ANNEX

METHODOLOGY AND PROCEDURES

(a) Methodology

Baldwin and Murray (1977) develop a partial equilibrium approach to estimating the trade creation and diversion effects of extending GSP tariff reductions to manufactured and semi-manufactured imports from developing countries. Following their method, Bourke (1988) estimates the impacts of removing trade barriers to forest product imports from developing countries into the Japanese market. The same methodology has been adapted and applied in this paper to calculate the potential effects of the Uruguay Round tariff agreement on selected forest products in a range of importing markets. Adaptations to the original approach of Baldwin and Murray were necessary as the Uruguay tariff reductions in import markets have affected both developing and developed country forest product trade, and in some markets have led to the elimination of differential GSP and MFN tariff rates. This annex illustrates the methodology employed and its adaptation for different import market conditions.

Following Baldwin and Murray, forest products of an importing market can be distinguished in terms of imports from developing countries, imports from developed countries and domestic production. It is assumed that the importing country is a price-taker for international forest products, and that both foreign and domestic supply are perfectly elastic over the small range of price changes considered. Moreover, given the relatively insignificant share of domestic and imported foreign products expenditure in the national income of the importing country, the income effects of tariff reductions are ignored.

These assumptions allow the trade creation and diversion effects of the Uruguay Round tariff removal in the importing market to be analyzed in terms of the respective demand and prices for the three forest product categories indicated above. That is, if Q is the domestic output of a forest product, M_a are imports of the product from developed countries and M_b from developing countries, then the respective demand curves can be represented by

Q = Q(P^d, P^m_a, P^m_b) (1)

M_a = M_a(P^d, P^m_a, P^m_b)

M_b = M_b(P^d, P^m_a, P^m_b),

where P^d, P^m_a and P^m_b are the prices of a forest product from domestic output, developed countries and developing countries respectively. It is assumed that the above demand curves display the normal own and cross-price properties. Also, value added tariffs on imports apply, so that P^m_a = P^W(1 + t^m_a) and P^m_b = P^W(1 + t^m_b), where P^W is the world (free trade) price for the forest product and t^m_a and t^m_b are the initial tariff rates for developed and developing country imports respectively.

Totally differentiating (1) with respect to import prices of developed and developing country forest products yields

dQ = (_Q/_ P^m_a)dP^m_a + (_Q/_ P^m_b)dP^m_b (2)

dM_a = (_M_a/_ P^m_a)dP^m_a + (_M_a/_ P^m_b)dP^m_b

dM_b = (_M_b/_ P^m_a)dP^m_a + (_M_b/_ P^m_b)dP^m_b .

In what follows, the above price effects can be distinguished in terms of trade creation and trade diversion.

Trade creation occurs when reductions in import prices lead to substitution between domestic output and imports. That is,

- dQ = - (_Q/_ P^m_a)dP^m_a - (_Q/_ P^m_b)dP^m_b = (_M_a/_ P^m_a)dP^m_a + (_M_b/_ P^m_b)dP^m_b (3),

or - dQ = TC = TC_a + TC_b ,

i.e. total trade creation, TC, equals substitution for domestic output from developed country imports,TC_a, and developing country imports, TC_b, respectively.

Trade diversion occurs when changes in import prices lead to substitution between developing and developed country imports. Let M^S_a,b denote the net substitution from developing country imports to developed country imports. Then the effect of a change in import prices is

>

M^S_a,b = (_M_a/_ P^m_b)dP^m_b - (_M_b/_ P^m_a)dP^m_a = 0 (4)

<

= dM^S_a - dM^S_b

and M^S_b,a = - M^S_a,b ,

where dM^S_a is the substitution effect on developed country imports, dM^S_b is the substitution effect on developing country imports and M^S_b,a is net substitution of developing country imports for developed country imports.

Thus the total trade effect of changes in import prices comprises both trade creation and substitution, or trade diversion, between imports. As the latter involves an increase in the share of imports by one group of countries at the expense of the other, it results in no net trade expansion.

However, the tariff reductions for forest products agreed in the Uruguay Round differ across import markets. Before the Round, some countries display similar demand and/or tariff structures for both developing and developed country imports; other countries do not. Tariff reductions are also not always equal across products. For example, some countries have reduced MFN rates more than GSP rates, and others have reduced only one rate (usually the MFN). The above relationships have therefore been adapted to fit the various import market and tariff reduction scenarios of the Uruguay Round tariff removals on forest products.

The revised methodology considers three different cases that must be allowed for:

· Case 1 Uniform import demand and tariff reduction - in this case there is relatively little difference in the demand for developed and developing country imports of a forest product and all imports face the same tariff structure; the only effect of tariff reduction will be trade creation, which can be distinguished by developed and developing country import share.

