THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
INTERNATIONAL MONETARY FUND
`
THE INTERACTION OF MONETARY AND
MACROPRUDENTIAL POLICIES—BACKGROUND PAPER
Approved By
Jan Brockmeijer
This paper was prepared by a staff team led by Erlend Nier,
comprising Heedon Kang, Tommaso Mancini, Heiko Hesse (all
MCM), Francesco Columba (WHD), Robert Tchaidze (EUR), and
Jerome Vandenbussche (EUR).
CONTENTS
I. INTRODUCTION ________________________________________________________________________________ 3
II. INTERACTIONS BETWEEN MONETARY AND MACROPRUDENTIAL POLICY ________________ 5
A. Policy Interactions––Conceptual Framework ____________________________________________________ 5
B. Monetary Policy and Side Effects on Financial Stability _________________________________________ 6
C. Macroprudential Transmission and Effects on Real Economic Outcomes _____________________ 11
III. EMPIRICAL ANALYSIS _______________________________________________________________________ 16
A. Macroprudential Policies—Effects on Credit, House Prices, and Output ______________________ 16
B. Effects of Macroprudential Policy Measures—Symmetric or Asymmetric? _______________________ 24
IV. COUNTRY CASES ____________________________________________________________________________ 28
A. Selected Central, Eastern, and South-Eastern Europe Countries ______________________________ 28
B. Brazil __________________________________________________________________________________________ 33
C. Turkey _________________________________________________________________________________________ 38
D. Korea __________________________________________________________________________________________ 44
E. United States __________________________________________________________________________________ 51
REFERENCES _____________________________________________________________________________________ 60
BOX
1. Case Studies on Monetary and Macroprudential Policies _______________________________________ 4
FIGURES
1. Number of Macroprudential Measures—Tightening or Loosening ____________________________ 25
2. Selected CESEE Countries: Foreign Currency Loans and Policy Interest Rate Spreads, 2005–11 ___ 29
3. Brazil: Macroeconomic Conjuncture and Policy Responses ____________________________________ 33
December 27, 2012
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
INTERNATIONAL MONETARY FUND
4. Brazil: Credit Expansion ________________________________________________________________________ 34
5. Brazil: Impacts of RRs Tightening (1 percent) on Credit Growth _______________________________ 36
6. Brazil: Effectiveness of Changes of Capital Requirements on Consumer Loan _________________ 36
7. Brazil: Monetary and Macroprudential Policy Coordination ___________________________________ 37
8. Turkey: Credit Growth and Current Account Deficit ___________________________________________ 39
9. Turkey: Interest Rates, Reserve Requirement Ratios, and Growth of Lending _________________ 42
10. Turkey: Inflation Expectations and Inflation Rates ____________________________________________ 43
11. Turkey: Cumulative Liquidity Injections ______________________________________________________ 44
12. Korea: House Prices and Household Debts ___________________________________________________ 45
13. Korea: Monetary Policy as a Countercyclical Tool ____________________________________________ 46
14. Korea: Effectiveness of Limits on LTV and DTI Ratios _________________________________________ 47
15. Korea: External Net Assets of Banking Sector ________________________________________________ 49
16. Korea: Foreign Exchange Rates and CDS Premium ___________________________________________ 50
17. United States: Interest Rates _________________________________________________________________ 52
18. United States: Inflation Rates and GDP Growth Rates ________________________________________ 52
19. United States: Recommended Policy Rates from Baseline Taylor Rule Responding to CPI
Inflation and A Variant Responding to the GDP Deflator ____________________________________ 54
20. United States: Recommended Policy Rates from Baseline Taylor Rule Responding to CPI
Inflation and A Variant Responding To Core PCE Inflation ___________________________________ 54
21. United States: Leverage Ratio ________________________________________________________________ 56
22. United States: Net Federal Funds and Security Repo Funding to Banks and Brokers-
Dealers _______________________________________________________________________________________ 57
23. United States: Standard and Poor Composite Home Price Index _____________________________ 58
TABLES
1. Monetary Policy Effects on Financial Stability ___________________________________________________ 7
2. Use of Macroprudential Measures Across Countries __________________________________________ 19
3. Effects of Macroprudential Measures on Credit Growth _______________________________________ 20
4. Effects of Macroprudential Measures on House Price Appreciation ___________________________ 21
5. Effects of Macroprudential Measures on Output Growth and Residential Investment ________ 22
6. Effects of Macroprudential Measures on Capital Inflows ______________________________________ 24
7. Number of Macroprudential Measures—Tightening or Loosening ____________________________ 24
8. Effects of Macroprudential Policy Stance on Credit Growth ___________________________________ 26
9. Effects of Macroprudential Policy Stance on House Price Appreciation _______________________ 27
10. Selected CESEE Countries: Inflation Target, Inflation Outturn, and Policy Rates, 2006–11 ____ 28
11. Selected CESEE Countries: Use of Macroprudential Instruments Addressing Foreign Currency
Loans, 2002Q1–2012Q1 ______________________________________________________________________ 31
12. Selected CESSE Countries: Determinants of the Share of Foreign Currency Loans, 2001Q1-
2012Q1 ______________________________________________________________________________________ 32
13. Brazil: Changes of Capital Requirements on Consumer Loans _______________________________ 35
14. Turkey: Macroprudential Measures Undertaken in 2008–11 _________________________________ 41
15. Korea: Changes of Limits on LTV and DTI Ratios _____________________________________________ 48
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
I. INTRODUCTION
1. This paper provides background material to support the Board paper on the
interaction of monetary and macroprudential policies. It analyzes the scope for and evidence on
interactions between monetary and macroprudential policies. It first reviews a recent conceptual
literature on interactive effects that arise when both macroprudential and monetary policy are
employed. It goes on to explore the “side effects” of monetary policy on financial stability and their
implications for macroprudential policy. It finally addresses the strength of possible effects of
macroprudential policies on output and price stability, and draws out implications for the conduct of
monetary policy.
2. The paper then presents empirical analysis of these issues. Using cross-country data on
the use of macroprudential policy tools from 2000 to 2011 in 36 countries, the paper assesses
empirically the effects of macroprudential policy on financial variables—such as credit and asset
prices—as well as their effects on the real economy. This analysis also investigates how the effects of
macroprudential policy tools may depend on financial and economic conditions and whether the
strength of effects of macroprudential tools depends on whether the tools are tightened or
loosened.
3. The paper finally collects a number of country case studies that were prepared to shed
light on the interplay between macroprudential and monetary policies in practice. These
examine the experience in Central, Eastern, and South-Eastern Europe, Brazil, Turkey, Korea, and the
United States. Brief summaries of these case studies are in Box 1.
4. The rest of the paper is organized as follows. Section II considers theory and existing
evidence on policy interactions. Section III presents empirical evidence. Section IV assembles country
case studies on the interplay between monetary and macroprudential policy tools.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
INTERNATIONAL MONETARY FUND
Box 1. Case Studies on Monetary and Macroprudential Policies
Central, Eastern, and South-Eastern Europe. A salient feature of the experience in Central, Eastern, and
South-Eastern Europe ahead of the crisis was a pronounced increase in foreign currency (FX) lending. This
case study examines the experience of five inflation targeting countries in the region and investigates
whether interest rate spreads stimulated the increase in FX lending. It also studies macroprudential policy
responses that were taken to reduce the systemic risk associated with such lending. The study finds that
where interest rates were low relative to advanced country rates, the increase in FX lending was less
pronounced, other things equal. It also finds that the strongest macroprudential measures were effective in
counteracting the increase.
Brazil. Brazil has been an active user of both monetary and macroprudential policies. Its experience during
the post-crisis period illustrates well the complementary relationship between the two policies. Monetary
policy was used countercyclically in macroeconomic management, and macroprudential instruments were
also used to contain the potential buildup of systemic risks from rapid credit growth. As these policies
leaned against the business and financial cycle, synchronized during this period, the policy mix was
appropriate to meet two objectives—price and financial stability—with two instruments.
Turkey. In the aftermath of the global financial crisis, the Turkish authorities faced a challenging
environment, characterized by widening current account deficits, strong short-term capital inflows, and rapid
credit growth. In response, the Central Bank of the Republic of Turkey (CBRT) adopted a new “policy mix”
that emphasized financial stability objectives, while other macroprudential measures were taken only with
some delay. This case study examines the policy outcomes and points to the importance of coordination and
clear communication in responding to building financial imbalances.
Korea. During the 2000s, Korea experienced housing price boom-busts and a sharp increase of short-term
foreign currency (FX) borrowing in its banking system. While the Bank of Korea focused on price and output
stability under a flexible inflation targeting framework, financial imbalances in the housing market were
addressed with targeted macroprudential policy measures, such as limits on loan-to-value (LTV) and debt-
to-income (DTI) ratios. More recently, restrictions on FX derivative positions, and a Macroprudential Stability
Levy were brought in to curb excessive short term foreign currency borrowing. This case study shows that
such macroprudential measures have clear advantages over the use of monetary policy, which is too blunt to
deal with housing market developments and can worsen external vulnerabilities in an economy with a fully
open capital account like Korea
United States. The United States offers prime terrain to study financial instability in the years leading up to
the financial crisis of late 2007. Did an overly loose monetary policy and absence of macroprudential
measures undermine financial stability? The study finds some, though weak, evidence that interest rates
were too low relative to an optimal monetary policy response. It also finds that a relaxation of regulations
and the absence of an institutional framework geared explicitly to financial stability contributed to the
growing leverage of large investment banks, though other factors may also have been at play. Moving
forward, it will be essential to improve the effectiveness of macroprudential policies in advanced economies.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
5
5
II. INTERACTIONS BETWEEN MONETARY AND
MACROPRUDENTIAL POLICY
A. Policy Interactions––Conceptual Framework
5. Recent advances in analytical modeling offer a simple conceptual framework for
thinking about policy interactions between monetary and macroprudential policies. This
literature examines the interaction between macroprudential and monetary policy in theoretical
(DSGE) models with borrower collateral constraints and a banking sector. In these models, monetary
policy controls the risk free interest rate and macroprudential policy the risk premium, or the spread
between lending rates and the risk free rate.
1
6. A basic result is that, in the presence of macroprudential policy, it is optimal for
monetary policy to stay focused on price stability. In particular, the optimal calibration of the
reaction of monetary policy to output and inflation does not change markedly when
macroprudential policy is also used, and instead remains close to that commonly found in traditional
models without financial sector distortions or macroprudential policy.
7. In practice, macroprudential policy may not be fully effective in containing systemic
risk. The assumption made by the models is that the available macroprudential instrument is
perfectly targeted and fully offsets financial shocks. In practice, this is unlikely to be the case. For
instance, political economy considerations may limit the deployment of certain, unpopular,
macroprudential instruments, in particular when use of the instrument has strong distributional
implications. In addition, institutional arrangements may limit the frequency with which
macroprudential policy may be used, as when parliamentary or political approval is required to reset
an instrument.
8. As a result, monetary policy may still need to respond to financial conditions. Indeed,
in models where macroprudential policy is absent or time invariant, but in the presence of financial
sector distortions, it is optimal for monetary policy to consider financial shocks.
2
In such contexts,
optimal monetary policy responds to the growth in credit (in addition to the output gap and
deviations of inflation from target).
3
By extension, when macroprudential policy is imperfectly
targeted, it can be desirable for monetary policy to respond to financial conditions.
1
This literature includes Baillu and others (2012), Kannan and others (2009), Unsal (2011), Angelini and others (2011),
Bean and others (2010), Christensen and others (2011), and Cecchetti and Kohler (2012).
2
See, for example, Woodford (2011). Kannan and others (2009), as well as Christensen and others (2011), also find
that optimal monetary policy responds to credit when macroprudential policy is switched off.
3
Where there is a response to the credit gap, the optimal sensitivity parameters of monetary policy to the output
gap and deviations of inflation from target do not change markedly.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
6
6
INTERNATIONAL MONETARY FUND
9. In these models, having two policies to achieve both price and financial stability
enhances welfare. Moreover, in practice, especially when capital accounts are fully open, achieving
both objectives with one instrument may not be feasible. For example, when capital inflows
appreciate the currency and lead to imbalances, increases in policy rates cannot reduce incentives
for (foreign exchange) wholesale funding or credit expansion. Conversely, the literature shows that
when monetary policy is constrained or absent, using macroprudential policy in the place of
monetary policy to control output and inflation is inefficient and costly, as it severely constrains the
financial sector and output.
4
10. More generally, the literature points to synergies, rather than conflicts, even if the
optimal policy mix can vary with the type of shock hitting the economy. In the presence of a
financial shock, most models imply that only macroprudential policy should be used since it is more
targeted at the distortion.
5
In the presence of aggregate demand (preference) shocks that induce an
increase in both credit and inflation, both policies are tightened, complementing each other in
responding to the shock. In the presence of productivity shocks, conflicts can arise since a positive
supply shock can lead asset prices and credit demand to rise but dampens goods market inflation.
The optimal policy mix then depends on the strength of the externality from increases in credit to
aggregate financial risks. If this externality is strong, the accommodative monetary policy response
to the productivity shock is complemented by targeted macroprudential policy to contain the build-
up of leverage that may be induced by the shock.
6
11. While structural models offer clear insights into policy interactions, their downside is
their simplicity. This includes the abstraction in the majority of cases from modeling the side
effects of monetary policy on financial stability, described in the next section; and the lack of realistic
modeling of the transmission of macroprudential instruments to financial and output stability, which
is explored further below. Moreover, the adaptation and calibration of models to country
circumstances is often yet to be undertaken and hindered empirically by limited (cross-country)
experiences with both policies.
B. Monetary Policy and Side Effects on Financial Stability
12. It has long been understood that monetary policy can affect financial stability. This
section offers a taxonomy of these––beneficial or adverse––effects. It also considers the factors that
may impact the strength of the effects and explores how well-designed macroprudential policies
have the potential to contain the adverse effects of monetary policy on financial stability.
4
See, e.g., Unsal (2011).
5
As in Baillu and others (2012), Kannan and others (2009), Unsal (2011), Angelini and others (2011), and Bean and
others (2010).
6
As in Christensen and others (2011).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
7
7
13. Building on the financial markets’ imperfections literature, there are a number of
channels by which monetary policy can affect financial stability. It can affect the tightness of
borrowing constraints and likelihood of default; the risk seeking incentives of intermediaries; and
externalities operating through aggregate price variables, such as asset prices and exchange rates.
Table 1 shows for each channel the prediction from theoretical models of the effects of changes in
the monetary policy stance on financial stability. It also summarizes related empirical evidence.
7
An
appendix in the main paper reviews the empirical evidence in more detail.
Table 1. Monetary Policy Effects on Financial Stability
1/
Sources of
Financial
Instability
Channel
Predicted Effect
( improves stability)
Selected Empirical Evidence
↓r ↑r
Borrowing
Constraints
Balance Sheet
(default)
Channel
Sengupta (2010)
Jiménez and others (2009)
Gertler and Gilchrist (1994)
Asea and Blomberg (1998)
↑r ,
↑r ,
↑r ,
↑r ,
Risky
Behavior of
Financial
Institutions
Risk-taking
Channel
Jiménez and others (2009)
Ioannidou and others (2009)
Merrouche and Nier (2010)
↓r ,
↓r ,
X
Risk-shifting
Channel
Gan (2004)
Landier and others (2011)
↑r ,
↑r ,
Externalities
through
Aggregate
Prices
Asset price
Channel
Altunbas and others (2012)
Del Negro and Otrok (2007)
IMF (2009)
↓r ,
↓r ,
X
Exchange rate
Channel
Hahm and others (2012)
Merrouche and Nier (2010)
Jonsson (2009)
↑r ,
↑r ,
↑r ,
Source: IMF.
1/ ↓r means a decrease of policy rates, ↑r means an increase of policy rates, “ ” means a decline of stability, “ ” an
improvement, and “X” no statistically significant effect.
14. Changes in the monetary stance can affect the tightness of borrowing constraints and
the likelihood of default. Monetary easing relaxes collateral constraints, mitigating financial
distortions both on the demand and supply side of credit. Conversely, a tightening of rates can
adversely affect borrowers’ quality, leading to higher default rates and potentially precipitating a
crisis (Allen and Gale, 2000; Illing, 2007, Goodhart and others, 2009).
15. Changes in the monetary stance can affect the risk-seeking behavior of financial
intermediaries in multiple ways. Two channels may move in opposite directions, as follows:
7
These channels are formalized in theoretical work, but empirical evidence on these effects faces challenges,
including the absence of a counterfactual path for monetary policy and difficulties in telling apart the effects on the
demand and supply of credit. Most of the papers cited in Table 1 can be interpreted as indirect evidence.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
8
8
INTERNATIONAL MONETARY FUND
Risk-taking. Low monetary policy rates can create incentives for banks to expand their balance
sheets and reduce efforts in screening borrowers (Borio and Zhu, 2008). They can also lead other
agents to seek more risks in order to achieve higher returns (Rajan 2006). These effects are likely
to be worse if monetary policy is (too) accommodative for too long during expansions. If
monetary policy is expected to be lowered during recessions to support the financial system,
this may create additional incentives to correlate risks (Farhi and Tirole, 2012).
Risk-shifting. Increases in policy rates can reduce intermediation margins, and lead lenders,
especially poorly capitalized intermediaries, to seek more risk (Bhattacharya, 1982). This channel
may be stronger just ahead of a crisis, when intermediary leverage is high and competition limits
the pass-through of policy rates to lending rates. More generally, a flattening of the yield curve
associated with increases in policy rates can lead banks to seek risk in order to maintain profits
(Merrouche and Nier, 2010).
16. Monetary policy can affect externalities operating through aggregate financial prices,
including asset prices and exchange rates. By affecting asset prices and exchange rates, monetary
policy affects the value of collateral, which influences the tightness of borrowing constraints.
Asset prices. Low interest rate can increase asset prices, which can trigger further increases in
leverage and lead to asset price booms, exacerbating the financial cycle (Bernanke and Gertler,
1989). Conversely, a tighter monetary stance can cause collateral constraints to bind, fire sales to
follow, with resulting adverse asset price externalities (Shin, 2005).
Exchange rates. In open economies, interest rate increases can attract capital flows,
appreciating the currency, leading to excessive borrowing in foreign currency and laying the
ground for exchange rate externalities during the depreciation phase (Bruno and Shin 2012,
Hahm and others, 2012).
