Transcript Aucun titre de diapositive - Organization of American States

ECONOMIC EFFECTS (I)
SESSION II:
THE ECONOMICS OF PTAs: I
(traditional effects)
JAIME DE MELO
This version, March 2004
1
ECONOMIC EFFECTS (I)
OUTLINE: SESSION II
PART I: THEORY
 Viner’s analysis: trade creation and Trade diversion
 What do we know about the welfare effects of RTAs?
(effects related to discriminatory trade preferences)
PART II: EMPIRICS
 Detecting (ex-post) trade diversion and trade creation
 Sophisticated anti-monde: the gravity model
 Effects on third countries
 Annex: Tools for trade policy analysis (you must master
these to follow discussion in this and next session)
2
ECONOMIC EFFECTS (I)
PART I: THEORY
3
ECONOMIC EFFECTS (I)
TRADE CREATION AND TRADE DIVERSION
(from overview)
3 countries: A (importer and country of interest here); B
(partner in RTA) and C (ROW). Here B and C both exporters)
Unit production costs assumed to be constant
Initially A applies identical tariffs on both imports (t=20%)
Let A and B form an FTA = no more tariffs on B imports
Welfare implications for A depend on cost conditions:
If the lowest cost producer is:
B: all is well as partner is low cost producer = trade
creation (as partner is low cost supplier initially)
C:  If C imports are now replaced by (less efficient) B
imports = trade diversion (partner is high cost supplier )
A: nothing happens! (A is not even importing initially).
4
ECONOMIC EFFECTS (I)
NUMERICAL EXAMPLE
(from overview)
Shoes
c
c (1+t0)
c (1+t1)
B
11
13.2
11 
A
13
13 
13
C
10
12
12
Textiles
c
c (1+t0)
c (1+t1)
18
21.6
18
15
15
15
20
24
24
neither Trade
Creation nor
Trade
Diversion
Computers
c
c (1+t0)
c (1+t1)
15
18
15 
17
17
17
16
23.2
23.2
Trade
Creation
Trade
Diversion
tB0 = tC0 = 20%; tB1 = 0; tC1 = 20%
5
ECONOMIC EFFECTS (I)
WELFARE ANALYSIS (for country A)
PTA: leads to effects that are both positive (reduction
of protection) and negative (discrimination)
Anything can happen = PTA is not necessarily welfare
increasing.
What does welfare effects depend on intuitively?
- Elasticities (in A, B, and C)
- Extent of tariff reduction in A (we assume B and C
don’t change their tariff)
- Extent of price reduction in A (for given tariff
reduction by A)
- cost differences between B and C
- amount of imports affected
6
ECONOMIC EFFECTS (I)
GRAPHICAL REPRESENTATION: CONSTANT COST CASE
FOR PARTNER ILLUSTRATED IN NUMERICAL EXAMPLE
P
ESB+t
1
5
2
6
ESC+t
7
ESB
6
4
3
8
ESC
EDA
MT
MFTA
RIA  gain in private welfare: +1275
loss in tariff revenue: - 1234
Imports of A
net change:
?
W=276 - 5634 > 0
N.B: What happens if ESB>ESC+t and what situation is it likely to represent?
7
ECONOMIC EFFECTS (I)
DETERMINANTS OF WELFARE CHANGE (I)
 RIA is likely to be welfare-enhancing for the
importing country if, all things equal:
•there is a small cost difference between B and C
•the import demand is elastic
•the initial tariff is high
•the initial amount of imports is small
Is the above configuration of welfare-enhancing
conditions for RIAs to be welfare enhancing likely to
be met South-South integration?
8
ECONOMIC EFFECTS (I)
DETERMINANTS OF WELFARE CHANGE (II)
Suppose that partner B is too small to satisfy
domestic demand by A
WA=(PS+ PC)+ GR
= (0) +(-1234)
WB= PS= +1246
 WB +WA=
-(236)
WC = 0
9
ECONOMIC EFFECTS (I)
In the absence of “sudden death” of imports from C,
then A necessarily loses (pure tariff revenue loss)
mercantilist flavor as gains go to the country that gets
market access!
