the dollar, bank leverage and deviations from covered interest rate parity · *bank for...

The Dollar, Bank Leverage and Deviations fromCovered Interest Rate Parity

Stefan Avdjiev*, Wenxin Du**, Catherine Koch* and Hyun Song Shin*

*Bank for International Settlements, ** Federal Reserve Board

Princeton JRCPPF, February 22, 2018

The views expressed in this presentation are those of the authors and not necessarily thoseof the Bank for International Settlements, the Federal Reserve Board of Governors, or the

Federal Reserve System

Avdjiev, Du, Koch and Shin (BIS and FRB) Dollar, Bank Leverage, CIP Deviations JRCPPF, February 22, 2018 1 / 25

Overview

The Role of the U.S. Dollar

Cross-border U.S. dollar-denomianted bank claims and liabilities ($trillions)

9/14

who hold a global portfolio of assets, the currency mismatch between the assets they hold and the commitments they have to their domestic stakeholders in yen, euros or Swiss francs looms large.

For instance, pension funds and life insurance companies have obligations to their beneficiaries and policyholders. These obligations are denominated in domestic currency – in euros, yen or Swiss francs. To the extent that investors face currency risk, they will hedge that risk. We know that investors from emerging economies with large funded pension systems hedge actively.10 However, institutional investors from rich economies will face the problem most acutely, as they have the largest portfolios of global assets. The hedging counterparty is typically a domestic bank, and the bank lays off its own currency risk by borrowing dollars. That way, dollar claims are counterbalanced by dollar debts. The upshot is that banks take on liabilities denominated in dollars in the process of providing hedging services, and dollar intermediation by the global banking system mirrors the underlying currency hedging demands.11

Cross-border US dollar-denominated credit, all sectors

In trillions of US dollars Graph 6

By residence By nationality1

1 The break in series between Q1 2012 and Q2 2012 is due to the Q2 2012 introduction of a more comprehensive reporting of cross-border positions. For more details, see www.bis.org/publ/qtrpdf/r_qt1212v.htm.

Source: BIS locational banking statistics, Tables A5 (by residence) and A7 (by nationality).

Graph 6 provides a window on the total dollar-denominated cross-border bank intermediation

arranged by jurisdiction. In both panels, upward-pointing bars indicate cross-border assets and downward-pointing bars indicate cross-border liabilities. The left-hand panel breaks out the total by residence, while the right-hand panel breaks out the total by nationality, meaning the location of the headquarters. So, for instance, the cross-border claims of a German bank office in London is classified as “UK” in the left-hand panel, but as “euro area” in the right-hand panel.

10 See F Avalos and R Moreno, “Hedging in derivatives markets: the experience of Chile”, BIS Quarterly Review, March 2013, pp 53–62.

11 C Borio, R McCauley, P McGuire and V Sushko, “The failure of covered interest parity: FX hedging demand and costly balance sheets”, BIS Working Papers, no 590, 2016.

–15

–10

–5

0

5

10

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16

Euro areaUnited Kingdom

Claims (+) and liabilities (-) of:SwitzerlandJapan

United StatesOther

–15

–10

–5

0

5

10

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16

Source: BIS Locational Banking Statistics

I Size of FX forward and swap markets:I Notional outstanding: $61 trillion; daily trading volume: $3 trillion


Overview

Covered Interest Rate Parity (CIP)I No-Arbitrage Argument for CIP:

I Borrow $1 at the U.S. dollar risk-free rate y$t ;

I Exchange $1 into St units of foreign currency;I Invest at foreign risk-free rate yt ;I Convert the proceeds back at the pre-determined forward rate Ft ;I No risk, no profits:

St(1 + yt)/Ft = 1 + y$t

or in logs:ρt ≡ ft − st = yt − y$

t .

I In other words,y$

t︸︷︷︸cash market

= yt − ρt︸︷︷︸,swap market

I In every international finance textbook, but no longer true.


Overview

Failure of CIP

I Cross-currency basis = cash market $ rate - swap market $ rate:

xt,t+n = y$t,t+n︸︷︷︸

cash market

− [yt,t+n − ρt,t+n]︸︷︷︸swap market

.

