Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
A network analysis of preferential trade agreements:1815 - 1914
Luca De Benedictis1 Silvia Nenci2
1University of Macerata (Italy)2University of Roma Tre (Italy)
Valencia FRESH meeting 2013Frontier Research in Economic and Social History
Valencia, 13-14 June 2013
Luca De Benedictis, Silvia Nenci 1 / 34
Outline of the presentation
1 Aims
2 Why using Network Analysis?
3 Background: Trade policy in the period 1815 - 1914
4 Focus: The ”‘network”’ of Cobden-Chevalier Treaties
5 Network Analysis: basics
6 Data sources and Stylized facts
7 First remarks
Luca De Benedictis, Silvia Nenci 2 / 34
Aims
The aim of this introductory work is twofold:
First: a) to describe and analyse the stucture of internationalpreferential trading relations (i.e. PTAs) from 1815 to 1913 and itsevolution over time by applying NA; b) to investigate in historicalperspective if trade relations and PTAs are strongly heterogeneousamong countries, and if specific countries play different roles overtime, by calculating some network statistics.
Second: to apply NA to the analysis of a debated issue in theeconomic history literature to show how this approach allows to getnew insights. The issue is: Can the Cobden-Chevalier Treaty beconsidered as a real turning point in the trade policy of the 19thcentury?
Luca De Benedictis, Silvia Nenci 3 / 34
Motivation: the interdependence issue
Economic history literature as well as trade literature generallyunderestimate the systematic ways in which one countrys choicesaffect the choices of others.
Trade integration is not a characteristic of a single country, but is theresult of the relations that every single country has with the economicsystem.
Interdependence between countries is a key factor in their choice tojoin a PTA (Egger and Larch, 2008; the Baldwin’s domino effect,1995).
Our presumption: international context has strong and systematiceffects on trade policy and therefore should not be ignored (see alsoAlapuro 1988; Lindberg 1983; Y. M. Werner 1989; Phare, 2008).
Luca De Benedictis, Silvia Nenci 4 / 34
Why using Network Analysis?
Interdependence is the essence of networks
Network Analysis focuses more on the properties of the system (i.e.structural dimension) rather than on individual characteristics.
NA is well-fitted to analyze the characteristics of single countries intheir relative position in the whole system of relations.
The algoritm: countries which are connected tend to stay close, whilecountries which are not connected tend to be placed far apart. But,the position of each country does not depend only on its link withanother country but also on the indirect effect of others: the partnersof its partners will contribute to the determination of the country’sposition in the network.
Luca De Benedictis, Silvia Nenci 5 / 34
A short review of the literature
Our work is not the first one on the application of NA to trade issues:Smith and White (Social Forces,1992)Serrano and Bogua (Phys. Rev., 2003)Garlaschelli and Loffredo (Physica A, 2005)Fagiolo, Reyez and Schiavo (S.Anna WP, 2007)Kali and Reyes (J. Intern. Bus. Studies, 2007)De Benedictis and Tajoli (World Economy, 2011)De Benedictis, Nenci, Santoni, Tajoli, Vicarelli (CEPII Working Paper,fortcoming)
If the application of NA to the study of trade relations is quite recent,its application to the same relation but in historical perspective is ararity (Flandreau and Clemens, 2005 apply NA to identify the coreand the periphery of the international monetary system).
The present study is a first attempt to analyse trade integration inhistorical perspective using NA.
Luca De Benedictis, Silvia Nenci 6 / 34
Background: Trade policy during the 19th century and thebeginning of the 20th
Since the end of the Napoleonic Wars, there has been an increasingnumber of PTAs aimed at reducing tariff protection between two ormore countries on selected traded products (Jacks, 2005; Shiue, 2005;Kaukiainen, 2001; Ejrns and Persson, 2000)
The liberalization process of the nineteenth century started verygradually and spread unevenly among Europe and outside.
The second half of the 19th century, especially the first twenty-fiveyears, can be described as a period of increasing tradeliberalization(Foreman-Peck, 1999). During this period the diffusionof free trade doctrine occurred in Europe and (partially) in the UnitedStates, and interested their partners as well (Colonial nations).
