OPTIMAL REGULATION OF BANK CAPITAL AND

LIQUIDITY

E Philip DavisNIESR and Brunel University

West Londone_philip_davis@msn.com

www.ephilipdavis.comgroups.yahoo.com/group/financial_stability

Course on Financial Instability at the Estonian Central Bank,9-11 December 2009 – Lecture 8

Abstract• Raising capital adequacy standards and introducing

binding liquidity requirements can have beneficial effects if they reduce the probability of a costly financial crisis, but may also reduce GDP by raising borrowing costs for households and companies.

• We estimate both benefits and costs of raising capital and liquidity, with the benefits being in terms of reduction in the probability of banking crises, while the costs are defined in terms of the economic impact of higher spreads for bank customers.

• Result shows a positive net benefit from substantial regulatory tightening, depending on underlying assumptions.

Structure

1. Introduction

2. Literature survey

3. Could we predict the crisis?

4. Impact of the crisis

5. Modelling the banking sector

6. Costs and benefits of tighter regulation

7. Conclusion

1 Introduction• NIESR research for FSA - background is recent

crisis and discussion of regulatory changes at global level – authors Ray Barrell, E Philip Davis, Tatiana Fic, Dawn Holland, Simon Kirby and Iana Liadze

• Review literature on bank behaviour relative to regulation of capital and liquidity

• Then assess whether a link of capital and liquidity to crisis probabilities can be traced

• Consider long run cost of crises (scarring)• Model behaviour of banks, impact on economy• Consider over what range there are positive net

benefits to regulatory tightening

2 Literature survey

• Much theoretical work abstracts from regulation• Most relevant is on bank capital buffers…• …and UK empirical work showing regulation (trigger

ratios) is main determinant of capital held – and also impacts loan supply

• Overall procyclicality of capital and balance sheet, and impact of introduction of Basel in early 1990s (credit crunch)

• Cross country work showing impact of capitalisation on margins

• Controversy regarding relevance of MM• Developing literature on crisis probabilities

UK bank capital adequacy

6.00

7.00

8.00

9.00

10.00

11.00

12.00

13.00

14.00

De

c-8

9

De

c-9

0

De

c-9

1

De

c-9

2

De

c-9

3

De

c-9

4

De

c-9

5

De

c-9

6

De

c-9

7

De

c-9

8

De

c-9

9

De

c-0

0

De

c-0

1

De

c-0

2

De

c-0

3

De

c-0

4

De

c-0

5

De

c-0

6

De

c-0

7

Perc

en

t o

f to

tal

as

se

ts

trigger ratio risk weighted capital ratio

3 Predicting crises

• Most work on predicting crises, such as Demirguc Kunt and Detragiache (1998) uses logit and estimates across global sample of crises (mainly in emerging market economies)

• Typically crises found to be correlated with macroeconomic, banking sector and institutional indicators:– Low GDP growth, high interest rates, high inflation,

fiscal deficits. – Ratio of broad money to Foreign Exchange reserves,

credit to the private sector/GDP ratio, lagged credit growth

– Low GDP per capita and deposit insurance.

Could they predict the crisis?

• Traditional crisis prediction models did not pick up the risks that were developing (Davis and Karim 2008)– The build up of debt was worrying– The house price bubble was a concern– The regulatory architecture was flawed– The dangers of securitisation were not seen

• None of these were under the control of the monetary authorities

• A crisis means credit rationing (Barrell et al 2006 Journal of Financial Stability)

Explaining Crises

log

p(crisis)-1

p(crisis) = - 0.333 LEV(-1) – 0.118 LIQ(-1) + 0.113 RHPG(-3) (3)

(-2.85) (-3.55) (2.8)

Probabilities of a crisis

FR 0.066 0.043 0.028 0.018GE 0.007 0.005 0.003 0.002IT 0.035 0.023 0.014 0.009SP 0.056 0.024 0.023 0.015UK 0.217 0.150 0.101 0.067US 0.006 0.004 0.003 0.002

Variables are taken with lag length (based on the model)

Initial probability of the cirsis

in 2007

After 1 percentage

point increase in LIQ and

LEV

After 2 percentage

point increase in LIQ and

LEV

After 3 percentage

point increase in LIQ and

LEV

Comments on results

• We use unadjusted capital adequacy due to data limitations over 1980-2006 sample

• Changes in capital alone twice as effective as liquidity according to estimate

• Results show substantial reduction in crisis probabilities from regulatory tightening in Europe using OECD logit

• Weaker effect in US, which we consider relates to omission of off balance sheet activity – work under way to rectify this

4 Impact of the crisis

• Seek to assess effect of current crisis on UK economy, as a benchmark for benefits of regulation

