Download sovereign sector

Oversight of Sovereign Exposure Concentration
Zlata Komárková and Václav Hausenblas
Financial Stability Department
Agenda
1.
2.
3.
4.
5.
Sovereign exposures and systemic risk
Motivation for Holding of Sovereign debt
Sovereign exposures in Czech banking sector
Policy options
CNB approach: enhanced Pillar 2 requirement
a) A limit on sovereign exposures
b) Sovereign risk indicator (ISR)
2
Main definitions
Sovereign credit risk: a situation in which a sovereign sector
(usually a national government) is unable to honour its pre-agreed
financial obligations unaided and the country therefore defaults.
Financial Stability:
a situation where the financial system
operates with no serious failures or undesirable impacts on the
present and future development of the economy as a whole, while
showing a high degree of resilience to shocks coming from various
sectors of the economy, the financial markets and macroeconomic
developments.
The aim of this part of the presentation is to justify the designation of
sovereign credit risk as systemic risk.
(1.)
Sovereign Exposures and Systemic Risk
The nexus of banks and sovereigns
The two-way interaction between the banking sector and
the sovereign sector:
 sovereign sector: the ultimate source of insurance, the
supply of risk-free GB – a basis for the pricing of other
assets,
 banking sector: the smooth flow of credit to the real
economy, an important creditor for sovereign sector.
A close two-way link between banking and sovereign
distress, with problems in the banking sector having a
negative effect on the sovereign sector, and sovereign
stress exacerbating the disruption in the banking system.
Transmission of banking sector risk to sovereigns
Three main transmission channels :
1. the provision of government support to banks, which increase
sovereign debt => forms: liquidity measures, direct capital increase,
troubled asset relief programs, government guarantees etc.
2. banking sector deleveraging, which, by amplifying the contraction in
overall economic activity, leads to falling budget revenues and rising
budget expenditures.
3. credit booms can give a one-off boost to government‘s fiscal
balances: the government‘s fiscal position appears much stronger
than it actually is, this may unjustifiably give governments the
confidence to pursue policies that result in increases in spending that
are unsustainable in the long run.
Transmission of sovereign credit risk to banks
Four main transmission channels :
1. direct portfolio exposures: the higher bond yields (lower bond
prices) associated with higher sovereign risk can hurt banks
through their holdings of sovereign debt.
2. funding conditions (sovereign securities as collateral): increases in
sovereign risk reduce the availability or eligibility of collateral, and
hence banks‘ funding capacity.
3. the perceived ability of the sovereign to provide a backstop to
banks under strain: being perceived in a weaker fiscal position (for
example rating cuts) provides less credible and valuable
guarantees or financial support => higher liquidity and credit risk
(more expensive funding) for banks.
4. crowding out – the possibility of government debt crowding out
private sector debt, sovereign distress increase the cost and/or
reduces the availability of bank funding through debt.
… two more channels
• The loss of availability of „the risk-free asset“ on whose existence
most market risk management models are based. A loss of power of
such models can in turn lead to a further increase in risk aversion
and to knock-on effects on banks‘ portfolios.
• The channel of change in the attitude to risk: sovereign tensions may
cause a rise in investors‘ risk aversion, which in turn may increase
the risk premiums demanded on sovereign and bank securities and
reduce banks‘ funding availability.
Financial Stability Review, April 2012, BdF
Sovereign exposures – diversity across the EU
Domestic sovereign exposures held by MFI
(in % of total assets, selected EU economies)
EE
FI
NE
LV
FR
DE
AT
BE
ES
IT
SI
CZ
SK
RO
0
5
Source: ECB
Note: As of November 2014.
10
15
20
(2.)
Motivation for Holding of Sovereign Debt
Sovereign securities are regarded as risk-free
risk–free = zero default risk
Government bond markets are generally the most liquid, therefore
government bonds are used as a store of value. There are two main
reasons:
1. intertemporal debt servicing: the sovereign sector as a debtor can
carry its debt forward from one period to the next (from one generation to
another).
2. government has power to change the degree of central bank
independence: government can „print money“ to pay off their debts. Their
ability to do so depends mainly on the degree of central bank independence
and on their ability to change it.
