ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Global Poverty, Inequality, and Aid Flows: A Rough Guide to Some Simple Justice

How one measures poverty and inequality has implications for a variety of policy interventions relating to fair allocation in a number of institutional settings. The distribution of international aid is an important case in point. This essay reasserts the importance of certain old-fashioned questions relating to international aid: what is the quantum of aid available in relation to the need for it? How may patterns of allocation, at both the dispensing and receiving ends of aid, be determined so as to take account of both poverty and inter-national inequality in the distribution of incomes? Can some simple and plausible rules of allocation be devised? If so, what correspondence does reality bear to such rules? The questions are addressed with the aid of some simple analytics relating to optimal budgetary intervention in the alleviation of poverty. The ideas discussed are clarified by means of data employed in elementary empirical illustrations.

SPECIAL ARTICLE

Global Poverty, Inequality, and Aid Flows: A Rough Guide to Some Simple Justice

S Subramanian

How one measures poverty and inequality has implications for a variety of policy interventions relating to fair allocation in a number of institutional settings. The distribution of international aid is an important case in point. This essay reasserts the importance of certain old-fashioned questions relating to international aid: what is the quantum of aid available in relation to the need for it? How may patterns of allocation, at both the dispensing and receiving ends of aid, be determined so as to take account of both poverty and inter-national inequality in the distribution of incomes? Can some simple and plausible rules of allocation be devised? If so, what correspondence does reality bear to such rules? The questions are addressed with the aid of some simple analytics relating to optimal budgetary intervention in the alleviation of poverty. The ideas discussed are clarified by means of data employed in elementary empirical illustrations.

I would like to thank, without implicating, Mark McGillivray for suggesting the problem to me. This essay is a revised version of the paper ‘International Aid in Light of Global Poverty and Inequality: Some Unsubtle Propositions’, which appeared in 2007 as Research Paper No 2007/31 in UNU-WIDER’s Research Papers Series, Helsinki. Thanks are due to UNU-WIDER for their hospitality, and for permission to rework the research paper in the preparation of this essay. I am considerably indebted to Thomas Pogge and Sanjay Reddy for helpful comments on an earlier version of the paper. The usual disclaimer applies.

S Subramanian (sspa1994@bsnl.in) is with the Madras Institute of Development Studies, Chennai.

T
his is an old-fashioned essay which will reassert some o ldfashioned views on international aid, global poverty, and inter-country inequality. The literature on aid allocation has become increasingly complex, nuanced, and fine-tuned, but often at the cost of disengagement with certain large and undeniable truths which are crucially germane to the issue. The present essay attempts to keep the broader picture in view while dealing with some simple rules of aid allocation which are motivated by considerations of “how much?”, “from whom?”, and “to whom?”. In the process, it addresses the following questions. How much poverty is there in the world? How much aid is available in r elation to the need for it? How onerous is the redistributive effort entailed in eradicating global poverty? What relation do the amounts of aid disbursed by different countries have to the relative capabilities of these donor countries? What relation does the pattern of aid receipt bear to the relative needs of beneficiary countries? These issues are addressed largely within the framework of a simple analysis of optimal budgetary intervention in the redress of poverty.

It should be emphasised that the “empirical” content of this essay is not distinguished by any particular care devoted to the complete reliability or “up-to-date-ness” of the data employed. The overall objective is no more than to present rough orders of magnitude of a set of indicators and statistics relevant to an assessment of the aid problem. This essay is based very considerably on Subramanian (2003), and especially on Subramanian (2007); and the idea is simply to convey an impressionistic p icture

– without being distracted by peripheral issues – of certain aspects of international aid in the context of global poverty and inequality. Insofar as it has proved possible, all formalisms have been relegated to footnotes which are likely to be of interest only to the specialist, and which therefore can be ignored by the nonspecialist. All tabulated data have been gathered together and placed at the end of the essay.

The Magnitude of Global Poverty

We use 2005 data on country-wise gross domestic product (GDP), population, aid disbursement, and aid receipt from the UNDP’s Human Development Report (HDR) 2007-08. The report presents information on a variety of socio-economic indicators for a set of 177 countries. Information on GDP, population, and aid receipt is available for a set of 174 countries which, together, we shall treat as constituting the “world”. Before proceeding further it is as well to point out that all monetary aggregates in this paper are in terms of nominal US dollars, uncorrected for inter-country price differentials: while there may be a case for employing purchasing

Economic & Political Weekly

EPW
NOVEMBER 15, 2008

Table 1: Cross-Country Data on GDP, Population and Aid Receipts (2005) Table 1: (Continued)
No Country GDP Per Capita GDP Aid Receipts Population No Country GDP Per Capita GDP Aid Receipts Population
($ billion) ($) ($ million) (million) ($ billion) ($) ($ million) (million)
1 Luxembourg 36.5 79,851 0.4571 60 Venezuela (Bolivarian Republic of) 140.2 5,275 48.7 26.5782
2 Norway 295.5 63,918 4.6231 61 Malaysia 130.3 5,142 31.6 25.3403
3 Iceland 15.8 53,290 0.2965 62 South Africa 239.5 5,109 700 46.8781
4 Qatar 42.5 52,240 0.8136 63 Mauritius 6.3 5,059 31.9 1.2453
Switzerland 367 49,351 7.4365 64 Turkey 362.5 5,030 464 72.0676
6 Ireland 201.8 48,524 4.1588 65 Saint Lucia 0.8 5,007 11.1 0.1598
7 Denmark 258.7 47,769 5.4156 66 Uruguay 16.8 4,848 14.6 3.4653
8 United States 12,416.50 41,890 296.4073 67 Panama 15.5 4,786 19.5 3.2386
9 Sweden 357.7 39,637 9.0244 68 Argentina 183.2 4,728 99.7 38.7479
Netherlands 624.2 38,248 16.3198 69 Costa Rica 20 4,627 29.5 4.3225
11 Austria 306.1 37,175 8.2340 70 Romania 98.6 4,556 21.6418
12 Finland 193.2 36,820 5.2471 71 Grenada 0.5 4,451 44.9 0.1123
13 United Kingdom 2,198.80 36,509 60.2262 72 Brazil 796.1 4,271 191.9 186.3966
14 Australia 732.5 36,032 20.3292 73 Dominica 0.3 3,938 15.2 0.0762
Japan 4,534.00 35,484 127.7759 74 Belize 1.1 3,786 12.9 0.2905
16 Belgium 370.8 35,389 10.4778 75 Kazakhstan 57.1 3,772 229.2 15.1379
17 France 2,126.60 34,936 60.8713 76 Saint Vincent and the Grenadines 0.4 3,612 4.9 0.1107
18 Canada 1,113.80 34,484 32.2990 77 Jamaica 9.6 3,607 35.7 2.6615
19 Germany 2,794.90 33,890 82.4698 78 Bulgaria 26.6 3,443 7.7258
Kuwait 80.8 31,861 2.5360 79 Dominican Republic 29.5 3,317 77 8.8936
21 Italy 1,762.50 30,073 58.6074 80 Fiji 2.7 3,219 64 0.8388
22 United Arab Emirates 129.7 28,612 4.5331 81 Algeria 102.3 3,112 370.6 32.8728
23 Singapore 116.8 26,893 4.3431 82 Belarus 29.6 3,024 53.8 9.7884
24 New Zealand 109.3 26,664 4.0992 83 Namibia 6.1 3,016 123.4 2.0225
Spain 1,124.60 25,914 43.3974 84 Suriname 1.3 2,986 44 0.4354
26 Hong Kong, China (SAR) 177.7 25,592 6.9436 85 Tunisia 28.7 2,860 376.5 10.0350
27 Cyprus 15.4 20,841 0.7389 86 Peru 79.4 2,838 397.8 27.9774
28 Greece 225.2 20,282 11.1034 87 Macedonia (TFYR) 5.8 2,835 230.3 2.0459
29 Israel 123.4 17,828 6.9217 88 Iran (Islamic Republic of) 189.8 2,781 104 68.2488
Bahrain 12.9 17,773 0.7258 89 Ecuador 36.5 2,758 209.5 13.2342
31 Bahamas 5.5 17,497 0.3143 90 Thailand 176.6 2,750 -171.1 64.2182
32 Portugal 183.3 17,376 10.5490 91 Colombia 122.3 2,682 511.1 45.6003
33 Slovenia 34.4 17,173 2.0031 92 Albania 8.4 2,678 318.7 3.1367
34 Brunei Darussalam 6.4 17,121 0.3738 93 Bosnia and Herzegovina 9.9 2,546 546.1 3.8885
Korea (Republic of) 787.6 16,309 48.2924 94 Guatemala 31.7 2,517 253.6 12.5944
36 Malta 5.6 13,803 0.4057 95 El Salvador 17 2,467 199.4 6.8910
37 Saudi Arabia 309.8 13,399 26.3 23.1211 96 Swaziland 2.7 2,414 46 1.1185
38 Czech Republic 124.4 12,152 10.2370 97 Maldives 0.8 2,326 66.8 0.3439
39 Barbados 3.1 11,465 -2.1 0.2704 98 Jordan 12.7 2,323 622 5.4671
Trinidad and Tobago 14.4 11,000 -2.1 1.3091 99 Samoa 0.4 2,184 44 0.1832
41 Hungary 109.2 10,830 10.0831 100 Tonga 0.2 2,090 31.8 0.0957
42 Antigua and Barbuda 0.9 10,578 7.2 0.0851 101 Angola 32.8 2,058 441.8 15.9378
43 Estonia 13.1 9,733 1.3459 102 Cape Verde 1 1,940 160.6 0.5155
44 Oman 24.3 9,584 30.7 2.5355 103 Ukraine 82.9 1,761 409.6 47.0755
Saint Kitts and Nevis 0.5 9,438 3.5 0.0530 104 China 2,234.30 1,713 1,756.90 1304.3199
46 Croatia 38.5 8,666 125.4 4.4426 105 Morocco 51.6 1,711 651.8 30.1578
47 Slovakia 46.4 8,616 5.3853 106 Turkmenistan 8.1 1,669 28.3 4.8532
48 Seychelles 0.7 8,209 18.8 0.0853 107 Armenia 4.9 1,625 193.3 3.0154
49 Poland 303.2 7,945 38.1624 108 Azerbaijan 12.6 1,498 223.4 8.4112
Lithuania 25.6 7,505 3.4111 109 Georgia 6.4 1,429 309.8 4.4787
51 Mexico 768.4 7,454 189.4 103.0856 110 Syrian Arab Republic 26.3 1,382 77.9 19.0304
52 Chile 115.2 7,073 151.7 16.2873 111 Bhutan 0.8 1,325 90 0.6038
53 Latvia 15.8 6,879 2.2968 112 Indonesia 287.2 1,302 2,523.50 220.5837
54 Libyan Arab Jamahiriya 38.8 6,621 24.4 5.8601 113 Congo 5.1 1,273 1,448.90 4.0063
Equatorial Guinea 3.2 6,416 39 0.4988 114 Paraguay 7.3 1,242 51.1 5.8776
56 Lebanon 21.9 6,135 243 3.5697 115 Egypt 89.4 1,207 925.9 74.0679
57 Botswana 10.3 5,846 70.9 1.7619 116 Sri Lanka 23.5 1,196 1,189.30 19.6488
58 Gabon 8.1 5,821 53.9 1.3915 117 Philippines 99 1,192 561.8 83.0537
59 Russian Federation 763.7 5,336 143.1222 118 Honduras 8.3 1,151 680.8 7.2111
54 NOVEMBER 15, 2008 Economic & Political Weekly
EPW

