Post-crisis bank regulations and financial market liquidity
Darrell DuffieGSB Stanford
Belgian Research Financial FormNational Bank of Belgium
Brussels, June, 2018
Based in part on research with Leif Andersen, Antje Berndt, Yang Song, and Yichao Zhu
Duffie Post-crisis bank regulations and financial market liquidity 1
A bank-intermediated over-the-counter market
c1
d1
c2
d2 d3c3
c4
c6
c7
c5
Duffie Post-crisis bank regulations and financial market liquidity 2
Implications of post-crisis regulations for market efficiency
1 More financial stability from higher bank capitalization and bail-in failure resolution.
2 Increased cost of access to bank balance sheets.
• The leverage-ratio rule has reduced incentives to intermediate safe assets.
• Bail-in failure resolution has significantly increased bank funding costs.
3 Market infrastructure and new competition rules lower the need for balance-sheet space.
Duffie Post-crisis bank regulations and financial market liquidity 3
Dealer balance sheet
assets
debt
equity
Duffie Post-crisis bank regulations and financial market liquidity 4
More equity to fund more assets
assets
debt
equity
old assets
new assets
debt
new equity
equity
Duffie Post-crisis bank regulations and financial market liquidity 5
Legacy shareholders have subsidized creditors
assets
debt
equity
old assets
new assets
debt
new equity
equity
Higher capitalization implies a value transfer from legacy shareholders to creditors.
Duffie Post-crisis bank regulations and financial market liquidity 6
Debt overhang
assets
debt
equity
old assets
new assets
debt
new equity
equity
For shareholders to break even, the new assets must be purchased at a profit that exceeds thevalue transfer to creditors. (Myers, 1977)
Duffie Post-crisis bank regulations and financial market liquidity 7
Leverage ratio rule is more binding than risk-based capital rulesResults of the Fed’s 2017 stress tests for the largest US dealer banks
JPM CITI BAML GS MS
CET1 (CCAR)CET1 (DFAST, adj.)SLR (CCAR)SLR (DFAST, adj.)
02
46
810
Exc
ess
capi
tal r
atio
(%
)
CCAR: stressed CET1 after assumed payouts, less 4.5%; stressed SLR less 3.0%.DFAST, adjusted: stressed CET1 (no payouts) less (4.5% + G-SIB surcharge); stressed SLR less the G-SIBminimum of 5%.
Data source: Board of Governors of the Federal Reserve, 2017.Duffie Post-crisis bank regulations and financial market liquidity 8
European banks reduce their balance sheets at quarter ends
Daily collateral outstanding in the tri-party repo market and the Federal
Reserve’s overnight reverse repo (ON RRP) facility
0
100
200
300
400
500
600
1/2016 4/2016 7/2016 10/2016 1/2017 4/2017
U.S. banks European banks Other banks Fed (ON RRP)
Figure Source: Egelhov, Martin, Zinsmeister, Federal Reserve Bank of New York, August, 2017.
Notes: Banks headquartered in the euro area and Switzerland report leverage ratios as a snapshot of their value on the last day of each quarter, while their U.S. counterparts report quarterly averages. Totals only include trades backed by Fedwire-eligible securities–that is, U.S. Treasury and agency securities.
Billions of dollars
Duffie Post-crisis bank regulations and financial market liquidity 9
Impact of the leverage-ratio regulationon repo intermediation costs to legacy shareholders
old assets
old debt
equity
repo asset repo claim
old assets
repo asset
old debt
repo claim
equity
new equitysafe assets
Duffie Post-crisis bank regulations and financial market liquidity 10
Impact of SLR on UST repo market efficiency
13−Q1 13−Q3 14−Q1 14−Q3 15−Q1 15−Q3 16−Q1 16−Q3 17−Q1 17−Q3 18−Q10
5
10
15
20
25
GC
F−
trip
arty
rat
e sp
read
(ba
sis
poin
ts)
(a) bid-ask spreads up39
Decline in GCF net lending volume
(b) inter-dealer positions down
Figure: (a) Average within-quarter difference between overnight GCF and Tri-party repo rates. Data sources:Bloomberg and BNY-Mellon. (b) Figure source: Antoine Martin, FRBNY (2016).