12
Financial markets with a risk–free sovereign
How will markets adjust to a loss of faith in the idea of risk–free
sovereign? Do they need it?
1. sovereign securities serve as the base asset or reserve asset of the
banking system: the low-volatility, low-credit-risk asset around which
bankers and investors build their balance sheets and portfolios,
2. sovereign securities serve as a benchmark that is a reference for value,
3. sovereign securities serve as (high quality liquid) suitable collateral,
4. sovereign securities are used to hedge away certain risks,
5. sovereign securities as a proxy for the risk–free rate that helps to
assess the riskiness of other assets.
13
The regulation gives preferential treatment to sovereign
exposures
The current and newly prepared European banking regulation treats
sovereign risk by essentially not admitting the possibility of the domestic
government defaulting on its debt, and the default risk associated with
government bonds is considered to be zero up to a certain threshold:
1. a low or zero capital requirement for sovereign exposures denominated in the
domestic currency,
2. a low capital requirement for exposures collateralized by government paper
given the very low haircuts required for sovereigns with high ratings,
3. the exclusion of sovereign exposures from the existing limits on large
exposures,
4. automatically to classify government bonds denominated in the domestic
currency as highly liquid, high-quality assets, and
5. moreover, the new Solvency II framework for the insurance sector envisages
a zero risk factor for sovereign exposures.
14
(3.)
Sovereign exposures in Czech banking sector
Sovereign exposures in Czech banking sector (1)
Significance of Czech public debt
• Czech government debt can be
regarded as sustainable, even though
its long-term trend is potentially risky.
• Czech financial system is bankdominated and relatively small.
• Banks operate in environment of
structural liquidity surplus.
• Czech capital market does not offer
enough investment alternatives in local
currency to domestic government
bonds.
• Czech government bonds are held
mainly by domestic banking sector and
mostly to maturity.
(in %)
50
18
45
16
40
14
35
12
30
10
25
8
20
6
15
10
4
5
2
0
0
2008
2009
2010
2011
2012
2013
Debt to GDP
Debt to MFI and insurance sector total assets
Debt held by MFI as a share in assets (right axis)
Source: ECB, Eurostat, Czech National Bank
Public debt to GDP
(in %, selected EU countries)
EE
FI
NE
LV
FR
DE
AT
BE
ES
IT
SI
CZ
SK
RO
0
20
Source: Eurostat
Note: As of 2014
40
60
80
100
120
140
16
Sovereign exposures in Czech banking sector (2)
• Shares of SE in assets of national banking sectors differ significantly.
These shares are not necessarily directly related to size of government
debt (see CZ and SK).
• The share of Czech government bonds in Czech banking assets is
above average by international comparison (EU average less than 4 %).
Domestic sovereign exposures held by MFI
Bank sector assets
(in % of total assets, selected EU economies)
(% of Total Equity, as of 30. 9. 2015)
Loans: Central bank
EE
FI
NE
LV
FR
DE
AT
BE
ES
IT
SI
CZ
SK
RO
129,5
Loans: Interbank
61,5
Loans: Clients
569,8
Bonds: CZ government
144,6
Bonds: Other government
15,7
Bonds: CZ corporate
69,0
Bonds: Other corporate
7,4
Other assets
157,4
0
0
5
Source: ECB
Note: As of November 2014.
10
15
100
200
300
400
500
600
20
Source: CNB
17
(4.)
Policy options
Policy options (1)
• Discussion about changing regulatory approach has already
started in EU and at global level.
• Regulation of banks:
•
•
•
•
•
Pillar 1 capital requirements – removing „domestic carve-out“ in the
SA, introducing a non-zero risk-weight floor in the SA, setting a
minimum floor in the IRB approach…,
Diversification requirements – removal of the examption to the large
exposures regime …,
Regulation of liquidity risk – haircuts that cater to credit risk…,
Macroprudential regulation – capital add ons, applied to all
sovereigns, a group of sovereigns or and individual sovereign, could
be activated when certain conditions are fulfilled (decision by
supervisors),
Enhanced Pillar 2 or Pillar 3 requirements.