Is it likely that this situation will prevail in South-South
integration?
What are implications of S-S integration among partners
of unequal size?
What should the small (usually poorer) partner do if it
cannot get compensation? Any implications for REPA
implementation?
10
ECONOMIC EFFECTS (I)
 What do we know about the welfare effects of RTAs? (I)
in theory one can design a welfare augmenting PTA (but
this is largely irrelevant from a practical point of view!!!)
A eliminates a quota on
imports from B, but keeps
same quota level on imports
from C ( = TOT with C
unchanged)
WA = PS + CS + GR
=(-3456) + (089) + 0
WB = PS = 4375
WB + WA = (376)+(089)
A welfare increasing FTA
WC = 0 (by construction)
11
ECONOMIC EFFECTS (I)
 What do we know about the welfare effects of RTAs? (II)
Economies of scale are not a “reason for PTAs (even if
welfare gains will be larger)
Best policy for A and B
0
Free trade
CSA=CSB=035
P
(1+1)
1
(1+2)
6
Pw = 1
3
 Minimum protection for
monopolists in A & B
(1+1)  CSA=CSB=(245)
 FTA with production in A
2
7
AC
4
8
5
DA
9
ESC = MC
(1+2)  CSA=CSB=-789
DA+B
… but free trade still superior
FTA with increasing returns to scale
12
ECONOMIC EFFECTS (I)
 What do we know about the welfare effects of RTAs? (III)
Staying out will be costly (the cold shower effect)
• Prior to RIA, firm faces
D0(P; P0R)
0 = 1 2 4 3
AC
MC
• After RIA, firm in
country that has not
joined faces D1(P;P1R)
1 = 5 6 7 8
Profit loss for non-member firm
13
ECONOMIC EFFECTS (I)
 What do we know about the welfare effects of RTAs? (IV)
In theory a PTA is preferable to an FTA
Suppose that three countries A,B,C produce three
differentiated goods (a,b,c). Good b and a are substitutes
in consumption in A
 A lowers progressively its tariff on B. For a small tariff
reduction, the welfare gain from reducing that distortion is
about the same as the welfare loss from the reduction in
imports from C (with less consumption of good a in A .
 By the budget constraint, exports of A will rise, so there
is a net expansion of imports at world prices. And there is
a welfare gain since the increase in imports from A
exceeds the decrease in imports from C.
 As tariff reduction continues height of TC gain on goods
from B falls while rectangle associated with TD on C
remains constant. So eventually a net loss as tariff !!!
14
ECONOMIC EFFECTS (I)
PART II: EMPIRICS
15
ECONOMIC EFFECTS (I)
 Are countries complementary?
Presumption is that countries with different endowments
will be complementary and hence import different goods
so that an RIA will have an effect
Michaely index (the higher the value, the more similar the
economies and potential for trade):
Cij 100  (ij M ik  E jk )2
Ejk = share of good k in export of j
Mik = share of good k in imports of i
Value of index for EU and NAFTA 12 times larger than for SSA and 3
times larger than for MERCOSUR
16
ECONOMIC EFFECTS (I)
 Trade shares in apparent consumption
Apparent consumption are total sales on the domestic
market where D=X- E
Compute partners (P) and non-partners (N) shares in
apparent consumption SP and SN
SP =(MP ) / (MP + MN+ D ) ; SN =(MN ) / (MP + MN+ D )
Simplistic presumption is that if
SP > SN  TC
And
SP < SN  TD
Following table shows trade effects 1 year before and 5
years after RIA
17
ECONOMIC EFFECTS (I)
Openness indices import shares increases TC, and if share
of non-partner decreases TD (see formula on previous slide
and figure 2.1, 2.2,2.3 from S-W). Shares increased
suggesting strong TC. But would have happened anyway
because of growth?
