assessing the impact of trade policy measures along global value … › sites › default › files...

Wiener Institut für Internationale Wirtschaftsvergleiche

The Vienna Institute forInternational EconomicStudies

www.wiiw.ac.at

Robert StehrerThe Vienna Institute for International Economic Studies – wiiw

www.wiiw.ac.at

Assessing the impact of trade policy measures along Global Value Chains

NIESR workshop“Global Value Chains: Current developments and implications for Europe”

June 6, 2019NIESR, London, UK

2

Some general remarks

GVCs and European integration dynamics

Using GVC data (MC IOT) for trade policy impacts: 3 research strands

Assessing the role of trade policies in GVCs- Focus on Non-Tariff Measures (NTMs)

Plan of the talk

3

Some general remarks

4

Does not draw a completely different picture of trade patterns, but a more nuanced view emphasizing specific aspects

- E.g. trade specialization measured by RCAs are not too different when measures in value added trade terms (with a few exceptions maybe)

You never walk alone …- Puts emphasis on cross-country/cross-industry linkages and spillover effects

- Indirect effects are important (particularly in highly integrated areas, like EU)

Reconsider relative importance of imports and exports- The larger the foreign content of exports, the lower is the domestic content and

therefore the (direct) impact on GDP

- Imports are in important factor of production

A country’s trade deficit (or surplus) is the same in gross and value added terms

- Bilateral imbalances might be different

Selected general lessons from “MC-IOT based GVC research”

5

“Trade in value added” is nothing new …- In standard models (e.g. simple Ricardo model) all trade is TiVA

Question is: How much is ‘domestic value added’ traded: Not 100% if intermediates are imported

“Trade in intermediates” is nothing new …- Trade in ancient times (e.g. along the Silk Road)

Spices shipped to Europe used in the preparation of a meal

- Question is: How important is trade intermediates?It is important: 2/3 – 3/4 of global trade is trade in intermediatesRole of intra-firm trade is important (but under-researched and lack of data)

Measures based on MC-IOTs is mostly based on ‘trade in intermediates’- However, GVCs might be considered as only a subset of this (activities of firms)

GVCs are highly correlated to FDI patterns

- Or: everything is ‘global value chain’ (the Leontief inverse is fully occupied)

Common misunderstandings …

6

Measures of the importance of ‘GVCs’ and their changes over time- … all kind of (newly) measures …

MC IOTs (EORA, OECD TiVA, WIOD, …) are widely used in research and allow for better and more detailed analysis

- Analysis, modelling, Footprint analysis, Impact analysis, …

Strong need to be combined with other (satellite) data to exploit the full potential of GVC based research (and ‘spillover/linkages effects’)

- Employment, emissions, resource use,

- R&D, innovation, …

- Trade policy (focus of this presentation)

What’s then “new”?

7

Value chains and the EU manufacturing core

8

Value added exports in % of GDP, 2006 and 2014

Source: Extended WIOD database, wiiw calculations; preliminary results

0

10

20

30

40

50

60

70

80

Luxe

mbo

urg

Irela

ndBe

lgiu

mN

ethe

rland

sG

erm

any

Denm

ark

Aust

riaSw

eden

Finl

and

Gre

ece

Port

ugal

Spai

nU

nite

d Ki

ngdo

mIta

lyFr

ance

Mal

taSl

ovak

iaHu

ngar

ySl

oven

iaLi

thua

nia

Czec

hia

Latv

iaEs

toni

aBu

lgar

iaCy

prus

Pola

ndRo

man

iaSe

rbia

Croa

tiaM

onte

negr

oM

aced

onia

Alba

nia

Bosn

ia a

nd H

erze

govi

naSw

itzer

land

Icel

and

Nor

way

Russ

iaTu

rkey

Ukr

aine

Taiw

anKo

rea

Cana

daCh

ina

Rest

-of-W

orld

Indo

nesia

Aust

ralia

Mex

ico

Japa

nIn

dia

Braz

ilU

SA

EU-15 EU-CEE EU-WB OtherEurope

WiderEurope

Other world

9

Source: WIOD database, wiiw calculations (Stehrer/Stöllinger, 2013 FIW).

15 pp

Austrian value added exports in % of GDP, 1995 to 2011

8 pp

10

Foreign content in % of gross exports, 2006 and 2014

0

10

20

30

40

50

60

70

80

Luxe

mbo

urg

Irela

ndBe

lgiu

mDe

nmar

kPo

rtug

alN

ethe

rland

sFi

nlan

dAu

stria

Gre

ece

Swed

enIta

lySp

ain

Ger

man

yFr

ance

Uni

ted

King

dom

Mal

taHu

ngar

ySl

ovak

iaEs

toni

aCz

echi

aBu

lgar

iaSl

oven

iaLa

tvia

Lith

uani

aCy

prus

Pola

ndRo

man

iaM

aced

onia

Bosn

ia a

nd H

erze

govi

naM

onte

negr

oSe

rbia

Alba

nia

Croa

tiaSw

itzer

land

Icel

and

Nor

way

Ukr

aine

Turk

eyRu

ssia

Taiw

anKo

rea

Rest

-of-W

orld

Mex

ico

Japa

nCa

nada

Indi

aIn

done

siaCh

ina

Aust

ralia

Braz

ilU

SA

EU-15 EU-CEE EU-WB OtherEurope

WiderEurope

Other world

Source: Extended WIOD database, wiiw calculations (preliminary results)

11

Car assembled in Slovak Republic

12

Value added generated in respective country due to German exports in transport equipment industryin % of GDP of respective country

Source: WIOD..

