5. Simulation results

This section reports assessments of the likely impacts of the three scenarios, representing the Harbinson, the US and the EU proposals. Most of the main results are expressed as percentage change from the base levels. Also reported is the information on "winners and losers", i.e. the number of individual countries that gain or lose in various ways.

It is useful at this stage to take note of the commonly stated scepticism about model-based results. The sceptics argue that global trade models often give different assessments, and messages, at times widely different for even seemingly similar reform. This is true to some extent, but not entirely so. For example, most studies of this type point to the same direction for changes in world market prices, at times even to similar magnitudes, e.g. the impacts on temperate-zone food products versus those on tropical products. But differences do also occur. There are several reasons for this, which are worth noting in reviewing the results.^[70] For example, models differ with respect to structure, specification and parameters. Models also differ greatly in terms of aggregation, both of countries and commodities, which tend to cancel out some effects that would be evident in more disaggregated models. The elements of a simulated reform package could also be different, and even weights attached to particular instruments could vary. Nevertheless, there is no alternative to model-based assessments.

5.1 Impact on world market prices

Often, it is changes in world market prices that drive other results, and so this has to be the starting point. Table 1 shows the impact on world market prices, measured as percentage change from the base period levels. There are three main messages here. First, on the whole, the impact is large on most temperate-zone commodities such as beef, sheepmeat, dairy products, sugar, wheat and vegetable oils, in all three scenarios. This is an expected result, as these are the commodities receiving high levels of protection and support. Other studies do also point to similar impacts. For example, the USDA study (Diao et al., 2001) showed large increases in world prices under full policy reform (18 percent for wheat, 15 percent for other grains, 22 percent for butter and 12 percent for beef). The impact was small (e.g. 3.4 and 1.4 percent for wheat and other grains) under slight reductions in domestic support and tariffs. The study of Diao et al., (2001) full policy reform scenario is closer to the US scenario in this paper. Similar effects were noted in a recent FAPRI study analyzing the impact of the Doha Round reforms (FAPRI, 2002). By contrast, impacts on tropical product prices were modest, which was expected result.^[71]

Second, looking at the results across the three scenarios, the impact is most pronounced under the US scenario, also expected in view of the sharp reductions in tariffs in particular but also elimination of domestic support and exports subsidies. Unlike the other two proposals, the US one is based on cuts in applied rates, which are much lower already for developing countries. These price rises are on average twice the levels under the EU proposal, while those under the Harbinson proposal lie in between the two scenarios. There are some cases where this pattern does not hold strictly.

Third, it is already possible to guess, based on the extent of the changes in world prices, the direction of the impact on some of the national level indicators, like trade and welfare. Thus, the change in the terms of trade would be unfavourable for developing countries to the extent they import basic foods and export tropical products.

All elements of the reform package contribute to the price change, but tariffs play a dominant role, especially the sharp reduction for developing countries resulting from the US scenario.

Table 1. Impact on world market prices (percent change from base levels)

Source: Simulation results.

5.2 Impact on total welfare

As explained in Section 3.1, total welfare or surplus (TS) in the ATPSM is the sum of producer and consumer surpluses (PS and CS respectively) and government revenue. This sub-section summarizes the overall impact of trade liberalization on country groups in terms of the TS; impacts on the three individual components of the TS are discussed separately in some detail in the following sub-sections and so these should be read for further insights. Table 2 shows the results, whereas Annex 3, Table 1 gives the results by country groups and commodities.

Table 2. Change in total welfare by country groups (billion US$)

Source: Simulation results. All monetary values are in US dollar terms. DD= Developed countries, RDC = Rest of developing countries, LDC =Least developed countries.

Results for both the Harbinson and EU scenarios show a similar pattern in terms of the direction of change (positive or negative), while this is somewhat different in the US scenario. Focusing on the developing countries and the former two scenarios, the results show both groups of countries losing in TS term although the magnitude of the loss is small. The loss in TS is the result of the sum of the losses in CS and government revenues exceeding the PS gain. In case of the LDCs, the main reason for the CS loss is that since they are not required to reduce tariffs, the high world prices are transmitted to consumers in full; had tariffs been lowered, these would have offset some of the effects of higher world prices.

In the case of the rest of the developing countries,^[72] the expectation would be that the effect of the higher world prices on CS would be moderate as these countries reduce import tariffs. But the CS still turned out to be negative for them, which is somewhat surprising. This means that the reduction in the operative tariff in the model (that is, the lower of the applied or bound tariffs) was not large enough to more than offset the effect of the higher world price. Producer surpluses for the two developing country groups are positive. This is easier to explain - as tariffs did not decline or did not fall sufficiently, the higher world prices were transmitted to domestic markets almost entirely, leading to positive PS. As for government revenues, the negative effect was because imports declined as world prices rose (while tariffs remained the same or fell only slightly).

