Capital Regulation, Monetary Policy,

and Financial Stability

Pierre-Richard Agenor,a Koray Alper,b
and Luiz Pereira da Silvac
a

University of Manchester and Centre for

Growth and Business Cycle Research
b
Central Bank of Turkey
c
Central Bank of Brazil

This paper examines the roles of bank capital regulation and monetary policy in mitigating procyclicality and
promoting macroeconomic and nancial stability. The analysis is based on a dynamic stochastic model with imperfect
credit markets. Macroeconomic stability is dened in terms
of a weighted average of ination and output-gap volatility, whereas nancial stability is dened in terms of three
alternative indicators (real house prices, the credit-to-GDP
ratio, and the loan spread), both individually and in combination. Numerical experiments associated with a housing
demand shock show that in a number of cases, even if monetary policy can react strongly to ination deviations from
target, combining a credit-augmented interest rate rule and a
Basel III-type countercyclical capital regulatory rule may be
optimal for promoting overall economic stability. The greater
the degree of policy interest rate smoothing, and the stronger

We would like to thank the editor, an anonymous referee, participants at the

ECB workshop on The Bank Lending Channel in the Euro Area: New Models
and Empirical Analysis (Frankfurt, June 2425), and particularly our discussant,
Javier Suarez, as well as participants at seminars at the European University
Institute and the University of Manchester, for helpful comments on a previous draft. Financial support from the PREM Network of the World Bank is
gratefully acknowledged. A more detailed version of this article, containing the
appendices, is available upon request. We bear sole responsibility for the views
expressed here. Agenor: Professor, School of Social Sciences, University of Manchester, United Kingdom, and co-Director, Centre for Growth and Business Cycle
Research; Alper: Director, Open Market Operations, Central Bank of Turkey;
Pereira da Silva: Deputy Governor, Central Bank of Brazil.

193

194

International Journal of Central Banking

September 2013

the policymakers concern with nancial stability, the larger is

the sensitivity of the regulatory rule to credit growth gaps.
JEL Codes: E44, E51, F41.

Preserving nancial stability is so closely related to the standard goals of monetary policy (stabilizing output and ination)
that it . . . seems somewhere between foolish and impossible to
separate the two functions.
Alan S. Blinder,
How Central Should the Central Bank Be? (2010, p. 12)
Ensuring nancial stability requires a redesign of macroeconomic as well as regulatory and supervisory policies with an
eye to mitigating systemic risks. For macroeconomic policies,
this means leaning against credit and asset price booms; for
regulatory and supervisory policies, it means adopting a macroprudential perspective.
Bank for International Settlements,
79th Annual Report (2009, p. 14)
1.

Introduction

The recent crisis in global nancial markets has led to a substantial number of proposals aimed at strengthening the nancial system
and at encouraging more prudent lending behavior in upturns. Many
of these proposals aim to mitigate the alleged procyclical eects of
Basel II capital standards. Indeed, several observers have argued
that by raising capital requirements in a contracyclical way, regulators could help to slow credit growth and choke o asset-price
pressures before a crisis occurs. A recent proposal along these lines,
put forward by Goodhart and Persaud (2008), involves essentially
adjusting the Basel II capital requirements to take into account and
act at the relevant point in the economic cycle.1 In particular, in the
1
Buiter (2008) extended the Goodhart-Persaud proposal by suggesting that
capital and liquidity requirements be applied to all highly leveraged nancial
institutions, not only banks.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

195

Goodhart-Persaud proposal, the capital adequacy requirement on

mortgage loans would be linked to the rise in both mortgage lending and house prices.2 The Turner Review (See Financial Services
Authority 2009) also favors countercyclical capital requirements, and
so do Brunnermeier et al. (2009), who propose to adjust capital adequacy requirements over the cycle by two multiplesthe rst related
to above-average growth of credit expansion and leverage, the second related to the mismatch in the maturity of assets and liabilities.3
At the policy level, there has been concrete progress toward establishing new standards in this area; the Basel Committee on Banking Supervision (BCBS) has developed a countercyclical framework
that involves adjusting bank capital in response to excess growth in
credit to the private sector, which it views as a good indicator of
systemic risk. On November 12, 2010, G20 leaders adopted BCBSs
proposal to implement a countercyclical capital buer ranging from
0 to 2.5 percent of risk-weighted assets, as part of the new Basel III
framework (see Basel Committee on Banking Supervision 2010).
At the same time, the global nancial crisis has led to renewed
calls for central banks (and regulators) to consider more systematically potential trade-os between the objectives of macroeconomic
stability and nancial stability, or equivalently whether the central
banks policy loss function (and therefore its interest rate response)
should account explicitly for a nancial stability goal. The issue is
not new; it has long been recognized, for instance, that an increase
in interest rates aimed at preventing the development of inationary pressures may, at the same time, heighten uncertainty and foster volatility in nancial markets. The debate (which predates the
recent crisis) has focused on the extent to which monetary policy
should respond to perceived misalignments in asset prices, such as

Goodhart and Persaud argue that their proposal could be introduced under
the second pillar of Basel II. While Pillar 1 consists of rules for requiring minimum capital against credit, operational, and market risks, Pillar 2 is supposed to
take into account all the additional risks to which a bank is exposed, in order to
arrive at its actual capital needs. However, by using only Pillar 2 at the discretion
of local regulators, it can allow banks to engage in regulatory arbitrage.
3
Although not as explicit, Blinder (2010) has also endorsed the view that
central banks should try to limit credit-based bubbles through regulatory instruments (rather than interest rates) and refers to it as possibly becoming the new
consensus on how to deal with asset-price bubbles.

196

International Journal of Central Banking

September 2013

real estate and equity prices.4 In that context, several observers have
argued that trying to stabilize asset prices per se is problematic for
a number of reasonsone of which being that it is almost impossible to know for sure whether a given change in asset values results
from changes in underlying fundamentals, non-fundamental factors,
or both. By focusing on the implications of asset-price movements
for credit growth and aggregate demand, the central bank may be
able to focus on the adverse eects of these movementswithout getting into the tricky issue of deciding to what extent they represent
changes in fundamentals.
This paper is an attempt to address both sets of issues in a unied framework. We examine the role of both monetary policy and
a capital regulation rule that bears close similarity to some recent
proposals to mitigate the procyclical tendencies of nancial systems,
and we evaluate their implications for macroeconomic stability and
nancial stabilitydened in terms of the combined volatility of
ination and the output gap, on the one hand, and the volatility
of a measure of potential nancial stress, on the other. We do so
under a Basel II-type regime, with endogenous risk weights on bank
assets. Among the issues that we attempt to address are the following: to promote nancial stability, how should countercyclical
bank capital requirement rules be designed? Instead of adding a
cyclical component to prudential regulation, shouldnt policymakers
use monetary policy to constrain credit growth directly? To what
extent should regulatory policy and monetary policy be combined to
ensure both macroeconomic and nancial stability? Put dierently,
are these policies complementary or substitutes?
Quantitative studies of these issues are important for a number of reasons. Regarding the design of countercyclical bank capital
rules, several observers have noted that there are indeed signicant
potential practical problems associated with their implementation
including the period over which relevant nancial indicators (credit
growth rates, for that matter) should be calculated. More important perhaps is the possibility that these rules may operate in
counterintuitive ways, depending on the degree of nancial-sector
4
See, for instance, Chadha, Sarno, and Valente (2004), Filardo (2004), Akram,
Bardsen, and Eitrheim (2006), Faia and Monacelli (2007), and Akram and
Eitrheim (2008). Wadhwani (2008) provides a brief overview of the literature.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

197

imperfections. In particular, in countries where bank credit plays a

critical role in nancing short-term economic activity (as is often
the case in developing economies), a rule that constrains the growth
in overall credit could entail a welfare cost. At the same time, of
course, to the extent that it succeeds in reducing nancial volatility
and the risk of a full-blown crisis, it may also enhance welfare. The
net benets of countercyclical bank capital rules may therefore be
ambiguous in general and numerical evaluations become essential.
Regarding the role of monetary policy, the key issue is whether
a central bank with a preference for output and price stability can
improve its performance with respect not only to these two objectives but also to nancial stability, by responding to excessive movements in credit and/or asset prices in addition to uctuations in
prices and activity. In a relatively complex model, understanding
the conditions that may lead to trade-os among objectives often
requires quantitative experiments.
To conduct our analysis, we extend the New Keynesian model
described in Agenor, Alper, and Pereira da Silva (2012). Important
features of that model are that it accounts explicitly for a variety of
credit market imperfections and bank capital regulation.5 A housing
sector is introduced, and the role of real estate as collateral examined. Specically, we establish a direct link between house prices and
credit growth via their impact on collateral values and interest rate
spreads on loans: higher house prices enable producers to borrow
and invest more, by raising the value of the collateral that they can
pledge and improving the terms at which credit is extended. This
mechanism is consistent with the evidence suggesting that a large
value of bank loans to (small) rms, in both industrial and developing countries, is often secured by real estate. To capture nancial
5
There is a small but growing literature on the introduction of capital regulation in New Keynesian models with banking; recent papers include Aguiar
and Drumond (2007), Dib (2009, 2010), Hirakata, Sudo, and Ueda (2009), Gerali
et al. (2010), and Meh and Moran (2010). Some contributions have also attempted
to integrate countercyclical regulatory rules in this type of model; they include
Kannan, Scott, and Rabanal (2009), Levieuge (2009), Angeloni and Faia (2010),
Covas and Fujita (2010), Darracq Pari`es, Kok Sorensen, and Rodriguez Palenzuela (2010), and Angelini, Neri, and Panetta (2011). However, as is made clear
later, our approach diers signicantly from all of these contributions, making
comparisons dicult.