Table 1: (Continued) power parity (PPP) exhange rates in order to present monetary

No Country GDP Per Capita GDP Aid Receipts Population
($ billion) ($) ($ million) (million)
119 Occupied Palestinian Territories 4 1,107 1,101.60 3.6134
120 Guyana 0.8 1,048 136.8 0.7634
121 Cameroon 16.9 1,034 413.8 16.3443
122 Bolivia 9.3 1,017 582.9 9.1445
123 Nicaragua 4.9 954 740.1 5.1363
124 Côte d’Ivoire 16.3 900 119.1 18.1111
125 Djibouti 0.7 894 78.6 0.7830
126 Papua New Guinea 4.9 840 266.1 5.8333
127 Lesotho 1.5 808 68.8 1.8564
128 Sudan 27.5 760 1,828.60 36.1842
129 Nigeria 99 752 6,437.30 131.6489
130 Mongolia 1.9 736 211.9 2.5815
131 India 805.7 736 1,724.10 1094.7011
132 Yemen 15.1 718 335.9 21.0306
133 Pakistan 110.7 711 1,666.50 155.6962
134 Senegal 8.2 707 689.3 11.5983
135 Moldova 2.9 694 191.8 4.1787
136 Comoros 0.4 645 25.2 0.6202
137 Viet Nam 52.4 631 1,904.90 83.0428
138 Solomon Islands 0.3 624 198.2 0.4808
139 Zambia 7.3 623 945 11.7175
140 Mauritania 1.9 603 190.4 3.1509
141 Chad 5.5 561 379.8 9.8039
142 Kenya 18.7 547 768.3 34.1865
143 Uzbekistan 14 533 172.3 26.2664
144 Benin 4.3 508 349.1 8.4646
145 Haiti 4.3 500 515 8.6000
146 Lao People’s Democratic Republic 2.9 485 295.7 5.9794
147 Ghana 10.7 485 1,119.90 22.0619
148 Kyrgyzstan 2.4 475 268.5 5.0526
149 Sao Tome and Principe 0.1 451 31.9 0.2217
150 Cambodia 6.2 440 537.8 14.0909
151 Bangladesh 60 423 1,320.50 141.8440
152 Mali 5.3 392 691.5 13.5204
153 Burkina Faso 5.2 391 659.6 13.2992
154 Timor-Leste 0.3 358 184.7 0.8380
155 Togo 2.2 358 86.7 6.1453
156 Tajikistan 2.3 355 241.4 6.4789
157 Guinea 3.3 350 182.1 9.4286
158 Central African Republic 1.4 339 95.3 4.1298
159 Mozambique 6.6 335 1,285.90 19.7015
160 Tanzania (United Republic of) 12.1 316 1,505.10 38.2911
161 Gambia 0.5 304 58.2 1.6447
162 Uganda 8.7 303 1,198.00 28.7129
163 Nepal 7.4 272 427.9 27.2059
164 Madagascar 5 271 929.2 18.4502
165 Zimbabwe 3.4 259 367.7 13.1274
166 Niger 3.4 244 515.4 13.9344
167 Rwanda 2.2 238 576 9.2437
168 Eritrea 1 220 355.2 4.5455
169 Sierra Leone 1.2 216 343.4 5.5556
170 Guinea-Bissau 0.3 190 79.1 1.5789
171 Malawi 2.1 161 575.3 13.0435
172 Ethiopia 11.2 157 1,937.30 71.3376
173 Congo (Democratic Republic of the) 7.1 123 1,827.60 57.7236
174 Burundi 0.8 106 365 7.5472
Aggregate 44043.5 7050.999 59584.6 6246.42

Source: Tables 14 and 18 of Human Development Report 2007-08. (Population figures have been obtained by dividing GDP figures by corresponding per capita GDP figures in Table 14 of HDR 2007-08.)

Economic & Political Weekly

EPW
NOVEMBER 15, 2008

aggregates in “real” terms, this exercise has not been undertaken in the present paper. For one thing, the paper would become disproportionately bloated if all the relevant money-metric statistics were to be re-tabulated and reinterpreted in PPP-dollar terms. For another, the emphasis in this paper, as already stated, is on broad approaches, principles, magnitudes, trends, and tendencies, rather than on precision-claiming complications of detail.

The per capita GDP in 2005 for the set of 174 countries under consideration, at current prices, is $7,051 (Table 1, p 54). It seems reasonable to suggest that a country should be deemed to be poor if its per capita GDP is less than $1,400. Unquestionably, this is an arbitrary judgment, but arguably not an unreasonable one. An international poverty line which is pitched at less than a fifth of the global per capita GDP can scarcely attract the criticism of excessive liberalism, considering that the national poverty line in many advanced western countries is set at one-half the median family income.

We shall, for the most part, not be concerned with the intracountry distribution of income: the assumption will be that within any country, each person receives its per capita income. (We shall, however, have occasion at a later stage to consider some relevant complications which arise from reckoning inequality in the intra-national distribution of income.) Each country can be described in terms of a pair of numbers, the first of which is its per capita GDP, and the second its population size. The global d istribution of income can then be represented by a list of these pairs for all countries of the world. Table 2 (p 56) presents information on the distribution of income for the poor countries of the world. One requires information on the global distribution of income and the poverty line in order to obtain an estimate of global p overty.1

A widely-employed means of measuring poverty is in terms of a family of poverty indices advanced by Foster, Greer and Thorbecke (1984). This is the so-called Pα family, where α is a parameter that assumes non-negative values and reflects a degree of “aversion to poverty”, with higher values of α signifying greater aversion to poverty.2 As is well-known, P0 is the headcount ratio, or proportion of the world’s population living in its poor countries. The headcount ratio violates the monotonicity axiom, which is the requirement that, other things equal, a diminution in any poor person’s income should increase poverty. This index also violates the transfer axiom, which is the requirement that, other things remaining the same, any equalising redistribution of income among the poor should reduce poverty. For positive values of α strictly less than 1, Pα satisfies monotonicity but violates transfer: in fact, it favours dis-equalising transfers among the poor. P1 is the per capita income-gap ratio, or the proportionate deviation of the average income of the poor from the poverty line, expressed in per person terms; this index also satisfies monotonicity without satisfying transfer: it is sensitive only to the aggregate poverty gap, and not to its inter-personal distribution. The index P2, by contrast, does attend to distributional considerations: it satisfies both the monotonicity and the transfer axioms.3 Using information on the global distribution of income (Table 2), the values of P0, P0.5, P1 and P2 can be computed. These

