RDP 9705: The Response of the Current Account to Terms of Trade Shocks: A Panel-data Study Appendix A: Further Results

Christopher Kent

October 1997

A.1 Correlation Episodes and Terms of Trade Persistence

Table A1: Summary of Country Correlations Episodes and TOT Persistence
Country
 		 Correlation of CAt with Episodes from Table 2 correlations Data exists ADF Inline Equation, (p = 3)
 
TOTt
 		 TOTt-1
 		 positive
 		 zero
 		 negative
 		 '70–'92
 		 Trend & constant Constant only
Average/Sum 0.21 −0.15 516 1,120 240   0.68 0.84
Algeria 0.57 −0.44 5 2 0 yes 0.89 0.92
Antigua & Barbuda 0.04 −0.28 1 16 2   0.64 0.67
Argentina −0.12 0.13 9 10 6 yes 0.75 0.97
Australia 0.13 −0.40 3 6 5 yes 0.47 0.98
Austria 0.24 −0.17 6 11 0 yes 0.75 0.91
Bahamas, The −0.06 0.62 2 14 4   0.79 0.78
Bahrain 0.13 0.20 3 9 3   0.78 0.78
Bangladesh 0.16 0.07 1 15 3 '71–'92 0.31 0.86
Barbados 0.28 −0.21 3 6 2 yes 0.64 0.92
Belgium 0.49 −0.21 9 3 0 yes 0.76 0.94
Bolivia 0.44 −0.49 4 9 0 yes 0.97 0.86
Botswana 0.00 −0.09 0 4 3 '71–'92 0.91 0.86
Brazil 0.13 0.02 9 10 2 yes 0.54 0.77
Bulgaria 0.87 −0.19 3 0 0      
Burkina Faso 0.31 −0.02 5 11 1 yes 0.68 0.94
Burundi 0.59 −0.04 7 10 0 '71–'92 0.54 0.87
Cameroon 0.37 −0.31 8 8 0 '71–'92 0.73 0.74
Canada −0.03 −0.27 3 11 5 yes 0.85 0.94
Cape Verde 0.44 −0.36 4 11 0   0.63 0.91
Centrl. African Rep. 0.12 −0.21 4 6 1 yes 0.71 0.71
Chad 0.08 0.06 3 13 1 yes 0.61 0.93
Chile 0.49 −0.19 7 11 1 yes 0.79 0.95
China 0.24 0.00 6 13 0 '71–'92 0.69 0.71
Colombia 0.22 −0.06 4 5 2 yes 0.89 0.83
Costa Rica 0.35 −0.31 6 15 2 yes 0.07 0.91
Cote d'Ivoire 0.39 −0.38 7 9 2 yes 0.81 0.79
Cyprus −0.06 −0.11 5 12 3 yes 0.79 0.99
Denmark 0.29 −0.31 5 6 1 yes 0.80 0.86
Dominica −0.27 −0.19 0 10 4   0.78 0.91
Dominican Republic 0.33 0.02 5 8 2 yes 0.30 0.59
Ecuador 0.18 −0.26 4 3 0 yes 0.88 0.92
Egypt, Arab Rep. −0.15 −0.22 2 5 3 yes 0.66 0.78
El Salvador 0.47 −0.41 5 8 0 yes 0.79 1.02
Equatorial Guinea 0.06 −0.09 5 10 0   0.10 0.11
Ethiopia −0.14 −0.03 3 9 2 yes 0.60 0.93
Fiji 0.40 −0.37 2 8 2 yes 0.69 0.84
Finland −0.31 −0.05 2 9 10 yes 0.64 0.63
France 0.58 −0.46 6 3 0 yes 0.75 0.75
Gabon 0.67 −0.28 4 3 0 yes 0.83 0.91
Gambia, The 0.29 −0.10 4 14 2 '71–'92 0.52 0.86
Germany 0.15 −0.06 5 8 2 yes 0.74 0.75
Ghana 0.07 −0.02 2 11 4 yes 0.81 0.83
Greece −0.22 0.23 3 14 3 yes 0.59 0.92
Grenada 0.05 −0.21 2 14 2   0.82 0.99
Guatemala 0.57 −0.54 5 10 0 yes 0.65 1.01
Guinea 0.31 −0.52 0 16 0   0.81 0.93
Guyana 0.38 −0.24 3 8 0 yes 0.49 0.81
Haiti 0.48 −0.15 9 8 1 yes 0.39 1.03
Honduras −0.03 −0.02 2 9 8 yes 0.31 1.09
Hong Kong 0.40 −0.29 8 12 2   0.82 1.04
Hungary −0.09 0.04 4 4 3   0.76 0.94
Iceland 0.09 −0.47 6 17 2 yes 0.70 0.73
India 0.09 0.02 5 8 3 yes 0.77 0.92
Indonesia 0.38 −0.48 5 1 1 yes 0.87 0.91
Iran, Islamic Rep. 0.55 −0.44 2 3 2 yes 0.85 0.91
Iraq 0.65 −0.32 1 6 0   0.84 0.92
Ireland 0.30 −0.52 5 9 0 yes 0.58 0.63