• Key background is approach of Hoggarth and Sapporta (2001) who saw costs of crisis as integral of output lost below previous trend

• Highlight that initial recession is only part of costs given the possible long run effect on capital stock (scarring)

Assessing long run costs of crises

Qt = γ (δKt-ρ + (1-δ)(Lt e techl

t) -ρ ) -1/ρ

Log (Kt /Qt) = a1 + σ Log (usert)

Output depends on the supply side as described by the production function

The equilibrium capital output ratio depends on the user cost of capital

The user cost is driven by weighted average cost of capital, linked in turn to risk free long real rates (lrr) and by the borrowing margin charged by banks (corpw) or the bond market (iprem) to reflect costs and risks

wacct = b1t(Et/Pt) +(1-b1t) (c1t (lrrt+corpwt)

+(1- c1t) (lrrt+ipremt))*(1-ctaxrt)

Risk premia on corporate borrowing rose in 2007-9

Spread between BAA corporate and government bonds

0

1

2

3

4

5

6

7

8

9

10

05

Ma

rch

20

00

05

Ju

ne

20

00

05

Se

pte

mb

er

20

00

05

De

ce

mb

er

20

00

05

Ma

rch

20

01

05

Ju

ne

20

01

05

Se

pte

mb

er

20

01

05

De

ce

mb

er

20

01

05

Ma

rch

20

02

05

Ju

ne

20

02

05

Se

pte

mb

er

20

02

05

De

ce

mb

er

20

02

05

Ma

rch

20

03

05

Ju

ne

20

03

05

Se

pte

mb

er

20

03

05

De

ce

mb

er

20

03

05

Ma

rch

20

04

05

Ju

ne

20

04

05

Se

pte

mb

er

20

04

05

De

ce

mb

er

20

04

05

Ma

rch

20

05

05

Ju

ne

20

05

05

Se

pte

mb

er

20

05

05

De

ce

mb

er

20

05

05

Ma

rch

20

06

05

Ju

ne

20

06

05

Se

pte

mb

er

20

06

05

De

ce

mb

er

20

06

05

Ma

rch

20

07

05

Ju

ne

20

07

05

Se

pte

mb

er

20

07

05

De

ce

mb

er

20

07

05

Ma

rch

20

08

05

Ju

ne

20

08

05

Se

pte

mb

er

20

08

05

De

ce

mb

er

20

08

05

Ma

rch

20

09

05

Ju

ne

20

09

US Euro Area UK

Trend UK output and the scar• Trend output has stepped

down – Risk premia rose in two

stages before and after Lehman’s –this reduced trend 3-4%

– Banking sector gains were illusary (1-2%)

– In-migration will fall reducing trend by ¾%

• There is only a 1 in 20 chance output will be at July 2008 projections in 2018

12.4

12.5

12.6

12.7

12.8

12.9

13

2000

2002

2004

2006

2008

2010

2012

2014

2016

2018

Out

put (

log s

cale)

Output projections and scarring from

the Crisis

July 2008 forecast

95, 90 and 80% confidence bounds

around April forecast

5 Modelling the banking sector• Shown that regulation has benefits in reducing

incidence and costs of crises• May also have a negative impact on output in both

the short and the long run by increasing borrowing costs and raising user cost of capital (an effective tax on banks has real effect, e.g. widens spreads)

• So model effects of regulation on output by constructing banking sector model and embedding in NiGEM

• Use the risk weighted capital adequacy, but correlation of 0.92 to unweighted

NiGEM model• NiGEM covers the OECD economies

– Supply and demand spelled out– Trade and capital account linkages– Stock flow consistent– Long run properties as a DSGE model

• Financial and exchange markets forward looking, as is the wage bargain

• Capital stock depends on user cost and on expected output 4 years ahead

• Increasing spread between borrowing and lending rates for individuals changes their incomes, and decision making on the timing of consumption, with possibility of inducing sharp short term reductions (no long run effect).

• Changing spread between borrowing and lending rates for firms may change user cost of capital and hence equilibrium level of output and capital in the economy in a sustained manner (short and long run effects).

UK banking sector model• Banking activity modelled as a set of supply (or price) and

demand curves. • Demand depends on levels of income or activity, and on

relative prices, whilst supply, or price, depends upon the costs of providing assets and on the risks associated with those assets.

• Banking sector has four assets– secured loans to individuals for mortgages, (morth) with a

borrowing cost (rmorth), – unsecured loans to individuals for consumer credit (cc) with a

higher borrowing cost or rate of return (ccrate), – loans to corporates (corpl) with a rate or return or cost of

borrowing (lrr+corpw) where lrr is the risk free long rate and corpw is the mark up applied by banks

– liquid assets (lar). – The categories subsume, along with deposits and risk weighted

capital adequacy itself (levrr), all on-balance sheet activity within the UK.

