(ANIKA) Budhwar Peth Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
Julia Wörz. The trade and demand nexus:Do global value chains matter?
1. The trade and demand nexus:
Do global value chains matter?
Julia Wörz
Foreign Research Division
Oesterreichische Nationalbank
Joint work with Alexander Al-Haschimi, Frauke Skudelny and
Elena Vaccarino (ECB)
Eesti Pank Seminar, 3 September 2015
www.oenb.at
WORK IN PROGRESS…
4. Rubric
www.oenb.at oenb.info@oenb.at4
Puzzle 1: Trade grew consistently faster than GDP.
Puzzle 2: The trade-GDP ratio declined.
Ratio of global import growth to GDP growth
(in PPP)
Source: ECB staff calculations.
Motivation
5. Rubric
www.oenb.at oenb.info@oenb.at5
Puzzle 1: Trade grew consistently faster than GDP.
Puzzle 2: The trade-GDP ratio declined since the mid-90s.
Ratio of global import growth to GDP growth
(in PPP) (in USD)
Source: ECB staff calculations.
Motivation
6. Rubric
www.oenb.at oenb.info@oenb.at6
• Standard trade models failed to explain changes in trade-GDP
growth ratio
• Reasons for these changes are unknown, cyclical or
structural?
• This paper analyses the role of GVCs in this context
(= possible structural reason).
• We include indicators for the participation in GVCs in a
standard import demand equation.
• Alternatively, one could analyse differences between gross
and value added trade.
Motivation
8. Rubric
www.oenb.at oenb.info@oenb.at10
• Yi (2001):
Falling trade barriers too weak and too early to explain high trade
growth, but vertical specialisation can explain the puzzle
• Eaton, Kortum & Romalis (2011):
During the crisis, spending shifted away from durable goods
composition effect hypothesis
• Alessandria, Kaboski & Midrigan (2010); Altomonte, di Mauro,
Ottaviano, Rungi & Vicard (2012); Bems, Johnson & Yi (2012):
Disproportionally large inventories in GVC trade, higher sensitivity of
trade to foreign income shocks, bullwhip effect supply chain effect
hypothesis
• Constantinescu, Mattoo & Ruta (2015):
decline in trade-GDP ratio started long before the crisis, thus reflecting
longer term structural reasons
• Ollivaud & Schwellnus (2015):
No decline in the ratio with correct GDP measurement and treatment of
intra-EU flows, thus weak global demand is responsible
Literature
10. Rubric
www.oenb.at oenb.info@oenb.at
Import demand function …
… augmented by an index of GVC participation:
Most G20 and EU countries
Estimations for 1980-2012 and 1995-2012
2 alternative GVC integration measures:
Vertical specialisation index (only backward integration)
GVC participation (based on decomposition of gross exports)
Empirical set-up
12
𝒍𝒏 𝑴𝒊𝒋𝒕 = 𝜶𝒊𝒋 + 𝜶 𝟏 𝒍𝒏 𝑻𝑭𝑬𝒊𝒕 + 𝜶 𝟐 𝒍𝒏
𝑷𝒋𝒕
𝑷𝒊𝒕
+ 𝜶 𝟑 𝒍𝒏 𝑬𝑹𝒊𝒋𝒕 + 𝜺𝒊𝒋𝒕
𝒍𝒏 𝑴𝒊𝒋𝒕 = 𝒊𝒋 + 𝟏 𝒍𝒏 𝑻𝑭𝑬𝒊𝒕 + 𝟐 𝒍𝒏
𝑷𝒋𝒕
𝑷𝒊𝒕
+ 𝟑 𝒍𝒏 𝑬𝑹𝒊𝒋𝒕
+ 𝟒 𝒍𝒏 𝑮𝑽𝑪_𝒑𝒂𝒓𝒕𝒊𝒕 + 𝜺𝒊𝒋𝒕
11. Rubric
www.oenb.at oenb.info@oenb.at13
Data sources
IMF DOTS
Value of imports between each country and its
trading partners
IMF WEO + IFS
GDP, deflators, prices, exchanges rates
WIOD
Available across 40 countries over the period
1995-2011
CEPII – CHELEM Database
Bilateral trade values, 79 countries, 1967-2012
GVC
participation
Index à la
Koopman et
al. (2014)
Vertical
Specialisation
Index (VSI) à
la Amador and
Cabral (2009)
Trade
elasticities
12. Rubric
www.oenb.at oenb.info@oenb.at
1st GVC-integration measure: VSI
14
Vertical Specialisation Index (VSI)
2010=100
Source: update of Amador & Cabral, 2009.
0
20
40
60
80
100
120
140
160
180
200
1967 1972 1977 1982 1987 1992 1997 2002 2007 2012
China
India
Brazil
0
20
40
60
80
100
120
140
160
180
200
1967 1972 1977 1982 1987 1992 1997 2002 2007 2012
United States
Germany
United Kingdom
We observe increasing vertical specialisation, especially by EMEs
13. Rubric
www.oenb.at oenb.info@oenb.at
2nd GVC-integration measure: GVC-participation
15
GVC participation index
2010=100
Source: authors’ calculations based on WIOD.
