RoO for PTAs: Implications for AFTA of US and EU Regimes

O. Cadot, J. de Melo, A. Portugal-Perez

WB-BFA Workshop, Hainan, June 26-27

Jaime de Melo: demelo@ecopo.unige.ch

2

Outline

• EU (ACP and GSP) and US (Mexico under NAFTA) RoO → a maze of RoO

• Preferences & utilization rates (EU-US)

• AFTA utilization rates: Why so low?

• Economics of RoO (summary of evidence)

• « New » results for AFTA

• Conclusions

3

…The Maze(classification of PSRO under NAFTA and PANEURO at HS-6 level

(Also see annex table 1 for comparison across PTAs)

% of tariff lines

“No other requiremen

t”or

“EXC”

“TECH”or

“TECH+EXC”

“VC”or

“VC+EXC”

“TECH+VC”or

“TECH+VC+

EXC”

“Wholly obt'd” (WH)

Other addit.

Requir.

SUBTOTAL

NAF PAN NAF PAN NAF PAN NAF PAN NAF PAN NAF PAN NAFPAN

No CTC

0.5 0.6 0.0 2.6 0.0 13.0 0.0 0.2 0.0 8.1 0.0 7.00.5

31.5

CI 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

CS 3.8 0.1 0.4 0.0 0.1 0.4 0.0 0.0 0.0 0.1 0.0 0.4 4.4 1.0

CH 36.316.5

0.2 7.8 4.1 12.8 0.1 0.1 0.0 0.3 0.0 13.940.7

51.3

CC 48.7 0.0 5.8 7.3 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.2 54.4 7.7

Altern 8.7 0.0 8.7

Total 89.317.2

6.4 17.7 4.2 26.1 0.1 0.4 0.0 8.5 0.0 30.1 100 100

4

Preferences & utilization rates

40%

50%

60%

70%

80%

90%

100%

2% 4% 6% 8% 10% 12% 14% 16% 18% 20%

Minimum preferential margin

Ave

rag

e u

tiliz

atio

n r

ate

ACP-37(HS-8)

GSP-92(HS-8)

NAFTA (HS-6)

5

EU-US utilization rates (I)

• Certainly not low by AFTA stds (see below)• Average pref. Margins =4.5% for NAFTA, 2.4%

for GSP, 4.6% for ACP (positive tariff line only).• For EU 62% of GSP-eligible trade at zero tariffs

and 80% for ACP.

• …but T&A utilization rates (ui) are rather low.

• Paper argues that strong evidence that RoO are the culprit.

6

EU-US utilization rates (I)(pref. Margins, util rates(#tariff lines)

GSP(92)b ACP(37) b NAFTAc

τ≥4% a 50.2% (1297)d 92.5% (1627)d 87.0% (1239) d

τ≥8 % a 52.5% (91) 94.3% (892) 86.0% (558)

τ≥12% a 66.2% (44) 96.4% (566) 82.8% (287)

7

Textiles & Apparel(ACP tolerance=15%, GSP=0%)

#Observations

Utilization rate Preferential

Margin

NAFTA (2001)

618 79.9 10.4

GSP (2004)16555(HS-8)12920 (HS-

6)52.2 1.8

ACP (2004)1370 (HS-8)1175(HS-6)

50 10.4

8

AFTA RoO: Why so low (I)

• EU has over 500 PSRO….and is considering moving to a value content (VC) only criterion to meet the sufficient processing requirement

• AFTA currently has a VC threshold = VNOM<=60% FOB price with close to full cumulation (last stage of processing has to be undertaken in country)

• …very low util. Rates 5%-10% for highest user (Philippines) in late 90s

9

AFTA RoO: Why so low (II)

• Intra regional trade mostly in intermediates « factory Asia » as Japan and Korea successively outsource labor-intensive stages

• Intra-regional trade in computer/machinery (HS-84) and electrical equipment (HS-85) where preferential market access is low

• …with trade in final products mostly to RoW & given that tariff on intms. are always lower