Israel 0.26 0.02 5 8 1 yes 0.64 0.63
Italy 0.58 −0.23 7 5 0 yes 0.88 0.87
Jamaica 0.20 −0.44 4 4 1 yes 0.38 0.44
Japan 0.39 −0.39 8 8 0 yes 0.89 0.84
Jordan −0.04 0.11 2 5 2 yes 0.53 0.65
Kenya 0.21 −0.43 6 3 4 yes 0.93 0.94
Korea, Republic of 0.48 −0.19 10 3 3 yes 0.73 0.74
Kuwait 0.11 −0.15 0 7 0   0.86 0.92
Lebanon 0.23 0.49 0 11 0   0.63 0.87
Liberia −0.14 −0.17 5 15 2   0.73 0.90
Libya 0.39 −0.44 4 1 2   0.85 0.91
Madagascar 0.05 0.29 0 14 4   0.67 0.85
Malawi 0.20 −0.39 5 14 1 yes 0.24 0.76
Malaysia 0.31 −0.68 0 15 0   0.75 0.77
Mali 0.26 0.24 7 12 2 yes 0.19 0.36
Malta 0.22 −0.11 5 7 5 yes 0.89 0.83
Mauritania −0.29 −0.02 2 7 6 yes 0.80 0.95
Mauritius 0.46 −0.33 5 8 0 yes 0.64 0.64
Mexico −0.03 −0.23 2 7 0 yes 0.82 0.84
Morocco 0.04 −0.60 1 1 4 yes 0.37 0.61
Mozambique 0.32 0.17 3 14 0   0.41 0.43
Namibia 0.15 0.32 2 13 0   0.38 0.83
Netherlands 0.22 −0.14 6 10 2 yes 0.82 0.90
Netherlands Antilles 0.61 0.47 2 21 0   0.86 0.87
New Zealand 0.20 −0.50 7 11 0 yes 0.61 0.75
Nicaragua −0.11 −0.08 2 12 2 yes 0.71 0.98
Niger −0.40 0.48 0 12 5 yes 0.63 0.94
Nigeria 0.55 −0.51 4 5 2 yes 0.90 0.91
Norway 0.59 0.00 8 6 0 yes 0.80 0.83
Oman 0.72 −0.52 2 2 0   0.87 0.87
Pakistan −0.24 0.25 2 8 2 yes 0.72 1.00
Panama −0.20 0.24 2 14 2 yes 0.84 0.81
Papua New Guinea 0.63 −0.05 9 12 0 '71–'92 0.10 0.80
Paraguay 0.58 0.18 4 17 0 yes 0.40 0.89
Peru 0.08 −0.33 5 18 5 yes 0.80 0.96
Philippines −0.18 −0.33 2 9 3 yes 0.68 0.87
Poland −0.14 0.84 2 12 1   0.36 0.55
Portugal 0.29 −0.42 9 8 2 '71–'92 0.22 0.95
Qatar 0.58 −0.53 3 3 0   0.83 0.91
Rwanda 0.16 −0.31 3 5 1 yes 0.78 0.87
St. Kitts and Nevis 0.23 0.49 4 15 0   0.84 0.83
St. Lucia −0.10 −0.25 2 13 2   0.77 0.76
Saudi Arabia 0.72 −0.34 4 2 0 yes 0.84 0.92
Senegal 0.25 0.08 4 3 2 yes 0.72 0.77
Sierra Leone 0.13 −0.24 5 12 4 yes 0.04 0.98
Singapore −0.10 0.12 6 4 2 yes 0.73 0.86
Solomon Islands −0.42 −0.19 0 7 5   0.68 0.85
South Africa −0.09 −0.29 7 10 6 yes 0.86 0.98
Spain 0.24 −0.37 6 12 6 yes 0.87 0.89
Sri Lanka 0.54 −0.24 9 6 0 yes 0.36 0.67
Sudan 0.34 0.11 2 8 0 yes 0.38 0.97
Swaziland 0.41 0.24 2 5 0   0.61 0.65
Sweden 0.28 −0.19 3 5 1 yes 0.82 0.83
Switzerland 0.12 −0.10 6 4 6 yes 1.10 0.81
Syrian Arab Rep. −0.25 −0.24 0 7 3 '70–'91 0.85 0.93
Tanzania 0.12 −0.57 5 6 3 yes 0.93 1.09
Thailand −0.31 0.12 0 13 2 yes 0.26 0.89
Tonga −0.39 0.01 0 10 2   0.67 0.82
Trinidad & Tobago 0.49 −0.20 5 7 1 yes 0.85 0.87
Tunisia 0.12 −0.38 0 10 0 yes 0.80 0.90
Turkey 0.33 −0.06 4 9 2 yes 0.82 0.88
Uganda 0.24 −0.28 7 6 2 yes 0.74 0.82
United Arab Emrts. 0.87 −0.32 6 1 0   0.85 0.90
United Kingdom 0.53 0.06 5 6 0 yes 0.67 0.66
United States 0.01 −0.58 0 7 9 yes 0.85 0.93
Uruguay 0.18 −0.21 5 5 3 yes 0.53 0.78
Venezuela 0.58 −0.34 6 1 0 yes 0.82 0.89
Western Samoa −0.06 −0.22 2 11 2   0.64 0.79
Zaire −0.12 0.26 2 16 3   0.72 0.96
Zambia 0.72 −0.35 10 13 0 yes 0.72 0.91
Zimbabwe −0.08 −0.19 3 15 3 '71–'92 0.53 0.95

