Transcript The „out-of-sample“ method

Excessive Credit Growth and
Countercyclical Capital Buffers in Basel III
An Empirical Evidence from Central and East
European Countries
Adam Geršl
Jakub Seidler
Česká a světová ekonomika po globální finanční krizi
International Scientific Conference VŠFS
Prague, CNB, 25 November 2011
Countercyclical capital buffer in Basel III
• should protect the banking sector from the credit cycle
• act against procyclicality in the financial system
• protect the banking sector from periods of excess aggregate credit
growth that have often been associated with the build up of systemwide risk
• the aim is to ensure that the banking sector in aggregate has the capital
on hand to help maintain the flow of credit in the economy when the
broader financial system experiences stress after a period of excess
credit growth
• may help to lean against the build-up phase of the cycle
• raising the cost of credit, and therefore dampening its demand, when
there is evidence that the stock of credit has grown to excessive
levels relative to the benchmarks of past experience
• this potential moderating effect on the build-up phase of the credit cycle
should be viewed as a positive side benefit, rather than the primary aim of
the countercyclical capital buffer regime
2
The role of credit-to-GDP gap
• the common reference guide is based on the aggregate private sector
credit-to-GDP gap
• a gap between the observed value and the calculated long-term
trend of private sector credit to GDP
Countercyclical capital buffer
• for calculation of the long-term
(% of RWA as a function of credit-to-GDP gap in pps)
trend, the Basel committee
-1,4
suggests
-1,3 using the Hodrick-1,2
Prescott
-1,1filter with a high
-1
smoothing
parameter
-0,9
-0,8
(lambda=400,000)
-0,7
-0,6
• buffer set
as a function of the
-0,5
-0,4
credit-to-GDP
gap
-0,3
• role for-0,2
macroprudential authority
-0,1
0
to assess,
decide and
0,1
0,2
apply/remove
4
3,5
3
2,5
2
1,5
1
0,5
0
-4
-2
0
2
4
6
8
10
12
14
Credit-to-GDP gap (in %)
3
The role of a macroprudential authority regarding CCB
• to monitor credit and its dynamics (and potentially other
indicators) and make assessments of whether system-wide
risks are being built up
• based on this assessment, to decide whether the CCB
requirement should be imposed (set above the zero value)
• to apply judgment to determine whether the CCB should
increase or decrease over time (within the range of zero to
2.5% of risk weighted assets, in very strong credit booms
even above 2.5%)
• to be prepared to remove the requirement on a timely basis if
the system-wide risk crystallizes
4
Buffer as a countercyclical instrument
Credit dynamics (e.g. y-o-y growth)
period of financial
exuberance
period of financial
distress
CCB set to zero again
time
CCB set at maximum
2,5 %
CCB set to zero
turning point (start of
crisis): credit growth
falls, lending conditions
tighten
5
Countercyclical capital buffer in Europe (CRD IV)
• a series of discussions and consultations in the EU to
• appropriately implement Basel III in the EU
• and, at the same time, reflect a unique nature of EU financial
markets (large integration, common rules = single rulebook)
• the discussions are still going on (draft proposal for CRR/CRD IV
published in July 2011)
• open issues related to the buffer
• cyclical versus structural buffer
• voluntary versus obligatory cross-border reciprocity
• national discretion versus single rulebook
• role of European Systemic Risk Board (ESRB) in setting
guidance (ex ante agreement within the ESRB on indicators
versus ex post reaction by ESRB via warnings and
recommendations)
6
Converging EU countries from the CEE region
• CEE countries experienced rapid credit growth in the pre-crisis
period 2002–2007
• policymakers in some countries assessed the credit boom as
excessive, applying a number of tools to limit it (with mixed
evidence as to their effectiveness)
Credit growth and number of tools applied to limit credit boom
(horizontal axes: number of tools;
vertical axes - average y-o-y real credit growth 2005-2007)
0.4
Tools to limit credit growth
Monetary policy tools