Duffie Post-crisis bank regulations and financial market liquidity 11
Cross-currency basis and bank funding costsFunding value adjustments now leave wider arbitrage bounds on the basis
Five-Year Cross-Currency Basis: G10 Currencies
−100
−50
050
Basis
Poin
ts
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
AUD CAD CHF DKK EUR
GBP JPY NOK NZD SEK
(a) 5-year USD cross-currency basis. Source: Du,Tepper, and Verdelhan (2017).
US banksEuropean banks
2004 2006 2008 2010 2012 2014 2016 2018
050
100
150
200
250
300
year
CD
S r
ate
(b) 5-year dealer credit spreads
Duffie Post-crisis bank regulations and financial market liquidity 12
CIP arbitrage can be costly to dealer shareholdersDebt overhang cost for funding synthetic dollar deposits
assets
debt
equity
old assets
EUR→USD
old debt
USD debt
equity
To benefit shareholders, the trade profit must exceed the funding value adjustment (FVA), adebt-overhang cost.
Duffie Post-crisis bank regulations and financial market liquidity 13
Funding cost to shareholders
old assets
EUR→USD
old debt
USD debt
equity
funding value adjustment (FVA)
A debt-funded safe arbitrage is not valuable to bank shareholders unless it’s excess yield isabove the bank’s credit spread. Source: Andersen, Duffie, Song (2018)
Duffie Post-crisis bank regulations and financial market liquidity 14
Credit spreads: funding-cost wedge and arbitrage bounds
2002 2004 2006 2008 2010 2012 2014 2016 2018
050
100
150
200
250
year
one−
year
IBO
R−
OIS
spr
ead
(bas
is p
oint
s) EURIBOR−OIS (Eonia)USD−LIBOR−OIS (Fed funds)
Figure: Spreads between one-year IBOR and OIS rates. Data source: Bloomberg.
Duffie Post-crisis bank regulations and financial market liquidity 15
At a given solvency level, big-bank credit spreads are higherTime fixed-effects in the relationship between CDS rates and distance to default
Big banksAll other firms
2002 2004 2006 2008 2010 2012 2014 2016
01
23
45
67
year
CD
S ti
me
fixed
effe
ct
Based on panel regression, from work in progress with Antje Berndt.
Duffie Post-crisis bank regulations and financial market liquidity 16
GSIB 5-year credit spread at annual default probability of 0.5%Before adjusting for pre-crisis bailout protection
Big banks
2002 2004 2006 2008 2010 2012 2014 2016
050
100
150
200
250
year
CD
S r
ate
(bas
is p
oint
s)
Based on panel regression, from work in progress with Antje Berndt.
Duffie Post-crisis bank regulations and financial market liquidity 17
Sovereign support has been removed from GSIB ratings
Other firmsGSIBs
2002 2004 2006 2008 2010 2012 2014 2016 2018
02
46
810
12
year
Ref
ined
rat
ing
Group medians of refined ratings. Data source: Moody’s Investor Services.
Duffie Post-crisis bank regulations and financial market liquidity 18
Average leverage of systemic banks and investment banks
Dealers: GS−MS−LEH−BSC−MERBanks: C−BAC−JPM*−WFC
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
05
1015
2025
3035
40
year
Leve
rage
Data source: 10K filings. J.P. Morgan includes preceding mergers, pro forma.Duffie Post-crisis bank regulations and financial market liquidity 19
Market-to-book equity ratios of systemic dealer banks
Dealers: GS−MS−LEH−BSC−MERBanks: C−BAC−JPM*−WFC
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
01
23
4
year
Mar
ket−
to−
book
equ
ity r
atio
Data source: 10K filings. J.P. Morgan includes preceding mergers, pro forma.Duffie Post-crisis bank regulations and financial market liquidity 20
Funding Costs to Dealer ShareholdersFrom work with Andersen and Song: The marginal increase in the value of the dealer’s equity per dollar of adebt-funded asset purchase is
p∗π − δCOV∗ − FVA,
where
I p∗ is the dealer’s risk-neutral probability of survival to term.
I π is the trade profit (P&L).
I δ is the risk-free discount.
I COV∗ is the risk-neutral covariance of the asset payoff and dealer default event.
I FVA is the funding value adjustment p∗δST , where S is the dealer’s credit spread and T is the term.
The extra marginal cost to dealer shareholders when a fraction α of the funding must be equity isα(1− p∗ − FVA), which annualizes to roughly αS (assuming a loss given default of 0.5).
For safe assets, the shareholder breakeven “arbitrage” yield is thus the total annualized funding cost to
shareholders of roughly (1 + α)S.