Trade intensity indices (TIA) is the trade intensity of A
Suppose A and B form an FTA. A’s imports will be biased
toward B if A’s trade share from B is greater than the ROW
trade share from B
TI indices correct for growth effects since fast growing
countries will tend to absorb more imports. In most cases
in table 2A1 below, Intra-bloc TI increased while extrabloc TI decreased
Trade propensity indices = (TI) (M/GDP). Compares B’s
share in A’s imports with B’s share in world trade
18
ECONOMIC EFFECTS (I)
Appendix Table 2.1: The Trade Effects of RIAs between Developing Countries
‘Before’: one year preceding; ‘after’: five years following implementation of internal preferences
Import
Shares %
Trade
Intensities
Import/GDP Ratios
Intra
bloc
Extrabloc
Intrabloc
Extrabloc
Intra
bloc
Extrabloc
Total
Impor
ts
Intrabloc
Extrabloc
1990
14.5
85.5
10.52
0.87
0.57
3.34
3.91
0.41
0.03
1996
20.2
79.8
14.12
0.81
1.64
6.48
8.10
1.15
0.07
Andean Pact I 1968
4.3
95.7
2.90
0.97
0.54
12.15
12.69
0.37
0.12
1974
7.2
92.8
6.16
0.94
1.01
13.14
14.15
0.87
0.13
1990
9.1
90.9
81.19
0.91
2.63
26.22
28.85
23.42
0.26
1996
12.6
87.4
64.33
0.88
3.78
26.29
30.07
19.34
0.26
93.2
6.84
0.94
0.86
11.77
12.63
0.86
0.12
13.6
86.4
15.55
0.87
2.01
12.80
14.81
2.30
0.13
1972
5.0
95.0
13.59
0.95
2.76
52.05
54.81
7.45
0.52
1978
3.8
96.2
7.66
0.97
3.10
78.01
81.11
6.21
0.78
1965
3.3
96.7
11.68
0.97
0.48
14.39
14.87
1.74
0.14
1971
4.5
95.5
20.78
0.96
0.91
19.28
20.19
4.20
0.19
1965
1.4
98.6
8.92
0.99
0.24
17.59
17.83
1.59
0.18
1971
4.0
96.0
30.09
0.96
0.79
18.82
19.61
5.90
0.19
Mercosur
CACM II
Andean Pact II 1990 6.8
1996
CARICOM
CEAO
UDEAC
%
Trade
Propensities
19
ECONOMIC EFFECTS (I)
25
20
15
10
5
0
Before
C
C
After
G
M
er
c
os
An ur
de
an
I
C
AC
M
An
I
de I
an
C
AR II
IC
O
M
C
EA
O
U
D
EA
C
AF
TA
percent
Figure 2.1 Share of Imports from partners
** For Mercosur this encompasses two years 1991 and 1996; Andean Pact I
1968 and 1974; Andean Pact II 1990 and 1996; CACM II 1990 and 1996;
CARICOM 1972 and 1978; CEAO 1965 and 1971; AFTA 1991 and 1996;
and the GCC 1980 and 1986.
20
ECONOMIC EFFECTS (I)
21
ECONOMIC EFFECTS (I)
22
ECONOMIC EFFECTS (I)
Figure 2.4
Intra-bloc import propensities
100
10
1
C
C
G
A
AF
T
C
U
D
EA
O
C
EA
M
er
co
su
r
An
de
an
I
C
AC
M
II
An
de
an
II
C
AR
IC
O
M
0.1
Strong increase in MERCOSUR and UDEAC
23
ECONOMIC EFFECTS (I)
Figure 2.5 Extra-bloc import propensities
1
0.1
Before
After
M
er
c
os
An ur
de
an
CA I
CM
An
I
de I
a
CA n II
RI
CO
M
CE
AO
UD
EA
C
AF
TA
G
CC
0.01
Also an increase but beware the scale is different so that extra-bloc
import propensities have been increasing less than intra-bloc. So the
countries in the RIAs, have been doing something right. But is it the RIA
itself, other policies, etc…? Need a mor sophistiacated anti-monde (or
counterfactual).