- interest rate increases
LT
0.35
LV
RO
- reserve requirements
Regulatory measures
0.3
- higher risk weights/capital charges
EE
0.25
BG
SI
- restrictions on LTV/LTI
CZ
0.2
- provisioning rate
SK
0.15
- tight regulation on large exposures
PL
- rules on collateral valuation
HU
0.1
Administrative measures
0.05
- quantitative restrictions on credit growth
- eligibility criteria for borrowers
0
0
2
4
6
8
10
12
- tax treatment of loan-related payments
- guidelines and recommendations
Source: IMF, national authorities' websites
7
Particular features of credit boom in the CEE countries
• despite rapid credit growth, the level of credit still relatively low in
many CEE countries in the pre-crisis year 2007 if compared to the
rest of the EU; however,
• some CEE countries already aproached levels observed in the euro area
Bank credit to the private sector in selected EU countries
(as % of GDP)
250%
CEE countries
200%
150%
100%
50%
0%
CY IE NL
GR BE FI
Source: IMF IFS, authors' calculations
EE LV SI
CZ SK PL RO
• may be underestimated (only
bank credit captured, but in the
CEE also cross-border and nonbank credit plays a role)
• FX lending prevalent in several
CEE countries (but not in all)
• credit booms funded via external
borrowing of domestic banks,
usually from parent companies
(foreign ownership of banking
sectors in the CEE)
8
CEE countries as a case study to try out the buffer calculation
• application of the suggested credit-to-GDP guide challenging for CEE
countries
• short time-series (Basel recommends 20 years of quarterly data)
• high stable rise of credit growth is incorporated in the trend
(convergence in credit to GDP)
Development of credit to GDP ratio in CEE countries
(%)
Development of credit to GDP ratio in CEE countries
(%)
120
100
90
100
80
70
80
60
60
50
40
40
30
20
20
10
0
1998Q1 1999Q3 2001Q1 2002Q3 2004Q1 2005Q3 2007Q1 2008Q3
Hungary
Poland
Romania
Slovenia
0
1998Q1 1999Q3 2001Q1 2002Q3 2004Q1 2005Q3 2007Q1 2008Q3
Bulgaria
Estonia
Source: IMF IFS, authors' calculations
Source: IMF IFS, authors' calculations
Latvia
Lithuania
9
CEE countries and buffer calculation
• initial undershooting and catching up hypothesis
• banking sector restructuring (bad assets, bank privatization) and
changes in the composition of credit to the private sector
• end-point bias as an obstacle to conduct of macroprudential policy
Development of credit to GDP ratio in CEE countries
(%)
Stock of bank credit to the real sector in the Czech Republic
(in CZK bil)
80
2000
70
1800
1600
60
1400
50
1200
40
1000
30
800
600
20
400
10
200
0
1998Q1 1999Q3 2001Q1 2002Q3 2004Q1 2005Q3 2007Q1 2008Q3
Czech Republic
Source: IMF IFS, authors' calculations
Slovak Republic
0
93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
Non-financial corporations
Source: Czech National Bank
Households
10
The Czech Republic
• using HP filter with recommended lambda would indicate excess credit
growth (since 2003 if estimated continuously)
• holds even if other reasonable denominators are used (assets)
• discretion of policymakers needed
Credit gaps in the Czech Republic with alternative denominators
(%)
Bank credit to the real sector in relative terms
(%)
70
20
60
15
50
10
40
5
30
0
20
-5
10
-10
0
1996
1998
2000
Source: CNB, CZSO
2002
2004
2006
2008
Credit-to-GDP
Credit-to-financial-assets
Credit-to-assets
2010
-15
03/98
03/00
03/02
03/04
03/06
03/08
Credit-to-GDP gap
Credit-to-GDP gap (actual period estimation)
Credit-to-financial-assets gap
Credit-to-assets gap
03/10
11
Source: CNB, authors' calculations
Methods used to find equilibrium credit
Frequently used
• Detrending data (HP – filter)
• Hilbers et al. (2005)
• VAR / Cointegration for individual country
• Hoffman (2001)
• Out-of-sample panel regression
• Cottareli et al. (2005), Égert et al. (2006)
• Panel VAR
• Eller et al. (2010)
Less frequently used – not used so far for excess credit
• Kalman filter, e.g. Bacchetta et al. (1999)
• Structural models / DSGE, e.g. Gerali et al. (2010)
12
Historical comparison
•
indicates that the Czech Republic has lower level of credit than selected core EU
countries when at similar level of economic development (in the 1980s)