Duffie Post-crisis bank regulations and financial market liquidity 21
When should a dealer arbitrage the USD-JPY CIP basis?
0100
200
300
400
12/31/14 3/31/15 3/30/15 9/30/15 12/30/15 3/31/16 6/30/16 9/30/16
1w deviation 1m deviation 3m deviation
(a) Level of Yen CIP Deviations
−300
−200
−100
0100
200
12/31/14 3/31/15 3/30/15 9/30/15 12/30/15 3/31/16 6/30/16 9/30/16
3m−1m deviation 1m−1w deviation
(b) Term Structure of Yen CIP Deviations
Figure 7: Illustration of Quarter-End Dynamics for the Term Structure of CIPDeviations: In both �gures, the blue shaded area denotes the dates for which the settlementand maturity of a one-week contract spans two quarters. The grey shaded area denotesthe dates for which the settlement and maturity dates of a one-month contract spans twoquarters, and excludes the dates in the blue shaded area. The top �gure plots one-week,one-month and three-month CIP Libor CIP deviations for the yen in red, green and orange,respectively. The bottom �gure plots the di�erence between 3-month and 1-month LiborCIP deviation for the yen in green and between 1-month and 1-week Libor CIP deviationfor the yen in red.
52
Source: Du, Tepper, and Verdelhan (2016).Duffie Post-crisis bank regulations and financial market liquidity 22
Central counterparties reduce need for balance-sheet space
c1
d1
c2
CCP
d2 d3c3
c4
c6
c7
c5
Duffie Post-crisis bank regulations and financial market liquidity 23
Compression eliminates space used for redundant swaps
d1
d2
d5
d3
d4
10
40
60
50
40
d1
d2
d5
d3
d4
10
40
20
10
0
Figure: Counterparty exposures and initial margin are reduced without changing market exposures. Providersinclude TriOptima, which has eliminated over $1 quadrillion notional of swaps.
Duffie Post-crisis bank regulations and financial market liquidity 24
A bank-intermediated bilateral OTC market
c1
d1
c2
d2 d3c3
c4
c6
c7
c5
Duffie Post-crisis bank regulations and financial market liquidity 25
Improving trade competitionObjective: Migration of actively traded products to all-to-all trade platforms
c1
CLOBd1
d2c2
c3
c5c4
c6
Duffie Post-crisis bank regulations and financial market liquidity 26
OTC competition after Dodd-Frank and MiFIDBuy-side firms request quotes at multilateral trading platforms
c1
MTP
d1 d2
Duffie Post-crisis bank regulations and financial market liquidity 27
Excessive fragmentation across platforms
c1MTP1
d1
d2
MTP2
d2
d3
Duffie Post-crisis bank regulations and financial market liquidity 28
Reducing fragmentation improves competition
c1
MTP
d1
d2
d3
Duffie Post-crisis bank regulations and financial market liquidity 29
At corporate bond platformsDealer competition lowers buy-side trade costs
-‐10
0
10
20
30
40
50
60
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Cost in Basis Points
Number of dealers responding
Investment Grade
High Yield
Source: Hendershoc and Madhavan (2014) Source: Hendershott and Madhavan (2016)Duffie Post-crisis bank regulations and financial market liquidity 30
Now typical fragmented two-tiered OTC markets
c1
d1
c2
CLOB
d2
MTP1
d3c3
MTP2
c6
c7
MTP3
c5
Duffie Post-crisis bank regulations and financial market liquidity 31
Appendix: How CIP arbitrage costs dealer shareholdersI Suppose the one-year USD risk-free rate is zero.
I Our bank has a one-year credit spread of 35 basis points.
I We borrow $100 with one-year USD commercial paper, promising $100.35.
I We invest $100 in one-year EUR CP, swapped to USD, with the same all-in credit quality as that of ourbank’s CP, and uncorrelated.
I Suppose the EUR CP, swapped to dollars, promises $100.60, for a basis of −25bps.
I We have a new liability worth $100 and a new asset worth $100.65/1.0035 ' $100.25, for a trade profit ofapproximately $0.25.
I However, the marginal value of the trade to our shareholders is negative, because, conditional on dealersurvival, the expected incremental payoff to equity is
$100.25 − $100.35 = − $0.10. Conditional on default, equity gets nothing.
Duffie Post-crisis bank regulations and financial market liquidity 32