17. The intensity of these effects can depend on the point in the financial cycle. As financial
imbalances build up, low monetary policy rates reduce current defaults, but can induce banks to
make riskier loans and increase leverage. When rates are increased close to the peak of the financial
cycle, this can induce risk-shifting and borrower defaults. Moreover, incentives to correlate risks due
to the expectation of future monetary easing can be stronger in the upswing of the financial cycle.
18. The strength of the effects can also depend on financial structure and capital account
openness. For example, securitization generally reduces the strength of the effects of monetary
policy on credit extension by banks (e.g., Altunbas, and others 2012). But the importance of risk-
taking and risk-shifting channels may not diminish, since they come to work through both banks
and non-banks. Moreover, in open and financially-integrated economies, domestic monetary policy
has a weaker influence over domestic long-term rates and asset prices, but exchange rate
externalities become more important.
In open economies, high policy rates can encourage capital inflows and foreign exchange
borrowing. In a number of countries in emerging Europe, foreign exchange (FX) lending to
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
9
9
households increased ahead of the crisis, with tighter monetary policy aggravating the situation,
as it provided further incentives for borrowing in FX (case study, Section IV). At the same time,
when the central bank lowers rates to support the economy in a downturn, this can lead to a
depreciation and worsen exchange rate externalities arising from tightening constraints.
As international financial integration increased over the past decades, domestic monetary
control has weakened for both advanced countries and emerging markets. Bernanke (2005)
argued that a global saving glut reduced long-term rates in advanced economies, and that as a
result the relationship between short rates and long rates had become weak (Greenspan
conundrum), thereby reducing the pass-through of policy rates to asset prices.
For emerging economies, similarly, the correlation between domestic short and long rates has
weakened (Moreno, 2008) and the importance of foreign factors strengthened. Increased cross-
border banking contributes to these effects since for cross-border banks, global monetary
conditions seem to matter more than local conditions (Cetorelli and Goldberg, 2012; Shin, 2011).
19. First principles suggest that well-targeted macroprudential policies have the potential
to contain the undesirable effects of monetary policy. Where the side effects of monetary policy
on financial stability are expected to be undesirable, this can create conflicts between financial and
price stability objectives. Appropriate macroprudential policies can attenuate these side effects,
thereby reducing policy dilemmas and creating additional “room for maneuver” for monetary policy.
For most of the channels discussed above, a range of specific macroprudential instruments may
reduce the effect when brought in ex ante.
The impact on defaults from a tightening of monetary policy can be contained by
macroprudential tools, such as a limit on DTI. A conservative DTI ratio may reduce the effect of
increases in policy rates on debt affordability, thereby lessening an unwanted transmission of
increased policy rates to household default rates (Igan and Kang, 2011). This in turn can help
protect bank balance sheets and reduce the force of fire-sale dynamics for asset-backed
securities.
A range of macroprudential measures can affect the risk-taking channel. Increases in capital
requirements or a tight leverage ratio can help contain increases in bank leverage in response to
low policy rates and reduce the incentives to take risk (Farhi and Tirole, 2012). Such measures
also create additional buffers to absorb risks from an erosion of lending standards. However,
where shadow banks are important providers of credit, macroprudential tools are needed that
control leverage both inside and outside of the banking system (United States case study,
section IV). The regulation of margin in securities lending is an example (Kashyap and others,
2010).
Risk-shifting incentives associated with increases in policy rates could also be addressed through
appropriate macroprudential tools. Liquidity measures, such as the Basel Net Stable Funding
ratio, encourage banks to seek stable and longer-term funding. Where funding is longer term,
this can reduce the impact of a monetary policy tightening on lending margins and profits and
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
0
0
INTERNATIONAL MONETARY FUND
attenuate the incentive for intermediaries (banks and non-banks) to seek further risk exposure in
response to increases in policy rates.
8
Capital buffers can also reduce risk-shifting incentives
from a compression of margins (Bhattacharya 1982, Hellmann and others, 2000).
When accommodative monetary policy drives up asset prices, macroprudential measures, such
as limits on LTV ratios, can tame house price boom-busts. When low policy rates encourage
borrowing and greater credit in turn drives up asset prices, a lower LTV ratio can counter this
effect. Some studies have found that a conservative LTV ratio can contain the feedback loops
between credit and house prices (IMF, 2011b). Moreover, studies have found that a tightening of
LTV ratios can slow the rate of house price appreciation, thereby reducing the potential for a
housing bubble to emerge (Crowe and others, 2011; Igan and Kang, 2011; Wong and others
2011).
The policy dilemma that may arise from interactions between domestic monetary policy and
capital flows can be addressed by macroprudential measures. Macroprudential measures can
affect gross flows and help change the composition of flows away from short-term and FX
denominated liabilities issued by banks, thereby reducing the systemic risk associated with
capital flows (Hahm and others, 2012).
9
Examples are FX reserve requirements (RRs)
implemented in Romania and the levy on FX denominated non-core liabilities introduced in
Korea. In addition, where high domestic rates encourage corporations or households to borrow
in FX, macroprudential measures can reduce heightened default risks, including higher risk
weights and tighter LTV ratios, as well as limits on FX lending, as applied in a number of
countries in emerging Europe (case study, section IV).
20. In sum, in its transmission, monetary policy can interact with financial distortions in
several ways, with the net effect on financial stability often ambiguous. Several channels may
be at work, operating simultaneously with their strength varying with the stage of the financial cycle,
financial structure, and other country characteristics. Where the side effects are expected to be
undesirably strong, well-designed macroprudential policies that are brought in ex ante can
attenuate these effects.
8
English and others (2012) find empirically that a flatter yield curve is associated with lower net interest margins, with
the size of the effect increasing in the maturity mismatch between bank assets and bank liabilities.
9
As long as the measures do not affect the size of the net inflow, overall leverage may continue to rise.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
1
1
1
1
C. Macroprudential Transmission and Effects on Real Economic Outcomes
21. This section presents a closer examination of existing literature on the transmission of
macroprudential policies and traces out implications for monetary policy. This analysis focuses
on tools that have traditionally been used most frequently in emerging markets and those that are
likely to be used most actively in future, in both advanced and emerging economies, such as
dynamic capital buffers, limits on loan-to-value ratios and RRs.
10
For each of these tools, the paper
traces out the transmission to reduced systemic risk, as well as real economic outcomes, and draws
out implications for monetary policy.
11
Capital buffers
22. The main objective of an increase in the dynamic capital buffer is to increase the
resilience of the banking system. The idea is that when credit grows strongly, the quality of the
credit portfolio is likely to deteriorate, increasing the likelihood of future losses. When high credit
growth triggers an increase in the dynamic capital buffer in good times, the buffer can cushion the
effect of losses on bank balance sheets and thus help maintain the flow of credit when losses
materialize.
23. The effect of increases in the dynamic capital buffers on aggregate credit is likely to be
weak, in principle. When credit growth is strong, banks have ample profits that can be used to
build up the buffer through retained earnings. In addition, theory suggests that asymmetric
information and the resulting adverse signaling effects are among the main reasons for banks’
reluctance to issue new equity (Myers and Majluf, 1984; Kashyap and others, 2010) or to cut
dividend payouts (Bhattacharya, 1979). These effects are likely to be weak in good times. Moreover,
since the increase in capital is mandated, adverse signaling effects from a decrease in the dividend
payout ratio or from issuing new equity are likely to be small. On the other hand, since banks must
fear losing profitable business if they increase lending rates or cut exposures outright, they may be
less likely to pursue these strategies to meet the capital buffer.
24. In practice, increases in the buffer may still reduce credit and output for a number of
reasons, which may need to be offset by monetary policy. First, if the increase in the buffer is
brought in fast, banks will not be able to accumulate it through retained earnings alone. Second,
further distortions, such as the tax benefits of equity may make banks reluctant to issue new equity.
In some countries, in addition, banks may not have easy access to capital markets, or are privately or
cooperatively held, making it difficult for these banks to issue new equity. In the presence of such
distortions, some effect on aggregate credit and output is to be expected. Existing evidence
suggests, however, that the effects on output of increases in capital requirements may be relatively
10
The analysis focuses on macroprudential tools whose benefit is seen as containing a “time dimension” of systemic
risk, or the risk of “procyclical” increases in the risk of financial instability.
11
CGFS (2012) provides further analysis of the transmission of macroprudential tools.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
2
2
INTERNATIONAL MONETARY FUND
modest (BIS 2010, Jimenez and others 2012). This implies that any dampening effect exerted by an
increase in the buffer can be countered by more accommodative monetary policy, if necessary, as
long as monetary policy is effective.
12
25. On balance, stronger effects are likely in bad times when the accumulated buffer helps
sustain the provision of credit to the economy. Existing evidence points to stronger effects of
capital buffers on credit in bad times. Nier and Zicchino (2008) find that a larger capital buffer
mitigates the adverse effect of loan losses on loan growth and that this effect is stronger in crisis
times. Jimenez and others (2012) show that the effects of varying dynamic provisions on credit in
Spain were much stronger in crisis times than they were ahead of the crisis. New results presented in
Section III are also consistent with stronger effects on credit in bad times.
26. Where a dynamic capital buffer is in place, this will therefore reduce the need for
monetary policy makers to offset the effects of tighter credit conditions on output. In
response to a tightening of the availability of credit from October 2008, many advanced country
central banks cut interest rates aggressively in an effort to support the financial system. When
capital buffers are built up ahead of the downturn, the buffers can help sustain the provision of
credit to the economy and reduce the depth of the downturn. The presence of a dynamic capital
buffer may then lessen the risk that monetary policy runs into the constraints posed by the lower
bound on nominal rates and complement monetary policy in bad times, resulting in a smoother
path of monetary policy through the cycle.
27. In addition, when capital buffers have been built up in the upswing of the financial
cycle, the buffers may help keep open the transmission of monetary policy. In the absence of
sufficient buffers, the erosion of capital may lead banks to reduce the supply of credit to the
economy. Even where policy rates are lowered aggressively, this may not be enough to counter
banks’ reluctance to lend. A bigger capital buffer that banks are allowed to run down can help
unblock the transmission of monetary policy to the provision of credit (Turner, 2012).
Loan-to-value ratios
28. Limits on loan-to-value (LTV) and debt-to-income (DTI) ratios are increasingly being
viewed as useful to contain potentially damaging boom-bust cycles in residential housing
markets (Igan and Kang, 2011; IMF, 2011b; Crowe and others, 2011). An LTV ratio imposes a cap on
the size of the loan relative to the value of the property, thereby imposing a minimum down
payment. In principle, even a static, but conservatively calibrated LTV ratio can strongly affect house-
price dynamics. Its effect can be enhanced when the calibration is varied with cyclical conditions in
the housing market (as in Korea and Hong Kong SAR), or when it is complemented with a DTI ratio
12
Complications can arise when capital requirements are tightened in bad (crisis) times, and when monetary policy is
already close to its lower bound. In this case, the risk of deleveraging is greater and cannot easily be countered by
monetary policy.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
1
1
3
3
(as in Poland and many other countries). A DTI ratio caps total loans to a fixed multiple of household
income and may help contain unsustainable increases in household debt more broadly.
29. These measures have been found successful in containing house price accelerations in
the upswing. Limits on LTV ratios can reduce financial accelerator mechanisms that otherwise lead
to a positive two-way feedback between credit and house prices. A number of cross-country studies
have found that positive shocks to household income or the size of the population translates into
larger house price increases where prevailing leverage ratios are higher (Almeida and others 2005;
IMF, 2011b). Moreover, a number of studies have found that a tightening of LTV ratios can slow the
rate of house price appreciation, thereby reducing the potential for a housing bubble to emerge
(Igan and Kang, 2011; Wong and others, 2011; Crowe and others, 2011). For example, Crowe and
others (2011) find that a 10 percentage point tightening of the LTV ratio leads to a decline in the
rate of house price appreciation of between 8 and 13 percentage points.
30. Where these measures limit house price acceleration and household indebtedness,
they may also dampen the associated increases in aggregate demand, in turn modifying
optimal monetary policy. LTV and DTI measures may reduce the response of residential
investment and household consumption to positive financial shocks. When they reduce the strength
of financial accelerator mechanisms in the upturn, this may allow monetary policy to be somewhat
looser than in the absence of these measures (IMF, 2008).
31. A growing body of evidence also points to the benefit of LTV and DTI ratios in
containing the severity of the property bust when the housing market turns.
In theory, where leverage is high, even a relatively small fall in house prices may lead borrowers
to become underwater. This creates incentives to default strategically, which in turn imparts
further downward pressure on prices. Consistent with this, IMF (2011b) show that across OECD
countries over the 1980 to 2010 period, conditional on a housing bust occurring, the fall in
property prices is less steep where LTV ratios are tight.
A housing bust can put stress on financial intermediaries’ engaged in mortgage credit, and tight
LTV ratios can reduce these impacts. High rates of default can reduce profitability and deplete
banks’ capital cushions. Wong and others (2011) document that, for a given fall in house prices,
the incidence of mortgage default is higher for countries without a LTV ratio limit than it is for
countries with such a tool. They also show that losses sustained by lenders for a given fall in
house prices are lower.
Stress on financial intermediaries can lead to a contraction of mortgage credit and credit more
broadly, adversely affecting both household consumption and business investment. Based on
1960–2007 cross-country data, Claessens, Kose, and Terrones (2008) show that output losses in
recessions accompanied by housing busts are two to three times larger than otherwise.
Moreover, housing busts tend to prolong recessions, as falling house prices act as a further drag
on household consumption and residential investment, while putting financial intermediary
balance sheets under stress.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
4
4
INTERNATIONAL MONETARY FUND
32. By reducing the depth and duration of the downturn, limits on LTV and DTI ratios can
also lessen the risk that monetary policy will run into its lower bound. When tight LTV and DTI
ratios contain mortgage defaults and losses sustained by financial intermediaries in the wake of a
fall in house prices, this can also reduce output losses from the property bust. A milder downturn
can in turn reduce the need for monetary easing that would otherwise be necessary to counter the
financial headwinds from the property bust. IMF (2008) shows formally that offsetting the
deflationary impact of a negative financial shock requires a larger accommodative monetary policy
response in an economy with a high LTV ratio and a smaller response where LTV ratios are tight.
33. These complementarities are further strengthened by the effects of LTV and DTI ratios
on the transmission of monetary policy. As seen in many advanced countries since the crisis
broke, when a large fraction of borrowers have high LTV mortgages, this can clog up the
transmission of lower policy rates on conditions in mortgage markets after the bust. After a fall in
house prices, high LTV borrowers will find themselves unable to refinance their loans since the
principal exceeds the value of their property. These borrowers will then not be able to take
advantage of lower mortgage rates that an easing of monetary policy may help bring about. A
tighter LTV constraint going into the property bust can mitigate this and help strengthen the
transmission of monetary policy after a house price falls (Geanakoplos, 2010).
34. Limits on DTI and LTV ratios can also affect developments in the composition of
output that are not easily controlled by monetary policy. First, a tightening of these measures
can, by slowing housing transactions and dampening house price growth, reduce the share of
residential construction in GDP (see Section III for empirical evidence). Second, to the extent that
these measures contain increases in household leverage, they can help control a rise in debt-
financed consumption spending that worsens the current account. Along with other policy measures
(such as structural and fiscal policies), LTV and DTI policies may thus contribute to a reduction of
external imbalances.
Reserve requirements
35. Central banks can use variations in the level of RRs to affect broader credit conditions.
When RRs are remunerated below the policy rate or are unremunerated, a variation in the level of
the requirement imposes a tax on lending. This tax is expected to increase the spread between
lending and deposit rates as banks pass on increased costs to their customers (Gray, 2011; Tovar
and others 2012; Glocker and Towbin, 2012a). Independent of the incidence, since the tax will lead
to a fall in deposit supply, or a contraction of loan demand, or both, the amount of credit provided
to the economy is expected to fall. By increasing the spread between lending and deposit rates, RRs
will then lower the amount of credit provided to the economy, acting as a “speed limit.”
36. Inflation targeting central banks will typically offset the impact on banking system
liquidity and interbank interest rates of a change in RRs. The volume of open market operations
can be adjusted to offset the impact on banking system liquidity and to keep interbank rates close
to the target rate. But even if the monetary effect of changes in RRs is sterilized, there can be a
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
1
1
5
5
macroprudential effect, which works through an increase in the spread between lending and deposit
rates. See further the case studies for Brazil and Turkey in Section IV of this paper.
37. Empirical studies tend to find evidence in support of an effect on credit. Vargas and
others (2011) study the experience in Colombia and find that RRs have a strong and lasting effect on
lending rates charged on business loans. Glocker and Towbin (2012b) estimate a structural vector
autoregressive (VAR) model for the Brazilian economy. They find that a one percentage point
increase in the RRs leads to a peak increase in the spread between lending and deposit rates of
80 basis points. Moreover, domestic credit falls on impact by about 2.5 percent and remains below
trend for close to two years. The empirical exercise reported in Section III is also consistent with a
significant effect on credit.
38. However, increases in RRs do not increase resilience and can have unintended side
effects. Unlike an increase in capital requirements, an increase in RRs has no impact on the
resilience of the banking system to loan losses. In addition, an increase in RRs can exacerbate risk-
shifting incentives. When RRs squeeze profitability this can lead banks to shift into higher margin,
but higher risk segments, in an effort to restore return on equity. In Turkey, for instance, relatively
aggressive increases in RRs in early 2011 may have further spurred banks’ consumer lending, which
was ultimately addressed by increases in regulatory risk weights on such lending (Turkey FSAP and
case study).
13
39. An increase in RRs can lead to nominal depreciation and affect capital inflows. In small
open economies, increases in the monetary policy rate will tend to attract capital inflows and lead to
an appreciation of the currency. Increases in RRs tend to have the opposite effects, since they will
tend to decrease returns on domestic and FX deposits (Glocker and Towbin, 2012a).
14
Using data
from Brazil, Glocker and Towbin (2012b) find that an increase in the reserves requirement by one
percentage point leads to a 2 percent depreciation of the domestic currency. Evidence presented in
Tovar and others (2012) confirms the effects of RRs on exchange rates, even if their results point to a
more transitory depreciation. New evidence presented in Section III suggests that RRs have the
potential to affect the composition of capital inflows, away from bank portfolio flows.
40. The effects of increases in RRs on output are ambiguous in theory. The rise in bank
lending rates should tighten credit and lead to a decline in investment spending. However, the fall in
deposit rates may decrease domestic savings and increase consumption. Moreover, any
depreciation resulting from the increase in RRs would lead to an increase in net exports that boosts
aggregate demand. This implies that while an increase in RRs unambiguously lowers aggregate
13
These adverse effects of increases in RRs contrast with bank capital requirements. An increase in the latter leaves
banks’ return on total assets unaffected and in general reduces rather than increases banks’ incentives to take risk.