198

International Journal of Central Banking

September 2013

instability, we consider three alternative indicators, both individually and in combination: real house prices, the credit-to-GDP ratio,
and the loan spread.6 This is also in line with the literature suggesting that nancial crises are often preceded by unsustainable developments in the real-estate sector and private-sector credit, and a large
increase in bank interest rate spreads.7
In this context we examine the implications of two alternative
policy rules for economic stability: a standard Taylor-type interest
rate rule augmented to account for credit growth and (in line with
the Basel III regime) a countercyclical regulatory rule that relates
capital requirements also to credit growth. Our numerical experiments show that even if monetary policy can react strongly to ination deviations from target, combining a credit-augmented interest
rate rule and a Basel III-type countercyclical capital regulatory rule
may be optimal for promoting overall economic stability. The greater
the degree of interest rate smoothing, and the stronger the policymakers concern with nancial stability (when measured in terms
of either the credit-to-GDP ratio and/or loan spread volatility), the
larger is the sensitivity of the regulatory rule to credit growth gaps.
The paper continues as follows. Section 2 presents the model. We
keep the presentation very brief, given that many of its ingredients
are described at length in Agenor, Alper, and Pereira da Silva (2012);
instead, we focus on how the model presented here departs from
that paper, especially with respect to the nancial sector and countercyclical policy rules. The equilibrium is characterized in section
3. Key features of the steady state and the log-linearized version, as
well as a brief discussion of an illustrative calibration, are discussed
in section 4. We present in section 5 the impulse response functions
associated with our base experiment: a temporary, positive housing
demand shock. Section 6 discusses the two alternative countercyclical rules alluded to earlier, involving an augmented monetary policy
rule and a capital regulatory rule, both dened in terms of deviations
6

The concept of nancial stability has remained surprisingly elusive in the

existing literature; see Goodhart (2006). Our focus in this paper is on alternative,
operational, quantitative measures of nancial instability.
7
See Calder
on and Fuentes (2011) and International Monetary Fund (2011).
Gerdesmeier, Reimers, and Roa (2010) found that credit aggregates also play
a signicant role in predicting asset-price busts in industrial countries.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

199

in credit growth and aimed at promoting stability. Section 7 investigates whether the use of these rules (taken in isolation) generates
gains in terms of both nancial and macroeconomic stability, that
is, whether they entail a trade-o among objectives; to do so we
present simulation results of the same housing demand shock under
both types of rules, for some specic parameter values. Section 8 discusses optimal policy rules, when the objective of the central bank is
to minimize volatility in a measure of economic stability, dened as
a weighted average of separate measures of macroeconomic stability
and nancial stability. Section 9 provides some sensitivity analysis.
The last section provides a summary of the main results and discusses the implications of our analysis for the ongoing debate on
reforming bank capital standards.
2.

Outline of the Model

The core model presented in this paper departs from Agenor, Alper,
and Pereira da Silva (2012) essentially by introducing a housing market and linking it with collateral and loans for investment purposes.
To save space, this section provides only a brief outline of the model
in most respectsexcept for bank regulation and the optimization
problem of the bank, for which a more formal analysis is presented.8
We consider a closed economy populated by six types of agents:
innitely lived households, intermediate good (IG) producers, a nal
good (FG) producer, a capital good (CG) producer, a monopoly
commercial bank, the government, and the central bank, whose mandate also includes bank regulation. The nal good is homogeneous
and can be used either for consumption or investment, although in
the latter case additional costs must be incurred.
There are two types of households, constrained and unconstrained.9 Constrained households do not participate in asset markets and follow a rule of thumb which involves consuming all their
8

A detailed, formal presentation of the model is available in the working paper

version of this article, which is available upon request.
9
As discussed later, the distinction between these two types of households
is useful to understand the dynamics of consumption following housing demand
shocks. See Agenor and Montiel (2008) for a discussion of why consumption
smoothing may be imperfect in the context of developing countries.

200

International Journal of Central Banking

September 2013

after-tax disposable wage income in each period. They also supply labor inelastically. Unconstrained households consume, can trade
in asset markets and hold nancial assets (including nominal debt
issued by the bank), and supply labor to IG producers. As in
Iacoviello (2005), Silos (2007), and Iacoviello and Neri (2008), housing services are assumed to be proportional to their stock, which
enters directly in the utility function. These households also make
their housing stock available, free of charge, to the CG producer, who
uses it as collateral against which it borrows from the bank to buy
the nal good and produce capital. At the beginning of the period,
unconstrained households choose the real levels of cash, deposits,
bank debt, government bonds, and labor supply to IG rms. They
receive all the prots made by the IG producers, the CG producer,
and the bank, and pay a lump-sum tax.
Optimization yields a standard Euler equation and a standard
labor supply function, which relates hours worked positively to the
real wage and negatively to consumption. It also yields three asset
demand equations: the rst relates the real demand for cash positively to consumption and negatively to the opportunity cost of
holding money, measured by the interest rate on government bonds;
the second relates the real demand for deposits positively to consumption and the deposit rate, and negatively to the bond rate; and
the third is the demand for bank debt, Vtd , which is given by
Vtd
= 1
V
Pt

iVt iB
t
1 + iB
t


,

(1)

where V 0 denotes an adjustment cost parameter that captures

transactions costs associated with changes in household holdings of
bank capital, Pt the price of the nal good, iB
t the bond rate, and
iVt the rate of return on bank debt. Thus, the demand for bank debt
depends positively on its rate of return and negatively on the bond
rate. Similarly, there is a demand function for housing services, from
which it can be established that a positive shock to preferences for
housing leads (all else equal) to a rise in todays demand for housing.
The FG producers optimization problem is specied in standard fashion. The nal good, which is allocated to private consumption, government consumption, and investment, is produced
by assembling a continuum of imperfectly substitutable intermediate

Vol. 9 No. 3

Capital Regulation, Monetary Policy

201

goods. The FG producer sells its output at a perfectly competitive

price. Given the intermediate goods prices and the nal good price,
it chooses the quantities of intermediate goods that maximize its
prots.
IG producers produce, using capital and labor, a distinct perishable good that is sold on a monopolistically competitive market.
At the beginning of the period, each IG producer rents capital from
the CG producer and borrows to pay wages in advance. Loans contracted for the purpose of nancing working capital (which are short
term in nature) do not carry any risk, and are therefore made at a
rate that reects only the marginal cost of borrowing from the central bank, iR
t , which we refer to as the renance rate. These loans
are repaid at the end of the period. IG producers solve a two-stage
problem. In the rst stage, taking input prices as given, they rent
labor and capital in perfectly competitive factor markets so as to
minimize real costs. This yields the optimal capital-labor ratio. In
the second stage, each IG producer chooses a sequence of prices so
as to maximize discounted real prots, subject to adjustment costs
`a la Rotemberg (1982). The solution gives the adjustment process
of the nominal price.
The CG producer owns all the capital in the economy and uses a
linear technology to produce capital goods. At the beginning of the
period, it buys the nal good from the FG producer. These goods
must be paid in advance; to purchase nal goods, the CG producer
must borrow from the bank. At the end of the period, loans are paid
in full with interest. Thus, the total cost of buying nal goods for
investment purposes includes the lending rate, iL
t . The CG producer
combines investment goods and the existing capital stock to produce
new capital goods, subject to adjustment costs. The new capital
stock is then rented to IG producers. The CG producer chooses the
level of the capital stock (taking the rental rate, the lending rate,
and the price of the nal good as given) so as to maximize the value
of the discounted stream of dividend payments (nominal prots)
to the household. The rst-order condition for maximization shows
that the expected rental rate of capital is a function of the current
and expected loan rates, the latter through its eect on adjustment
costs in the next period.
The commercial bank (which is owned by unconstrained households) also supplies credit to IG producers, who use it to nance their

202

International Journal of Central Banking

September 2013

short-term working capital needs. Its supply of loans is perfectly elastic at the prevailing lending rate. To satisfy capital regulations, it
issues nominal debt at the beginning of time t, once the level of
(risky) loans is known.10 It pays interest on household deposits (at
rate iD
t ), the liquidity that it borrows from the central bank (at rate
iR
),
and
its debt. The maturity period of both categories of bank
t
loans and the maturity period of bank deposits by unconstrained
households is the same. In each period, loans are extended prior to
activity (production or investment) and paid o at the end of the
period. At the end of each period, the bank is liquidated and a new
bank opens at the beginning of the next; thus, all its prots are
distributed, bank debt is redeemed, and new debt is issued at the
beginning of the next period to comply with prudential regulatory
rules.
Formally, and abstracting from required reserves and holdings of
government bonds, the banks balance sheet is
B
LF
t = Dt + Vt + Lt ,

(2)

B
where Dt is household deposits, LF
t total loans, Lt borrowing from
the central bank (which covers any shortfall in resources), and Vt
total capital held by the bank, given by

Vt = VtR + VtE ,

(3)

with VtR denoting capital requirements and VtE excess capital.

Given that LF
t and Dt are determined by private agents behavior, the balance sheet constraint (2) can be used to determine borrowing from the central bank:
F
LB
t = Lt Dt Vt .

(4)

10
Thus, capital consists therefore, in the Basel terminology, solely of supplementary or tier 2 capital; there is no core or tier 1 capital, that is, ordinary
shares. In practice, to meet capital requirements, banks have often issued hybrid
securities that are more like debt than equity. Data from the International Monetary Fund show that at the end of 2008 the average ratio of equity made up of
issued ordinary shares to assets was only 2.5 percent for European banks and 3.7
percent for U.S. banks. However, under the new Basel III regime, the denition
of capital in terms of common equity has been considerably tightened.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

203

The bank is subject to risk-based capital requirements, imposed

by the central bank. It must hold an amount of capital that covers
an endogenous percentage of its risky loans.11 Loans for working
capital need bear no risk and are subject to a zero weight in calculating capital requirements. Thus, with tF denoting the risk weight
on loans to the CG producer, capital requirements are given by
VtR = t tF LF,I
t ,

(5)

where t (0, 1) is the overall capital ratio, dened later, and LF,I
t
is loans for investment. As in Agenor and Pereira da Silva (2012),
and in line with the foundation variant of the internal ratingsbased (IRB) approach of Basel II, we relate the risk weight to the
repayment probability of the CG producer estimated by the bank,
qtF (0, 1), because it reects its perception of default risk:12

tF

qtF
qF

q
,

(6)

where q > 0 and qF is the steady-state value of qtF . In the steady

state, the risk weight is therefore normalized to unity.13
The bank is risk neutral and chooses both the deposit and lending
rates, and excess capital, so as to maximize the present discounted
11
Because the bank is liquidated at the end of each period, it does not accumulate capital through retained earnings. This is in contrast to Angelini, Neri,
and Panetta (2011), for instance.
12
Under the IRB approach, the estimated credit riskand the associated risk
weightsare assumed to be a function of four parameters: the probability of
default (PD), the loss given default, the exposure at default, and the loans maturity. Banks adopting the advanced variant of this approach must provide all four
of these parameters from their own internal ratings models; those adopting the
foundation variant provide only the PD parameter, with the other three parameters set externally by regulators. Appendix C in Agenor, Alper, and Pereira
da Silva (2012) provides a justication for the reduced-form, constant elasticity
specication shown in (6), based on Basel II formulas. See also Angeloni and
Faia (2010), Covas and Fujita (2010), and Darracq Pari`es, Kok Sorensen, and
Rodriguez Palenzuela (2010).
13
The standardized approach in Basel II can be modeled by making the risk
weight a function of the output gap, under the assumption that ratings are procyclical, in a manner similar to Zicchino (2006) and Angeloni and Faia (2010), for
instance. See Drumond (2009) and Panetta and Angelini (2009) for a discussion
of the evidence on this issue.

