2001.04.26 全球收入分配不公

shiyi18

BY INVITATION
Winners and losers
The global distribution of income is becoming ever more unequal. That should be a matter of greater concern than it is, argues Robert Wade


Apr 26th 2001

AP



ANYBODY interested in the wealth and poverty of nations must be interested in what is happening to the global distribution of income, one would suppose. A lot turns on the question. If the world's income distribution has become more equal in the past few decades, this would be powerful evidence that globalisation works to the benefit of all. It would give developing countries good reason to integrate their economies closely into the world economy, as the IMF and the World Bank—and their mostly rich-country shareholders—urge them to do. It would answer some of the fears of the anti-globalisation protesters. And it would help to settle a crucial and long-standing disagreement in economic theory, between the orthodox view that economic growth naturally delivers “convergence” of rich and poor countries, and alternative theories which, for one reason or another, say the opposite.


Despite its importance, this issue has received rather little attention within the fields of development studies, international relations and (until very recently) international economics. Neither the World Bank nor the IMF has devoted significant resources to studying it. Many analysts apparently take it for granted that global inequality is falling. Others think it sufficient to focus on poverty, and ignore inequality as such. Both these views need to be challenged. New evidence suggests that global inequality is worsening rapidly. There are good reasons to worry about that trend, quite apart from what it implies about the extent of world poverty.

Distributing the spoils
What exactly does “world income distribution” mean? This article is concerned with distribution among the planet's 6.2 billion people, regardless of country or region. World income distribution can be thought of as the combination of (a) the internal income distributions for all the countries and (b) the distribution of average incomes across countries. Most of the inequality in world incomes reflects inequality in country averages rather than inequality within countries.



Chart 1 shows the distribution of the world's population by average income of each country (using compatible data from 1993, the most recent year available). Income is measured in terms of purchasing power over comparable bundles of goods and services, or purchasing-power parity (PPP), rather than in terms of actual exchange rates. China and India, with between them almost 40% of world population, are each divided into urban and rural sectors and treated as four separate countries.


The distribution has two poles. One, at the bottom end, is at an average income of less than $1,500 a year. It contains the populations of most of Africa, India, Indonesia and rural China. At the other pole, with average PPP incomes of more than $11,500, are the United States, Japan, Germany, France, Britain and Italy. Some of the space between $1,500 and $11,500 is occupied by countries such as urban China, Russia and Mexico. But notice the strange “missing middle”: relatively few people live in countries with average PPP incomes that fall between $5,000 and $11,500. If incomes were measured using actual exchange rates, the range from poorest to richest would be much larger.

Nobody denies that world income distribution became vastly more unequal after the industrial revolution. On this timescale divergence dominates, big time. But what happened in the past three or four decades?

Having ignored world income distribution for decades, international economics has lately seen a burst of interest. But the statistical difficulties are so formidable that the debate has so far revolved around questions of econometric technique. Standing back from the fray, we can see that much of the controversy concerns how to compare income in different countries.

The answer to what is happening to world income distribution turns out to depend heavily on whether countries are weighted by population, and whether income in different countries is measured in PPP terms or by using actual exchange rates. These two criteria can be set out in a matrix of four cells, with countries treated equally or weighted by population on one axis, and income measured in current exchange rates or in PPP terms on the other. Table 2 summarises the findings of the existing literature, dividing the studies among the four cells.



If countries are treated equally (not weighted by population) and average income is measured in PPP terms, most studies find that world income distribution has become more unequal in the past few decades. If countries are weighted by their populations (so that China's change in average income counts for many times more than Uganda's), the world's PPP income distribution over recent decades shows little change.

It may seem obvious that one should weight countries by population, giving every individual in the world equal weight. For some purposes, though, it may make sense to weight countries equally. To test the growth theories mentioned earlier, or to assess the effectiveness of policy advice on how to raise growth rates, one can treat each country as a laboratory trial—an observation of a set of policies, institutions and resources—and take the average over all the trials. Likewise in the context of understanding co-operation in multilateral organisations such as the United Nations, where small-population states have more of a voice than their share of world population would warrant, it may make sense to look at world income distribution with countries weighted equally.

