Transcript DISCUSSION by David Vavra

The Determinants of Cross-border
Bank Flows to Emerging Markets: New
Empirical Evidence on the Spread of
Financial Crises
by Herrmann and Mihaljek
Discussion
David Vavra
16th Dubrovnik Economic Conference
June 24 2010
Discussion Points




What I have learnt
What I would like to learn
What puzzled me
Data and specification issues
What I have learnt (I)


Global and country specific factors are
significant determinants of cross-border flows
Global risk aversion and market volatility
most important factors behind the decline in
flows in the crisis
What I have learnt (II)


Deeper financial integration and sounder
banking systems limited the decline in CESE
Fixed exchange rate limited the decline in
CESE
Long-Term Impact of Crisis
GDP Level and Potential Output (CNB‘s Forecast)
4000
Potential GDP (PF-Approach)
Real GDP (s.a.)
Linear trend before crisis (10Y)
3500
3000
2500
2000
1999
2001
2003
2005
2007
2009
2011
2013
Permanent loss in output level, but not in growth rate (in line with IMF, WEO,
Oct 2009)?
Real Convergence



Real convergence observed only in the past 10 years
Question: will it resume?
Importance of free capital flows
Czech GDP in PPP as % of EU27
82
80
78
76
74
72
70
68
66
64
62
1997
1998
Source: Eurostat
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Real Convergence and bank
cross-border flows



Unique data set on bank flows
Banks were the main source of capital for catch-up growth in CESE
Financial markets in CESE dominated by banks
External positions of BIS reporting banks
(ex rate edjusted, q-o-q, in % GDP)
What I would like to learn
about: Policy Implications

What is the effect of regulation on these flows?


What can policy do about affecting these determinants?


What was the effect of specific policy actions preventing the
decline in cross-border flows. e.g. the Vienna initiative.
How come that CESE performed better in region specific
factors in the crisis than other EMs, when CESE flows
were actually hit more?


Could we distinguish between a cross-border flows to banking
and non-banking sectors?
The openness and bank health story not too convincing
Fixed ex rate as a shelter in the crisis

Conducive to more flows and perhaps also more growth?
Floaters’ Output Performance on
Average Better than Peggers’
Max Extent of Depreciation since July 2008*
50
Poland
45
40
35
Romania
30
25
Hungary EA 15
20
CR
15
10
5
0
-20
Latvia
-15
Lithuania
Estonia
-10
Bularia
Slovenia Slovakia
-5
0
5
GDP Growth 2009
* to Euro, EA15 to USD
Source: Eurostat, CNB
9
Other comments: The choice
of explanatory variables

Endogeneity and causality


Risk-return characteristics of loans





Changing sings, decreasing significance with lag effects
UIP, or excess return performance?
Should really a high interest rate differential imply higher flows?
Exchange rate depreciation reducing flows?
Nominal vs real rates of return
CA deficit and credit growth as explanatory variables – why
surprised by positive signs
Other comments: Data
issues




The flow data are adjusted for exchange rate
movements – wouldn’t some real measure or relative to
GDP be even more appropriate?
How does the database incorporate NPLs – does it keep
loans at a book value?
Data are gross – e.g. Vienna initiative made participating
banks keep their exposures – this effect is not captured
Each type of flow finances different activity and should
have different determinants



Is capital included – an important type of cross-border flow
Subsidiary flow – mortgages and small business
Non-bank flows – businesses