Peter Gallagher

The setup

Let’s imag­ine that 38 coun­tries agree to elim­i­nate their bar­ri­ers to imports of cere­als (wheat, maize, rice, bar­ley, sorghum). The coun­tries are: Argentina, Aus­tralia, Brazil, Canada, Chile, China, Colom­bia, EC, Egypt, India, Indone­sia, Israel, Japan, Jor­dan, Korea, Kuwait, Malaysia, Mex­ico, Morocco, New Zealand, Nige­ria, Nor­way, Peru , Philip­pines, Qatar, Saudi Ara­bia, Sin­ga­pore, South Africa, Switzer­land, Tai­wan, Thai­land, Turkey, Ukraine, United Arab Emi­rates, Uruguay, US, Venezuela, Vietnam.

Together, these coun­tries account for just over 80% of world trade inthese cere­als.

The results

The fol­low­ing is a brief sum­mary of the key results of the sim­u­la­tion. I’ll be pub­lish­ing more details later in the year in con­junc­tion with my fellow-researcher Andrew Stoler of the Insti­tute for Inter­na­tional Trade.

As in most agri­cul­tural trade lib­er­al­iza­tion sce­nar­ios, the over­all pat­tern of the model pro­jec­tions is an increase in devel­op­ing exports to devel­oped coun­tries (Table 3) where mod­er­ate to high applied rates of duty are matched by rel­a­tively high lev­els of mar­ket con­sump­tion. In the CMA sce­nario, how­ever, some of the largest (India, China, Indone­sia) and most highly pro­tected (Indone­sia for rice) devel­op­ing coun­try mar­kets are also opened up.

As mar­ket bar­ri­ers fall, con­sump­tion increases in mar­kets where the bar­ri­ers for­merly held domes­tic prices above the world mar­ket price. Demand on the world mar­ket rises due to the higher total level of imports push­ing world mar­ket prices up by between 3% for bar­ley and 10% for wheat (Table 2).

Elim­i­nat­ing import bar­ri­ers to these grains in the thirty-eight CMA coun­tries lifts world export rev­enues by 23% and the export rev­enues of the CMA coun­tries as a group by 18% (Table 3). The sta­tic increase in the group’s total wel­fare is almost $4 bil­lion due to both the higher level of pro­ducer sur­plus in CMA export­ing coun­tries (higher world-market prices) such and to higher con­sumer sur­pluses in for­merly pro­tected CMA import­ing coun­tries. Although world mar­ket prices rise, grains cost less than they used to in mar­kets such as Japan and West­ern Europe when the bar­ri­ers are pulled down, so con­sumers in those large mar­kets are better-off.

Least devel­oped coun­tries see the largest increase in exports (from a low base) but reduced total wel­fare because import prices are now higher. The higher world prices result in a net wel­fare loss Sub-Saharan Africa of $246 mil­lion. Most of this loss in the model, how­ever, is due to reduced gov­ern­ment rev­enue due to lower tar­iffs and con­se­quent loss of pro­ducer sur­plus mostly on rice and wheat in Nige­ria and, to a small extent, in South Africa. As you would expect, some of the biggest trade gains among the CMA coun­tries accrue to the com­pet­i­tive, low pro­tec­tion, grains exporters such as the United States, Canada, Aus­tralia and Argentina (Table 4).

India enjoys the biggest growth in net-exports of grains (exports minus imports) fol­low­ing CMA lib­er­al­iza­tion, out­strip­ping even the United States as a result of a switch to exports from domes­tic sales (pro­duc­ers see bet­ter oppor­tu­ni­ties on world mar­kets once domes­tic prices are no longer pro­tected). India’s high export sur­plus also reflects its strong export sales in the ‘base’ period used in this ver­sion of the model. Although it expe­ri­ences a small growth in net imports of $102 mil­lion, China’s over­all cere­als trade is in near bal­ance. It’s no sur­prise taht Indone­sia sees a $400 mil­lion increase in rice imports once bor­der pro­tec­tion is elim­i­nated. This fully accounts for its increase in cereal imports. Japan, too, sees cereal imports increase by almost a bil­lion dol­lars, mostly in rice.

The biggest absolute changes occur in the EU where a $3 bil­lion dol­lar increase in wheat imports dom­i­nates its own (and the global) projections.

Elim­i­nat­ing sub­si­dies, too

Sup­pose that a CMA agree­ment, in addi­tion to cut­ting import bar­ri­ers to zero were to include an agree­ment to elim­i­nate export sub­si­dies on cere­als and trade-distorting domes­tic sup­ports on cereals.

Although this sounds like a rad­i­cal an
exten­sion of the CMA idea, it is not. If import bar­ri­ers are elim­i­nated, trade-distorting domes­tic sup­ports, which take the form of mar­ket sup­port sub­si­dies, would be for all prac­ti­cal pur­poses infeasible.

Own­ing to the national-treatment oblig­a­tion of GATT, any mar­ket sup­ports offered in an open mar­ket would ‘leak’ imme­di­ately to imports cre­at­ing an un-stoppable hole in the sup­port bud­get. So it is possible—even likely—that a CMA agree­ment to elim­i­nate bor­der bar­ri­ers would have to be com­ple­mented by an agree­ment to elim­i­nate trade-distorting domes­tic supports.

Almost all export sub­si­dies are offered by the Euro­pean Union, which has already planned their elim­i­na­tion as part of its inter­nal mar­ket reforms.

