Motivation Model Calibration/estimation Results Conclusions
The impact of the credit crunch on the Polisheconomy
Michaª Brzoza-Brzezina and Krzysztof Makarski
Emerging Markets Workshop, Banco de España
21-22 September 2009
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Question
What was the role of the banking sector in the propagation ofthe �nancial crisis?
How much of GDP decline in Poland was generated by�nancial frictions?
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in Poland
Credit crunch
Increase in lending restrictions (and spreads)Decline in loans to households and �rms
Decline in external demand
E�ect: Decrease of exports and lower growth of GDP
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in PolandLending restrictions and credit creation in Poland. (Source: NBP)
Fig 2 Wykres 1
Page 1
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
I 2004 I 2005 I 2006 I 2007 I 2008 I 20090
5000
10000
15000
20000
25000
30000
35000
40000
45000
LTV_Enterprise LTV_Mortgage Enterprise loans Household loans
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in Poland
Credit crunch
Increase in lending restrictions (and spreads)Decline in loans to households and �rms
Decline in external demand
E�ect: Decrease of exports and lower growth of GDP
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in PolandDecline in external demand: GDP yoy in the EA
-5%
-4%
-3%
-2%
-1%
0%
1%
2%
3%
4%
1q2004 1q2005 1q2006 1q2007 1q2008 1q2009
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in Poland
Credit crunch
Increase in lending restrictions (and spreads)Decline in loans to households and �rms
Decline in external demand
E�ect: Decrease of exports and lower growth of GDP
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Short history of the crisis in PolandExport and GDP in Poland
Fig 1 Wykres 3
Page 1
-10
-5
0
5
10
15
2004 2005 2006 2007 2008 2009
GDP Exports
%
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Literature
The interest in models with �nancial frictions is growingrapidly.
Two main strands of literature:- Bernanke et al. (1999): �nancial accelerator through prices(interest rates). The cost of monitoring increases spreadsbetween the lending and the risk free rate.- Iacoviello (2005): �nancial accelerator through quantities.Collateral puts a constraint on the volume of lending.
Work in progress!!!
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Key Features
Small open economy new Keynesian model of business cycle.
Non-neutrality of monetary policy due to: - Nominal rigidities -prices and wages are set according to a Calvo scheme (Calvo,1983)
Monetary Policy - Taylor rule (Taylor, 1993)
Financial frictions - following Iacoviello (2005) and Gerali et al.(2008):
collateral constraints on HH and enterprises,stochastic shocks to credit constraints and spreads.
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Structure
Patient HH
Impatient HH
Entrepreneurs
Final good producers
Banks Central bank ROW
Housing/ Capital good producers
Deposits
Labour
OMO
Loans
Wholes. goods
Retailers
Interm. goods
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Households
There are two types of households: patient, P , and impatient,I .
In equilibrium patient HHs save and impatient HHs borrow.
Patient HHs maximise utilityE0 ∑t(βP)tu(cPt , χP
t , nPt , εt),subject to the budget constraint.
Impatient households maximise utilityE0 ∑t(βI )tu(c It , χI
t , nIt , εt),
subject to the budget constraint and the credit constraintRH
L,tLHt ≤ mH
t Et
[Pχ,t+1 (1− δχ) χI
t
]where mH
t ∼ AR(1) denotes LTV ratio. Thus to get a loanthey need collateral (housing).
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Entrepreneurs (wholesale good producers)
Do not work, run �rms. Sell their product in a competitivemarket.
Are impatient (βE = βI ).Own capital and use it as collateral for loans.
Maximise utilityE0 ∑t(βE )tu(cEt , εt),
subject to the �ow of funds and the credit constraintRF
L,tLFt ≤ mF
t Et [Pk,t+1 (1− δk) kt ]where mF
t ∼ AR(1) denotes LTV ratio.
Other (intermediate and �nal good) producers are standard.
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Banking
Monopolistically competitive markets - nominal rigidities(Calvo) slow down the adjustment of the interest rates.
Banks have access to the domestic and international interbankmarket.
Deposit banks- Collect deposits from patient households and deposit them inthe interbank market. zD,t ∼ AR(1) a�ects the spreadbetween the interbank rate Rt and the HHs deposit rate RH
D,t .