Figure 1: 3M (left) and 5Y (right) cross-currency basis

−25

0−

200

−15

0−

100

−50

050

Bas

is P

oint

s

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

−10

0−

500

50B

asis

Poi

nts

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

Source: Du, Tepper, Verdelhan (Forthcoming)


Overview

Level of CIP Deviations at Quarter EndsWhy is not the basis arbitraged away? – Constraints on banks’ balance sheet.

Figure 1. Quarter-End Anomaly for the CIP Deviations (Yen basis)

010

020

030

040

050

060

070

080

090

0B

asis

Poi

nt

12/31/143/31/15

6/30/159/30/15

12/30/153/31/16

6/30/169/30/16

12/31/163/31/17

6/30/179/30/17

12/29/17

1w deviation 1m deviation 3m deviation

Source: Du, Tepper, Verdelhan (Forthcoming)


Overview

This Paper: Spot-Basis-XB Flows Triangle

Broad Dollar

Cross-Currency Basis XB flows

I Stronger dollar, wider CIP deviations, lower cross-border lending indollars

I The dollar is a risk barometer in global capital markets: stronger dollar,higher shadow cost of banks’ balance sheet capacity and lower bankleverage.


Overview

Broad Dollar and the Basis Post-GFC

117

104

91

78

20

0

–20

–40

20172016201520142013201220112010200920082007

FRB broad dollar index (lhs) Mean cross-currency basis swap spread (rhs)

I Cross-currency basis = cash market $ rate - swap market $ rate.I The cross-currency basis is the mirror image of dollar strength.


Overview

Broad Dollar and the Basis (2000-2017)

−40

−20

020

40B

asis

Poi

nts

6080

100

120

140

Jan

1997

=10

0

2002 2004 2006 2008 2010 2012 2014 2016 2018

FRB broad dollar index (lhs) Mean cross−currency basis spread (rhs)


Overview

Related LiteratureI Failure of CIP:

I CIP deviations during the crisis : Baba and Packer (2009), Baba, Packer andNagano (2008), Baba, McCauley and Ramaswamy (2009), Bottazzi et al.(2012), Coffey, Hrung and Sarkar (2009), McGuire and von Peter (2009),Goldberg, Kennedy and Miu (2011), Griffoli and Ranaldo (2011), Ivashina,Scharfstein and Stein (2015)

I Recent work: Du, Tepper and Verdlehan (Forthcoming), Borio et al. (2016)and Sushko et al. (2016), Liao (2016), Iida et al. (2016), Rime et al. (2016)

I Intermediary and margin-based asset pricing: Bernanke and Gertler(1989), Holmstrom and Tirole (1997), Brunnermeier and Pedersen(2009), Garleanu and Pedersen (2011), He and Krishnamurthy(2012,2013), Brunnermeier and Sannikov (2014), Adrian and Shin(2014), Adrian et. al. (2014).

I Risk-taking channel of exchange rates: Bruno and Shin (2015ab)I FX spot determination in the presence of financial market frictions:

Gabaix and Maggiori (2016)I Dollar factor in exchange rates and equity markets: Verdelhan

(Forthcoming), Brusa, Ramadorai and Verdelhan (2017).


Spot and Basis

Spot and Basis

Broad Dollar



Spot and Basis

Broad Dollar and the Basis

117

104

91

78

20

0

–20

–40

20172016201520142013201220112010200920082007

FRB broad dollar index (lhs) Mean cross-currency basis swap spread (rhs)

I Cross-currency basis = cash market $ rate - swap market $ rate.I The cross-currency basis is the mirror image of dollar strength.


Spot and Basis

Daily Regression of 3M Basis on the DollarTable 2: Regression results of the 3-month cross-currency basis (daily frequency)

(1) (2) (3) (4) (5) (6)ΔDollart -2.610*** -3.007*** -2.964*** -2.229*** -2.291***

(0.539) (0.775) (0.767) (0.658) (0.684)ΔBERit -0.585** 0.284 0.303 0.173 0.286

(0.263) (0.324) (0.316) (0.266) (0.269)lnV IXt 0.0531 -0.00259 0.0273

(0.411) (0.399) (0.409)Δ lnV IXt -0.0195 -0.00189 -0.0163

(0.0188) (0.0197) (0.0189)ΔCurrV olit -0.192*** -0.179***

(0.0576) (0.0571)ΔRRit 1.040 1.173

(0.806) (0.827)Δ(yit − yUS

it ) 0.0888**(0.0370)

