Asymmetric Trade Liberalizations and Current Account Dynamics (Alessandro Barattieri, Collegio Carlo Alberto and ESG UQAM)

Asymmetric Trade Liberalizations
and Current Account Dynamics
Alessandro Barattieri
Collegio Carlo Alberto and ESG UQAM
European University Institute
Fiesole, 20 April 2017
Alessandro Barattieri (ESG UQAM & CCA ) Asymmetric Trade Liberalizations 20 April 2017 1 / 30

Motivation: Current Account Dynamics in Europe
-.15-.1-.050.05
CA_GDP
2000 2002 2004 2006 2008
TIME
DEU ESP
PRT GRC

Motivation: Inside European Current Accounts (I)
-.2-.10.1
2000 2002 2004 2006 2008
TIME
DEU Goods DEU Serv
GRC Goods GRC Serv

Motivation: Inside European Current Accounts (II)
-6000-4000-200002000
1995 2000 2005 2010
year
MANUF SERV
Portugal bilateral net trade ﬂows with Germany

Contribution
I solve analytically a simple model where the current account is
determined by present and future relative changes in the trade
costs.

Contribution
costs.
I build relative trade liberalization measures and test the key
prediction of the simple model using two diﬀerent datasets.

Contribution
costs.
I build relative trade liberalization measures and test the key
prediction of the simple model using two diﬀerent datasets.
I propose a two country DGE model featuring trade costs (and
productivity) in order to explore the relevance of the mechanism in
explaining the dynamics of the German trade surplus.

Main Findings
1 In both samples, contemporaneous high relative trade
liberalizations are correlated with current account deﬁcits, while
future high relative trade liberalization with surpluses. Moreover,
I cannot reject the speciﬁc parameter restriction suggested by the
model.

Main Findings
model.
2 German data reveals a large asymmetries between the
manufacturing and the service sector in terms of trade costs and
productivity dynamics over the period 2000-2007. When fed with
the actual asymmetries found in the data, the model produces a
peak surplus of about 5% of GDP.

Main Findings
model.
2 German data reveals a large asymmetries between the
manufacturing and the service sector in terms of trade costs and
productivity dynamics over the period 2000-2007. When fed with
the actual asymmetries found in the data, the model produces a
peak surplus of about 5% of GDP.
3 When only fed with asymmetric trade costs dynamics, the model
still delivers sizeable trade surpluses. However, when fed only
with asymmetric productivity dynamics, the model fails to
replicate the dynamics of the German surplus.

Outline of the Presentation
A short piece of Theater...in 3 Acts
Act I: A simple model
Act II: Empirical evidence
Act III: A Quantitative Investigation of the German Surplus

Related Literature
1 Industrial specialization and current account (Jin, 2012; Ju and Wei,
2012; Barattieri, 2014)
2 Determinants of current account dynamics in Europe (Lane, 2013;
Kollmann et al, 2015)
3 Trade costs dynamics and the current account (Reyes-Heroles, 2016;
Alessandria and Choi, 2016; Alessandria, Choi and Lu, 2016)

A Simple Model (I)
Outline
2 Periods. 2 Countries. 2 Goods (endowment). Complete
specialization.

A Simple Model (I)
Outline
specialization.
Consumer maximizes a standard 2 periods CRRA Utility function
(discount β). A riskless bond B.

A Simple Model (I)
Outline
specialization.
Consumption basket: standard CES
Demand functions: standard CES

A Simple Model (I)
Outline
specialization.
Pt = ph
t
1−θ
+ τf
t pf ∗
t
1−θ
1
1−θ
; P∗
t = τh
t ph
t
1−θ
+ pf ∗
t
1−θ
1
1−θ

A Simple Model (I)
Outline
specialization.
Pt = ph
t
1−θ
+ τf
t pf ∗
t
1−θ
1
1−θ
; P∗
t = τh
t ph
t
1−θ
+ pf ∗
t
1−θ
1
1−θ
Goods mkts clearing: Y h
t = Ch
t + τh
t Ch∗
t ; Y f ∗
t = τf
t Cf
t + Cf ∗
t

A Simple Model (I)
Outline
specialization.
Pt = ph
t
1−θ
+ τf
t pf ∗
t
1−θ
1
1−θ
; P∗
t = τh
t ph
t
1−θ
+ pf ∗
t
1−θ
1
1−θ
Goods mkts clearing: Y h
t = Ch
t + τh
t Ch∗
t ; Y f ∗
t = τf
t Cf
t + Cf ∗
t
Bond mkt clearing: B1 + B∗
1 = 0

A Simple Model (II)
Log-Linearized Model and Solution
I log-linearize the model around a symmetric equilibrium.

