Davide Furceri (OECD) (264 KB )

advertisement
Discussion of “Fiscal Shocks and the Real
Exchange Rate”
by
Agustin S. Bénétrix & Philp R. Lane
Davide Furceri
OECD, University of Palermo
Banco de Espana Conference
Interactions between monetary and fiscal policies
Madrid, 25-26 February 2010
Aim of the paper
• Estimate the impact of shocks to government spending on the real
exchange rate for a panel of EMU member countries:
•The authors estimate a panel VAR for eleven EMU countries using a threevariable system (spending, output, real exchange rate);
•Differentiating among components of government spending:
i) government consumption; i.a) non-wage component; 1.b) wage
component; ii)government investment.
Main results
• A shock to government absorption (consumption + investment)
appreciates the real exchange rate;
• Shocks to government investment have larger and more persistent
effects than government consumption;
•Among government consumption the government wage
component has larger effect than the non-wage one;
• The results are robust to several tests.
Major Comments
• The paper provides new interesting findings on the effect of
spending shocks on the real exchange rate;
• It is very enjoyable to read….
• …but at the same time
• more efforts could have been done to convince the reader about the
robustness of the results and the appropriateness of the methodology
• additional analysis could test for non-linearity
Robustness test
• One of the (classical) comment is the appropriateness of the
identification strategy. In particular, one could be skeptical about
the absence of simultaneous response of government spending to
output. To provide more convincing support two strategies could be
adopted:
• 1) Perform an a analysis similar to Beetsma et al. (2006) using quarterly
data.
• 2) Take out from government absorption those components that are more
likely to act as automatic stabilizers (Derby and Melitz, 2009; Furceri, 2010).
Robustness test
• Although it is to be good to parsimonious (specially in VAR!) as
robustness test the authors should include taxes in the set of
variables, to indeed be sure about the correct identification of
government spending shocks;
• Are we sure that the variables in the VAR are integrated of the
same order? One could argue that this must be the case for the
variables before the “difference” with respect to the EMU.
• To this purpose unit roots tests (preferably on the quarterly series) should
be carried out.
Robustness test
• The construction of data in relative terms is based on differencing
the variables with the respect to the EMU-j aggregate. Would this
difference for large countries affect also the nominal exchange
rate?
• Subdivide the analysis between small and large countries
• Construct the difference with respect to the EMU
• The authors claim that the real depreciation result found for nonEMU countries is due to different exchange rate regimes. To
convince the reader, a similar test could have been done focusing
on the EU before 1991.
Additional analysis
• It would be interested to see if the results differ between:
• Large vs. small countries
• Large trade weight vs. small trade weight
• High debt (refinancing ratio) vs. small debt countries
• EU sample before 1991
Minor points
• The authors assume that total expenditure is equal to government
consumption + investment. In contrast, transfers are a considerable
part of the total;
• As in the introduction, the discussion in the rest of the paper
should focus more on the real exchange rate and downplay the
discussion of the real output results.
• Are time fixed effects significant?
Thanks!
Fiscal Shocks and the Real Exchange Rate
by
Agustin S. Bénétrix & Philp R. Lane
Discussion by:
Davide Furceri
OECD, University of Palermo
Banco de Espana Conference
Interactions between monetary and fiscal policies
Madrid, 25-26 February 2010
Download