A.2 Impulse Response Functions: Confidence Intervals

Confidence intervals around impulse response functions (discussed at the end of Section 4.4) were estimated by using a bootstrap technique. For each model estimated, new current account data was created by assuming the original model estimate to be true and generating errors from an independently identically distributed normal distribution with a mean of zero and a variance estimated from the original error estimates. New estimates of the model were produced using this new current account series, and from this a new impulse response function was constructed. This procedure was repeated one thousand times. A 95 per cent confidence interval was constructed by eliminating the upper 25 and lower 25 extreme impulse responses point by point along the impulse response function.[37]

Figure A2: Current Account Response to a 1 Per Cent Terms of Trade Shock

A.3 Panel-data Results Including Government Fiscal Balances

Table A2: Panel Data FGLS – Highest versus Lowest TOT Persistence
Dependent variable – ΔCA, period of estimation – 1970 to 1992
11 countries with highest TOT persistence 9 countries with lowest TOT persistence(a)
Full model(b) Parsimonious model Full model(b) Parsimonious model(b)
Variable Lag Coefficient (t-statistics)
ΔCA 1 −0.252 (−3.02) −0.192 (−2.61) −0.322 (−3.91) −0.398 (−2.97)
2 −0.138 (−2.08) – −0.379 (−5.35) −0.428 (−5.71)
3 −0.206 (−3.58) −0.182 (−3.66) −0.114 (−1.94) −0.126 (−2.33)
4 −0.096 (−1.56) – −0.167 (−2.54) −0.166 (−2.75)
ΔTOTs 0 0.061 (2.19) 0.060 (2.11) 0.137 (2.81) 0.134 (2.77)
1 −0.049 (−1.83) −0.053 (−2.07) 0.001 (0.01) –
2 −0.037 (−1.78) −0.043 (−2.41) −0.001 (−0.04) –
3 −0.009 (−0.51) – 0.009 (0.18) –
4 −0.055 (−2.67) −0.047 (−2.08) 0.010 (0.32) –
ΔGDPs 0 −0.199 (−3.79) −0.205 (−3.94) −0.161 (−2.06) −0.161 (−2.09)
1 −0.019 (−0.47) – −0.169 (−2.15) −0.195 (−2.29)
2 −0.028 (−0.62) – −0.108 (−1.89) −0.201 (−1.98)
3 0.020 (0.51) – −0.046 (−0.54) –
4 0.028 (0.74) – 0.030 (0.37) –
ΔGVT 0 0.112 (1.58) 0.100 (1.55) 0.189 (2.04) 0.229 (2.38)
1 −0.008 (−0.12) – 0.063 (0.69) –
Inline Equation   −0.087
 		 (−2.10)
 		 −0.083
 		 (−2.50)
 		 0.155
 		 (1.71)
 		 0.134
 		 (2.77)
 
Inline Equation   −0.198
 		 (−1.78)
 		 −0.205
 		 (−3.94)
 		 −0.454
 		 (−2.10)
 		 −0.556
 		 (−3.36)
 
Inline Equation   0.104
 		 (0.84)
 		 0.100
 		 (1.55)
 		 0.252
 		 (1.83)
 		 0.229
 		 (2.38)
 
  No. Obs 209 No. Obs 209 No. Obs 171 No. Obs 171
  R2 = 0.38 R2 = 0.35 R2 = 0.47 R2 = 0.39
Notes: (a) Jamaica is excluded because it is missing a lot of GVT data.
(b) Indicates standard errors corrected for heteroskedasticity and for serial correlation.

Footnote

The alternative was to construct a metric that estimated the distance of the impulse response functions from the initial estimate of the function and then eliminate extreme functions to create a confidence interval. [37]