The results under the US scenario for the two developing country groups are interesting in that TS is positive. Importantly, the positive TS gain for the LDC is due to CS gain since the impact on PS is negative, which is in contrast to the other scenarios (see Table 4 below). This is explained by the fact that the LDC do reduce tariff in the US scenario. As a result, domestic prices do not rise by as much, and so the CS is positive while PS is negative.

It is relatively straightforward to explain the impact on developed countries. They experience TS gains in all three scenarios, due to gains in both CS and government revenues which together more than offset the large PS losses. These are all expected results, and are explained further below. Briefly, the PS loss is due to the sharp reduction in tariffs which lowered domestic prices despite the increased world prices. The significant increases in government revenues reflect reduced expenditures on domestic and export subsidies. The bulk of the welfare gains to developed countries come from commodities receiving high support and protection such as wheat, rice, and livestock products see Table 4. The much larger impacts in the US scenario are due to both elimination of subsidies and steep reduction in tariffs.^[73]

Table 3 shows the number of countries experiencing TS gain and loss. Among the three scenarios, the Harbinson proposal shows the highest number of losers, a total of 114 countries out of 161, 73 of them being rest of the developing countries. However, the magnitude of the loss itself is small, as PS and CS typically tend to cancel out. For example, the TS loss is only US$0.32 billion for Mexico and US$0.3 billion for China. The EU scenario results I the lowest number of losers, only 97 countries, of which 69 are from the rest of the developing countries group, 21 LDCs and 7 developed countries.

Table 3. Winners and losers: total welfare

Source: Simulation results. DD= Developed countries, RDC =Rest of developing countries, LDC = Least-developed countries.

5.3 Consumer and producer surpluses

Consumer and producer surpluses are useful welfare indicators on their own. These are also interesting from the standpoint of the political economy. Governments often attach different importance to these gains. In general, producer gains are preferred in practice. Although agriculture is at times found to be taxed, many developing country policy makers would like to support the sector in view of its importance for economic growth and poverty reduction. Most developed countries also attach greater weight to the PS gain, perhaps for other reasons.

Table 4 shows estimated PS and CS values for the three scenarios. The developed countries are seen to experience loses in PS terms and gains in CS terms in all three proposals. The gains more or less offset the losses, with one exception in the US scenario where the CS gain is markedly higher than the PS loss. It is straightforward to explain this outcome. With liberalization, the high level of support and protection currently accorded to agriculture in these countries declines, leading to massive PS losses (lower producer prices and reduced production). This is also obvious from the estimates of the gains and losses associated with individual commodities.

Table 4. Change in consumer and producer surplus (in billion US$)

Source: Simulation results.

The opposite is the case for rest of the developing countries - they lose in CS but gain in PS in all three scenarios. Both the Harbinson and US results show similar PS gains, about US$21 billion, but only half that amount in the EU scenario. For LDCs, there is an outlier to this pattern - they lose in PS term in the US scenario, and significantly so, about US$3 billion, but gain (about US$1.6 billion) in the other two scenarios. The main reason for this loss, as noted before, is that in the US scenario the LDCs also reduce tariffs and so farm prices decline. By contrast, in the other two scenarios, although tariffs decline they do not fall so far as to lower domestic prices much given the rise in world prices.

Thus, overall, if developing countries were to place greater importance on the PS gain, the Harbinson and US proposals appear attractive, as these gains are about twice the levels under the EU proposal. For the LDC on the other hand, the best outcome from the PS standpoint seems to be the Harbinson proposal, as they lose in the US proposal. Tables 2 and 3 in Annex 3 show the results for producer and consumer surplus by country group and commodity.

Table 5 shows the number of winners and losers. Considering the case of the two developing country groups, the results show that an overwhelming majority of them (about 115 countries) gain in PS term under the EU and Harbinson scenarios, versus only 88 gainers under the US proposal. Out of 42 LDCs, 39 gain under the EU and Harbinson proposals. The opposite result holds somewhat in the case of the CS gain - 56 developing countries combined gain on this account under the US proposal and only half as many under the EU and Harbinson scenarios.

Table 5. Winners and losers: consumer and producer surplus

Source: Simulation results.

In summary, two points may be noted. First, the results show differing gains and losses under the three proposals, at times markedly so, and mostly in opposite directions. This is useful information for trade negotiators and policy makers to assess the outcomes, as they often attach different weights to the two sources of welfare gains. Second, the results raise the issue of compensation, as is well known in welfare economics. There are two dimensions to this issue. One is compensating winners and losers within a country (i.e. between the PS and CS gainers and losers). The other is compensation at the world level. Given the results, it is clear that governments need to take into account the issue of social cost to particular population groups during the reform process. At the world level, the framework for compensation is not as developed. Nevertheless, this is one area worth pursuing within the WTO framework.^[74]

5.4 Government revenue

In the ATPSM, a change in government revenue results from changes in tariff revenues, expenditures (namely, domestic and export subsidies) and the part of the quota rent not received by exporters. For a large number of developing countries, tariff revenue accounts for a large share of total revenue - in the 10-20 percent range for many, and considerably more in several cases (Weisbrot and Baker, 2002). The impact on revenue of reduced domestic support and export subsidies is obvious - these show up in the model in the form of increased revenue. Reduced import tariffs (export taxes) typically lower border revenues but it may also be the case that these reductions could actually lead to higher border revenues if lower tariffs lead to higher imports (or more exports) to the extent that the overall revenue is higher. In the model, changes in trade flow also depend upon world prices, and thus there are some interesting interactions at play.