turn out to be, respectively, 0.44, 0.29, Table 2: Distribution of Income in Poor Countries (2005) ensuring that the resulting global distri
0.21, and 0.12. Familiarity with corresponding values of these indices for No 1 Country Syrian Arab Republic Per Capita GDP Population ($) (million) 1,382 19.0304 bution of income cannot be “Lorenzdominated”, as economists put it, by any
known poor countries suggests that the 2 Bhutan 1,325 0.6038 other distribution in which the presently
extent of global poverty is very consider 3 Indonesia 1,302 220.5837 poor countries are just enabled to escape
able. This leads to our First Observation: There is a lot of poverty 4 5 6 Congo Paraguay Egypt 1,273 1,242 1,207 4.0063 5.8776 74.0679 poverty? This problem has been considered by Jayaraj and Subramanian (1996)
in the world. 7 Sri Lanka 1,196 19.6488 in the context of within-country poverty
Aid in Relation to Its Need Let Di stand for the ith poorest country’s 8 Philippines 9 Honduras 10 Occupied Palestinian Territories 11 Guyana 1,192 1,151 1,107 1,048 83.0537 7.2111 3.6134 0.7634 eradication. The solution to the problem can be described as follows. Let the per capita income of the richest country be
deficit, or total shortfall of income from 12 Cameroon 1,034 16.3443 reduced to that of the next richest coun
what is required in order to escape poverty. The aggregate global deficit is D, which is obtained by summing the defi 13 Bolivia 14 Nicaragua 15 Côte d’Ivoire 16 Djibouti 1,017 954 900 894 9.1445 5.1363 18.1111 0.7830 try. If the resulting tax revenue is sufficient to meet the aggregate poverty d eficit D, then that is all that needs to be
cits of all the poor countries.4 Table 3 17 Papua New Guinea 840 5.8333 done. If not, reduce the per capita
(p 57) provides information on the country-wise and total deficit for the set of 18 Lesotho 19 Sudan 20 Nigeria 808 1.8564 760 36.1842 752 131.6489 incomes of the two richest countries to the per capita income of the third richest
poor countries. The aggregate deficit D is 21 Mongolia 736 2.5815 country. If the resulting tax revenue is
of the order of $1,838.5 billion. Data on 22 India 736 1094.7011 sufficient to meet the deficit D, then the
aid disbursed by various countries, avail 23 Yemen 24 Pakistan 718 21.0306 711 155.6962 exercise stops at this stage. If not, the per
able in Table 17 of HDR 2007-08, suggest 25 Senegal 707 11.5983 capita incomes of the three richest coun
that the total quantum of aid disbursed 26 Moldova 694 4.1787 tries should be reduced to the level of the
in 2005 was of the order of $110 billion. The amount of aid available, as a propor 27 Comoros 28 Viet Nam 29 Solomon Islands 645 0.6202 631 83.0428 624 0.4808 fourth richest country’s per capita income…and so on, down the line, until
tion of aid required to eradicate global 30 Zambia 623 11.7175 we reach that marginal country for
p overty, works out to 6 per cent. This 31 Mauritania 603 3.1509 which the aggregate revenue raised is
leads to our Second Observation: The quantum of aid 32 Chad 33 Kenya 34 Uzbekistan 561 9.8039 547 34.1865 533 26.2664 just equal to the aggregate poverty d eficit D. What is entailed is the imple
available, in relation to the need for it, is 35 Benin 508 8.4646 mentation of a “lexicographic maximin

vanishingly small. 36 Haiti 500 8.6000 solution” – the L-M scheme, for short – to 37 Lao People’s Democratic Republic 485 5.9794

the optimal taxation problem.5 (For dis

38 Ghana 485 22.0619

The International Burden of Poverty cussion of a related, though distinct,

39 Kyrgyzstan 475 5.0526 As we have seen, the aggregate poverty 40 Sao Tome and Principe 451 0.2217 problem, see Anand’s (1983) Redressal of

41 Cambodia 440 14.0909

deficit, D, is in the region of $1,839 bil- Poverty Rule.)

42 Bangladesh 423 141.8440

lion. From Table 1, it can be verified that The tax-cum-transfer scheme just

43 Mali 392 13.5204

the aggregate GDP of all the non-poor described can, as it happens, be inter

44 Burkina Faso 391 13.2992

countries – call this Y – is in the region of 45 Timor-Leste 358 0.8380 preted as reflecting an “agglomerative”

46 Togo 358 6.1453

$42,082 billion. The ratio of D to Y is 4.4 view of three principles of distributive

47 Tajikistan 355 6.4789

per cent, a number scarcely suggestive of justice – those of “sufficientarianism” (on

48 Guinea 350 9.4286 an insuperable burden of international 49 Central African Republic 339 4.1298 which see, in particular, Frankfurt 1987),

poverty. Indeed, the Brandt Commission 50 Mozambique 335 19.7015 “prioritarianism” (on which see, in par51 Tanzania (United Republic of) 316 38.2911

on North-South Relations, in 1980, had ticular, Parfit 1998), and “egalitarian

52 Gambia 304 1.6447

recommended an international tax-cum- ism”. (This was pointed out to the author

53 Uganda 303 28.7129 transfer arrangement, and it is worth 54 Nepal 272 27.2059 in personal communication by Sanjay

considering the simple arithmetic of 55 Madagascar 271 18.4502 Reddy.) 56 Zimbabwe 259 13.1274

eradicating global poverty through aid Sufficientarianism is, essentially, the

57 Niger 244 13.9344

disbursements consistent with the imple- view that inequality in itself is objection

58 Rwanda 238 9.2437 mentation of a specific scheme of redis-59 Eritrea 220 4.5455 able only when, and to the extent that, it

60 Sierra Leone 216 5.5556 61 Guinea-Bissau 190 1.5789

tributive taxation, as discussed below. coexists with the failure of some people

Suppose the objective is to ensure that to achieve a standard of living that is

62 Malawi 161 13.0435

every presently poor country is enabled compatible with adequacy, or sufficiency,

63 Ethiopia 157 71.3376

to reach the poverty line of $1,400 per 64 Congo (Democratic Republic of the) 123 57.7236 or freedom from need. Prioritarianism is

65 Burundi 106 7.5472

capita. What would be a maximally equi- the view that just distribution should be

Aggregate 722.68 2714.3564

table tax-transfer scheme which will informed by the desirability of attaching

A “poor country” is one with a per capita GDP of less than $1,400. realise this objective, in the sense of Source: Table 14 of HDR 2007-08. greater weight, or priority, to those that

NOVEMBER 15, 2008

Table 3: Poverty Deficits of Poor Countries (2005) Table 3: (Continued)

No Country Poverty Line Per Capita Per Capita Population Total Poverty No Country Poverty Line Per Capita Per Capita Population Total Poverty

($) GDP ($) Deficit (Million) Deficit ($) GDP ($) Deficit (Million) Deficit

(Poverty Line - (Population (Poverty Line - (Population

Per Capita Times Per Capita Per Capita Times Per Capita

GDP in $) Poverty Deficit GDP in $) Poverty Deficit

in $ billion) in $ billion) 1 Syrian Arab Republic 1400 1382 18 19.03 342.55 35 Benin 1400 508 892 8.46 7550.39 2 Bhutan 1400 1325 75 0.60 45.28 36 Haiti 1400 500 900 8.60 7740.00 3 Indonesia 1400 1302 98 220.58 21617.20 37 Lao People’s Democratic Republic 1400 485 915 5.98 5471.13 4 Congo 1400 1273 127 4.01 508.80 38 Ghana 1400 485 915 22.06 20186.60 5 Paraguay 1400 1242 158 5.88 928.66 39 Kyrgyzstan 1400 475 925 5.05 4673.68 6 Egypt 1400 1207 193 74.07 14295.11 40 Sao Tome and Principe 1400 451 949 0.22 210.42

7 Sri Lanka 1400 1196 204 19.65 4008.36 41 Cambodia 1400 440 960 14.09 13527.27 8 Philippines 1400 1192 208 83.05 17275.17 42 Bangladesh 1400 423 977 141.84 138581.56

9 Honduras 1400 1151 249 7.21 1795.57 43 Mali 1400 392 1008 13.52 13628.57 10 Occupied Palestinian Territories 1400 1107 293 3.61 1058.72 44 Burkina Faso 1400 391 1009 13.30 13418.93 11 Guyana 1400 1048 352 0.76 268.70 45 Timor-Leste 1400 358 1042 0.84 873.18

12 Cameroon 1400 1034 366 16.34 5982.01 46 Togo 1400 358 1042 6.15 6403.35 13 Bolivia 1400 1017 383 9.14 3502.36 47 Tajikistan 1400 355 1045 6.48 6770.42 14 Nicaragua 1400 954 446 5.14 2290.78 48 Guinea 1400 350 1050 9.43 9900.00 15 Côte d’Ivoire 1400 900 500 18.11 9055.56 49 Central African Republic 1400 339 1061 4.13 4381.71 16 Djibouti 1400 894 506 0.78 396.20 50 Mozambique 1400 335 1065 19.70 20982.09 17 Papua New Guinea 1400 840 560 5.83 3266.67 51 Tanzania (United Republic of) 1400 316 1084 38.29 41507.59

18 Lesotho 1400 808 592 1.86 1099.01 52 Gambia 1400 304 1096 1.64 1802.63 19 Sudan 1400 760 640 36.18 23157.89 53 Uganda 1400 303 1097 28.71 31498.02 20 Nigeria 1400 752 648 131.65 85308.51 54 Nepal 1400 272 1128 27.21 30688.24 21 Mongolia 1400 736 664 2.58 1714.13 55 Madagascar 1400 271 1129 18.45 20830.26 22 India 1400 736 664 1094.70 726881.52 56 Zimbabwe 1400 259 1141 13.13 14978.38 23 Yemen 1400 718 682 21.03 14342.90 57 Niger 1400 244 1156 13.93 16108.20