Luca De Benedictis, Silvia Nenci 7 / 34
The the Cobden-Chevalier Treaty and the Network ofPTAs of the 1860s
The most important initiative in the diffusion of trade liberalizationpolicy in the second half of the century was the Cobden-ChevalierTreaty signed by France and the UK in 1860 (Bairoch, 1976, 1989)
Following this treaty, which included the ”most favoured nation”(MFN) clause, most European countries, through treaties signed withFrance or the UK, became part of a network of more than 50 bilateraltreaties, known as ”the network of Cobden-Chevalier Treaties”.
The main focus of the bilateral treaties was on reducing tariffs onmanufactured articles, such as textiles, ironware, articles of leatherand rubber, and especially on wines and stronger alcoholic beverages.
This network guaranteed the development of free trade among themain trading powers in Europe for around 20 years, lowering averagetariffs to a degree unmatched in international trade until the GATTera.
Luca De Benedictis, Silvia Nenci 8 / 34
The network of Cobden-Chevalier Treaties
Experts in nineteenth century trade policy (Bairoch, 1977, 1989; Nye,1991, 2007; Irwin, 1993; and ORourke and Williamson, 1999) arguethat the Cobden-Chevalier treaty was decisive in reducing tariffprotection by spreading bilateral agreements containing the MFNclause.
For these authors, the Cobden-Chevalier agreement has beenconventionally portrayed as a milestone in 19th century liberalisationin effect as the mother of all free trade.
A growing research activity has been devoted recently to the analysisof the Cobden-Chevalier network ’s impact on trade growth of theEuropean countries during the nineteenth century (Bairoch, 1976,1989, 1993; Pahre, 2008; Accominotti and Flandreau, 2005; Irwin,1993; Stein, 1984; Lazer, 1999; Lampe, 2009).
Luca De Benedictis, Silvia Nenci 9 / 34
Was the Cobden-Chevalier Treaty a real turning point inthe trade policy of the 19th century?
This work aims at analysing the importance of this Agreement not interms of trade growth but in terms of its impact on:
The structural dimension of trade preferential relations in the secondhalf of the 19th century;
The role and centrality of countries in the trade integration process ofthe 19th century and their evolution over time.
To do this we apply the NA’s standard tools: graphs and indicators
Luca De Benedictis, Silvia Nenci 10 / 34
Network Analysis: Graph
G = (V ,L)
unit: vertex , node ∈ Vlink: line ∈ L
directed (arc) ∈ A;undirected (edge) ∈ E ;L = A∪ EL is a binary set [0,1]
In this work, unit is the countryand link is the PTA
In this work:
undirected network;
unweighted network.
Luca De Benedictis, Silvia Nenci 11 / 34
Structural properties of network: Degree Centrality
Degree centrality is the simplest measure of the position of a node ina network. If the network is unweighted, it measures the centrality ofa node by the number of connections the node has.
This indicator ranges from 0 to 1; the more it’s degree is close to 1,the more a country is connected in the network.
The degree centrality is essentially a local centrality measure. It takesinto consideration only the direct links of a node, its nearestneighborhood, respectless to the position of the node in the network’sstructure.
Luca De Benedictis, Silvia Nenci 12 / 34
Structural properties of network: Eigenvector Centrality
Eigenvector centrality index measures a country’s centrality lookingat the importance of its neighbors (what really matters is thecentrality of the countries linked to him)
In general, countries with a high value of eigenvector centrality arethe ones which are connected to many other countries which are, inturn, connected to many others.
The Eigenvector centrality is essentially a global centrality measure.
Luca De Benedictis, Silvia Nenci 13 / 34
Data: the Phare’s database
Comprehensive data on preferential agreements all starts around 1950(data prior to this date are scattered and fragmented).
We use historical data on PTAs arising from the Pahre s (2008)database that includes intra-European, intra-American, andinterregional agreements signed from 1815 to 1914.
The database contains a wide range of agreements; the main typesare conventions and treaties.
Units included in the database are 92 territories (48 current states).
Some territorial divisions (such as the German Zollverein, Modenaand Parma before the Italys unification) or special forms ofintegration (such as the German Tax Union or the North GermanConfederation) are treated as separate States in the Phare’s database(since they had the power to negotiate commercial treaties on behalfof their members).
Luca De Benedictis, Silvia Nenci 14 / 34
Annual Treaties in Effect: European Great Powers, 18151913
The Figure suggests a steadily expanding network of cooperation. Thisvisual continuity reflects the expansion of the network to include mostcountries by century s end.