What about the second measurement issue: actual exchange rates or PPP? When incomes in different countries are compared using actual exchange rates, the evidence shows that world income distribution has become much more unequal over the past several decades, and that inequality accelerated during the 1980s, whether countries are treated equally or weighted by population. When incomes are compared using PPP calculations, the degree of inequality shrinks, and so does the rate of widening.


Many assume that PPP measures are always superior. Certainly the exchange rate is a flawed measure of purchasing power: it fails to reflect the large amount of non-monetary exchange in developing countries, or money payment for services that are not subject to international competition. Also, exchange rates are much affected by capital flows and by monetary policy.

But PPP measures have their drawbacks as well. Different methods of measuring purchasing-power parity, all plausible in themselves, yield different results. Comprehensive estimates of PPP incomes for developing countries, based on actual data on prices of comparable goods and services, go back only to the 1970s. This makes longer-run analysis difficult. Finally, incomes based on actual exchange rates may be a better measure than PPP of relative national power and national modernity—matters of more interest to sociologists and political scientists, no doubt, than to economists.

In any event, the bulk of the evidence on trends in world income distribution runs against the claim that world income inequality has fallen sharply in the past half-century and still faster in the past quarter-century. None of the four cells in the table supports this idea.

From poorer to richer
So much for the existing research. However, we can now go further. Two very recent studies based on new data challenge the finding in the fourth cell that world PPP-income distribution weighted by countries' population shows little change over the past few decades. The new studies show, on the contrary, a rapid rise in inequality.

These new studies differ from the others in being based solely on household income and expenditure surveys. The earlier ones either used average GDP, ignoring inequality within each country, or used indirect methods to estimate within-country inequality, including production surveys and revenue surveys, which typically miss important components of household incomes. Branko Milanovic at the World Bank assembled the database, using the Bank's formidable statistical organisation to obtain household survey data from just about all the Bank's members, covering 85% of the world's population, for the years 1988 and 1993. The result is probably the most reliable data set on world income distribution.

Then Mr Milanovic computed the Gini coefficient for world income distribution, combining within-country inequality and between-country inequality, and measuring it in PPP terms. (The Gini coefficient is a commonly used measure of inequality: 0 signifies perfect equality, 100 means that one person holds all the income.) The results are startling. World inequality increased from a Gini coefficient of 62.5 in 1988 to 66.0 in 1993. This is a faster rate of increase of inequality than that experienced within the United States and Britain during the 1980s. By 1993 an American on the average income of the poorest 10% of the population was better off than two-thirds of the world's people.



The other new study, by Yuri Dikhanov and Michael Ward, uses the same data set with a different methodology. It confirms that world income distribution became markedly more unequal between 1988 and 1993. Like the Milanovic study, it finds that the Gini coefficient increased by about 6%. It finds, further, that the share of world income going to the poorest 10% of the world's population fell by over a quarter, whereas the share of the richest 10% rose by 8%. The richest 10% pulled away from the median, while the poorest 10% fell away from the median, falling absolutely by a large amount. In short, we have to revise cell 4. World PPP income distribution with countries weighted by population (and China and India split into urban and rural) became “much more unequal” between 1988 and 1993 (see table 3).

Why has global inequality increased? The answer is in four parts: (1) faster economic growth in developed OECD countries than developing countries as a group; (2) faster population growth in developing countries than in OECD countries; (3) slow growth of output in rural China, rural India, and Africa; and (4) rapidly widening output and income differences between urban China on the one hand, and rural China and rural India on the other. The income of urban China grew very fast during 1988-93, which reduced the gap between China's average income and that of the middle-income and rich countries, and so reduced the world Gini coefficient; but the widening gaps between rural China and urban China and between urban China and rural India increased world inequality by even more.

These trends in turn have deeper causes. Technological change and financial liberalisation result in a disproportionately fast increase in the number of households at the extreme rich end, without shrinking the distribution at the poor end. Population growth, meanwhile, adds disproportionately to numbers at the poor end. These deep causes yield an important intermediate cause that makes things worse: the prices of industrial goods and services exported from high-income countries are increasing faster than the prices of goods and services exported by low-income countries, and much faster than the prices of goods and services produced in low-income countries that do little international trade.