Global Crisis induced a reversal in GVC participation intensity
𝑮𝑽𝑪 𝑷𝒂𝒓𝒕𝒊𝒄𝒊𝒑𝒂𝒕𝒊𝒐𝒏𝒊𝒕 = 𝑽𝑺𝟏𝒊𝒕 + 𝑭𝑽𝒊𝒋𝒕
21. Rubric
www.oenb.at oenb.info@oenb.at
Conclusions and way forward
23
• Countries with stronger involvement in GVCs tend to import more,
beyond the effect of the demand variable
• Omitting these variables leads to higher estimates of income elasticities
larger difference for EME and for longer sample
Confirming our hypothesis that GVCs explain part of the high trade to
GDP growth ratio
• Income elasticity higher and GVC elasticity lower in advanced than
emerging countries
• Sensitivity of imports to GVC-integration rising for EMEs, but declining for
AEs
• On the agenda: more (and proper) robustness checks, allow for time-
varying effect of GVC-participation, differentiate demand components
Ganges et al: trade collapse 2009 because GVCs have increased sensitivity to external shocks;
@ composition and the supply chain effect:
The composition effect postulates that global value chain trade is concentrated in durable goods industries, which are known to have high income elasticities. As a result, aggregate trade becomes more sensitive to foreign income shocks when the importance of GVCs in trade rises.
The supply chain effect assumes that income elasticities in global value chain trade are higher than in regular trade due to structural characteristics of GVCs such as higher inventory holdings
All variables index and log
Real imports, deflated by X-deflator of partner
Nominal TFE (WEO)
Relative prices and ER (IFS)
Producer prices of exporter divided by importers’ producer prices
No bilateral import prices available…
GVC participation (WIOD/CHELEM)
VSI from Amador & Cabral (2009): 79 countries, 1967-2012
GVC participation index a la Koopman (2010), based on WIOD data
“Vertical specialization across the world: A relative measure” - North American Journal of Economics and Finance 20 (2009) 267–280
VS: the use of imported inputs to produce goods that are afterwards exported. A relative measure of VS-based trade that combines information from Input-Output matrices and international trade data.
Multiple Threshold approach:
1st threshold: how important a particular input i is in producing output j
2nd threshold: export share of product j and import share of product i simultaneously and significantly exceed the respective world average shares -> value of intermediate imports above the threshold = relative measure of VS activities for each country/product pair.
LIMITS:
It only looks at the imports side: It accounts for backward integration but not precisely for forward (just when dealing with the second threshold).
THIS IS THE MAIN DIFFERENCE WITH GVC PARTICIPATION INDEX THAT IDEALLY ACCOUNTS FOR BOTH
It underestimates the true VS activities when the world threshold is increasing
Alternative index based on the decomposition of gross exports proposed by Koopman et al. (2014).
sums up the following two ratios that capture both, forward and backward, integration in GVCs:
𝐺𝑉𝐶 𝑃𝑎𝑟𝑡𝑖𝑐𝑖𝑝𝑎𝑡𝑖𝑜 𝑛 𝑖𝑡 =𝑉𝑆 1 𝑖𝑡 +𝐹 𝑉 𝑖𝑗𝑡
𝑉𝑆 1 𝑖𝑡 is the share of exports from country i sent indirectly through third countries (corresponds to the VS1 ratio in Hummels et al. (2001))
𝐹 𝑉 𝑖𝑗𝑡 is foreign value added in gross exports from country i to country j.
Data from WIOD database, hence limited to 1995-2011.
We do not use the bilateral version of the index but aggregate all partner countries j using a year-specific trade-weighted average. to minimize a potential endogeneity bias in the estimation; bilateral measure might entail too much endogeneity as GVC-participation is likely to boost import demand.
PMG estimator: allows intercept, short-run coefficients and error variances to differ, but long-run coefficients contraint to be equal across groups
Income elasticity 0.8-1.5 for ADV, 0.3 – 1.4 for EME
Prices and ER neg. (expected, increase in relative prices and bil. xr = depreciation of the importing country’s currency, dampens real imports)
Interaction term always positive -> GVCs increase impact of income, with impact higher in EME than ADV
Income elasticity decreases when interaction term is included -> without, it takes up some of the effect of GVCs (in EMEs: income effect not sig. when controlling for vertical specialisation)
Nobs ADV around 1,200 EME around 1,000
Country pairs ADV 84 EME 72
Income elasticity 1-1.5 for ADV, 0.4 – 1.6 for EME (higher now!)
Prices and ER still neg.
Interaction term again positive
GVC elasticity increases in more recent time period for EMEs, but decreases for AEs
!!! Here nominal imports used, therefore positive sign for prices
Still positive effect from GVC-participation
ADV and EME increasing in 2004, rather stable around 1.5 from 2004 onwards, slight decline for EME
Similar picture with GVC part interaction term for demand variable
Interact
- for ADV sign only for sample including at least 2008;
- always sign for EME and increase over 90s and decrease over 2000s
- Increasing during the crisis
Similar picture for income elasticity with interaction term
Link to Chart 1: showed that ratio was largely above 1 in early period and then declining