13

Value added structure of SK transport equipment exportsin % of gross exports

Source: WIOD..Notes: EU-14 is EU-15 without Germany; EU-12 means EU-12 without the respective country.

14

European Manufacturing CoreValue added exports in % of GDP

Source: WIOD database, own calculations.

HUN

DEUAUT

IRL

POL

CZEMLT

LUX

GRC

DNKLTU

ROU

BGR

ESP

SWE

NLD

SVN

GBR

SVK

ITA

BEL

FRA

EST

PRT

LVA

CYP

FIN

0

10

20

30

40

50

60

2011

0 10 20 30 40 50 601995

15

Sectoral differences in GVC participationHigh-tech industries are involved relatively more in global (regional) production networks

Source: WIOD database, own calculations.

AtBC

15t16

17t181920

21t22

24

25

26

27t28

29

30t33

34t35

36t37

E

F

505152

H

60

61

62

6364

J

70

71t74L

M

N O

P05

10

15

20

200

7

0 5 10 15 201995

Machinery, electrical engineering, transport equipment

16

These „manufacturing core „ countries have also managedto specialise in high-tech industries

17

Source: wiiw Handbook of Statistics 2016

Share of manufacturing in GDP, 2015

West-East integration of production particularly in medium-high and high-tech industries

Technology spillovers via FDI and intermediates trade and favourable conditions for catching-up processes in CEECs

Production and quality upgrading

Export upgrading in medium-quality segment

However, CEECs show low share of business relatedservices (KIBS) of which large shares are imported

EU as a whole could maintain ratherstrong manufacturing base

Strong specialisation patterns in EuropeEmergence of „EU manufacturing core“ and „EU services core“

‚Left-outs‘ (with no clear specialisation patterns); mostly peripheral countries

Imbalances problems in EU

The emergence of the Central European Manufacturing Core

18

Climbing up the value chain – the Smile curve

R&D

productionlogistics

support services

value added creation

value chain functions

pre-production production post-production

headquarter services

19

To smile ...

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on

1 HQ 2 R&D 3 Production 4 Logistics5 Support services

AT

0

.5

1

1.5

2

relat

ive fu

nctio

nal s

pecia

lisati

on


DE

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


FR

0

.5

1

1.5

2

relat

ive fu

nctio

nal s

pecia

lisati

on


UK

0

1

2

3

relat

ive fu

nctio

nal s

pecia

lisati

on


NL

0

.5

1

1.5

2

relat

ive fu

nctio

nal s

pecia

lisati

on


IT

Functional specialisation in selected Western EU core countries, averages 2003-2015

Source: R. Stöllinger, work in progress

20

... or not to smile?

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


CZ

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


HU

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


PL

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


SK

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


RO

0

.5

1

1.5

relat

ive fu

nctio

nal s

pecia

lisati

on


BG

Source: R. Stöllinger, work in progress

Functional specialisation in Visegrad-4 countries, Romania and Bulgaria, averages 2003-2015

21

Trade policy and GVCs

22

Measuring the direct and indirect effects - Leontief approach (IO modelling)- CGE, etc. modelling

Structural gravity modelling- MC IOTs are heavily used in such models which rely on consistency and information

on intra-country flows- New advances in methods

Effects of trade policy measures along GVCs- Cascading effect of tariffs if (intermediate) products cross borders multiple times

Trade policy aspects: 3 important approaches

23

Indirect effects Trade impacts on GDP are smaller than usual openness measures would

suggest- e.g. Austria: Exports/GDP ratio ~ 60%; Value added exports to GDP ratio ~ 33%

- Thus openness has to be defined broader to include imports

Trade impacts are larger as ‚indirect‘ effects become more important- US tariff on cars not only impacts on Germany, but also suppliers (Eastern Europe)- Brexit not only impacts on direct exporters to UK, but also on all upstream countries

Trade diversion effects can be quite complex

24

Value added generation due to US ‘car’ imports, in % of each country’s GDP

Source: WIOD release 2016; wiiw calculations.