204

International Journal of Central Banking

September 2013

value of its real prots.14 Because the bank is liquidated and debt is
redeemed at the end of each period, this optimization program boils
down to a period-by-period maximization problem, subject to several constraintsthe loan demand function from the CG producer,
total credit, the balance sheet constraint (4), the denition of total
capital (3), and the capital requirement constraint (5). In addition,
the bank internalizes the fact that the demand for loans by the CG
producer (supply of deposits by unconstrained households) depends
negatively (positively) on the lending (deposit) rate, and takes the
repayment probability of the CG producer, the value of collateral,
capital requirements, prices, and the renance rate as given.
Expected, end-of-period t prots in real terms, which are distributed at the beginning of period t + 1, are given by




F,W
F,I
L
L
t
t

(1 + iR
+ qtF (1 + iL
+ (1 qtF )pH
t )
t )
t H + dt
Pt
Pt
 E E
 B
 
Vt
Lt
Vt
D
R
V
(1 + it )
+ 2V V
(1 + it )dt (1 + it )
,
Pt
Pt
Pt
(7)
is the exogenous supwhere (0, 1), V V 0, E (0, 1), and H
15
ply of housing. The second term in this expression on the right1 F,I
hand side, qtF (1 + iL
t )Pt Lt , represents expected repayment on
loans to the CG producer if there is no default. The third term represents what the bank expects to earn in case of default, that is,
eective collateral, given by a fraction (0, 1) of raw collateral, that is, the housing stock. Coecient can be viewed as a
measure of eciency of enforcement of debt contracts (see Djankov
14
To simplify matters, we solve only for the loan rate applicable to the CG producer. In principle, even if loans to IG producers carry no risk and are extended
at the marginal cost of funds (the renance rate), it should be assumed that the
bank also determines it as part of its optimization problemin which case the
elasticity of the demand for working capital loans would aect the markup over
the renance rate. For simplicity, we have assumed directly that the cost of these
loans is only iR
t .
15
Housing supply could be endogenized by adding a construction sector to the
model. This would reduce the volatility of housing prices by allowing the housing
stock to respond to demand shocks. However, given the time frame of the model,
the assumption of exogenous supply is quite reasonable.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

205

et al. 2008) or an inverse measure of anti-creditor bias in the judicial

system (see Cavalcanti 2010). Note also that collateral is marked to
market, a practice that has become prevalent in recent years and
tends to magnify procyclicality in leverage.
The fourth term, dt , represents the reserve requirements held at
the central bank and returned to the bank at the end of the period
(prior to its closure). The term (1 + iD
t )dt represents repayment
of deposits (principal and interest) by the bank, whereas the term
1 B
(1+iR
t )Pt Lt represents gross repayments to the central bank. The
term (1 + iVt )Vt represents the value of bank debt redeemed at the
end of the period plus interest.16 The last term, 2V V (VtE /Pt )E ,
captures the view that maintaining a positive capital buer generates some benetsit represents a signal that the banks nancial
position is strong and reduces the intensity of regulatory scrutiny
(or degree of intrusiveness in the banks operations), which in turn
reduces the pecuniary cost associated with providing the information required by the supervision authority; the restriction E < 1
ensures a sensible solution (see Agenor, Alper, and Pereira da Silva
2012).17
The rst-order conditions for maximization give

iD
t

1 + iL
t =

1
1+
D

1
iR
t ,

V
(1 t t )(1 + iR
t ) + t t (1 + it )
,
(1 + F1 )qtF

(8)
(9)

16
In the full version of the model presented in the working paper version of this
article, we add a linear term in Vt , to capture the cost associated with issuing
shares, which includes the cost of underwriting, issuing brochures, etc. The cost
of issuing equity could of course be dened in a more general way, to account
for the fact that (i) a positive capital buer provides a signal to markets that
the bank is healthy, and (ii) in recessions (expansions), market funding is more
dicult (easier) to obtain.
17
In Dib (2009, 2010), holding bank capital in excess of the required level also
generates gains. An alternative approach, which has yet to be implemented (as
far as we know) in New Keynesian models of the type discussed here, would be to
introduce a precautionary demand for excess capital along the lines proposed
by Repullo and Suarez (2009). In their framework, banks are unable to access
equity markets in every period; by holding capital buers today, they mitigate
the possibility that their ability to lend in the future may be compromised by
shocks to their earnings or aggregate economic conditions.

206

International Journal of Central Banking

VtE
=
Pt

V V
V
it iR
t

September 2013

1/E
,

(10)

where D is the interest elasticity of the supply of deposits to the

deposit rate and F is the interest elasticity of the CG demand for
loans (or investment) to the lending rate.
Equation (8) shows that the equilibrium deposit rate is a markup
over the renance rate. Equation (9) indicates that the gross lending
rate depends negatively on the repayment probability and positively
on a weighted average of the marginal cost of borrowing from the
central bank (at the gross rate 1 + iR
t ) and the cost of issuing debt.
Weights on each component of funding costs are measured in terms
of the ratio of required capital to (risky) loans, t t and 1 t t .
Equation (10) indicates that an increase in the cost of issuing debt, iVt , reduces excess capital, whereas an increase in V V
raises excess capital. With V V = 0, holding capital beyond what is
required brings no benet, so VtE = 0 for all t. Finally, an increase
in required capital, by raising the cost of issuing bank debt iVt , has
an indirect, negative eect on the desired level of excess capital.
There is therefore some degree of substitutability between the two
components of bank capital.
As in Agenor, Alper, and Pereira da Silva (2012), the repayment probability qtF is taken to depend positively on the eective
collateral-CG loan ratio (which mitigates moral hazard on the part
of borrowers), the cyclical position of the economy (as measured by
the output gap), and the banks capital-to-risky-assets ratio, which
increases incentives for the bank to screen and monitor its borrowers:
1 
2

H
V
P
H
t
t
qtF =
(ytG )3 ,
(11)
F,I
LF,I
L
t
t
with i > 0 i and ytG = Yt /Yt is the output gap, with Yt denoting the frictionless level of aggregate output.18 Our specication is
18
This semi-reduced-form approach to modeling the loan spread has been
adopted in some other contributions, such as C
urdia and Woodford (2010). In
Agenor and Pereira da Silva (2011), the repayment probability is endogenously
determined as part of the banks optimization process. Specically, they assume
that the bank can aect the repayment probability on its loans by expending

Vol. 9 No. 3

Capital Regulation, Monetary Policy

207

thus consistent with the double moral hazard framework developed in Chen (2001), Aikman and Paustian (2006), and Meh and
Moran (2010), among others, according to which banks have greater
incentives to screen and monitor borrowers when more of their capital (relative to their outstanding loans) is at stake.19 The novelty
here is that we assume explicitly that greater monitoring translates
into a higher probability of repayment.20 Indeed, equations (9) and
(11) imply a negative relationship between the capital-to-risky-assets
ratio and bank lending spreads; direct support for this link is provided by Fonseca, Gonz
alez, and Pereira da Silva (2010) in a study
of pricing behavior by more than 2,300 banks in ninety-two countries
over the period 19902007. Note also that if net worth values are procyclical, both the collateral and the output-gap eects are consistent
with the evidence suggesting that price-cost margins in banking, or
lending spreads, behave countercyclically (see, for instance, AliagaDaz and Olivero 2010 and Fonseca, Gonz
alez, and Pereira da Silva
2010, tables 6 and 9). Thus, in the model, the bank capital channel operates through two eects on the lending rate: a cost eect
(through iVt ) and a monitoring incentive eect (through qtF ).
The central banks assets consist of a xed stock of government
bonds and loans to the commercial bank, LB
t , whereas its liabilities consist of currency supplied to (unconstrained) households and
rms. Any income made by the central bank from bond holdings
and loans to the commercial bank is transferred to the government
at the end of each period. Monetary policy is operated by xing
eort to select (ex ante) and monitor (ex post) its borrowers; the higher the eort,
the safer the loan. Assuming that the cost of monitoring depends (inversely) not
only on the collateral-investment loan ratio but also on the cyclical position of
the economy and the capital-loan ratio yields a specication similar to (11).
19
However, there are signicant dierences in the way bank capital is modeled; here, bank capital is motivated by regulatory requirements rather than by a
pure moral hazard problem. Also, in Meh and Moran (2010) for instance, bank
capital consists mostly of retained earnings.
20
Mehran and Thakor (2011) and Allen, Carletti, and Marquez (2011) provide
rigorous microfoundations for the link between bank capital and monitoring. In
Allen, Carletti, and Marquez (2011), a monopoly bank holds capital because it
strengthens its monitoring incentive and increases the borrowers success probability, whereas in Mehran and Thakor (2011), bank capital increases the future
survival probability of the bank (as in Repullo and Suarez 2009), which in turn
enhances the banks monitoring incentives. The reduced-form approach that we
use can be viewed as a convenient shortcut for macroeconomic analysis.

208

International Journal of Central Banking

September 2013

the renance rate, iR

t , and providing uncollateralized loans (at the
discretion of the commercial bank) through a standing facility.21
In the baseline experiment, the renance rate is determined by a
contemporaneous, Taylor-type policy rule:
R
iR
r + t + 1 (t T ) + 2 ln ytG ] + t ,
t = it1 + (1 )[

(12)

where r is the steady-state value of the real interest rate on bonds,

T 0 the central banks ination target, (0, 1) a coecient
measuring the degree of interest rate smoothing, and 1 , 2 > 0
the relative weights on ination deviations from target and the output gap, respectively, and t is a stochastic shock, which follows a
rst-order autoregressive process.
Finally, the government purchases the nal good and issues nominal riskless one-period bonds. It collects tax revenues and all interest
income that the central bank makes from its lending to the commercial bank and its holdings of government bonds. It adjusts lump-sum
taxes to balance its budget.
Flows between agents (abstracting from the dierence between
constrained and unconstrained households) and the links between
regulatory capital, the repayment probability, and the lending rate,
are summarized in gures 1 and 2.
3.

Equilibrium

In a symmetric equilibrium, rms producing intermediate goods are

identical. Equilibrium conditions must be satised for the credit,
deposit, goods, labor, housing, bank debt, and cash markets. Because
the supply of loans by the commercial bank and the supply of
deposits by households are perfectly elastic at the prevailing interest
21
Standing facilities are now commonly used in both high- and middle-income
countries to create (narrow) corridors to bound departures of short-term moneymarket interest rates from target, with open-market operations used for the secondary objective of smoothing liquidity and moderating interest rate uctuations.
These facilities make the quantity of central bank cash endogenous by providing
unlimited accesssubject to collateral requirements and institutional rules on
who is eligible to maintain current balances with the central bankto extra cash
at the posted interest rate. For simplicity, we abstract from collateral requirements (typically low-risk and low-yield assets such as government securities) and
open-market operations, and consider a zero-width band around the target rate.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

209

Figure 1. Flows between Agents

Figure 2. Regulatory Capital, Repayment Probability,

and Lending Rate

rates, the markets for loans and deposits always clear. Equilibrium
in the goods markets requires that production be equal to aggregate
demand. The equilibrium condition of the market for bank debt is
obtained by equating (1) and (3):
Vtd = VtR + VtE .

(13)

210

4.