· Case 2 Differentiated import demand and tariff reduction for some imports - in this case the demand and tariff structures for developed and developing country imports of a forest product are the same, but reduction in the tariff for imports from only one of the source occurs (e.g. a change in the MFN rate only); trade creation occurs only for imports from the source affected (e.g. developed country imports), and trade diversion consists entirely of substitution of imports from one source to the other (e.g. from developing country to developed country imports).

· Case 3 Differentiated import demand and tariff reduction for all imports - in this case the demand for forest product imports from developed and developing country imports is differentiated, and both sources of imports will have their tariff rates reduced but by different amounts; total trade creation will be the sum of the trade creation effects of the resulting price changes on both sources of imports, and trade diversion will be a net effect depending on the relative tariff changes and import demands.

Case 1 Uniform import demand and tariff reduction

If there is relatively little difference in the demand for developed and developing country imports of a forest product and all imports face the same tariff structure, then (1) becomes

Q = Q(P^d, P^m) (5)

M = M(P^d, P^m).

As a consequence, import tariff reduction is uniform and has the following effect on the import price

dP^m = -P^wdt^m . (6)

The only trade effect is trade creation, which is

- dQ = - (_Q/_ P^m)dP^m = dM = (_M/_ P^m)dP^m = (_M/_ P^m)P^wdt^m . (7)

Denoting the own price import elasticity of demand by _^m, (7) can be rewritten as

- dQ =TC = M_^m_^m, where _^m = dt^m/(1+ t^m), (7)

or - dQ =TC = _^m_^m(M_a + M_b)/M,

to distinguish trade creation by developed and developing country market share, M_a and M_b, respectively. Aggregating across each forest product market, i, yields total trade creation equal to

__iTC_i = _^m_i_^m_i(M_ia + M_ib)/M_i . (8)

Case 2 Differentiated import demand and tariff reduction for some imports

This scenario of tariff reduction is similar to the case examined in detail by Baldwin and Murray (1977) through graphical and empirical analysis. Demand and tariff structures differ for developed and developing country imports of a forest product. Thus demand is characterized by (1), and assume that t^m_a >t^m_b . However, suppose that just t_a is reduced so that only the price of these imports are affected

dP^m_a = - P^wdt^m_a . (9)

Total trade creation is now

- dQ =TC =TC_a = M_a_^m_a_^m_a, (10)

with _^m_a = dt^m_a/(1+ t^m_a) and _^m_a and is the own price elasticity of demand for developed country imports. Across all forest product imports total trade creation would therefore be

__iTC_i = __iTC_ia = __iM_ia_^m_ia_^m_ia (11)

across all forest products.

Because there is a unilateral import price reduction, trade diversion occurs and is equal to

M^S_a,b = - (_M_b/_ P^m_a)dP^m_a = (_M_b/_ P^m_a)P^wdt^m_a = dM^S_b > 0 (12)

M^S_a,b = - dM^S_b .

That is, trade diversion consists entirely of substitution for developing country imports by developed country imports. An important assumption by Baldwin and Murray is that the substitution response to a change in P^m_a will be similar for both domestic output and developing country imports. This suggests that the cross-price elasticities are the same and thus in this case

(dQ/Q)/(dP^m_a/P^m_a) = (dM^S_b/M_b)/(dP^m_a/P^m_a) (13)

or dQ/Q = dM^S_b/M_b .

From (10) it is clear that - dQ/Q = (TC_a)/Q and thus combining with (13) yields

dM^S_b = (M_b/Q)TC_a = (M_b/Q)M_a_^m_a_^m_a (14)

or __idM^S_ib = __i(M_bi/Q_i)TC_ia

across all forest products.

Condition (14) is equivalent to the expression derived by Baldwin and Murray. It is extremely useful for it allows trade diversion (in this case) from developing to developed country imports to be estimated without employing cross-price elasticities.

Case 3 Differentiated import demand and tariff reduction for all imports

In many instances, not only will the demand and tariff structure for forest product imports from developed and developing countries be differentiated but also the Uruguay Round tariff reductions may also differ for these two categories of imports. For example, MFN rates may be reduced more than GSP rates. This implies that (in absolute value terms)

dP^m_a = P^wdt^m_a > dP^m_b = P^wdt^m_b . (15)

Total trade creation is therefore

-dQ =TC =TC_a + TC_b , (16)

with TC_a = M_a_^m_a_^m_a

TC_b = M_b_^m_b_^m_b ,

where _^m_b = dt^m_b/(1+ t^m_b) and _^m_b and is the own price elasticity of demand for developing country imports. Across all forest product imports trade creation would therefore be

__iTC_i = __iTC_ia + __iTC_ib = __iM_ia_^m_ia_^m_ia + __iM_ib_^m_ib_^m_ib . (17)

From (4) it is clear that net trade diversion from developing country imports to developed country imports will be

M^S_a,b = (_M_a/_ P^m_b)dP^m_b - (_M_b/_ P^m_a)dP^m_a (18)

= - (_M_a/_ P^m_b)P^wdt^m_b - (_M_b/_ P^m_a)P^wdt^m_a

= dM^S_b - dM^S_a ,

and M^S_b,a = - M^S_a,b .

