How would one go about estimating these “appropriate” levels of bindings? As said earlier, the main justification for appropriate tariffs was concerns about potential disruptions to agricultural sectors that are vital for food and livelihood security. The major source of the disruption mentioned is external shock, manifested in terms of depressed world prices or/and import surges. Other arguments made in the proposals are vulnerability, level of food insecurity, level of development, and domestic and export subsidies by trading partners. Among these, only world market prices provide an objective basis with high data frequency (e.g. on a monthly basis) to quantify tariff rates, besides also being commodity-specific.
Ignoring other factors that do not normally vary with world market prices but nevertheless create a wedge between these and domestic prices, e.g. exchange rates, transportation and marketing margins etc., a simple expression (1) relates the world market price and tariff to the domestic price.
Pd = Pw * (1+t) (1)
Pd = domestic market price
Pw = world market price
t = ad valorem tariff
Rearranging (1), the tariff rate or “t” in percentage terms is:
t = [(Pd - Pw )*100/ Pw ] (2)
In a tariff-based stabilization framework, Pd is the target or reference price that a policy maker is striving to maintain or stabilize with the help of the tariff, and thus is a pre-determined or given number. In that case, it is the tariff, “t” that varies with each change in Pw (e.g. on a monthly basis if monthly world prices are used for analysis, as is done in this paper). The interest here is in instances when the world market prices are lower than the reference price, i.e. when tariffs are positive.8 For complete stabilization of domestic prices over some periods (when world prices are fluctuating), applied tariffs would be varied depending upon the price gap - from a level of zero percent to a maximum when world market prices are lowest relative to the reference price. The maximum of these duties would then be the tariff that is required for complete price stabilization all the time. We may call this maximum the bound rate in the WTO sense. The purpose here is to compute this number.
To implement this method, a key parameter that must be known in advance is the reference or stabilized price that is being defended. In theory, this price can be any number but it obviously needs to be “reasonable”. As an example, if one were to choose the maximum world price of wheat observed for any month during 1990-2000, which was $263 per tonne, a maximum tariff of 150 percent would be required for complete price stabilization during those 132 months9. Obviously, choosing a highest-ever world price for reference purposes is unreasonable.
The analysis is based on monthly world market prices for 132 months, January 1900 through December 2000, with the exception of poultry and pork for which the data series begin in January 1993. Although this period is history now and the outcome of the new negotiations is for future implementation, what matters is price fluctuations around a trend. Studies have shown that there are no particular trends in price fluctuations for most commodities, and so the computed maximum tariffs will remain valid for the future as well.
Is $150 per tonne a reasonable reference price for wheat in the 1990s that import-competing countries should desire or aspire to maintain for their farmers, or is it $120? Obviously, there is no unique answer to this question.10 In what follows, a number of reference prices that are considered reasonable are illustrated and maximum tariffs computed. The objective is not to advocate a particular reference price or such a scheme but to derive a range of maximum tariffs that would be required for price stabilization. Five alternative references are explored - using the full range of trend values of world prices for 1990-2000, average world price for 1990-2000, average world price for 1990-94, 36-months and 24-months moving averages of world prices.11
Trend world market prices as reference price. One option to consider would be to stabilize domestic prices along a “smoothed” line of opportunity values (i.e. world market prices), which makes some economics sense. Figure 1 illustrates, for raw sugar, 12 the pattern of applied tariffs that stabilize domestic prices under this option. Tariffs were triggered in the early 1990s and during 1998-2000 when world prices were below their trend values, while tariffs were not required for the middle years. The maximum tariff required for complete price stabilization during 1990-00 was 75 percent (which was required for March 2000 - Figure 1), although applied rates would be much lower most of the time.
Fixed reference prices derived from past world prices. An alternative to trend reference prices is a fixed reference price. Two variants are reviewed here. In one, the reference price is the average of actual world prices during 1990-2000, and, in the other, the average of 1990-94 prices. There is some rationale in these choices. During the 1990s, world prices of many basic foods underwent through three distinct phases - a period of relatively stable prices in the early years, one or two spikes around 1994 and 1995 and a phase of depressed prices. As a result, the price series do not show particular strong trends but seem more to behave like random fluctuations around some fixed value - hence the rationale for averaging the entire 1990-2000 prices. The 1990-94 average for the reference price is also interesting in that world prices for many basic foods were relatively stable then, prior to the spikes of the mid-1990s and depressed prices thereafter.
Figure 2 illustrates one of these options. With fixed reference price of $226/mt, the maximum tariff required for complete stabilization is 101 percent (needed for March 2000). The second option also gives a fairly close maximum tariff of 106 percent (also for March 2000). Although the two reference prices turned out to be close for raw sugar, this is only by chance. For many other products, the reference prices differ and so do the computed maximum tariffs.
Reference prices based on moving averages of world prices. Moving averages (MAs) have the effect of smoothing the otherwise volatile time-series, such as world market prices. Moving averages of world prices were used extensively in the price band schemes (and their minimum price variants) implemented by several countries in Central and Latin America in the 1980s and 1990s. They are still being used by some countries (FAO 2001b). The number of months used for these MAs is somewhat arbitrary - while a greater smoothing is achieved with longer price series, there is a risk that the resulting MAs may deviate too far from current world price trends. In this paper, the effects of two MAs are reviewed - 36 months (MA36) and 24 months (MA24).
Figure 3 illustrates the case of the MA36 for raw sugar. Since the 1990-92 data were used up for computing the first MA observation, the MA series begins only in January 1993. Except for two months early on, the MA36 prices lie below actual world prices until the beginning of 1996 and so tariffs remained zero. Tariffs were triggered from around mid-1996 with the maximum tariff reaching 86 percent in April 1999. With MA24, the maximum tariff was only 75 percent (also triggered in April 1999).
8 The reverse case is not relevant here. When Pw exceeds Pd, the resulting "t" is negative, which corresponds to an import subsidy, i.e. an import subsidy has to be given to stabilize (reduce, in this case) domestic prices. Developing countries rarely subsidize imports.
9 This level of the tariff would have been required for the month of December 1999 when the world price of wheat fell to its lowest level of $105 per tonne.
10 It is worth remembering that the policy objective is price stabilization and not price protection. Price protection would require a higher level of tariff than is sufficient for price stabilization.
11 Many countries fix domestic farm support prices based on considerations such as cost of production and inflation, in addition to world market prices. Although such prices may be "fair" to farmers, relatively speaking, they would not provide the right signal to farmers if the support prices deviate too far from world price trends - the opportunity values. As countries open up their markets increasingly to trade, reference prices that diverge too far from world price trends are unlikely to be an attractive option. Therefore this option is not pursued in this paper.
12 Raw sugar prices have been used for illustrating various reference prices, although any of the 22 price series used in this paper would do the job equally well.