1. Standard fiscal measures and bank rescue measures

The Effects of Bank Rescue Measures in the recent Financial Crisis

Jan in 't Veld DG-ECFIN, European Commission

Werner Roeger DG-ECFIN, European Commission

September, 2011

The views expressed in this paper are those of the authors and should not be attributed to the European Commission.

1. Introduction There has been an intense debate about the effects of fiscal measures in the recent

financial crisis.

The Debate concentrates on the effects government purchases and transfers to

households, and of tax cuts (see, e.g., Coenen et al. (2010), Corsetti et al. (2010)

and Drautzburg and Uhlig (2010)).

However, a key dimension of the fiscal policy response to the crisis were sizable

government interventions in the banking system, in the form of bank asset

purchases, loan guarantees and bank recapitalization.

As documented below, these ‘unconventional’ fiscal interventions were actually larger

than the changes in standard fiscal instruments enacted during the crisis.

Surprisingly, the macro-economic effects of these unconventional fiscal measures

directed at the financial sector have, so far, received little attention in the literature.

This paper seeks to fill this gap, by analyzing the effect of state-aid to banks, using

the Commission's macroeconomic model QUEST III augmented by a financial sector.

Our Approach:

We make a simple extension of a standard DSGE model by distinguishing between

three types of households

1 Risk averse savers (deposits, government bonds)

2 Less risk averse savers (equity of banks and non financial firms)

3 (Mortgage) Debtors

Losses originate from debtor households and are borne by equity owners.

Government intervention may be helpful in spreading risks across all households

Other Approaches: There are other models which emphasise moral hazard (Gertler and Karadi (2010)), adverse selection (Ikeda (2011)), asymmetric information and monitoring costs (Hirakata et al. (2009))

Also Angeloni and Faia (2010) have build a Diamond/Rajan (2000) banking model

into a DSGE model and have used it for analysing alternative fiscal and monetary

policies

Krishnamurthy (2009) provides an overview of the various models.

He makes a distinction between two alternative amplification channels, namely via

balance sheet effect and via uncertainty (as emphasised by Caballero (2009) or

Bean (2010)).

Standard fiscal measures

US: American Recovery and Reinvestment Act (ARRA) – about 2% of GDP ,

EU: European Economic Recovery Plan (EERP)- about 0.8% of GDP

Table 1: Conventional fiscal stimulus measures (as % of GDP) US EU 2009 2010 2009 2010 Total fiscal stimulus 1.98 1.77 0.83 0.73 of which Government expenditure 0.67 0.80 0.30 0.15 Transfers 0.64 0.20 0.24 0.09 Tax reductions 0.67 0.77 0.29 0.49 Source: Coenen et al (2010).

1. Standard fiscal measures and bank rescue measures

Government support to banking sector

(1) recapitalisation

(2) guarantees on banks' liabilities

(3) purchases of toxic or impaired assets, "bad banks".

Recapitalisations and asset purchases combined amounted to roughly 5% of EU GDP

in total over the crisis.

Liability guarantees were much larger almost 8% at its peak.

Table 2: State aid for financial sector (as % of GDP)

Feb-09 May-09 Aug-09 Dec-09 Oct-10 Dec-10 Apr-11 IAR 0.43 0.45 0.75 2.84 2.15 2.00 1.94 Recap 1.09 1.45 1.67 1.88 2.17 2.21 2.11 Guarantees 6.56 7.30 7.95 7.79 5.80 5.61 5.07 Total 8.08 9.19 10.38 12.51 10.12 9.82 9.12

Source: Commission services (survey based)

3. The Model

2 Regions, the EU and the RoW.

Standard DSGE Model with:

Three types of households: savers, investors/equity holders and debtors.

Corporate sector consists of banks and non financial firms.

3.1 Corporate Sector

3.1.1 Non Financial Corporate Sector

Production function with capital tK and labour tN

(1) Y

tttt ZNKY

1

, .