Assuming condition (13) still holds in this case, it follows that

dM^S_b = (M_b/Q)TC_a = (M_b/Q)M_a_^m_a_^m_a . (19)

A similar condition can also be derived for dM^S_a. That is, if the substitution response to a change in P^m_b is also assumed to be similar for both domestic output and developed country imports, then the cross-price elasticities are the same, i.e.

(dQ/Q)/(dP^m_b/P^m_b) = (dM^S_a/M_a)/(dP^m_b/P^m_b) (20)

and dQ/Q = dM^S_a/M_a .

Note that in (19) and (20) dQ represents the change in domestic output that is attributed to substitution by developed and developing country imports respectively. From (15) and (20) the following condition is derived

dM^S_a = (M_a/Q)TC_b = (M_a/Q)M_b_^m_b_^m_b. (21)

Trade diversion is then

M^S_a,b = (M_b/Q)TC_a - (M_a/Q)TC_b (22)

or __iM^S_a,b = __i(M_bi/Q_i)TC_ia - __i(M_ai/Q_i)TC_ib

across all forest products.

Rules (16) and (22) of Case 3 clearly represent the trade creation and diversion rules in their most generalized form. If tariff removal is only partial (i.e. dt^m_b = 0), then (16) and (22) reduce to rules (10) and (14) respectively of Case 2. Similarly, if _^m_a = _^m_b and _^m_a = _^m_b then under rule (22) there will be no trade diversion and trade creation will reduce to rule (7) of Case 1. Thus Cases 1 and 2 are essentially specific versions of the more general Case 3.

The above three cases provide the full set of rules for estimating the trade creation and diversion impacts of the Uruguay Round tariff reductions on forest product imports. These rules are the basis for the various calculations in this paper of tariff changes for selected forest products and import markets.

(b) Procedures

Utilizing the above methodology and the Uruguay Round tariff changes, the trade creation and diversion effects were estimated for selected forest product imports in key developed and developing country markets. Given data limitations, it was possible to estimate only the trade impacts on imports of logs, sawnwood, veneer, particleboard, fibreboard, plywood, wood pulp and newsprint. Table 4 in the report shows the summarized results of the analyses by developing and developed country imports and also the total trade effects.

For each forest product in an importing country market a tariff effect was calculated using the Uruguay Round tariff reductions for the product (if any). For developing country imports into a developed country markets the tariff effect for both MFN rates and GSP rates were included. Although it was generally assumed that all developing countries may benefit from GSP rates on imports in major developed country markets, this was not always the case.

Accurate estimates of own-price elasticity of demand for each type of forest product were not always available, particularly estimates that distinguish developed from developing country imports. Where possible, the demand elasticity estimates used were verified against empirical studies of import demand for specific forest products in various markets.

The level of imports for each forest product in an importing market was based on 1991 forest product trade flows contained in FAO (1994), which were disaggregated by import source using the UNCTAD TRAINS data base. As domestic production and crucial trade data for furniture, builders' joinery, packing cases and some pulp and paper products were not included in these statistics, trade creation and diversion effects could not be estimated for any of these products experiencing tariff reductions.

The trade impacts were estimated for each forest product in an import market experiencing a tariff reduction. The trade changes were expressed in terms of the total value of imports, and where possible, also in terms of the total volume of imports. For both developed and developing country sources of imports, the trade changes differ depending on the estimates of import demand elasticities used for each forest product. In the case of developing country imports, the trade changes also differed because of the assumption that all developing country imports face GSP rates as opposed to MFN rates in developed country markets.

Given that solid wood imports are recorded in different physical units (cubic metres, or cum) than paper products (metric tonnes, or mt), it was not always possible to depict the total trade change estimates in volume terms. Thus for comparison the total trade change was calculated in value terms.

Given the assumptions behind the calculations of trade effects, a range of six different estimates of the total value of the trade changes were estimated for each importing market, corresponding to low, medium and high assumptions concerning the price elasticity of import demand, and in addition, whether it was assumed that all developing country imports faced MFN or GSP rates initially in developed country markets. Two of the middle estimates in the range were used in the summary table (Table 4). Scenario A in that table assumes a medium elasticity of demand for forest product imports and that pre-Uruguay Round tariffs on imports in developed country markets are at GSP rates for developing country imports and MFN rates for developed country imports. Scenario B assumes a medium elasticity of demand for forest product imports, and pre-Uruguay Round tariffs in all markets are at MFN rates.