19
Policy options (2)
• Concentration of SE in banks’ balance sheets important issue for
Czech banking sector and for the CNB as bank supervisor:
•
The CNB ready to use Pillar 2-based approach until regulatory
discussion delivers different options.
• Constrained discretion
concentration:
necessary
in
responding
to
SE
•
fundamental macroprudential analysis of structural features of
government debt required,
•
rules with some flexibility should be pre-announced.
• Assessment needed when deciding whether to employ tools:
•
economic conditions (government debt characteristics, financial
market possibilities, volume of savings, monetary regime, etc.),
•
initial situation of sectors and intensity of links between them
(absorption capacity, room to adjust, size of buffers, etc.).
20
(5.)
CNB approach: enhanced Pillar 2 requirement
CNB approach: enhanced Pillar 2 requirement (1)
• No scope for any radical solution to SE concentration (trying to avoid
side-effect impacts).
• Our new prudential supervisory tool to be applied for gradually
reducing the risk of systemic concentration of SE under Pillar 2 –
systemic important SE.
• The tool is based on the fact that banks have to ensure consistent
and effective management of concentration risk in their risk
management system.
• The CNB has to review and assess whether banks ensure proper
management and coverage of this risk – from now on according to
the new tool.
• Aim of the tool is not to directly limit investment in sovereign
bonds, but to motivate banks to behave in forward-looking and
prudent manner far enough in advance.
22
CNB approach: enhanced Pillar 2 requirement (2)
• SE defined as exposures to individual governments and their
agencies in EU Member States in all currencies, to which
exemptions apply under Pillar 1,
• An important SE defined as exposures held by a bank with a
minimum ratio of 100% to its eligible capital and become
systemic if the assets of banks with important exposures exceed
5% of their total assets.
• Fundamental principle of the tool: create additional capital from
above-limit SE if increased sovereign risk is indicated three years
ahead.
• The add-on is applied with three-year time delay to enable banking
and sovereign sectors to react in time.
23
Starting points for use of tool after detecting systemic
important SE
• Three main variables for computing the additional capital requirement:
•
Limit for determining above-limit part of systemic important SE at given
time.
•
Estimate for a sovereign risk indicator (ISR as a degree of riskiness) three
years ahead => monitoring of whether threshold value of the indicator is
exceeded.
•
Loss given default (LGD, Article 161 of CRR for FIBR => 45%).
24
Above-limit sovereign exposure
• SE limit is a decreasing function of the ISR in 0–100%.
• As ISR rises, share of below-limit part of SE decreases and share of
above-limit part increases.
•
If the ISR is 0%, the limit is
at its upper bound of 222%
of eligible capital, whereas
if the ISR is 100%, the limit
is at the lower bound of
zero.
•
As the limit decreases
linearly between the two
ISR bounds, its amount
can be expressed for any
ISR value as
((1-ISR)*eligible capital)/0,45.
25
Above-limit sovereign exposure
• Lower limit is 0% when ISR = 100%: the entire exposure is above
limit and expected loss when LGD = 45% is covered by capital.
• Lower limit highly unlikely to be reached – tool will send out timely
signals needed for bank balance sheet adjustment.
26
Additional capital requirement
• WHEN? The CNB will indicate setting additional capital
requirements
• at the earliest when the riskiness of sovereign exposures (the
ISR ) three-year outlook exceeds a „soft“ threshold of 5% and the
results of our analysis confirm necessity,
• or if the ISR three-year outlook exceeds a „hard“ threshold of 8%,
unconditionally.
• WHO? Three years after the indication a bank will have to create
the additional capital requirement:
•
if the bank holds the above-limit SE,
•
and its already allocated capital to cover risk stemming from the SE is not
sufficient.
• HOW HIGH?
•
•
The above-limit part of SE for individual banks according the current ISR
and eligible capital,
The RWA derived from the standard formula set out of the CRR (instead PD
the current ISR will be used).