0

0.5

1

1.5

2

2.5

3

3.5

Mex

ico

Kor

eaC

anad

aJa

pan

Ger

man

yH

unga

ryS

lova

k R

epub

licTa

iwan

Cze

ch R

epub

licA

ustri

aS

love

nia

Pol

and

Italy

Sw

eden

Rom

ania

Chi

naU

nite

d K

ingd

omB

elgi

umN

ethe

rland

sLu

xem

bour

gS

pain

Bul

garia

Turk

eyU

SA

Indo

nesi

aN

orw

ayS

witz

erla

ndR

ussi

aP

ortu

gal

Irela

ndFr

ance

Finl

and

Est

onia

Cro

atia

Den

mar

kLi

thua

nia

Aus

tralia

Mal

taIn

dia

Latv

iaB

razi

lC

ypru

sG

reec

e

Direct value added Indirect value added

25

Employment generation due to US ‘car’ imports, in % of each country’s GDP

Source: WIOD release 2016; wiiw calculations.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Mex

ico

Can

ada

Kor

eaJa

pan

Ger

man

yH

unga

ryS

lova

k R

eplic

Cze

ch R

epub

licTa

iwan

Aus

tria

Italy

Slo

veni

aP

olan

dS

wed

enR

oman

iaLu

xem

bour

gB

elgi

umU

nite

d K

ingd

omC

hina

Spa

inN

ethe

rland

sTu

rkey

Bul

garia

Cro

atia

Sw

itzer

land

Fran

ceP

ortu

gal

Est

onia

Finl

and

Irela

ndM

alta

Den

mar

kIn

done

sia

Nor

way

Aus

tralia

Latv

iaLi

thua

nia

Rus

sia

Bra

zil

Cyp

rus

Indi

aG

reec

e

Direct employment Indirect employment

26

Recent methodological advances in gravity modellingsuggest to

include intra-country flows plus other technical recommendations

Scenario: Accession of Western Balkan countries to EU

Scenario: EU-JPN partnership agreement

Source: Reiter and Stehrer, 2017 Source: Grübler, Reiter and Stehrer, 2017

27

Still wanted: Deeper insights in the effects of trade policy measures along GVCs (ongoing research)

Still existing challenges to understand complexity of impacts- Models discussed so far hide channels how trade policy impacts along GVCs- In GVCs product standards and regulations become more important for proper

functioning- Product standards and regulations are an important (and debated) issue in trade

negotiationsConsumer safety issues, environmental protection, health standards, etc

Research challenges- What is the impact of non-tariff measures and how this can be quantified?- How big is the ‚cascade effect‘ along GVCs?- Differentiation between price, quantity and quality effect

28

Non-tariff measures trickling through GVCs(joint work with Mahdi Ghodsi,, Julia Grübler and Oliver Reiter; work in progress)

1 The increasing importance of non-tariff measures

Research was done in the project "Productivity, Non-Tariff Measures and Openness" (PRONTO) funded by the European Commission under the 7th Framework Programme, Theme SSH.2013.4.3-3 "Untapped Potential for Growth and Employment -Reducing the Cost of Non-Tariff Measures in Goods, Services and Investment", Grant agreement No. 613504.

29

Number of PTAs over time

The Political Economy of non-tariff measures (NTMs)Depth of PTAs over time

Trends in Tariffs and NTMs

0

500

1,000

1,500

2,000

2,500

3,000

0

2

4

6

8

10

12

14

1995

1997

1999

2001

2003

2005

2007

2009

2011

United StatesSimple avg. over all countriesGermanyNTM notifications

Note: Applied tariff rate, weighted mean, all products (%)Sources: WTO I-TIP (NTM data), World Bank (WITS tariff data), wiiw calculations. Dür et al, 2014 (Number and Depth of PTAs).

30

“Non-tariff measures (NTMs) are policy measures, other than ordinary customs tariffs, that can potentially have an economic effect on international trade in goods, changing quantities traded, or prices or both.” (UNCTAD, 2017)

Non-tariff measure classification by MAST chapterA: Sanitary and phytosanitary measures B: Technical barriers to tradeC: Pre-shipment inspection and other formalities D: Contingent trade-protective measures E: Non-automatic licensing, quotas, prohibitions and quantity-control measures other than for SPS or TBT reasons F: Price-control measures, including additional taxes and charges G: Finance measures H: Measures affecting competition I: Trade-related investment measures J: Distribution restrictions K: Restrictions on post-sales services L: Subsidies (excluding export subsidies under P7) M: Government procurement restrictions N: Intellectual property O: Rules of origin P: Export-related measures

What are non-tariff measures (NTMs)?

31

Sanitary and phytosanitary measures (SPS)- ~30% of NTM notifications Bilateral SPS measure of the EU blocking the import of dried beans from Nigeria due to

pesticide residues at levels exceeding the reference dose as stated by the European Food Safety Authority [WTO Document: G/SPS/N/EU/131, 29 June 2015]

Measures to prevent the spread of transmissible diseases, such as spongiform encephalopathies [WTO Document: G/SPS/N/EU/67, 4 March 2014]

Chlorinated chicken

Technical Barriers to Trade (TBT)- ~45% of NTM notifications Energy labelling requirement for storage cabinets, including those used for refrigeration. The

stated aim of the EU is to pull the market towards more environmentally friendly products by providing more information to end-users.[WTO Document: G/TBT/N/EU/178, 28 January 2014]

Examples of NTMs

32

Data collected from WTO trade notifications WTO I-TIP- and harmonised ...