14
Since an increase in RRs will lead to a decline in deposit rates, under uncovered interest parity, net capital inflows
will fall. An alternative explanation is that the tax reduces total expected return for foreign investors (Gray 2011).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
6
6
INTERNATIONAL MONETARY FUND
credit, its net effect on output may be relatively small (Glocker and Towbin, 2012a).
15
Empirical
results presented in Section III suggest that there is no measurable effect of an increase in RRs on
output.
41. RRs provide a potential way to curb excessively strong credit growth, while effects on
other economic variables are quite different from that of monetary policy. In contrast to
increases in policy rates, an increase in RRs can reduce excessive credit growth without attracting net
capital inflows and appreciating the exchange rate. Moreover, when increases in RRs dampen capital
inflows, this can give greater room for maneuver for monetary policy to increase interest rates, as
has been the experience in Peru (Tovar and others 2012).
42. Equally, in economic downturns, a relaxation of RRs can stimulate credit growth
without this leading to a depreciation of the exchange rate or capital outflows. This contrasts
again with the effects of an easing of the monetary policy rate, which is likely to contribute to a fall
in the currency and capital outflows, especially in bad times (Federico, Vegh, and Vuletin, 2012).
43. In sum, the transmission of macroprudential policy tools and the implication for the
conduct of monetary policy may differ with the tools considered.
Some tools, including the dynamic capital buffer and limits on LTV ratios, increase the resilience
of the economy against aggregate shocks, mitigating the effects on output of a credit crunch
and housing bust. This can reduce the need to for accommodative policy in such scenarios and
makes it less likely that monetary policy will hit the constraint imposed by the lower bound. It
can also help keep open monetary transmission channels in a downturn scenario.
When capital requirements and LTV ratios are tightened in upturns, they may have effects on
credit and asset prices, and hence, potentially on aggregate output. Where these effects are
sizable, they can be offset, as necessary, by more accommodative monetary policy.
RRs may be a useful complement to monetary policy, especially in open economies, since use of
this tool can control credit growth. This can give greater room for maneuver for monetary policy
in the face of potentially destabilizing capital flows. The effect of an increase in RRs on output is
ambiguous, though, and empirically found to be small.
III. EMPIRICAL ANALYSIS
A. Macroprudential Policies—Effects on Credit, House Prices, and Output
44. This section describes empirical analysis of the effects of macroprudential policy
measures. Several empirical studies show that macroprudential policy instruments can be effective
15
Empirically, Glocker and Tobin (2012b) find that increases in RRs increase unemployment in Brazil, but that this
effect is half that from an equivalent variation in the monetary policy rate.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
1
1
7
7
in addressing systemic risk externalities, if used appropriately. Lim and others (2011) find that
macroprudential instruments may reduce the correlation between credit growth and GDP growth,
and several studies show that limits on LTV and DTI ratios can curb the feedback loop between
mortgage credit availability and house price appreciation. However, few studies consider differential
effects across macroprudential tools. Moreover, there is to date only a very limited analysis of any
macroeconomic effects of the use of macroprudential tools.
45. Our analysis expands on the existing literature in several ways.
First, the direct effects of macroprudential measures on financial outcomes—credit growth and
housing price appreciation—are tested.
Second, we perform analysis of “side effects” of the macroprudential tools on economic growth,
residential investment, and capital inflows.
Third, we investigate whether the strength of these effects differs with measures of the
economic and financial cycle.
Model specification and data
46. We focus on (varying) capital requirements (CR), limits on LTV ratios, caps on DTI
ratios, and RRs. For each macroprudential instrument, an index variable is constructed. This index
increases by 1 whenever an instrument was introduced or tightened and decreases by 1 whenever
the instrument is loosened, resulting in a series that looks like a step function. The index variable
captures both semi-quantitative effects and the average treatment effects of the instrument.
Countries and periods in which instruments are not used are included as counterfactuals and help in
identifying the effects of key control variables. The information required on the use of the
instruments is obtained and extended from the 2010 IMF survey (Lim and others, 2011).
16
47. A fixed-effect dynamic panel regression is used with the following specification:
∆
,
·
,
·∆
,
·
,
··
,
,
For each country ,
,
and
,
represent changes of outcome variables and a time-series index
of one of the four macroprudential measures respectively, where the coefficient captures the
effects of macroprudential measures on the outcome.
,
denotes a set of control variables and
·
,
is included to capture the interaction between the control variables and the
macroprudential instruments. Throughout, we include time-fixed effects, to account for common
variation across countries, as well as country-fixed effects, to account for time-invariant country-
16
We would like to acknowledge Ivo Krznar’s contribution to this section. The data comes mostly from Krznar and
others (forthcoming) and the regressions extend the framework in Arregui and others (forthcoming).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
1
1
8
8
INTERNATIONAL MONETARY FUND
characteristics. Our regressions also control for monetary policy rates and dummy variables
denoting phases of credit and economic cycles.
17,18
48. The measurement of the effects of policy changes on both financial and aggregate
variables is subject to well-known endogeneity problems. This issue is shared by most existing
studies on the effects of macroprudential policy study (e.g., Lim and others, 2011). When
macroprudential policy responds to credit and asset prices, rather than output, this bias should in
principle be stronger when measuring the effect on credit and asset prices—as does much of the
existing literature—and weaker when investigating aggregate effects—which is the focus here.
Moreover, as long as the problem does not differ across tools considered it may not affect
comparisons across tools in their relative effects. Throughout, we lag all policy variables by one
quarter in an attempt to address endogeneity biases.
19
Nonetheless, we take the results as only
suggestive of the relative strength of the effects across tools, rather than as a reliable measure of
the size of each effect.
20
49. Quarterly data from 2000 to 2011 were used for 36 countries,
21
including 21 emerging
market economies (EMEs) and 15 advanced economies (AEs).
22
Most of the data are collected
from official and commercial sources, such as IFS, central banks, national statistical offices, Haver
Analytics, and CEIC database, being specified along the results. Detailed information on countries
which have used macroprudential policies can be found in Table 2.
17
Variables in the form of dummies are constructed to control for the stages of financial and economic cycles. First, a
credit bust is classified as a stage with either of the two following conditions being satisfied: (i) the deviation from a
HP filtered trend is smaller than 1.5 times its standard deviation; or (ii) the quarterly credit growth rate is lower than a
long-run average by 1.5 times its standard deviation. Second, a recession dummy is equal to one on the quarters
when the output gap, using the HP filter, is negative for previous six consecutive quarters.
18
In addition, interactions between monetary and macroprudential policies are analyzed, but no significant empirical
evidence is found. We created dummy variables indicating whether the monetary stance is tight, or whether it is
loose and estimated interactions between the macroprudential tools and the monetary policy dummies. This result is
in line with results obtained by Dell’Ariccia and others (2012). Coefficients on the interaction terms are unstable and
rarely significant across all macroprudential tools. Similarly we do not find that effectiveness of macroprudential
policy depends on the monetary and FX regime, echoing results already documented by Lim and others (2011).
19
In order to try to reduce the endogeneity problem, one quarter lagged policy variables are used for the main
results, and a robust test is conducted with concurrent variables, which shows similar results.
20
As mentioned in Lim and others (2011), the estimation of a dynamic panel by OLS with country and time fixed
effects will be biased, since by construction there is a positive correlation between the lagged dependent variable
and the unobserved individual level effects. We dropped the lagged dependent variable as a robustness check, and
found that the main results do not change materially.
21
The countries in the sample are as follows: Argentina, Austria, Brazil, Bulgaria, Canada, Chile, China, Colombia,
Croatia, Estonia, Hong Kong SAR, Hungary, India, Indonesia, Ireland, Israel, Italy, Latvia, Malaysia, Mexico,
Netherlands, Norway, Peru, Poland, Romania, Russia, Serbia, Singapore, Slovak Republic, South Korea, Spain, Sweden,
Thailand, Turkey, Uruguay, and U.S.
22
As of September 2012, the number of countries in the sample with each de facto exchange rate regime is as
follows: free floating (10), floating (15), fixed (3) (of which currency board (2) and conventional peg (1)), and others (8)
(of which crawl like (3), other managed (4), and stabilized (1)).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
1
1
9
9
Table 2. Use of Macroprudential Measures Across Countries
Sources: Lim and others (2011) and IMF staff.
Advanced Economies Emerging Market Economies
(Free)
Floating
Fixed
Other
Managed
(Free) Floating Fixed
Stabilized or
Other
Managed
Capital
requirement
Estonia,
Israel,
Korea
Ireland,
Spain
—
Argentina, Brazil,
Mexico, India,
Thailand, Turkey
Bulgaria
China,
Malaysia,
Croatia
Limits on LTV
ratio
Canada,
Norway,
Korea,
Sweden
Hong Kong
SAR,
Netherlands
Singapore
India, Thailand,
Hungary,
Romania, Turkey
Bulgaria,
Latvia
China,
Malaysia
Caps on DTI
ratio
Canada,
Korea,
Norway
Hong Kong
SAR
—
Thailand, Poland,
Romania,
Hungary, Serbia
— —
Reserve
requirements
Korea — —
Argentina, Brazil,
Chile, Colombia,
Peru, Uruguay,
India, Indonesia,
Romania, Serbia
Bulgaria
China,
Croatia,
Russia
Results—effect on financial variables
50. Investigating the effects on credit growth, we find statistically significant effects for
both (varying) capital requirements and RRs (Table 3). For capital requirements in particular, we
find that the effect on credit growth is stronger during credit busts. For a subsample containing
EMEs only, we find that limits on LTV and DTI ratios are also associated with lower credit growth. In
this subsample, the effect of RRs is little changed relative to the full sample. The effect of capital
requirements on credit growth weakens, but remains stronger during credit busts.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
0
0
INTERNATIONAL MONETARY FUND
Table 3. Effects of Macroprudential Measures on Credit Growth
1/
Source: IMF staff estimates.
1/ Green, orange, and yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively.
51. When investigating effects on house price appreciation rates we find statistically
strong effects for limits on LTV ratios and capital requirements, but not for RRs. This suggests
that in our sample, variation in capital requirements might have been specifically targeted at
housing credit in a number of countries, such as higher risk weights on mortgage loans, while RRs
by construction do not differentiate between asset classes, and are therefore less likely to have an
effect on a particular asset price. Interestingly, the effects of macroprudential tightening (or
loosening) on house prices is estimated stronger during recessions across most measures, while for
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
2
2
1
1
house prices the stage of the credit cycle appears to play less of a role. Throughout, the evidence for
direct effects of variation in the DTI ratio on house prices is in general not as strong (Table 4).
Table 4. Effects of Macroprudential Measures on House Price Appreciation
1/
Source: IMF staff estimates.
1/ Green, orange, and yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
2
2
INTERNATIONAL MONETARY FUND
Results—effect on macro variables
52. Turning to the effects on output, the results of the main panel regressions suggest
that limits on LTV ratio have an impact on output growth, and that this may work through a
negative impact on investments in construction (Table 5). After controlling for monetary policy
rates and foreign exchange rates, the coefficient on limits on LTV ratio across different regression
equations is statistically significant in the whole sample. Especially for EMEs, a higher LTV ratio is
associated with smaller investments in construction.
Table 5. Effects of Macroprudential Measures on Output Growth and Residential
Investment
1/
Source: IMF staff estimates.
1/ Green, orange, and yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
2
2
3
3
53. For other macroprudential tools, we find little evidence of a direct effect on output.
The coefficients on capital requirements, caps on DTI ratio and RRs are not statistically significant,
indicating weaker effects on output than found for the LTV ratio. It is possible that for capital and
RRs statistically significant effects could be picked up in larger samples or using a different study
design. However, the results also confirm existing studies that tend to find modest effects of these
tools on output. See BIS (2010) and Glocker and Towbin (2012b), respectively.
54. By contrast, throughout, we find that variation in the policy rate has a statistically
strong effect on output growth. These findings suggest that some macroprudential policy tools
may be able to separately target a specific component of domestic demand, unlike monetary policy,
but that the effect of these macroprudential tools on aggregate output is more limited. Conversely,
the policy rate affects all economic activity regardless of which sector is vulnerable to systemic risks,
and may then have stronger effect on output growth.
55. It is possible that the effects are too small to show up as statistically significant in our
panel analysis. An indirect way of gauging the effects on output is to extrapolate from the effect on
credit growth. A number of empirical studies show the effects of credit supply shocks on output
growth, ranging from 0 to 0.34, that is, a one percent decline in credit supply induces a drop in the
GDP growth rate of up to 34 basis points. Thus, combining with results in Table 5, some
macroprudential policy instruments may still affect output growth to a meaningful degree.
Results—effects on capital flows
56. We finally turn to an investigation of the effect of macroprudential measures on
capital inflows. Monetary policy is often said to be constrained in open economies since policy rate
hikes to contain financial exuberance are likely to trigger more capital inflows. Some
macroprudential tools, on the other hand have been found in the literature to be useful to affect
gross flows and the composition of capital inflows.
57. We specifically investigate the determinants of portfolio investments. The
specification is similar to those employed before. However, we now account for the effect of the
spread between the domestic rate and the federal funds rate. In particular, we create a dummy
variable that indicates whether this spread is unusually high, relative to the average in the country
concerned.
58. The results of this exercise are contained in Table 6. We find that where the interest
spread is high, this stimulates portfolio inflows. Moreover we do not find that capital requirements,
LTV and DTI ratios have any effect on the strength of portfolio inflows. By contrast, we find
statistically strong evidence that increases in RRs reduce portfolio inflows in emerging economies
with floating exchange rates. These results chime with those found elsewhere in the literature:
increases in RRs lead to a depreciation of the currency. They can also, more mechanically, arise when
RRs apply to FX debt securities issued by banks.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
4
4
INTERNATIONAL MONETARY FUND
Table 6. Effects of Macroprudential Measures on Capital Inflows
1/
Source: IMF staff estimates
1/ Green, orange, and yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively.
B. Effects of Macroprudential Policy Measures—Symmetric or Asymmetric?
59. This section describes additional empirical tests to investigate potential asymmetries
in the effectiveness of macroprudential measures. From a policy perspective, it is important to
ascertain whether the size of the effect of a tightening of a macroprudential policy tool differs from
that of a loosening of the tool, or whether the effectiveness of macroprudential measures is
symmetric.
60. Table 7 documents the number of instances, for each macroprudential tool, in which
the measures were tightened and loosened. This suggests that for most measures, the number of
tightening events is far greater than that of loosening events. This is a key limitation for the
empirical analysis, since if there are few instances of loosening this will reduce the power of any test
of differential effects. Inspection of the table suggests that this is a major issue in particular for
capital requirements and DTI ratios. For RRs in particular, the situation is somewhat better with a
ratio of tightening to loosening events roughly 3 to 1.
Table 7. Number of Macroprudential Measures—Tightening or Loosening
Total Sample Tightening Loosening
Capital Requirements 1728 21 3
Limits on DTI Ratio 1728 17 2
Limits on LTV Ratio 1728 34 5
Reserve Requirements 1728 28 9
Sources: Lim and others (2011) and author’s extension.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
2
2
5
5
61. The potential for asymmetric effects is gauged by interaction exercises. The basic
strategy followed is to interact the measure of the tightness of the macroprudential tool (the step
function used previously) with dummy variables that indicate, for each change in the step function,
whether the change is an increase (tightening) or a decrease (loosening). We examine the issue of
symmetry for the key financial variables investigated before. That is, our dependent variables are
credit growth as well as the growth of asset prices.
62. Figure 1 suggests that loosening events tend to occur from 2008, when in many
countries there would have been financial stress as a result of the global financial crisis. This
suggests that countries are more likely to loosen macroprudential policy tools when the financial
system is in need of support. Our empirical analysis takes account of the resulting potential
measurement bias by including dummy variables that indicate financial stress, such as credit bust,
asset price busts and recession, as well as interactions with these variables that capture differential
effects of macroprudential policy in times of stress.
63. Overall we cannot reject the hypothesis that the effect of macroprudential policy tools
is symmetric, rather than asymmetric. The results on credit growth suggest that, if anything, a
loosening of RRs has a stronger effect on credit growth than a tightening of RRs (Table 8). However,
Figure 1. Number of Macroprudential Measures—Tightening or Loosening
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
2000Q1 2002Q3 2005Q1 2007Q3 2010Q1
Number of Tightening or Loosening of Capital
Requirements (Sample: 36 countries)
Source: IMF staff calculations.
Tightening
Loosening
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
2000Q1 2002Q3 2005Q1 2007Q3 2010Q1
Number of Tightening of Caps on Debt-to-Income Ratio
(Sample: 36 countries)
Source: IMF staff calculations.
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
2000Q1 2002Q3 2005Q1 2007Q3 2010Q1
Number of Tightening of Limits on Loan-to-Value Ratio
(Sample: 36 countries)
Source: IMF staff calculations.
-5.0
-4.0
-3.0
-2.0
-1.0
0.0
1.0
2.0
3.0
2000Q1 2002Q3 2005Q1 2007Q3 2010Q1
Number of Tightening of Reserve Requirements
(Sample: 36 countries)
Source: IMF staff calculations.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
6
6
INTERNATIONAL MONETARY FUND
for the other macroprudential tools, there are no measurable differences. The results on house
prices suggests that, if anything, a tightening of LTV ratios appears to have a stronger effect than a
loosening does, while, for other macroprudential tools there is no measurable difference (Table 9).
64. These results need to be interpreted with great caution. As pointed out above, for some
of the macroprudential tools, in particular capital requirements and DTI ratios, the incidence of
Table 8. Effects of Macroprudential Policy Stance on Credit Growth
Source: IMF staff estimates.
1/ Green, orange, yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively.