24 Pakistan 1400 711 689 155.70 107274.68 58 Rwanda 1400 238 1162 9.24 10741.18 25 Senegal 1400 707 693 11.60 8037.62 59 Eritrea 1400 220 1180 4.55 5363.64 26 Moldova 1400 694 706 4.18 2950.14 60 Sierra Leone 1400 216 1184 5.56 6577.78 27 Comoros 1400 645 755 0.62 468.22 61 Guinea-Bissau 1400 190 1210 1.58 1910.53

28 Viet Nam 1400 631 769 83.04 63859.90 62 Malawi 1400 161 1239 13.04 16160.87

29 Solomon Islands 1400 624 776 0.48 373.08 63 Ethiopia 1400 157 1243 71.34 88672.61 64 Congo (Democratic

30 Zambia 1400 623 777 11.72 9104.49 Republic of the) 1400 123 1277 57.72 73713.01

31 Mauritania 1400 603 797 3.15 2511.28 65 Burundi 1400 106 1294 7.55 9766.04

32 Chad 1400 561 839 9.80 8225.49 Aggregate 722.68 677.32 2714 1838.50

33 Kenya 1400 547 853 34.19 29161.06

A “poor country” is one with a per capita GDP of less than $1,400. 34 Uzbekistan 1400 533 867 26.27 22772.98 Source: Derived from Tables 1 and 2 of this paper.

are worse off. Egalitarianism is an expression of the belief that an by any other scheme of redistribution. The L-M scheme, thus, equal distribution of benefits is, in itself, good, or that striving for would appear to be rationalisable in terms of the “moral matheequality, on grounds of justice or fairness (or some other value), matics” (a phrase due to Reddy 2007) of redistribution that can is the right thing to do. That the L-M scheme is compatible with be derived from each of the principles of sufficientarianism, prithese three principles of distributive justice is not hard to see. If oritarianism, and egalitarianism. “sufficiency” in income space is identified with the attainment of Using the data provided in Table 1, it can be verified that only a “poverty line” level of income, then the L-M scheme can be seen the richest 11 countries of the world – Luxembourg, Norway, to be compatible with sufficientarianism in the sense of support-I celand, Qatar, Switzerland, Ireland, Denmark, US, Sweden, ing a pattern of redistribution which ensures inequality is reduced N etherlands, and Austria – would be involved in the redistributive to the extent that those who are currently poor are just enabled exercise described above. The per capita incomes of these counto escape poverty. The L-M scheme is prioritarian in the sense tries would have to be reduced to $37,045, just a little below the that priority is accorded only to the poor, and, among the poor, Austrian per capita income of $37,175. The details are provided in greater weight is attached to the poorer, as reflected in the fact Table 4 (p 58), and the figures in this table suggest the following. that transfers are greater for those with larger poverty deficits, The post-tax-cum-transfer per capita GDP of the seven richest and zero for the non-poor. The demands of egalitarianism are countries taken together will be over 87 per cent of their pre-taxserved by the L-M scheme in the following restricted sense: the cum-transfer per capita GDP, while the post-tax-cum-transfer per objective of raising all those who are poor to the poverty line by capita GDP of the 65 poorest countries taken together will be over redistribution is accomplished by the L-M scheme through a 193 per cent of their pre-tax-cum-transfer per capita GDP. From p attern of income distribution that cannot be Lorenz-dominated an impartial, “arithmetical” point of view, a relatively small

Economic & Political Weekly

EPW
NOVEMBER 15, 2008 57

s acrifice by a small number of rich countries could yield a disproportionately large benefit to a large number of poor countries. The size of the population in the “sacrificing” countries is 353 million, or 13 per cent of the size of the population, at 2,714 million, of the beneficiary countries. There need be no fear that the transfers will be anything like remotely immiserising: at the end of the redistributive exercise, the 11 richest countries will enjoy an average standard of living very near that of Austria; and the

Table 4: Redistributive Taxation for Eradicating Global Poverty, 2005: The ‘Lexicographic Maximin’ Solution

No Country Per Capita Income Level Per Capita Population Total Tax GDP Tax-GDP
GDP ($) (x*) to which Tax (million) (Per Capita ($ billion) Ratio
Per Capita (Per Capita Tax Times (%)
GDP must be GDP – x*) Population)
Reduced ($) ($) ($ billion)
1 Luxembourg 79,851 37,045 42,806 0.4571 19.57 36.5 53.62
2 Norway 63,918 37,045 26,873 4.6231 124.24 295.5 42.04
3 Iceland 53,290 37,045 16,245 0.2965 4.82 15.8 30.49
4 Qatar 52,240 37,045 15,195 0.8136 12.36 42.5 29.09
5 Switzerland 49,351 37,045 12,306 7.4365 91.51 367.0 24.94
6 Ireland 48,524 37,045 11,479 4.1588 47.74 201.8 23.66
7 Denmark 47,769 37,045 10,724 5.4156 58.08 258.7 22.45
8 USA 41,890 37,045 4,845 296.4073 1,436.09 12,416.5 11.57
9 Sweden 39,637 37,045 2,592 9.0244 23.39 357.7 6.54
10 Netherlands 38,248 37,045 1,203 16.3198 19.63 624.2 3.15
11 Austria 37,175 37,045 130 8.2340 1.07 306.1 0.35
Aggregate 1,838.5 14,922.3 12.32

The quantity x* is defined in footnote 4. Source: Derived from Tables 1, 2 and 3 of this paper.

per capita GDP of the richest country ($37,045) will still exceed the per capita GDP of the poorest country ($1,400) by a factor of over 2600 per cent.

The upshot of the preceding discussion leads us to our Third Observation: While the magnitude of global p overty is large, the international burden of poverty is small.

global surplus, S, is then obtained by summing the surpluses of all countries in the set A.6 A reasonably equitable scheme of taxation would be one in which, from among the set A of rich countries, the ith poorest country’s share in total aid disbursed is si, where si = Si/S. That is, a country’s share in aid disbursed is equated to its share in the aggregate surplus. One could refer to si as country i’s “normative share” in aid disbursement.

Table 1 indicates that there are 35 countries constituting the set

A: Luxembourg, Norway, Iceland, Qatar, Switzerland, Ireland, Denmark, United States, Sweden, Netherlands, Austria, Finland, United Kingdom, Australia, Japan, Belgium, France, Canada, Germany, Kuwait, Italy, United Arab Emirates, Singapore, New Zealand, Spain, Hong Kong China (SAR), Cyprus, Greece, Israel, Bahrain, Bahamas, Portugal, Slovenia, Brunei Darussalam, and Republic of Korea. Of these, 22 countries belong to the Development Assistance Committee (DAC) of the OECD. The HDR 2007-08 (Table 17) furnishes information for 2005 on the aid disbursed by each of the DAC countries. Using these data, and data provided in Table 1, Table 5 in this paper presents information, for each of the DAC countries, on its actual share ai of aid disbursed by the DAC countries, and its normative share si. The figures in this table suggest that for all but 5 of the 22 DAC countries, the actual aid share ai is in excess of the normative share si: of salience are the cases of Sweden, Netherlands, Norway, and Denmark, for each of which countries the ratio of ai to si is in excess of 2. New Zealand, I reland, Japan, Australia, and the US are the countries for which the ratio of ai to si is less than unity. Particularly noteworthy, and

Table 5: Actual and Normative Aid Shares in Disbursement for the DAC Countries (2005)

No Country Per Capita Per Capita Population Total Surplus Aid Actual Share Normative Acutal
GDP Surplus (million) (Population Disbursed in Aid Share in Share/
($) (Per Capita Times ($ billion) Disbursed Aid Normative
GDP - 1400) Per Capita (Aid Disbursed Share
($) Surplus) Disbursed/ (Total Surplus/
($ billion) Aggregate Aggregate
Aid Total
Disbursed) Surplus)
(%) (%)
The Disbursement of Aid in Relation to 1 Norway 63918 49918 4.6231 230.78 2.79 2.6092 1.1527 2.2635
Donor Capability 2 Australia 3 Canada 36032 34484 22032 20484 20.3292 32.2990 447.89 661.61 1.68 3.76 1.5734 3.5176 2.2372 3.3047 0.7033 1.0644
The redistributive tax system described in the previ 4 Ireland 48524 34524 4.1588 143.58 0.72 0.6734 0.7172 0.9389
ous section could attract the criticism of being 5 Sweden 39637 25637 9.0244 231.36 3.36 3.1486 1.1556 2.7246
extreme in its insistence on a certain sort of stringent 6 Japan 35484 21484 127.7759 2745.14 13.15 12.3127 13.7119 0.8980
egalitarianism. In this scheme, only 11 of the richest 7 Netherlands 38248 24248 16.3198 395.72 5.12 4.7904 1.9766 2.4235
countries are called upon to bear the burden of inter 8 France 34936 20936 60.8713 1274.40 10.03 9.3898 6.3656 1.4751
national poverty. In particular, only countries with a per capita GDP equalling or exceeding the Austrian per capita GDP of $37,175 are required to disburse aid. 9 Finland 10 United States 11 Spain 12 Denmark 36820 41890 25914 47769 22820 27890 11914 33769 5.2471 296.4073 43.3974 5.4156 119.74 8266.80 517.04 182.88 0.90 27.62 3.02 2.11 0.8448 25.8691 2.8265 1.9752 0.5981 41.2925 2.5826 0.9135 1.4124 0.6265 1.0944 2.1622
There may well be a case for a more broad-based 13 Austria 37175 23175 8.2340 190.82 1.57 1.4732 0.9532 1.5456
spreading of the overheads of global deprivation. The 14 United Kingdom 36509 22509 60.2262 1355.63 10.77 10.0837 6.7714 1.4892
criterion for “aid liability” can be significantly relaxed 15 Belgium 35389 21389 10.4778 224.11 1.96 1.8384 1.1194 1.6423
– by requiring, for instance, that the burden of aid 16 Luxembourg 79851 65851 0.4571 30.10 0.26 0.2398 0.1504 1.5946
should be borne by countries with a per capita GDP in 17 New Zealand 26664 12664 4.0992 51.91 0.27 0.2566 0.2593 0.9896
excess of $14,000 (which is itself 10 times the inter 18 Italy 30073 16073 58.6074 942.00 5.09 4.7679 4.7053 1.0133
national poverty line of $1,400). Let A be the set of these countries. For every country i in the set A, let Si be country i’s surplus, or the total excess of income 19 Germany 20 Greece 21 Portugal 22 Switzerland 33890 20282 17376 49351 19890 6282 3376 35351 82.4698 11.1034 10.5490 7.4365 1640.32 69.75 35.61 262.89 10.08 0.38 0.38 1.77 9.4422 0.3596 0.3531 1.6549 8.1934 0.3484 0.1779 1.3131 1.1524 1.0322 1.9848 1.2603
over what it would be if its per capita GDP were exactly Aggegrate 20020.09 106.78
equal to the cut-off level of $14,000. The aggregate Source: Derived from Table 1 of this paper and Table 17 of HDR 2007-08.
58 NOVEMBER 15, 2008 Economic & Political Weekly
EPW