15 out of 16 classifications concern imports thereof 3 are technical in nature:

- TBT… technical barriers to trade- SPS … sanitary and phytosanitary measures- Pre-shipment inspection and other formalities

thereof 12 are non-technical measures:- antidumping (ADP)- countervailing Duties (CVD)- (special) safeguards (SG) - quantity control measures (e.g. quotas, licensing) (QRS)- …

Specific trade concerns (SPS & TBT)

Note: Total number of notifications (37,982) to the WTO between 1979 and March 2015; Graph excludes 899 Specific Trade Concerns. Sources: UNCTAD, 2013 (Classification), WTO I-TIP (NTM data), wiiw calculations.

Non-tariff measures (NTMs): Overview

Other Counteracting measures (OCA)

TBT50%

SPS35%

QRS2%

ADP10%

SG1%

SSG1%

CVD1%

33

Source: Ghodsi et al. (2017)

NTM notifications over time

0

500

1,000

1,500

2,000

2,500

3,000

3,500

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

TBT STC(TBT) SPS STC(SPS) QRS ADP OCA

34

Non-tariff measures trickling through GVCs1 The increasing importance of non-tariff measures2 Assessing the impact of non-tariff measures on trade

35

Ad-valorem equivalents (AVEs) of NTMsor: How to make NTMs directly comparable to tariffs?

Price gap method Indirect Approach

Wha

t? Deriving AVEs from an analysis of the price wedge due to the implementation of NTMs

Deriving AVEs with a Gravity estimation approach from the impact on import quantities and import demand elasticities

Who

? • Dean et al., 2009

• Nimenya et al., 2012

• Cadot and Gourdon, 2015

• Kee et al., 2009

• Bratt, 2014

• Beghin et al, 2014

Issu

es

› Necessity to compare different prices along the production and supply chain

› Neglect of product quality differences

› Price data availability usually restricts to few countries for a small set of products

› Based on import demand elasticities, which are themselves estimates

› Neglect of product quality differences

New data set (WTO I-TIP complemented by Ghodsi et al) Types and intensity of NTMs Using a panel structure Assess effects bilaterally Assess impacts in GVCs

36

Based on GDP-function approach:

- estimation of bilateral import price elasticities at sectoral level 𝜀𝜀ij,mbased on detailed 6-digit HS data

Gravity estimation to assess effect of NTM on trade flows βn,i,m,t

and transformation to AVEs:

Calculating bilateral indirect effects of tariffs and NTMs for each industry by taking into account inter-country and inter-industry linkages using WIOD

Assessing the impacts of these on exports

Method (in a nutshell)

𝜕𝜕 ln mii,m

𝜕𝜕NTMn,ij,m=

𝜕𝜕 ln mij,m

𝜕𝜕 ln pij,m

𝜕𝜕 ln pij,m𝜕𝜕NTMn,ij,m

= εij,mAVEn,ij,𝑚𝑚

τn,ij,m,t =eβn,i,m,t − 1

εij,m

Ad-valorem equivalent of country i‘s importsof good from j at time t

37

-100 -50 0 50 100 150

Meat of fowls of species Gallus domesticus, not cut in…

Meat & edible meat offal of ducks/geese/guinea fowls…

Meat of turkeys, not cut in pieces, frozen

Cuts & edible offal of species Gallus domesticus, frozen

Meat of ducks/geese/guinea fowls, not cut in pieces, frozen

Fatty livers of ducks/geese/guinea fowls, fresh/chilled

Cuts & edible offal of turkey, frozen

Meat of fowls of species Gallus domesticus, not cut in…

Cuts & edible offal of turkey, fresh/chilled

Cuts & edible offal of species Gallus domesticus, fresh/chilled

Meat of ducks/geese/guinea fowls, not cut in pieces,…

Meat & edible meat offal of ducks/geese/guinea fowls…

Meat of turkeys, not cut in pieces, fresh/chilled

TBT SPSExample: AVEs on chicken products


38

Estimated ad-valorem equivalents (simple statistics)

Sample Mean

Mean AVE>0

#

AVE>0Mean AVE<0

#

AVE<0

#

AVE=0

SPS 2.2% 41.4% 73,087 -38.0% 43,055 508,218

TBT 5.2% 36.7% 127,289 -23.4% 61,384 435,687

Tariff 2.9%

TRI 10.3% 28.3% 316,223 -33.7% 74,909 233,228

Note: Number of observations is (44*44*22 – 44*22)*15 = 624360

Source: Ghodsi and Stehrer (2019)

39

“TBT/SPS measures do not unambiguously increase or decrease trade. In general, TBT/SPS measures have positive effects for more technologically advanced sectors, but negative effects on trade in fresh and processed goods. As economic theory suggests, the introduction of a new TBT/SPS measure yields a trade-off between higher costs of adaption to new requirements for producers and lower information costs for consumers, who can be confident about the quality of the product in question.“ (WTO (2012), World Trade Report)

NTMs should not be seen as “trade costs”, implying that removal of those will –similar as tariffs in standard trade models – bring further beneficial effects (‘welfare gains’)

- NTMs are often important and beneficial itself as e.g. safety standards, environmental protection, rules, etc.