24
ECONOMIC EFFECTS (I)
Gravity trade model:
« Accounts » for 70% of variation in bilateral trade intensity
across countries so best available model to build
counterfactual.
Works particularly well for developed countries (data is
better).
Formulation:
Tij = Imports or average trade between i and j
Dij average distance (or better) average
transport costs between countries
25
ECONOMIC EFFECTS (I)
Gravity trade model formulation: Here we have IN= index of
quality of infrastructure: R= remoteness index; E= landlock
index; L= common language
In estimates of (8) by Carrère all coefficients have expected sign,
plausible values.
26
ECONOMIC EFFECTS (I)
Add to the above model the following dummy variables
Following figures only report statistically significant coefficients for ’s
27
ECONOMIC EFFECTS (I)
Fig. 2 Evolution of dummies for MERCOSUR
28
ECONOMIC EFFECTS (I)
Conclusions from Carrère study:
…basically a difficult business to build the counterfactual, and need
good techniques (here panel estimates) to get plausible results
29
ECONOMIC EFFECTS (I)
ANNEX:
TOOLS FOR TRADE POLICY
ANALYSIS
30
ECONOMIC EFFECTS (I)
OPTIMALITY PROPERTIES OF FREE TRADE
Definition: A situation is Pareto optimal (PO), if no change
can improve the welfare of some agent without reducing the
welfare of at least one other agent
Implication: PO implies that marginal benefits and costs are
equal across all agents (in case of externalities, PO occurs
when social marginal benefits and social marginal costs are
equal across all agents )
Notation:
MBp (MCP) = private marginal benefit (cost)
MBs (MCS) = social marginal benefit (cost)
31
ECONOMIC EFFECTS (I)
Closed economy
NO MARKET FAILURE ON DOMESTIC MARKETS
For now we consider only the case of no externalities (then
private benefit and costs coincide with social benefits and
costs)
MBP = MBS ; MCP = MCS
PO in closed economy  MB = MC
32
ECONOMIC EFFECTS (I)
Checking PO condition in closed economy
Wheat market under autarky
P
MB>MC: one
more unit
leads to
more
benefits than
costs
MC
MB
MB<MC: one
unit less saves
more costs
than the loss
in benefits
Q
optimality condition is:
MC = MB
33
ECONOMIC EFFECTS (I)
Checking PO in closed economy with PS and CS
Show that sum of producer surplus (PS) and consumer
surplus (CS) is maximized when PO holds
 Compute total consumption gain (CG) less total
production costs (PC) for Q*, Q0 and Q1
PS + CS = CG - PC
1
P
2
MC
6
9
7
3
5
0
4
Q0
8
Q*
Q1
MB
Q
 Q=Q0
CG = area 0124
PC = area 0534
PS + CS = area 1235
 Q=Q1
CG = area 0178
PC = area 0568
PS + CS = area (195-967)
 Q=Q*
CG = area 019Q*
PC = area 059Q*
PS + CS = area 195
34
ECONOMIC EFFECTS (I)
Open economy
 Now take into account:
- either the marginal cost through trade, MCT =for imports
- or the marginal benefit through trade, MBT=for exports
 for a SMALL OPEN ECONOMY (SOE), these marginal values
through trade are exogenous, given by the price imports(p* )
m
*
or exports ( p ).