for reasons opposite to those which make the Nordic countries remarkable, is the case of the US, whose actual aid share is just 63 per cent of its normative share. Indeed, at the levels of aid commitment of New Zealand, Ireland, Japan, Australia, and the US, if the remaining DAC countries decided to disburse aid in accordance with their normative shares, then the total aid disbursement of the DAC countries would be just a little over four-fifths of the present (and already low) level. Clearly, a disproportionate aid effort has had to be put in by one group of countries in order to offset the aid reluctance of countries like the US and Australia which account, respectively, for the highest and next highest share in the aggregate global surplus. This leads us to our

Fourth Observation: The relative contributions to aid bear little relation to the relative capabilities of donor countries.

The Receipt of Aid in Relation to Beneficiary Need

Bourguignon and Fields (1990) – see also Gangopadhyay and Subramanian (1992) – is one of the earliest efforts at engaging explicitly with the question of optimal budgetary intervention in the alleviation of poverty. Their approach can be adapted to the context of an aid allocation exercise. Specifically, if a budget B is available for aid disbursement, and if the objective of aid transfers is to minimise poverty, how much aid Bi should be allocated to the ith poorest country in the set of poor countries? The answer would depend on how one specifies the objective function (or equivalently, in the present case, on how one measures poverty) and also on the constraints under which the optimisation exercise is carried out. Bourguignon and Fields (as adapted to our present concerns) consider different members of the Foster-Greer-Thorbecke Pα family of poverty measures, and they seek to minimise poverty as measured by each of these indices subject to the constraints (a) that the sum of aid transfers does not exceed the budgeted outlay B, (b) that no country receives aid in excess of its poverty deficit, and (c) that aid transfers are always non-negative. Suppose we add a mildly “equality-preferring” fourth constraint which demands that, for every pair of poor countries the share of the poorer country in the combined poverty deficit of the pair should not become larger after the aid transfers have been made. Effectively, this constraint is compatible with the requirement that the poorer (in terms of poverty deficit) of two countries should not receive a smaller transfer. Suppose, further, that p overty is measured by the index P0.5. Subramanian (2006) has demonstrated that the solution to this constrained optimisation problem is a proportional allocation rule, whereby each country receives aid in proportion to its share in the aggregate poverty deficit. That is, the share di of the ith poorest country in the budgeted outlay of B will simply be the country’s share in the aggregate poverty deficit, namely, the quantity Di/D.

7

The proportionality rule just described is, we shall maintain, a reasonably rational guide to aid allocation decisions. It should be mentioned here that Milanovic (2006) has advanced a proposal for the creation of an independent international agency to p reside over the global redistribution of income. The schedule of aid d isbursement by donor countries (reviewed in the preceding s ection), and the schedule of aid-receipts by recipient countries (reviewed in this section), are examples of elementary rules in

Economic & Political Weekly

EPW
NOVEMBER 15, 2008

terms of which Milanovic’s proposed international agency could be guided in its global redistribution exercise.

We shall refer to di as country i’s normative share in aid receipts. How has the pattern of actual country shares in aid receipts – call these the bi – compared with the normative shares? Table 6 (p 60) of this paper, based on 2005 data available in Table 18 of the HDR 2007-08, furnishes information on the amount of aid received by each country for which data are available on GDP, population, and aid receipt. (It is to be noted that of the total aid disbursement of $110 billion in 2005, only about 54 per cent – or $59.3 billion – have been allocated as receipts to specific countries.) Table 6 reveals that, if a poverty line of $1,400 per capita is accepted as an international poverty line, then as many as 60 non-poor countries out of a total of 125 countries have been recipients of positive net aid. Specific examples thrown up by Table 6 are worth noting. If we describe a country by an ordered pair of (per capita GDP, per capita aid received), then here are some pairs of numbers for selected countries, which suggest that it would be hard to find any need-related rationale for aid allocations. (The suggestion, it may be clarified, is not that the precise disproportionalities reported in what follows for the year 2005 are reflective of some secular pattern that is replicated on a year-to-year basis. That is, the claim is about a certain generic inter-temporal commonality of idiosyncratic aid receipts without, necessarily, a specific inter-temporal identity of country-wise patterns.) G renada, with a per capita GDP of $4,451 receives a per capita aid of $400, and Croatia, with a per capita GDP of $8,666, receives a per capita aid of $28: both of these are non-poor countries. C ontrast these figures with those for a pair of poor countries: India, with a per capita GDP of $736, receives a per capita aid of $1.58, while Pakistan, with a per capita GDP of $711, receives a per capita aid of $4.43. Grenada’s per capita GDP is nearly 12 times that of India, while India’s aid receipt per capita is 0.056 times that of Grenada.

Finally, and confining ourselves to the set of poor countries, it is instructive to look at the pattern of actual shares bi in aid receipts in relation to the corresponding normative shares di. Table 7 (p 62) presents the relevant information. A generous margin of deviation from unity of the actual-to-normative-share ratio would be the interval [0.5,1.5]. As it happens, and as Table 7 reveals, only 20 of the 65 poor countries fall within this band. For the rest, we have a wide range of variation in the ratio of actual aid share to normative aid share, with the polarities described by Congo (110) at one end of the spectrum, and Uzbekistan (0.09) at the other end.

In the light of the preceding discussion, we are led to our Fifth Observation: The relative receipts of aid bear little relation to the relative needs of beneficiary countries.

Allowing for Intra-Country Inequality in the Distribution of Income

It should be stressed that even if the total quantum of aid is equal to the aggregate poverty deficit D, and if actual shares in aid disbursement and aid receipt are precisely equal to their respective normative shares, aid would be a successful means of eradicating global income poverty only on the assumption that there is no

Table 6: Aid, Income and Poverty Deficit Data for All Aid Receiving Countries (2005) Table 6: (Continued)

No Country Per Capita Per Capita Population Total Poverty Aid No Country Per Capita Per Capita Population Total Poverty Aid

GDP ($) Deficit (million) Deficit Receipts GDP ($) Deficit (million) Deficit Receipts

(Poverty Line - (Population ($ million) (Poverty Line - (Population ($ million)

Per Capita Times Per Capita Per Capita Times Per Capita

GDP in $) Poverty Deficit GDP in $) Poverty Deficit

in $ million) in $ million)

(A) Non-Poor Countries 54 Tonga 2090 -690 0.0957 -66.03 31.8 1 Saudi Arabia 13399 -11999 23.1211 -277430.42 26.3 55 Angola 2058 -658 15.9378 -10487.07 441.8 2 Barbados 11465 -10065 0.2704 -2721.46 -2.1 56 Cape Verde 1940 -540 0.5155 -278.35 160.6 3 Trinidad and Tobago 11000 -9600 1.3091 -12567.27 -2.1 57 Ukraine 1761 -361 47.0755 -16994.26 409.6 4 Antigua and Barbuda 10578 -9178 0.0851 -780.88 7.2 58 China 1713 -313 1304.3199 -408252.13 1756.9 5 Oman 9584 -8184 2.5355 -20750.33 30.7 59 Morocco 1711 -311 30.1578 -9379.08 651.8 6 Saint Kitts and Nevis 9438 -8038 0.0530 -425.83 3.5 60 Turkmenistan 1669 -269 4.8532 -1305.51 28.3 7 Croatia 8666 -7266 4.4426 -32280.29 125.4 61 Armenia 1625 -225 3.0154 -678.46 193.3 8 Seychelles 8209 -6809 0.0853 -580.62 18.8 62 Azerbaijan 1498 -98 8.4112 -824.30 223.4

