Does WTO Accession Raise Income?

1

Does WTO Accession Raise Income?

May 22, 2006

M.K. Tang Shang-Jin Wei

Personal views, not necessarily shared by the IMF

2

Motivation IAccessions often involve extensive and

sometime politically difficult policy changes (in recent years)

It took China 14 years to join

The median time gap b/n application and accession was 71 months for countries that joined after 1995

3

Motivation IIDiverse Views on the Consequences Beneficial (if sometimes bitter)?

“WTO accession provides a predictable business environment and gives a powerful guarantee to investors that there will be no policy reversals.” An advisor to the Ethiopian government

Irrelevant? Andrew Rose (2004): “Do we really know that the WTO has

promoted trade?” American Economic Review

Worse than irrelevant? Developing countries squeezed for the benefits of rich and

powerful nations, Jawara and Kwa (2003) Behind the Scenes at the WTO: the Real World of International Trade Negotiations

4

Regime change for developing country applicants

Before the Uruguay Round: Very little policy change needed

After the Uruguay Round: Much more extensive policy changes typically demanded by existing members (except for countries qualified for GATT Article XXVI 5(c))

5

Value of accession Market access in exchange for political support from

reform-opposition Commitment against future policy reversal

Previous work: on trade volume Subramanian and Wei, 2004

This paper: Effects on growth and investment

6

Methodology:

Difference-in-differences in spirit Change in growth before/after 43 accessions Change in growth for non-accession developing

countries during the same period

Results: Higher annual growth and Inv/GDP than before,

especially for those undergoing more rigorous accession procedures

7

Addressing Endogeneity Concerns

Accessions with and without reforms Application vs. actual accession Formal statistical test using Heckman

selection equation Varying effects along commitment/

governance dimensions

8

Roadmap

1. Brief Description of Data and Specification

2. Statistical Results

3. Extensions and Robustness Checks

4. Conclusions

9

Data and Specifications

Data Per capita GDP, gross fixed capital

formation, trade from WEO Institution quality from WB’s “Governance

Matters” 1981-2003, exclude all industrial and

OPEC countries

10

Empirical Specifications

i: country, t: year s: “Time-profile” of accession, #years from

accession {null, -2, -1, 0, 1, 2, 3, 4, 5, beyond} Country and year fixed effects included Treatment group: accession countries Control group: other developing countries Beginning period: 8 years before accession

,*)log( ,,01,, ytitis ststi capitaperGDPGrowth

.)/log( ,,,, ytitis sstiGDPInv

11

Accessions: 1990-2001

Results for All Accessions Growth sped up from 1 year before

accession to 3 years after Little effect on Inv/GDP though

Positive coef. est. but insignificant

12

Coef est. t-stat. Coef est. t-stat.

Lagged log(GDP per capita) -0.115 -4.74

s = -2 0.009 0.91 0.045 0.64

-1 0.024 2.57 0.043 0.69

0 0.019 2.58 0.024 0.38

1 0.026 3.72 0.050 0.83

2 0.033 4.91 0.073 1.17

3 0.015 1.92 0.105 1.46

4 -0.002 -0.18 0.053 0.70

5 0.012 1.89 -0.011 -0.16

beyond -0.002 -0.37 -0.078 -1.09

Country fixed effects

Year fixed effects

2

Table 3. Changes in Growth and Investment Around Accessions

Annual Growth Rate Log(Pri Inv/GDP)

Y

Y

Y

Y

1

13

-2 -1 0 1 2 3 4 5-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

Years Since Accession

Ch

an

ge

in G

row

th in

Pe

rce

nta

ge

Po

ints

Figure 1. Change in Growth Around Accession

90% C. I.

14

-2 -1 0 1 2 3 4 5-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25


Pe

rce

nta

ge

Ch

an

ge

in I

nv

/GD

P R

atio

Figure 2. Change in Inv/GDP Around Accession

90% C. I.