Dividends

(2) tItttt

NFt JpNwYdiv )(

Max problem:

(3)

10

0 0

100

)1(

)1(

tJttt

tt

NFjt

t

t

j

Ejt

NF

KZJKE

divrEVMax

3.1.2 The banking sector

(2)

))())((

)()1()(

)()1(

))(1(())((

2

11111

111

11111

BGjt

BPjtjt

Gjtjtjt

jtGtjtjt

Djt

Gjtjtjt

Gjtjt

Pjtjt

Ajt

Gjt

Gjtjt

Ljtt

BGjt

BPjt

Bjtt

SSqLLD

DLLDrLLguar

LLdefLdefrtox

LLrESSdivE

Government bailout policies:

Recapitalisation measures take the form of a purchase of newly issued bank shares

at the current market price ( 0 Gtt Sq ).

Government purchases of toxic assets consists of purchasing loans ( GjtL ) and taking

over loan losses at rate Gjtrtox

Guarantees on loans at rate jtguar provide insurance to the banking system.

Max problem:

(3) jt

Bjt

t

t

j

Ejt

B SdivrEVMax

1

0 0

100 )1(

.

From these FOCs we obtain the following loan interest rate rule (5) D

tE

tL

t rrr 1)1(

3.2.1 Savers

(8)

10

0

,1

,

00

0,

11

11,

0

000

)1(

)1(

)1(

)1(

),,,(

tLandt

Landtt

ts

t

st

st

sHt

HsHt

st

ts

t

st

st

ts

tsH

tHt

Landt

Lt

stt

st

Dt

stt

sst

Constrt

sHt

Ht

st

Cttss

t

t

st

st

st

sst

stss

LandgJLand

HJH

TJpJpNwDr

BrDBJJpCp

DHNsCUVMax

t

Savers have preferences over consumption, labour, housing and deposits (liquidity

service.)

Savers supply banking system with deposits.

3.2.2 Debtors

1) Higher rate of time preference ( sc )

2) Collateral constraint on their borrowing tL .

3) Banks impose a loan to value ratio tcc

t z .

Max problem

(13)

ct

Ht

ct

Lt

tct

ct

ct

HcHt

ct

tc

t

ct

ct

t

ct

cttt

At

Ltt

cHt

Ht

ct

Ct

tcct

t

ct

ct

ct

ctcc

HpLrE

HJH

TNwLdefrLJpCp

HNCUVMax

t

)1(

)1(

))1(

),1,(

,0

1,

00

01

,0

000

3.2.3 Equity owners

et

ct

NFt

NFt

BPt

Bt

Bt

BPt

Bt

tet

et

e

t

teE CpdivqSqdivSqECUEVMax )1()()( 111,

00

,00

(Inverse of the) stochastic discount factor for corporate investment

(17) )1(11,

, Etec

te

tC

ct

etC

t rpU

pUE

Note: Dividends are not exogenous to the equity owners. Corporate sector makes

decisions in order to optimise the dividend stream. However, optimisation is

constrained by capital adjustment costs (NFF) and capital requirements (Banks).

3.5 Fiscal Policy

Standard, except for income and expenditures related to bank rescue measures.

Gt

Lt

Gttt

Gt

Bt

Gt

Gtttttt LrSdivTSqDEFLGUARGBrB 11111 )1(

4. Calibration

For the non-financial sector we use parameter estimates from Ratto et al. (2009) and

In 't Veld et al. (2011) for the Euro area and the US respectively.

Skewed wealth distribution:

10% equity owners, savers and debtors represent 45% of the population.

Luxembourg Wealth Study: top 10% of the population in the EU own roughly 50% of

total net worth (financial assets + dwellings + consumer durables - liabilities)

The equity owners in the model own roughly 65% of total net worth. (Note: we do not

consider consumer durables).

Individual household types have different rates of time preference and risk aversion

parameters:

Savers and equity owners have the same discount factor of 0.99 but differ in their

degree of risk aversion.

Savers have a 1h , while equity owners are slightly less risk averse ( 75.h ).

Debtors have a higher discount factor ( 75.0c ) but log utility like savers.

5. Simulations

5.1 Baseline (no government intervention scenario)

1) Risk Premium Shock on housing investment: bursting of the housing bubble.

2) Tightening of lending conditions.

3) Default Shock on matured loans.

EU: EUR 500bn, (around 4% of EU GDP).

RoW EUR 900bn, (around 2.5% of GDP).