Δ(tsit − tsUSit ) -0.0900*

(0.0501)

Observations 25,166 26,513 25,166 24,604 24,530 23,426R-squared 0.015 0.004 0.015 0.016 0.019 0.024


Spot and Basis

Quarterly Regression of 5Y Basis on the DollarTable 3: Regression results of the 5-year cross-currency basis (quarterly frequency)

(1) (2) (3) (4) (5) (6)ΔDollart -1.180*** -0.996** -0.806** -0.774** -0.771**

(0.276) (0.408) (0.334) (0.366) (0.325)ΔBERit -0.522*** -0.131 -0.140 -0.0899 -0.273*

(0.102) (0.123) (0.112) (0.145) (0.145)lnV IXt -2.440 -2.226 -2.086

(2.172) (2.123) (2.041)Δ lnV IXt -0.0389 -0.0258 -0.000603

(0.0237) (0.0274) (0.0304)ΔCurrV olit -0.0490 -0.0247

(0.0440) (0.0388)ΔRRit -0.0215 -0.184

(1.008) (0.932)Δ(yit − yUS

it ) -0.0696***(0.0227)

Δ(tsit − tsUSit ) 0.00148

(0.0148)

Observations 430 430 430 430 430 411R-squared 0.144 0.101 0.147 0.159 0.168 0.208


Spot and Basis

Dollar Beta by Currency

I Dollar beta by currency: ∆xit = αi + βi ∆Dollart + εit

Table 4: Dollar beta by country(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)AUD CAD CHF DKK EUR GBP JPY NOK NZD SEK

Panel (A): 3-month basis, daily frequency∆Broadt -1.122*** -1.303*** -1.519*** -2.193*** -2.511*** -0.903*** -2.060*** -1.243*** 0.339 -1.993***

(0.344) (0.301) (0.354) (0.386) (0.312) (0.311) (0.297) (0.358) (0.340) (0.329)

Panel (B): 5-year basis, quarterly frequency-0.523 -1.563*** -1.904*** -1.825*** -1.310*** -2.283*** -1.177*** -0.0953 -1.113***∆Broadt 0.00326

(0.379) (0.546) (0.374) (0.608) (0.479) (0.292) (0.520) (0.278) (0.408) (0.235)

Notes: This table reports regression coefficients of changes in the cross-currency basis of currency i on the changes in the broad dollar. Panel Ashows regressions based on daily changes using the three-month cross-currency basis and Panel B shows results based on quarterly changes using thefive-year cross-currency basis. We winsorize the daily changes in the cross-currency basis at 1 percent on both tails. The sample period is January 1,2007 to December 8, 2017. Robust standard errors are shown in the parentheses. ***p<0.01, **p<0.05, *p<0.1.

Table 5: Dollar beta and the cross-currency basis(1) (2)

Mean 3M Basis Mean 5Y Basisβi 16.89*** 32.06***

(2.24) (4.94)Constant 2.154 24.17***

(4.64) (6.58)

Observations 10 10R-squared 0.567 0.948

Notes: This table reports the regression coefficients of the mean cross-currency basis in basis points on the dollar beta across countries. Column1 reports results based on the three-month basis and Column 2 reports results based on the five-year basis. e winsorize the daily changes in thecross-currency basis at 1 percent on both tails. The sample period is January 1, 2007 to December 8, 2017.Bootstrapped standard errors for thetwo-pass regressions based on 10,000 replications are shown in the parentheses. ***p<0.01, **p<0.05, *p<0.1.

19


Spot and Basis

Cross-Currency Basis vs. Dollar Beta

I The dollar beta is strongly correlated with the level of the basis.I The dollar is a risk factor pricing the cross-section of CIP arbitrage

returns.I Reversal roles of “safe-haven” and “risky” currencies.

