Volume 1 April 2007

China in international imbalances
Yu Yongding

International trade and poverty: cause or cure?
L. Alan Winters

Making the boom pay
John Freebairn

Reforming Australian industrial relations
Joe Isaac

Minimum wages and inequality
Andrew Leigh

Does the Fair Pay Commission decision matter?
Mark Wooden

The corporate political environment and big business response
Geoff Allen

Stock return predictability in rational markets
Bruce D. Grundy

Passive profits from accounting indicators
John D. Lyon

A 'Battle of Ideas'
Tom Elliott

On painting one's life picture
Peter Yates

Print this page

By L. Alan Winters

This paper asks how we might think about the effects of international trade on poverty, how we might measure these effects and what the results of such measurement efforts have been. The basic argument is that trade liberalisation generally stimulates medium-term economic growth and that this generally alleviates poverty. But it usually creates some losers, some of whom may be poor or pushed into poverty. For public policy, such effects should make us careful to protect the poor, but not stop liberalising our trade regimes.

Economic growth

Economic growth is the key to sustained poverty alleviation. It creates the resources to raise incomes, and even if, improbably, ‘trickle-down’ were insufficient to bring any immediate benefit to the poor, it at least gives governments scope for financing investment, by complementary or redistributive measures, to help them.

Economic theory offers many reasons to expect a country’s own trade liberalisation to stimulate its economic growth: for example, specialising in goods for which world prices exceed home prices, reaping economies of scale, improving performance in the face of new competition, and benefiting from better inputs and technologies available from abroad.

None of these is guaranteed, however, so ultimately whether trade stimulates growth is an empirical matter. Over the 1990s several highly visible global cross-country studies suggested that openness was good for growth, but the technical hurdles to absolute proof are very high. These hurdles include: measuring the restrictiveness of trade policy accurately; establishing causation – opening up may stimulate growth, but growing countries might be more willing to liberalise; isolating the effects of trade reform from other reforms that typically come with it, especially if it depends on those reforms to work – e.g., having flexible labour markets in order to respond to trade incentives; and allowing for the effects of openness on other policies and institutions, the latter of which are now universally believed to be critical to growth.

Despite these difficulties, the weight of experience and evidence seems to strongly support the idea that openness enhances growth. And even the critics concede that there is no coherent evidence that openness is bad for growth.

The evidence also strongly suggests that economic growth reduces absolute poverty. This is not the same as saying that it reduces income inequality. Indeed, some argue that for low-income countries trade liberalisation could be unequalising.

This probably arises because only the relatively well off are equipped with the skills or locational advantages to capitalise on the opportunities that trade opens up. But these results do not necessarily mean that the poor actually get poorer: all incomes could be rising.

Direct effects of trade shocks

Treating the household as the basic unit over which poverty is defined, the question is posed as to how trade reforms impinge on poor households. This is primarily a matter of how the price changes generated by the reforms affect their consumption and sources of income. Three channels of causation may be distinguished – the prices of goods and services, the market for labour and the role of taxation and government expenditure.

The last is important but not very sophisticated analytically. There is no simple link between trade reform and tariff revenues. In many cases tariff reductions increase revenues since they are associated with reductions in exemptions, less evasion and an enlargement of the tax base as trade increases. But, of course, as tariffs fall towards zero, revenue eventually falls to zero. But even this does not necessarily translate into worse services or higher taxes for the poor. That is essentially a political decision.

A more interesting link is between world prices and trade policy, and the prices of the goods that poor households consume and produce. The bulk of the world’s poor are self-employed in either low-level agriculture or the informal sector of the economy. Thus their incomes are directly affected by price changes induced by international trade. An increase in the price of something that the household sells will increase its real income, while a decrease reduces it. Equally important are the prices the poor pay for what they consume.

An important question, then, is whether trade policy changes, which occur on the border, get transmitted into price changes for poor or nearly poor households. This depends on factors such as transport costs and other costs of distribution; the structure of markets; and domestic taxes and regulations. Some impediments, such as transport costs, are inevitable although they may be increased by policy decisions such as to levy fuel taxes or provide inadequate infrastructure. Others, on the other hand, represent direct economic inefficiency, such as permitting marketing monopsonies or monopolies.

Price transmission is likely to be particularly ineffective for poor people living in remote rural areas and could prevent families from making market transactions almost completely. Such isolation saves the poor from any negative shocks emanating from the international economy but it also prevents them from experiencing the positive shocks and secular benefits that I discussed above. Their problem is too little globalisation, not too much.

A household’s ability to adjust to a trade shock – say to switch production towards goods whose prices have risen – clearly affects the size of any impact it suffers. Hence one needs to ask whether the poor are able to make such adjustments. For many of them, a major constraint on improvements in agricultural productivity following the external liberalisation has been shown to be the absence of key productive assets (draft animals, implements), capital, credit, or information.