Key equations

• Corporate spread related to capital adequacy and inverse of headroom to trigger ratio

• Household spread related to capital adequacylendw = lendw(-1) - 0.000128 - 0.446002*(lendw(-1) +0.024201 (-0.87) (-4.41) (5.83) - 0.063713*levrr(-3) (2.84) - 0.006035*arr(-4) + 0.003159*nwpi(-1)) (8.75) (-7.28)

corpw = -0.196809 + 0.131227*(log(y)-log(ycap))*100 (-0.46) (4.69) + 0.841752*invhead(-1) + 0.522302*insolr + 0.194533*levrr (5.32) (5.72) (5.60)

Model details

• Each spread feeds into borrowing cost and then credit volume, with appropriate signs

• No direct role for liquidity (effect is calibrated based on estimated results for US)

• One to one relation of bank spread to bond spread (arbitrage)

• Adjustment equations for balance sheet and capital

Tighter regulation and margins

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2009Q1 2010Q1 2011Q1 2012Q1 2013Q1 2014Q1 2015Q1 2016Q1 2017Q1 2018Q1

Ab

so

lute

ch

an

ge

in

co

rpo

rate

bo

rro

win

g c

os

ts

One percentage point rise in capital Two percentage point rise in capitalThree percentage point rise in capital

Note – changes to target capital only

Tighter regulation and output

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

2009Q1 2010Q1 2011Q1 2012Q1 2013Q1 2014Q1 2015Q1 2016Q1 2017Q1 2018Q1

Pe

rce

nt

ch

an

ge

in

GD

P

One percentage point rise in capital Two percentage point rise in capitalThree percentage point rise in capital

Note – equal changes to target and actual capital in this and following

(6) Costs and benefits of tighter regulation

• Compare the gains from tighter regulation to the costs of regulation– Gains are the change in the probability of a crisis

times lost output in this crisis

– Costs are regulatory impacts

• Look at steady state impacts in 2018

Borrowing costsImpacts on borrowing costs of increases in regulatory constraints Changes in Capital and liquidity Requirements Corporate borrowing costs (absolute change)

Year One percentage point increase

Two Percentage point increase

Three percentage point increase

Four percentage point increase

1 0.29 0.59 0.88 1.17 2 0.29 0.58 0.87 1.17 5 0.29 0.58 0.87 1.16

10 0.29 0.58 0.87 1.16 User cost of capital (percent change)

1 0.63 1.26 1.88 2.50 2 0.77 1.54 2.28 3.02 5 0.85 1.70 2.52 3.33

10 0.85 1.70 2.50 3.30 Changes in Liquidity Requirements Corporate borrowing costs (absolute change)

1 0.10 0.19 0.29 0.39 2 0.10 0.20 0.30 0.40 5 0.11 0.22 0.33 0.43

10 0.11 0.23 0.34 0.45 User cost of capital (percent change)

1 0.21 0.42 0.63 0.84 2 0.27 0.53 0.79 1.05 5 0.32 0.64 0.93 1.23

10 0.33 0.67 0.97 1.27

Long run - regulation on output

-0.9

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

pe

rce

nta

ge

po

ints

of

GD

P

one two three four five six seven

Cost benefit analysis - combined

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

1 2 3 4 5 6 7

Percentage points increase in capital adequacy and liquidity standards

ne

t p

res

en

t v

alu

e a

s a

pro

po

rtio

n o

f G

DP

7 Conclusions

• Lax regulation raises the risk of a crisis• Raising capital adequacy standards and

introducing binding liquidity requirements beneficial if they reduce probability of costly financial crisis

• Also costs - any effective banking regulation works as a tax on bank activity. Hence regulations may reduce output through impacts on borrowing costs for households and companies

• Find short run impact on consumption and short and long run impact on investment

• Estimates of impacts on costs are upper bound as structure of portfolios and of relative prices may change if regulations significantly tighten

• When capital adequacy standards tightened by 1 pcp, banks contract balance sheets by 1.2 per cent and reduce riskiness of portfolio, with their risk weighted assets falling by 1.6 per cent - results contrary to the Modigliani Miller theorem of irrelevance of the debt equity choice

• Caution needed with rapid regulatory tightening owing to immediate impact on margins

• Positive net benefit from regulatory tightening, with a 2 to 6 percentage point increase in capital and liquidity ratios increasing welfare, depending upon assumptions

• Separate tightening of capital or liquidity ratios offers benefit up to 10 percentage points

References

• Barrell R, Davis E P, Fic T, Holland D, Kirby S, Liadze I (2009), "Optimal regulation of bank capital and liquidity: how to calibrate new international standards”, FSA Occasional Paper No 38