27
Additional capital requirement
28
Public finance stress tests
• To assess the riskiness of sovereign exposures in a
forward-looking manner CNB conducts a public sector
debt stress test:
•
•
•
•
Since 2015,
On a yearly basis,
On 3Y horizon,
To obtain stressed ISR based of projection of macro/fiscal
fundamentals and market prices under sever stress.
29
Motivation behind ISR and its stress-testing
1. Stress-test allows for longer-term horizons.
2. Avoids reliance on and/or interference with rating
agencies and market sentiment.
3. Stress-testing is a well founded tool in financial stability.
Failure in stress-test is a clear and well-understood
trigger for supervisor’s intervention.
30
Motivation behind ISR and its stress-testing (2)
• Stress-testing framework for banks has relatively clear
measures and benchmarks: capital ratio and its
regulatory minimums (8% + add-ons)
• What measures and benchmarks can we use for public
finances? Traditional measures not suitable:
• Government debt – low explanatory power.
• Primary budget balance – good power, but very cyclical on
its own.
• CDS spreads – volatile, contains noise and other credit
premiums
• → Sovereign risk indicator (ISR)
31
Motivation behind ISR and its stress-testing (3)
• ISR is (intentionally) not PD, although it is loosely related to
1Y PD.
• Estimating PDs for sovereigns is art rather then science.
• Many type of information cannot not be taken into account
using quantitative methods.
• It needs heavy support from expert judgement.
• That would not be very transparent and therefore not suitable
for communication.
• ISR is a benchmarking measure tailor-made for CNB’s
stress-testing framework.
32
Construction of the sovereign risk indicator (ISR)
• ISR is constructed as a composite indicator recalibrated to
0–100% scale.
• Composite of several sub-indicators (variables):
•
•
•
•
Macroeconomic conditions,
Fiscal conditions and debt structure (quality),
Financial conditions,
Quality of institutions.
• Careful selection of indicators to meet the following criteria:
• High goodness of fit.
• Easily connects to stress-test.
• Variables that are part of the scenario should not drive the ISR
too much.
33
Construction of the sovereign risk indicator (ISR) (2)
• Each variable has its critical threshold. Breaching the
threshold sends a signal.
• Thresholds were estimated statistically using historical
data of 40 economies with/without fiscal stress event.
• Noise-to-signal method:
• Find a threshold that minimize both Type I and Type II
errors (i.e. maximize 1-fpr-fnr).
• Method used e.g. by the EC for S0 indicator of fiscal
stress.
• Other methods tends to be highly biased towards minimizing
false positive rate, as the data are heavily unbalanced (to few
fiscal stress observations).
34
Reminder: Type I and Type II errors
• Type I error: indicator signals crisis when there’s no crisis
coming
• Type II error: indicator misses the crisis approaching
outcome
negative (0)
positive (1)
True positive
False positive
(Type I error)
negative (0)
False negative
(Type II error)
True negative
test
positive (1)
35
Construction of the sovereign risk indicator (ISR) (3)
• Each signal has its weight.
• The better the sub-indicator, the higher the weight.
1.
2.
3.
4.
Find the signaling sub-indicators.
Take the sum of the weights of signaling sub-indicators.
Get a number (CI) from 0–100 interval.
This is a great indicator of stress but its relationship to
PD is highly nonlinear, so
5. recalibrate it using the formula:
•
•
ISR = exp(-8.1+10.1CI)/(1+exp(-8.1+10.1CI))
Constant term (-8.1) and slope (10.1) in the logit
expression were estimated using logistic regression.
36
Construction of the sovereign risk indicator (ISR) (4)
Critical limit
• Several kinds of driving factors:
• Macroeconomic conditions,
• Fiscal conditions and debt
structure (quality),
• Financial conditions,
• Quality of institutions.
• Careful selection to meet the
following criteria:
• Has high goodness of fit.
• Easily connects to stress-test.