These price trends mean that the majority of the population of poor countries are able to buy fewer and fewer of the goods and services that enter into the consumption patterns of rich-country populations. The poorer countries and the poorer two-thirds of the world's population therefore suffer a double marginalisation: once through incomes, again through prices. Hence the figures in table 3, which show that the gap between the richest 10% of the world's population and the median is widening, and the gap between the median and the poorest 10% is also widening.

Polarised zones
Income divergence helps to explain another kind of polarisation taking place in the world system, between a zone of peace and a zone of turmoil. The regions of the wealthy pole in chart 1 show a strengthening republican order of economic growth and liberal tolerance (except towards immigrants), with technological innovation able to substitute for depleting natural capital. The regions of the lower- and middle-income pole contain many states whose capacity to govern is stagnant or eroding, mainly in Africa, the Middle East, Central Asia, Russia, and parts of East Asia. Here, a rising proportion of people find their access to basic necessities restricted at the same time as they see people on television driving Mercedes cars.


The result is a lot of unemployed and angry young people, to whom new information technologies have given the means to threaten the stability of the societies they live in and even to threaten social stability in countries of the wealthy zone. Economic growth in these countries often depletes natural capital and therefore future potential. More and more people see migration to the wealthy zone as their only salvation.

It is remarkable how unconcerned the World Bank, the IMF and other global organisations are about these trends. The Bank's World Development Report for 2000 even said that rising income inequality “should not be seen as negative” if the incomes at the bottom do not fall and the number of people in poverty falls. Such lack of attention shows that to call these world organisations is misleading. They may be world bodies in the sense that almost all states are members, but they think in state-centric rather than global ways. They neglect not only matters of world income distribution, but also world inflation, world exchange rates, and world interest rates; and, in the case of the World Bank, the global environmental issues of the oceans, the atmosphere, and nuclear waste.

It is striking that most of the organised opposition to more globalisation comes from North America, Western Europe and Oceania. Why have elites from developing countries for the most part subscribed to the globalisation agenda that western states, businesses, and multilateral organisations have been promoting, if a case can be made that the gains of free markets for goods and capital tend to be concentrated in the top levels of the income distributions of their countries? Why are they doing so little to integrate their economies into the world economy in a strategic way, not open-endedly?

Part of the reason may be that elites in developing countries, like their counterparts in the rich world, are content to believe either that world inequality is falling, or that inequality is good because it is the source of incentives. They, like the multilateral economic organisations (and the reformers of Victorian England), worry about poverty. But they see no link between widening world income distribution and poverty; and they think that poverty can be fixed by providing the poor with welfare and opportunities without changing larger structures like income and asset distributions. Academic analysts have a responsibility to counter the current neglect by analysing the relationship between trends in world income distribution and poverty as a way of getting distribution issues on to the world agenda.

Growing inequality is analogous to global warming. Its effects are diffuse and long-term, and there is always something more pressing to deal with. The question is how much more unequal world income distribution can become before the resulting political instabilities and flows of migrants reach the point of directly harming the well-being of the citizens of the rich world and the stability of their states. Before that point is reached we should mobilise our governments, the multilateral organisations, and international NGOs to establish as an overarching priority a more equal world income distribution—and not just, as now, fewer people in poverty.



Robert Wade is professor of political economy at the London School of Economics and a temporary fellow at the Wissenschaftskolleg in Berlin. He is the author of “Governing the Market”, a study of economic development in East Asia

The author would like to thank Michael Ward and Branko Milanovic for discussions about income distribution. This week's Economics focus comments on this article