International Journal of Central Banking

September 2013

Steady State and Calibration

The steady state of the model is derived in appendix A (available

from the authors upon request). With an ination target T equal to
zero, the steady-state ination rate
is also zero. Beyond standard
results (on the steady-state value of the marginal cost, for instance),
the key results on the steady-state values of interest rates are as
follows (with
F = 1, by implication of (6)):
1

1
1
B
R
D
= = 1 = r, = 1 +
R ,

D
V > B ,

for V > 0

and
1 + L =

(1 )(1 + R ) + (1 + V )
,
(1 + F1 )
qF

where (0, 1) is a discount factor.

From these equations it can be shown that B > D . We also
have V > B for V > 0 because holding bank debt is subject
to a cost; from the perspective of the household, the rate of return
on that debt must therefore compensate for that and exceed the
rate of return on government bonds. Of course, when V = 0, then
V = B . From the above results, and because B > D , we also have
V > D for V > 0; bank capital is more costly than deposits (or,
equivalently, households demand a liquidity premium), as in Aguiar
and Drumond (2007). The reason here is that holding bank debt is
costly.
To analyze the response of the economy to shocks, we loglinearize the model around a non-stochastic, zero-ination steady
state. The log-linearized equations are summarized in appendix B.
A key property of the model is that deviations of real investment
from its steady-state level depend on deviations in the lending rate,
which depend themselves on deviations in the capital-to-risky-assets
ratio. Thus, regulatory policy, just like monetary policy, has a direct
eect on aggregate demand.
The calibration of the model, which we view as illustrative, dwells
on Agenor and Alper (2012) and Agenor, Alper, and Pereira da Silva
(2012) and is described in detail in appendix C. Table 1 summarizes

Vol. 9 No. 3

Capital Regulation, Monetary Policy

211

Table 1. Benchmark Calibration: Parameter Values

Parameter

Value

Household

N
x
H

0.985
0.6
1.5
0.02
0.02
0.82
0.3

Production

K
Bank

K
1

0.3

Discount Factor
Elasticity of Intertemporal Substitution
Relative Preference for Leisure
Relative Preference for Money Holdings
Relative Preference for Housing
Share Parameter in Index of Money Holdings
Adjustment Cost Parameter, Holdings of
Bank Debt
Share of Constrained Households

10.0
0.65
74.5
0.01
14

Elasticity of Demand, Intermediate Goods

Share of Labor in Output, Intermediate Good
Adjustment Cost Parameter, Prices
Depreciation Rate of Capital
Adjustment Cost Parameter, Investment

0.06
0.0
0.03

Eective Collateral-Loan Ratio

Weight of Capital Stock in Total Collateral
Elasticity of Repayment Probability with
Regard to Collateral
Elasticity of Repayment Probability with
Regard to Capital-to-Risky-Assets Ratio
Elasticity of Repayment Probability Regard to
Cyclical Output
Elasticity of the Risk Weight with Regard to
Probability of Repayment
Cost of Adjustment, Bond Holdings
Cost of Issuing Bank Capital
Benet of Holding Excess Bank Capital
Capital Adequacy Ratio (Deterministic
Component)

0.0

1.5

0.05

B
V
V V
D

0.05
0.08
0.004
0.08

Central Bank

1
2
Shocks

H , H

Description

0.1
0.0
1.5
0.15
0.6
0.6, 0.002

Reserve Requirement Rate

Degree of Interest Rate Smoothing by Central
Bank
Response of Renance Rate to Ination
Deviations
Response of Renance Rate to Output Gap
Degree of Persistence, Monetary Policy Shock
Persistence/Standard Deviation, Housing
Demand Shock

212

International Journal of Central Banking

September 2013

parameter values.22 A few parameters are worth noting here; in particular, the adjustment cost parameter for holdings of bank debt,
V , is set at 0.3. The share of capital in output of intermediate
goods is set at 0.35, whereas the elasticity of demand for intermediate goods is set at 10. Both values are close to those used by Medina
and Soto (2007) for Chile. The adjustment cost for transforming the
nal good into investment, K , is set at 14. The elasticities of the
repayment probability with respect to the collateral-to-risky-loans
ratio, the bank capital-to-risky-assets ratio, and cyclical output are
set at 1 = 0.03, 2 = 0.0, and 3 = 0.15, respectively.23 Thus, in
the benchmark calibration, we abstract from the monitoring incentive eect of the bank capital channel identied earlier. The value
of 1 is close to the elasticity of the external nance premium
the inverse of the repayment probability hereto leverage used by
Liu and Seeiso (2012) for South Africa. The elasticities of the risk
weight with respect to the repayment probability, q , as well as the
cost parameter V V are set at low values, 0.05 and 0.004, respectively. Because V V is a parameter that could potentially inuence
in important ways the transmission eects of capital requirements
(and thus the performance of the countercyclical regulatory rule),
we will later consider alternative values. Our specication of the
risk weight (6) implies that its value is unity in the steady state; we
set the overall capital adequacy ratio to 0.08. We also calibrate the
excess capital-to-risky-assets ratio to be equal to 0.04. This implies
that the steady-state ratio of total bank capital to risky loans is set at
about 12 percent (so that V E /V R = 0.53), in line with the evidence
reported in Agenor and Pereira da Silva (2012). For the monetary
policy rule, we set 1 = 1.2, 2 = 0.2, and = 0.0 initially, which is
in line with the results for Latin America reported by Moura and de
Carvalho (2010) and Cebi (2011) for Turkey. The share of government spending in output is set at 20 percent, a value consistent with
the evidence for a number of middle-income countries, such as South
22
A more complete table is provided in a more detailed version of this article,
available upon request.
23
Higher values of 1 destabilize the model very quickly. While this may not
be inconsistent with recent facts, in the model this is also related to the fact that
only housing is used as collateral. If physical capital could also be used for that
purpose, it would be possible to increase 1 quite substantially, as can be inferred
from the results in Agenor and Alper (2012).

Vol. 9 No. 3

Capital Regulation, Monetary Policy

213

Africa (see Liu and Seeiso 2012). Our calibration also implies a total
(corporate) credit-to-output ratio of about 60 percent, which is consistent with data for several middle-income countries where nancial
intermediation is bank based and consumer lending remains limited.
The proportion of constrained households, , is set to 0.3.24 For the
degree of persistence of the housing demand shock, we assume a
value of 0.6.
5.

Housing Demand Shock

To illustrate the functioning of the model when concerns with nancial stability are absent, we consider as a base experiment a positive temporary shock to housing preferences that translates into an
impact increase in real house prices of 1.6 percentage points.25 The
results are summarized in gure 3, under two cases: the rst under
the calibration described earlier for the repayment probability, and
the second where the probability of repayment is constant (so that
1 = 2 = 3 = 0). This allows us to compare in a simple manner
the response of the model with and without credit market frictions.
The immediate eect of the shock is to raise housing prices.
In turn, this raises the value of collateral and, with credit market
frictions, thus the repayment probability. The lending rate therefore drops, thereby stimulating investment in the rst period. The
increase in aggregate demand is matched by an increase in supply
(given sticky prices) and this stimulates the demand for labor. Over
time, the increase in investment raises the capital stock; this raises
24
This value is substantially lower than some of the results reported in Agenor
and Montiel (2008). However, much of the early literature relates to developing countries in general (including therefore low-income countries) and does not
account for the nancial liberalization that has occurred in many middle-income
countries over the past two decades. As a matter of comparison, note that the
proportion of constrained households in the euro area estimated by Coenen and
Straub (2005) varies between 0.25 and 0.37.
25
Our goal is to illustrate the dynamics induced by a fundamental shock
to asset prices, rather than unsustainable changes in expectations. Although the
size and the persistence of the eect of our shock on house prices is not comparable to the much more persistent movements in these prices observed during
typical booms, the qualitative eects are quite similar to those of a standard
house-price bubble. A more formal attempt to model asset-price bubbles, involving more persistent shocks, could follow along the lines of Bernanke and Gertler
(1999), as done in Levieuge (2009).

214

International Journal of Central Banking

September 2013

Figure 3. Base Experiment: Positive Housing

Demand Shock
(Deviations from Steady State)

Note: Interest rates, ination rate, and the repayment probability are measured
in absolute deviations; that is, in the relevant graphs, a value of 0.05 for these
variables corresponds to a 5-percentage-point deviation in absolute terms.

the marginal product of labor and therefore gross wages. At the same
time, the increase in the capital stock tends to lower the rental rate
of capital.
The increase in output tends to raise immediately the policy
rate; combined with the increase in the gross wage, this tends to

Vol. 9 No. 3

Capital Regulation, Monetary Policy

215

raise the eective cost of labor for the producers of intermediate

goods.26 Because the rental rate of capital does not change on impact
(due to the one-period lag in capital accumulation), marginal costs
unambiguously increase in the rst period. Ination therefore rises,
putting further upward pressure on the policy rate. Over time, the
reduction in the rental rate of capital induced by the boom in investment tends to lower marginal costs and ination.
The higher policy rate (which translates into a higher deposit
rate) is also accompanied by a higher interest rate on government
bonds. The reason is that the increase in the deposit rate raises
demand for these assets; this translates into a reduction in bank
borrowing from the central bank. The reduction in money supply
must be matched by a lower demand for currency, which is brought
about by a higher bond rate. This leads to a shift in consumption from the present to the future for unconstrained consumers,
who do not benet directly from the collateral eect induced by
higher house prices because they do not borrow from the banking
system; moreover, the intertemporal eect in consumption dominates any wealth eect associated with a positive (but temporary)
housing price shock. This downward eect dominates the positive
response of spending by constrained consumers (given the increase
in their labor income), so that aggregate consumption falls.27 This
mitigates the increase in aggregate demand induced by the initial
investment boom. The drop in consumption reduces the marginal
utility of leisure and induces unconstrained households to supply
more labor, thereby mitigating the upward pressure on real wages.
The increase in the repayment probability lowers the risk weight
under the Basel II-type regulatory capital regime that we consider,
which tends to reduce capital requirements. However, the increase
in risky assets (loans to the CG producer) dominates, which implies
that required capital increases; in turn, this tends to raise the rate
of return on bank debt. Given that the policy rate increases, the net
26

Recall that the eective cost of labor for IG producers is the gross renance
rate 1 + iR times the real wage.
27
The fall in consumption, which lowers the demand for currency, attenuates
the initial increase in the bond rate. A positive response of aggregate consumption
could be obtained by increasing signicantly the share of unconstrained households, compared with the base calibration. However, a consumption boom is not
a stylized fact associated with periods of sustained increases in house prices; see
Detken and Smets (2004) and International Monetary Fund (2011).