4) Panic: perceived losses (1 year ahead)

Table 4: Crisis scenario : no government intervention 2008 2009 2010 2011 2012 2020 GDP -2.3 -8.17 -6.89 -3.9 -3.04 -4.89 Consumption -1.91 -4.9 -4.44 -3.18 -3.28 -6.18 Corp. investment -12.96 -48.82 -38.21 -15.39 -4.93 -2.55 Res. Investment -3.22 -9.12 -9.54 -10.03 -10.45 -9.53 Real wages -0.76 -4.23 -5.05 -3.74 -2.28 -0.23 Employment -1.85 -6.8 -5.15 -2.17 -1.35 -3.71 Value of banks -34.88 -61.62 -39.3 -21.28 -14.45 -12.09 Real interest rate (5y) (bps) 286.32 378.88 312.9 77.6 7.1 -17.78 Labour income tax (pp) 0.19 1.55 2.67 3.15 3.47 7.25 Gov. debt (% of GDP) 2.12 10.76 13.36 12.3 12.33 22.15

Note: % deviations from baseline values, or basispoints (bps). Table 3 : Stylised facts: EU27 2008 2009 2010 GDP growth 0.5 -4.2 1.8 Consumption growth 0.7 -1.7 0.8 Corp.investment growth 2.3 -20.0 4.3 Res. investment growth 1.2 -9.3 -5.2 Employment growth Debt/GDP

0.9 62.3

-1.9 74.4

-0.5 80.2

5.2 Bank rescue measures Asset purchases Governments buy assets (loans) from banks and take over a share of losses associated

with these loans.

By partially taking over bank losses, the government can effectively smoothen the

dividend stream of corporate banks, provide more consumption smoothing and

consequently a smaller increase in the equity premium.

In contrast to standard fiscal measures, which tend to crowd out private investment, these

state aid measures support corporate investment.

In terms of effectiveness, the fiscal multiplier of state support in the form of asset

purchases is positive but well below one. Total asset purchases amounting to roughly

2.8% of GDP boost GDP by around 1%.

Table 6: Government intervention: asset purchases

2008 2009 2010 2011 2012 2020 GDP 0.27 0.97 0.43 -0.36 -0.31 0.03 Consumption 0.26 0.6 0.31 -0.19 -0.14 0.15 Corp. investment 2.44 12.29 5.83 -1.16 -1.33 -0.23 Res. Investment 0.06 0.11 0.11 -0.01 -0.09 0.05 Real wages 0.11 0.68 0.81 0.46 0.19 0.04 Employment 0.21 0.73 0.16 -0.53 -0.47 -0.02 Value of banks 8.9 11.47 2.38 -1.17 -0.54 0.22 Real interest rate (5y) (bps) -38.33 -63.06 -36.64 12.05 9.7 1.3 Labour income tax (pp) -0.02 0.56 0.86 0.7 0.55 0.01 Gov. debt (% of GDP) -0.24 1.35 2.69 2.55 1.89 -0.35

Note: % difference from no-intervention

Recapitalisations

Government recapitalisation measures have similar effects compared to asset

purchases (Table 7).

Both measures have different distributional consequences:

The fiscal costs of asset purchases will be shared equally across all household types,

while the purchase of bank stock and the subsequent sale shifts the burden of the

fiscal costs more onto equity owners and increases the equity premium and reduces

investment.

In terms of effectiveness, a "stimulus" in the form of recapitalisations of around 2.2 %

of GDP at its peak, give a positive GDP effect of 1.7%.

Table 7: Government intervention: recapitalisations

2008 2009 2010 2011 2012 2020 GDP 0.49 1.7 0.45 -0.13 -0.08 0.24 Consumption 0.33 1.01 0.4 0.05 0.13 0.31 Corp. investment 3.33 14.23 3.01 -1.79 -1.99 -0.01 Res. Investment 0.19 0.9 1.03 0.78 0.67 0.12 Real wages 0.13 0.66 0.61 0.25 0.03 -0.09 Employment 0.4 1.38 0.27 -0.28 -0.2 0.24 Value of banks 1.64 3.49 0.56 -0.53 -0.42 0.11 Real interest rate (5y) (bps) -60.76 -89.58 -20.01 12.31 11.52 2.42 Labour income tax (pp) -0.03 0.1 0.17 0.13 0.03 -0.4 Gov. debt (% of GDP) -0.41 -0.79 0.31 0.57 0.18 -1.34

Note: % difference from no-intervention

Conventional fiscal stimulus measures

Table 5 Fiscal multipliers of conventional stimulus measures EU

Without

collateral constraint

s.