应邀参加
赢家和输家
全球的收入分配越来越不平等。罗伯特-韦德认为，这应该是一个比现在更值得关注的问题。


2001年4月26日

AP



任何对国家的财富和贫困感兴趣的人，都会对全球收入分配的情况感兴趣，我们认为。很多事情都取决于这个问题。如果世界的收入分配在过去几十年里变得更加平等，这将是全球化对所有人有利的有力证据。这将使发展中国家有充分的理由将其经济紧密地融入世界经济，正如国际货币基金组织和世界银行--以及它们的大部分富国股东所敦促的那样。它将回应反全球化抗议者的一些担忧。而且，这将有助于解决经济理论中的一个关键和长期的分歧，即正统观点认为经济增长自然会带来富国和穷国的 "趋同"，而其他理论由于这样或那样的原因，则表示相反的观点。


尽管这个问题很重要，但在发展研究、国际关系和（直到最近）国际经济学领域，这个问题却很少得到关注。世界银行和国际货币基金组织都没有投入大量资源来研究它。许多分析家显然认为，全球不平等现象正在减少是理所当然的。其他人则认为只需关注贫困问题，而忽略了不平等问题本身。这两种观点都需要受到挑战。新的证据表明，全球不平等正在迅速恶化。我们有充分的理由担心这种趋势，除了它所暗示的世界贫困的程度之外。

分配战利品
"世界收入分配 "到底是什么意思？本文关注的是地球上62亿人之间的分配，不分国家或地区。世界收入分配可以被认为是（a）所有国家的内部收入分配和（b）各国平均收入分配的结合。世界收入的大部分不平等反映了国家平均水平的不平等，而不是国家内部的不平等。



图1显示了按每个国家的平均收入计算的世界人口分布情况（使用1993年的兼容数据，这是最新的一年）。收入是以可比商品和服务的购买力，或购买力平价（PPP）来衡量的，而不是以实际汇率来衡量的。中国和印度的人口几乎占世界人口的40%，它们各自分为城市和农村部门，并被视为四个独立的国家。


这种分布有两个极点。一个是处于底层的国家，年平均收入不到1500美元。它包含了非洲大部分地区、印度、印度尼西亚和中国农村的人口。另一极是美国、日本、德国、法国、英国和意大利，它们的平均购买力平价收入超过11,500美元。1,500美元和11,500美元之间的一些空间被中国城市、俄罗斯和墨西哥等国家占据。但是请注意奇怪的 "中间缺失"：相对来说，很少有人生活在平均购买力平价收入在5000美元和11500美元之间的国家。如果用实际汇率来衡量收入，从最穷到最富的范围会大得多。

没有人否认，在工业革命之后，世界收入分配变得更加不平等。在这个时间尺度上，分歧占主导地位，很大。但是，在过去的三、四十年里发生了什么？

在忽视世界收入分配几十年后，国际经济学最近出现了兴趣的爆发。但统计上的困难是如此之大，以至于到目前为止，辩论一直围绕着计量经济学技术的问题进行。退一步讲，我们可以看到，大部分的争论是关于如何比较不同国家的收入。

世界收入分配情况的答案在很大程度上取决于各国是否按人口加权，以及不同国家的收入是按购买力平价还是用实际汇率来衡量。这两个标准可以在一个由四个单元组成的矩阵中列出，其中一个轴线上的国家被平等对待或按人口加权，另一个轴线上的收入是以当前汇率或购买力平价衡量的。表2总结了现有文献的研究结果，将这些研究划分为四个单元。



如果各国被平等对待（不按人口加权），平均收入以购买力平价衡量，大多数研究发现世界收入分配在过去几十年中变得更加不平等。如果各国按人口加权（因此，中国平均收入的变化比乌干达的变化多出许多倍），近几十年来世界的购买力平价收入分配没有什么变化。

看起来很明显，我们应该按人口对国家进行加权，给世界上的每一个人以同等的权重。不过，对于某些目的来说，对国家进行同等的加权可能是有意义的。为了检验前面提到的增长理论，或者评估关于如何提高增长率的政策建议的有效性，我们可以把每个国家当作一个实验室试验--对一套政策、机构和资源的观察，并取所有试验的平均值。同样，在理解联合国等多边组织的合作方面，人口少的国家拥有比其在世界人口中所占份额更多的话语权，因此，研究世界收入分配并对各国进行平均加权可能是有意义的。