Lending banks- Take loans in the interbank market and grant loans toimpatient HHs and �rms. zL,t ∼ AR(1) a�ects the spreadbetween the interbank rate Rt and the lending rate RL,t .- Borrowing constraints tied to housing (HHs) and capital(�rms)
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Calibration/estimation
Parameters that are widely discussed in the literature andsteady state parameters were calibrated.
Remaining parameters Bayesian Estimation,
Quarterly data 1q1997 - 1q2009,
14 time series:
PL: GDP, C, I, G, Inf, WIBOR3M, HH loans, Enterprise loans,Spread on HH loans, Spread on enterprise loans, Spread on HHdeposits.EUR: GDP, Inf, EURIBOR3M.
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
HHs loans LTV shock
10 20 30 40−1
0
1
2
3x 10
−3 gdp
10 20 30 40−5
0
5
10
15x 10
−3 c
10 20 30 40−2
0
2
4x 10
−3 i
10 20 30 40−2
0
2
4
6x 10
−4 pi
10 20 30 400
0.5
1x 10
−3 R
10 20 30 40−5
0
5
10
15x 10
−4 l_f
10 20 30 40−0.05
0
0.05
0.1
0.15l_h
10 20 30 40−5
0
5x 10
−5 spread_f_l
10 20 30 40−5
0
5x 10
−5 spread_h_l
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
HHs loans spread shock
10 20 30 40−4
−2
0
2x 10
−4 gdp
10 20 30 40−2
−1
0
1x 10
−3 c
10 20 30 40−4
−2
0
2
4x 10
−4 i
10 20 30 40−10
−5
0
5x 10
−5 pi
10 20 30 40−15
−10
−5
0
5x 10
−5 R
10 20 30 40−2
0
2
4x 10
−4 l_f
10 20 30 40−15
−10
−5
0
5x 10
−3 l_h
10 20 30 40−2
−1
0
1x 10
−5 spread_f_l
10 20 30 40−1
0
1
2
3x 10
−3 spread_h_l
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Historical realisations of selected shocks
10 20 30 40-0.04
-0.03
-0.02
-0.01
0
0.01v_y_star
10 20 30 40-0.01
-0.005
0
0.005
0.01v_pi_star
10 20 30 40-6
-4
-2
0
2x 10
-3 v_R_star
10 20 30 40-0.015
-0.01
-0.005
0
0.005
0.01v_z_h_d
10 20 30 40-0.015
-0.01
-0.005
0
0.005
0.01v_z_h_l
10 20 30 40-0.01
-0.005
0
0.005
0.01v_z_f_l
10 20 30 40-0.3
-0.2
-0.1
0
0.1
0.2v_m_f
10 20 30 40-0.2
-0.1
0
0.1
0.2v_m_h
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Simulation
The goal of the simulation is to separate out the impact of thecredit crunch on GDP,
We run counterfactual simulation: hypothetical GDP without�nancial shocks,
Selected shocks are set to zero in the period 3q2008-1q2009,
We group shocks - shocks to spreads and to LTVs,
To obtain a better visualisation of the impact of the creditcrunch we make an unconditional forecast for 20 quarters (itshould be treated as a visualisation of the strength of theimpact rather than a real forecast),
Comparison with the impact of external shocks.
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Scenario 1: No spread shocks
0 10 20 30 40 50 60 70-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
Counterfactual scenario and forecastData and forecastImpact
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Scenario 2: No LTV shocks
0 10 20 30 40 50 60 70-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
Counterfactual scenario and forecastData and forecastImpact
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Scenario 3: No �nancial shocks
0 10 20 30 40 50 60 70-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
Counterfactual scenario and forecastData and forecastImpact
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Scenario 4: No external shocks
0 10 20 30 40 50 60 70-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
0.025
Counterfactual scenario and forecastData and forecastImpact
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Plan of the Presentation
1 Motivation
2 Model
3 Calibration/estimation
4 Results
5 Conclusions
M. Brzoza-Brzezina and K. Makarski Credit crunch
Motivation Model Calibration/estimation Results Conclusions
Conclusions
Model allows to identify the impact of the �nancial turmoil onthe Polish economy,
The role of direct credit restrictions is greater than the role ofthe price of credit (spreads),
Financial shocks have a quantitatively important e�ect,nevertheless their impact is smaller than the impact of externalshocks,
Still work in progress: moment conditions, endogeneity of thebanking sector, entry-exit of �rms etc.
M. Brzoza-Brzezina and K. Makarski Credit crunch
Thank you.