the macroeconomic impact of money market disruptions...the macroeconomic impact of money market...

Overview The model Analysis Numerical results Conclusions Figures

The Macroeconomic Impact of Money MarketDisruptions

Fiorella De Fiore Marie Hoerova Harald Uhlig

ECB & CEPR ECB & CEPR U. Chicago & CEPR

Preliminary

ESSIM, Tarragona23 May 2017

The views expressed are solely those of the authors.


What do we do?

I Document some stylized facts on euro area money marketsduring the recent financial and sovereign crisis

I Build a DSGE model of bank liquidity management withsecured and unsecured money markets, and collateralizedcentral bank funding.

I Calibrate the model and use it to provide a macroeconomicassessment of the observed money market disruptions


Stylized factsI Declining role of interbank markets for EA banks:

I Interbank liabilities to assets: 30% before 2008, 20% by 2013

I Drying up of unsecured markets and shift to securedmarket funding Secured vs unsecured funding

I Increase in private haircuts on some sovereign bonds, farbeyond the increase in ECB haircuts ECB vs private haircuts

I Fears of deposit runs in some stressed countries.I In May 2012, ”Greek depositors withdrew 700 million euros

from banks on Monday” [WSJ], i.e. 0.4% of total depositsin one day.

I Increased recourse to central bank funding Eurosystem funding

I Eurosystem funding to MFI’s deposit liabilities: since 2008,0% in core countries, >25% in peripheral countries


The environment

I HouseholdsI Infinitely-lived; hold money and deposits; consume and

work

I FirmsI Final good producers: owned by banks, use labor and

capital (Cobb Douglas technology)

I Capital good producers: buy existing capital from banks;create new capital goods; sell it back to banks

I GovernmentI Finances expenses through labor taxed and long-term debt.

I Uses income taxes to stabilize debt at targeted level B∗.


The central bank

I Hold discount bonds issued by govt, BCt , with fixed

repayment rate, κ, and price Qt.

I Provides collateralized loans to bank l, Ft,l ≤ ηt QtBFt,l,

applying haircut ηt.

I Issues money and transfers seigniorage to govt

I CB balance sheet at t:

Assets Liabilities

QFt F t (loans to banks) M t (currency in circulation)

QtBCt (govt bond holdings) St(seigniorage)

I CB chooses M t, BCt , Q

Ft , and ηt


Banks: timingI Morning:

I a “type”shock realizes. With prob ξt, a bank is“connected”, can access unsecured mkt

I after type lknown, choice of assets (capital kt,l , bonds Bt,l,money Mt,l, and dividends φNt,l) and liabilities (depositsDt,l, CB loans Ft,l, and net worth Nt,l)

I Afternoon (liquidity management):I iid liquidity shock: deposits reshuffled across banks

I C(onnected) banks: riase liquidity in unsecured mkt

I U(nconnected) banks: raise liquidity in secured mkt up toηtQt(Bt,u −BF

t,u)

I End of period:I Reverse liquidity shock occurs, loans are repaid

I All banks return earnings to mother bank, which allocatesnet worth equally to all banks in t+ 1.


Banks: “morning”problem

Maximize end-of period bank value

Vt,l = ψt,kPtkt,l + ψtBt,l + ψt,MMt,l − ψt,DDt,l − ψt,FFt,l

s.t. for l = c, u,

Vt,l ≥ λ (Ptkt,l +QtBt,l +Mt,l)

Ptkt,l +QtBt,l + φNt = Dt,l +QFt Ft,l +Nt

Ft,l ≤ ηtQtBFt,l

0 ≤ Bt,l −BFt,l

and, for U(nconnected) banks l = u, afternoon constraint:

ωmaxDt,u −Mt,u ≤ ηtQt

(Bt,u −BF

t,u

)


Foreign sector

I Foreign sector allows U banks to change their collateralposition independently from the govt’s stock of debt.