Credit Growth Rate (%, q-o-q)
Credit Growth Rate (-1) 0.40 0.39 0.39 0.39 0.39 0.39 0.39 0.40 0.40 0.39 0.39 0.39 0.39
GDP Growth Rate 0.23 0.23 0.23 0.23 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.21
Interest Rate (-1) -0.04 -0.05 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04 -0.04
Credit Bust 0.88 0.43 0.55 0.49
House Price Bust -0.69 -0.66 -0.71 -0.58
Recession
-0.62 -0.56 -0.65 -0.59
Capital Requirement (-1) -0.33 -0.49 -0.56
Limits on DTI Ratio (-1) -0.25 -0.20 -0.17
Limits on LTV Ratio (-1) -0.07 -0.04 -0.04
Reserve Requirements (-1) -0.28 -0.23 -0.29
Other Measures(-1) -0.13
-0.23 -0.27 -0.21 -0.14 -0.22 -0.29 -0.21 -0.14 -0.23 -0.31 -0.22
CR(-1)*Credit Bust -1.44
DTI(-1)*Credit Bust 1.90
LTV(-1)*Credit Bust 0.11
RR(-1)*Credit Bust 0.41
CR(-1)*House Price Bust 0.74
DTI(-1)*House Price Bust 0.82
LTV(-1)*House Price Bust 0.45
RR(-1)*House Price Bust -0.10
CR(-1)*Recession 0.20
DTI(-1)*Recession -0.02
LTV(-1)*Recession 0.24
RR(-1)*Recession 0.14
CR(-1)*CR_tight(-1) 0.20 0.13 0.14
DTI(-1)*DTI_tight(-1) -0.11 0.10 0.08
LTV(-1)*LTV_tight(-1) 0.14 0.11 0.12
RR(-1)*RR_tight(-1) 0.19 0.16 0.17
CR(-1)*CR_loose(-1) -0.77 -1.11 -0.74
DTI(-1)*DTI_loose(-1) -0.67 -0.12 -0.07
LTV(-1)*LTV_loose(-1) -0.48 -0.35 -0.41
RR(-1)*RR_loose(-1)
1.78 1.83 1.62
T-test (H0: two coefficients equal) 0.94 0.70 0.65 1.69 1.00 0.28 0.49 1.75 0.84 0.19 0.57 1.53
P-value 0.35 0.49 0.51 0.09 0.32 0.78 0.63 0.08 0.40 0.86 0.57 0.13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
2
2
7
7
loosening events may be too small to detect differences that are statistically significant in a small
sample. That is, it is quite possible that differences could be detected in samples with a greater
number of observations.
Table 9. Effects of Macroprudential Policy Stance on House Price Appreciation
House Price Appreciation Rate (%, q-o-q)
House Price (-1) 0.39 0.38 0.39 0.40 0.39 0.38 0.38 0.38 0.38 0.38 0.39 0.40 0.39
GDP Growth Rate 0.48 0.48 0.48 0.48 0.48 0.39 0.41 0.40 0.41 0.48 0.48 0.47 0.49
Interest Rate (-1) -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02 -0.02
Credit Bust -0.26 -0.03 0.31 -0.20
House Price Bust -4.89 -4.88 -5.09 -4.99
Recession 0.20 0.02 -0.18 0.39
Capital Requirement (-1) -0.65 -0.53 -0.39
Limits on DTI Ratio (-1) 0.37 0.29 0.39
Limits on LTV Ratio (-1) 0.04 -0.28 -0.21
Reserve Requirements (-1) -0.01 0.00 0.18
Other Measures(-1) 0.03 -0.25 -0.04 -0.17 -0.07 -0.29 -0.11 -0.22 0.03 -0.25 -0.05 -0.12
CR(-1)*Credit Bust 0.70
DTI(-1)*Credit Bust -0.43
LTV(-1)*Credit Bust -0.97
RR(-1)*Credit Bust 0.46
CR(-1)*House Price Bust -1.41
DTI(-1)*House Price Bust -2.57
LTV(-1)*House Price Bust -0.28
RR(-1)*House Price Bust -0.72
CR(-1)*Recession -0.72
DTI(-1)*Recession -0.20
LTV(-1)*Recession 0.38
RR(-1)*Recession -1.10
CR(-1)*CR_tight(-1) -0.02 0.13 0.11
DTI(-1)*DTI_tight(-1) 0.11 -0.01 0.03
LTV(-1)*LTV_tight(-1)
-0.79 -0.36 -0.67
RR(-1)*RR_tight(-1) -0.20 -0.25 -0.20
CR(-1)*CR_loose(-1) -1.06 1.51 -0.68
DTI(-1)*DTI_loose(-1) 0.16 -0.12 0.11
LTV(-1)*LTV_loose(-1) -0.29 -0.59 -0.42
RR(-1)*RR_loose(-1) -1.90 -0.90 -1.41
T-test (H0: two coefficients equal)
0.90 0.06 0.23 1.38 1.06 0.13 0.23 0.55 0.68 0.09 0.49 0.98
P-value 0.37 0.95 0.82 0.17 0.29 0.89 0.82 0.58 0.50 0.93 0.62 0.33
Source: IMF staff estimates.
1/ Green, orange, and yellow color in each cell indicate significance at 1, 5, and 10 percent level, respectively
.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
2
2
8
8
INTERNATIONAL MONETARY FUND
IV. COUNTRY CASES
A. Selected Central, Eastern, and South-Eastern Europe Countries
23
A salient feature of the experience in Central, Eastern, and South-Eastern Europe (CESEE)
ahead of the crisis was a pronounced increase in foreign currency (FX) lending. This case
study examines the experience of five inflation targeting countries in the region and
investigates whether interest rate spreads stimulated the increase in FX lending. It also
studies macroprudential policy responses that were taken to reduce the systemic risk
associated with such lending. The study finds that where interest rates were low relative
to advanced country rates, the increase in FX lending was less pronounced, other things
equal. It also finds that the several macroprudential measures were effective in
counteracting the increase.
65. This case study focuses on five inflation-targeting countries in Central, Eastern and
South-Eastern Europe (CESEE): The Czech Republic, Hungary, Poland, Romania, and Serbia. The
first three countries joined the European Union in 2004, the fourth in 2007, and the fifth became an
EU accession candidate in 2012. The five countries have strong linkages to the Euro Area and have
banking sectors dominated by large Euro Area banking groups.
24
The euro is their domestic
currencies’ natural cross.
66. The dispersion in the five countries’ monetary policy rates has narrowed over time but
remains large. The Czech Republic was the earliest inflation targeting adopter and has managed to
maintain low inflation and low policy rates over the past several years, suggesting a high degree of
policy credibility (Table 10). At the other end of the spectrum, Serbia has struggled to meet
increasingly more ambitious inflation targets, with inflation overshooting the target by more than
5 percentage points in 2011, and policy rates remaining close to double-digits.
Table 10. Selected CESEE Countries: Inflation Target, Inflation Outturn, and Policy Rates, 2006–11
23
Prepared by Jerome Vandenbussche (EUR).
24
See Chapter 4 of IMF (2011).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
2
2
9
9
67. Higher policy rates are associated with a higher share of foreign currency (FX) loans
across this group of countries (Figure 2). FX lending has been a long standing feature in a large
part of CESEE and increased further during the credit boom of the last decade.
25
While there are
multiple demand and supply factors that explain the currency composition of credit, and each of
them is likely to have played a role in favoring the growth of FX loans in the region over time, it is
striking to see that among our group of five inflation targeters, the level of the monetary policy rate
is very strongly associated with the share of foreign currency loans.
26
Except for the Czech Republic,
interest rates on domestic currency loans are generally higher than on FX loans, due to lower
monetary policy credibility and/or higher inflation volatility in the domestic economy. The lower
interest rate charged on FX loans may be too salient a feature for the typical unhedged borrower to
appropriately factor the risks of FX appreciation into his or her decision. Indeed in many countries in
the CESEE region, especially those with fixed or appreciating exchange rates, FX loans were
perceived to be cheaper. This was especially the case for mortgages: mortgages in Euros and Swiss
francs (and even in some cases in Japanese yen) carried a much lower interest rate—and longer
maturity—than in local currency.
Figure 2. Selected CESEE Countries: Foreign Currency Loans and Policy Interest Rate Spreads,
2005–11
25
See Dell’Ariccia and others (2012) for a description of the 2003-2008 credit boom in CESEE.
26
Besides interest rates, other demand-side determinants include expectations of euro adoption, underestimation of
foreign currency risk, and natural hedges. Major supply side determinants include deposit euroization and foreign
funding of the banking system. Some determinants, such as institutional quality and exchange rate volatility operate
both through the demand and the supply side. See, among others, Rosenberg and Tirpak (2008); Pann, Seliger and
Ubelies (2010); Zettelmeyer, Nagy and Jeffrey (2010); and Steiner (2011).
0
2
4
6
8
10
12
0
10
20
30
40
50
60
70
80
Czech Republic Poland Hungary Romania Serbia
Foreign currency loans to total loans (non-financial corporations, percent, 2011)
Foreign currency loans to total loans (households, percent, 2011)
Average policy rate spread to euro (2005-2011, pps, right scale)
Sources: IMF BSA database, Haver, and author's calculations.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
3
3
0
0
INTERNATIONAL MONETARY FUND
68. To prevent the emergence of FX loans on a large scale, policy intervention may be
needed. There are three main reasons for this. First, large aggregate unhedged FX exposures create
negative externalities because they are a significant source of systemic risk in the banking system
during crisis times (as greater installments increase the probability of default), and they generate
greater macroeconomic volatility and limit macroeconomic policy options (because policy makers
internalize the adverse balance sheet effects of devaluations or large depreciations on unhedged FX
borrowers). Indeed, the Czech National Bank was able to reduce its policy rate by 150 bps between
end-June and end-December 2008, while the Serbian National Bank increased its policy rate by
200 bps during the same period. Second, such exposures are subject to moral hazard related to
implicit bailout guarantees. Third, a FX loan may expose the borrower (whether hedged or
unhedged) to greater liquidity risk than a domestic currency loan if the bank supplying the loan is
funded through international wholesale markets rather than more stable sources of funding (such as
domestic deposits). All of these considerations may therefore justify policy action—on
macroeconomic management and financial stability grounds—to limit the extent of FX borrowing in
the economy. In addition, policy intervention may also be required for customer protection motives
if some borrowers misunderstand and/or are not properly alerted to exchange rate risks.
69. Across the five countries, policy-makers addressed the risks associated with FX loans
differently and at different stages of the recent boom-bust cycle (Table 11).
27
The Czech
policy-makers did not have to intervene, as FX loans in their country were mostly to hedged
corporations and remained stable throughout the past decade. Romania and Serbia, which have a
large share of euroized liabilities, increasingly differentiated the rate of RRs by currency starting in
2004/05.
28
They also differentiated loan classification and provisioning rules by currency (in 2005 in
Romania and in 2008 in Serbia). Higher risk-weights on FX loans above a certain threshold amount
were introduced in Serbia in 2006, and higher risk-weights on FX mortgages were introduced in
Poland in 2008. Poland (in 2006) imposed stricter debt-to-income (DTI) and loan-to-value (LTV)
ratios on new FX mortgage holders (through the so-called “Recommendation S”). Romania imposed
a maximum ratio of FX loans to unhedged borrowers to own funds between 2005Q3 and its entry
into the European Union in 2007Q1, and tightened DTI limits for households for a short period in
2008-09. As the macroeconomic and financial costs of FX loans to unhedged borrowers became
apparent during the post-Lehman bust, Hungary introduced LTV and DTI regulation differentiated
by currency before banning FX mortgages altogether in 2010. More recently, Poland further
increased risk-weights on FX household loans while Romania introduced differentiated LTV limits by
currency. Across the CESEE region, there is now greater consciousness among policymakers of the
need to develop local currency capital markets so that banks can decrease their reliance on FX
27
For a comprehensive description of the recent boom bust cycle in CESEE and each of its individual countries, see
Bakker and Klingen (2012).
28
A significant part of FX loans in Hungary and Poland were funded through FX swaps, making differentiated RRs by
currency in those two countries a less relevant possible instrument.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
1
1
funding for long maturities,
29
while, at the European Union level, the European Systemic Risk Board
has published a set of recommendations on lending in FX (ESRB, 2011).
Table 11. Selected CESEE Countries: Use of Macroprudential Measures Addressing Foreign
Currency Loans, 2002Q1–2012Q1
Sources: Vandenbussche-Vogel-Detragiache (2012) database and national central banks' websites.
70. A panel regression analysis confirms that greater interest rate spreads increase the
share of FX loans within countries (Table 12). Explanatory variables included in the regression
include the spread between the domestic policy rate and the policy rate of that currency’s natural
cross, the volatility between the domestic currency and that cross currency, and the past
appreciation of the domestic currency relative to the cross currency. The natural cross currency is
taken to be the euro in all cases. While higher spreads and greater recent appreciation are expected
to stimulate demand for FX loans, exchange rate volatility is expected to reduce their attractiveness.
Regression results, both for FX loans to non-financial corporations and to households, are consistent
with these priors but only the interest spread is consistently significant.
71. At the same time, the several macroprudential measures have been effective in
counteracting that effect. The various types of macroprudential measures discussed above are also
included in the regression. Because policy-makers are likely to take measures against unhedged FX
loans when they anticipate that unhedged FX borrowing would otherwise be strong, endogeneity
likely biases the estimates for the effect of these measures. In spite of endogeneity, we do find that
the strongest measures—a maximum ratio of FX loans to own funds as in Romania, and quantitative
restrictions on the share of FX mortgages in Hungary (0 percent of the flow)—and that stricter debt-
to-income ratios for FX loans had an impact.
30
The availability of funding in FX, captured by the
change in the share of FX deposits and by the itraxx index (which is correlated with funding
pressures of large Western European banks), does not enter significantly in the regression results.
29
See European Bank Coordination (“Vienna”) Initiative (2011).
30
It is likely that more conservative LTV limits for FX loans helped keep default rates relatively low, even if—at least
according to our analysis—they may not have done much to slow FX lending.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
3
3
2
2
INTERNATIONAL MONETARY FUND
Table 12. Selected CESSE Countries: Determinants of the Share of Foreign Currency Loans,
2001Q1–2012Q1
1/
Sources: Haver, IFS, Vandenbussche-Vogel-Detragiache (2012) database, national central banks' websites, and authors’
calculations.
1/ The dependent variable is the quarter-on-quarter change in the logistic transformation of the share of foreign currency
loans (adjusted for exchange rate movements). The unbalanced panel covers the Czech Republic, Hungary, Poland, Romania and
Serbia during 2002Q1-2012Q2 and contains 160 observations. The estimation method is fixed effects with robust standard
errors. All explanatory variables are lagged one period. One (resp. two, three) stars indicates significance at the 10 (resp. 5, 1)
percent confidence level. A "+" or "-" indicates the sign of the estimated coefficient. The strength of each type of
macroprudential measure is measured using the same method as Vandenbussche-Vogel-Detragiache (2012). A dummy for
Hungary in 2012q1 is included to account for the drop in the share of household foreign currency loans by about 6 percentage
points because of the mortgage early repayment scheme initiated by the government.
2/ The euro is used as the cross currency in all cases but two. Because most FX loans to households in Poland and Hungary are
in Swiss franc, the Swiss franc is used in those two cases.
72. We conclude that policy rate differentials have been one of the key drivers of changes
in FX lending in our group of five CESEE inflation targeters, and that at least some
macroprudential measures can contain vulnerabilities from FX lending by reducing the extent
of the build-up. The case study confirms that strong monetary and macroprudential policies can
have mutually reinforcing effects. If a country has a credible monetary policy regime, policy rates can
stay relatively low, reducing the incentive for unhedged FX borrowing. Conversely, strong
macroprudential policies can help enrich the set of feasible monetary policy options and sustain
monetary policy transmission in small open economies. In the case of countries with a high degree
of foreign ownership of the banking system, as in the CESEE region, circumvention of domestic
macroprudential measures can be a relatively greater concern, and close home-host supervisory
cooperation is therefore a requirement to enhance the effectiveness of both macroprudential and
monetary policies.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
3
3
B. Brazil
31, 32
Brazil has been an active user of both monetary and macroprudential policies. Its
experience during the post-crisis period illustrates well the complementary relationship
between the two policies. Monetary policy was focused on ensuring price stability, and
macroprudential instruments were used to contain the potential buildup of systemic risks
from rapid credit growth. As these policies leaned against the business and financial cycle,
synchronized during this period, the policy mix was appropriate to meet two objectives—
price and financial stability—with two instruments.
Macroeconomic conjuncture and monetary policy
73. Brazil experienced a short but sharp swing in economic activity and inflation after the
global financial crisis. Until early 2011, the real economy had rebounded strongly from the global
financial crisis, even showing signs of overheating, and inflation was driven by buoyant domestic
demand and high food and commodity prices. But the economy subsequently slowed sharply, and
inflation dropped from the second half of 2011 on the back of policy tightening and in a globally
more uncertain environment.
74. Monetary policy was used countercyclically in macroeconomic management during
the post-crisis period. In response to rising inflation and the fast-paced economic rebound, the
Banco Central do Brasil (BCB) raised the policy (Selic target) rate by 200 bps in 2010, which was
followed by a 175 bps increase in the first half of 2011, amounting to a cumulative rate hike of
375 bps. But as global economic deterioration adversely affected confidence and trade, contributing
to the sharp slowdown of economic activity, the policy rate was recently eased substantially, by
525 bps.
Figure 3. Brazil: Macroeconomic Conjuncture and Policy Responses
31
Prepared by Heedon Kang (MCM).
32
The case study draws on the 2012 Article IV staff report, the latest FSAP update, Financial Stability Reports of the
BCB, several published papers from the IMF and the BCB.
-6
-4
-2
0
2
4
6
8
10
12
14
16
2008M1 2009M1 2010M1 2011M1 2012M1
Output Gap
Inflation (IPCA)
Policy Rate (Selic Target)
Monetary Policy, Inflation, and Output Gap
(In percent)
Sources: Banco Central do Brasil, IMF staff calculations.
-2
-1
0
1
2
3
4
5
2008 2009 2010 2011
Primary Balance
Structural Balance including policy lending
Structural Primary Balance
Fiscal Policy
(In percent of GDP)
Source: IMF staff reports.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
3
3
4
4
INTERNATIONAL MONETARY FUND
Credit expansion and macroprudential policies
75. Credit expanded rapidly since 2004, supporting economic growth and financial
inclusion, but also posing risks, particularly to the household sector (Figure 4).
Credit-to-GDP in Brazil has risen fast, even if from a low base (from 24 percent of GDP in 2004 to
51 percent in 2012). To date though, its level remains relatively low by international standards.
33
Moreover, the pace of credit expansion has moderated recently, shrinking the estimated credit-
to-GDP gap significantly.
While household debt is still in line with that of regional peers, debt service-to-income and NPL
ratios are high in comparison to those peers, reflecting high lending rates and short maturities,
which are sources of vulnerability. Even though the growth of credit to households decelerated
somewhat during the post-crisis period, the debt service-to-income ratio rose to 22 percent
(18 percent at end-2008), and the NPL ratio increased to 8 percent.
33
See Dell’Ariccia and others (2012) for an international comparison of the credit-to-GDP ratio.