Figure : Annual Treaties, European Great Powers, 1815 - 1913
Source:Phare (2008)Luca De Benedictis, Silvia Nenci 15 / 34
Stylized facts: the evolution of PTAs’ network
Figure : PTAs, 1815
ARG
COL
HZH
HZS
PRS
PRT
SWD
UK
US
WRT
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 16 / 34
Stylized facts: PTAs in 1850
Figure : PTAs, 1850
BAD CR
HZS
ICE
PRT PRU
SWZ
URG
VEN
WRT
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 17 / 34
Stylized facts: PTAs in 1855
Figure : PTAs, 1855
AH ARG
BAD
BLGBOL
BRMBRW
BRZ
CAN
CHLCOL
CR
DENDR
ECU
ELS
FRK
FRN
GRC
GUA
HAMHAN
HSC
ITL
LIE
LIP
LUB
LUX
MCS
MEC
MEX
MOD
NAS
NICNTH OLD
PAR
PRG
PRS
PRT
PRU
RUS
SCGSPN
SRD
SWD
SWZ
SXH
SXWTCU
TUR
UK
URG
US
VEN
WLD
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 18 / 34
Stylized facts: PTAs in 1861
Figure : PTAs, 1861
ADS
AH
ANB
ANH
ANKARG
BAD
BAV
BLG
BOL
BRM
BRN
BRW
BRZ
CAN
CHL
COLCR
DEN
DR
ECU
ELS
FRK
FRN
GRC
GUA
HAM
HAN
HOH
HOM
HON
HSC
HSE
HSG
ITL
LIE
LIP
LUB
LUX
MCS
MEC
MEX
NAS
NIC
NTH
OLD
PRG
PRS
PRT
PRU
RUS
SBRSBS
SCG
SPN
SWD
SWZ
SXC
SXHSXWTCU
TUR
UK
URG
US
VENWLD
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 19 / 34
Stylized facts: PTAs in 1865
Figure : PTAs, 1865
ADS
AH
ANB
ANH
ANK
ARG
BAD
BAV
BLG
BOL
BRM
BRN
BRW
BRZ
CAN
CHL
COL
CR
DENDR
ECU
ELS
FRK
FRN
GRC
GUA
HAI
HAM
HAN
HOH
HOM
HON
HSC
HSE
HSG
ITL
LIE LIP
LUB
LUX
MCS
MEC
MEX
NAS
NIC
NOR
NTH OLDPRG
PRS
PRT
PRU
RUS
SBR
SBS
SCG
SPN
SWDSWZ
SXC
SXH
SXW
TCU
TUR
UKURGUS
VEN
WLD
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 20 / 34
Stylized facts: PTAs in 1880
Figure : PTAs, 1880
AH
ARGBLGBOL
BRZCHL
COLCR
DEN
DRECU
ELS
FRN
GRC
GUA
HAI
HAW
HON
ITL
JPN
LIE
LUX
MEC
MEX
NIC
NOR
NTH
OFS
OLD
PRG
PRS
PRT
PRU
ROM
RUS
SPN
SRB
SWD
SWZ
TRV
TURUK
URG
US
VEN
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 21 / 34
Stylized facts: PTAs in 1913
Figure : PTAs, 1913
AH
ARG
BLG
BOL
BRZ
BUL
CAN
CHL
CHN
COL
CRCUB
DEN
DR
ECU
ELS
FRN
GRC
HAI
HAW
HON
ICE
IRN
ITL
JPN
LIE
LUX
MEC
MEX
MNT
NIC
NOR
NTH
OFS
OLD
PRG
PRS
PRT
PRU
ROMRUS
SPNSRB
SWD
SWZ
TUR
UK
URG
US
VEN
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 22 / 34
Stylized facts: PTAs Network density 1815-1913
Figure : Density
Fundamental years
Den
sity
0.00
0.01
0.02
0.03
0.04
1815 1850 1855 1861 1865 1880 1901 1913
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 23 / 34
Stylized facts: PTAs and degree centrality of countries in1861
Figure : degree centrality, 1861
ADS
AH
ANB ANH
ANK
ARG
BAD
BAV
BLG
BOL
BRM
BRN BRW
BRZ
CAN
CHL
COLCR DENDR
ECU
ELS
FRK
FRN
GRC
GUA
HAM
HAN
HOHHOM
HON
HSC
HSE
HSG
ITL
LIE
LIP
LUB
LUX
MCS
MEC
MEX
NAS
NIC
NTHOLD PRG
PRS
PRT
PRU RUS
SBR
SBS
SCG
SPN
SWDSWZ
SXC
SXH
SXWTCU
TUR
UKURG
US
VEN
WLD
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 24 / 34
Stylized facts: Degree centrality for UK and France
Figure : Degree centrality
Fundamental years
UK
and
Fra
nce
degr
ee c
entr
ality
010
2030
4050
60
1815 1850 1855 1861 1865 1880 1901 1913
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 25 / 34
Stylized facts: Eigenvector centrality vs degree centralityfor UK and France
Figure : Eigenvector centrality vs degree centrality
Fundamental years
Deg
ree
and
eige
nvec
tor
cent
ralit
y F
ranc
e an
d U
K
01
23
45
1815 1850 1855 1861 1865 1880 1901 1913