other features of credit growth in the Czech Republic also do not indicate the build up of
system-wide risk (no FX loans to households; no external funding; high deposit-to-loan
ratio; low LTV ratios)
contrasts with Latvia that had comparatively much higher stock of credit in 2008, several
„dangerous“ features of credit boom and lower level of GDP per capita
•
•
Credit to GDP for similar level of economic development
(GDP per capital in 2005 USD = 17 ths USD; in %)
Credit to GDP for similar level of economic development
(GDP per capital in 2005 USD = 14 ths USD; in %)
100
100
90
90
80
80
70
70
60
60
50
50
40
40
30
30
20
20
10
10
0
0
CZ
(2009)
IT
(1986)
NL
(1984)
AT
(1984)
Source: IMF IFS, authors' calculations
FR
(1985)
ES
(1990)
DE
(1983)
Latvia
(2008)
IT (1984)
ES (1988)
Source: IMF IFS, authors' calculations
AT (1985)
DE (1985)
13
The „out-of-sample“ method
• a way how to form judgment about the sustainable level of credit in the
economy, based on (see Backe, Egert and Zumer 2006; Kiss, Nagy
and Vonnák 2006):
• regressing the credit to GDP on a range of economic fundamentals
(GDP per capita; households consumption; inflation etc.), using
data for developed countries
• applying the estimated elasticities „out of sample“, i.e. on CEE
countries to calculate „equilibrium credit“
• in contrast to HP filter, the out-of-sample method takes into account
the economic fundamentals influencing the level of credit in the
economy
14
Model specification
•
Dynamic nonstationary heterogenous panel estimator
• Pesaran et al. (1999)
• Short-run effect different for cross-sections
• The same long-run cointegration relationship for all countries
• SR effect:
• inflation (-),
• ∆ (consumption to gdp)(+)
• LR effect:
• consumption to gdp(+),
• gdp per capita (+),
- other determinants tested
15
The „out-of-sample“ method: results
• the out-of-sample estimation leads to different credit-to-GDP gaps
• Czech Republic does not seem to have been in excessive credit situation
in 2009, while Estonia and Latvia do
• for results for other countries see Annex
Czech Republic
Latvia
Estonia
20%
50%
40%
15%
40%
30%
10%
20%
30%
10%
5%
20%
0%
0%
10%
-5%
-10%
0%
-10%
-20%
-15%
-10%
-20%
-20%
-40%
-25%
2000q1 2002q1 2004q1 2006q1 2008q1
-30%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
-50%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
HPgap
OUTgap
-30%
HPgap
OUTgap
HPgap
OUTgap
16
The „out-of-sample“ method
• different estimations of „gap“ lead to different levels of buffer
• to be fair, the out-of-sample method also has a number of drawbacks
• fundamental variables selection (housing prices seem to be a relevant
variable, but can themselves suffer from bubbles)
• different fundamentals in out-of-sample countries and countries of our
interest at the current stage of development
• estimation method suffers by losing country-specific constant
Countercyclical capital buffer
(% of RWA)
Out-of-sample
HP filter
Out-of-sample
10.8
2.5
2.5
-15.0
2.4
0.0
27.9
1.0
2.5
-8.3
1.5
0.0
19.6
0.0
2.5
-10.7
0.0
0.0
-23.3
0.3
0.0
-27.3
1.3
0.0
-22.8
1.3
0.0
5.5
1.1
1.1
Credit-to-GDP gap (%)
HP filter
BG
11.4
CZ
9.5
EE
5.3
LT
6.9
LV
1.0
HU
-1.4
PL
3.0
RO
6.1
SK
6.1
SI
5.4
Source: authors' calculations
17
Countries above equilibrium credit
• however, it seems that for CEE countries, the out-of-sample method better
predicts the problem countries
• empirical evidence shows that the four countries identified as being above
equilibrium credit (LV, BG, EE, SI) and the two close to the border (HU and LT)
did not show particularly high Tier 1 capital ratios before crisis in 2008 (except
Bulgaria) and some of them experienced relatively high drop in RoE of banks
Credit-to-GDP gap via out-of-sample and Tier 1 ratio in 2008
(gap in pps; Tier 1 capital ratio in 2008)
Credit-to-GDP gap via out-of-sample and change in RoE
(gap in pps; change in RoE of banking sector in pps)
10
CZ
13
0
RO
12
-40.0
BG
CZ
-20.0
RO
0.0
HU
PL
-10
SI
20.0
40.0
BG
11
SK
SK
PL
-20
LV
10
LT
HU
-30
EE
SI
9
-40
LV
8
-50
LT
7
6
-40.0
-20.0
0.0
Source: IMF, authors' calculations
20.0
40.0
-60
-70
Source: IMF, authors' calculations
EE
18
Thank your for your attention!