Figure 4. Brazil: Credit Expansion
Latin America—consumer indebtedness, 2011
0
5
10
15
20
25
2008M1 2009M1 2010M1 2011M1 2012M1
Principal Interest
Household Debt Service to Income Ratio
(In percent)
Source: Banco Central do Brasil.
0
2
4
6
8
10
0
5
10
15
20
25
30
35
2008M1 2009M1 2010M1 2011M1 2012M1
NPL Ratio (Right Axis)
Household Credit Growth Rate (Left Axis)
Household Credit Growth Rate and NPL Ratio
(In percent (YoY), in percent)
Sources: Banco Central do Brasil, Haver Analytics.
-2
0
2
4
6
8
10
0
10
20
30
40
50
60
2004M1 2006M1 2008M1 2010M1 2012M1
Credit-to-GDP Gap
Credit-to-GDP
Credit-to-GDP and Credit-to-GDP Gap
(In percent of GDP)
Sources: Haver Analytics, IMF staff calculations.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
5
5
76. The BCB used macroprudential measures (MaPPs), such as reserve and capital
requirements, to manage liquidity and contain the potential buildup of systemic risks from
rapid credit growth and household leverage.
The BCB changed RRs frequently to manage credit cycles in a countercyclical manner (Figure 5).
Right after the onset of the global financial crisis, the BCB eased RRs to avoid a credit crunch in
the financial system. Since then, the BCB used tighter requirements as speed limits, to slow
down overall credit growth, in the conjunction with more targeted measures on consumer loans.
Capital requirements on new loans to households were tightened in December 2010 (Table 13).
The measure focused on vehicle financing, payroll-deducted loans, personal credits, involving
longer maturities and higher loan-to-value (LTV) ratio. In November 2011, the BCB recalibrated
the measure by removing the LTV condition on vehicle loans and lowering risk weights on
shorter-term loans. A hike of the minimum payment for credit card bills to 15 percent from
10 percent was announced in December 2010 and implemented in June 2011, and the IOF (a
financial transaction tax) on consumer credit operations was hiked to 3 percent from 1.5 percent
in April 2011.
34
Table 13. Brazil: Changes of Capital Requirements on Consumer Loans
Source: Banco Central do Brasil.
34
This measure was to be tightened further to 20 percent in December 2011, but was not implemented.
Operation
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
3
3
6
6
INTERNATIONAL MONETARY FUND
77. Increases in RRs were temporarily effective in raising interest rate spreads and
curtailing credit growth. The recent FSAP report shows impulse responses to a one percent shock
in weighted average RRs that suggest a moderate but transitory slowing of credit growth (Figure 5).
Glocker and Towbin (2012b) document that, when tightened in a bank-based economy like Brazil,
RRs act as a tax on banks, increasing lending rates relative to deposit rates, contributing to a
depreciation of the currency, and thereby reinforcing the dampening effect of tighter RRs on
credit.
35
Figure 5. Brazil: Impacts of RRs Tightening (1 percent) on Credit Growth
78. Increases in the capital requirements on consumer loans contributed to reducing the
speed of household credit growth (Figure 6). After the implementation of the December 2010
measure, the growth rate of credit to households decreased 11 percentage points (from 22 percent
in December 2010 to 11 percent in December 2011).
36
And the proportion of vehicle loans with
maturity higher than 60 months in total vehicle loans fell by about 20 percentage points.
Figure 6. Brazil: Effectiveness of Changes of Capital Requirements on Consumer Loan
35
The BCB sterilizes impacts on overall liquidity of changes in RRs via open market operations, in order to maintain
the Selic rate close to its target, as decided by the Monetary Policy Committee (COPOM).
36
It should be noted that the decrease occurred against the backdrop of a slowing economy.
0
10
20
30
40
50
60
70
2008M1 2009M1 2010M1 2011M1 2012M1
On Demand Deposits
On Saving Deposits
On Time Deposits
Marginal RR Ratios
Source: Banco Central do Brasil.
Statutory Reserve Requirement Ratios
(In percent)
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
1357911131517192123
Response of Credit Growth Rates (MoM)
(In percent, ± 2 standard error)
Source: IMF staff calculation.
-10
0
10
20
30
Jan-03 Jan-05 Jan-07 Jan-09 Jan-11
Households
Legal Entities
Total
Growth Rate of Credit Granting
(In percent, YoY)
Source: Banco Central do Brasil.
Jan-08 Mar-09 Mar-10 Mar-11
0
10
20
30
40
50
60
70
maturity ≤ 3 years
3 < maturity ≤ 4
4 < maturity ≤ 5
maturity > 5 years
New Concession Granted : Vehicle Loans by Maturity
(In percent)
Source: Banco Central do Brasil
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
7
7
Policy coordination
79. Since the fast economic recovery was synchronized with a rapid credit expansion,
especially in the household sector, monetary and macroprudential policies appeared to be
complementary.
37
From the first half of 2010 to mid-2011, the BCB raised its policy rate by 375 bps
and tightened reserve and capital requirements. The policy mix was effective in managing aggregate
demand and inflation pressures, as well as in reducing the pace of credit growth (Figure 7).
80. The BCB played a leading role in macroprudential policies with a broad range of tools
at its disposal as both the monetary and supervisory authority. Based on guidelines of the
national monetary council (CMN), the BCB executes monetary policy, and regulates and supervises
the banking sector. It established a financial stability committee (COMEF) in May 2011 to separate
more clearly the prudential policy function from the monetary policy function. The committee
monitors sources of systemic risk, defines strategies to mitigate such risks, and allocates
responsibilities among departments within the BCB.
38
Figure 7. Brazil: Monetary and Macroprudential Policy Coordination
Source: Haver Analytics.
Source: Haver Analytics.
37
The credit growth rate is strongly and positively correlated with the output gap and the inflation rate, from the
second half of 2009. The correlation coefficients are 0.62 and 0.52, respectively.
38
The BCB’s mission is defined by its Board of Governors: “to ensure the stability of the purchasing power of the
currency and the soundness and efficiency of the financial system.”
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
3
3
8
8
INTERNATIONAL MONETARY FUND
C. Turkey
39
In the aftermath of the global financial crisis, the Turkish authorities faced a challenging
environment, characterized by widening current account deficits, strong short-term
capital inflows, and rapid credit growth. In response, the Central Bank of the Republic of
Turkey (CBRT) adopted a new “policy mix” that emphasized financial stability objectives,
while other macroprudential measures were taken only with some delay. This case study
examines the policy outcomes and points to the importance of coordination and clear
communication in responding to building financial imbalances.
81. In the aftermath of the global financial crisis, the Turkish authorities faced
a challenging environment. From late 2010, strong capital inflows led the Turkish lira (TL) to
appreciate, undermining competitiveness, and fueled a credit boom, adding to inflationary pressures
and increasing imports, leaving the economy exposed to the risk of a sudden capital flow reversal.
Were the inflows to dry up—either in response to Turkey’s imbalances or because of changes in the
global risk appetite—the lira would have rapidly depreciated, adding to inflationary pressures (this
time through the exchange rate pass-through), affecting balance sheets of banks and corporates
that had been borrowing in FX, and undermining overall confidence. Indeed, earlier examples of
such reversals led to sharp contractions of output.
40
82. The CBRT, operating as an inflation-targeting central bank since 2006, became
increasingly vocal about financial stability in mid-2010. It pointed to rapid expansion of
domestic credit and an increase in external borrowing by corporates and banks,
41
with a significant
share of it happening on a short-term basis. Together with real appreciation of the Turkish lira—
significantly above what the CBRT considered to be consistent with the fundamentals and driven
both by inflation differentials and nominal strengthening of the currency—this initiated a sharp
widening of the current account deficit (Figure 8).
83. In absence of an active and timely response from the financial supervisor (BRSA),
the CBRT was prompted to employ less traditional tools. This was done with an explicit purpose
of addressing both price and financial stability concerns. The strategy was built around several
instruments, with the emphasis shifting among them as ongoing developments provided insights
into the usefulness of various elements of the continuously evolving framework.
84. A cornerstone of the strategy employed until mid-2011 was the policy-induced
uncertainty in short-term market rates. Initially, in order to prevent fast appreciation of the lira,
the CBRT relied on the preannounced FX purchasing auctions, the volume of which was increased as
capital flows strengthened. Seeing them as ineffective, the CBRT lowered sharply the overnight
39
Prepared by Robert Tchaidze (EUR) and Heiko Hesse (MCM).
40
In 2001, seasonally-adjusted GDP fell by 11 percent from peak to trough and in 2009 by 13 percent.
41
Households are not allowed to borrow in FX since 2009.
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
3
3
9
9
borrowing rate—from 6.5 percent in September 2010 to 1.75 percent in November—and, by varying
the volume of liquidity provided via the repo auctions, opted to generate a lot of volatility in the
overnight market rate.
Figure 8. Turkey: Credit Growth and Current Account Deficit
Source: Central Bank of the Republic of Turkey.
85. This “interest rate corridor”—formed by the CBRT’s overnight borrowing and lending
rates—came to serve as a signaling device. In May 2010, the CBRT introduced a policy rate, at
which it would provide liquidity through quantity repo auctions. Initially, the overnight interbank
rates used to be kept close to this policy rate, and thus, the latter was indicating what a short-term
investor was likely to earn. As the interest rate corridor was “opened down,” the volatility of the
market rates was allowed to increase proportionally with it, even if on average they continued to
remain close to the policy rate. Thus, now it was the floor of the corridor that came to indicate a
guaranteed rate of return. This, it was hoped, would deter speculative inflows and reverse
appreciation of the lira.
86. In addition, in order to impact lending directly, the CBRT turned to the RRs. While it
saw them as the last tool to be used (after short-term interest rates and liquidity management)
when price stability was at stake, it also saw them as the first line of defense when it came to
financial stability. The CBRT argued that elasticity of demand for credit to interest rates is low and
that instead it had to rely on the RRs. Thus, the CBRT started to increase them (in June 2010 for FX
denominated liabilities and in November 2010 for TL), ceased to remunerate them
(September 2010),
42
and finally started to differentiate them by maturity (since February 2011 for TL
42
On TL liabilities. RRs on FX denominated liabilities have been unremunerated since December 2008.
-20
-10
0
10
20
30
40
50
60
70
-20
-10
0
10
20
30
40
50
60
70
Jan-09
Jun-09
Nov-09
Apr-10
Sep-10
Feb-11
Jul-11
Dec-11
Constant Exchange Rate-Adjusted Total Credit
Growth (In percent change)
year-on-year
13-week ma, annualized
0
5
10
15
0
2
4
6
8
10
12
2009 2010 2011 2012
Credit Growth and Current Account Deficit
(In percent of GDP)
Current account deficit
(12-month rolling sum)
Change in nominal credit
stock (right scale)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
0
0
INTERNATIONAL MONETARY FUND
and since June 2011for FX). By June 2011, the RRs on short-term TL liabilities were increased to
16 percent, from 5 percent in October 2010.
87. By then, the policy response had predominantly been led by the CBRT, and
macroprudential tools that were in the domain of the BRSA had been relatively underutilized.
In response to the global financial crisis, the BRSA took some steps in 2008–09 that helped
safeguard the domestic financial sector: in October 2008, banks’ dividend payouts were sharply
curtailed to bolster bank retained earnings and capital (renewed in following years); and in
June 2009, banks were prohibited to lend to consumers in FX. The BRSA took some further steps in
December 2010, when it introduced de jure loan-to-value limits on real estate loans and allowed
regulatory forbearance measures introduced following the global crisis to lapse in March 2011.
Finally, the authorities used moral suasion to target a uniform 25 percent increase on banks’ annual
loan growth for 2011, which appeared to have become binding for some banks in mid-2011.
88. The key macroprudential measures were introduced by the BRSA in June 2011. It
increased risk weights for new general purpose (consumer) loans and raised general provisioning
requirements for banks with high levels of consumer loans or non-performing consumer loans.
While the June measures on consumer loans were brought in with a delay, together with the credit
growth cap and the worsening external market conditions, they have contributed to the sharp slow-
down in credit growth in the second half of 2011. The BRSA also limited credit card payments in
July 2011, introduced capital surcharges for large exposures to interest rate risk (August 2011), and
amended minimum capital requirements for banks with strategic foreign shareholders
(September 2011).
89. Even though the macroprudential measures helped slow down credit growth, in late
2011 the monetary framework had to be significantly altered. At that time, in part reflecting
jitters in the global financial markets, the currency depreciated beyond what had been the intention,
and the inflows weakened. The depreciation fueled inflation, which by end-year stood at
10.4 percent against the target of 5 percent. The CBRT switched to FX selling auctions and even
undertook direct interventions. The RRs were lowered. The overnight borrowing rate was raised back
to 5 percent in August, while the overnight lending rate was increased in October from 9 percent to
12.5. In spite of all this, the current account deficit ended up at 10 percent of GDP, the second
highest in the world in dollar terms.
90. In principle, changes to the RRs could have impacted credit growth in two different
ways, monetary and macroprudential. First, and in principle, a tightening of RRs withdraws
liquidity from the market and may thus increase interbank market rates, thereby providing for
monetary tightening. However, by injecting liquidity through open market operations, the CBRT
chose to offset this monetary effect, in an effort to maintain market rates close to the policy rate, on
average. Second, and even if the monetary effect is offset, there can be a macroprudential effect: a
tightening of RRs forces banks to widen net interest margins, since it is a tax levied on banks’
liabilities, which can be offset either by lowering deposit rates or by raising lending rates or both. In
both cases, lending would be negatively affected, either through a reduction in available funds or
through weakened demand for credit.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
1
1
INTERNATIONAL MONETARY FUND
Table 14. Turkey: Macroprudential Measures Undertaken in 2008–11
Measure Description Adoption Date
Dividend policy
Restrictions on FX lending
Loan-to-value (LTV) ceilings
Guidance to Credit Growth Cap
High risk weights for consumer loans
Increased provisions for consumer loans
Limits to credit card payments
Interest Rate Risk
Changes to minimum Capital Adequacy
Requirements
Requires banks to seek approval from the BRSA before
distributing dividends. The maximum dividend payout for
CAR>18 percent is 20 percent, for 16 percent<CAR<14
percent is 15 percent and for 13 percent<CAR<16 percent
is 10 percent.
Allows non FX-earnings companies to borrow in FX from
local banks (previously, only FX-earning companies could
borrow FX), provided FX loan amount is greater than
US$5 million and maturity date is longer than a year; bans
consumers from taking out FX-linked loans
Implements loan-to-value ceilings on housing loans to
consumer (at 75 percent) and on purchases of commercial
real estate (at 50 percent)
The authorities provided guidance to banks that credit
growth (adjusted for FX movements) in 2011 should not
exceed 25 percent.
Higher risk weights introduced for fast growing consumer
loans. For new general purpose loans with maturities
below two years, the risk-weighting increased to 150
percent (from 100 percent). For new general purpose
loans with maturity greater than two years, the risk-
weight increased to 200 percent (from 100 percent)
For new (performing) general purpose loans, general
provisions were increased from 1 percent to 4 percent.
Specific provisions for (pre-nonperforming) loans
increased from 2 percent to 8 percent. The higher
provisioning requirements are conditional on banks
having a consumer loan portfolio exceeding 20 percent of
total loans or having a general purpose loan NPL greater
than 8 percent.
If three or more monthly payments within a calendar year
are less than half of the outstanding balance for the
period, the individual credit card limits cannot be
increased and cash advances for such credit cards cannot
be permitted, unless the outstanding balance for the
period is fully covered.
Announced by the BRSA to contain interest rate risk
through capital changes on large maturity mismatches,
discouraging duration gaps. Effective from 2012.
Amended by the BRSA in September 2011 to apply to
banks with foreign strategic shareholders as of January
2012. The minimum ratio would depend on various
factors such as the CDS spread of the parent and its
sovereign, EBA stress test results and the public debt ratio
in the country of origin.
October 2008;
extended in 2010 for
2009 profits; and
again in 2011
June 2009
December 2010
Spring 2011
June, 2011
June 2011
June 2011
August 2011
September 2011
Sources: Turkish authorities and IMF staff.
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
2
2
INTERNATIONAL MONETARY FUND
91. The effectiveness of macroprudential measures in controlling credit growth is not easy
to judge from the data. When the RRs started to increase, lending rates were at first declining,
while the deposit rate decreased with a lag and only marginally. As the RRs were increased further
and became differentiated by maturity, lending rates started to rise. However, so did the deposit
rate, suggesting that banks continued to expand credit, possibly moving further into the higher
return, higher risk segments, and with that in mind, trying to attract more deposits. By contrast, as
the BRSA measures were introduced, the impact on lending became more pronounced, with
the credit growth rates declining and the lending rates increasing much faster. This suggests that
the BRSA measures were more effective in curbing credit growth, in particular in 2011Q4. Increases
in lending rates also suggest that the slowdown in credit can be at least partly attributed to these
measures, rather than implying a slowdown in credit demand as a result of the cooling economy.
92. The mechanics of the framework had also negatively affected the primary objective of
the CBRT—inflation and inflationary expectations. In spite of favorable developments, with
inflation dropping in early 2011 to a 40-year low, the core inflation measures started to accelerate
again from October 2010, when the new framework was put in place, a trend changed only in early
2012. As for the inflationary expectations, until the inception of the new framework, they were well
approximated by an average of the inflation target and the latest observation. Under the new
framework, expectations have remained broadly flat at around 7 percent, suggesting that the link
broke and that survey participants simply reported the numbers around the top of the CBRT’s
inflation target band.
Figure 9. Turkey: Interest Rates, Reserve Requirement Ratios, and Growth of Lending
Sources: Central Bank of the Republic of Turkey and BRSA.
0
10
20
30
Jan-10
Mar-10
May-10
Jul-10
Sep-10
Nov-10
Jan-11
Mar-11
May-11
Jul-11
Sep-11
Nov-11
Lira Interest Rates and RR Ratios
(In percent)
Consumer lending Corporate lending
Deposit Average RR ratio
BRSA's
MaPP
measures
TL RR
differen-
tiation
by
maturity
Increases
in RR
ratios
-40
0
40
80
120
-40
0
40
80
120
Jan-09
Jun-09
Nov-09
Apr-10
Sep-10
Feb-11
Jul-11
Dec-11
Growth in Lending
(In percent, YoY)
Housing Automobile
Credit Cards Other
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
3
3
INTERNATIONAL MONETARY FUND
93. While it is hard to pinpoint factors that led to these developments, two hypotheses
can be put forward. First, the reliance on the RRs to deliver both financial and price stability was
miscalculated, particularly since an increase in the volume of the open market operations offset the
liquidity withdrawal, so that the overnight market interest rates would remain on average close to
the policy rate by (a strategy abandoned in 2012). The second is that with too many objectives and
too many tools, market participants became disoriented and confused when trying to deduce the
prioritization of the various objectives and the ultimate goal of these policies. For instance, in
August 2010, the CBRT cut its policy rate by 50 basis points, something one would expect either in
the face of a slowdown or strong inflows, but the RRs were kept high and the FX purchasing
auctions had already been canceled.