The dif­fer­ences in this sce­nario are due in part to world price changes that are sig­nif­i­cantly larger than in the access-only sce­nario; wheat prices rise by 12%. But the biggest changes—positive and neg­a­tive in dif­fer­ent regions—are due to the elim­i­na­tion of huge sub­si­dies to pro­duc­tion and export, espe­cially in the devel­oped economies.

The global wel­fare gain is strongly pos­i­tive and larger than in the ‘barrier-only’ sce­nario by more than $2 bil­lion (Table 5, com­pare Table 2). “Win­ners” far out­num­ber losers, sug­gest­ing that some over­all com­pen­sa­tion trans­fer mech­a­nism would be pos­si­ble that would ensure no one is worse off as a result of the changes. The CMA mem­bers again receive a larger ben­e­fit than the world-as-a-whole despite (or because of) the fact that they are cut­ting sub­si­dies as well as access barriers.

Devel­op­ing coun­tries as a group see net wel­fare ben­e­fits that are almost four times the ben­e­fits in the first sce­nario on account of larger pro­ducer sur­pluses (higher pro­ducer prices). But net importers, such as the coun­tries of sub-Saharan Africa, face slightly larger net wel­fare losses because the loss of export subsidies

About the ATPSM model

The ATPSM model that I used for these sim­u­la­tions was con­structed by econ­o­mists in the Trade Divi­sion of the UN Con­fer­ence on Trade and Devel­op­ment (UNCTAD). There is a down­load­able ver­sion of this model, which runs under recent ver­sions of the Win­dows oper­at­ing sys­tem. It uses trade data cre­ated for the model by the FAO and UNCTAD con­tained in a set of Microsoft Access data­base tables. Pro­duc­tion and con­sump­tion data are an aver­age of actual data for 2002 — 2004. Prices are FAO esti­mates from 2006, a year in which prices were about 1% higher than the aver­age for 2004– 2007.

In the jar­gon, the ATPSM model is a deter­min­is­tic, par­tial equi­lib­rium, com­par­a­tive sta­tic model. It begins with a ‘snap­shot’ descrip­tion of the world, con­tained in the Access data­base tables. This ‘toy’ world is based on the actual pro­duc­tion, con­sump­tion and prices of 35 com­modi­ties in 151 coun­tries. The model works by apply­ing a trade pol­icy ‘shock’ to this set of facts in the form of a change in trade bar­ri­ers cho­sen by the mod­eler. Trade bar­ri­ers work, mostly, by chang­ing rel­a­tive prices and the level of gov­ern­ment taxes. In reac­tion to the ‘shock’, the model forces its rep­re­sen­ta­tion of domes­tic and world mar­kets to find a new set of prices where excess (too much or too lit­tle) sup­ply and demand ini­tially caused by the trade-policy shock ‘clears’. That is, the model finds a new set of prices at which global pro­duc­tion and con­sump­tion are roughly equal after tak­ing account of imports and exports. Now we have a new ‘snap­shot’ of the world, com­pris­ing a new set of ‘equi­lib­rium’ prices along with new pat­terns of pro­duc­tion, con­sump­tion and trade flows. The model allows the user to make a com­par­i­son between the ‘pre-shock’ and the ‘post-shock’ snapshots.

You can prob­a­bly guess that this sort of math­e­mat­i­cal exer­cise nec­es­sar­ily leaves out a lot of things that hap­pen in ‘real’ mar­kets. To keep the sim­u­la­tions man­age­able, the model takes no account, for exam­ple, of the vast dif­fer­ences in infor­ma­tion avail­able to peo­ple in dif­fer­ent parts of the mar­ket. It ignores the impact of monop­oly pric­ing by state enterprises—although not where this pric­ing is directly related to a trade barrier—and the many sub­tle influ­ences of investors’ and spec­u­la­tors’ behav­ior in com­modi­ties mar­kets. ATPSM also ignores the impact of stocks.

But even with these sim­pli­fi­ca­tions the ATPSM model can still show us impor­tant and non-obvious rela­tion­ships between pro­duc­ers, con­sumers and gov­ern­ment poli­cies as a result of changes in trade bar­ri­ers. What is most inter­est­ing is the shape of the changes: who is better-off and who is worse-off. The scale of the change is also indica­tive, although the absolute val­ues are not likely to be reliable.

One final remark about the lim­its of the ATPSM sim­u­la­tions. The pro­jected results are very likely to be much too small com­pared to the ‘real’ world impacts of the sim­u­lated changes. The ATPSM does not cap­ture the ‘dynamic’ fac­tors of change in the com­mod­ity mar­kets that we are simulating.

One exam­ple of a ‘dynamic’ fac­tor is learn­ing or the impact of tech­nol­ogy. When imports take a larger share of a mar­ket because of a change in trade bar­ri­ers, one of the effects that goes hand-in-hand with the tougher com­pet­i­tive con­di­tions for local pro­duc­ers is the import of new tech­nolo­gies, or designs or plant vari­eties or fla­vors or… any aspect of learn­ing and inno­va­tion that pro­duc­ers else­where have adopted to improve their com­pet­i­tive­ness. Local pro­duc­ers imi­tate or per­haps acquire the new tech­nol­ogy in order to stay com­pet­i­tive with imports, lift­ing their pro­duc­tiv­ity. They are likely, too, to build on and improve the tech­nol­ogy for local conditions.

These fac­tors have a strong and valu­able impact on growth and trade, typ­i­cally account­ing for a much big­ger impact over time (even at dis­counted prices) than the ini­tial impacts of the trade bar­rier reduc­tion. ATPSM does not cap­ture these results.