fiscal policy, relative prices and net exports in a ...€¦ · greece ireland italy netherlands...
TRANSCRIPT
Discussion
“Fiscal Policy, Relative Pricesand Net Exports in a Currency Union”
Luisa Lambertini Christian Proebsting
Benjamin Born
Frankfurt School of Finance & Management and CEPR
Fourth ECB biennial conference on fiscal policy and EMU governance
Frankfurt, December 19, 2019
Fiscal policy and intra-Eurozone competitiveness. . . some people have strong opinions
What Europe needs is austerity in the south and inflationary growth in the north toimprove the competitiveness of the south and to structurally improve the currentaccount imbalances.
Hans-Werner Sinn (2014)
1/12
Fiscal policy and intra-Eurozone competitiveness. . . naive correlation doesn’t show much
5 10 15 20 25 30 35
Government consumption (%-change)
-4
-2
0
2
4
6
Real exchange r
ate
(%
-change)
2001Q1-2015Q4
Belgium
Finland
France
Germany
Ireland
Italy
NetherlandsPortugal
Austria
Greece
Spain
=-0.091
2/12
Overview
Research question:
What are the effects of fiscal policy on net exports, relative prices and expenditureswitching?
First things first
• Very nice paper on a highly policy-relevant topic
• Authors dig deep into the data to get to the bottom of things
Plan for this discussion
• Brief summary
• Few comments, mostly on the empirics
3/12
Overview
Research question:
What are the effects of fiscal policy on net exports, relative prices and expenditureswitching?
First things first
• Very nice paper on a highly policy-relevant topic
• Authors dig deep into the data to get to the bottom of things
Plan for this discussion
• Brief summary
• Few comments, mostly on the empirics
3/12
Overview
Research question:
What are the effects of fiscal policy on net exports, relative prices and expenditureswitching?
First things first
• Very nice paper on a highly policy-relevant topic
• Authors dig deep into the data to get to the bottom of things
Plan for this discussion
• Brief summary
• Few comments, mostly on the empirics
3/12
This paper: Empirics
Approach:
• Identify exogenous shocks to government spending and consumption taxes
• Estimate cumulative fiscal multipliers via local projections
• Augment this by industry-level regressions
Findings:
• Both spending cuts and tax hikes raise net exports
→ mostly driven by fall in imports
• Spending cuts are deflationary
→ driven by non-traded goods prices
• Wages and prices in traded-goods industries react much less to spending cuts
4/12
This paper: Empirics
Approach:
• Identify exogenous shocks to government spending and consumption taxes
• Estimate cumulative fiscal multipliers via local projections
• Augment this by industry-level regressions
Findings:
• Both spending cuts and tax hikes raise net exports
→ mostly driven by fall in imports
• Spending cuts are deflationary
→ driven by non-traded goods prices
• Wages and prices in traded-goods industries react much less to spending cuts
4/12
This paper: ModelApproach:
• SMOPEC model of monetary union a la Galı/Monacelli (2005) with a number ofextensions to capture the empirically-observed relative price movements
→ restricted factor mobility between sectors
→ strong home-bias in government purchases
→ distribution services for traded retail good
Findings:
• Overall, model with extensions (!) can account for empirical evidence quite well
• However, mobility friction and home-bias lead to large output costs of currentaccount correction through fiscal policy
→ might be worthwhile for policy to tackle these issues
5/12
This paper: ModelApproach:
• SMOPEC model of monetary union a la Galı/Monacelli (2005) with a number ofextensions to capture the empirically-observed relative price movements
→ restricted factor mobility between sectors
→ strong home-bias in government purchases
→ distribution services for traded retail good
Findings:
• Overall, model with extensions (!) can account for empirical evidence quite well
• However, mobility friction and home-bias lead to large output costs of currentaccount correction through fiscal policy
→ might be worthwhile for policy to tackle these issues
5/12
# 1: Austerity vs. stimulus• Authors sometimes consider cuts in G and sometimes hikes
→ inconsequential as they have a linear framework
• What if the world is non-linear?
→ e.g. if nominal wages are downwardly rigid (Schmitt-Grohe/Uribe, 2016)
5 10 15 20 25 30 35
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
15
Rea
l excha
ng
e r
ate
(%
-ch
an
ge
)
2001Q1-2007Q4
Austria
Belgium
Finland
FranceGermany
Greece
Ireland
Italy NetherlandsPortugal
Spain
=-0.796
-20 -15 -10 -5 0 5 10
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
152010Q1-2015Q4
Austria
BelgiumFinland
France
Germany
Greece
Ireland
Italy Netherlands
Portugal SlovakiaSloveniaSpain
=-0.157
6/12
# 1: Austerity vs. stimulus• Authors sometimes consider cuts in G and sometimes hikes
→ inconsequential as they have a linear framework
• What if the world is non-linear?