Table 6 reports impacts on government revenues. Turning first to the simpler case of developed countries, revenues increase substantially under all three scenarios. The explanation is simple - most of this is due to reductions in government expenditures on domestic and export subsidies. This is also clear from Annex 3, Table 4, which shows that most of the increases in government revenues came from products that received large subsidies in the base period. Reduced import tariffs also play some role, but a relatively small one for these countries. Increased revenue is about US$13 billion under the Harbinson proposal, and US$9 and US$6 billion under the EU and the US scenarios respectively.

Table 6. Impact on government revenue (in billion US$)

Source: Simulation results

In the case of the two developing country groups, the impact of all three proposals is negative, i.e. there are revenue losses, although magnitudes differ for obvious reasons. Since these countries grant few domestic and export subsidies (none in the case of the LDCs), almost all of the observed effects are due to changes in border revenues, notably tariff revenues. As said earlier, the net effect is the outcome of two factors, average tariff rate and trade volume. The former is obviously lower after liberalization. As for the second, while higher world prices discourage imports, lower tariffs typically have the opposite effect. Given that the revenues declined, it is obvious that import volumes did not rise enough to more than offset the effect of the lower tariffs. In particular, it seems that the substantively reduced revenues for the rest of the developing countries (US$5.8 billion) under the US scenario was due to both marked declines in import volumes (as import prices rose sharply) and much lower tariff rates after the reform. In general, the decline in government revenue is least under the EU proposal, i.e. one-third less than in the Harbinson proposal and several times less under the US scenario.

These results show the possibility of significant negative effects that many developing countries have been expressing as a matter of concern for them in the context of trade liberalization. Faced with these prospects, they are often advised to diversify into other forms of taxation. But this is often more easily said than done for a majority of these countries if one takes into account administrative and other costs involved in raising these taxes. In any case, the results show that it is important that this aspect is taken into account in considering alternative proposals for trade liberalization.

Table 7 provides further information in terms of the number of countries that lose or gain on this account. It shows that 81 percent, 74 percent and 71 percent of the developing country group face reduced government revenues under all three proposals, while a majority of developed countries show gains.

Table 7. Winners and losers: government revenue

Source: Simulation results.

5.5 Impact on export earnings, import cost and trade balance

In the model, the two key factors that determine changes in export earnings and import bills following a policy reform are changes in world market prices and tariff rates, which influence trade flows. Table 8 shows the results. With the exception of two cases for LDCs, both export earnings and import costs increased for all country groups and scenarios. Although the changes in world prices are common to all country groups, the impact on trade flows varies according to changes in tariff rates and to differences in commodity composition. The developed country group experiences deteriorations in trade balance in all three scenarios, significantly so in the Harbinson and US proposals, as import bills rise markedly over export earnings. This is explained mainly by generalized contractions in export volumes as production shrinks when farm support is lowered and the market is opened. Import bills would have increased even if import volumes did not increase much, due to the effect of the world price.

Table 8. Change in export revenues, import costs and trade balance(billion US$)

Source: Simulation results.

As for the rest of the developing countries, they too experience increases in both export revenues and import bills in all three scenarios, as above. However, and in contrast, for them the increase in export revenue more than offsets the rise in import bills and, as a result, the trade balance improves. As above, the increases are significantly higher under the Harbinson and US scenarios. The main reason for the modest rises in import bills is contraction in consumption and hence in imports, due to higher prices.^[75] By contrast, export earnings rose because both exports and world prices increased.

In the case of the LDCs, the direction of the impact differs in three places relative to the case of the rest of the developing countries (negative signs in the table). The underlying reasons behind these effects on LDCs have been discussed throughout this section. Since the LDCs do not reduce tariffs, higher world prices raise domestic prices fully. This causes consumption, and thus imports, to contract. Despite the high world price, the net effect was reduced food import bill in the two scenarios. In the case of the US proposal, as said earlier, even the LDCs reduce tariffs and hence the effect is similar to that seen for other countries. The reason for higher export earnings is obvious.

Table 9 reports the number of winners and losers under the three scenarios. These show 75 winners in the rest of the developing country group under the US proposal, 76 under the EU proposal and 75 under the Harbinson proposal. Overall, thus, there are significantly more winners than losers.

Table 9. Winners and losers in terms of trade balance

Source: Simulation results