第二个测量问题是什么：实际汇率还是购买力平价？当使用实际汇率比较不同国家的收入时，证据显示，在过去几十年里，世界收入分配变得更加不平等，而且在20世纪80年代，无论各国被平等对待还是按人口加权，不平等都在加速。当使用购买力平价计算法比较收入时，不平等的程度缩小了，扩大的速度也缩小了。


许多人认为，购买力平价的衡量方法总是优越的。当然，汇率是一个有缺陷的购买力衡量标准：它没有反映出发展中国家大量的非货币交换，或者没有受到国际竞争的服务的货币支付。另外，汇率受资本流动和货币政策的影响很大。

但购买力平价的衡量方法也有其缺点。衡量购买力平价的不同方法，本身都是合理的，但却产生了不同的结果。根据可比商品和服务价格的实际数据，对发展中国家购买力平价收入的全面估计，只能追溯到1970年代。这使得长期分析变得困难。最后，基于实际汇率的收入可能比购买力平价更好地衡量相对国力和国家现代性--毫无疑问，社会学家和政治学家比经济学家更感兴趣。

无论如何，关于世界收入分配趋势的大部分证据都与世界收入不平等在过去半个世纪中急剧下降，在过去四分之一世纪中仍在快速下降的说法相悖。表中的四个单元格都不支持这一观点。

从穷到富
现有的研究就这么多了。然而，我们现在可以更进一步。两项基于新数据的最新研究对第四格中的结论提出了质疑，即按国家人口加权的世界购买力平价收入分配在过去几十年中没有什么变化。新的研究显示，相反，不平等现象迅速上升。

这些新的研究与其他的研究不同，它们完全基于家庭收入和支出调查。早期的研究要么使用平均国内生产总值，忽略了每个国家内部的不平等，要么使用间接方法来估计国家内部的不平等，包括生产调查和收入调查，这些调查通常会错过家庭收入的重要组成部分。世界银行的布兰科-米拉诺维奇（Branko Milanovic）组建了这个数据库，利用世界银行强大的统计组织，从几乎所有的银行成员那里获得了1988年和1993年的家庭调查数据，涵盖了世界上85%的人口。其结果可能是关于世界收入分配的最可靠的数据集。

然后，Milanovic先生计算了世界收入分配的基尼系数，结合了国内不平等和国家间不平等，并以购买力平价衡量。(基尼系数是一个常用的不平等衡量标准：0表示完全平等，100表示一个人拥有所有的收入）。结果是惊人的。世界不平等从1988年的62.5的基尼系数增加到1993年的66.0。这比美国和英国在1980年代经历的不平等的增长速度要快。到1993年，一个拥有最贫穷的10%人口的平均收入的美国人，比世界上三分之二的人生活得更好。



另一项新的研究，由尤里-迪哈诺夫和迈克尔-沃德进行，使用相同的数据集和不同的方法。它证实了世界收入分配在1988年至1993年间明显变得更加不平等。与Milanovic的研究一样，它发现吉尼系数增加了大约6%。它还发现，世界收入中最贫穷的10%人口所占的份额下降了四分之一以上，而最富有的10%人口所占的份额则上升了8%。最富有的10%拉开了与中位数的距离，而最贫穷的10%则远离了中位数，绝对下降了很多。简而言之，我们必须修改单元格4。1988年至1993年期间，按人口加权的国家（中国和印度分为城市和农村）的世界购买力平价收入分配变得 "更加不平等"（见表3）。

为什么全球的不平等会加剧？答案有四个部分。(1）发达的经合组织国家的经济增长比发展中国家作为一个群体要快；（2）发展中国家的人口增长比经合组织国家要快；（3）中国农村、印度农村和非洲的产出增长缓慢；以及（4）中国城市与中国农村和印度农村之间的产出和收入差异迅速扩大。1988-93年期间，中国城市的收入增长非常快，这缩小了中国的平均收入与中等收入国家和富裕国家之间的差距，因此降低了世界基尼系数；但中国农村与中国城市之间以及中国城市与印度农村之间的差距扩大，使世界不平等程度进一步加剧。