• Variables that are part of the
scenario should not drive the
ISR too much.
fpr
(%)
fnr
(%)
Weight
(%)
M acroeconomic variables
Real GDP grow th (%)
<
-2,3
4,8
65,0
6,3
<
-1,8
47,6
10,0
8,9
<
19,3
26,5
35,0
8,1
External debt (% of GDP)
>
99,6
18,0
56,0
5,5
Difference betw een real GDP
grow th and real GB yield (pp)
>
6,3
9,0
69,3
4,5
Government debt (% of GDP)
>
64,7
28,0
60,0
2,5
Primary balance (% of GDP)
<
-3,2
17,1
45,8
7,8
10Y government bond yield (%)
>
10,8
30,5
34,1
7,4
>
19,0
23,9
62,3
2,9
>
21,7
52,7
24,6
4,8
>
27,1
31,0
47,2
4,6
>
34,9
49,9
20,2
6,3
<
1,0
33,9
16,8
8,3
Political stability (WGI score)
<
0,8
40,4
13,4
7,8
Rule of law (WGI score)
<
1,2
44,9
10,5
7,5
Banking crisis
>
0
6,4
75,0
3,9
Past sovereign defaults
>
0
21,5
65,0
2,8
Current account balance (% of
GDP)
Gross national savings (% of
GDP)
Fiscal variables
Government debt maturing w ithin
one year (% of GDP)
Share of government debt
maturing w ithin one year (%)
Share of foreign currency debt
(%)
Share of non-residents in debt
holdings (%)
Institutional variables
Government effectiveness (WGI
score)
37
Modelling technicalities: Fiscal model
Overview
38
Modelling technicalities: Fiscal model
Primary balance – calculations summary
Optimistic Variant
Pessimistic Variant
Income
Direct taxes
GDP x Ratio to GDP assumed in fiscal prognosis
Indirect taxes
GDP x Ratio to GDP assumed in fiscal prognosis
Social insurance
GDP x Ratio to GDP assumed in fiscal prognosis
Other current income
GDP x Ratio to GDP assumed in fiscal prognosis
Sales income
GDP x Ratio to GDP assumed in fiscal prognosis
Capital income
GDP x Ratio to GDP assumed in fiscal prognosis
Expenditure
Social payments: healthcare
Social payments: illness and disability
Social payments: pensions
Social payments: unemployment
Social payments: other
Level assumed in fiscal prognosis
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Last value x Index of increase in unemployment rate
GDP x Ratio to GDP assumed in fiscal prognosis
Subsidies
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Other current income
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Interest payable
Not enter Primary Balance
Employees' remmuneration
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Consumption
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Investment
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Net acquisition of assets and inventories
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
Capital transfers
GDP x Ratio to GDP assumed in fiscal prognosis
Level assumed in fiscal prognosis
39
Modelling technicalities: Fiscal model
Interest payments
•
On the other hand, it is very sensitive
to fiscal situation and accelerates
effects of sovereign risk
•
•
CZ (2014): 1.3% GDP; 3.1% total
government expenditures
Portugal 5% of GDP, Italy 4.7%, Greece
3.9% (2014 data)
Thus it demands a proper modelling
How to model interest payable?
• Interest expense in our models
consists of 2 segments
•
•
Old debt – its structure and interest rate
have already been determined; its
development is thus largely automatic
New debt – its interest rates and maturity
structure have to be modelled
Avg past bond yield
Debt inherited from
previous periods
• Our approach uses 15 yearly
maturity buckets
• Data on maturity structure as to the
initial year taken from Bloomberg
Principal
•
Old debt
1Y
2Y
3Y
15Y
Residual maturity
New debt
Debt originating due to
1.
Old debt maturing (1Y bucket from
previous years)
2.
Current period budget deficit
(current expenditures exceeding
revenues)
Current bond yield
3.
Stock-Flow adjustments
(accumulation or dissolution of
government financial reserves)
Principal
Why to model interest payable?
• Interest expense is rather small
component of budget balance
1Y
2Y
3Y
4Y
5Y
15Y
40
Modelling technicalities: Fiscal model
Interest payments
Interest rates on new debt
Maturity structure of new debt
•
•
To determine interest rates on new debt,
we use a panel regression (35 OECD
countries and 24 years of annual data),
and using variables of
•
•
•
•
•
•
•
•
The analysis is conducted for 1Y, 5Y, and
10Y maturities
•
•
Change in public debt to GDP over 2
years*
3M money market rate
Inflation
US government yield at relevant horizon
Issuer credit rating
GDP p.c.