AUD

CAD

CHF

DKK

EUR

GBPJPY

NOK

NZD

SEK

−60

−40

−20

020

3M b

asis

(bp

s)

−3 −2 −1 0 1Dollar beta

3M basis vs. dollar beta (corr=75%)

AUD

CAD

CHF

DKK

EUR

GBP

JPY

NOK

NZD

SEK

−60

−40

−20

020

5Y b

asis

(bp

s)

−2.5 −2 −1.5 −1 −.5 0Dollar beta

5Y basis vs. dollar beta (corr=97%)


Spot and Basis

The Dollar and Basis after the U.S. ElectionI The relationship between the dollar beta and the basis is strongly

confirmed in the event study after the U.S. election.

Changes in the broad dollar index and three-month cross-currency basis since the US election

Cross-currency basis vs dollar beta2

Currency 8/11/ 2016

29/11/ 2016 change

dollar

beta1

Broad dollar 122.8 127.6 4.8 (3.9%) AUD 5.5 8.0 2.5 bps 0.64 CAD –30.0 –40.0 –10.0 bps –2.56 CHF –55.3 –70.8 –15.5 bps –3.97 DKK –67.5 –83.7 –16.2 bps –4.14 EUR –46.4 –61.0 –14.7 bps –3.75 GBP –31.8 –38.8 –7.0 bps –1.79 JPY –70.3 –90.5 –20.3 bps –5.18

NZD 7.5 10.3 2.8 bps 0.70 NOK –32.6 –40.8 –8.2 bps –2.10 SEK –43.2 –54.9 –11.6 bps –2.98

1 The dollar beta is calculated as the ratio of changes in the three-month cross-currency basis over changes in the broad US dollar index between 8 November and 29 November 2016. 2 The vertical axis shows the three-month cross-currency basis expressed in basis points on 8 November 2016, while the horizontal axis indicates the dollar beta.

Sources: Board of Governors of the Federal Reserve System; Bloomberg; BIS calculations.


Spot and Flows

Spot and XB Flows

Broad Dollar



Spot and Flows

XB Dollar Flows and the Board Dollar Index

2008Q4

−10

−5

05

10G

row

th r

ates

in U

SD

lend

ing

(%)

−5 0 5 10Change in broad dollar index (%)


Spot and Flows

Panel Regression of Bilateral $XB Flows on the Dollar

∆xblit = αi + β∆Dollart + γ∆BERit + δCONTRLt + εit

Table 8: Panel Regressions of dollar cross-border lending on the broad dollar(1) (2) (3) (4) (5) (6)

benchmark equal-weight since GFC DM EM OFCΔDollart -0.426*** -0.468*** -0.488*** -0.263** -0.794*** -0.728***

(0.117) (0.100) (0.134) (0.115) (0.155) (0.143)ΔBERit -0.174*** -0.0724** -0.175* -0.235*** -0.0957*** -0.311**

(0.0524) (0.0308) (0.0936) (0.0808) (0.0348) (0.115)lnV IXt -1.996** -2.788*** -1.566* -1.976** -3.770*** -1.054**

(0.869) (0.637) (0.917) (0.944) (1.094) (0.377)Δ lnV IXt 0.00681 -0.00296 0.0250** 0.00393 0.00357 0.0156

(0.0131) (0.00930) (0.0113) (0.0146) (0.00998) (0.0320)ΔIRUS,t 0.102 -1.058** 0.436 0.181 -0.975 0.527

(0.650) (0.439) (0.811) (0.681) (0.692) (1.262)

Observations 6,759 6,759 4,133 1,943 4,096 720R-Squared 0.066 0.058 0.110 0.058 0.106 0.084


Spot and Flows

Evidence from the Structural Panel VAR

I Structural panel VAR based on the Cholesky decomposition, whereyit = [∆irUS,t ,∆iri,t , ln VIXt ,∆xblit ,∆Dollart ]

′.

I We rank ∆xblit before ∆Dollart to tilt the odds against finding effects ofthe exchange raters on xb-flows. Results are robust to alternativeordering.

−1.

5−

1−

.50

0 1 2 3 4 5 6 7 8 9 10

All countries−

2−

1.5

−1

−.5

0

0 1 2 3 4 5 6 7 8 9 10

Developed markets

−1.