216

International Journal of Central Banking

September 2013

eect on the demand for excess bank capital is ambiguous in general;

given our calibration, it actually increases. The increase in the cost
of issuing bank debt mitigates the downward impact on the lending
rate associated with the collateral eect. By contrast, in the absence
of credit market frictions, the increase in real house prices exerts no
collateral eect; the lending rate does not fall on impact, implying
that investment does not increase. Because the shock is temporary,
the wealth eect associated with higher house prices is muted, and
the model does not display any signicant dynamics.
Thus, the results of this experiment suggest that with endogenous credit market frictions the model is consistent with the view
that a demand-induced boom in housing prices may lead, through
a nancial accelerator mechanism that operates through collateral
values and borrowing costs, to rapid increases in investment, an
expansion in output, inationary pressures, and debt accumulation.
Conversely, a bust in housing prices, through the same asset-price
channel, can lead to a credit crunch, a contraction in output and
investment, and deation. Because our analysis focused on a temporary shock, the simulation results do not identify explicitly any
tendency for instability; however, it is clear from our description
of the transmission mechanism that a shock of sucient duration
could well induce unsustainable movements in real and nancial variables, and thus economic instability, in the absence of a timely policy
response. We now turn to an examination of the policies that could
help to mitigate nancial instability, and the extent to which doing
so may entail trade-os with respect to macroeconomic stability.
6.

Policy Rules for Economic Stability

We now consider two alternative approaches to mitigating nancial instability, in line with some recent proposals. The rst involves
adjusting the renance rate in response to a nancial stability indicator, whereas the second focuses on reducing the degree of procyclicality of the nancial system through discretionary regulation
of bank capital, in line with the new Basel III regime.

6.1

An Augmented Interest Rate Rule

In the rst approach that we consider, the central bank adjusts its
policy rate directly in response to changes in an indicator of nancial

Vol. 9 No. 3

Capital Regulation, Monetary Policy

217

stability. Specically, we replace the interest rate rule (12) with the
augmented rule
R
iR
r + t + 1 (t T ) + 2 ln ytG
t = it1 + (1 )[

3 ( ln ltF,I

(14)

ln lF,I )] + t ,

where ln ltF,I (with ltF,I = LF,I

t /Pt ) is the growth rate of real credit
to the CG producer, and ln lF,I is the steady-state value of that
variable.28 Thus, in line with the discussion in the introduction,
the central bank sets its renance rate in part to lean against the
wind. Following a positive shock to housing prices, for instance, and
an increase in collateral values, the lending rate drops and stimulates investment; an increase in the renance rate tends to mitigate
the drop in the cost of bank borrowing and therefore to dampen
the investment boom. We can analyze the performance of alternative interest rate rules (that is, dierent values of 3 > 0) in terms of
specic indicators of macroeconomic stability and nancial stability,
and compare them with the base case where 3 = 0.

6.2

A Countercyclical Regulatory Capital Rule

The second rule can be introduced by decomposing the overall capital ratio, t , into a deterministic component, D , and a cyclical
component, C
t :
t = D + C
t .

(15)

Thus, the component D can be viewed as the minimum capital

adequacy ratio imposed under Pillar 1 of the Basel regime, whereas
the component C
t can be viewed as the discretionary adjustment that now forms part of the Basel III regime.29 The experiment
D
presented in the previous section assumed that C
t = 0 and t =
t.
28
A possible alternative to real credit growth is nominal credit growth. However, the latter includes an inationary component; thus, the benets of responding to it via an interest rate rule might result from that component.
29
In the ongoing debate about the implementation of Basel III, there is still
discussion as to whether the countercyclical component should be made mandatory or left to the discretion of local regulators. See Repullo and Saurina (2011)
and the nal section for a critical discussion and some further comments.

218

International Journal of Central Banking

September 2013

Adjustment of the cyclical component follows a simple dynamic

rule; we relate it only to deviations of the growth rate of real credit
to the CG producer from its steady-state value:
F,I
C
C
ln lF,I ),
t = ( ln lt

(16)

where C > 0 is the adjustment parameter.30 Thus, the macroprudential rule is designed so as to directly counter the easing of
lending conditions that induces borrowers to take on more debt as
house prices increase.
Suppose that in period t there is an increase in housing prices due
to a demand shock. As discussed earlier, the rise in the value of collateral tends to raise the repayment probability immediately, which
reduces the lending rate and stimulates borrowing for investment by
the CG producer. The increase in house prices is therefore procyclical. A rule like (16), by imposing higher capital requirements, tends
to raise directly the cost of issuing debt by the bank, thereby mitigating the initial expansionary eect associated with higher collateral
values. Thus, it dampens procyclicality of the nancial system. It
is consistent with the spirit, if not the letter, of some of the recent
proposals to amend or reform Basel II capital standards, such as
Goodhart and Persaud (2008), as mentioned in the introduction,
and the recently adopted rule under Basel III.31
Nevertheless, in the general equilibrium framework, whether the
eect on the lending rate is positive or negative depends also on the
net eect on the repayment probability, which depends (as noted
earlier) not only on the collateral-CG loan ratio, but also on the
cyclical position of the economy and the bank capital-to-risky-assets
ratio. In particular, under the risk-sensitive regulatory regime that
we consider, the increase in the repayment probability induced by
the improvement in the collateral-to-risky-loans ratio lowers the risk
weight and tends to reduce capital requirements. If the countercyclical regulatory rule is not too aggressive (in the sense that C is
Although the rule for C
t is not a backward-looking rule, we assume that the
bank takes t as given when solving its optimization problem. Thus, the bank
pricing rules derived earlier remain unchanged.
31
The second dimension of the rule proposed by Brunnermeier et al. (2009),
related to the mismatch in the maturity of assets and liabilities, cannot be
implemented in our setup.
30

Vol. 9 No. 3

Capital Regulation, Monetary Policy

219

not too high), the capital-to-risky-assets ratio will fall, and this will
tend to mitigate the initial rise in the repayment probability and
the drop in the loan rate. If so, then, the monitoring incentive eect
will operate in the same direction as the cost eect. By contrast, if
the regulatory rule is very aggressive (high C ), total capital may
increase by more than the increase in risky loans, and this may lead
to a higher repayment probability and a lower lending rate, making
the regulatory rule (16) more, rather than less, procyclical. Thus the
conicting eects of the two dimensions of the bank capital channel
may make the policy counterproductive. Alternatively, if the benet
from holding capital buers (as measured by V V ) is not too large,
the regulatory capital rule may be even more eective. To assess the
most likely outcomes requires numerical simulations, based on some
optimality criteria.

6.3

Stability Measures and Policy Loss Function

We assume in what follows that the central bank is concerned with

two objectives, macroeconomic stability and nancial stability. Both
concepts, as noted earlier, can be dened in dierent ways; in this
paper, we dene macroeconomic stability in terms of a weighted
average of the coecient of variation of ination and the output
gap, with equal weights at rst, and nancial stability in terms of
the coecient of variation of three alternative indicators: real housing prices, the credit-to-GDP ratio, and the loan spread.32 Empirical
studies of banking crises in developing countries have indeed documented that large increases in credit often precede the occurrence of
these crises (see Demirguc-Kunt and Detragiache 2005 and Agenor
and Montiel 2008). We also consider composite measures involving
all three indicators, and the last two only, given that theyunlike
the rstare more directly related to the nancial system. In addition, we also dene a composite index of economic stability, dened
rst with a weight of 0.8 for macroeconomic stability and 0.2 for
nancial stability and second with a weight of 0.9 for macroeconomic stability. Thus we consider the case (which we believe to be

32

In turn, coecients of variations are based on the asymptotic (unconditional)

variances of the relevant variable.

220

International Journal of Central Banking

September 2013

the more relevant one in practice) where the central bank remains
mainly focused on macroeconomic stability.33
7.

Model Dynamics and Policy Trade-Os

Before we study optimal policies, we begin with a simple examination of how the two rules (14) and (16) operate, independently of
each other.
The upper panels in gure 4 show 3D diagrams of macroeconomic
and nancial stability indicatorsmeasured, in the latter case, using
the various indicators described earlier, individually and in combination. In all diagrams, 3 varies between 0 and 2.5 (northwest horizontal axis) and C varies between 0 and 10 (northeast horizontal axis).
The outer contour of each graph corresponds to the corner
cases where either 3 = 0 or C = 0. They therefore show how
the two rules aect our measures of (in)stability when the underlying shock is the same as described earliera positive shock to
housing preferences. The outer contours suggest that there is no
trade-o among policy objectives: a more aggressive response to
credit growth gaps leads, in either case, to a monotonic reduction in
both macroeconomic and nancial volatility, regardless of the measure used.34 Thus, from the perspective of either policy objective,
monetary and countercyclical regulatory policies appear to be substitutes rather than complements.35 However, the curves have a convex shape, which indicates that the marginal benet of either policy
diminishes as it becomes more aggressive. Indeed, the upper panels
suggest that beyond a value of 3 = 0.6, the gain in terms of reduced
33

The policy loss function could also incorporate the volatility of the change in
the policy rate, to capture the cost attached to nominal interest rate uctuations.
Because this would lead to more persistence, we account for that cost directly by
considering dierent values of the parameter in our experiments.
34
Although not reported, the volatility of the renance rate falls as well, as
either policy rule becomes more aggressive. This is because a model with forwardlooking private agents, such as this one, has strong expectational eects
households anticipate a stronger reaction from the central bank and factor it
into their decision-making process. The result is that monetary policy works
partly through the threat of a stronger response instead of actually delivering
that stronger response.
35
Note that the absence of trade-os relates only to policy responses with
respect to housing demand shocks and depends on the fact that we abstract from
output-ination trade-os by focusing on the volatility of nominal income.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

221

Figure 4. Housing Demand Shock: Policy Rules and

Macroeconomic/Financial Stability Trade-os

volatility become smaller; a similar result holds in the lower panels for C above 2 when real house prices are used, 2.5 when the
credit-to-GDP ratio is used, and 4 when the loan spread is used.
The next step is to examine to what extent their combination
leads to lower variability in either target. This is also illustrated in
gure 4. The upper-level diagrams suggest clearly that, given our
base calibration, the marginal contribution of the regulatory capital
rule, once an augmented interest rate is in place, decreases rapidly,

222

International Journal of Central Banking

September 2013

in terms of either macroeconomic stability or nancial stability. The

lower-level diagrams also show the results for the economic stability
index described earlier, when nancial stability is measured in terms
of either the credit-to-GDP ratio or a combination of the creditto-GDP ratio and the loan spread. Not surprisingly, the results are
qualitatively similar: given our base calibration, the marginal contribution of the regulatory capital rule to economic stability decreases
rapidly once the augmented interest rate rule is used.
Put dierently, countercyclical bank regulation may not provide
very large benets if monetary policy can be made more reactive
to an indicator of nancial stability. However, to the extent that
monetary policy is constrained (because the central bank fears destabilizing markets by raising interest rates too sharply while ination
is subdued, for instance), there may be some degree of complementarity between the two rules: a countercyclical regulatory rule can
help to improve outcomes with respect to both objectives. Put differently, the convexity of the curves shown in gure 4 suggests that,
if there is a cost in implementing large changes in the policy rate, it
may be optimal to combine the two policies.
8.