Lagged variable of past yield
For other maturities up to 15 years,
interpolation is used
The model should be easy enough and work
with variables we are able to project
•
Governments might optimize maturity structure
of new debt according to current and expected
financial market conditions
Especially if credit risk worsens and longmaturity debt commands large premium,
issuers might shift to shorter maturities
•
•
•
•
This reduces interest expense
But increases future refinancing need, which is
penalized in the ISR
Government response to the market
conditions is inherently difficult to model
Our model thus assumes previous year‘s
maturity structure of new issues prevails
in future
* Notice the positive feedback loop between public debt
raising interest rates, which raises public debt further due to
higher interest expenses.
41
Results: FSR 2015/2016
Public finance stress test
• FSR 2015/2016 sovereign stress test
results:
• Two sub-indicators send signals
already for 2015.
• These drives the ISR under very
severe macro stress test scenario:
• Real GDP growth,
• GB yields and
• Primary deficit,
• ISR up to 0.8 and then back to 0.27
on three-year horizon.
• This is well bellow the 5% threshold.
Adverse Scenario
M acroeconomic variables
Real GDP grow th (%)
Current account balance
(% of GDP)
Gross national savings
(% of GDP)*
External debt
(% of GDP)*
Difference betw een real GDP
grow th and real 10Y GB
yield (pp)
Fiscal variables
Government debt
(% of GDP)
Primary balance
(% of GDP)
10Y government bond
yield (%)
Government debt maturing
w ithin one year
(% of GDP)
Share of government debt
maturing w ithin one
year (%)
Share of foreign currency
debt (%)
Share of non-residents in
debt holdings (%)*
Institutional variables
Government effectiveness
(WGI score)*
Political stability (WGI
score)*
Rule of law (WGI score)*
Banking crisis*
Past sovereign defaults*
Sovereign risk indicator
(ISR, %)
2015
2016
2017
2018
Critical
lim it
4.3
-2.7
-3.3
-0.7
<
-2.3
0.6
3.3
2.6
1.0
<
-1.8
28.2
28.2
28.2
28.2
< 19.3
70.0
70.0
70.0
70.0
> 99.6
-4.0
5.0
10.4
4.4
>
41.1
45.0
55.5
64.6
> 64.7
0.7
-0.5
-3.8
-5.8
<
0.7
1.7
2.4
3.1
> 10.8
6.7
8.9
12.0
13.4
> 19.0
16.4
19.7
21.6
20.7
> 21.7
15.4
15.5
15.8
11.7
> 27.1
38.4
38.4
38.4
38.4
> 34.9
1.02
1.02
1.02
1.02
<
1.0
1.0
1.0
1.0
1.0
<
0.8
1.1
1.1
1.1
1.1
<
1.2
No
No
No
No
> Yes
No
No
No
No
> Yes
-
0.23
0.80
0.27
Source: CNB, CZSO, ECB, WB, CNB calculation
42
6.3
-3.2
References
• CNB (2016): Financial Stability Report 2015/2016
• Komárková, Dingová, Komárek (2013): Fiscal sustainability and
financial stability, CNB FSR 2012/2013
• Interní metodika ČNB určená pro účely přezkumu a vyhodnocování
rizika koncentrace svrchovaných expozic, CNB website (Czech
only)
• http://www.esrb.europa.eu/pub/pdf/other/esrbreportregulatorytreat
mentsovereignexposures032015.en.pdf?c0cad80cf39a74e20d9d5
947c7390df1
• http://www.cnb.cz/miranda2/export/sites/www.cnb.cz/cs/financni_st
abilita/zatezove_testy/download/InterniMetodika_SvrchovaneExpo
zice.pdf
• http://www.cnb.cz/miranda2/export/sites/www.cnb.cz/cs/financni_st
abilita/zpravy_fs/fs_2012-2013/fs_2012-2013_clanek_ii.pdf
43
Thank you for your attention
www.cnb.cz
Financial Stability Department
zlatuse.komarkova@cnb.cz
vaclav.hausenblas@cnb.cz
44