9 Mexico 7454 -6054 103.0856 -624080.17 189.4 63 Georgia 1429 -29 4.4787 -129.88 309.8 10 Chile 7073 -5673 16.2873 -92397.79 151.7 Aggregate 2329.9 11660.1 11 Libyan Arab Jamahiriya 6621 -5221 5.8601 -30595.80 24.4 (B) Poor Countries

1 Syrian Arab Republic 1382 18 19.0304 342.55 77.9

12 Equatorial Guinea 6416 -5016 0.4988 -2501.75 39.0 2 Bhutan 1325 75 0.6038 45.28 90.0

13 Lebanon 6135 -4735 3.5697 -16902.44 243.0 3 Indonesia 1302 98 220.5837 21617.20 2523.5

14 Botswana 5846 -4446 1.7619 -7833.36 70.9 4 Congo 1273 127 4.0063 508.80 1448.9

15 Gabon 5821 -4421 1.3915 -6151.88 53.9 16 Venezuela 5 Paraguay 1242 158 5.8776 928.66 51.1

(Bolivarian Republic of) 5275 -3875 26.5782 -102990.52 48.7 6 Egypt 1207 193 74.0679 14295.11 925.9 17 Malaysia 5142 -3742 25.3403 -94823.53 31.6 7 Sri Lanka 1196 204 19.6488 4008.36 1189.3 18 South Africa 5109 -3709 46.8781 -173870.72 700.0 8 Philippines 1192 208 83.0537 17275.17 561.8 19 Mauritius 5059 -3659 1.2453 -4556.57 31.9 9 Honduras 1151 249 7.2111 1795.57 680.8 20 Turkey 5030 -3630 72.0676 -261605.37 464.0 10 Occupied Palestinian Territories 1107 293 3.6134 1058.72 1101.6 21 Saint Lucia 5007 -3607 0.1598 -576.31 11.1 11 Guyana 1048 352 0.7634 268.70 136.8 22 Uruguay 4848 -3448 3.4653 -11948.51 14.6 12 Cameroon 1034 366 16.3443 5982.01 413.8 23 Panama 4786 -3386 3.2386 -10965.94 19.5 13 Bolivia 1017 383 9.1445 3502.36 582.9 24 Argentina 4728 -3328 38.7479 -128952.96 99.7 14 Nicaragua 954 446 5.1363 2290.78 740.1 25 Costa Rica 4627 -3227 4.3225 -13948.56 29.5 15 Côte d’Ivoire 900 500 18.1111 9055.56 119.1 26 Grenada 4451 -3051 0.1123 -342.73 44.9 16 Djibouti 894 506 0.7830 396.20 78.6 27 Brazil 4271 -2871 186.3966 -535144.72 191.9 17 Papua New Guinea 840 560 5.8333 3266.67 266.1 28 Dominica 3938 -2538 0.0762 -193.35 15.2 18 Lesotho 808 592 1.8564 1099.01 68.8 29 Belize 3786 -2386 0.2905 -693.24 12.9 19 Sudan 760 640 36.1842 23157.89 1828.6 30 Kazakhstan 3772 -2372 15.1379 -35907.00 229.2 20 Nigeria 752 648 131.6489 85308.51 6437.3 31 Saint Vincent and the

21 Mongolia 736 664 2.5815 1714.13 211.9

Grenadines 3612 -2212 0.1107 -244.96 4.9

22 India 736 664 1094.7011 726881.52 1724.1 32 Jamaica 3607 -2207 2.6615 -5873.91 35.7

23 Yemen 718 682 21.0306 14342.90 335.9 33 Dominican Republic 3317 -1917 8.8936 -17048.99 77.0

24 Pakistan 711 689 155.6962 107274.68 1666.5 34 Fiji 3219 -1819 0.8388 -1525.72 64.0

25 Senegal 707 693 11.5983 8037.62 689.3 35 Algeria 3112 -1712 32.8728 -56278.15 370.6

26 Moldova 694 706 4.1787 2950.14 191.8 36 Belarus 3024 -1624 9.7884 -15896.30 53.8

27 Comoros 645 755 0.6202 468.22 25.2 37 Namibia 3016 -1616 2.0225 -3268.44 123.4

28 Viet Nam 631 769 83.0428 63859.90 1904.9 38 Suriname 2986 -1586 0.4354 -690.49 44.0

29 Solomon Islands 624 776 0.4808 373.08 198.2 39 Tunisia 2860 -1460 10.0350 -14651.05 376.5

30 Zambia 623 777 11.7175 9104.49 945.0 40 Peru 2838 -1438 27.9774 -40231.57 397.8

31 Mauritania 603 797 3.1509 2511.28 190.4 41 Macedonia (TFYR) 2835 -1435 2.0459 -2935.80 230.3

32 Chad 561 839 9.8039 8225.49 379.8 42 Iran (Islamic Republic of) 2781 -1381 68.2488 -94251.64 104.0

33 Kenya 547 853 34.1865 29161.06 768.3 43 Ecuador 2758 -1358 13.2342 -17972.08 209.5

34 Uzbekistan 533 867 26.2664 22772.98 172.3 44 Thailand 2750 -1350 64.2182 -86694.55 -171.1

35 Benin 508 892 8.4646 7550.39 349.1 45 Colombia 2682 -1282 45.6003 -58459.58 511.1

36 Haiti 500 900 8.6000 7740.00 515.0 46 Albania 2678 -1278 3.1367 -4008.66 318.7

37 Lao People’s Democratic 47 Bosnia and Herzegovina 2546 -1146 3.8885 -4456.17 546.1 Republic 485 915 5.9794 5471.13 295.7 48 Guatemala 2517 -1117 12.5944 -14067.90 253.6 38 Ghana 485 915 22.0619 20186.60 1119.9 49 El Salvador 2467 -1067 6.8910 -7352.66 199.4 39 Kyrgyzstan 475 925 5.0526 4673.68 268.5 50 Swaziland 2414 -1014 1.1185 -1134.13 46.0 40 Sao Tome and Principe 451 949 0.2217 210.42 31.9 51 Maldives 2326 -926 0.3439 -318.49 66.8 41 Cambodia 440 960 14.0909 13527.27 537.8 52 Jordan 2323 -923 5.4671 -5046.10 622.0 42 Bangladesh 423 977 141.8440 138581.56 1320.5 53 Samoa 2184 -784 0.1832 -143.59 44.0 43 Mali 392 1008 13.5204 13628.57 691.5

NOVEMBER 15, 2008

Table 6: (Continued)

No Country Per Capita Per Capita Population Total Poverty Aid GDP ($) Deficit (million) Deficit Receipts

(Poverty Line - (Population ($ million) Per Capita Times Per Capita GDP in $) Poverty Deficit

in $ million)

44 Burkina Faso 391 1009 13.2992 13418.93 659.6
45 Togo 358 1042 6.1453 6403.35 86.7
46 Timor-Leste 358 1042 0.8380 873.18 184.7
47 Tajikistan 355 1045 6.4789 6770.42 241.4
48 Guinea 350 1050 9.4286 9900.00 182.1
49 Central African Republic 339 1061 4.1298 4381.71 95.3
50 Mozambique 335 1065 19.7015 20982.09 1285.9
51 Tanzania (United Republic of) 316 1084 38.2911 41507.59 1505.1
52 Gambia 304 1096 1.6447 1802.63 58.2
53 Uganda 303 1097 28.7129 31498.02 1198.0
54 Nepal 272 1128 27.2059 30688.24 427.9
55 Madagascar 271 1129 18.4502 20830.26 929.2
56 Zimbabwe 259 1141 13.1274 14978.38 367.7
57 Niger 244 1156 13.9344 16108.20 515.4
58 Rwanda 238 1162 9.2437 10741.18 576.0
59 Eritrea 220 1180 4.5455 5363.64 355.2
60 Sierra Leone 216 1184 5.5556 6577.78 343.4
61 Guinea-Bissau 190 1210 1.5789 1910.53 79.1
62 Malawi 161 1239 13.0435 16160.87 575.3
63 Ethiopia 157 1243 71.3376 88672.61 1937.3
64 Congo (Democratic
Republic of the) 123 1277 57.7236 73713.01 1827.6
65 Burundi 106 1294 7.5472 9766.04 365.0
Aggregate 2714.4 47652.5

A “poor country” is one with a per capita GDP of less than $1,400. Source: Derived from Table 1 of this paper.

intra-country inequality in the distribution of GDP. As a matter of fact, however, per capita GDP for a poor country could rise through an infusion of aid without making much of an impact on poverty if the aid is not employed appropriately to address preexisting domestic inequality and injustice, or is employed in projects – controlled by domestic dictators, for instance – which can exacerbate the mal-distribution of income. By looking only at GDP per capita, one overlooks the problem of targeting aid within the recipient country, and therefore the problem of how to persuade a beneficiary country to attend to a poverty-eradicating distribution of the aid on offer to it – a problem which has often been held out as a reason for the reluctance of wealthy countries to increase the size of overseas development assistance. (This complication has been brought to the author’s attention in p ersonal communication by Thomas Pogge.)