A Simple Model (II)
I solve analytically for the current account of the Home Country
( ˆB1) in terms of the trade costs and the endowments.
ˆB1 = −η ˆτh
1 − ˆτf
1 + η ˆτh
2 − ˆτf
2 + ν ˆY h
1 − ˆY f ∗
1 − ν ˆY h
2 − ˆY f ∗
2
(1)
η and ν are functions of the structural parameters of the model.

A Simple Model (II)
I solve analytically for the current account of the Home Country
( ˆB1) in terms of the trade costs and the endowments.
ˆB1 = −η ˆτh
1 − ˆτf
1 + η ˆτh
2 − ˆτf
2 + ν ˆY h
1 − ˆY f ∗
1 − ν ˆY h
2 − ˆY f ∗
2
(1)
η and ν are functions of the structural parameters of the model.
Consumption smoothing drives current account dynamics
An asymmetric trade liberalization process for which
ˆτh
1 − ˆτf
1 < 0 or ˆτh
2 − ˆτf
2 > 0 pushes Home’s CA toward
surplus.

A Simple Model (III)
The Parameter Values
Deﬁne: sh = 1
1+τ1−θ , sf = τ1−θ
1+τ1−θ , α0 = sh−sf
β > 0, α1 = sf sh(θ − 1) > 0,
α2 = 2α1 + sf > 0
Then:
η =
α0
α2
+ (σ − 1) (sh − sf )α1
α2
+ sf
(1 + β) 2
β + α0β
α2
+ (1 + β) (σ − 1)(sh − sf )α0
α2
> 0 (2)
ν =
1 − sf
α2
− (σ − 1)(sh − sf ) sf
α2
(1 + β) 2
β + α0β
α2
+ (1 + β) (σ − 1)(sh − sf )α0
α2
> 0 (3)
Where: η is positive as long as σ > 1 −
α1
α2
(sh−sf )
α1
α2
+sf
. ν is positive if θ > 1
and σ < 1 + 2sh(θ−1)
(sh−sf ) .

Empirical Analysis - OECD Countries + BRICS (I)
Stage 1: the Constructed Home Bias Index
Structural gravity for trade in services and manufacturing:
Xk
ij =
Y k
i Ek
j
Y k
tk
ij
Pj Πi
1−θk
(4)

Xk
ij =
Y k
i Ek
j
Y k
tk
ij
Pj Πi
1−θk
(4)
The constructed home bias index (CHB) is the ratio of realized
internal trade relative to the frictionless case
CHBik =
Y k
i Ek
i
Y k
tk
ii
Pi Πi
1−θk
Y k
i Ek
i
Y k
=
tk
ii
Pi Πi
1−θk
. (5)

Xk
ij =
Y k
i Ek
j
Y k
tk
ij
Pj Πi
1−θk
(4)
The constructed home bias index (CHB) is the ratio of realized
internal trade relative to the frictionless case
CHBik =
Y k
i Ek
i
Y k
tk
ii
Pi Πi
1−θk
Y k
i Ek
i
Y k
=
tk
ii
Pi Πi
1−θk
. (5)
Using (4) and (5), I compute the CHB as:
CHBik =
tk
ii
Pi Πi
1−θk
=
Xk
ii Y k
Y k
i Ek
i
. (6)

Empirical Analysis - OECD Countries + BRICS (II)
Stage 2: Relative Trade Liberalization Measure
I base the analysis on the CHB index introduced before. Using data
for 24 OECD countries + BRICS.
I divide the sample into “goods-oriented” and “service-oriented”
countries. Then I deﬁne a measure of relative trade liberalization
as:
(ˆτh
t − ˆτf
t ) = ∆
i
ωi CHBh
i
t
− ∆CHBf
it (7)
For Germany h is manufacturing, f services. For Spain h is services
and f manufacturing.
A high relative trade liberalization: country i CHB in the
importing sector decreased by more (or increased by less) than the
CHB of the trading countries in its the export sector.