15

Article XXVI 5(c) of GATT 1947

“If any of the customs territories, in respect of which a contracting party has accepted this Agreement, possesses or acquires full autonomy in the conduct of its external commercial relations and of the other matters provided for in this Agreement, such territory shall, upon sponsorship through a declaration by the responsible contracting party establishing the above-mentioned fact, be deemed to be a contracting party.”

16

Article XXVI5c vs. Non-Article XXVI5c Co’s Heterogeneity in accession procedures Article XXVI 5(c) – 18 countries

Upon becoming independent, former colonies of GATT members were entitled to formally join GATT almost automatically

The option can be exercised any time (up to the founding of the WTO)

Non-Article XXVI5(c) – 25 countries Acceded by normal procedures (Article XII),

meaning lengthy negotiations!

17

Results for AXXVI5c vs. non-AXXVI5c Faster growth after accession than before:

72% non-Article XXVI5c

vs.

47% Article XXVI5c Higher Inv/GDP after accession than before:

59% non-Article XXVI5c

vs.

38% Article XXVI5c

18

Non-Article XXVI 5c countries Growth became stronger since 1 year before

accession Inv/GDP became higher since 2 years after

accession Article XXVI 5c countries

Growth: Weakly positive, insignificant effects on change in growth

Inv/GDP: Negative, insignificant effects on change in Inv/GDP

19


Lagged log(GDP per capita) -0.113 -4.66

s = -2 0.007 0.53 0.012 0.17

-1 0.027 3.29 0.056 0.95

0 0.024 2.76 0.088 1.51

1 0.043 4.84 0.098 1.48

2 0.047 5.52 0.135 1.96

3 0.038 3.47 0.177 2.27

4 0.027 3.18 0.159 2.44

5 0.020 2.48 0.110 1.54

beyond 0.020 3.42 0.148 2.70

s*AXXVI5c Dummy: -2 * AXXVI5c 0.006 0.31 0.083 0.60

-1 * AXXVI5c -0.007 -0.30 -0.029 -0.22

-0 * AXXVI5c -0.013 -0.88 -0.152 -1.09

1 * AXXVI5c -0.045 -3.33 -0.106 -0.86

2 * AXXVI5c -0.039 -3.01 -0.137 -1.07

3 * AXXVI5c -0.054 -3.98 -0.149 -1.03

4 * AXXVI5c -0.063 -3.30 -0.200 -1.44

5 * AXXVI5c -0.019 -1.67 -0.225 -1.81

beyond * AXXVI5c -0.041 -4.61 -0.401 -3.40


Year fixed effects

Article XXVI 5(c) vs. Non-Article XXVI 5(c) Countries

Y

Y

Y

Y

Table 5. Changes in Growth, Investment and Trade:


1 2

20

-2 -1 0 1 2 3 4 5-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08


Ch

an

ge

in G

row

th in

Pe

rce

nta

ge

Po

ints

Figure 3. Change in Growth: Article XXVI 5(c) vs. Other Countries

Article XXVI 5(c) Countries

Non-Article XXVI 5(c) Countries

90% C.I.

90% C.I.

21

-2 -1 0 1 2 3 4 5-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4


Pe

rce

nta

ge

Ch

an

ge

in I

nv

/GD

P R

atio

Figure 4. Change in Inv/GDP: Article XXVI 5(c) vs. Other Countries

90% C.I.

90% C.I.