One possible way of addressing this complication might be to conceive of an institutional setting in which a special international agency, as prescribed by Milanovic (2006), is fully entrusted with the task of intra-country allocation of aid – even at the apparent cost of interfering with national sovereignty. At an analytical level, the problem of aid allocation can be viewed as a two-stage problem. In the first stage, the concern is with the determination of the country-specific levels of aid provision to be made. A possible formula is the “proportionality rule” we have already considered. If B is the total quantum of aid available, if D is the aggregate poverty deficit of all the poor countries taken together, and if Di is the poverty deficit of the ith poorest country,

Economic & Political Weekly

EPW
NOVEMBER 15, 2008

then the proportionality rule will dictate that the allocation to the ith poorest country which will minimise the poverty index P0.5 is given by B*i (Di/D)B.

At the second stage of the allocation problem, one can determine the intra-country allocation of the optimal budget B*i for the ith poorest country. At this stage of the problem, one can relax the assumption of an equal within-country distribution of incomes, and consider the actual (unequal) distribution that obtains. The problem of within-country allocation can now again be set up as a programming exercise. Indeed, and by way of example, the problem to be solved at the intra-national level can be formulated in terms which are formally identical to the way in which the problem was formulated, in the first stage, at the international level. That is to say, if Pi0.5 is the poverty level, as measured by the index, P0.5, of the ith poorest country, the objective could be to minimise Pi0.5 subject to the constraints (a) that the sum of aid transfers to country i’s citizens does not exceed the country’s budgeted outlay B*i, (b) that no person receives aid in excess of her poverty deficit, (c) that aid transfers are always nonnegative, and (d) that for every pair of poor individuals the share of the poorer person in the combined poverty deficit of the pair should not become larger after the aid transfers have been made. The solution to this problem will, of course, again be the proportionality rule, in a within-country context: no non-poor person will receive any aid, while each poor person will receive aid in proportion to her share in the aggregate poverty shortfall of all the poor persons taken together. The usual problems of information and enforcement cannot, of course, be eliminated in this imperfect world, but presumably the problem of a deliberate, centrally-sponsored sub-optimal targeting of benefits can be sought to be minimised.

Alternatively, or additionally, one could favour an independent international agency which imposes severe conditionalities on the appropriate internal directing of aid. If the Bretton Woods agencies could successfully impose conditionalities on lending to sustain structural adjustment, it is not clear why conditionalities cannot be imposed – or will not be accepted, especially by the “radical” constituency of commentators – in the good cause of sensible and scrupulous within-country aid allocation. It would be hard to deny the case for a stringent monitoring of where the aid goes, and for making subsequent rounds of aid severely contingent on past performance. Briefly, the problem of fair and rational intra-country aid allocation need not, in principle, jeopardise the prospect of fair and rational inter-country aid allocation.

Having said this, and so that one does not veer over wholly to the other side, it would be fair to note that even if aid is abused, this should not entail any overwhelming financial cost for the donor countries: possibly more wheat is dumped in the ocean to keep international prices up than is allocated to aid. In other words, while it is undeniable that some countries would certainly hesitate to offer aid when it simply ends up financing dictators or corrupt governments, it is also true that for many other countries this serves as an alibi for plain flintiness and ungenerosity (especially considering how much is spent in keeping various corrupt regimes going). As Pogge (2003) points out, an important

Table 7: Actual and Normative Aid Shares of Aid-Receiving Poor Countries (2005) Table 7: (Continued)
No Country Total Poverty Aid Actual Share Normative Actual No Country Total Poverty Aid Actual Share Normative Actual
Deficit Receipt in Aid Share in Aid Aid Deficit Receipt in Aid Share in Aid Aid
(from Table 6) ($ million) Received Received Share/ (from Table 6) ($ million) Received Received Share/
in $ million (Aid Received/ (Total Normative in $ million (Aid Received/ (Total Normative
Aggregate Deficit/ Aid Share Aggregate Deficit/ Aid Share
Aid Received) Aggregate Aid Received) Aggregate
(%) Total Deficit) (%) Total Deficit)
(%) (%)
1 Congo 508.80 1448.90 3.04 0.03 109.87 35 Sierra Leone 6577.78 343.40 0.72 0.36 2.01
2 Bhutan 45.28 90.00 0.19 0.00 76.68 36 Mali 13628.57 691.50 1.45 0.74 1.96
3 Occupied Palestinian Territories 1058.72 1101.60 2.31 0.06 40.14 37 Burkina Faso 13418.93 659.60 1.38 0.73 1.90
4 Solomon Islands 373.08 198.20 0.42 0.02 20.50 38 Benin 7550.39 349.10 0.73 0.41 1.78
5 Guyana 268.70 136.80 0.29 0.01 19.64 39 Chad 8225.49 379.80 0.80 0.45 1.78
6 Honduras 1795.57 680.80 1.43 0.10 14.63 40 Madagascar 20830.26 929.20 1.95 1.13 1.72
7 Nicaragua 2290.78 740.10 1.55 0.12 12.46 41 Guinea-Bissau 1910.53 79.10 0.17 0.10 1.60
8 Sri Lanka 4008.36 1189.30 2.50 0.22 11.45 42 Cambodia 13527.27 537.80 1.13 0.74 1.53
9 Syrian Arab Republic 342.55 77.90 0.16 0.02 8.77 43 Uganda 31498.02 1198.00 2.51 1.71 1.47
10 Timor-Leste 873.18 184.70 0.39 0.05 8.16 44 Burundi 9766.04 365.00 0.77 0.53 1.44
11 Djibouti 396.20 78.60 0.16 0.02 7.65 45 Tanzania (United Republic of) 41507.59 1505.10 3.16 2.26 1.40
12 Bolivia 3502.36 582.90 1.22 0.19 6.42 46 Tajikistan 6770.42 241.40 0.51 0.37 1.38
13 Sao Tome and Principe 210.42 31.90 0.07 0.01 5.85 47 Malawi 16160.87 575.30 1.21 0.88 1.37
14 Mongolia 1714.13 211.90 0.44 0.09 4.77 48 Philippines 17275.17 561.80 1.18 0.94 1.25
15 Indonesia 21617.20 2523.50 5.30 1.18 4.50 49 Gambia 1802.63 58.20 0.12 0.10 1.25
16 Zambia 9104.49 945.00 1.98 0.50 4.00 50 Niger 16108.20 515.40 1.08 0.88 1.23
17 Senegal 8037.62 689.30 1.45 0.44 3.31 51 Viet Nam 63859.90 1904.90 4.00 3.47 1.15
18 Papua New Guinea 3266.67 266.10 0.56 0.18 3.14 52 Kenya 29161.06 768.30 1.61 1.59 1.02
19 Sudan 23157.89 1828.60 3.84 1.26 3.05 53 Congo (Democratic
20 Mauritania 2511.28 190.40 0.40 0.14 2.93 Republic of the) 73713.01 1827.60 3.84 4.01 0.96
21 Nigeria 85308.51 6437.30 13.51 4.64 2.91 54 Zimbabwe 14978.38 367.70 0.77 0.81 0.95
22 Cameroon 5982.01 413.80 0.87 0.33 2.67 55 Yemen 14342.90 335.90 0.70 0.78 0.90
23 Haiti 7740.00 515.00 1.08 0.42 2.57 56 Ethiopia 88672.61 1937.30 4.07 4.82 0.84
24 Eritrea 5363.64 355.20 0.75 0.29 2.56 57 Central African Republic 4381.71 95.30 0.20 0.24 0.84
25 Moldova 2950.14 191.80 0.40 0.16 2.51 58 Guinea 9900.00 182.10 0.38 0.54 0.71
26 Egypt 14295.11 925.90 1.94 0.78 2.50 59 Pakistan 107274.68 1666.50 3.50 5.83 0.60
27 Lesotho 1099.01 68.80 0.14 0.06 2.42 60 Nepal 30688.24 427.90 0.90 1.67 0.54
28 Mozambique 20982.09 1285.90 2.70 1.14 2.36 61 Togo 6403.35 86.70 0.18 0.35 0.52
29 Kyrgyzstan 4673.68 268.50 0.56 0.25 2.22 62 Côte d’Ivoire 9055.56 119.10 0.25 0.49 0.51
30 Ghana 20186.60 1119.90 2.35 1.10 2.14 63 Bangladesh 138581.56 1320.50 2.77 7.54 0.37
31 Paraguay 928.66 51.10 0.11 0.05 2.12 64 Uzbekistan 22772.98 172.30 0.36 1.24 0.29
32 Lao People’s Democratic Republic 5471.13 295.70 0.62 0.30 2.09 65 India 726881.52 1724.10 3.62 39.54 0.09
33 Comoros 468.22 25.20 0.05 0.03 2.08 Total 1838498.89 47652.50 100.00 100.00
34 Rwanda 10741.18 576.00 1.21 0.58 2.07 A “poor country” is one with a per capita GDP of less than $1,400. Source: Derived from Tables 1 and 2 of this paper.
c omponent of what he calls the “Purely Domestic Poverty not see themselves as being under a moral obligation to assist
T hesis” (PDPT) is the view that it is not so much an unfair global poor countries. In respect of the last complication, an argu
order as the prevalence of corrupt and brutal governments and ment that is often held out is that poor countries do not have a
elites in poor countries which accounts for so much global right to aid. Even setting aside the counter-view that aid is no
d eprivation: a comforting construction that omits to acknowl more than a reparation for historical and contemporary
edge the grave complicity of the world order itself in the corrup wrongs such as colonialism and unfair trade practices, it is
tion and brutality of the systems of power that prevail in many worthwhile to remind oneself of T immermann’s (2004) obser
poor countries. vation: “Rights imply duties, but there can be duties without
corresponding rights”.
Concluding Observations Despite all the direct plainness of the observations made
As threatened at the outset, this has been an unsubtle paper. e arlier – namely that there is a great deal of poverty in the world,
There are a number of complications8 we have not taken on that the quantum of aid available is very small in relation to
board: the possibility that income is not the only indicator of the magnitude of the poverty problem, that the redistributive
deprivation; the possibility that there are inter-country varia effort that would be required to eradicate poverty is quite small,
tions in the ability to effectively “absorb” aid; the possibility that there is little relationship between actual and normative aid
that aid allocations are sometimes influenced by the historical shares at the dispensing end, and similarly little relationship
specificity of events like colonialism which mediate bilateral between actual and normative aid shares at the receiving end –
relations; and, of course, the possibility that rich countries do the orders of magnitude reviewed do not suggest that a
62 NOVEMBER 15, 2008 Economic & Political Weekly
EPW