Optimal Policies

We now consider how the parameters 3 and C in the two rules (14)
and (16) can be determined optimally, so as to minimize economic
instability, as dened earlier. As can be inferred from gure 4, the
fact that there is no trade-o among policy objectives, and that the
augmented interest rate rule is more powerful in mitigating volatility, means that, in general, if there is no restriction on the value of
3 , the optimal policy always implies setting C = 0. To generate a
role for regulatory policy, suppose that the central bank chooses not
to react beyond a certain point to changes in credit growth, out of
concern that large changes in interest rates can generate instability.
To account for this, we perform a set of experiments in which we
arbitrarily limit the value of 3 to a plausible upper limit, which we
set at 2.5. At the same time, we impose a higher limit on the parameter C , equal to 10. Thus, our analysis is best described as a search
for constrained optimal policies, with a relatively less aggressive
potential response of monetary policy to credit growth gaps.
Tables 2 and 3 report the results. We calculate optimal values
based on dierent sets of two other key parameters: the response

Vol. 9 No. 3

Capital Regulation, Monetary Policy

223

Table 2. Housing Demand Shock: Optimal Policy

Parameters (Relative Weight on Financial Stability: 0.1)
1
(3 , c )

1.2

1.5

1.8

Financial Stability Indicator: Real House Prices

0
0.4
0.8

2.5, 0
2.5, 0
2.5, 0

2.5, 0
2.5, 0
2.5, 0

2.5, 0
2.5, 0
2.5, 0

Financial Stability Indicator: Credit-to-GDP Ratio

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 7
2.5, 7.5
2.5, 9

2.5, 3
2.5, 4
2.5, 7

Financial Stability Indicator: Loan Spread

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 10
2.5, 10
2.5, 10

2.5, 6
2.5, 7
2.5, 10

Financial Stability Indicator: Real Credit and Spread

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 8.5
2.5, 9
2.5, 10

2.5, 4.5
2.5, 5.5
2.5, 10

Financial Stability Indicator: Real Credit, Spread, Real House Prices

(weight 0.1)
0
0.4
0.8

2.5, 4
2.5, 4
2.5, 4

2.5, 0
2.5, 0
2.5, 2.5

2.5, 0
2.5, 0
2.5, 2

of the policy rate to deviations of ination from target, 1 , and

the degree of persistence in the policy rate, . In addition to the
baseline value 1 = 1.2, we consider two other values, 1.5 and 1.8,
which capture a more aggressive stance toward ination. For the
smoothing parameter, we consider, in addition to the baseline value
= 0.0, values of = 0.4 and = 0.8; again, these alternative values capture the view that underlying preferences reect a concern

224

International Journal of Central Banking

September 2013

Table 3. Housing Demand Shock: Optimal Policy

Parameters (Relative Weight on Financial Stability: 0.2)
1
(3 , c )

1.2

1.5

1.8

Financial Stability Indicator: Real House Prices

0
0.4
0.8

2.5, 0
2.5, 0
2.5, 0

2.5, 0
2.5, 0
2.5, 0

2.5, 0
2.5, 0
2.5, 0

Financial Stability Indicator: Credit-to-GDP Ratio

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 7.5
2.5, 8
2.5, 10

2.5, 3.5
2.5, 4.5
2.5, 8.5

Financial Stability Indicator: Loan Spread

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 10
2.5, 10
2.5, 10

2.5, 10
2.5, 10
2.5, 10

Financial Stability Indicator: Real Credit and Spread

0
0.4
0.8

2.5, 10
2.5, 10
2.5, 10

2.5, 10
2.5, 10
2.5, 10

2.5, 6.5
2.5, 8
2.5, 10

Financial Stability Indicator: Real Credit, Spread, Real House Prices

(weight 0.1)
0
0.4
0.8

2.5, 0
2.5, 0
2.5, 2

2.5, 0
2.5, 0
2.5, 0.5

2.5, 0
2.5, 0
2.5, 0.5

with movements in policy rates that are too large, possibly because
the central bank believes that large movements can destabilize nancial markets.36 We calculate the value of the loss function dened
36
Alternatively, high persistence in the policy rate is viewed as desirable
because, as argued by Woodford (1999), it allows the government to commit
to future ination targets.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

225

earlier for all values of 3 varying between 0 and 2.5, and C varying
between 0 and 10. We perform a grid search at intervals of 0.0125
for 3 and 0.5 for C , which is quite large but sucient to illustrate
our main points.
When nancial stability is measured in terms of nancial indicators (the credit-to-GDP ratio and the loan spread, individually
or in combination) the results can be summarized as follows. First,
regardless of the degree of persistence in the policy rate, and regardless of the strength with which monetary policy can react to ination
deviations from its target value, it is always optimal to use monetary policy to its limit. Second, the stronger monetary policy can
react to ination deviations from its target value (the higher 1 is),
the lower should be the reliance on macroprudential policy, particularly so if the weight attached to nancial stability in the policy
loss function is small (0.1, rather than 0.2). Third, the greater the
degree of interest rate smoothing (the higher is, possibly because
the central bank fears destabilizing markets by raising interest rates
too sharply), the stronger should be the countercyclical regulatory
responseeven if monetary policy can react strongly to ination.
Fourth, the stronger the policymakers concern with nancial stability, the stronger should be the sensitivity of countercyclical regulation to the credit growth gap. The second result suggests that
monetary and regulatory policies are partial substitutes, whereas
the third result suggests that they are partial complements; even
with aggressive responses to ination and credit growth gaps, it is
optimal to rely also on the countercyclical regulatory rule as the
degree of interest rate smoothing increases.
When nancial stability is measured in terms of real house prices
(either individually or in combination with the other indicators), the
rst result continues to hold. The second and the third do not when
the indicator is used individually, but when it is used in combination
with a relatively low weight (0.1, compared with 0.45 for the others),
the second holds when the degree of interest rate smoothing is high
( = 0.8 ) and so does the third as increases, regardless of the
value of 1 . Regarding the fourth, again it does not hold when real
house prices are used individually, but when it is used in combination with the others (again, with a relatively low weight) and with
a high degree of interest rate smoothing, we get what appears to
be a counterintuitive result: the stronger the policymakers concern

226

International Journal of Central Banking

September 2013

with nancial stability, the weaker should be the sensitivity of countercyclical regulation to the credit growth gap. However, it is not
clear that much weight should be attached to this result, because
real house prices may not be the most reliable indicator of nancial
stability.37
9.

Sensitivity Analysis

To assess the sensitivity of the previous results, we consider several

additional experiments: a higher elasticity of the repayment probability to the capital-to-risky-assets ratio and the solution of optimal
policy parameters in response to productivity and demand shocks.

9.1

Response to Capital-to-Risky-Assets Ratio

For the rst experiment, we account for a monitoring incentive eect

of the bank capital channel (as described earlier) by increasing the
elasticity 2 from its initial value of zero to 0.05. This alternative
value is within the two-standard-error condence interval for the
elasticity of the bank loan spread with respect to the capital-torisky-assets ratio estimated by Fonseca, Gonzalez, and Pereira da
Silva (2010) for developing countries.
The results (which are not reported here to save space) indicate
that a stronger bank capital channel (operating through a monitoring incentive eect) strengthens the countercyclical regulatory rule
in the presence of risk-sensitive weights. The reason is that following the housing price shock and the initial increase in the repayment
probability (as discussed earlier), the weight on risky assets in our
Basel II-type regime tends to fall; this lowers capital requirements
and therefore tends to reduce the capital-to-risky-assets ratio. In
turn, this mitigates the banks incentives to monitor and reduces
the repayment probabilitythereby osetting the initial increase in
that variable and dampening the initial drop in the lending rate.
37
Another proxy for nancial (in)stability, as proposed by Granville and Mallick
(2009), is the volatility of the deposit-loan ratio. For more micro-based approaches
to measuring nancial instability, see De Graeve, Kick, and Koetter (2008), Blavy
and Souto (2009), and Segoviano and Goodhart (2009).

Vol. 9 No. 3

Capital Regulation, Monetary Policy

227

Thus, a stronger bank capital channel (operating through monitoring and screening incentives) makes the countercyclical regulatory
rule more eective.

9.2

Productivity and Demand Shocks

To assess the sensitivity of optimal policy choices to the nature of

shocks, we considered rst a positive productivity disturbance, taking the form of a one-percentage-point increase in the technology
parameter that characterizes the production function of intermediate goods. The results (which are not shown to save space) indicate
that in almost all cases monetary policy should be used to the fullest
(3 = 2.5), and macroprudential policy not at all, regardless of the
values of 1 and . The reason is fairly intuitive. With the housing preference shock discussed earlier, a change in real house prices
has two immediate eects on the repayment probability: a collateral eect and a cyclical output eect. The countercyclical regulatory rule operates through both channels, as noted earlier. With
a productivity shock, however, the rst eect is absent on impact
and remains muted subsequently, given that in our base calibration
the elasticity of the repayment probability with respect to collateral
is relatively small. The main channel through which macroprudential policy aects the repayment probability is through the cost of
issuing bank capitalbut its coecient is relatively small in the
loan interest-rate-setting condition. Thus, monetary policy (which
has a large, direct eect on the cost of credit) is more eective to
ensure macroeconomic and nancial stability. This result continues
to hold even with substantially higher values of the elasticity of
the repayment probability with respect to the cyclical component of
output.
Second, we considered an aggregate demand disturbance, taking
the form of a temporary increase in the share of government spending in GDP, from an initial value of 0.2 to 0.22. The results (which,
again, are not shown to save space) indicate that, as before, monetary policy should be used to the fullest, whereas in most cases
macroprudential policy should be usedexcept, importantly, when
the credit-to-GDP ratio is used as a measure of nancial stability, in
which case a mild response of the countercyclical capital rule appears
to be optimal. The intuition for the general result is similar to what

228

International Journal of Central Banking

September 2013

occurs under the productivity shock; the absence of an initial collateral eect mitigates the magnitude of the impact eect of the
shock. Thus, a general implication of this analysis is that the results
obtained earlier with a housing preference shock, which militate in
favor of the intensive use of macroprudential policy in a variety of
policy parameter congurations, hold largely because the shock has
a direct eect on collateral values.38
Finally, we also conducted some sensitivity analysis with respect
to some of the parameters of the model. For lack of space we do
not report the full simulation results here, but rather provide a brief
intuition of their implications. First, we considered a higher adjustment cost parameter for investment, K . This is a critical parameter
because it aects investment demand and thus credit growth and
output uctuations. A higher value of K implies a muted response
of credit growth and output in response to shocks, thereby reducing the need for countercyclical monetary and regulatory policies.
The value of K also matters for the relative eectiveness of the
two policy instruments. Because the policy rate aects the whole
structure of interest rates in the economy whereas the regulatory
rule directly impinges on the loan rate only, a change in the renance rate becomes more detrimental for consumption when the
adjustment cost parameter is smaller. This helps to illustrate the
view that changes in the policy rate may be too brutal to mitigate
nancial instability.
Second, we considered a higher value of the adjustment cost
parameter for households holdings of bank debt, V . In the model,
as discussed earlier, V is the main determinant of the spread
between the returns on bank debt and government bonds. Considering the fact that the interest rate on risky loans is a weighted
average of the cost of bank capital and the policy rate (see (9)), and
that the regulatory rule determines the weight attached to the cost
of bank capital in the loan rate, this parameter plays a major role
in the eectiveness of the regulatory rule. Indeed, simulation results