这些趋势又有更深层次的原因。技术变革和金融自由化导致极端富裕阶层的家庭数量不成比例地快速增加，而贫困阶层的分布却没有缩小。同时，人口的增长也使贫困家庭的数量不成比例地增加。这些深层原因产生了一个重要的中间原因，使情况变得更糟：高收入国家出口的工业产品和服务的价格增长速度比低收入国家出口的产品和服务的价格增长速度快，而且比很少进行国际贸易的低收入国家生产的产品和服务的价格增长速度快。

这些价格趋势意味着，穷国的大多数人口能够购买的进入富国人口消费模式的商品和服务越来越少。因此，较贫穷的国家和世界上较贫穷的三分之二人口遭受双重边缘化：一次是通过收入，一次是通过价格。因此，表3中的数字显示，世界上最富有的10%人口与中位数之间的差距正在扩大，而中位数与最贫穷的10%人口之间的差距也在扩大。

极化区
收入差距有助于解释世界体系中发生的另一种两极分化，即和平区和动荡区之间。图1中的富人区显示了经济增长和自由宽容（除了对移民）的共和秩序的加强，技术革新能够替代正在消耗的自然资本。中低收入地区有许多国家的治理能力停滞不前或受到侵蚀，主要在非洲、中东、中亚、俄罗斯和东亚部分地区。在这里，越来越多的人发现他们获得基本必需品的机会受到限制，同时他们在电视上看到有人开着奔驰汽车。


其结果是大量的失业和愤怒的年轻人，新的信息技术给他们提供了威胁他们生活的社会稳定的手段，甚至威胁到富裕地区国家的社会稳定。这些国家的经济增长往往耗尽自然资本，因此也耗尽了未来的潜力。越来越多的人将移民到富裕区视为他们唯一的救赎。

令人惊讶的是，世界银行、国际货币基金组织和其他全球组织对这些趋势并不关心。世界银行2000年的《世界发展报告》甚至说，如果底层的收入不下降，贫困人口的数量不减少，收入不平等的上升 "不应该被视为消极的"。这种不重视表明，把这些世界组织称为世界组织是一种误导。从几乎所有国家都是成员的意义上来说，它们可能是世界性机构，但它们的思维方式是以国家为中心的，而不是全球性的。它们不仅忽视了世界收入分配的问题，也忽视了世界通货膨胀、世界汇率和世界利率；就世界银行而言，还忽视了海洋、大气和核废料等全球环境问题。

令人震惊的是，大多数有组织的反对更多的全球化的声音来自北美、西欧和大洋洲。既然可以证明商品和资本的自由市场的收益往往集中在他们国家收入分配的最高层，为什么发展中国家的精英们大部分都赞同西方国家、企业和多边组织一直在推动的全球化议程？为什么他们在以战略方式而不是无限制地将其经济融入世界经济方面做得这么少？

部分原因可能是，发展中国家的精英们，像他们在富裕世界的同行一样，满足于相信世界不平等正在下降，或者不平等是好的，因为它是激励的来源。他们和多边经济组织（以及维多利亚时代英国的改革者）一样，担心贫困问题。但他们认为世界收入分配的扩大与贫困之间没有联系；他们认为贫困可以通过向穷人提供福利和机会来解决，而不需要改变收入和资产分配等更大的结构。学术分析家们有责任通过分析世界收入分配趋势和贫困之间的关系来反击当前的忽视，以此作为将分配问题纳入世界议程的一种方式。

日益严重的不平等现象类似于全球变暖。它的影响是分散的和长期的，而且总是有一些更紧迫的事情需要处理。问题是，在由此产生的政治不稳定和移民潮达到直接损害富裕世界公民的福祉和国家稳定的程度之前，世界收入分配能变得多不平等。在达到这一点之前，我们应该动员我们的政府、多边组织和国际非政府组织，将更平等的世界收入分配作为首要任务，而不仅仅是像现在这样，减少贫困人口。



罗伯特-韦德是伦敦经济学院的政治经济学教授和柏林Wissenschaftskolleg的临时研究员。他是《治理市场》一书的作者，该书是对东亚经济发展的研究。

作者要感谢Michael Ward和Branko Milanovic对收入分配的讨论。本周经济学焦点对本文的评论