• → preferential market access irrelevant

10

Figure 2: Growing Vertical integration in the World Economy

Import content of exports (%)

0

5

10

15

20

25

30

35

40

45

50

E. Asia & Pacific Europe & C. Asia

Latin America & Caribbean

Middle East & N. Africa

S. Asia Sub-Saharan Afria

Indirect

Direct

Sources : WB (2005), GEP, figure

11

AFTA RoO: Why so low (III)

• Paper argues that with new regionalism in AFTA preferential market access will become more relevant

• Protectionist pressures will most likely be felt in RoO set-up (the security valve par excellence outside the scope of WTS)

• VA=10%;VNOM=40% in mfg.►VO=64% • VA=10%;VNOM=60% in mfg.►VO=46% • …so RoO could be binding…

12

Economics of RoO: basics (I)

• Details on set up in the paper

Unit costs= Undistorted+ distorted+ Adminis+rent sharing

Total Unit costs = Undistorted + compliance costs

0 0D A Ci i i i i i ic c c c c c

13

Economics of RoO: Quantification?

• Huge heterogeneity (more at HS-6 level?)

• Limited data (=utilization rates at HS-6 level +Yes or No for each PSRO + variation in VC where applicable (often VC is not the only criterion when it is used…)

• …crazy to try and quantify???

• Yes, but here is a summary…

14

Economics of RoO: Evidence (I)

• R-index at HS-6 level(1<Ri<7). Ordinal index based on simple obs. rule (like an ERP index)

• Non-parametric (based on Ri)

• Econometric (reduced-form estimates)

• Simulations (based on models with well-accepted behavior---see the paper)

15

Economics of RoO: Evidence (II)Ri=1 (not restrictive)=7(very restrictive)

Tariff peaks =3 times average tariff rateNumber of tariff lines (in parenthesis) 

R-Index valuea,b

NAFTA PANEURO

Tariff peaksc 6.2 (257) 5.18(780)

Low tariffsd 4.8 (1432) 3.92(3241)

Total number of tariff lines 3555 4961

16

Economics of RoO: Evidence (III)

• Non-parametric based on revealed preference and R-index values (see discussion in text & table 4 –perhaps 3% for administrative for admin. Costs).

• …but difficult to be confident about these results since preferential status often asked when preferential margin is zero (i.e. zero tariffs)

17

Economics of RoO: Evidence (IV)

• Strong and consistent results across NAFTA (Mexican exporters to US); EU (GSP & ACP).

• Utilization rates lower when multiple criteria (after controlling for preferential rates); technical requirements associated with high compliance costs…which are easily manipulated by special interests (see discussion in paper on evidence about market power in T&A and pass-trough under NAFTA)

18

New Results for AFTA(EU data;RVC is only criterion)RVC=min.origin. value; tariff lines;util rate; preference margin)

RVC rateb Frequency Percent ui (GSP)d ui (ACP)d

50 4,000 20.55 38% (2.6) 42% (2.6)

60 14,185 72.89 16% (1.1) 18% (2.7)

70 213 1.09 53% (2.3) 29% (3.0)

80 1,012 5.20 50% (2.9) 93% (4.3)

85 52 0.27 44% (3.6) 0% (3.7)

Total 19,642 100 N.A. N.A.

τi ≥ 5% c 36% 34%

τi ≥ 10% c 34% 58%

19

New Results for AFTA (II)(see annex table 2 for details)

(effects on util. Rates –percentage points)

GSP ACP

Rvc1=rvc0-10

(58% to 48%)

2.5 (17%). 8.2 (13%)

Simulations

assumed values

rvc= 40%

(10%) [4.5%]

rvc= 40%

(10%) [4.5%]

Predicted utilization rates

(37%) [23%] (45%) [40%]

20

Conclusions• Scientific approach to RoO may work (has worked in the past on

trade incentives in the Region)…but formidable task as recognized by Mr. Qiang

• EU & US. « Business owned » rather than «business friendly» ?