94. Altogether, Turkey’s 2010–11 experience demonstrates the importance of policy
coordination. Had macroprudential measures been employed in Turkey in a more timely fashion,
they could have relieved pressure on the central bank, which was trying to address financial stability
concerns while attempting to prevent the emergence of imbalances, manifested in heightened
inflation and the large current account deficit.
95. Other lessons can be drawn as well. Firstly, a clear assignment of tools to the policy
objectives is needed: interest rate based policies, part of a central bank’s tool-kit, are better suited to
dealing with the price stability objective, while the macroprudential policies, which apart from the
RRs are in a financial supervisor’s domain, are better suited to dealing with the financial stability
objective. Secondly, communication, in particular in times of frequent and large policy changes, is of
utmost importance to guide market’s expectations. The expectations channel of monetary policy
breaks down easily and repairing it requires time and significant efforts when the authorities
abandon a clear assignment of tools to objectives. Finally, to ensure financial stability,
Figure 10. Turkey: Inflation Expectations and Inflation Rates
Sources: Central Bank of the Republic of Turkey and IMF staff calculations.
4
5
6
7
8
9
10
11
Jan-06
May-06
Sep-06
Jan-07
May-07
Sep-07
Jan-08
May-08
Sep-08
Jan-09
May-09
Sep-09
Jan-10
May-10
Sep-10
Jan-11
May-11
Sep-11
Inflation Expectations
(In percent change, YoY)
Average of the Target and the latest Obs.
Expectations 12m ahead
0
2
4
6
8
10
12
14
Jan-09
Mar-09
May-09
Jul-09
Sep-09
Nov-09
Jan-10
Mar-10
May-10
Jul-10
Sep-10
Nov-10
Jan-11
Mar-11
May-11
Jul-11
Sep-11
Nov-11
Headline and Core Inflation
(In percent change, YoY)
Headline G-core
H-core I-core
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
4
4
INTERNATIONAL MONETARY FUND
macroprudential policy action needs to be timely and well-coordinated among several agencies,
ensuring that the chosen policy-mix is most effective in dealing with emerging financial imbalances.
D. Korea
43, 44
During the 2000s, Korea experienced housing price boom-busts and a sharp increase of
short-term foreign currency (FX) borrowing in its banking system. While the Bank of
Korea (BOK) focused on price and output stability under a flexible inflation targeting
framework, financial imbalances in the housing market were addressed with targeted
macroprudential policy measures, such as limits on loan-to-value and debt-to-income
ratios. More recently, restrictions on FX derivative positions, and a Macroprudential
Stability Levy were brought in to curb excessive short term foreign currency borrowing.
This case study shows that such macroprudential measures have clear advantages over
the use of monetary policy, which is too blunt to deal with housing market developments
43
Prepared by Heedon Kang (MCM).
44
The case study draws on the IMF Article IV staff reports, Monetary Policy Reports and Financial Stability Reports of
the Bank of Korea (BOK), working papers from the IMF and the BOK, and several journal articles.
Figure 11. Turkey: Cumulative Liquidity Injections
Source: Central Bank of the Republic of Turkey.
-15
-10
-5
0
5
10
15
20
25
30
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
Oct-10 Jan-11 Mar-11 May-11 Aug-11 Oct-11
Cumulative Liquidity Injections Since Oct., 2010
(In billion of Turkish lira)
Government deposits
Change in lira reserve requirements
FX interventions
Net CBT lending
Net liquidity supply
(rhs)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
5
5
INTERNATIONAL MONETARY FUND
and can worsen external vulnerabilities in an economy with a fully open capital account
like Korea.
Housing price boom-bust and limits on LTV and DTI ratios
96. Korea experienced two housing price booms during a period of macroeconomic
tranquility prior to the financial crisis. During these periods, both house prices and mortgage
loans increased in a synchronized fashion, demonstrating the two-way feedback loop (Figure 12).
45
The booms began with a sharp house price increase in a prime location, the “Gangnam” district in
the southern part of Seoul, followed by increases in other parts of Seoul and other regions in Korea
with time lags. The expansion of bank credit to households, especially mortgage loans, was also
fuelled by other factors, such as the liberalization of the housing finance system,
46
the resulting
severe competition for mortgage market share, and the preferred treatment of mortgage loans in
the BIS capital adequacy ratio calculation (Igan and Kang, 2011; Lee, 2012).
97. The authorities were aware of the dangers of house-price boom busts and responded.
Since the house price booms were financed through mortgage loans, the balance sheet of
households was likely to deteriorate sharply when the boom turned into bust, which in turn could
lead to a credit crunch, with negative consequences for real economic activity (Crowe and others,
2011). One of the lessons, which policymakers in Korea took away from the Asian crisis in 1997–98,
was that a credit boom-bust can trigger fire-sales of assets and a devastating recession, and it paved
the way for strong policy reactions by Korean authorities to tame the house price booms, in order to
prevent the recurrence of financial crises.
Figure 12. Korea: House Prices and Household Debts
45
During the 2000s, prior to the financial crisis, the economic growth rate remained around 5 percent and CPI
inflation stabilized around 3 percent.
46
The housing finance system was deregulated from the second half of 1990s. Particularly, commercial banks actively
provided mortgage loans in 1996, followed by the 1997 privatization of the Korean Housing Bank (KHB), the
monopolistic provider of low-interest and long-term housing loans.
-10
0
10
20
30
40
2000M1 2003M1 2006M1 2009M1 2012M1
Korea Seoul Gangnam
Real House Price Appreciation Rates
(In percent, yoy)
Source: Kookmin Bank.
-10
0
10
20
30
40
50
2001M1 2003M7 2006M1 2008M7 2011M1
GDP Growth Rate
Household Debts
Inflation
House Prices in Gangnam
Sources: Bank of Korea and Kookmin Bank.
Growth Rate of Household Debts
(In percent, yoy)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
6
6
INTERNATIONAL MONETARY FUND
98. However, the sector- and region-specific feature of the booms made it difficult to use
a policy tool, which affects the economy at large, such as monetary policy. As a flexible
inflation targeting central bank, the Bank of Korea (BOK) used its policy rates as a countercyclical
tool to achieve price and output stability, rather than as a tool to influence house price dynamics
(Figure 13). Between January 2000 and September 2008, the correlation coefficient between policy
rates and the output gap (inflation gap) is calculated as 0.68 (0.19), while the one between house
prices and policy rates is only 0.03. This also reflects the fact that house price cycles were weakly or
even negatively correlated with inflation and output cycles (-0.40 and 0.24) during the period.
Figure 13. Korea: Monetary Policy as a Countercyclical Tool
99. The Korean authorities used limits on loan-to-value (LTV) and debt-to-income (DTI)
ratios on mortgage lending to contain systemic risks from the housing price boom and the
associated household debt growth. Since their launch in September 2002 and August 2005, the
LTV and DTI limits have targeted speculative regions in the real estate market, rather than the whole
housing market on a nationwide basis. Their specific conditions have also been flexibly adjusted in
terms of maximum limits, loan types, and covered financial institutions. The measures were
tightened six and five times respectively, and loosened four times (Table 15).
100. Several studies show that the two measures reduced house price volatility, tamed
speculative incentives, and promoted the soundness of financial institutions, by keeping
households’ default rates low, but that they were less effective in curbing rising household
debts.
Igan and Kang (2011) present econometric results that house price appreciation slowed down
and transaction activity dropped in the six-month and three-month window following a
tightening of LTV and DTI limits. In addition, they find that the impact worked through the
expectations channel: tightening measures curbed households’ expectations about future gains
from house price appreciation, which reduced housing demand and alleviated pressures on
house prices. However, the authorities had to adjust them frequently and close loopholes in
order to maintain their effectiveness.
0
1
2
3
4
5
6
-6
-4
-2
0
2
4
2000M1 2004M1 2008M1 2012M1
Output Gap
Inflation Gap
Policy Rates
Sources: Bank of Korea and IMF staff calculations.
Output Gap, Inflation Gap, and Policy Rates
(In percent)
-2
0
2
4
6
8
10
12
-20
0
20
40
60
2000M1 2004M1 2008M1 2012M1
Household Debt
National House Prices
Policy Rates
Sources: Bank of Korea and Kookmin Bank.
House Prices, Household Debt, and Policy Rates
(In percent, yoy)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
4
4
7
7
Housing prices fell from 2008, but the delinquency ratio on household loans remained below
1 percent even until September 2012. This implies that strict implementation of limits on LTV
and DTI ratios prevented households’ defaults even as house prices fell, thus reducing financial
institutions’ credit risks. From this standpoint, the measures were helpful to secure the
soundness of financial institutions (Lee, 2012).
Household debt continued to expand even under the LTV and DTI limits, mainly for two reasons.
First, incentives to expand mortgage loans were strong from banks’ perspectives due to profits
from higher lending rates than on corporate loans, and the preferred treatment of household
loans in the BIS capital adequacy ratio calculation (Lee, 2012). Second, nonbanks expanded
credit to households, for which the authorities needed to close gaps by extending the
regulations to nonbanks. While the growth of bank credit to households declined to less than
10 percent from 2007, nonbank credit continued to expand briskly at 16 percent until the
financial crisis began (IMF, 2012).
Indeed, it is conceivable that the expectations that macroprudential measures would be taken to
support house prices in downturns may have indirectly contributed to the continuing growth of
household debt. Greater reliance on market driven price corrections might have helped prevent
excessive leverage (see 2012 Article IV Staff Report, Box 2). While the targeted use of LTV and
DTI ratios can be seen as successful in Korea, asymmetric use of such measures can be a
potential pitfall.
Figure 14. Korea: Effectiveness of Limits on LTV and DTI Ratios
House Price Appreciation Rates (Q-o-Q)
Quarterly Provision of Household Loans
Source: Lee (2012).
Source: Lee (2012).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
4
4
8
8
INTERNATIONAL MONETARY FUND
Table 15. Korea: Changes of Limits on LTV and DTI Ratios
Date Limits on LTV Ratio Coverage Direction
Sep.
2002
Introduced limits on LTV ratio as 60 percent. Banks Inception
Jun.
2003
Reduced the LTV ratio from 60 to 50 percent for loans of three years and
shorter maturity to buy houses in the speculative zones.
Banks Tighten
Oct.
2003
Reduced the LTV ratio from 50 to 40 percent for loans of 10 years and
shorter maturity to buy houses in the speculative zones.
Banks Tighten
Mar.
2004
Raised the LTV ratio from 60 to 70 percent for loans of 10 years or
longer maturity and less than one year of interest-only payments.
All Financial
Institutions
Loosen
Jun.
2005
Reduced the LTV ratio from 60 to 40 percent for loans of 10 years and
shorter maturity to buy houses worth W 600 million and more in the
speculative zones.
Banks Tighten
Nov.
2006
Set the LTV ratio as 50 percent for loans of 10 years and less maturity to
buy houses worth W 600 million and more in the speculative zones and
originated by all financial institutions.
All Financial
Institutions
Tighten
Nov.
2008
Removed all areas, except “Gangnam Three” districts, off the list of
speculative zones.
All Financial
Institutions
Loosen
Jul.
2009
Reduced the LTV ratio from 60 to 50 percent for loans to buy houses
worth W 600 million and more in the metropolitan area.
Banks Tighten
Oct.
2009
Extended the LTV regulations to all financial institutions for the
metropolitan area.
All Financial
Institutions
Tighten
Date Caps on DTI Ratio Coverage Direction
Aug.
2005
Introduced caps on DTI ratio as 40 percent for loans to buy houses in
the speculative zones only if the borrower is single and under the age of
30 or if the borrower is married and the spouse has debt.
All Financial
Institutions
Inception
Mar.
2006
Set the DTI ratio as 40 percent for loans to buy houses worth
W 600 million and more in the speculative zones.
All Financial
Institutions
Tighten
Nov.
2006
Extended the DTI regulation to overheated speculation zones in the
metropolitan area.
All Financial
Institutions
Tighten
Feb.
2007
Set the DTI ratio as 40–60 percent for loans to buy houses worth
W 600 million and less.
Banks Tighten
Aug.
2007
Set the DTI ratio as 40–70 percent for loans originated by non-bank
financial institutions.
All Financial
Institutions
Tighten
Nov.
2008
Removed all areas, except “Gangnam Three,” districts off the list of
speculative zones.
All Financial
Institutions
Loosen
Sep.
2009
Extended the DTI regulation to the non-speculative zones in Seoul and
the metropolitan area (“Gangnam Three” 40 percent, non-speculative
zones in Seoul 50 percent, the other metropolitan areas 60 percent).
Banks Tighten
Aug.
2010
Exempted the loans to buy houses in the non-speculative zones of the
metropolitan area if the debtor owns less than two houses (set to expire
by end-March 2011).
All Financial
Institutions
Loosen
Apr.
2011
Reinstated as planned: 40 percent in “Gangnam Three” districts,
50 percent in other Seoul metropolitan areas, and 60 percent in non-
Seoul metropolitan areas.
Banks Tighten
Source: Igan and Kang (2011).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
4
4
9
9
External vulnerability and FX-related macroprudential policies
101. In Korea, FX liquidity and maturity mismatches were another key vulnerability in the
years leading up to the financial crisis. From 2005, Korean banks rapidly increased short-term
non-core FX liabilities, creating FX liquidity and maturity mismatches. The key underlying structural
reason was that they bought dollar forward from exporters and asset management companies who
expected trend appreciation of the Korean won, and then hedged their long dollar positions with
short-term external FX borrowing. The aggregate short-term net external debt of Korean banks rose
to U$106 billion in the third quarter of 2008 from U$12 billion at end-2005 (Figure 15).
Figure 15. Korea: External Net Assets of Banking Sector
102. Given the FX mismatches, a sudden stop in capital flows hit Korean banks and the
macroeconomy hard after the Lehman Brothers bankruptcy, and Korea came close to
suffering another currency crisis. When the international wholesale funding market froze,
domestic banks and foreign banks’ branches were unable to roll over their maturing short-term
external liabilities, and short-term net external debt was sharply curtailed by U$50 billion within two
quarters. The Korean won depreciated rapidly and the CDS premium on Korean government bonds,
after having been at a level similar to those in China, Malaysia, and Thailand, rose to become much
higher, similar to that of Indonesia and the Philippines (Figure 16).
47
47
The rollover ratio of short-term external debt fell rapidly right after the financial crisis, from 99.8 percent in
September 2008 to 39.9 percent in October 2008 (Lee, 2012).
-120
-100
-80
-60
-40
-20
0
20
2000Q1 2003Q1 2006Q1 2009Q1 2012Q1
Net Asset (Domestic Banks)
Net Asset (Foreign Branches)
Total Net Asset
Short-Term FX Net Asset
(In U$ billions)
Source: Bank of Korea.
-120
-100
-80
-60
-40
-20
0
20
2000Q1 2003Q1 2006Q1 2009Q1 2012Q1
Short-Term Net FX Asset
Long-Term Net FX Asset
Short-Term and Long-Term FX Net Asset
(In U$ billions)
Source: Bank of Korea.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
0
0
INTERNATIONAL MONETARY FUND
Figure 16. Korea: Foreign Exchange Rates and CDS Premium
103. Domestic monetary policy may not be an effective tool to lean against the excessive
growth of the short-term non-core FX liabilities in an economy with fully open capital account.
Raising policy rates would encourage more carry trades and capital inflows, which can in turn fuel
credit booms. Moreover, there is evidence that Korean banks’ non-core liabilities are much more
(negatively) related to the United States policy rate than to the domestic policy rate. When global
liquidity conditions are lax, financial intermediaries are more engaged in the carry trade of
borrowing at low foreign interest rates and investing at higher domestic interest rates (Hahm and
others, 2012).
104. To address these vulnerabilities, the authorities implemented an array of
macroprudential measures, such as ceilings on banks’ FX derivative positions (June 2010 and
July 2011) and a macroprudential stability levy on non-core FX liabilities (August 2011).
48
The
combination of these two measures was meant to target both the source and the costs of the
excessive dependence on short-term non-core FX borrowings, and to encourage long-term and
stable sources of funding. The levy is adjustable and can be used as a countercyclical tool when
capital flow surges seriously threaten financial stability, with the maximum rate being 50 bps. Its
proceeds flow into the Foreign Exchange Stabilization Fund, which is separate from the government
budget and can be used as a buffer in the event of financial crises.
48
FX derivative positions were limited to 50 percent of capital for domestic banks and 250 percent for foreign banks’
branches in June 2010, and the limits were lowered to 40 percent and 200 percent in June 2011. The macroprudential
stability levy is currently charged at between 2–20 bps, depending on the maturities of debts: 20, 10, 5, and 2 bps on
short-term FX liabilities with maturity of one year or less, less than three years, less than five years, and more than
five years, respectively. The levy base is calculated as the daily average outstanding balance of banks’ eligible liability
for the year at each maturity.
800
900
1000
1100
1200
1300
1400
1500
1600
2000M1 2003M1 2006M1 2009M1 2012M1
Foreign Exchange Rate
(Won/U$)
Source: Bank of Korea.
0
200
400
600
800
1000
1200
1400
1/1/2006
7/1/2006
1/1/2007
7/1/2007
1/1/2008
7/1/2008
1/1/2009
7/1/2009
1/1/2010
7/1/2010
1/1/2011
7/1/2011
1/1/2012
7/1/2012
Korea
China
Malaysia
Thailand
Indonesia
Philippines
CDS Premium
(5 years)
Source: Bloomberg.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
1
1
INTERNATIONAL MONETARY FUND
105. The measures appear to have been reasonably effective in curbing banks’ reliance on
short-term FX funding and in reducing vulnerabilities from FX liquidity mismatches and their
links to exchange rate volatility (IMF, 2012; Ree and others, 2012).
49
Short-term net external debt, mostly of foreign banks’ branches, declined steadily from
US$65 billion in June 2010 to US$43 billion in June 2012. On the other hand, long-term net
external debt increased during the period to US$25 billion.
Rollover risks for domestic banks have contracted as residual maturities of their external debt
have increased (IMF, 2012).