- NTMs can also lead to reduction of trade costs (e.g. due to harmonisation)

Related research shows that “Ad-Valorem Equivalents (AVE)” of NTMs are in almost 50% of cases negative, i.e. trade enhancing- Bratt (2014); Beghin et al. (2014); Ghodsi et al. (2017)

Non-tariff measures are not necessarily Non-tariff barriers

40

Simple mean of bilateral trade restrictive indices, 2000–2014


-5%

0%

5%

10%

15%

20%

25%

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

SPS TBT T TRI

41

Trade-weighted mean of bilateral trade restrictive indices, 2000–2014

-2%

0%

2%

4%

6%

8%

10%

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

SPS TBT T TRI


42

Trade restrictions by country (simple averages)


-15%

-10%

-5%

0%

5%

10%

15%

20%

IND

BR

AK

OR

CH

NS

WE

GR

CM

EX

LTU

RO

WB

ELB

GR

RO

UTW

NH

RV

EST

CY

PC

AN

AU

SC

ZEFR

AJP

NP

RT

US

AE

SP ITA

DN

KD

EU

SVK IR

LFI

NC

HE

NLD IDN

LUX

LVA

GBR

TUR

RU

SH

UN

AU

TP

OL

SVN

NO

RM

LT

Tariff rate TBT SPS Total TPIM

43

Size of AVEs of TBTs and SPS are broadly comparable to tariffs

TBT and SPS can have trade enhancing effects (negative AVEs), but on average hampering trade (positive AVEs)

Crisis effect stronger for NTMs

Important findings

44

Non-tariff measures trickling through GVCs1 The increasing importance of non-tariff measures2 Assessing the impact of non-tariff measures on trade3 The impact of NTMs (and tariffs) along GVCs

45

Example question:

What‘s the impact of a specific NTM (TBT, SPS,...)

imposed by the EU

on imports from goods produced in the Chinese electrical equipment industry

imported from the Slovak Republic machinery industry

on the exports of the German automotive industry?

Considering the effect of NTMs (and tariffs) alongGVCs makes it even trickier

46

3. Cumulative AVEs in GVCs:𝛕𝛕t,GVC = (𝐈𝐈𝐂𝐂⨂𝐞𝐞1N) × 𝐁𝐁t × 𝐈𝐈 − 𝐀𝐀t −1 ′, ∀𝑡𝑡 = 1, … ,𝑇𝑇

- 𝐀𝐀t is the NC × NC matrix of technical coefficients in each year 𝑡𝑡- 𝐁𝐁𝐭𝐭 is a NC × NC matrix of element-by-element multiplication of each countries’ technical

coefficients 𝐀𝐀t with a vector of annual bilateral AVEs denoted by 𝛕𝛕tijW,n (of dimension NC ×1) [assuming that using industries’ k in a country i are facing the same AVE when importing ‘good’ W from country j]

- 𝐈𝐈C is the identity matrix with number of countries C as its dimension- 𝐞𝐞1N is a row vector of ones with the number of industries N ,

- .

⇒ The resulting 𝛕𝛕t,GVC is a NC × C matrix indicating the indirect implied AVEs of NTMs and tariffs at the bilateral-industry level, ie. 𝜏𝜏ij,k⇒ Tariffs and AVEs of NTMs imposed by country i industry k on imports from j

Methodology

47

Mean of indirect bilateral trade restrictive indices, 2000–2014

-0.05%

0.00%

0.05%

0.10%

0.15%

0.20%

0.25%

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

SPS TBT T TRI

Note: Includes services sectors


48

Non-tariff measures trickling through GVCs1 The increasing importance of non-tariff measures2 Assessing the impact of non-tariff measures on trade3 The impact of NTMs (and tariffs) along GVCs4 Impact on gross trade flows

49

Channel 1 Trade policy measures imposed by the destination country and faced by the

exporting country/industry – standard trade policy measures in the literature (e.g. import tariffs, AVEs of NTMs on imported products)

Channel 2 Trade policy measures imposed by the home country against imports

(protectionism):

- higher import costs

- lowering competition

- improve safety and quality requirements of imported products

Channel 3 Trade policy measures levied against the intermediary inputs of an industry which

accumulate over stages of production (i.e. backward linkages)

Channels of trade policy impacts along GVCs

50

Applying a structural gravity model with the Poisson pseudo maximum likelihood

(PPML) developed by Yotov et al. (2016) and Larch et al. (2018):

xtijW = e ∑n γn1 τn,tjiW+ ∑n γn2 τn,tijW+∑n γn3 τn,tijW,GVC+γiWt+γjWt+γijW+ϕtijW

𝑛𝑛 ∈ 𝑇𝑇, 𝑆𝑆𝑆𝑆𝑆𝑆,𝑇𝑇𝑇𝑇𝑇𝑇

Data: AVEs of NTMsAVEs for tariffs with priority PRF, MFN, and AHS from TRAINS, and WTOPTA from the WTOGravity data from WDI of the World Bank, PWT, FAO, and CEPIIWIOD (Timmer et al., 2015)Note: 𝑥𝑥𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 … Exports of (i,k) to j𝜏𝜏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 … tariffs levied by country j on imports from (ik)𝜏𝜏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 … tariffs levied by country i on imports from j on products/industry k𝜏𝜏𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 … tariffs levied by country i on imports from j on products/industry k via upstream linkages

Assessing the diverse impact of tariffs and NTMs in GVCs on export :

51

Impact on bilateral export performance (preliminary results)

Robust standard errors clustered by bilateral-sectors in parentheses; * p<0.1, ** p<0.05, *** p<0.01All estimations are including importer-sector-time, exporter-sector-time, and bilateral-sector fixed effects.