• Uniform VC content by-passes PRSO thereby avoiding capture by special interests….but not easy to implement for small firms in Low-income partners

• Keep it simple and avoid opening the Pandora’s box of technical requirements, exceptions and other PSRO that are likely to be captured by special interests in the high income partners (but Asia could be an exception… again and muster the collective action necessary to avoid the US-EU outcome?)

• …but best strategy of all is to make RoO irrelevant by continuing to reduce tariff barriers while avoiding tariff peaks (extending preferences to many trading partners lowers the stakes as value of preferences is reduced)

21

ANNEX

1.Description of RoO under PTAs2. Econometric Estimation

3. Results for AFTA

22

Table A1: Rules of Origin in a Sample Free Trade and Preferential Trade Agreements

CTC

Value AddedSpecific

Manufacturing Process

Cumulation ToleranceAbsor-ptionDomestic or

Import Content

Implied Import Content

A. Agreements Involving the EU

EUPanEu

roYes (4,2)

Yes - Import (50-30%)

50-30% YesBilateralDiagonal

Yes 10%b Yes

EU GSPYes (4,2)

Yes - Import (50-30%)

50-30% YesBilateralDiagonala Yes 10%b Yes

EUCoton

ouYes (4,2)

Yes - Import (50-30%)

50-30% Yes Full Yes 15%b Yes

EU - ChileYes (4,2)

Yes - Import (50-30%)

50-36% Yes Bilateral Yes 10% Yes

EU - MexicoYes (4,2)

Yes - Import (50-30%)

50-30% Yes Bilateral Yes 10% Yes

EU – South Africa

Yes (4,2)

Yes - Import (50-30%)

50-30% Yes

BilateralDiagonal

ACP)Full (SACU)

Yes 15% Yes

23

Econometric Approach

• Remember huge heterogeneity in data…

Hold your breadth and estimate

'( ); (.) 0i i iu f c f

' (.) 0i ic g RoO g

i i ik ik ku RoO

24

Row 1 (GSP+ACP) 2 (GSP) 3 (ACP) 4 (GSP+ACP) 5 (GSP) 6 (ACP)

Dependent varb. iu iu iu iu iu iu

Constant 1 1.58 1.7 14.88 0.43 0.97 10.61 (4.52)*** (4.99)*** (10.68)*** -1.01 (2.36)** (6.38)***

i (tar. Pref.)(GSP) 2

0.077 0.066 0.05 0.026

(12.74)*** (10.89)*** (5.35)*** (2.80)***

ln ivc 3

-0.632 -0.59 -4.069 -0.309 -0.351 -2.936

(7.47)*** (7.14)*** (11.66)*** (3.06)*** (3.56)*** (6.96)***

i (tar. Pref.)(ACP) 4

0.043 0.03

(10.00)*** (6.06)*** Number of Obser. 5 19261 13448 5857 5958 4305 1697 Likelihood Ratio a 6 263.43(0) 198.83(0) 222.3(0) 47.48(0) 25.5(0) 71.68(0)

Sample restriction

7 2%i 2%i 2%i 5%i 5%i 5%i

1 0 10i i irvc rvc u

8 2.0% 2.5% 8.2% 5.2% 1.7% 7.9

Mean margin 9 3.74 3.79 5.09 5.14 5.22 9.78 Mean Utilization Rate 10 0.12 0.17 0.13 0.17 0.22 0.20 Mean value content 11 58.8 58.9 58.6 58.2 58.6 57.5

Simulated utilization rates

Assumed value for irvc 12

40 40 40 40 40 40

Mean MFN tariff (China) 13 10 10 10 10 10 10 Predicted utlilization rate 14 0.32 0.37 0.45 0.25 0.28 0.40

Mean MFN tariff (Philippines) 15 4.5 4.5 4.5 4.5 4.5 4.5

Predicted utlilization rate 16 0.19 0.23 0.40 0.18 0.24 0.36