Transcript The Financial Accelerator, Globalization and Output Growth Volatility

Comments on
“The Financial Accelerator, Globalization and
Output Growth Volatility”
Bruno Ćorić, Geoff Pugh
by Saša Žiković
Motivation





Considerable evidence that US GDP growth in early ‘80
became less volatile than in previous decades – “Great
Moderation” process.
The paper builds upon the premise of similar behaviour
around the world.
Authors ask a very interesting question – “Why did so
many countries across the world experience reduced
output volatility in the decades preceding the current
global downturn?”
Investigation from a global perspective into the
possibility that international diversification of economic
agents’ net worth influenced the strength of the
Financial Accelerator which, in turn, contributed to
changes in GDP growth volatility.
The contribution of the present paper is to test this
hypothesised explanation by econometric analysis of 85
countries over the period 1970 to 2004.
Motivation




The authors opt for the view that effect of changes in
aggregate economic activity on net worth may have
weakened the Financial Accelerator, thereby reducing
output growth volatility.
First hypothesis (H1): International net worth
diversification reduced output growth volatility.
The hypothesized mechanism depends on the
assumption of asymmetric economic shocks across
national economies; that is, of incomplete GDP growth
synchronization.
Second hypothesis (H2): Moderating or stabilizing
influence of net worth diversification may have been
merely transient, as its capacity to stabilize the global
economy
was
eventually
undermined
by
the
synchronizing effects of globalization
Good points of the paper







Very good written and potentially an
important paper
Builds upon good theoretical and empirical
basis
Relevant and up to date literature
Uses adequate data and variables
Robustness checks in place
Econometrics are done carefully and every
detail is taken care of
All around an excellent piece of work
Findings




Negative coefficients on net worth diversification (NWD)
= increased international diversification of agents’ net
worth connected to lower GDP growth volatility.
Positive coefficients on inflation volatility = higher
monetary disturbance on average leads to higher GDP
growth volatility.
Positive coefficients on trade openness = increase in
international trade intensity is associated with higher
growth volatility.
Results are robust with respect to a different definition
of output growth volatility as well as to different
estimation techniques, unlikely to be driven by
endogeneity or spurious regression.
Findings?



The original question “...Why did so many countries
across the world experience reduced output volatility...”
is not answered!
Is this the case of authors’ failure to answer the
question they start with or is the intro poorly written.
I trust it’s the second – rewrite the introduction since it
is misleading the reader
What about causality – what is the cause and what is
the effect?
 Is NWD:
1) a piece of a puzzle that is causing the moderation or
2) is it that the moderation is causing investors to
diversify and seek new opportunities?

Turning to financial accelerator hypothesis




The financial accelerator hypothesis says that credit
market distortions magnify economic shocks.
Disturbances that would be small if markets were
efficient are exaggerated and prolonged due to
imperfections in credit and loan markets. Credit
market distortions destabilize the economy.
The financial accelerator hypothesis is important
because standard business cycle models require large,
persistent disturbances to mimic the business cycles
observed in the data.
Because the financial accelerator amplifies and
propagates shocks, it can potentially explain why
business cycles are so significant even though the
observed shocks are not.
A word of warning
A problem related to the existing literature applies to
the empirical evidence giving support to the financial
accelerator hypothesis.
 Most of these studies are either:
1) based on reduced form analysis not aimed at revealing
the causal structural interplay among the variables, or
2) structural specifications not given support by data in
the sense of being misspecified.


Some recent examples: Lown and Morgan (2006),
Swiston (2008) and Bayoumi and Melander (2008)
A word of warning


Misspecification - In the case of estimating simultaneous
equation models that have been exactly identified
through e.g. imposing a priori restrictions on their
contemporary causal structure and assuming a diagonal
structural covariance matrix, one risks ending up with
models that do not adequately represent the causal
structure of the data and thus induce a simultaneity bias
in estimation through imposing an improper causal
structure.
The reason for this is related to the fact that one never
can test for the exactly identifying restrictions of a
structural model.
Questions and suggestions







The paper is successful at finding a significant factor
to “moderation” but its explanatory power is really
low!
The paper identifies only a piece of a puzzle - that is
why the original question cannot be answered!
Hypothesis 1 proven partially – we cannot be certain
about the direction of causality?!
Hypothesis 2 – weak evidence, for a publishing
purposes it would be better left for the time being.
Methodology standard and correct – I am worried a bit
about oversmoothing.
The paper is measuring unconditional volatility!
The authors could use ARCH models for modelling
conditional volatility, even EWMA would be better than
SMA standard deviation.
Questions and suggestions


Construction of NWD var: it is constructed to decrease variance
but the problem of doubling FDI remains!
Example: A world is composed of only 2 countries. US invests
5% of its FDI/GDP in China and vice versa. What is the amount
USA
China
of FDI in the world?
A
5%
L
5%
A
5%
10%
10%
20% ?
FDI/GDP:
5%
+ 5%
10%
L
5%
Questions and suggestions

Policy implications are still left unclear – should some
action be taken? What should a country do?
The paper should try to give a better insight into the
mechanism behind this relationship – why are the two
thing moving in opposite directions!?

Is the “Great Moderation” a good thing?

If NO, why?
If YES, for who?



Is it playing in favour of US and EU? Should they be
worried about BRICK growth and closing of the gap!?
Sustainable development – the developing can never
catch the developed
It would be interesting to single out BRICK countries to
see the specific effects on them
Something to think about




The main part of world “moderation” happened after
the two oil shocks – are these the crucial moments that
shaped future GDP movement?
Slowing down in world trade in the wake of the
industrial country recession – decrease in exports from
less developed countries
Significant support is found for the positive link
between bilateral trade intensity and business cycle
comovement in a cross-section of industrialized country
pairs (trade-comovement puzzle) - see Drozd, Nosal
(2008)
Is this the financial accelerator or stabilizer equilibria
we were seeing at work? – see House (2006)