x
* (importing SOE)
PO : either MB = MC = MCT = pm
or MB = MC = MBT = p*x (exporting SOE)
Free trade is PO for SOE
35
ECONOMIC EFFECTS (I)
Pareto optimality of free trade for SOE
S
IMPORTS
p*
m
MBT
ES*
MC
MB
D
S
*
px
MBT
MB
ED*
MC
EXPORTS
D
36
ECONOMIC EFFECTS (I)
IMPORT TARIFF
pd
Consumers lose: 1349
S
Producers gain: 1289
Gov ’t gains: 2356
p*+t 1
9
p*
2
ES*+t
3
4
8
6
5
QF
QT
CT
ES*
D
Efficiency loss in
production: 286
Consumption: 345
CF
Tariff at rate t:
1) production & gov ’t rev.; consumption & imports
2)Efficiency loss or dead weight loss (DWL) is 286 + 345
3) price measures (but also NTBs) redistribute income
Exercise 1: Analyze an export tax
37
ECONOMIC EFFECTS (I)
EXERCISE 2: COMPARE
EXPORT TAX & EXPORT SUBSIDY
pd
pd
ES
p*+s
1 2
p*-t 5
4
p*
ET
3
EF
export tax
ED*
ED*-t
p*
ES
2
1
5
ED*+s
4
3
X
EF
ED*
ES
X
export subsidy
Q1) Fill the blanks below
GR =
GR =
(PS+CS)=
(PS+CS)=
W=
W=
Q2) (more difficult) Should a large country subsidize its
exports?
38
ECONOMIC EFFECTS (I)
Open economy
 Large exporter has monopoly power
Exports   world price, pW 
pW   MR = MBT < pw
PO  MB = MC= MBT as before
PO  MB=MC = MR < pw
Free trade is not PO for country with
monopoly power!
39
ECONOMIC EFFECTS (I)
MONOPOLY POWER IN EXPORTS (1)
Vietnam (rice); Mongolia (cashmere)
Small export tax (t0):
ES
t1
t0
a
b
d
c
ED*
• increases H welfare as
revenue gain (a+b)
larger than private
losses (b+c)
• decreases F welfare
(a+d)
As wedge becomes larger (t0t1) revenue gain gets smaller
while private loss becomes larger
 scope for an “optimum tax” at intermediate range.
 But F could retaliate  no longer “optimal”
40
ECONOMIC EFFECTS (I)
MONOPOLY POWER IN EXPORTS (2)
pd
ES
p1w
1
p0w
5
1
pw -t
6
2
4
8
A large exporting
country can improve
its welfare by
restricting its exports
provided it faces no
retaliation.
3
7
MR
ED*-t
ED*
exports
With tax t, GR = 1276; (PS+CS) = (-5376):
W = (1245)+ (-437) which can be >0 or <0.
But:
1)
H and F together lose 237
2)
if H chooses t so that ED*-t crosses ES at intersection with MR
curve, then H gains ( this is “optimum tariff”: Can you see why?)