Dep. Var.: 𝐥𝐥𝐥𝐥𝐗𝐗𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖

Channel 1 Channel 1 & 2 Channel 1, 2, & 356 sectors Non-services 56 sectors Non-services 56 sectors Non-services

𝛕𝛕𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 -2.03*** -2.03*** -1.21*** -1.21*** -1.22*** -1.22***(0.23) (0.23) (0.20) (0.20) (0.20) (0.20)

𝛕𝛕𝐓𝐓𝐁𝐁𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 -0.032*** -0.032*** -0.034*** -0.034*** -0.034*** -0.034***(0.0068) (0.0068) (0.007) (0.007) (0.0074) (0.0074)

𝛕𝛕𝐒𝐒𝐒𝐒𝐒𝐒,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 0.0024 0.0024 0.0021 0.0021 0.0012 0.0012 (0.0089) (0.0089) (0.0089) (0.0089) (0.0086) (0.0086)

𝛕𝛕𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 -1.488*** -1.487*** -1.97*** -1.98***(0.19) (0.19) (0.18) (0.18)

𝛕𝛕𝐓𝐓𝐁𝐁𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 0.0043 0.0043 0.0043 0.0043 (0.0061) (0.0061) (0.0061) (0.0061)

𝛕𝛕𝐒𝐒𝐒𝐒𝐒𝐒,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖 0.0012 0.0012 0.00045 0.00043 (0.0069) (0.0069) (0.0068) (0.0068)

𝛕𝛕𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖,𝐆𝐆𝐆𝐆𝐂𝐂 16.4*** 16.7***(3.12) (3.36)

𝛕𝛕𝐓𝐓𝐁𝐁𝐓𝐓,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖,𝐆𝐆𝐆𝐆𝐂𝐂 -0.34*** -0.31***(0.094) (0.098)

𝛕𝛕𝐒𝐒𝐒𝐒𝐒𝐒,𝐭𝐭𝐢𝐢𝐢𝐢𝐖𝐖,𝐆𝐆𝐆𝐆𝐂𝐂 0.20*** 0.19***(0.063) (0.064)

Observations 1491770 629912 1491770 629912 1491770 629912 R-squared 1.000 0.999 1.000 0.999 1.000 0.999

52

Channel 1- Tariffs have the expected negative impact on a country’s exports- TBTs imposed on a country’s exports have negative (but much smaller) effect- SPS no significant impact

Channel 2- Tariffs a country is imposing on its imports is negatively affecting exports (importance

of intermediates trade and production integration)- However, we find no significant impact by imposing NTMs on imports

Channel 3- The effect of tariffs along GVCs is significantly positive (tbc)

Increase in costs and therefore price of exported goods, Quality effect (?)

- TBTs along GVCs impact negatively- SPS along GVCs impact positively

Quality impact on exports, costs

Summary of these findings

53

Standards and regulations in trade have multiple aims- Consumer safety, harmonisation of technical standards, health effects, ...- Complying with European standards and values

Impact of NTMs on trade are diverse- due to complexity of technical regulations, etc.- even if negative impacts occur, NTMs serve various other purposes

Evidence suggests that on average impact is comparable to tariffs (on average)

Trade measures accumulate when international production sharing is important

Even if some NTMs have trade-impeding effects, these have to weighed off with other important effects of promoting standards and consumer needs

Summary: NTMs, important but tricky

54

GVC perspective provides useful additional insights

Additional research requires compatible set of satellite data

Spillover effects (in general) are measurable but difficult to track in detail

Trade policy aspects- NTMs are important but heavily under-researched- Challenge to differentiate quantity, price and quality effects

Concluding remarks

Wiener Institut für Internationale Wirtschaftsvergleiche

The Vienna Institute forInternational EconomicStudies

www.wiiw.ac.at

Thanks for attention!

Visit us on:

https://www.youtube.com/channel/UCyNuSfsgzgBsIf7AaH8Vk-Q/videos

https://twitter.com/wiiw_news/

https://www.facebook.com/wiiw.economic.studies/

https://wiiw.ac.at/

https://wiiw.ac.at/mailings.html

56

Complexity of NTMs: Rosendorff (1996), Otsuki et al. (2001), Vandenbussche and Zanardi (2008), Moore and Zanardi (2011), de Almeida et al. (2012), Beghin et al. (2012), Sanjuán López et al. (2013), Ghodsi (2015a,b, 2019)

Elasticity: Kee et al. (2008), Ghodsi et al. (2016b)

AVE for NTMs: Kee et al. (2009), Beghin et al. (2015), Bratt (2017), Ghodsi et al. (2016a), Niu et al.(2018)

GVC: Gereffi et al. (2005), Gereffi and Sturgeon (2013), Timmer et al. (2013), Antras et al. (2012), Backer and Miroudot (2013), Rouzet and Miroudot (2013), Miroudot et al. (2013); Johnson and Noguera (2012); OECD (2013); Timmer et al. (2014); Koopman et al. (2014).