I Foreign demand for domestic bonds:

BWt = Pt

[κ − 1

%logQtπt

],


Analysis: steady state

I Analytical characterization of three different cases:

I “Normal”times (all banks unconstrained)

I Binding afternoon constraint for U banks, no CB funding

I Binding afternoon constraint for U banks, CB funding


Case 1: “normal”times (all banks unconstrained)

I Slack afternoon constraint for U banks

I C and U banks have identical values

I Banks do not borrow from the CB, do not hold money,Fl = Ml = 0

I Returns on capital and bonds equalized:

ψk =ψb

Q


Case 2: Binding afternoon constraint, no CB funding

I Binding afternoon constraint for U banks. Requires that

ψk >ψb

Q

I U banks hold bonds for their collateral valueI U banks hold both bonds and capitalI C banks do not hold bonds

I Sufficient condition for no CB funding: private haircutssufficiently low relative to CB funding cost, η ≥ ηQFωmax

I U banks may hold cash to relax the afternoon constraint(self-insurance)


Case 3: Binding afternoon constraint, CB funding

I Borrowing from CB requires afternoon constraint to bind:

ωmaxDu = ηQ(Bu −BF

u

)+Mu

I Increasing deposits further tightens afternoon constraint

I For low η and QF and η sufficiently close to 1, U banks topup deposit funding with CB funding in the morning

I U banks then hold money Mu to cover afternoonwithdrawals


Calibration

I Key parameters:I Fraction of ’connected’ banks, ξ: 0.42 pre-crisis. Euro Area

Money Market Survey.

I Max withdrawal as share of deposits, ωmax: 0.1. Infoembedded in Liquidity Coverage Ratio: HQLA able to cover30-days liquidity needs under stress over total assets.

I Private and ECB haircuts, η and η: 0.97. Data from LCHClearnet and ECB for 2010.


I Six free parameters: φ, λ, χ, g, BC , B∗. Set to minimize

squared log-deviation of six model predictions fromempirical pre-crisis counterparts

Variable Data Model

Debt/GDP 0.57 0.61

Bank leverage 6.00 5.81

Loan spread (annual) 0.021 0.021

Share bonds unconnected banks 0.23 0.23

Share bonds foreign sector 0.64 0.61

Inflation (annual) 0.020 0.021

I List all parameter values


Unsecured money market freeze

I Comparative statics: reduced access to unsecured mktResults: unsecured market freeze

I Reducing share of unsecured to secured market volumesfrom 42% to 24% generates an output contraction of 0.4%

I The same reduction when fears of deposit withdrawalsdouble (ωmax increases from 0.1 to 0.2) generates an outputcontraction of 4%


Secured money market freeze

I Comparative statics: decrease in η Results: secured market freeze

I Increase in private haircuts from 3% to 25% generates anoutput contraction of 0.3% (0.8% when ωmax = 0.2)

I Impact would be much more severe in absence of CB


Run on deposits

I Comparative statics: increase in ωmax Results: depositor run

I Increase in customer withdrawals from 10% to 20% of totaldeposits generates an output contraction of 3%


Conclusions

I Money markets help manage short-run liquidity needs

I Calibrated DSGE model to assess the macroeconomicimpact of money market frictions

I Three exercises, aiming to capture unsecured and securedmoney market disruptions, and depositor runs

I Output contractions are sizeable (0.4% to 4% in ourscenarios)


To do list

I Policy experiments:

I Shortage of safe assets

I Enlarging the set of collateral accepted at the CB

I Asset purchases by the CB


Quarterly turnover in the euro money market

0

5

10

15

20

25

30

35

40

45

0

5

10

15

20

25

30

35

40

45

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

unsecured secured

Source: Euro Area Money Market Survey. Cumulative quarterly turnover in the euro moneymarket (EUR trillion).The panel comprised 98 euro area credit institutions.

Back


ECB vs private haircuts on sovereign bonds

ECB Private

CQS1-2 CQS3 Germany Portugal

2010 2.8 7.8 2.7 8.12011 2.8 7.8 3.0 10.12012 2.8 7.8 3.0 80.02013 2.8 7.8 3.0 80.02014 2.2 9.4 3.0 80.0

Data source: ECB dataset and LCH Clearnet

Back


Eurosystem funding in total deposit liabilities

0

5

10

15

20

25

0

5

10

15

20

25

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Source: ECB dataset, own calculations.Back


Comparative statics: unsecured money market freeze

9

0.1 0.2 0.3 0.4 0.524

25

26Deposits (d)

9

0.1 0.2 0.3 0.4 0.526

28

30Capital (k)

9

0.1 0.2 0.3 0.4 0.52.65

2.7Output (y)

9

0.1 0.2 0.3 0.4 0.50

1

2Money H (mh)