Source: Authors’ elaboration on Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 26 / 34
Stylized facts: Centralities, 1855
Figure : Local and global centralities, 1855
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Degrees − 1855
Eig
enve
ctor
s −
185
5
Anhalt−Dessau
Austria Hungary
Anhalt−BernburgAnhaltAnhalt−K.then
Argentina
Baden
Bavaria
BelgiumBolivia
Bremen
Brunswick
BrunswickBrazil
Bulgaria
Canada
Chile
China
ColumbiaCosta Rica
Cuba
Denmark
Dominican Republic
Ecuador
El Salvador
Frankfurt
France
Greece
Guatamala
Haiti
HamburgHannover
HawaiiHohenzollernHomburgHonduras
Hesse−Cassel (Electorate)
Hesse−Electoral (looks like a duplicate with Hesse−Cassel)Hesse−Grossherzogtum (Duchy)Hohenzollern−HechingenHohenzollern−SigmaringenIcelandIran
Italy
JapanLiechtenstein
Lippe
L.beck
LuxembourgMecklenburg−Schwerin
MecklenburgMexico
Middle German UnionMontenegroModena
Nassau
North German Confederation
Nicaragua
Norwary
Netherlands
Orange Free State
Oldenburg
Parma
Paraguay
Prussia
Portugal
Peru
Romania
Russia
SaxonySchwarzburg−RudolstadtSchwarzburg−SondershausenSaxe−Coburg−Gotha
Spain
Serbia
Sardinia (before 1860)
SteuervereinSweden
Switzerland
Saxe−HildburghausenSaxony−MeiningenSaxe−Weimar−Eisenbach
Thuringian Custom Union
Transvaal
Turkey
UK
Uruguay
USA
Venice
Waldeck
W.rttemberg
Source: Authors’ elaboration on Phare’s data (2008)Luca De Benedictis, Silvia Nenci 27 / 34
Stylized facts: Centralities, 1865
Figure : Local and global centralities, 1865
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Degrees − 1865
Eig
enve
ctor
s −
186
5
Anhalt−Dessau
Austria Hungary
Anhalt−BernburgAnhaltAnhalt−K.then
Argentina
BadenBavaria
Belgium
Bolivia
Bremen
BrunswickBrunswick
Brazil
Bulgaria
Canada
Chile
China
Columbia
Costa Rica
Cuba
Denmark
Dominican Republic
Ecuador
El Salvador
Frankfurt
France
Greece
Guatamala
HaitiHamburg
Hannover
Hawaii
HohenzollernHomburg
Honduras
Hesse−Cassel (Electorate)Hesse−Electoral (looks like a duplicate with Hesse−Cassel)Hesse−Grossherzogtum (Duchy)
Hohenzollern−HechingenHohenzollern−SigmaringenIcelandIran
Italy
JapanLiechtenstein
Lippe
L.beck
LuxembourgMecklenburg−Schwerin
Mecklenburg
Mexico
Middle German UnionMontenegroModena
Nassau
North German Confederation
Nicaragua
Norwary
Netherlands
Orange Free State
Oldenburg
Parma
Paraguay
Prussia
Portugal
Peru
Romania
Russia
Saxony
Schwarzburg−RudolstadtSchwarzburg−Sondershausen
Saxe−Coburg−Gotha
Spain
SerbiaSardinia (before 1860)Steuerverein
Sweden
Switzerland
Saxe−HildburghausenSaxony−Meiningen
Saxe−Weimar−EisenbachThuringian Custom Union
Transvaal
Turkey
UK
Uruguay
USA
Venice
Waldeck
W.rttemberg
Source: Authors’ elaboration on Phare’s data (2008)Luca De Benedictis, Silvia Nenci 28 / 34
First Remarks
This work is a first step to visualize the PTAs network and assesshistorical trade integration emphasizing the importance of thestructural dimension
Preliminary remarks are:
The structure of the PTAss network changes over the timeThe Cobden Chevalier Treaties Network has not a remarkable densitySpecific countries play different roles in the trade integration networkstructure over time (see UK, France, USA, Prussia)Further steps:to extend the analysis to other datato define a research hypothesis and make some inference
Luca De Benedictis, Silvia Nenci 29 / 34
Thank you
Thanks!
