Macroeconomic Shocks and Foreign Bank Assets

1
Macroeconomic Shocks and Foreign Bank Assets

• Andrea Schertler (IfW, Kiel)
• Claudia M. Buch (University of Tübingen and IfW)
• Kai Carstensen (IfW, Kiel)
• London
• March 24, 2006

2
Aim of the paper

• Foreign bank assets accounted for 130 percent of
  world GDP at the end of 2003
• These financial linkages are a potential channel
  through which domestic shocks are transmitted
  between countries
• Recent literature has focused on bank lending to
  emerging markets (e.g. Jeanneau and Micu 2002)
• We analyze how foreign bank assets in OECD
  countries respond to domestic and foreign
  macroeconomic shocks

3
Extending the literature

• Bi-directional transmission of macroeconomic
  shocks between the OECD countries
• Three main macroeconomic shocks supply, demand,
  and monetary shocks
• Two different methods Observable macroeconomic
  indicators and empirical shocks derived from a
  structural vector autoregressive (VAR) model

4
What are the main questions?

• Do cross-border assets of banks contribute to the
  transmission of shocks across countries?
• Do cross-border assets react to macroeconomic
  developments in a way predicted by theory?
• How important is the degree of financial
  openness?

5
Empirical test

• Baseline regression
• Regress changes in banks foreign assets on
  changes in macroeconomic developments at home and
  abroad
• Extensions
• Separate Euro zone
• Proxies for macroeconomic shocks from panel VAR
• Impact of transaction costs

6
Data on foreign asset holdings of banks

• Bank for International Settlements
• Consolidated banking statistics
• Comprehensive measures of foreign assets
• Bilateral data
• No breakdown by currency, type of asset, or time
• Sample
• 1999-2003
• Quarterly data
• 17 OECD reporting and recipient countries

7
Foreign assets of domestic banks (mean in of
total foreign assets)
8
Foreign liabilities of domestic banks (mean in
of total foreign liabilities)
9
Specification of the panel model

• Dependent variable dlog (banks foreign assets)
  (USD)
• Explanatory variables
• Difference in domestic and foreign GDP growth
• Difference in domestic and foreign interest rates
• Difference in domestic and foreign inflation
• US GDP growth
• Bilateral exchange rate

10
Specification of the panel model

• Dynamic panel data model
• GMM estimator (Blundell and Bond 1998)
• Finite sample correction (Windmeijer 2005)
• Test on over-identifying restrictions
• Test on serial correlation
• Full set of country fixed effects
• Quarterly dummies
• Annual dummies
• general-to-specific methodology

11
Expected signs of the explanatory variables

• Macroeconomic developments (Sutherland 1996)
• GDP growth differential ()
• Interest rate differential ()
• Inflation rate differential ()
• Control variables
• US GDP growth ()
• Relative exchange rates (quantity notation) ()

12
Baseline Specification
13
Extensions

• Euro zone versus non-Euro zone
• Most results driven by the Euro area
• Valuation effects insignificant even outside the
  Euro area
• Estimation of a panel VAR model for the
  identification of shocks Similar results
• Impact of transaction costs
• Distance helps in explaining cross-boarder assets
• Interaction terms insignificant

14
Summary

• Does cross-border lending of banks contribute to
  the transmission of shocks across countries?
  YES.
• Foreign asset holdings react to macroeconomic
  shocks.
• Some bilateral financial linkages are important.
• Overall explanatory power is low.
• Does cross-border lending react to macroeconomic
  developments in a way predicted by theory? YES.
• Reaction to supply and monetary shocks is in line
  with theory.
• Reaction to demand shocks is insignificant.
• Is the degree of financial openness important?
  NO.
• Financial openness matter for for asset holdings.
• Financial openness matters not for shock
  transmission.