→ e.g. if nominal wages are downwardly rigid (Schmitt-Grohe/Uribe, 2016)
5 10 15 20 25 30 35
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
15
Rea
l excha
ng
e r
ate
(%
-ch
an
ge
)
2001Q1-2007Q4
Austria
Belgium
Finland
FranceGermany
Greece
Ireland
Italy NetherlandsPortugal
Spain
=-0.796
-20 -15 -10 -5 0 5 10
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
152010Q1-2015Q4
Austria
BelgiumFinland
France
Germany
Greece
Ireland
Italy Netherlands
Portugal SlovakiaSloveniaSpain
=-0.157
6/12
# 1: Austerity vs. stimulus• Authors sometimes consider cuts in G and sometimes hikes
→ inconsequential as they have a linear framework
• What if the world is non-linear?
→ e.g. if nominal wages are downwardly rigid (Schmitt-Grohe/Uribe, 2016)
5 10 15 20 25 30 35
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
15
Real exchange r
ate
(%
-change)
2001Q1-2007Q4
Austria
Belgium
Finland
FranceGermany
Greece
Ireland
Italy NetherlandsPortugal
Spain
=-0.796
-20 -15 -10 -5 0 5 10
Government consumption (%-change)
-30
-25
-20
-15
-10
-5
0
5
10
152010Q1-2015Q4
Austria
BelgiumFinland
France
Germany
Greece
Ireland
Italy Netherlands
Portugal SlovakiaSloveniaSpain
=-0.157
6/12
# 1: Austerity vs. stimulusEuro area: adjustment to fiscal shocks asymmetric (Born et al., 2019)
Gov spending Output Exchange rateN
egat
ive
sho
ckP
osit
ive
sho
ck
quarter quarter quarter
7/12
# 1: Austerity vs. stimulus
• Distinguishing between hikes and cuts might also help you understand theheterogeneous responses of traded and non-traded industries
• If the heterogeneity is coming from cuts, it might be differences in downwardnominal wage rigidity
• Technically, given that you already have a two-stage approach, it might be as easyas including positive and negative shocks separately
→ cumulative multipliers might be an issue
8/12
# 2: Consumption tax variation
1Development of the overall tax revenue in the European Union
21DG Taxation and Customs Union | Taxation Trends in Europe 2019
Graph!9: Decomposition of the implicit tax rate on consumption 2017 (%)
0%
5%
10%
15%
20%
25%
30%
35%
EU-28EA-19
VAT component Tobacco and alcohol component Energy component Residual
DKHU SE LUFI EE NL SI CZ IE BG AT LV BEMTFR DE ELPL SK ITPT ROLT UKCY ES
Source: DG Taxation and Customs Union, based on Eurostat data. NB:!EU-27 data represent all EU Member States except Croatia.
Graph!10: Development of average standard VAT rate, EU-28, 2000-2019 (%)
17%
18%
19%
20%
21%
22%
23%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Source: DG Taxation and Customs Union.
Figure 2: Consumption Tax Changes and Announcements
Note: Figure depicts consumption tax changes in our sample. Dark bars correspond to tax changes that were announced less than 6 months prior toimplementation; grey bars correspond to tax changes that were announced at least 6 months before implementation; white bars are tax changes wherewe miss information on the announcement date. Bars show the sum of all tax changes across countries that fall in the same category out of six categories(announced, unannounced, missing information ⇥ positive change, negative change). Only tax changes amounting to 0.5 percentage points or more aredisplayed.
48
• Consumption tax changes dominated by “permanent” hikes after FinancialCrisis/Euro Area Sovereign Debt Crisis
→ this might be a very special time
→ and we know that fiscal multipliers depend on many things
9/12
# 2: Consumption tax variation
1Development of the overall tax revenue in the European Union
21DG Taxation and Customs Union | Taxation Trends in Europe 2019
Graph!9: Decomposition of the implicit tax rate on consumption 2017 (%)
0%
5%
10%
15%
20%
25%
30%
35%
EU-28EA-19
VAT component Tobacco and alcohol component Energy component Residual
DKHU SE LUFI EE NL SI CZ IE BG AT LV BEMTFR DE ELPL SK ITPT ROLT UKCY ES
Source: DG Taxation and Customs Union, based on Eurostat data. NB:!EU-27 data represent all EU Member States except Croatia.
Graph!10: Development of average standard VAT rate, EU-28, 2000-2019 (%)
17%
18%
19%
20%
21%
22%
23%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Source: DG Taxation and Customs Union. Figure 2: Consumption Tax Changes and Announcements
Note: Figure depicts consumption tax changes in our sample. Dark bars correspond to tax changes that were announced less than 6 months prior toimplementation; grey bars correspond to tax changes that were announced at least 6 months before implementation; white bars are tax changes wherewe miss information on the announcement date. Bars show the sum of all tax changes across countries that fall in the same category out of six categories(announced, unannounced, missing information ⇥ positive change, negative change). Only tax changes amounting to 0.5 percentage points or more aredisplayed.