5−

1−

.50

0 1 2 3 4 5 6 7 8 9 10

Emerging markets


Model

Triangle


Model

Extended Triangle


Bank Equities

Negative Impact of A Stronger Dollar on Bank Equities

Regressions of bank equity returns on the broad dollar movements(1) (2) (3)

Bank Equity Return Bank Equity Return Bank Equity Return

∆Broadt -2.016*** -0.268** -0.0303(0.127) (0.103) (0.0838)

∆Broadt × bst 2.875***(0.808)

∆Markett 1.246*** 1.236***(0.0527) (0.0524)

Constant -0.00444*** -0.00762*** -0.00728***(3.25e-05) (0.000122) (0.000166)

Observations 3,755 3,755 3,755R-squared 0.102 0.452 0.459


Bank Equities

Dollar, Bank Equity and the BasisBank equity’s dollar beta (normalized by market’s dollar beta) vs. the basis

in the dollar after controlling for the market returns. For countries with a positive basis,

such as the Australia, bank equities actually respond positively to a dollar appreciation

after controlling for benchmark equity index returns.

Figure 8: Sensitivity of bank equity returns to the dollar vs. cross-currency basis

AUDCAD

CHF

DKK

EUR

GBP

JPY

NOK

SEK

USD

.51

1.5

22.

5B

ank

dolla

r bet

a / M

arke

t dol

lar b

eta

-30 -20 -10 0 10 20Mean 5Y Basis (bps)

Notes: On the x-axis, we plot the mean 5-year cross-currency basis by currency. On they-axis, we plot the average ratio of the regression beta of changes in the bank equities onchanges in the broad dollar index over the regression beta of changes in the benchmarkequity index on changes in the broad dollar index by currency. The sample period is2000-2016.

Figure 8 visualize the relationship between the bank equity beta and the cross-

currency basis. On the horizontal axis, we plot the mean 5-year cross-currency basis

by currency. On the vertical axis, we plot the mean ratio of individual bank equity’s

dollar beta over the respective equity index’s dollar beta across banks headquartered in

each currency area. We can see that the relative sensitivity of bank equities over equity

index’s sensitivity with respect to the dollar decreases in the level of cross-currency basis.

In countries with positive bases, such as Australia and Canada, bank equities are less

sensitive to dollar fluctuation than their respective equity indices. In countries with very

negative bases, such as Denmark, Switzerland and Japan, bank equities have significantly

higher sensitivity to dollar fluctuation than their respective equity indices.

In summary, we find that a stronger dollar has a negative impact on bank equities,

and the effect is particularly pronounced for banks headquartered in countries with a

32

I A stronger dollar has a negative effect on bank equities.I The effect is stronger for banks in countries with a more negative

cross-currency basis, or a more severe dollar shortage.


Conclusion

Conclusion

I Triangular relationship among

I The value of the dollar

I CIP deviations

I XB lending denominated in the dollar

I The U.S. dollar is a barometer of risk-bearing capacity in global capitalmarkets.

I A dollar appreciation corresponds to higher shadow cost of bank balancesheet capacity and lower bank leverage.


Conclusion

Appendix Regression results of the cross-currency basis on dollar exchange rates

(1) (2) (3) (4)Daily Δxt,t+3M Daily Δxt,t+3M Quarterly Δxt,t+5Y Quarterly Δxt,t+5Y

ΔDollarEMt -3.115*** -2.339** -0.924** -0.801**

(0.988) (1.014) (0.415) (0.394)ΔDollarDM

t -0.413 -0.418 -0.0312 -0.0406(0.434) (0.440) (0.302) (0.257)

ΔBERit 0.140 0.166 -0.190 -0.315**(0.279) (0.237) (0.123) (0.143)

lnV IXt 0.0938 0.0546 -2.070 -1.928(0.409) (0.409) (2.152) (2.045)

Δ lnV IXt -0.00459 -0.00737 -0.0265 0.00649(0.0192) (0.0199) (0.0295) (0.0340)

ΔCurrV olit -0.159*** -0.0190(0.0567) (0.0376)

ΔRRit 1.252 -0.251(0.817) (0.913)

Δ(yit − yUSit ) 0.0889** -0.0692***

(0.0371) (0.0226)Δ(tsit − tsUS

it ) -0.0903* 0.00774(0.0500) (0.0148)

Observations 24,604 23,426 430 411R-squared 0.019 0.025 0.167 0.211


the dollar, bank leverage and deviations from covered interest rate parity · *bank for...

Documents