defaultingCRA gets a private signal θ 2 fg , bg about the true stateω:
Pr(θ = g j ω = g) = Pr(θ = b j ω = b) = e
Pr(θ = g j ω = b) = Pr(θ = b j ω = g) = 1� e
e measures the precision of the signal (known)
Fees can be set conditional on report of CRA: φm , wherem = G ,B.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Investment technology
W=g
R
W=b
R
rp
1p
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Signalling strategy: bad state of nature
g
b
W=b
Signal
e
1e
G
B
G
B
p
1p
p
1p
p
1p
p
1p
R
r
R
r
R
r
R
r
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
After observing the report, the issuer can purchase it, orrefuse to purchase it (shopping).
Once the rating is announced (or not), the issuer sets auniform price T
Investors can be sophisticated (1� α) or naive (α)
We assume exogenous reputation costs: in the event of adefault after a good rating, investors �nd out whether theCRA lied and punish by withdrawing their business.
Discounted sum of future pro�ts= ρ
Assumption A0: There is a tiny amount of uncertainty onthe part of the CRA about the actual value of ρ, i.e.ρ 2 [ρ̃� ε, ρ̃+ ε] such that ε ! 0. This uncertainty isresolved when the CRA receives its signal.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Investors are risk neutral and purchase 0, 1 or 2 unitsReservation utility: u on the �rst unit and U on thesecond unit, where U > uCuto¤ p� is de�ned by the indi¤erence condition:
(1� p�)R + p�r = U.Also:
(A1) (1� p)R + pr > u
(A2) (1� e)p < p�
(A3)p2> p�
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Timing
1 The CRA posts its fee of φm , where m = G ,B.
2 The issuer asks for the signal to be retrieved or not.
3 Given a request by the issuer, the CRA receives the signaland then makes a report of m = G or m = B,
4 The issuer observes the report and decides whether to buyand distribute it or not. The issuer then sets a price T fora unit of the investment.
5 Investors observe the price T and the CRA rating if thereis any and decide how much of the investment to purchase.
6 The return is realized.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Notation
V 0 =Value of the investment absent any (credible)informationV G =Value of the investment if g is obtained and truthfullyannounced G=gV B =Value of the investment if b is obtained and truthfullyannounced B=b
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Information Regimes
Lemma
For a given set of fees φm , the CRA�s reporting strategy is:
1 For φG � φB > epρ, the CRA always reports �G�
2 For 0 < φG � φB < epρ, the CRA reports the truth,relaying its signal perfectly.
3 For φG � φB < 0, the CRA always reports �B�
Lemma
The issuer never buys a m = B report. This implies that theCRA�s actual reporting strategy is:
1 For φG > epρ, the CRA always reports m = G, and
2 For 0 < φG < epρ, the CRA reports the truth, relaying itssignal perfectly.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Equilibrium Fees
Proposition
There are two equilibria of the fee setting game:
1 If α2V G � V 0 > epρ, the CRA sets φG = α2V G � V 0,always reports m = G, and has pro�ts
α2V G � V 0 + (1� ep2)ρ,
2 If α2V G � V 0 < epρ, the CRA setsφG = min[2V G +max[αV 0,V B ]� 2V 0, epρ], reportstruthfully, and has pro�ts
12min[2V G +max[αV 0,V B ]� 2V 0, epρ] + ρ.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
The cuto¤ α2V G � V 0 � epρ determines which informationregime prevails
Low reputation costs or many naive investors makeoverstating more likely
Low p makes overstating more likely. This could be relatedto the business cycle.
Higher precision means higher current pro�ts, but morelikely to get caught
In truthtelling, fees are bounded above by expectedreputation costs
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Two CRAs
Game has same structure, fees are posted simultaneously
De�ne V GG ,V BB
Extra assumption:
(A4) α2V G � V 0 > 2(V GG � V G )
The discounted sum of future pro�ts for each CRA notcaught lying is ρD
Information regimes are the same as before
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Equilibrium Fees
Proposition
The Nash equilibrium of the fee setting subgame is:
1 If α2(V GG � V G ) > epρD , both CRAs always report G
2 If α2(V GG � V G ) < epρD , both CRAs report truthfully,and
1 If α 2 [32V
0
2V G , 1], the issuer hires both CRAs
2 If α 2 [ V 02V G ,32V
0
2V G ), the issuer only hires one CRA and
φRk = φIk = 0, k = 1, 2
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Results - Two CRAs vs. One CRA
The cuto¤ α2(V GG � V G )� epρD determines which regimethe CRAs are in
Current payo¤s are larger with one CRA, but so are futurecosts if ρ � ρD
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Total Ex-Ante Surplus
Proposition
A truthtelling duopoly (when both CRAs are hired) is lesse¢ cient than a truthtelling monopoly.
Moreover, TS is the same in monopoly or duopoly whenthe CRA(s) always report G
Proposition is the same for Investor Surplus. IS is strictlylarger in monopoly when the CRA(s) always report G
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
RegulationThe Cuomo Plan
Issuers must pay CRAs upfront (not conditional on thereport)
With one or two CRAs: selects the truthtelling regime ofthe model
Eliminating shopping from the Cuomo plan
Two CRAs: Issuer only purchases one report and CRA feesare zero.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
RegulationAn Investor�s Pay Solution
Investor�s pay system disappeared because of free-ridingproblem in the early 1970s.
It could be re-established through taxation (of investors orissuers).
If this prohibited shopping, it could be as good as theCuomo plan without shopping.
However, there are additional regulatory costs: choosingthe optimal tax, monitoring CRA performance andchoosing how many CRAs are active.
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
RegulationMoral Hazard
If precision e is chosen by the CRA after receiving paymentsand is non-contractible, the CRAs would choose the minimumprecision in all three regulatory plans.In the main model, the CRAs would choose an e > 1
2Therefore, these plans would need oversight of minimumanalytical standards
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Regulatory reform
Dodd Frank Act Committee�More independent directors for CRAs, penalties and Internal�rewalls reduce con�icts of interest�The Franken amendment also addresses shoppingGlobal Financial Stability 10 Recommendations�Improve market discipline by improving information
The CreditRatings Game
Bolton,Freixas, andShapiro
Introduction
MonopolyCRA
Duopoly
Welfare
Regulation
Regulatoryreform
Conclusions
Conclusions
Higher pro�ts, more naive investors and lower reputationcosts foster in�ation rating
One CRA can be better than two in terms of TS and ISdue to shopping
Prohibiting shopping from Cuomo plan is optimal, butneed to monitor analytical quality
Restructuring reduces surplus