The sensitivity of exchange rate volatility to changes in VIX declined substantially since the
financial crisis, reflecting lower FX liquidity mismatches (Ree and others, 2012).
The measures has reduced the channeling of global banking funds into Korea and also
moderated the sensitivity of capital flows to external financial conditions (Bruno and Shin,
2012b).
Macroprudential policy coordination
106. To support the macroprudential policy function, a formal coordination committee was
newly set up in July 2012, the “Macroeconomic and Financial Committee (Macroprudential
Committee).” Prior to July 2012, systemic risks in the foreign exchange market and domestic
financial sector were assessed separately by the “FX Market Stabilization Committee” and the
“Economic and Financial Market Monitoring Committee.” Four different agencies are members of
the new committee–the Ministry of Strategy and Finance (MOSF), the Bank of Korea, the Financial
Services Commission, and the Financial Supervisory Service. It is led by the MOSF and convened
every quarter. The Macroprudential Committee assesses external and domestic systemic risks and
coordinates the use of macroprudential instruments. In the meantime, each agency still conducts its
primary policy independently.
E. United States
50
The United States offers prime terrain to study financial instability in the years leading up
to the financial crisis of late 2007. Did an overly loose monetary policy and absence of
macroprudential measures undermine financial stability? The study finds some, though
weak, evidence that interest rates were too low relative to an optimal monetary policy
response. It also finds that a relaxation of regulations and the absence of an institutional
framework geared explicitly to financial stability contributed to the growing leverage of
49
Since these measures were brought in only recently, firm conclusions on their effectiveness would need more
thorough analysis as more data become available.
50
Prepared by Francesco Columba (WHD) and Tommaso Mancini-Griffoli (MCM).
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
2
2
INTERNATIONAL MONETARY FUND
large investment banks, though other factors may also have been at play. Moving
forward, it will be essential to improve the effectiveness of macroprudential policies in
advanced economies.
Monetary policy
107. Monetary policy began the 2000s with a sharp cut in interest rates to help the
economy emerge from the 2001 recession (Figure 17). Rates remained low until 2004, and then
rose steadily until 2006. As the financial crisis hit, rates were cut starting in 2007Q3, gradually at first,
then sharply, until they hit the zero lower bound in 2008Q4. Up to the financial crisis, inflation
remained relatively stable and the output gap closed, though with output somewhat above potential
(Figure 18).
Figure 17. United States: Interest Rates
Source: Haver Analytics.
Figure 18. United States: Inflation Rates and GDP Growth Rates
Source: Haver Analytics.
Sources: Haver Analytics and IMF staff estimates.
-2
-1
0
1
2
3
4
5
6
7
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
(In percent)
Fed Funds rate 10 yr gov Bond Spread
-2
-1
0
1
2
3
4
5
6
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
(In percent)
Inflation (CPI)
-8
-6
-4
-2
0
2
4
6
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
(In percent)
GDP Growth
GDP Gap
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
5
5
3
3
108. Some have asked whether persistently low interest rates, especially in the lead up to
the financial crisis, induced excessive borrowing and undermined financial stability. Indeed,
the risk taking and asset price channels of monetary policy suggest—in theory—that low interest
rates may lead banks to expand their balance sheets and take on additional risks, especially if asset
prices boost the net worth of lenders and borrowers.
109. Were interest rates “too low for too long?” A measure of the potential misalignment
between the policy rate and an optimal rate can be gauged with Taylor rules. Such simple rules
are widely shown to replicate well optimal monetary policy especially under model uncertainty
(Williams, 2003; Levin, Wieland and Williams, 2003; Orphanides and Williams, 2006, 2008; Taylor and
Wieland, 2009). But as Bernhard and Mancini-Griffoli (2011) point out, there is a wide variety of
possible Taylor rule specifications. More importantly, Taylor rules recommend significantly different
interest rate paths depending on the data fed into them.
110. We investigate a range of Taylor rules, where the optimal rate is formulated from a
baseline and various alternative Taylor rules, following Bernhard and Mancini-Griffoli (2011).
The baseline rule responds contemporaneously to CPI inflation deviations from target and
deviations of output from potential as computed by the CBO, with common sensitivity parameters
of 0.5 on each deviation.
51
A variety of other Taylor rules is also considered, allowing interest rates to
respond to (i) inflation and output gap forecasts; (ii) other measures of inflation (PCE, CPI and PCE
core and the GDP deflator); (iii) realized (revised) data; as well as (iv) different sensitivity parameters.
111. Results show some evidence that interest rates were somewhat “too low for too long,”
at least between 2002 and 2004. Figure 19 shows that the Federal Funds rate recommended by
the baseline Taylor rule is noticeably higher than the actually targeted Federal Funds rate. Yet,
recommended interest rates come down markedly when the Taylor rule is optimized to best capture
the Fed’s reaction function (by minimizing the root mean squared error) over the 1990–10 sample
(Figure 19).
52
If the Taylor rule is instead optimized to fit the 2002–07 period, the gap between the
recommended and targeted interest rate shrinks further (Figure 20).
53
In all cases, though, a gap
larger than was ever the case since 1990 arises, at least between 2002 and 2004.
112. These results appear to be quite robust. The “best fitting” Taylor rules shown in Figures A4
and A5 are robust to changes in specification and data. There are no statistically significant gains
from responding to forecasts of inflation or the output gap, or indeed revising the rule’s sensitivity
51
The baseline rule can be written as: i
t
= r + P
t
+ 0.5(P
t
– P
t
*
) + 0.5(Y
t
– Y
t
*
) where r is the natural real interest rate
consistent with inflation at target and a closed output gap, r is taken to be 2.5, the average CBO estimate of potential
output growth from 1990, P
t
is inflation, P
t
*
is the inflation target, taken to be 2 percent, and (Y
t
– Y
t
*
) is the output
gap.
52
The corresponding rule responds to the GDP deflator, a more stable inflation rate.
53
The resulting rule responds to core PCE inflation.
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
4
4
INTERNATIONAL MONETARY FUND
parameters (according to Giacomini and White, 2008, tests of predictive ability; details are available
in Bernhard and Mancini-Griffoli, 2011).
54
Figure 19. United States: Recommended Policy Rates from Baseline Taylor Rule Responding to
CPI Inflation and A Variant Responding to the GDP Deflator
Sources: The Federal Reserve Bank of St. Luis (Alfred), Federal Reserve Bank of Philadelphia (PHIL), Federal Reserve
Board, Congressional Budget Office, IMF staff calculations.
Figure 20. United States: Recommended Policy Rates from Baseline Taylor Rule Responding to
CPI Inflation and A Variant Responding to Core PCE Inflation
Sources: The Federal Reserve Bank of St. Luis (Alfred), Federal Reserve Bank of Philadelphia (PHIL), Federal Reserve
Board, Congressional Budget Office, IMF staff calculations.
54
Instead, Taylor rules responding to inflation rates other than the CPI can be shown to be statistically different from
the baseline rule in their ability to forecast the Federal Funds rate.
-6
-4
-2
0
2
4
6
8
10
1990Q1 1992Q1 1994Q1 1996Q1 1998Q1 2000Q1 2002Q1 2004Q1 2006Q1 2008Q1 2010Q1
Interest rate (In percent)
Target FFR Baseline
Best rule (GDP deflator) Best rule (GDP deflator), adjusted assumptions
-6
-4
-2
0
2
4
6
8
10
1990Q1 1992Q1 1994Q1 1996Q1 1998Q1 2000Q1 2002Q1 2004Q1 2006Q1 2008Q1 2010Q1
Interest rate (In percent)
Target FFR Baseline
Best rule (PCE core) Best rule (PCE core), adjusted assumptions
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
5
5
5
5
Leverage and house prices
113. There is little evidence, though, that the “too low” policy rates from 2002 to 2004
were transmitted to higher leverage and house prices. In theory, there are good reasons not to
expect low interest rates to strongly undermine financial stability through leverage. Low interest
rates come at a time of economic downturn; not the environment in which one would expect the
risk taking or asset price channels of monetary policy to be particularly strong. In addition, the
steeper yield curve typically accompanying a cut in interest rates, and observed in the United States
between 2002 and 2004 (Figure 17), supports bank profits through maturity transformation, taking
pressure off banks to seek profits elsewhere, in more risky assets.
55
114. Empirically, the correlation between deviations from Taylor rules and increases in
leverage is weak. Indeed, it is interesting to observe that leverage of United States investment
banks—which if excessive can lead to financial instability—did not increase until after interest rates
were raised towards their optimal rate in 2004. This is illustrated in Figure 21, focusing on
investment banks as opposed to commercial banks, which mostly kept leverage relatively low
throughout the period.
56
115. The tri-party repo market also shows that leverage did not rise in the period of
particularly expansive monetary policy. The tri-party repo market is a major source of wholesale
funding for investment banks, which anecdotal evidence suggests account between 65 percent and
80 percent of the total United States repo market. Repo funding of broker dealers displayed a
spectacular boom between 2004 and 2008—at a time of rising policy rates—while repo funding of
commercial banks increased at a pace mostly unchanged since 1990 (Figure 22).
55
A steeper yield curve comes from (i) expectations of higher inflation and thus higher short rates once the economy
recovers; and (ii) a higher risk premium induced by the downturn (Cochrane, 2007). Both, argue Rudebusch, Sack and
Swanson (2007), account for the steepening of the U.S. yield curve between 2002 and 2004.
56
The role of excessive leverage in the financial sector in propagating financial instability has been underscored by a
number of studies (Adrian and Shin, 2010; Brunnermeier, 2009; Brunnermeier and Pedersen, 2009; Panetta et al.
2009).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
6
6
INTERNATIONAL MONETARY FUND
Figure 21. United States: Leverage Ratio
Source: SNL Financial.
Note: Commercial banks refer to the largest 19 U.S. bank holding companies. The top 6 bank holding
companies are Bank of America, Citigroup, Goldman Sachs, JPMorgan, Morgan Stanley and Wells Fargo.
Goldman Sachs and Morgan Stanley were investment banks under SEC supervision prior to the crisis. Bank
of America and JPMorgan acquired respectively the investment banks Merrill Lynch in 2009 and Bear
Stearns in 2008.
Source: SNL Financial.
1/ Investment banks: GS = Goldman Sachs, LB = Lehman Brothers, ML = Merrill Lynch,
MS = Morgan Stanley, BST = Bear Stearns, Commercial banks (for comparison): BAC = Bank of America,
Citi = Citigroup, JPM = JP Morgan Chase.
-7
-5
-3
-1
1
3
5
7
7
9
11
13
15
17
19
21
1990 1993 1996 1999 2002 2005 2008 2011
Commercial banks excluding top six banks
Top six banks
GDP growth (yoy, right)
Leverage Ratio (Total Assets/Capital)
(In percent)
2012Q1
5
10
15
20
25
30
35
40
45
50
1995 1997 1999 2001 2003 2005 2007 2009 2011
GS BAC Citi MS
JPM ML BST LB
Leverage Ratio (Average Total Assets/Average Total Common Equity)
1/
(In percent)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
5
5
7
7
Figure 22. United States: Net Federal Funds and Security Repo Funding to Banks and
Brokers-Dealers
Source: Haver Analytics - Federal Reserve Board Flow of Funds.
1/ Private Depository Institutions: U.S. chartered depository institutions, foreign banking
offices in U.S., and credit unions.
116. A number of studies find a link—though small—between United States policy rates
and house prices and credit. Overall, while most studies find an effect of policy rates on house
prices, the economic magnitude is often found to be quite small. Indeed, as pointed out by Shiller
(2007), the recent United States house price boom started as far back as July 1998 and continued
exponentially until its peak in June 2006, without an obvious impact of changes in policy rates on
the time profile of house price increases (Figure 23). Formal evidence often finds that the impact of
policy rates on asset prices is small and that other factors are likely to have been at play.
Del Negro and Otrok (2007) found that the impact of accommodative monetary policy on
United States house prices had been small relative to the overall housing price increase.
Cardarelli and others (2008) find that the stance of monetary policy had an effect on the housing
market in the United States although its effect was magnified by the loosening of lending
standards.
57
Bean and others (2010) present a counterfactual experiment that has policy rates roughly
200 basis points higher than the actual path over the 2003–06 period. They find that the impact
on house prices and credit is quite weak, overall. Real house prices would have peaked around
7.5 percent lower. But the stock of real credit would have been just 3 percent lower by the end
of 2006, compared with an actual expansion in credit of almost 30 percent.
57
Similar findings are emphasized in Taylor (2007) and Iacoviello and Neri (2010).
0
200
400
600
800
1,000
1,200
1,400
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
brokers and dealers Banks 1/
(In U$ billions)
CORRECTED: 01/09/13
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
5
5
8
8
INTERNATIONAL MONETARY FUND
Macroprudential policy and leverage
117. If monetary policy was not the main culprit behind the bubble in United States real-
estate prices and the financial crisis, perhaps the lack of a macroprudential policy framework
did play a role? As the introduction of the Dodd-Frank Act notes, the financial crisis of 2007–08
generated a consensus on the need to introduce a regulatory framework to preserve financial
stability.
118. There is some evidence that a relaxation of regulatory standards did affect leverage in
the United States, and thus undermined financial stability. The leverage of investment banks
rose significantly after 2004 (Figure 21), while leverage at commercial banks remained relatively low
heading into the crisis.
The divergence can be explained in part by the different regulatory
provisions on leverage. Commercial banks had to respect a leverage ratio unchanged since 1991,
58
while in 2004 a change in SEC regulation allowed investment banks to raise their leverage from 15:1
to 40:1.
59
Indeed, evidence from other countries also suggests that regulation does affect leverage
. 60
58
The maximum leverage ratio is 33:1 for banks rated “strong” and 25:1 for all other banks. Banks are also subject to
prompt corrective action rules requiring them to maintain a minimum leverage ratio of 20:1 in order to be
considered well capitalized.
59
Broker-dealers were subject to supervisory rules limiting the debt to net equity ratio to 15:1 until 2004, when
investment banks opted for consolidated oversight (requiring that capital and risks be computed on a group-wide
basis) that allowed them to increase leverage to 40:1 in some cases.
60
Caps were imposed since the early 1980s on banks and other deposit-taking institutions in Canada. Similar limits
have been adopted in Switzerland in 2008 and have been considered in the United Kingdom.
Figure 23. United States: Case-Shiller Home Price Index
Source: Haver Analytics.
0
50
100
150
200
250
Jan-87 Jan-91 Jan-95 Jan-99 Jan-03 Jan-07 Jan-11
(Jan 2000 = 100)
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
5
5
9
9
The rapid rise of shadow banking, largely outside the reach of regulation, also explains the high
leverage of the United States financial system between 2005 and 2007 (Pozsar and others, 2010).
61
119. There is evidence also that an erosion of lending standards and increases in household
leverage contributed to the house-price boom-bust. Geanakoplos (2010) documents that the
average downpayment on Alt-A and subprime mortgage loans fell steadiliy over the pre-crisis
period. By 2006, the average downpayment had fallen to 2.9 percent, from 13 percent in 2000,
corresponding to an increase in household leverage for these borrowers from 7.7 to 37. Sengupta
(2010) documents significant loosening of underwriting standards and increased offering of
adjustable rate contracts in subprime and Alt-A mortgages, especially in mortgages originated from
2004, which led to sharp increases in defaults on these vintages as interest rates increased. Wider
empirical evidence suggests that the relaxation of regulatory constraints and lending standards,
combined with financial innovation and changes in market structure, including increased
competition, may have fueled the recent boom and bust pattern in the United States (Dell’Ariccia
and others, 2012a; Favara and Imbs, 2010; Mian and Sufi, 2009).
120. Insufficiently tight regulation was not the only cause of the increase in leverage. The
recovery that had emerged by 2004 probably also contributed to the increase in leverage. Indeed,
especially among investment banks, leverage has been documented to be pro-cyclical (Adrian and
Shin, 2010). The post 2004 peak of the largest banks’ leverage, or that just of investment banks,
merely returned to late 1990s/ early 2000s levels when tighter leverage requirements were in place
(Figure 21).
61
Shadow banking can be seen to include products associated with securitization (MBS, ABS, and other GSE
liabilities), as well as short-term money market transactions that are not backstopped by deposit insurance, including
repo and commercial paper.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
6
6
0
0
INTERNATIONAL MONETARY FUND
References
Adrian, Tobias, and Hyun Song Shin, 2010, “Liquidity and Leverage,” Journal of Financial
Intermediation, 19 (3), 418–437.
Almeida, Heitor, Murillo Campello, and Crocker Liu, 2005, “The Financial Accelerator: Evidence from
the International Housing Markets” (unpublished, New York University).
Aiyar, Shekhar, Charles W. Calomiris and Tomasz Wieladek, 2012, “Does Macroprudential Leak?
Evidence from a U.K. Policy Experiment,” Bank of England Working Paper 445, (London: Bank
of England).
Allen, Franklin and Douglas Gale, 2000, “Bubbles and Crises,” The Economic Journal, Vol. 110, No.
460, pp. 236–55.
Altunbas, Yener, Leonardo Gambacorta, and David Marques-Ibanez, 2012, “Do Bank Characteristics
Influence the Effect of Monetary Policy on Bank Risk?” ECB Working Paper Series, No. 1427,
March (Frankfurt am Main: European Central Bank).
Angelini, Paolo, Stefano Neri, and Fabio Panetta, 2011, “Monetary and Macroprudential Policies,”
Bank of Italy Working Paper 801 (Rome: Banca d’Italia).
Arregui, Nicolas, Jaromir Benes, Ivo Krznar, Srobona Mitra, and Andre Santos, 2013, “The Costs,
Benefits, and Side-effects of Macroprudential Policies,” forthcoming (Washington:
International Monetary Fund).
Asea, Patrick and Brock Blomberg, 1998, "Lending cycles," Journal of Econometrics, vol. 83(1-2),
pages 89-128.
Baillu, Jeannine, Cesare Meh, and Yahong Zhang, 2012, “Macroprudential Rules and Monetary Policy
When Financial Frictions Matter,” Bank of Canada Working Paper 12–06 (Ottawa: Bank of
Canada).
Bakker, B. and C. Klingen, 2012, How Emerging Europe Came Through the 2008/09 Crisis: An Account
by the Staff of the IMF’s European Department (Washington: International Monetary Fund).
Bank for International Settlements, 2010, “Macroeconomic Assessment Group: Assessing the
Macroeconomic Impact of the Transition to Stronger Capital and Liquidity Requirements,
Interim Report, August (Basel: Bank for international Settlements).
