The document discusses a workshop on monetary policy spillovers and independence. It aims to examine how accommodative monetary policy in advanced economies may impact financial stability risks in emerging markets through various transmission channels. Specifically, it seeks to analyze potential correlations between financial cycles in advanced and emerging economies, assess how emerging markets responded to supportive monetary policies, and map macroprudential policy tools to different financial stability risks. Key challenges include generating long-term financial cycle data for emerging market countries.

1. BENJAMIN HUSTON
SANAA NADEEM
INCI OTKER-ROBE
Workshop on Monetary Policy: Spillovers and Independence
(December 18, 2014)
ACCOMMODATIVE MONETARY POLICY:
BREATHING SPACE OR BREEDING RISKS FOR
EMERGING MARKETS?
THE ROLE FOR MACROPRUDENTIAL POLICY

2. Outline
 Motivation
 Questions we ask
 What we do?
 Contribution and Challenges
 Next steps

3. Motivation
 Major AE central banks have been implementing extraordinarily
supportive monetary policy to restore the functioning of markets
and support economic activity/soundness of the financial system
 These supportive policies through CMP and UMP tools (+)
 Helped alleviate the market turmoil and reduce tail risk
 Benefitted the global economy (including EMDEs) through lower borrowing
costs for the private and public sectors
 But also raised policy challenges for EMFMs (-)
 Challenges related to buildup of financial stability risks in EMFMs (e.g., low
interest, high liquidity environment in AEs and search-for-yield in EMFMs)
 Challenges related to normalization of monetary policy and potential impact
on EMFMs (e.g., through capital flow reversals, balance sheet effects)

4. Questions we raise
 Risks? Are there indications of rising financial stability risks in
EMFMs associated with CMP/UMP in AEs
 Extent? How does the degree of vulnerability/risks depend on
how the policy space was used
 Policy? What can EMFMs do to protect against the adverse
consequences of UMP normalization/prolonged low interest rates
 MaPP? Is there a role of macroprudential policy to mitigate
potential financial stability risks of loose monetary policy or its
normalization
Implications for Policymakers/Macro-financial surveillance

5. What we aim to do (1)
 Consider the transmission channels of AE monetary policy to EMFMs
 Examine correlations of AE interest rates with EMFM short and long-term
interest rates, ERs, equity/house prices, capital inflows, credit growth…
 Control for EMFM’s macro-financial characteristics and policy frameworks
(e.g., degree of financial openness, ER regime, capital account openness…)
 Derive estimates of EMFM financial cycles
 Explore co-movements/correlations across financial cycles (between
AE and EM cycles) and between EMFM financial cycles and AE rates
 Assess the indication of rising financial risks in EMFMs based on the
position in the financial cycle
 For a sample of ~40 EMFMs

6. What we aim to do (2)
 Assess how EMFMs used the supportive monetary policy
space (indicators before/after UMP)—stylized facts
 To address macro and financial imbalances?
 To enhance financial/macro resilience and build policy space? or
 To build leverage/expand credit to nonproductive sectors?
 Go more granular into the likely sources of financial risks
(credit, FX, liquidity, etc)—key FSIs
 Map the risks to MaPP tools to address the buildup of
different risks, as well as to insure against adverse
consequences for EMFMs of AE exit

7. Contributions, challenges, next steps
 Key Contributions:
 The angle to look into the role of MaPP in mitigating emerging financial
risks and mapping the tools to risks for the sample EMFMs
 Generation of financial cycles for EMDEs
 Analyzing evolution and correlations across financial cycles
 Key Challenges:
 Generation of financial cycles for EMDEs—Absence of long data series
for the key components of financial cycles
 Innovation is to use imputation technique to derive the underlying data
series to compute the financial cycles
 Next steps: Preliminary draft around Spring 2015

8. Channels of Transmission—Example
 Rolling correlations of EM ST
rates with AE rates
 Rolling correlations of EM LT
rates with AE rates
-1.00
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
1.00
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
US EU Japan
EM rolling correlations - Long term interest rates
-1.00
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
1.00
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
US EU Japan
EM rolling correlations - short term interest rates

9. Transmission channels controlling for ER regime
 Correlations of short term rates  Correlations of long term rates
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
2000-2006 2007-2009 2010-2014
Pegged Managed Floating
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
2000-2006 2007-2009 2010-2014
Pegged Managed Floating

10. Transmission channels controlling for financial openness
(Correlations of EM short-term rates with AE rates)
-1.00
-0.80
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
1.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
2000-2007
2008-2014
Financial Openness (percentage of GDP)

11. Replicating BIS Financial Cycles

12. IMF results
Financial cycles derived as the simple average of the cyclical component of Credit,
Housing price, and Credit-to-GDP data
BIS results
Method: Apply “bandpass filter” to transformed time series of credit,
housing, and credit/GDP ratios (Borio et al 2012)

13. Two Approaches to Financial Cycles
 Frequency Analysis
 Apply a “bandpass filter” to transformed time series of credit,
housing, and credit/gdp ratios
 Take the simple average of the cyclical component identified in these
three series to get the “financial cycle”
 Turning Point Analysis
 Apply a modified version of the Bry and Boschan Quarterly
algorithm to transformed time series of credit, housing, and
credit/GDP ratios to identify series-specific “peaks” and “troughs”
 Pool series-specific peaks and troughs and repeat the process to
identify “financial cycle” peaks and through.
*The Frequency Analysis approach is used almost exclusively in practice
and its results are synonymous with the term financial cycle

14. Example: Advanced Economy Financial Cycles

15. Financial Cycle Interconnectivity
 Illustrative Granger-Causality
Network (1% significance level;
1985-2014)
 The network of global financial
cycles is highly connected
 Exhibits high levels of “feedback”
(mutual granger-causation) among
advanced economies and between
advanced and emerging economies

16. Addressing Data Constraints

17. The financial cycle imputation workflow
17
Utilize iterated univariate procedures* (“chained equations”) that work in parallel
to predict (“impute”) missing values of the response variable from previously
observed response and predicator values
Use chained
equations to impute
missing data from
observed data many
times
Gather
(incomplete)
data
Perform
subsequent
analysis
Apply bandpass
filter to extract
financial cycles
from each
imputed dataset
Pool results into a
single financial
cycles for each
country
*van Buuren S, Brand JPL, Groothuis-Oudshoorn CGM, Rubin DB (2006b). “Fully Conditional Specification in Multivariate Imputation.” Journal of Statistical Computation and
Simulation, 76(12), 1049–1064.

18. Experimental Imputed EM Housing Price Data
*Black line represents observed data. Red line represents data imputed using
chained equations.
Short imputation chain Long imputation chain

19. Appendix

20. The United Kingdom
and the Euro Area are
the most connected in
the overall network
The closer a node is to the center the more “important”, as measured by
betweeness centrality, the node is to the structure of the network,
Betweenness Centrality Graph of Granger-Causality Network
(1% Significance Level; 1985-2014)

21. The U.S. and Korea
are most similar in
terms of in terms of
connectivity.
The other countries
(excluding Germany)
have similar levels of
connectivity.
Nodes groups denote “similarity”, as measured by betweenness centrality
scores. Nodes within each group are most similar along this dimension.
Betweenness Community of Granger-Causality Network
(1% Significance Level; 1985-2014)

22. A betweenness centrality-
based cost function is
used to determine “splits”
Partitions formed by splits
represent distinct
"communities”
First Split
Second Split
Betweenness Community Dendogram of Granger-Causality Network
(1% Significance Level; 1985-2014)