greater accommodation of complexity will make substantial largely be a matter of arranging the deck-chairs, by remote condents in the truth of these observations. The justification for a trol, on someone else’s Titanic. There is no doubt that the “moral certain absence of subtlety derives from the persistence of the mathematics” of this paper lack finesse, but it is to be hoped that truths it reflects. Fussy sophistication in the discourse on aid they at least assist in preventing one from losing one’s way in

which does not directly address these stubborn truths could exquisite by-lanes.

Notes t axation are just sufficient to bridge the aggregate
1 Let z stand for the international poverty line, xi for the per capita GDP of the ith poorest country, and pi for its population. There are m countries, 6 poverty deficit D. That is, for all i∈A: Si ≡ pi(xi–14,000), and S = Σ Si . i∈A
2 m and total population is p (=Σ pi). We let N stand i=l for the set of all countries, and Q for the of poor countries, where Q ≡ {i∈N| xi < z}. The cardinality of N is m, and that of Q is q. The global dis tribution of income is represented by the partitioned vector (x;p) = (x1,…,xm;p1,…,pm), with the countries indexed in non-decreasing order of per capita GDP, that is, such that xi ≤ xi+1, for all i = 1,…,m-1. For every country i, (xi,pi) represents a combination of the per capita GDP and population size of that country. For every partitioned vector (x;p) and poverty line z, the Pα family of poverty indices is given by: 7 All of this can be stated more formally as follows. Let j and k be two poor countries with aggregate poverty deficits Dj and Dk respectively. Let σj,k≡ Dj/(Dj + Dk), i e, σj,k is the share of j in the combined poverty deficits of j and k. Obviously, if Dj ≥ Dk, then σj,k ≥ ½. Let σ′ j,k be the value of σj,k after the aid transfers have been made. Then, a preference for equality in aid distribution is compatible with the requirement that if j is the country with the larger poverty deficit, the index of pairwise inequality σj,k should not become larger after the distribution, that is, we would require that σ′ j,k ≤ σj,k. Let poverty be measured by the index P0.5. Then, the aid allocation problem can be set up formally as a p rogramming exercise of the following type: (*) Minimise P0.5(D1–B1,…, Dq – Bq;z) = (1/pz0.5)
Pα((x;p);z) = (1/p) Σ pi [( z–xi)/z]α, α ≥ o. Σ (Di–Bi)0.5
i∈Q i∈Q
3 To see what is involved, note first that a reduction {Bi}i∈Q
in a poor person’s income will leave the number of subject to
poor persons, and therefore the headcount ratio (a) Σ Bi ≤ B;
(Po), unchanged; however, the average income of i∈Q
the poor will decline, and therefore the per capita (b) Bi ≤ Di ∀ i ∈ Q;
income-gap ratio (P1) will rise, which explains (c) Bi ≥ 0 ∀ i ∈ Q; and
why Po fails, and P1 satisfies, the monotonicity (d) ∀ j,k∈Q, if σj,k ≥ ½, then σ′j,k ≤ σj,k.
condition. A transfer of income from a poor person to a poorer one will leave both the headcount ratio and the income-gap ratio (or proportionate shortfall of the average income of the poor from the poverty line) unchanged (notice that a pure redistribution is not going to affect the average income of the poor), which is why the transfer axiom is violated by both Po and P1 (which latter is just the product of the headcount and the incomegap ratios). If H is the headcount ratio, I the income-gap ratio, and C2 the squared coefficient variation in the distribution of poor incomes – this a well-known summary measure of relative i nequality – then it can be shown that the poverty index P2 can be written as: P2 = H[I2 + (1–I)2C2]; a progressive income transfer will reduce the value of C2, and therefore of P2. This poverty measure, consequently, satisfies the transfer axiom. The parameter α in the Pα family of i ndices is a register of “equity-consciousness”: the p overty measure increases in its distributionsensitivity with an increase in α until, in the limit, as α goes to infinity, Pα converges on a sort of “Rawlsian” measure such that distributions are ranked solely by the poverty of the poorest individual. 8 Subramanian (2006) has shown that the optimal solution to problem (*) is given by Bi * = diB, where di ≡ Di/D, ∀ i∈ Q. A particular complication not considered in this paper relates to the host of economy-wide macro effects which policies of taxation and transfer can have on both donor and recipient countries. Ram Reddy, in personal correspondence, has pointed out that aid mobilised through income taxation can result in a sort of forced saving for the donor country, and the consequentially altered aggregate saving rate can affect its rate of growth of national income and therefore the dynamics of the patterns and quantum of aid flows. Depending on how aid is utilised in the recipient country, there could be smaller or larger multiplier effects on income-generation with implications for the magnitude of future flows of aid. Taxation and transfer can also have incentive effects at both the giving and receiving ends of aid, and these are not easy to predict. These, and related issues, are clearly important ones to address in any c om prehensive treatment of the subject of aid, but they are somewhat outside the purview of the rather more limited objectives of the present paper.

4 To put it precisely: for every i∈Q, Di ≡ pi(z-xi), and D = Σ Di.

i∈Q

5 Formally, let x* be a level of income, and q* the poorest of the rich countries, such that these are determined through the following equation:

m

Σ pi(xi–x*) = D.

i=q*

Then, the optimal tax scehdule {T*i}i∈N, as described in the text, is given by: T*i = 0 ∀ i∈{1,…,q*–1};

= pi(xi – x*) ∀ i∈{q*,…,m}. Under the solution described above, the per capita incomes of the richest (m-q*) countries are equalised, through reduction, to a level of income x* such that, the proceeds from this scheme of

References

Anand, S (1983): Inequality and Poverty in Malaysia: Measurement and Decomposition, Oxford University Press, New York.

Bourguignon, F and G S Fields (1990): ‘Poverty M easures and Anti-Poverty Policy’, Recherches Economiques de Louvain, 56: 409-27.

Foster, J E, J Greer and E Thorbecke (1984): ‘A Class of Decomposable Poverty Measures’, Econometrica,

52: 761-66. Frankfurt, H G (1987): ‘Equality as a Moral Ideal’, E thics, 98: 21-43. Gangopadhyay, S and S Subramanian (1992): ‘Optimal Budgetary Intervention in Poverty Alleviation Programmes’ in S Subramanian (ed), Themes in Development Economics: Essays in Honour of Malcolm Adiseshiah, Oxford University, Delhi. Jayaraj, D and S Subramanian (1996): ‘Poverty Eradication through Redistributive Taxation: Some Elementary Considerations’, Review of Development and Change, 1(1): 73-84. Milanovic, B (2006): ‘Global Income Inequality: What It Is and Why It Matters’, DESA Working Paper No 26, United Nations Department of Social and E conomic Affairs, New York, USA. Parfit, D (1997): ‘Equality and Priority’, Ratio (new series), December : 202-221. Pogge, T W (2003): “ ‘Assisting’ the Global Poor”, available at http://www.etikk.no/globaljustice/ papers/GJ2003_Thomas_Pogge_Assisting_the_ Global_Poor.DOC. (Downloaded April 7th 2008.)

Reddy, S G (2007): ‘International Debt: The C onstructive Implications of Some Moral Mathematics’, Ethics and International Affairs, 21(1):

33-48. Subramanian, S (2003): ‘Aspects of Global Deprivation and Disparity: A Child’s Guide to Some S imple-Minded Arithmetic’ in F Carlucci and F Marzano (eds): Poverty, Growth and Welfare in the World Economy in the 21st Century, Peter Lang, Bern.

  • (2006): ‘Poverty Measures and Anti-Poverty Policy under an Egalitarian Constraint’ in S Subramanian, Rights, Deprivation, and Disparity: Essays in Concepts and Measurement, Oxford U niversity Press, Delhi.
  • (2007): ‘International Aid in Light of Global P overty and Inequality’, UNU-WIDER Research Paper No 2007/31, Helsinki.
  • Timmermann, J (2004): ‘The Individualist Lottery: How People Count, But Not Their Numbers’, A nalysis, 64: 106-112.

    Unbound Back Volumes of Economic and Political Weekly from 1976 to 2007 are available.

    Write to: Circulation Department,

    Economic and Political Weekly

    320-321, A to Z Industrial Estate, Ganpatrao Kadam Marg, Lower Parel, Mumbai 400 013. Circulation@epw.in

    Economic & Political Weekly

    EPW
    NOVEMBER 15, 2008

    Dear Reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Back to Top