Structural Gravity: Helpman et al. (2008); Head and Mayer (2014); Yotov et al. (2016) and Larch et al. (2018)

Selected related literature

57

Ghodsi, M., J. Grübler, O. Reiter and R. Stehrer (2017), The Evolution of Non-Tariff Measuresand their Diverse Effects on Trade, wiiw Research Report 418, May 2017.

Ghodsi, M., J. Grübler, and R. Stehrer (2016), Import Demand Elasticities Revisited, wiiwWorking Paper 132, November 2016.

Ghodsi, M., J. Grübler, and R. Stehrer (2016), Estimating Importer-Specific Ad Valorem Equivalents of Non-Tariff Measures, wiiw Working Paper 129, September 2016

Ghodsi, M., and R. Stehrer (2019), EU Trade Regulations and Imports of Hygienic Poultry, wiiwWorking Paper 135, April 2017; forthcoming.

Ghodsi, M. and R. Stehrer (2019), Quality impacts of NTMs, in progress.

Ghodsi, M. and R. Stehrer (2019), Non-tariff measures trickling through global value chains, in progress.

Related own research

58

Appendix

59

EU as a whole is rather well positioned to maintain competitiveness in manufacturing

• The ‘EU manufacturing divide’ poses a formidable challenge as leading to erosion of national manufacturing systems in part of the E Reduced long term growth prospects for EU-MS outside the manufacturing core

(except maybe those with strong business services sector) Recurring current account imbalances

Diverging specialisation patterns hard to counteract (even with targeted policy measures)

Several policy options but politically difficult to implement European industrial policy initiatives is the least likely policy option

Fiscal transfers

Increased EU internal labour mobility

Implications for EU industrial policy?

60

NTM notifications by product groups

Sec. XXISec. XIVSec. XIXSec. VIIISec. XIISec. XSec. IXSec. XISec. XIIISec. XVIISec. XVIIISec. XXSec. VSec. XVSec. IIISec. VIISec. XVISec. VISec. IVSec. ISec. II

0 2,000 4,000 6,000 8,000

Works of art and antiquesPearls, precious stones and metals; coin

Arms and ammunitionHides, skins and articles; saddlery and travel goods

Footwear, headgear; feathers, artif. flowers, fansPaper, paperboard and articles

Wood, cork and articles; basketwareTextiles and articles

Articles of stone, plaster; ceramic prod.; glassVehicles, aircraft and vessels

Instruments, clocks, recorders and reproducersMiscellaneous manufactured articles

Mineral productsBase metals and articles

Animal and vegetable fats, oils and waxesResins, plastics and articles; rubber and articles

Machinery and electrical equipmentProducts of the chemical and allied industries

Prepared foodstuff; beverages, spirits, vinegar; tobaccoLive animals and products

Vegetable products

TBT STC(TBT) SPS STC(SPS) QRS ADP OCA Summe


61

NTM notifications by imposing country


62

NTM notifications by affected countries


63

NTM notifications by level of development


64

1. Bilateral import demand elasticities:- Extending the approach proposed by Kee et al. (2008)

- 𝑠𝑠𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ,𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡, 𝑣𝑣𝑡𝑡𝑡𝑡𝑡𝑡𝑡 = 𝛼𝛼0 + ∑𝑡𝑡=1𝐽𝐽 𝛼𝛼𝑡𝑡𝑡𝑡

�ln 𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

+ ̂�̅�𝜂𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ + 𝛼𝛼𝑡𝑡𝑡𝑡𝑡2 + 𝛼𝛼𝑡𝑡𝑡𝑡𝑡2 + 𝛼𝛼𝑡𝑡𝑡𝑡𝑡 + 𝜚𝜚𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ,

- ∀𝑊𝑊 = 1, . . . , ℎ,𝐻𝐻,𝐻𝐻′ , 𝑖𝑖 ≠ 𝑗𝑗

- 𝜀𝜀𝑡𝑡𝑡𝑡𝑖𝑖 ≡ 𝜕𝜕𝑞𝑞𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑝𝑝𝑡𝑡 ,𝑣𝑣𝑡𝑡𝜕𝜕𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑞𝑞𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

= 𝛼𝛼𝑡𝑡𝑡𝑡̅𝑠𝑠𝑡𝑡𝑡𝑡𝑡𝑡

+ �̅�𝑠𝑡𝑡𝑡𝑡𝑡𝑡 − 1, 𝑠𝑠𝑡𝑡𝑡𝑡𝑡𝑡𝑖𝑖 < 0 , 𝜀𝜀𝑡𝑡𝑡𝑡𝑖𝑖 �< −1 𝑖𝑖𝑖𝑖 𝛼𝛼𝑡𝑡𝑡𝑡 > 0

= �̅�𝑠𝑡𝑡𝑡𝑡𝑡𝑡 − 1 𝑖𝑖𝑖𝑖 𝛼𝛼𝑡𝑡𝑡𝑡 = 0> −1 𝑖𝑖𝑖𝑖 𝛼𝛼𝑡𝑡𝑡𝑡 < 0

, 𝑖𝑖 ≠ 𝑗𝑗

- 𝑠𝑠𝑡𝑡𝑡𝑡𝑡𝑡𝑡 is the share of the import value of product ℎ shipped from country j to country i in the GDP of country i at time t;

- �ln 𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑝𝑝𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

is the fitted values of price index from first-stage instrumental variable (IV) approach;

- Three instruments are used: simple and distance weighted average of world price index; and export (free on board, f.o.b.) price.