With

collateral constraint

s .

With collateral constraint

s and

monetary accommo

-dation government purchases 0.78 0.81 1.03

general transfers 0.20 0.41 0.53

transfers targetted to collateral-constrained hh.

- 0.67 0.86

labour tax 0.22 0.44 0.55

Source: Roeger and in 't Veld (2010)

Government guarantees:

We restrict ourselves to analysing the effect of government guarantees in a

segmented financial market.

Essentially the value added of government guarantees in such an environment

consists of redistributing losses from a fraction of households (shareholders,

households owning risky assets) to all households (and thereby effectively removing

the market segmentation).

This has a macroeconomic benefit in terms of reducing the increase of the bond rate,

but it has also costs because the government support has to be financed by

distortionary taxes.

Given that EU governments have guaranteed bonds in the order of magnitude of 8%

of GDP, we create a default scenario where the financial sector expects loan losses

to accumulate to 8% of GDP

We compare two extreme cases.

1) No government guarantees are given,

2) Government guarantees to take over all future losses.

As can be seen from these tables, the government guarantees can prevent economic

activity from collapsing in the first two years.

However, these guarantees also have negative effects as higher labour taxes have a

negative impact on employment and corporate investment in the medium term.

Table 8: No Government Guarantees 2009 2010 2011 2012 2020 GDP -6.93 -3.46 -0.68 -0.84 -1.2 Consumption -6.25 -2.93 -0.73 -1.09 -1.5 Corp. investment -31.56 -15.96 -1.8 -0.61 -1.26 Res. Investment -2.16 -2.27 -1.16 -0.98 -0.78 Real wages -1.83 -2.58 -1.34 -0.58 -0.32 Employment -5.55 -2.73 -0.01 -0.14 -0.62 Value of banks -31.31 -8.28 -1.18 -1.43 -1.61 Real interest rate (5y) (bps) 503.62 144.2 -22.74 -5.69 2.38 Labour income tax (pp) 0.71 1.38 1.29 1.23 1.53 Gov. debt (% of GDP) 6.95 6.6 4.25 3.99 4.1

Note: % difference from no-defaults baseline

Table 9.: Defaults with government guarantees 2009 2010 2011 2012 2020 GDP -0.35 -0.7 -0.8 -0.82 -0.7 Consumption -0.46 -0.85 -0.94 -0.96 -0.85 Corp. investment -0.5 -1.11 -1.34 -1.36 -0.97 Res. Investment -0.2 -0.53 -0.6 -0.58 -0.37 Real wages 0.14 0.3 0.33 0.29 0.01 Employment -0.37 -0.73 -0.83 -0.83 -0.56 Value of banks -0.81 -0.97 -1.02 -1.02 -0.9 Real interest rate (5y) (bps) -4.07 0.32 2.14 2.33 1.98 Labour income tax (pp) 1.01 1.97 2.05 1.99 1.35 Gov. debt (% of GDP) 3.56 6.83 6.96 6.54 3.01

Note: % deviation from no default baseline As can be seen from these tables, the government guarantees can prevent economic

activity from collapsing in the first two years.

However, these guarantees also have negative effects as higher labour taxes have a

negative impact on employment and corporate investment in the medium term.

6. Conclusions

This paper has assessed the cost and benefits of state aid to the financial system in

an economy which is hit by a severe financial shock and is subject to financial market

imperfections.

Our analysis has shown that state interventions can stabilise the economy.

Multipliers are lower than those for government consumption, but generally larger

than those for transfers to households.

State support to the banking sector has helped to stabilise corporate investment,

which is the component of aggregate demand most severely affected from the

financial shock when there are financial frictions.

This feature also distinguishes state aid from conventional fiscal interventions (like an

increase in government spending or transfers), which primarily target non-investment

demand categories and rather crowd out private capital formation.

While asset purchases and recapitalisations are effective in the case of actual losses,

government guarantees play an important role in stabilising pessimistic financial

markets driven by excessively strong loss expectations.