19
Annex: comparison of HP and out-of-sample methods
Bulgaria
Slovakia
20%
20%
15%
10%
10%
0%
5%
0%
-10%
-5%
-20%
-10%
-15%
-30%
-20%
-40%
-25%
-50%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
HPgap
-30%
2000q1 2002q1 2004q1 2006q1 2008q1
OUTgap
Lithuania
HPgap
OUTgap
Hungary
20%
15%
10%
10%
5%
0%
0%
-10%
-5%
-10%
-20%
-15%
-30%
-20%
-25%
-40%
-30%
-50%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
HPgap
OUTgap
-35%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
HPgap
OUTgap
20
Annex: comparison of HP and out-of-sample methods
Poland
Romania
15%
10%
10%
0%
5%
-10%
0%
-5%
-20%
-10%
-30%
-15%
-20%
-40%
-25%
-50%
-30%
-35%
2000q1 2001q4 2003q3 2005q2 2007q1 2008q4
HPgap
OUTgap
-60%
2000q1 2002q1 2004q1 2006q1 2008q1 2010q1
HPgap
OUTgap
Slovenia
20%
15%
10%
5%
0%
-5%
-10%
-15%
-20%
-25%
-30%
2000q1 2002q1 2004q1 2006q1 2008q1 2010q1
HPgap
OUTgap
21
References
•
Bacchetta, P., Gerlach, S. (1997): Consumption and credit constraints: International evidence, Journal of Monetary Economics,
Volume 40, Issue 2, Monetary Policy and Financial Markets, October 1997, pp. 207–238.
•
Boissay, F., Calvo-Gonzales, O. and Kozluk, T. (2006): Is Lending in Central and Eastern Europe Developing too Fast?, Finance
and Consumption Workshop presentation, June 2006.
•
Brzoza-Brzezina, M. (2005): Lending Booms in Europe’s Periphery: South-Western Lessons for Central Eastern Members.
EconWPA, February.
•
Calza, A., Gartner, Ch. and Sousa, J. (2003): Modelling the Demand for Loans to the Private Sector in the Euro Area, Applied
Economics, Vol. 35, No. 1, pp. 107–117.
•
Cottarelli, C., Giovanni D. and Vladkova-Hollar, I. (2005): Early birds, late risers, and sleeping beauties: Bank credit growth to the
private sector in Central and Eastern Europe and in the Balkans, Journal of Banking & Finance 29, no. 1 (January): 83–104.
•
Égert, B., Backé, P. and Zumer T. (2006): Credit growth in Central and Eastern Europe - new (over)shooting stars?, European
Central Bank WP, No. 687, October 2006.
•
Eller, M., Frömmel, M., and Srzentic, N. (2010): Private Sector Credit in CESEE: Long-Run Relationships and Short-Run Dynamics,
Focus on European Economic Integration, no. 2.
•
Gerali, A. & Neri, S., Sessa, L. and Signoretti, F. (2010): Credit and banking in a DSGE model of the euro area, Working papers No.
740, Bank of Italy, Economic Research Department.
•
Hilbers, P., Otker-Robe, I., Pazarbasioglu, C. and Johnsen, G. (2005): Assessing and Managing Rapid Credit Growth and the Role
of Supervisory and Prudential Policies, IMF Working Paper, Vol. 151, No. 5, pp. 1–59.
•
Hofmann, B. (2001): The determinants of private sector credit in industrialized countries: Do property prices metter? BIS Working
Paper 108.
•
Kiss, G., Nagy, M. and Vonnák, B. (2006): Credit Growth in Central and Eastern Europe: Convergence or Boom?, MNB Working
Papers 2006/10, Magyar Nemzeti Bank (The Central Bank of Hungary).
22