- ̂�̅�𝜂𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ is inverse mills ratio from the sample selection equation (Heckman, 1979; Helpman et al., 2008; Semykina and Woolridge, 2010)

Methodology

65

2. Annual bilateral AVE for NTMs:- Extending the approach proposed by Kee et al. (2009):

- ln 𝑞𝑞𝑡𝑡𝑡𝑡𝑡𝑡𝑡 = 𝛽𝛽0 + 𝛽𝛽𝑇𝑇𝑡𝑡 ln 1 + 𝑇𝑇𝑡𝑡𝑡𝑡𝑡𝑡𝑡 + ∑𝑡𝑡=1𝐼𝐼 ∑𝑛𝑛=1𝑁𝑁 𝛽𝛽𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡 �𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡𝑡 + 𝛽𝛽1�̂�𝑧𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ + 𝛽𝛽2�̂�𝑧𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ 2 +𝛽𝛽3�̂�𝑧𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ 3 + 𝛽𝛽4 ̂�̅�𝜂𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ + 𝛽𝛽𝑡𝑡𝑡𝑡𝑡2 + 𝛽𝛽𝑡𝑡𝑡𝑡𝑡2 + 𝛽𝛽𝑡𝑡𝑡𝑡𝑡 + 𝜓𝜓𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ,

- ∀𝑊𝑊 = 1, . . . ,ℎ,𝐻𝐻,𝐻𝐻′ , ∀𝑡𝑡 = 1, … ,𝑇𝑇, 𝑛𝑛 ∈ 𝑇𝑇𝑇𝑇𝑇𝑇, 𝑆𝑆𝑆𝑆𝑆𝑆 , 𝑖𝑖 ≠ 𝑗𝑗

- 𝜏𝜏𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡𝑡𝑡 = 1𝜀𝜀𝑡𝑡𝑡𝑡𝑡𝑡

𝜕𝜕 ln 𝑞𝑞𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝜕𝜕𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

= 𝑒𝑒𝛽𝛽𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡−1𝜀𝜀𝑡𝑡𝑡𝑡𝑡𝑡

× 100, 𝑛𝑛 ∈ 𝑇𝑇𝑇𝑇𝑇𝑇, 𝑆𝑆𝑆𝑆𝑆𝑆 , 𝑖𝑖 ≠ 𝑗𝑗

- where ln 𝑞𝑞𝑡𝑡𝑡𝑡𝑡𝑡𝑡 is the natural log of the import quantity of product h to country i from country j at time t; ̂�̅�𝜂𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ and �̂�𝑧𝑡𝑡𝑡𝑡𝑡𝑡𝑡∗ are respectively inverse mills ratio and probability of exporting (firm heterogeneity measure) from the sample selection equation; �𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡𝑡 is the fitted value from the first stage of instrumental variable (IV) approach with three instruments of:

1. Reciprocal NTM: 𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡𝑡

2. 𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑡𝑡𝑡𝑢𝑢 = ∑𝑘𝑘

𝑢𝑢𝑡𝑡𝑡𝑡𝑡𝑡𝑡∑𝑡𝑡 𝑢𝑢𝑡𝑡𝑡𝑡𝑡𝑡𝑡

𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑘𝑘𝑡 , 𝑘𝑘 ≠ 𝑗𝑗 ∧ 𝑖𝑖 ≠ 𝑗𝑗 ∧ 𝑘𝑘 ≠ 𝑖𝑖, ∀𝑛𝑛 ∈ 𝑇𝑇𝑇𝑇𝑇𝑇, 𝑆𝑆𝑆𝑆𝑆𝑆

3. 𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑛𝑛𝑡𝑡𝑖𝑖𝑡𝑢𝑢 = ∑𝑡𝑡 ∑𝑘𝑘

𝑢𝑢𝑡𝑡𝑡𝑡𝑡𝑡𝑡∑𝑡𝑡 𝑢𝑢𝑡𝑡𝑡𝑡𝑡𝑡𝑡

𝑁𝑁𝑇𝑇𝑁𝑁𝑡𝑡𝑡𝑡𝑘𝑘𝑡 , 𝑘𝑘 ≠ 𝑗𝑗 ∧ 𝑖𝑖 ≠ 𝑗𝑗 ∧ 𝑘𝑘 ≠ 𝑖𝑖, ∀𝑛𝑛 ∈ 𝑇𝑇𝑇𝑇𝑇𝑇,𝑆𝑆𝑆𝑆𝑆𝑆

Methodology

assessing the impact of trade policy measures along global value … › sites › default › files...

Documents