48
• Consumption tax changes dominated by “permanent” hikes after FinancialCrisis/Euro Area Sovereign Debt Crisis
→ this might be a very special time
→ and we know that fiscal multipliers depend on many things
9/12
# 2: Consumption tax variation
1Development of the overall tax revenue in the European Union
21DG Taxation and Customs Union | Taxation Trends in Europe 2019
Graph!9: Decomposition of the implicit tax rate on consumption 2017 (%)
0%
5%
10%
15%
20%
25%
30%
35%
EU-28EA-19
VAT component Tobacco and alcohol component Energy component Residual
DKHU SE LUFI EE NL SI CZ IE BG AT LV BEMTFR DE ELPL SK ITPT ROLT UKCY ES
Source: DG Taxation and Customs Union, based on Eurostat data. NB:!EU-27 data represent all EU Member States except Croatia.
Graph!10: Development of average standard VAT rate, EU-28, 2000-2019 (%)
17%
18%
19%
20%
21%
22%
23%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Source: DG Taxation and Customs Union. Figure 2: Consumption Tax Changes and Announcements
Note: Figure depicts consumption tax changes in our sample. Dark bars correspond to tax changes that were announced less than 6 months prior toimplementation; grey bars correspond to tax changes that were announced at least 6 months before implementation; white bars are tax changes wherewe miss information on the announcement date. Bars show the sum of all tax changes across countries that fall in the same category out of six categories(announced, unannounced, missing information ⇥ positive change, negative change). Only tax changes amounting to 0.5 percentage points or more aredisplayed.
48
• Consumption tax changes dominated by “permanent” hikes after FinancialCrisis/Euro Area Sovereign Debt Crisis
→ this might be a very special time
→ and we know that fiscal multipliers depend on many things
9/12
# 2: Consumption tax variation
Figure 6: Empirical Government Spending Multipliers
Note: Figures depict the estimated government spending multipliers cMgh from regression (3.2), as a function of
the horizon h. 90 percent and 95 percent confidence intervals are displayed, based on Driscoll-Kraay standarderrors clustered at the country and time level.
Figure 7: Empirical Consumption Tax Multipliers
Note: Figures depict the estimated consumption tax multipliers cM⌧h from regression (3.2), as a function of
the horizon h. See Figure 6 for confidence intervals.
51
• This might also explain the large and persistent fall in GDP after a tax hike
• But would we expect the mirror image for a tax cut in Germany?
10/12
# 3: Forecasts in first stage
Table 1: First-Stage Regression
� ln Gi,t �⌧i,t
(1) (2) (3) (4)
Ft�1� ln Gi,t 0.56 0.85(0.06) (0.04)
� ln Gi,t�1 0.14 0.25(0.04) (0.04)
� ln Gi,t�2 0.03 0.15(0.04) (0.04)
�⌧i,t�1 0.17(0.05)
�⌧i,t�2 �0.09(0.05)
� ln Yi,t�1 0.02 0.03 0.00(0.04) (0.05) (0.01)
� ln Yi,t�2 0.14 0.12 0.01(0.04) (0.05) (0.01)
�ui,t�1 �0.34 �0.56 0.04(0.13) (0.14) (0.02)
�ui,t�2 0.29 0.44 �0.04(0.12) (0.13) (0.02)
R2 0.51 0.44 0.40 0.09Obs 523 549 535 513
Notes: Table displays the regression coe�cient of regression (2.1)
and (2.2). Notice that for tax changes, we only keep tax changes
announced less than six months prior to their implementation in
the sample.
44
• How large is the R2 just including fixed effects?
• How do IRFs look if estimated with shocksfrom specification (3)?
• If similar, could also run a quarterlyspecification without forecasts
• Oxford Economics has quarterly governmentconsumption forecasts starting in the 1990s
→ see Born et al. (2020) for details
11/12
# 3: Forecasts in first stage
Table 1: First-Stage Regression
� ln Gi,t �⌧i,t
(1) (2) (3) (4)
Ft�1� ln Gi,t 0.56 0.85(0.06) (0.04)
� ln Gi,t�1 0.14 0.25(0.04) (0.04)
� ln Gi,t�2 0.03 0.15(0.04) (0.04)
�⌧i,t�1 0.17(0.05)
�⌧i,t�2 �0.09(0.05)
� ln Yi,t�1 0.02 0.03 0.00(0.04) (0.05) (0.01)
� ln Yi,t�2 0.14 0.12 0.01(0.04) (0.05) (0.01)
�ui,t�1 �0.34 �0.56 0.04(0.13) (0.14) (0.02)
�ui,t�2 0.29 0.44 �0.04(0.12) (0.13) (0.02)
R2 0.51 0.44 0.40 0.09Obs 523 549 535 513
Notes: Table displays the regression coe�cient of regression (2.1)
and (2.2). Notice that for tax changes, we only keep tax changes
announced less than six months prior to their implementation in
the sample.