For the self-employed, the main determinant of income is the prices commanded by their output and paid for their inputs, but for employees it is wages and employment opportunities. Obtaining employment is one of the surest ways out of poverty, while the loss of a job is probably the most common reason for the precipitate declines into poverty that catch most public attention. The structure of the labour market is critical to how trade liberalisation gets translated into wage and employment changes. Maybe everyone’s wages are bid up a bit, but if wages are inflexible a positive shock is more likely to show up in the creation of a few new jobs at wages well above subsistence levels. If so, who gets the jobs is important. Sometimes they will be the poor, but other times the jobs will go to additional workers in households that already have formal employment or to more skilled workers.

Where unskilled labour is relatively abundant, the goods it produces will be plentiful and hence relatively cheap in the absence of international trade. These are the goods, then, that will be exported when trade increases and hence trade liberalisation will generally relieve poverty. However, not all developing countries fall into this class. For example, many Latin American and African countries have large endowments of natural resources and so liberalisation will stimulate these sectors rather than labour-intensive ones.

Testing and measuring poverty effects

The obvious questions – at least to an empiricist – are can we test or measure these effects and how important are they? Even the simple direct effects are difficult to pin down with complete confidence because (a) we don’t know what would have happened in the absence of trade liberalisation and (b) it is difficult, as argued above, to separate trade liberalisation from all the other reforms that dynamic governments pursue at the same time.

One effort from my own work concerns Vietnam, which liberalised and experienced dramatic declines in poverty over the 1990s. Observing the same 4,300 households both early in the process (1992/3) and later (1997/8), we looked at the influence of certain obviously trade-related factors on households’ ability to escape from poverty. For example, rice production, where the removal of Vietnam’s export restrictions caused an increase in prices by 29 per cent and a huge export boom; fertilizer use, where trade liberalisation reduced real prices by 23 per cent over our sample period; coffee production – Vietnam hugely increased coffee output and exports, partly responding to a large increase in world prices; and employment in one of the booming manufactured/processed export sectors – seafood, food processing, footwear and clothing. All of these proved statistically significant in explaining the escape from poverty and overall, it was estimated somewhat heroically that they cut poverty by about 17 per cent.

A second approach to estimating the importance of trade reform is simulation modeling, in which the possible consequences of future trade reform using numerical models are explored. Remembering, of course, that these are hypothetical not factual exercises, this is also attractive in allowing the effects of trade liberalisation to be isolated precisely – it is the only change made in the simulations. One recent exercise tries to estimate the poverty effects of reforming developed countries’ agricultural policy – both a total liberalisation and what we might get if the Doha Round comes to a successful close.

Effects on different countries

The current apology for preserving European and American agricultural protection and support is that it supports poor farmers in the North and that liberalisation would benefit only rich land owners in the South. Both assertions contain grains of the truth but the predominant effects are the very opposite.

In the US, only two per cent of the population lives in farm households and on average only eight per cent of their income comes from the farm. If we array farms by total wealth, the poorest half make almost no income from farming, but the higher wealth brackets get a large share of their income from it.

The richest farmers specialising in the most protected commodities – rice, cotton, sugar or dairy – get up to 90 per cent of their income from farming. Thus liberalisation hits richer farmers far harder than poor ones (up to 20 per cent losses of income), which explains the strong political opposition to reform.

Turning to developing countries, which vary considerably, and recognising the many different ways in which the poor earn their living, there are great differences between countries. Rich country agricultural trade liberalisation raises world agricultural prices and boosts real returns in agriculture – for land and farm workers. In Brazil, for example, this greatly reduces poverty, whereas in Mexico, with its much larger non-agricultural workforce (who now pay more for food) it increases poverty. In Zambia, farmers gain, and there are lots of them, but they start off so poor that very few actually make it above the poverty line as a result of these reforms. Overall, agricultural liberalisation would be poverty reducing, but country experience varies hugely.

Another interesting result is that the package of reforms discussed in the Doha Round could be made more poverty alleviating. It contains the abolition of rich countries’ export subsidies – which does little or nothing for poverty in most developing countries because the poor are either isolated or net purchasers of the commodities affected. It also calls for little liberalisation by developing countries themselves – which we find would have had very strong poverty-reducing effects in nearly all countries.

One final observation: Not only are the poverty effects of trade reform heterogeneous, but they are also far from sufficient to remove the scourge of poverty altogether. Trade reform will be more effective if it is accompanied by policies to relax constraints on the productivity of the poor – for example, access to credit and information – and long-run growth requires advances in other areas such as education, security of property, infrastructure and financial development. Even the staunchest advocates of trade liberalisation recognise that it is not the only thing needed for development.

Top ^