Non-Article XXVI 5(c) Countries

Article XXVI 5(c) Countries

22

Extensions and Robustness ChecksTiming of Spurs Median interval b/w application and accession:

71 months after 1995, or 55 months since 1990

Pickups in Growth and Investment after application die down gradually

But Growth and Investment rise again around accession

23

Coef est. t-stat. Coef est. t-stat.Lagged log(GDP per capita) -0.101 -3.89Year from application 0 0.013 0.92 -0.005 -0.04

1 0.024 1.50 -0.003 -0.032 0.029 2.57 0.017 0.123 0.026 2.25 0.109 1.404 0.009 0.71 0.126 2.035 0.006 0.53 0.068 1.266 -0.001 -0.07 0.005 0.10

s = -2 0.014 1.11 -0.031 -0.39-1 0.030 3.52 0.006 0.100 0.030 3.09 0.034 0.501 0.050 4.69 0.056 0.722 0.057 5.56 0.120 1.523 0.048 3.59 0.159 1.674 0.043 3.67 0.164 1.975 0.035 2.84 0.106 1.16

beyond 0.028 2.89 0.163 2.17

Country fixed effectsYear fixed effects

YY

YY

Table 7. Changes in Growth and Investment for Non-AXXVI 5(c) Countries Around Application and Accession

Annual Growth Rate Log(Pri inv/GDP)1 2

24

0 2 4 6 -2 0 2 4

-0.01

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08C

ha

ng

e in

Gro

wth

in P

erc

en

tag

e P

oin

tsFigure 5. Change in Growth Following Application and Around Accession

90% C.I.

Years Since Application Years Since Accession

25

0 2 4 6 -2 0 2 4-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4P

erc

en

t C

ha

ng

e in

In

v/G

DP

Ra

io

Figure 6. Change in Inv/GDP Following Application and Around Accession

90% C.I.

Years Since Application Years Since Accession

26

Testing Selection Bias Formally Does the endogenous nature of

membership status bias our results? Employ 2-stage estimation a la Heckman

1st stage: Probit estimation of membership status as a function of lagged GDP and Trade/GDP ratio

2nd stage: Add in the inverse Mills ratio as an independent variable

Coef. est. on the inverse Mills ratio are insignificant => No evidence that selection biase is quantitatively important

27

Coef est. t-stat. Coef est. t-stat. Coef est. t-stat.

Inverse Mills Ratio -0.002 -0.97 -0.033 -1.17 -0.003 -1.48

s = -2 -0.012 -0.82 0.053 0.58 -0.002 -0.16

-1 0.009 1.28 0.108 1.06 0.014 1.99

0 0.085 1.06 0.136 1.41 0.014 1.56

1 0.027 4.14 0.143 1.58 0.029 3.79

2 0.027 4.07 0.177 1.92 0.026 3.53

3 0.015 1.74 0.185 1.88 0.014 1.39

4 0.005 0.66 0.237 2.67 0.009 1.22

5 -0.001 -0.13 0.160 1.78 0.001 0.09

beyond -0.001 -0.35 0.169 4.33 0.001 0.15

s*AXXVI5c Dummy: -2 * AXXVI5c 0.035 1.82 0.015 0.10 0.025 1.26

-1 * AXXVI5c 0.032 2.28 -0.163 -0.86 0.032 2.33

-0 * AXXVI5c 0.001 0.06 -0.242 -1.14 -0.001 -0.06

1 * AXXVI5c -0.030 -2.46 -0.119 -0.77 -0.025 -1.87

2 * AXXVI5c -0.013 -1.37 -0.161 -0.94 -0.008 -0.71

3 * AXXVI5c -0.024 -2.06 -0.148 -0.89 -0.022 -1.65

4 * AXXVI5c -0.034 -1.60 -0.278 -1.47 -0.041 -1.81

5 * AXXVI5c 0.011 0.94 -0.280 -1.75 0.007 0.64

beyond * AXXVI5c -0.014 -2.25 -0.463 -4.62 -0.012 -1.89

Country fixed effects a/

Year fixed effects

Bera-Jarque-Lee LM test statistics

N

Y

N

Y

18.38 0.67


1 2

Table 8. Testing for Selection Bias

3

Annual Growth Rate

N

Y

0.67

28

Effects of Policy Commitments Sample: Accessions after WTO formation

More standardized documents 15 accessions

Commitments touch on general policies Wide-ranging: e.g., TRIPs, transparency,

competition policy, price controls, etc. Might spillover to other non-trade areas Clearly incorporated in Protocol – enforceable

thru’ WTO dispute settlement mechanism

29

Measure of degree of policy commitments # commitments explicitly stated in Protocol