Luca De Benedictis, Silvia Nenci 30 / 34
Data: the Phare’s database
Comprehensive data on preferential agreements all starts around 1950(data prior to this date are scattered and fragmented).
We use historical data on PTAs arising from the Pahres (2008)database that includes intra-European, intra-American, andinterregional agreements signed from 1815 to 1914.
The database contains all agreements by which two states agree tolower trade barriers between them (the number of treaties that eachcountry has in effect at a given moment). The DB codes treaties ineffect by dyad.
The main sources for the Phares database are government documentsand similar compilations (Calvo 1862; Cardoso de Oliveira 1912; First International American
Conference 1889 1890; Hertslet 1875-; Macgregor 1846; Ministerio degli Affari Esteri 1865 1899; Recueil des Traites et
Conventions 1858; Triepel, 1900 1945; U.S. Department of State 1890; U.S. Tariff Commission 1940; Vial Solar 1903).
The secondary literature has been used as well.
Luca De Benedictis, Silvia Nenci 31 / 34
Country list
Source: Phare’s data (2008)
Luca De Benedictis, Silvia Nenci 32 / 34
Types of trade agreements in 1815 - 1914
Table : Types of trade agreements in 1815-1913
Formal title Uses in this periodConvention on Commerce (and Navigation) Common instrument for long-term agreements, but also used for temporary extensions while negotiating a convention or treatyTreaty of Commerce (and Navigation) Common, though less common than conventions.Treaty of Amity/Friendship and Commerce (and Navigation) (and Establishment) Common for newly independent states or a first treaty between a pair of states.Declaration on Commercial Relations Common during renegotiations, especially for Italy; used very rarely for mutual grant of MFN without making specific tariff concessions.Extension of Treaty (or Convention) Typically short-term extensions of six months to three years while negotiating renewal or a new treaty (convention).Protocol on Extension of Commercial Treaty Often used during renegotiations, especially by Italy.Exchange of Notes on Commercial Relations Unusual; when used, normally grants reciprocal MFN without making additional tariff concessions; sometimes used to extend an earlier convention or treaty, especially during negotiations.Commercial Arrangement Unusual; typically grants MFN provisionally while negotiating over details.Provisional commercial arrangement Unusual; sometimes an extension of an earlier treaty or convention with modifications, sometimes a temporary agreement concerning pecific goods.Commercial Agreement Rare; typically covers trade in specific goods, sometimes used to extend earlier treaties indefinitely, rarely clarifies the position of a colony with respect to existing treaties.Declaration Concerning Treaty (or Convention) (or Provisional Agreement) Less common instrument used to extend previous arrangements during negotiations; also occasionally used to modify or clarify terms of revious treaties.Treaty of Peace (etc.) Rare but often important.Tariff Agreement Rare; all cases involve Ottoman Empire.Commercial Accord RareCustoms Union Rare except in Germany before unification.Customs Treaty RareModus vivendi; Exchange of Notes on modus vivendi Rare; most examples involve Spain or BrazilAgreement on Commercial Relations Only example is Canada-France, 1893.Treaty of Reciprocity RareProvisional Convention Rare.Provisional Commercial Arrangment RareAdditional Convention to Treaty (or Convention) Unusual choice of instruments to modify tariff reductions or add MFN provisions to an earlier treaty.
Source: Phare, 2008
Luca De Benedictis, Silvia Nenci 33 / 34
The Zollverein
Figure :
zollverein.png
Source:
Luca De Benedictis, Silvia Nenci 34 / 34