44
• How large is the R2 just including fixed effects?
• How do IRFs look if estimated with shocksfrom specification (3)?
• If similar, could also run a quarterlyspecification without forecasts
• Oxford Economics has quarterly governmentconsumption forecasts starting in the 1990s
→ see Born et al. (2020) for details
11/12
# 3: Forecasts in first stage
Table 1: First-Stage Regression
� ln Gi,t �⌧i,t
(1) (2) (3) (4)
Ft�1� ln Gi,t 0.56 0.85(0.06) (0.04)
� ln Gi,t�1 0.14 0.25(0.04) (0.04)
� ln Gi,t�2 0.03 0.15(0.04) (0.04)
�⌧i,t�1 0.17(0.05)
�⌧i,t�2 �0.09(0.05)
� ln Yi,t�1 0.02 0.03 0.00(0.04) (0.05) (0.01)
� ln Yi,t�2 0.14 0.12 0.01(0.04) (0.05) (0.01)
�ui,t�1 �0.34 �0.56 0.04(0.13) (0.14) (0.02)
�ui,t�2 0.29 0.44 �0.04(0.12) (0.13) (0.02)
R2 0.51 0.44 0.40 0.09Obs 523 549 535 513
Notes: Table displays the regression coe�cient of regression (2.1)
and (2.2). Notice that for tax changes, we only keep tax changes
announced less than six months prior to their implementation in
the sample.
44
• How large is the R2 just including fixed effects?
• How do IRFs look if estimated with shocksfrom specification (3)?
• If similar, could also run a quarterlyspecification without forecasts
• Oxford Economics has quarterly governmentconsumption forecasts starting in the 1990s
→ see Born et al. (2020) for details
11/12
# 3: Forecasts in first stage
Table 1: First-Stage Regression
� ln Gi,t �⌧i,t
(1) (2) (3) (4)
Ft�1� ln Gi,t 0.56 0.85(0.06) (0.04)
� ln Gi,t�1 0.14 0.25(0.04) (0.04)
� ln Gi,t�2 0.03 0.15(0.04) (0.04)
�⌧i,t�1 0.17(0.05)
�⌧i,t�2 �0.09(0.05)
� ln Yi,t�1 0.02 0.03 0.00(0.04) (0.05) (0.01)
� ln Yi,t�2 0.14 0.12 0.01(0.04) (0.05) (0.01)
�ui,t�1 �0.34 �0.56 0.04(0.13) (0.14) (0.02)
�ui,t�2 0.29 0.44 �0.04(0.12) (0.13) (0.02)
R2 0.51 0.44 0.40 0.09Obs 523 549 535 513
Notes: Table displays the regression coe�cient of regression (2.1)
and (2.2). Notice that for tax changes, we only keep tax changes
announced less than six months prior to their implementation in
the sample.
44
• How large is the R2 just including fixed effects?
• How do IRFs look if estimated with shocksfrom specification (3)?
• If similar, could also run a quarterlyspecification without forecasts
• Oxford Economics has quarterly governmentconsumption forecasts starting in the 1990s
→ see Born et al. (2020) for details
11/12
To sum up
• Very interesting paper
→ looking forward to reading future versions
• Three recommendations:
I Look at potential asymmetries between cuts and hikes
I Focus on government spending
→ you do that to a certain degree already
I Maybe also consider a quarterly sample
12/12
References I
Born, Benjamin, Francesco D’Ascanio, Gernot J. Muller, and Johannes Pfeifer (2019).“Government spending, downward nominal wage rigidity, and exchange ratedynamics”. Mimeo.
Born, Benjamin, Gernot J. Muller, and Johannes Pfeifer (2020). “Does austerity payoff?” Review of Economics and Statistics.
Galı, Jordi and Tommaso Monacelli (2005). “Monetary Policy and Exchange RateVolatility in a Small Open Economy”. Review of Economic Studies 72 (3), 707–734.
Schmitt-Grohe, Stephanie and Martın Uribe (2016). “Downward nominal wage rigidity,currency pegs, and involuntary unemployment”. Journal of Political Economy 124(5), 1466–1514.
Sinn, Hans-Werner (2014). “Austerity, growth and inflation: remarks on the eurozone’sunresolved competitiveness problem”. The World Economy 37 (1), 1–31.
References A 13/1