38
Another experiment that would have a direct eect on collateral values is a
shock to , the proportion of physical assets that can be pledged as collateral.
Note also that with lower values of 1 and higher values of 3 , the repayment
probability would be relatively more responsive to movements in cyclical output,
and macroprudential regulation could be more eective.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

229

indicate that the higher the value of the adjustment cost of holding
bank debt, the more eective the regulatory rule becomes.
Third, we considered a higher share of constrained households,
. This parameter has important implications for the responsiveness of consumption to interest rates. Small changes in the value
of do not have a large, quantitative impact on the results. However, large changes do matter for the eectiveness of countercyclical
policiesespecially changes in the policy interest rate. The reason,
related to the point made earlier, is that the larger is, all else
equal, the lower is the eect of interest rates on consumption. In
turn, when consumption becomes less responsive to changes in the
policy rate, an interest rate rule operates mainly through investment demandas is the case for a countercyclical regulatory rule.
Thus, with a substantially higher share of constrained households in
the economy, monetary and regulatory rules become more similar in
terms of their eects on macroeconomic and nancial stability.
Fourth, we examined the eects of a higher elasticity of the risk
weight with respect to the probability of repayment, q . This parameter plays a key role in determining the degree of procyclicality of a
Basel II-type regime (see Agenor, Alper, and Pereira da Silva 2012):
a higher elasticity with respect to the repayment probability magnies uctuations in regulatory capital. In turn, this translates into
higher loan rates in bad times and lower loan rates in good times,
resulting in a stronger response of both the policy rate and the countercyclical capital requirement, as both aggregate demand and credit
growth uctuate more. Moreover, the countercyclical capital regulation tool becomes more eective than the policy rate in smoothing
out these uctuations as it, in part, osets procyclical movements in
bank capital.
10.

Summary and Policy Implications

The purpose of this paper has been to examine the roles of monetary
policy and bank capital regulatory policy in mitigating procyclicality
and promoting macroeconomic and nancial stability. The analysis was based on a dynamic, structural optimizing macroeconomic
model with imperfect credit markets and Basel-type bank capital
regulation, with both cost and monitoring incentive eects. The

230

International Journal of Central Banking

September 2013

model incorporates also an asset-price/nancial accelerator mechanism, by which induced changes in asset prices aect the value of the
collateral pledged by borrowers and hence the cost of loans. Macroeconomic stability is dened in terms of the volatility of ination
and the output gap, whereas nancial stability is dened in terms of
the variability of three alternative indicators, used both individually
and in combination: real house prices, the credit-to-GDP ratio, and
the loan spread. Our basic experiment showed that with endogenous
credit market frictions, a positive housing demand shock, through
its eect on collateral values, leads to a credit expansion and an
investment boom. This is consistent with the evidence and provides
the proper background for discussing the roles of monetary and
regulatory policies in achieving economic stability.
We next considered two policy rules aimed at mitigating nancial
instability: a credit-augmented interest rate rule and a Basel IIItype countercyclical regulatory capital rule, both based on a credit
growth gap measure. The premise for the rst rule is that a central banks response to credit growth may serve to stabilize market
conditions (namely, the lending rate) by osetting the expansionary
(balance sheet) eect induced by a positive shock to asset prices.
The second rule is motivated by the view that capital regulation
should be operated in a more exible manner to account for changes
in systemic risk over the business cycle. In both cases, the underlying view is that the expansion of credit is an essential ingredient in
the buildup of imbalances in the nancial system.
Numerical experiments with a housing demand shock showed,
rst, that there are no trade-os between macroeconomic and nancial stability objectives when each instrument is used in isolation.
This is related to the fact that in our framework capital regulation is a fairly eective tool for constraining the growth in aggregate
demand, because all investment is nanced through bank loans. Second, if monetary policy cannot react suciently strongly to ination
deviations from targetsdue to concerns about interest rate volatility feeding uncertainty about economic fundamentals, or because
of fears that sharp changes in interest rates while ination is low
could induce volatility in price expectations and destabilize markets, for instancecombining a credit-augmented interest rate rule
and a countercyclical capital regulatory rule is optimal for mitigating

Vol. 9 No. 3

Capital Regulation, Monetary Policy

231

economic instability. This result also holds if monetary policy can

respond aggressively to ination, as long as the degree of interest
rate smoothing is high. Third, when nancial stability is measured
solely in terms of nancial indicators, the greater the concern with
nancial stability (compared with macroeconomic instability) is, the
larger the role of countercyclical regulation. Sensitivity analysis suggests that these results are specic to shocks that have a direct eect
on collateral values. Put dierently, whether macroprudential policy (in the form of a countercyclical capital rule) is eective or not
depends on the nature of the shocks to which the economy is subject.
Although our analysis has been conducted in a middle-income
country context, our results are useful for the current debate on
the role of monetary policy in fostering nancial stability and on
reforming bank capital standards in both industrial and developing
countries.39 First, some observers have argued that, to the extent
that credit growth aects output (as is the case in our model), there
may be no reason for monetary policy to react above and beyond
what is required to stabilize output and ination. Our results suggest
that this precept is not generally true. Second, as noted in the introduction, in November 2010 the G20, under proposition by the Basel
Committee on Banking Supervision (2010), adopted a countercyclical capital buer rule based on a credit-to-GDP gap measure. The
fundamental idea underlying countercyclical regulation is that nancial markets, left to themselves, are inherently procyclical. But even
though (excessive) risk becomes apparent in bad times, it is mainly
generated in boom times. Thus, the time for regulators to intervene
is precisely during good times, to prevent excessive risk taking and
moderate the growth in bank credit. By operating in a countercyclical fashion, nancial regulation therefore helps ensure that banks
build up resources in good times to help cushion adverse shocks in
bad times. At the same time, it is important to implement countercyclical regulation through relatively simple rules that cannot be
easily changed by regulators so they will not become captured
by the general exuberance that characterizes booms nor by vested
39
Indeed, the credit market imperfections emphasized in this paper are a matter
of dierences in degree between developed and developing countries. See Agenor
and Pereira da Silva (2010) for a more detailed discussion.

232

International Journal of Central Banking

September 2013

interests (see Brunnermeier et al. 2009). Our analysis suggests, however, that the benets of these rules may also be less than believed
if monetary policy can endogenously respond to (excessive) credit
growth. It also casts doubt on the wisdom of making a countercyclical capital requirement component mandatory under Pillar 1 of the
new Basel III regime; large structural dierences across countries
may make the attempt to impose a uniform rule problematic.
The analysis can be extended in several directions. First, as noted
earlier, although nancial instability is commonly associated with
periods of booms and busts in asset prices and credit, there is no
widely accepted denition and (hence) indicator of nancial stability; in our analysis we used several indicators, both alternatively and
in combination. But other indicators are also possible, as suggested
by the empirical evidence.40 More generally, it could be argued that
nancial instability is associated with uctuations in several nancial
and real economic variables rather than in just asset prices. Financial stability may therefore be dicult to assess by merely focusing
on a single or a limited set of either nancial or real economic variables. A useful extension would be consequently to consider broad
(or composite) indicators and examine how our results are altered. In
particular, one could combine real house prices with a credit growth
gap measure and bank lending rates to derive a composite indicator of nancial instability, with weights on each individual variable
based on the literature that measures the relative importance of each
in predicting either banking crises (see Demirguc-Kunt and Detragiache 2005, and Agenor and Montiel 2008) or periods of nancial
stress (see Misina and Tkacz 2009).
Second, in the denition of our countercyclical regulatory capital rule, we included only loans to capital goods producers, on the
grounds that loans for working capital needs are not risky. In practice, however, a legitimate question is whether such a distinction
between components of credit can be meaningfully implemented,
given well-known fungibility problems and the risk that dierences in
40
See Segoviano and Goodhart (2009). In De Graeve, Kick, and Koetter (2008)
for instance, nancial stability is dened and measured as a banks probability
of distress according to the supervisors denition of problem banks used for
supervisory policy.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

233

regulation may encourage banks to engage in distortive practices (see

Hilbers et al. 2005). This is a particularly important consideration
for middle-income countries, where the institutional and regulatory
environment is often weak to begin with. If so, then, there may be
little choice but to apply the regulatory rule on total creditwith
possible adverse welfare eects, in countries where working capital
needs represent a large share of bank loans.
Third, a more formal analysis of optimal rules, based on conditional discounted utility of the dierent categories of agents, would
allow a more comprehensive study of the welfare eects of a creditresponsive monetary policy and countercyclical regulatory capital
rules, in the presence of both macroeconomic and nancial stability
objectives.41
Finally, we must point out that our assessment of the benets
of countercyclical regulatory rules did not address implementation
issues. In general, the implementation of macroprudential regulation
requires stronger coordination (in countries where they are independent to begin with) between central banks and supervisory authorities. The issue of how best to achieve such coordination in practice
remains a matter of debate. Another practical issue to consider is
the extent to which the introduction of macroprudential rules may
adversely aect the anti-ination credibility of the central banka
particularly important concern in countries where such credibility
remains precarious. The risk is that the announcement of greater
reliance on macroprudential regulation could give rise to expectations that the central bank may now pursue a more accommodative
monetary policy in the face of inationary pressures, thereby weakening its credibility. A transparent communications strategy that
claries the nature of macroprudential rules and the way they are
expected to operate, while emphasizing the fundamental complementarity between the traditional objectives of monetary policy
and the goal of nancial stability, may thus be essential.

41
Faia and Monacelli (2007), for instance, found that monetary policy should
respond to increases in asset prices by lowering interest rates. In addition, when
monetary policy responds strongly to ination, the marginal welfare gain of
responding to asset prices vanishes. However, Angeloni and Faia (2010) found
opposite results.