Empirical Analysis - OECD Countries + BRICS (III)
Relative Trade Liberalization Measures, Selected Countries-.02-.010.01.02.03
meanofRELATIVE_CHB
ESP GRC PRT DEU

Empirical Analysis - OECD Countries + BRICS (IV)
Stage 3: Current Account Dynamics
Following (1), I evaluate the empirical merit of the model using the
following speciﬁcation:
∆
CA
GDP it
= η0 +η1(ˆτh
t −ˆτf
t )+
S
s=1
ηs+1(ˆτh
t+s −ˆτf
t+s)+ψZit +δi +δt + it
(8)

∆
CA
GDP it
= η0 +η1(ˆτh
t −ˆτf
t )+
S
s=1
ηs+1(ˆτh
t+s −ˆτf
(8)
Zit: openness, growth, GDP and per capita GDP, and a proxy for
ﬁnancial development.
Model predictions: i) η1 < 0

∆
CA
GDP it
= η0 +η1(ˆτh
t −ˆτf
t )+
S
s=1
ηs+1(ˆτh
t+s −ˆτf
(8)
Model predictions: i) η1 < 0 ii) ηs > 0

∆
CA
GDP it
= η0 +η1(ˆτh
t −ˆτf
t )+
S
s=1
ηs+1(ˆτh
t+s −ˆτf
(8)
Model predictions: i) η1 < 0 ii) ηs > 0 and iii) η1 + S
s=1 ηs+1 = 0

Empirical Analysis: OECD Countries + BRICS (V)
Regression Results
Dep. var: ∆ CA
GDP t
(1) (2) (3) (4) (5) (6)
(ˆτ h − ˆτ f )t -4.906*** -5.826*** -4.453*** -4.100*** -1.331
(0.918) (0.960) (1.031) (1.176) (4.910)
(ˆτ h − ˆτ f )t+1 2.477*** 3.666*** 3.897*** 3.857*** 10.361**
(0.953) (1.039) (1.027) (1.114) (5.221)
(ˆτ h − ˆτ f )t+2 1.928* 2.521*** 2.270** -5.526
(1.024) (0.956) (1.026) (4.611)
(ˆτ h − ˆτ f )t+3 0.609
(0.966)
OPENNESS 0.893** 3.935*** 1.412**
(0.434) (1.514) (0.579)
Real P.C. GDP -0.087 -4.998 -0.205
(0.123) (4.364) (0.153)
Real GDP 0.230** 5.073 0.318**
(0.106) (4.372) (0.140)
GROWTH -0.159*** -0.271*** -0.254**
(0.048) (0.070) (0.110)
CREDIT -0.004 -0.009 -0.004
(0.002) (0.006) (0.003)
DELTA INFL. 0.002 0.001 0.002
(0.009) (0.010) (0.011)
Time + Country FE No No No No Yes No
N 433 404 346 369 369 329
Adj R-squared 0.060 0.014 0.117 0.136 0.132 0.091
Log-Likelihood -917.859 -843.780 -706.129 -742.935 -722.106 -708.141
Sargan-Hansen stat 0.051
p-value 0.821
Endog (ˆτ h − ˆτ f )s 5.904
p-value 0.116
P-value of Test
η1 + η2 + η3 = 0 0.87 0.24 0.31 0.61

Empirical Analysis: Developing Countries (I)
Set-up
Sample of 13 developing countries (with tariff data from 1995-2012)
that are: 1) highly specialized in the export of agriculture 2)
specialized in the import of manufacturing.
I build relative trade liberalization measures ˆτh
t − ˆτf
t using tariff
data on Agricultural (ˆτh
t ) and Manufacturing (ˆτf
t ) goods
A high relative trade liberalization: country i tariff in the
importing sector decreased by more (or increased by less) than the
protection in its the export sector.