Bean, Charles, Matthias Paustian, Adrian Penalver, and Tim Taylor, 2010, “Monetary Policy after the
Fall,” in Macroeconomic Challenges: The Decade Ahead: Proceedings of the Federal Reserve Bank
of Kansas City Economic Symposium at Jackson Hole (Kansas City: Federal Reserve Bank).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
6
6
1
1
Bernanke, Ben S., 2005, “The Global Saving Glut and the U.S. Current Account Deficit,” Speech,
April 14 (Washington: Federal Reserve Board).
Bernanke, Ben and Mark Gertler, 1989, “Agency Costs, Net Worth, and Business Fluctuations,”
American Economic Review, Vol. 19, No. 4.
Bernhard, S., and T. Mancini-Griffoli, 2011, “Gauging Taylor rules: Do differences make a difference?”
mimeo (Zurich: Swiss National Bank).
Bhattacharya, Sudipto, 1979, “Imperfect Information, Dividend Policy, and "The Bird in the Hand"
Fallacy, The Bell Journal of Economics, 10 (1), 259–270.
Bhattacharya, Sudipto, 1982, “Aspects of Monetary and Banking Theory and Moral Hazard,” Journal
of Finance, 37, 371-384.
Borio, Claudio, and Haibin Zhu, 2008, “Capital Regulation, Risk-taking and Monetary Policy: A
Missing Link in the Transmission Mechanism?” BIS Working Paper 268, December (Basel: Bank
for International Settlements).
Brunnermeier, M., 2009, “Deciphering the 2007–08 Liquidity and Credit Crunch,” Journal of Economic
Perspectives, Vol. 23, No. 1, pp. 77–100.
Brunnermeier, M. and L. Pedersen, 2009, “Market Liquidity and Funding Liquidity,” Review of
Financial Studies, Volume 22, Number 6, p.2201–2238.
Bruno, Valentina, and Hyun Song Shin, 2012, “Capital Flows and the Risk-Taking Channel of
Monetary Policy,” presented at the 11th BIS Annual Conference (Basel: Bank for International
Settlements).
_________, 2012b, “Assessing Macroprudential Policies: Case of Korea,” presented at the symposium
of the Scandinavian Journal of Economics on Capital Flows.
Cardarelli, R., D. Igan and A. Rebucci, 2008, “The changing housing cycle and the implications for
monetary policy,” IMF World Economic Outlook, April (Washington: International Monetary
Fund).
Cecchetti, Stephen, and Marion Kohler, 2012, “When Capital Adequacy and Interest Rate Policy are
Substitutes (and When They are Not),” BIS Working Paper 379 (Basel: Bank for International
Settlements).
Cetorelli, Nicola, and Linda S. Goldberg, 2012, “Banking Globalization and Monetary Transmission,”
Journal of Finance, Vol. 67, No. 5, pp. 1811–843.
Christensen, Ian, Cesaire Meh, and Kevin Moran, 2011, “Bank Leverage Regulation and
Macroeconomic Dynamics,” Bank of Canada Working Paper 11–32 (Ottawa: Bank of Canada).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
6
6
2
2
INTERNATIONAL MONETARY FUND
Claessens, Stijn, M. Ayhan Kose, and Marco E. Terrones, 2008, “What Happens During Recession,
Crunches and Busts?” IMF Working Paper 08/274 (Washington: International Monetary Fund).
Cochrane, J., 2007, “Comments on ‘Macroeconomic implications of changes in the term premium’ by
Glenn Rudebusch, Brian Sack and Eric Swanson,” FRB of St. Louis Review, July/August (St. Louis:
Federal Reserve Bank).
Committee on the Global Financial System, 2012, “Operationalising the Selection and Application of
Macroprudential Instruments,” CGFS Papers 48 (Basel: Bank for International Settlement).
Crowe, Christopher W., Deniz Igan, Giovanni Dell’Ariccia, and Pau Rabanal, 2011, “How to Deal with
Real Estate Booms,” IMF Staff Discussion Note 11/02 (Washington: International Monetary
Fund).
Dell’Ariccia, G., D. Igan, and L. Leaven, 2012a, “Credit Booms and Lending Standards: Evidence from
the Subprime Mortgage Market,” Journal of Money, Credit and Banking, vol. 44, pages 367–
384.
Dell’Ariccia G., D. Igan, L. Laeven, and H. Tong, with B. Bakker and J. Vandenbussche, 2012b, “Policies
for Macrofinancial Stability: How to Deal with Credit Booms”, IMF Staff Discussion Note 12/06
(Washington: International Monetary Fund).
Del Negro, Marco and Christopher Otrok, 2007, “99 Luftbllons: Monetary Policy and the House Price
Boom Across U.S. States,” Journal of Monetary Economics, Vol. 54, No. 1962–985.
English, William B., Skander J. Van den Heuvel, and Egon Zakrajsek, 2012, “Interest Rate Risk and
Bank Equity Valuations,” Finance and Economics Discussion Series 2012–26, Division of
Research & Statistics and Monetary Affairs (Washington: Federal Reserve Board).
European Bank Coordination (“Vienna”) Initiative, 2011, “Report by the Public-Private Sector
Working Group on Local Currency and Capital Market Development.”
European Systemic Risk Board, 2011, “Recommendation of the ESRB of September 21, 2011 on
Lending in Foreign Currency.”
Farhi, Emmanuel and Jean Tirole, 2012, “Collective Moral Hazard, Maturity Mismatch and Systemic
Bailouts,” American Economic Review, February 2012, Vol. 102(1).
Favara, G., and S. Imbs, 2010, “Credit Supply and the Price of Housing” CEPR Discussion Papers 8129
Federico, Pablo, Carlos A. Vegh and Guillermo Vuletin, 2012, “Macroprudential Policy Over the
Business Cycle,” mimeo, University of Maryland.
Gan, Jie, 2004, “Banking Market Structure and Financial Stability: Evidence from the Texas Real Estate
Crisis in the 1980s,” Journal of Financial Economics, Vol. 73, pp. 567–601.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
6
6
3
3
Gertler, Mark and Simon Gilchrist, 1994, “Monetary Policy, Business Cycles, and the Behavior of Small
Manufacturing Firms,” The Quarterly Journal of Economics, vol. 109(2), pages 309–40.
Geanakoplos, John, 2010, “Solving the Present Crisis and Managing the Leverage Cycle,” FRB of New
York Economic Policy Review, August, 101–131 (New York: Federal Reserve Bank).
Giacomini, R., and H. White, 2006, “Tests of Conditional Predictive Stability," Econometrica, 74(6),
1545–1578.
Glocker, Christian and Pascal Towbin, 2012a, “Reserves Requirements for Price and Financial
Stability: When Are They Effective?” International Journal of Central Banking, 8 (1), 66–113.
________, 2012b, “The Macroeconomic Effects of Reserve Requirements,” Document de Travail 374
(Paris: Banque de France).
Goodhart, Charles, Dimitrios P. Tsomocos, and Alexandros P. Vardoulakis, 2009, "Foreclosures,
Monetary Policy and Financial Stability,” Conference Proceedings of the 10th International
Academic Conference on Economic and Social Development, Moscow.
Gray, Simon, 2011, “Central Bank Balances and Reserves Requirements,” IMF Working Paper 11/36
(Washington: International Monetary Fund).
Hahm, Joon-Ho, Frederic S. Mishkin, Hyun Song Shin, and Kwanho Shin, 2012, “Macroprudential
Policies in Open Emerging Economies,” NBER Working Paper Series 17780 (Cambridge,
Massachusetts: National Bureau of Economic Research).
Hellman, Thomas, F., Kevin C. Murdock and Joseph Stiglitz, 2000, “Liberalization, moral hazard in
banking and prudential regulation: are capital requirements enough?” American Economic
Review, 90, 147–165.
Hildebrand, P. M., 2008, “Is Basel II enough? The Benefits of a Leverage Ratio,” Speech at the
Financial Markets Group Lecture, December 15, Available at
http://www.bis.org/review/r081216d.pdf
(London: London School of Economics).
Iacoviello, Matteo, and Stefano Neri, 2010, “Housing Market Spillovers: Evidence from an Estimated
DSGE Model,” American Economic Journal, Macroeconomics, Vol. 2, No 2, pp. 125–64, April.
Igan, Deniz, and Heedon Kang, 2011, “Do Loan-to-Value and Debt-to-Income Limits Work? Evidence
from Korea,” IMF Working paper 11/297 (Washington: International Monetary Fund).
Illing, Gerhard, 2007, “Financial Stability and Monetary Policy―A Framework,” CESifo Working Paper
1971.
International Monetary Fund, 2008, “The Changing Housing Cycle and the Implications for Monetary
Policy,” World Economic Outlook, April, Chapter 3.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
6
6
4
4
INTERNATIONAL MONETARY FUND
_________, 2009, “Lessons for Monetary Policy from Asset Price Fluctuations, World Economic
Outlook,” October, Chapter 3.
_________, 2011a, “Housing Finance And Financial Stability—Back to Basics?” Chapter 3, Global
Financial Stability Report, April.
_________, 2011b, Regional Economic Outlook Europe, October.
_________, 2012, “Republic of Korea: 2012 Article IV Consultation Staff Report,” September.
Ioannidou, Vasso P., Steven Ongena and José-Luis Peydró, 2009, “Monetary Policy and Subprime
Lending: a Tall Tale of Low Federal Funds Rates, Hazardous Loans and Reduced Loan Spreads,”
European Banking Centre Discussion Paper 2009–045.
Krznar, Ivo, Akira Otani, Xiaoyong Wu and Cheng Hoon Lim, 2013, “The Macroprudential Framework:
Policy Responsiveness and Institutional Arrangements,” forthcoming, (Washington:
International Monetary Fund).
Jiménez, Gabriel, Steven Ongena, José-Luis Peydró and Jesus Saurina, 2009, “Hazardous Times for
Monetary Policy: What do Twenty-Three Million Bank Loans Say about the Effects of Monetary
Policy on Credit Risk-taking?” Bank of Spain Working Papers 833 (Madrid: Banco de España).
Jiménez, Gabriel, Steven Ongena, José-Luis Peydró and Jesus Saurina, 2012, “Macroprudential Policy,
Countercyclical Bank Capital Buffers and Credit Supply: Evidence from the Spanish Dynamic
Provisioning Experiments,” mimeo, Tilburg University.
Jonsson, Asgeir, 2009, “Why Iceland?” (New York: McGraw-Hill).
Kannan, Prakash, Pau Rabanal, and Alasdair Scott, 2009, “Monetary and Macroprudential Policy
Rules in a Model with House Price Booms,” IMF Working Paper 09/251(Washington:
International Monetary Fund).
Kashyap, Anil K., Jeremy Stein and Samuel Hanson, 2010, “An Analysis of Substantially Heightened
Capital Requirements on Large Financial Institutions,” mimeo, University of Chicago.
Landier, Augustin, David Sraer, and David Thesmar, 2011, “The Risk-Shifting Hypothesis: Evidence
from Sub-Prime Originations,” presented at the IMF 12th Jacques Polak Annual Research
Conference (Washington: International Monetary Fund).
Lee, Jong Kyu, 2012, “The Operation of Macroprudential Policy Measures: The Case of Korea,”
Mimeo (Seoul: Bank of Korea).
Levin, A., V. Wieland, and J. C. Williams, 2003, “The Performance of forecast-based Monetary Policy
Rules Under Model Uncertainty," American Economic Review, 93(3), 622–645.
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
6
6
5
5
Lim, C. H., F. Columba, A. Costa, P. Kongsamut, A. Otani, M. Saiyid, T. Wezel, X. Wu, 2011,
“Macroprudential Policy: What Instruments and How Are They Used? Lessons from Country
Experiences”, IMF Working Paper 238 (Washington: International Monetary Fund).
Merrouche, Ouarda, and Erlend Nier, 2010, “What Caused the Global Financial Crisis? Evidence on
the Drivers of Financial Imbalances 1999–2007,” IMF Working Paper 10/265 (Washington:
International Monetary Fund).
Mian, A., and A. Sufi, 2009, “The Consequences of Mortgage Credit Expansion: Evidence from the
U.S. Mortgage Default Crisis,” The Quarterly Journal of Economics, Vol. 124, No. 4, pp. 1449–
1496.
Moreno, Ramon, 2008, “Monetary Policy Transmission and the Long-term Interest Rate in Emerging
Markets,” BIS Working Papers 35 (Basel: Bank for International Settlements).
Myers, Stewart C. and N. Majluf, 1984, “Corporate Financing and Investment Decisions When Firms
have Information That Investors Do Not Have,” Journal of Financial Economics, Vol. 13, pp.
187–222.
Nier, Erlend W. and Lea Zicchino, 2008, “Bank Losses, Monetary Policy and Financial Stability—
Evidence on the Interplay from Panel Data,” IMF Working Paper No. 08/232 (Washington:
International Monetary Fund).
Orphanides, A., and J. C. Williams, 2006, “Monetary Policy with Imperfect Knowledge," Journal of the
European Economic Association, 4(2–3), 366–375.
_________, 2008, “Learning, Expectations Formation and the Pitfalls of Optimal Control Monetary
Policy," Journal of Monetary Economics, 55S, S80–S96.
Panetta F., P. Angelini –coordinators- and U. Albertazzi, F. Columba, W. Cornacchia, A. Di Cesare, A.
Pilati, C. Salleo and G. Santini, 2009, “Financial Sector Pro-cyclicality: Lessons from the Crisis,”
Bank of Italy Occasional Papers, No. 44 (Roma: Banca d’Italia).
Pann J., R. Seliger and J. Ubelies, 2010, “Foreign Currency Lending in Central, Eastern and
Southeastern Europe:the Case of Austrian Banks,” Austrian National Bank Financial Stability
Report, December 2010, pp. 60–80 (Vienna: Oesterreichische Nationalbank).
Poszar, Zoltan, Tobias Adrian, Adam Ashcraft, and Hayley Boesky, 2010, “Shadow Banking,” FRB of
New York, Staff Report, No. 458 (New York: Federal Reserve Bank).
Ree, Jack Joo K., Kyungsoo Yoon, and Hail Park, 2012, “FX Funding Risks and Exchange Rate
Volatility-Korea’s Case,” IMF Working Paper 12/268 (Washington: International Monetary
Fund).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
6
6
6
6
INTERNATIONAL MONETARY FUND
Rosenberg C., and M. Tirpák, 2008, “Determinants of Foreign Currency Borrowing in the New
Member States of the EU,” IMF Working Paper 08/173 (Washington: International Monetary
Fund).
Rudebusch, G., B. Sack and E. Swanson, 2007, “Macroeconomic Implications of Changes in the Term
Premium,” FRB of St. Louis Review, July/August (St. Louis: Federal Reserve Bank).
Sengupta, Rajdeep, 2010, “Alt-A: The Forgotten Segment of the Mortgage Market,” Federal Reserve
Bank of St. Louis Review, Vol. 92(1), pp. 55–71 (St. Louis: Federal Reserve Bank).
Shiller, Robert, J., 2007, “Historic Turning Points in Real Estate,” Cowles Foundation Discussion Paper
No. 1610 (New Haven: Yale University).
Shin, Hyun Song, 2005, “Financial System Liquidity, Asset Prices and Monetary Policy,” in The
Changing Nature of the Business Cycle, RBA Annual Conference Volume, No. acv 2005-16
(Sydney: Reserve Bank of Australia).
_________, 2011, “Global Banking Glut and Loan Risk Premium,” presented 12th Jacques Polak Annual
Research Conference (Washington: International Monetary Fund).
Steiner, K., 2011, “Households’ Exposure to Foreign Currency Loans in CESEE EU Member States and
Croatia”, Focus on Economic Integration, Q1 2011, pp. 6–24.
Tarullo D. K., 2012, “Shadow Banking After the Financial Crisis,” speech at the Federal Reserve Bank
of San Francisco Conference on Challenges in Global Finance: The Role of Asia, June 12, 2012
(San Francisco: Federal Reserve Bank).
Taylor, John B., 2007, “Housing and Monetary Policy,” in Housing, Housing Finance, and Monetary
Policy: Proceedings of Federal Reserve Bank of Kansas City Economic Symposium at Jackson
Hole, August 31–September 1 (Kansas City: Federal Reserve Bank).
Taylor, J. B., and V. Wieland, 2009, “Surprising Comparative Properties of Monetary Models: Results
from a New Database,” NBER Working Paper Series 14849 (Cambridge, Massachusetts:
National Bureau of Economic Research).
Tovar, Camilo E, Mercedes Garcia-Escribano, and Mercedes Vera Martin, 2012, Credit Growth and
the Effectiveness of Reserves Requirements and Other Macroprudential Instruments in Latin
America,” IMF Working Paper 12/142 (Washington: International Monetary Fund).
Turner, Adair, 2012, “Macroprudential Policy in Deflationary Times,” speech during the Financial
Policy Committee Regional Visit to Manchester, July 20 (London: Financial Services Authority).
Unsal, Filiz, 2011, “Capital Flows and Financial Stability: Monetary Policy and Macroprudential
Responses,” IMF Working Paper 11/189 (Washington: International Monetary Fund).
THE INTERACTION OF MONETARY AND MACROPRUDENTIAL POLICIES: BACKGROUND PAPER
INTERNATIONAL MONETARY FUND
6
6
7
7
Vandenbussche, J., U. Vogel and E. Detragiache, 2012, “Macroprudential Policies and Housing
Prices—A New Database and Empirical Evidence for Central, Eastern and Southeastern
Europe,” IMF Working Paper 12/303 (Washington: International Monetary Fund).
Vargas, Hernando, Carlos Varela, Yanneth Betancourt and Norberto Rodriguez, 2010,“Effects of
Reserve Requirements in an Inflation Targeting Regime: the Case of Colombia,” Borradores de
Economía, No. 587 (Bogotá: Banco de la República).
Williams, J. C., 2003, “Simple Rules for Monetary Policy," FRB of San Francisco Economic Review (San
Francisco: Federal Reserve Bank).
Wong, Eric, Tom Fong, Ka-fai Li and Henry Choi, 2011, “Loan-to-Value Ratio as a Macroprudential
Tools—Hong-Kong’s Experience and Cross-Country Evidence,” HKMA Working Paper, 01/2011
(Hong Kong: Hong Kong Monetary Authority).
Woodford, Michael, 2011, “Inflation Targeting and Financial Stability,” mimeo, Columbia University.
Zettelmeyer, J., P. Nagy and S. Jeffrey, 2010, “Addressing Private Sector Currency Mismaches in
Emerging Europe,” EBRD Working Paper No 115 (London: European Bank for Reconstruction
and Development).