Transparent, non-subjective Not one-size-fits-all:

“…accessions to the WTO will be ‘on terms to be agreed’… essentially a process of negotiation,” WTO.org

Alternative measure: # words in WP Report

30

# Commitments in WPRs Governance Index # Words in WPRsBulgaria 27 2.83 24542Albania 29 2.84 38829China 147 2.49 78641Croatia 27 2.39 38479Ecuador 21 2.56 25835Estonia 24 4.76 22920Georgia 29 1.32 27139Jordan 29 3.26 36608Kyrgyz 29 2.15 32149Latvia 22 3.72 25717Lithuthania 28 3.23 43029Moldova 28 2.87 43859Mongolia 17 2.91 12055Oman 26 4.73 24695Panama 24 3.91 19558

mean incl. China 33.8 3.06 32937 excl. China 25.7 29672

median 27 2.87 27139standard dev. incl. China 31.5 0.92 15537 excl. China 3.7 9370

Table 9. Extent of Commitments and Governance Quality of Acceding Countries

31

Differential effects of policy commitments First View: Countries w/ weak governance

May lack incentives to implement reforms May lack credibility to follow through

Second View: Accession commitments may be a partial remedy for weak governance

So, effect of #commitments is stronger or weaker for countries w/ weak governance?

32

Differential Effects: Results Interact the commitment variable w/

governance measure Interaction terms negative and significant

in both growth and investment regressions

Support the interpretation that accession-related commitment may partly offset the shortcomings of weak governance

33


s*log(# Com.)*Gov Quality: a/ -2 * log(#Com.) * gov. -1.549 -2.10 -4.976 -2.68

-1 * log(#Com.) * gov. -1.351 -2.98 -7.917 -5.62

-0 * log(#Com.) * gov. -0.971 -2.08 -6.845 -4.00

1 * log(#Com.) * gov. -1.489 -3.27 -8.309 -4.19

2 * log(#Com.) * gov. -1.476 -3.18 -8.378 -4.56

3 * log(#Com.) * gov. -2.620 -5.94 -9.656 -5.20

4 * log(#Com.) * gov. -1.475 -1.88 -9.383 -4.05

5 * log(#Com.) * gov. -0.870 -1.52 -1.966 -0.88

beyond * log(#Com.) * gov. -3.933 -4.24 17.081 3.98


Year fixed effects

Tables 13-14. Differential Effects of Policy Commitments on Changes in Growth and Inv/GDP

Y

Y

Annual Growth Rate Log(Inv/GDP)

Y

Y

34

Discrete categorization Governance

Above world median (3): Good governance Otherwise: Poor governance

Policy commitments Above sample median (27): Many com. Alternatively, #words > sample median: Many com. Otherwise: Few com.

Basic results hold: Policy commitments have stronger effects among poor-gov countries (Table 13)

35

Conclusions Accessions entail wide-ranging reform

commitments We yield affirmative results of accession

benefits by1. Distinguishing countries by the degree of accession

requirements

2. Tracing out the trajectories of performance changes

3. Considering the endogenous nature of membership

4. Showing the differential effects of accession requirements on countries with different governance quality

36

~The End~

Thank You

37

Stronger effects among poor-gov. countries on both growth and Inv/GDP Difference in year 2 after accession among:

Poor-gov.: 7.3 perc. points (change in growth)

40% (change in Inv/GDP)

Good-gov.: -0.8 perc. points (change in growth)

-52% (change in Inv/GDP)

Basic results hold with #words Caveat: based on only small sample

Documents

Does WTO Accession Raise Income?