Empirical Analysis: Developing Countries (II)
Current Account vs. Relative Trade Liberalization Index (Averages)
ARG
BRA
CHL
COL
CRI
ECU
HND
MWI
NIC
PRY
PER
URY
ZMB
y=-2 -10x, R2=0.60
-15-10-50
AverageCurrentAccountoverGDP
-.5 0 .5 1
Average Relative Liberalization 1995-2009
AV_CA_GDP Fitted values

Empirical Analysis: Developing Countries (III)
Econometric results
Dep. var: ∆ CA
GDP t
(1) (2) (3) (4)
(ˆτ h − ˆτ f )t -0.372** -0.268* -0.272* -0.299*
(0.144) (0.138) (0.141) (0.156)
(ˆτ h − ˆτ f )t+1 0.070 0.056 0.073
(0.143) (0.147) (0.164)
(ˆτ h − ˆτ f )t+2 0.374** 0.357** 0.465***
(0.151) (0.156) (0.172)
(ˆτ h − ˆτ f )t+3 -0.193 -0.220 0.018
(0.162) (0.160) (0.181)
ln(Real GDP) 0.078 11.284
(0.238) (9.055)
ln(Real P.C. GDP) -0.327 -6.275
(0.461) (9.960)
Growth -0.161** -0.188*
(0.067) (0.096)
Credit -0.003 -0.041
(0.017) (0.038)
Country FE No No No Yes
Year FE No No No Yes
R-squared 0.032 0.071 0.105 0.259
N 201 193 193 193
P-value of Test
η1 + η2 + η3 + η4 = 0 0.96 0.83 0.55

A Quantitative Analysis of the German Surplus (I)
A BKK-type Model with Trade Costs: Set-up (1)
Two Countries (1 and 2). Two intermediate goods (a and b).
Complete specialization (1 → a, 2 → b)

Intermediate output produced with capital and labor:
yit = Aitkα
1tn1−α
it (9)

yit = Aitkα
1tn1−α
it (9)
Final output (consumption and investment) is a CES aggregate of a
and b:
h1t = (1 − ω)
1
θ (a1t)
θ−1
θ + ω
1
θ (b1t)
θ−1
θ
θ
θ−1
(10)
h2t = (1 − ω)
1
θ (b2t)
θ−1
θ + ω
1
θ (a2t)
θ−1
θ
θ
θ−1
(11)

yit = Aitkα
1tn1−α
it (9)
Final output (consumption and investment) is a CES aggregate of a
and b:
h1t = (1 − ω)
1
θ (a1t)
θ−1
θ + ω
1
θ (b1t)
θ−1
θ
θ
θ−1
(10)
h2t = (1 − ω)
1
θ (b2t)
θ−1
θ + ω
1
θ (a2t)
θ−1
θ
θ
θ−1
(11)
Goods Market Clearing requires:
y1t = a1t + τ1ta2t y2t = b2t + τ2tb1t (12)

A Quantitative Analysis of the German Surplus (II)
Inﬁnite horizon utility maximization problem. Fixed Labor supply.
Incomplete Markets:
max ∞
t=0 βt c1−σ
1t
1−σ
subject to
k1t+1 + B1t+1 + QtB∗
1t+1 + c1t +
η
2
B2
1t+1 + Qt
η
2
B∗2
1t+1 =
r1tB1t + Qtr2tB∗
1t + rk
1tk1t + w1tn1t + (1 − δ)k1t−1 + T1t

Incomplete Markets:
max ∞
t=0 βt c1−σ
1t
1−σ
subject to
k1t+1 + B1t+1 + QtB∗
1t+1 + c1t +
η
2
B2
1t+1 + Qt
η
2
B∗2
1t+1 =
r1tB1t + Qtr2tB∗
1t + rk
1tk1t + w1tn1t + (1 − δ)k1t−1 + T1t
The Current Account is deﬁned as the change in Net Foreign Asset
position:
CA1t = ∆NFA1t = (B1t+1 − B1t) + Qt B∗
1t+1 − B∗
1t (13)

Incomplete Markets:
max ∞
t=0 βt c1−σ
1t
1−σ
subject to
k1t+1 + B1t+1 + QtB∗
1t+1 + c1t +
η
2
B2
1t+1 + Qt
η
2
B∗2
1t+1 =
r1tB1t + Qtr2tB∗
1t + rk
1tk1t + w1tn1t + (1 − δ)k1t−1 + T1t
The Current Account is deﬁned as the change in Net Foreign Asset
position:
CA1t = ∆NFA1t = (B1t+1 − B1t) + Qt B∗
1t+1 − B∗
1t (13)
The model is solved as a nonlinear forward looking deterministic
system using a Newton method.