234

International Journal of Central Banking

September 2013

References
Agenor, P.-R., and K. Alper. 2012. Monetary Shocks and Central Bank Liquidity with Credit Market Imperfections. Oxford
Economic Papers 64 (3): 56391.
Agenor, P.-R., K. Alper, and L. Pereira da Silva. 2012. Capital
Requirements and Business Cycles with Credit Market Imperfections. Journal of Macroeconomics 34 (3): 687705.
Agenor, P.-R., and P. J. Montiel. 2008. Development Macroeconomics, 3rd ed. Princeton, NJ: Princeton University Press.
Agenor, P.-R., and L. Pereira da Silva. 2010. Reforming International Standards for Bank Capital Requirements: A Perspective
from the Developing World. In International Banking in the
New Era: Post-Crisis Challenges and Opportunities, ed. S. Kim
and M. D. McKenzie, 191254. Emerald (Bingley).
. 2011. Macroprudential Regulation and the Monetary
Transmission Mechanism. Working Paper No. 254, Central
Bank of Brazil (November).
. 2012. Cyclical Eects of Bank Capital Requirements with
Imperfect Credit Markets. Journal of Financial Stability 8 (1):
4356.
Aguiar, A., and I. Drumond. 2007. Business Cycle and Bank Capital: Monetary Policy Transmission under the Basel Accords.
FEP Working Paper No. 242, Universidade do Porto (June).
Aikman, D., and M. Paustian. 2006. Bank Capital, Asset Prices
and Monetary Policy. Working Paper No. 305, Bank of England
(August).
Akram, Q. F., G. Bardsen, and O. Eitrheim. 2006. Monetary Policy
and Asset Prices: To Respond or Not? International Journal of
Finance and Economics 11 (3): 27992.
Akram, Q. F., and O. Eitrheim. 2008. Flexible Ination Targeting
and Financial Stability: Is It Enough to Stabilize Ination and
Output? Journal of Banking and Finance 32 (7): 124254.
Aliaga-Daz, R., and M. Pa Olivero. 2010. Is There a Financial
Accelerator in US Banking? Evidence from the Cyclicality of
Banks Price-Cost Margins. Economics Letters 108 (September): 16771.
Allen, F., E. Carletti, and R. Marquez. 2011. Credit Market Competition and Capital Regulation. Review of Financial Studies
24 (4): 9831018.

Vol. 9 No. 3

Capital Regulation, Monetary Policy

235

Angelini, P., S. Neri, and F. Panetta. 2011. Monetary and Macroprudential Policies. Working Paper No. 801, Banca dItalia
(March).
Angeloni, I., and E. Faia. 2010. Capital Regulation and Monetary
Policy with Fragile Banks. Unpublished, European Central
Bank (July).
Basel Committee on Banking Supervision. 2010. Basel III: A
Global Regulatory Framework for More Resilient Banks and
Banking Systems. Report No. 189 (December).
Bernanke, B., and M. Gertler. 1999. Monetary Policy and Asset
Price Volatility. Economic Review (Federal Reserve Bank of
Kansas City) Fourth Quarter (October): 1751.
Blavy, R., and M. Souto. 2009. Estimating Default Frequencies
and Macronancial Linkages in the Mexican Banking Sector.
IMF Working Paper No. 09/109 (May).
Blinder, A. S. 2010. How Central Should the Central Bank Be?
Journal of Economic Literature 48 (1): 12333.
Brunnermeier, M., A. Crockett, C. Goodhart, A. D. Persaud, and H.
Shin. 2009. The Fundamental Principles of Financial Regulation.
Geneva: International Center for Monetary and Banking Studies.
Buiter, W. H. 2008. Central Banks and Financial Crises. Discussion Paper No. 619, Financial Markets Group, London School
of Economics (September).
Calder
on, C., and J. R. Fuentes. 2011. Characterizing the Business
Cycles of Emerging Economies. Unpublished, World Bank
(January).
Cavalcanti, M. A. F. H. 2010. Credit Market Imperfections and the
Power of the Financial Accelerator: A Theoretical and Empirical
Investigation. Journal of Macroeconomics 32 (1): 11844.
Cebi, C. 2011. The Interaction between Monetary and Fiscal Policies in Turkey: An Estimated New Keynesian DSGE Model.
Working Paper No. 11/04, Central Bank of Turkey (January).
Chadha, J. S., L. Sarno, and G. Valente. 2004. Monetary Policy
Rules, Asset Prices, and Exchange Rates. IMF Sta Papers 51
(3): 52952.
Chen, N.-K. 2001. Bank Net Worth, Asset Prices, and Economic
Activity. Journal of Monetary Economics 48 (2): 41536.
Coenen, G., and R. Straub. 2005. Does Government Spending
Crowd in Private Consumption? Theory and Empirical Evidence
for the Euro Area. International Finance 8 (3): 43570.

236

International Journal of Central Banking

September 2013

Covas, F., and S. Fujita. 2010. Procyclicality of Capital Requirements in a General Equilibrium Model of Liquidity Dependence.
International Journal of Central Banking 6 (4): 13786.
C
urdia, V., and M. Woodford. 2010. Credit Spreads and Monetary Policy. Journal of Money, Credit and Banking 42 (s1):
335.
Darracq Pari`es, M., C. Kok Sorensen, and D. Rodriguez Palenzuela.
2010. Macroeconomic Propagation under Dierent Regulatory
Regimes: Evidence from an Estimated DSGE for the Euro Area.
Unpublished, European Central Bank (June).
De Graeve, F., T. Kick, and M. Koetter. 2008. Monetary Policy and
Financial (In)Stability: An Integrated MicroMacro Approach.
Journal of Financial Stability 4 (3): 20531.
Demirguc-Kunt, A., and E. Detragiache. 2005. Cross-Country
Empirical Studies of Systemic Bank Distress: A Survey.
National Institute Economic Review 192 (1): 6883.
Detken, C., and F. Smets. 2004. Asset Price Booms and Monetary
Policy. ECB Working Paper No. 364 (May).
Dib, A. 2009. Banks, Credit Market Frictions, and Business
Cycles. Unpublished, Central Bank of Canada (August).
. 2010. Capital Requirements and Financial Frictions in
Banking. Unpublished, Central Bank of Canada (June).
Djankov, S., O. Hart, C. McLiesh, and A. Shleifer. 2008. Debt
Enforcement around the World. Journal of Political Economy
116 (6): 110549.
Drumond, I. 2009. Bank Capital Requirements, Business Cycle
Fluctuations and the Basel Accords: A Synthesis. Journal of
Economic Surveys 23 (5): 798830.
Faia, E., and T. Monacelli. 2007. Optimal Interest Rate Rules,
Asset Prices, and Credit Frictions. Journal of Economic
Dynamics and Control 31 (10): 322854.
Filardo, A. J. 2004. Monetary Policy and Asset Price Bubbles: Calibrating the Monetary Policy Trade-Os. BIS Working Paper
No. 155 (June).
Financial Services Authority. 2009. The Turner Review: A Regulatory Response to the Global Banking Crisis. London: FSA.
Fonseca, A. R., F. Gonz
alez, and L. Pereira da Silva. 2010. Cyclical
Eects of Bank Capital Buers with Imperfect Credit Markets:
International Evidence. Unpublished, University of Oviedo and
World Bank (April).

Vol. 9 No. 3

Capital Regulation, Monetary Policy

237

Gerali, A., S. Neri, L. Sessa, and F. M. Signoretti. 2010. Credit and

Banking in a DSGE Model of the Euro Area. Journal of Money,
Credit and Banking 42 (s1): 10741.
Gerdesmeier, D., H.-E. Reimers, and B. Roa. 2010. Asset Price
Misalignments and the Role of Money and Credit. International
Finance 13 (3): 377407.
Goodhart, C. E. 2006. A Framework for Assessing Financial
Stability? Journal of Banking and Finance 30 (12): 3415
22.
Goodhart, C., and A. Persaud. 2008. A Proposal for How to Avoid
the Next Crash. The Financial Times, January 31.
Granville, B., and S. Mallick. 2009. Monetary and Financial Stability in the Euro Area: Pro-Cyclicality versus Trade-O. Journal
of International Financial Markets, Institutions and Money 19
(4): 66274.
Hilbers, P., I. Otker-Robe, C. Pazarbasioglu, and G. Johnsen. 2005.
Assessing and Managing Rapid Credit Growth and the Role of
Supervisory and Prudential Policies. IMF Working Paper No.
05/151 (July).
Hirakata, N., N. Sudo, and K. Ueda. 2009. Capital Injection, Monetary Policy, and Financial Accelerators. Unpublished, Bank of
Japan (October).
Iacoviello, M. 2005. House Prices, Borrowing Constraints, and Monetary Policy in the Business Cycle. American Economic Review
95 (June): 73964.
Iacoviello, M., and S. Neri. 2008. Housing Market Spillovers: Evidence from an Estimated DSGE Model. Working Paper No. 659,
Bank of Italy (January).
International Monetary Fund. 2011. Lessons for Monetary Policy from Asset Price Fluctuations. In Global Financial Stability Report, chapter 3. Washington, DC: International Monetary
Fund.
Kannan, P., A. Scott, and P. Rabanal. 2009. Monetary and Macroprudential Policy Rules in a Model with House Price Booms.
IMF Working Paper No. 251 (November).
Levieuge, G. 2009. The Bank Capital Channel and CounterCyclical Prudential Regulation in a DSGE Model. Recherches
economiques de Louvain 75 (June): 42560.
Liu, G., and N. E. Seeiso. 2012. Basel II Procyclicality: The Case
of South Africa. Economic Modelling 29 (3): 84857.

238

International Journal of Central Banking

September 2013

Medina, J. P., and C. Soto. 2007. The Chilean Business Cycle

through the Lens of a Stochastic General Equilibrium Model.
Working Paper No. 457, Central Bank of Chile (December).
Meh, C., and K. Moran. 2010. The Role of Bank Capital in the
Propagation of Shocks. Journal of Economic Dynamics and
Control 34 (3): 55576.
Mehran, H., and A. Thakor. 2011. Bank Capital and Value in the
Cross Section. Review of Financial Studies 24 (4): 101967.
Misina, M., and G. Tkacz. 2009. Credit, Asset Prices, and Financial
Stress. International Journal of Central Banking 5 (4): 95122.
Moura, M. L., and A. de Carvalho. 2010. What Can Taylor
Rules Say about Monetary Policy in Latin America? Journal
of Macroeconomics 32 (1): 392404.
Panetta, F., and P. Angelini, coordinators. 2009. Financial Sector
Pro-cyclicality: Lessons from the Crisis. Occasional Paper No.
44, Bank of Italy (April).
Repullo, R., and J. Saurina. 2011. The Countercyclical Capital
Buer of Basel III: A Critical Assessment. Unpublished, Bank
of Spain (March).
Repullo, R., and J. Suarez. 2009. The Procyclical Eects of Bank
Capital Regulation. Discussion Paper No. 2010-05S, European
Banking Center (October).
Rotemberg, J. J. 1982. Monopolistic Price Adjustment and Aggregate Output. Review of Economic Studies 49 (4): 51731.
Segoviano, M. A., and C. A. E. Goodhart. 2009. Banking Stability
Measures. IMF Working Paper No. 09/4 (January).
Silos, P. 2007. Housing, Portfolio Choice and the Macroeconomy.
Journal of Economic Dynamics and Control 31 (8): 2774801.
Wadhwani, S. 2008. Should Monetary Policy Respond to Asset
Price Bubbles? Revisiting the Debate. National Institute Economic Review 206 (1): 2534.
Woodford, M. 1999. Optimal Monetary Policy Inertia. Manchester
School 67: 135.
Zicchino, L. 2006. A Model of Bank Capital, Lending and the
Macroeconomy: Basel I Versus Basel II. Manchester School 74
(s1): 5077.