41
ECONOMIC EFFECTS (I)
MEETING A REVENUE CONSTRAINT
pd
S
A
B
F
p*
ES*+tm
G
J
I
K
C
E
ES*+tc
ES*
H
D
QF
QT
CT
CC
CF
To raise revenue, gov’t can either rely on :
a) A consumption tax, tc (area FGHI= area ABCE)
b) A tariff on imports, tm (area ABCE= area FGHI)
DWL under a) W= area AKC + area BEJ
DWL under b) W= area GHJ
42
ECONOMIC EFFECTS (I)
DETERMINANTS OF THE COST OF PROTECTION
p*(1+t)
p*
4
6
7 5
ES*(1+t)
2
1
ES*
3
The less elastic,
the demand for
imports, the more
efficient is the tax
EDA
EDB
MA=MB
(a)
Compare the effects of tariff at rate t on A and B
1)
2)
B: inelastic demand  large revenue, small DWL
A: elastic demand  small revenue, large DWL
Efficiency of tax collection: REVENUE/DWL
43
ECONOMIC EFFECTS (I)
DETERMINANTS OF THE COST OF PROTECTION
p*(1+2t)
p*(1+t)
ES*(1+2t)
4
6
2
ES*(1+t)
1
p*
5
3
ES*
Uniform taxation
that gives same
average level of
protection is less
costly (by area
2364)
MA=MB
(b)
Now both importers have the same elasticity of demand:
a)
b)
If both are taxed at tax t, loss W = 2 times (-123)
If only one sector is taxed loss W = (-154)
44
ECONOMIC EFFECTS (I)
ASSESSING A COUNTRY ’S TRADE REGIME
- Terms of Trade
- Nominal and effective rates of protection
Definition: Terms of Trade (TOT)
1)
The (barter) TOT is the ratio of an index of export prices
to import prices measured at border prices:
TOT = PX/PM = M/E
2)
The income terms of trade (which is equal to the barter
TOT times a quantity index of exports (QE)) is a measure
of the purchasing power (in terms of imports) of a
country ’s exports:
Income TOT = TOT x QE
45
ECONOMIC EFFECTS (I)

The profitability of import-substituting activities and of
exporting activities (other than natural-resource-based
activities that are virtually always profitable) depends on
their relative prices.
Consider an export tax, tX and an import tariff, tM

Relative profitability of each group of activities will be
the same if there is either:
a) An export tax of, say, 20%
b) Or an import tariff at the same 20% rate

So tariff is equivalent to an export tax at same rate.
46
ECONOMIC EFFECTS (I)
NOMINAL RATE OF PROTECTION (NRP)
EFFECTIVE RATE OF PROTECTION (ERP)
The NRP is the percentage difference between the price of X (pX)
in domestic currency and the price of X at the border (p*X )
Note that:
The official (or formal) tax, tx, is not a good measure of the
protection afforded to a sector, as:
1)
2)
3)
there are intermediate inputs (see next slide)
when rate is high, collected rate is less because
exemptions are high
when rate is high, there is smuggling
47
ECONOMIC EFFECTS (I)
THE EFFECTIVE RATE OF PROTECTION (ERP)
Definition: Percentage excess of domestic value-added in
activity X (VAX) over corresponding value-added at world
prices (VA*X)
ERPX = (VAX- VA*X)/ VA*X
If effects of ALL trade measures on domestic prices are
properly measured, then ERP is an indicator of the NET
incentives to produce activity X
ERPX >0  activity is protected
ERPX <0  activity is taxed
48
ECONOMIC EFFECTS (I)
AN EXAMPLE: CLOTHING
ERP in clothing (shirts). Textiles serves as an input
tt (tc) :ad-valorem tariff on textiles (clothing)
VAD (VAw) = Value-added at domestic (world) prices
Free trade  VAD = VAW .
Numerical example. Under free trade:
Sales value= 100;Textile input purchase= 40
Value-added (VA) in shirt-making : 100 - 40=60
Case I: tc= 20%; tt = 10%  VAD = 100(1.2)-40(1.1)=76
 ERP =VAD-VAW = (76-60)/60 = 27%
 Case II: tc= 10%; tt = 30%  VAD = 100(1.1)-40(1.3)=58
 ERP = VAD-VAW = (58-60)/60 = -3%
49
ECONOMIC EFFECTS (I)
Net incentives and home-market bias estimates
Compute for each activity, i
Effective rates of protection for export sales (e) and for
domestic-market (d) sales  (ERPei) and (ERPdi)
 Effective rates of subsidies to take into account effects of
fiscal policies  (ERSei) and (ERSdi)
Bias: B>1  home-market bias
Bti = (1+ERPei)/(1+ERPdi)  tariffs
Bsi = (1+ERSei)/(1+ERSdi)  subsidies
circa 1980
Argentina
Korea
Taiwan (China)
Bti (all sectors)
1.94
1.10
1.10
Bsi (all sectors)
1.86
1.01
0.89
50