A Quantitative Analysis of the German Surplus (III)
Asymmetric Productivity Dynamics: Germany (2000=1)
11.051.11.151.2
2000 2002 2004 2006 2008
year
MANUFACTURING SERVICES
TOTAL
Productivity Dynamics (TFP) Germany (2000=1)

A Quantitative Analysis of the German Surplus (IV)
Asymmetric Trade Costs Dynamics: The CHB in Germany (2000=1)
.8.85.9.951
2000 2002 2004 2006 2008
Time
MANUFACTURING SERVICES
CHB Index (2000=1), Germany

A Quantitative Analysis of the German Surplus (V)
Selected Endogenous Variable Dynamics, Symmetric Trade Liberalization and
Productivity Increase
1990 2000 2010 2020 2030
0.25
0.3
0.35
0.4
0.45
Exports
1990 2000 2010 2020 2030
0.25
0.3
0.35
0.4
0.45
Imports
1990 2000 2010 2020 2030
0.1
0.15
0.2
0.25
0.3
Investment
1990 2000 2010 2020 2030
0.1
0.15
0.2
0.25
0.3
Savings
1990 2000 2010 2020 2030
0.6
0.7
0.8
0.9
1
Consumption
1990 2000 2010 2020 2030
-1
-0.5
0
0.5
1
CA/GDP
1990 2000 2010 2020 2030
1
1.05
1.1
1.15
1.2
A1
and A2
1990 2000 2010 2020 2030
2.2
2.4
2.6
2.8
τ1
and τ2

A Quantitative Analysis of the German Surplus (VI)
Selected Endogenous Variable Dynamics, Asymmetric Trade Liberalization and
Productivity Increase
1990 2000 2010 2020 2030
0.25
0.3
0.35
0.4
0.45
Exports
1990 2000 2010 2020 2030
0.25
0.3
0.35
0.4
0.45
Imports
1990 2000 2010 2020 2030
0.15
0.2
0.25
0.3
0.35
Investment
1990 2000 2010 2020 2030
0.15
0.2
0.25
0.3
Savings
1990 2000 2010 2020 2030
0.6
0.7
0.8
0.9
1
Consumption
1990 2000 2010 2020 2030
-0.05
0
0.05
CA/GDP
1990 2000 2010 2020 2030
1
1.05
1.1
1.15
1.2
A1
and A2
1990 2000 2010 2020 2030
2.2
2.4
2.6
2.8
τ!
and τ2

A Quantitative Analysis of the German Surplus (VII)
Current Account over GDP, German Data and Model
-.04-.020.02.04.06
2000 2005 2010 2015
year
German Data Asymmetric Trade and Prod
Symmetric Trade and Prod

A Quantitative Analysis of the German Surplus (VIII)
Current Account over GDP, German Data Post 2007
.02.04.06.08.1
2000 2005 2010 2015
TIME
OVERALL INTRA EU
Goods Balance over GDP, Germany (Eurostat)

A Quantitative Analysis of the German Surplus (IX)
Current Account over GDP, Diﬀerent Models
-.04-.020.02.04.06
2000 2005 2010 2015 2020
year
Asymmetric Trade NO Prod Asymmetric Trade and Prod
Asymmetric Prod NO Trade

Concluding Remarks
Service trade liberalization possible further adjustment margin
for European rebalancing. Relatively neglected to date, relevant?
However, the link between the insight and economic policy can
be subtle: example of German tourism to Greece.
Asymmetric globalization can have signiﬁcant eﬀect for current
account dynamics (potentially relevant for the US and Trump?).

For further suggestions, comments, complaints:
barattieri.alessandro@carloalberto.org
Thanks!

Asymmetric Trade Liberalizations and Current Account Dynamics (Alessandro Barattieri, Collegio Carlo Alberto and ESG UQAM)

Recommended

Recommended

More Related Content

What's hot

What's hot (14)

Similar to Asymmetric Trade Liberalizations and Current Account Dynamics (Alessandro Barattieri, Collegio Carlo Alberto and ESG UQAM)

Similar to Asymmetric Trade Liberalizations and Current Account Dynamics (Alessandro Barattieri, Collegio Carlo Alberto and ESG UQAM) (20)

More from ADEMU_Project

More from ADEMU_Project (20)

Recently uploaded

Recently uploaded (20)

Asymmetric Trade Liberalizations and Current Account Dynamics (Alessandro Barattieri, Collegio Carlo Alberto and ESG UQAM)