Speaker:

L. Alan Winters, Director of the Development Research Group, The World Bank

Discussants:

Sheila Page, Senior Research Associate, International Economic Development Group, ODI;
Sherman Robinson, Professor of Economics, University of Sussex.

Chair:

Dirk Willem te Velde, Research Fellow, International Economic Development Group, ODI

The chair opened the meeting saying that it would discuss the book Poverty and the WTO: Impacts of the Doha Development Agenda, edited by Thomas Hertel and Alan Winters. He said that it was fortunate that this meeting was being held for a number of reasons:

• To discuss modelling approaches underlying estimates of the effects of trade
  

• To discuss the effects of the Doha round, and why the estimates have been changing over time
  

• To discuss what the book means for developing country negotiators, e.g in terms of the effects of agriculture negotiations.

Alan Winters spoke in a personal capacity at this meeting. He began his presentation by stating that the motivation for the study was the World Bank's (WB) objective to reduce poverty, focused on the Millennium Development Goals (MDGs). The WB believes that increased trade has the potential to create resources to meet the MDGs. It believed that negotiations on agriculture are likely to have the most impact on poverty, as it is the sector in which most poor people are engaged. Winters stressed that negotiations at the global level must be translated into results at the household level if we are to understand their impact on poverty on the ground. Throughout the presentation the potential impacts of the Doha Development Agenda (DDA) were compared with full liberalisation (an ideal).

Winters gave a summary of the tariff cuts that it is assumed will be negotiated under the DDA in agriculture and non agriculture market access for developed, developing and least developed countries.

Winters warned that if developed countries maintain current tariffs on even a small proportion of tariff lines (2%) under the rubric of special and sensitive products, the world will lose two thirds of the poverty-reducing benefits of the DDA.

Winters highlighted the difference between bound and applied rates of tariffs. The bound rate is the rate a country agrees in WTO negotiations, whereas the applied rate is the rate the country actually uses. The bound rate is sometimes many times higher than the applied rate, and so in order to have any effects large cuts in bound tariffs must be negotiated.

The presentation then moved to the impact of trade reforms on world exports. Projections based on the ideal scenario of full liberalisation and the more realistic scenario which is DDA liberalisation were compared. For example, the volume of rice exports could be expected to rise 120% with full liberalisation compared to 20% with DDA liberalisation. Figures for apparel exports are 20% vs. 7% - an indication that the DDA round will do less for manufacturing exports.

The case study of Mexico demonstrated the importance of estimating price transmission (whether price changes filter through the economy to the prices the poor pay). Poverty is affected by the prices that households receive or pay, not directly by export or import prices on the border. In Mexico, trade liberalisation has the largest impact on areas near the US border where the biggest price changes occur. The smallest price changes occur in the South, in the poorest regions, indicating little impact from trade liberalisation on the poorest. The DDA will provide limited gains in real income. Supplemented by other policies, such as enhanced productivity and price transmission, trade liberalisation could provide increases in incomes to poorer areas.

Negotiations and agreements do not occur in a policy vacuum. The Zambia case shows that other policy decisions can be crucial. Since privatisation of one marketing board in 1994, cotton production in Zambia has shifted strongly towards poorer farmers so that any world price rise arising from the DDA will now be more strongly pro-poor than it would have been before.

The case study of Brazil shows that the assertion by European negotiators that all the benefits of trade liberalisation go to rich landowners is not entirely true. The figures presented on slide 19 show that, although the richest decile of agricultural employers could see an 8.2% increase in income, the poorest decile of non-agricultural workers could benefit from an 8.5% increase income. This is the result of increased jobs in agriculture going to the previously unemployed.

Conclusions

1. The DDA round must be ambitious to have an impact on development. Winters makes a criticism that DDA should have focused on tariffs, rather than export subsidies, as tariffs have more impact on poverty.

2. The DDA says little on developing countries' own liberalisation (i.e. tariff cuts), which Winters explains would have the most impact on poverty.

3. He also concluded from his comparison of full liberalisation and DDA liberalisation that DDA is less development friendly than full liberalisation.

Sheila Page raised 3 main points:

1. There is a danger that the impact trade can have on development is being overemphasised. The World Bank itself lowered its estimates of the impact trade can have during each of the Uruguay and Doha rounds. Hopes should not be raised in vain.

2. There are important findings in the book which have political implications, e.g. that the poor as well as the rich stand to gain in Brazil. She emphasised that this was a "political book" and that studies of this nature are not undertaken for their own sake, but to influence policy outcomes

3. Policy is still the most crucial element in development. It is an important tool to offset the negative effects and reinforce the positive effects of liberalisation.

Sherman Robinson made the following points:

1. Heterogeneity is a key factor - it is very difficult to generalise from the country studies. The impact on unemployment in Brazil for example was not replicated in Indonesia. Trade is only part of the story and must be complemented by other policies.

2. Infrastructure building still needs attention - the gains made from market liberalisation can be lost in the high cost of getting products from the border to the market.

3. In the aid for trade debate, he suggested aid could be used at the micro level to provide assistance to raise productivity and therefore improve countries' ability to trade.

Floor debate
The discussion covered the following points:

1. The impact of price increases on the urban poor. The speakers replied that a trade-off is necessary between gains for the rural poor and costs for the urban poor.

2. Trickle-down wealth is sufficient to lift the poor out of poverty. Winters replied that countries with poor policies and a weak tax base frequently fail to benefit from trickle-down, as evidenced in many Latin American countries.

3. The message Winters would give to the UK House of Lords, debating the G8 development outcomes this week. Winters replied that reducing tariffs is a cheap way to get the poor some benefits ($38bn in the DDA). However, trade reform must be accompanied by other policies. Moreover, the G8 must accept adjustment costs.

4. Whether more benefits come from developing countries own liberalisation or from other countries' liberalisation. Winters replied that it varies from country to country and he did not have precise figures, but could be roughly estimated at a 40% own to 60% others' split. This book employs modelling techniques that emphasise the benefits from other countries' liberalisation. [Subsequent to the meeting Winters clarified that later work by Hertel and Ivanic based on their book chapter showed that the share of overall poverty reduction stemming from the own liberalisation component of a full liberalisation was estimated at 100% for Bangladesh, 40% for Indonesia and 60% for Vietnam.]

5. Whether the study compared the gains for the poor with the non-poor. Winters replied that it did not and that trade negotiations do not have a great effect on income distribution.

6. The definition of near term. Winters replied that the models used are static models, but that one might think of approximately 3 to 4 years.