9

0.1 0.2 0.3 0.4 0.524

25

26Deposits U (du)

9

0.1 0.2 0.3 0.4 0.5

28

30

Capital U (ku)

9

0.1 0.2 0.3 0.4 0.50

0.5

1Money U (mu)

9

0.1 0.2 0.3 0.4 0.51

2

3Bonds U (bu)

9

0.1 0.2 0.3 0.4 0.5

4

4.5Bonds Foreign (bw)

9

0.1 0.2 0.3 0.4 0.524

25

26Deposits C (dc)

9

0.1 0.2 0.3 0.4 0.5

28

30

Capital C (kc)

9

0.1 0.2 0.3 0.4 0.50

0.5

1Bonds pledged CB (bFu )

9

0.1 0.2 0.3 0.4 0.51

1.02

1.04Inflation (:)

9

0.1 0.2 0.3 0.4 0.50.95

1Bond discount factor (Q)

9

0.1 0.2 0.3 0.4 0.50.01

0.02

0.03Coll premium ( ~AK ! ~AB=Q)

Back


Comparative statics: secured money market freeze

~20.4 0.6 0.8

24.5

25

25.5Deposits (d)

~20.4 0.6 0.8

28

29

30Capital (k)

~20.4 0.6 0.8

2.67

2.68

2.69Output (y)

~20.4 0.6 0.8

0

1

2Money H (mh)

~20.4 0.6 0.8

23242526

Deposits U (du)

~20.4 0.6 0.8

26283032

Capital U (ku)

~20.4 0.6 0.8

0

2

4Money U (mu)

~20.4 0.6 0.8

1

2

3Bonds U (bu)

~20.4 0.6 0.8

4

4.5

5Bonds Foreign (bw)

~20.4 0.6 0.8

23242526

Deposits C (dc)

~20.4 0.6 0.8

26283032

Capital C (kc)

~20.4 0.6 0.8

0

1

2Bonds pledged CB (bFu )

~20.4 0.6 0.8

1

1.02

1.04Inflation (:)

~20.4 0.6 0.8

0.9

0.95

1Bond discount factor (Q)

~20.4 0.6 0.8

0.01

0.02


Back


Comparative statics: increased risk of depositor run

!max

0.1 0.2 0.3 0.4 0.510

20

30Deposits (d)

!max

0.1 0.2 0.3 0.4 0.520

25

30Capital (k)

!max

0.1 0.2 0.3 0.4 0.52.4

2.6

2.8Output (y)

!max

0.1 0.2 0.3 0.4 0.50

1

2Money H (mh)

!max

0.1 0.2 0.3 0.4 0.50

20

40Deposits U (du)

!max

0.1 0.2 0.3 0.4 0.50

50

Capital U (ku)

!max

0.1 0.2 0.3 0.4 0.50

1

2Money U (mu)

!max

0.1 0.2 0.3 0.4 0.50

2

4Bonds U (bu)

!max

0.1 0.2 0.3 0.4 0.5

4567

Bonds Foreign (bw)

!max

0.1 0.2 0.3 0.4 0.50

20

40Deposits C (dc)

!max

0.1 0.2 0.3 0.4 0.50

50

Capital C (kc)

!max

0.1 0.2 0.3 0.4 0.50

0.1

0.2Bonds pledged CB (bFu )

!max

0.1 0.2 0.3 0.4 0.51

1.02

1.04Inflation (:)

!max

0.1 0.2 0.3 0.4 0.50.8

1

1.2Bond discount factor (Q)

!max

0.1 0.2 0.3 0.4 0.50

0.02


Back


List of all parameter values

Parameter Description Value

θ Capital share 0.33

δ Depreciation rate 0.02

β Discount rate H 0.999

χ Coeff money in utility 0.0022

g Government spending 1.0081

κ−1 Aver maturity bonds 9

φ Fraction net worth as dividends 0.0191

ξ Fraction ’connected’ banks 0.42

η Private haircut on bonds 0.97

η CB haircut on bonds 0.97

λ Run-away coefficient 0.105

ωmax Max withdrawal as share of deposits 0.3

BC Bonds held by central bank 0.711

% Parameter foreign bond demand 0.09

B∗ Stock of debt 0.711

Back

the macroeconomic impact of money market disruptions...the macroeconomic impact of money market...

Documents