An analysis of triggers for the Special Safeguard Mechanism^[12]

Ramesh Sharma

That a Special Safeguard Mechanism (SSM) would be established for developing countries was agreed in the 2004 Framework Agreement of the WTO, as well as in other key framework texts. This was a response to many negotiating proposals put forward by these countries that brought to light the problems posed by import surges and depressed import prices and the need for a simpler-to-use safeguard until capability is developed for general trade remedy measures. This paper contributes to one key building-block of such a SSM - the mechanism for triggering a safeguard response. Two alternative trigger schemes are evaluated - the Agreement on Agriculture’s Special Safeguard (SSG) formula and one based on moving averages of prices and imports. There are advantages and disadvantages with both schemes. The three-year moving average reference price does a good job most of the time in triggering a safeguard when prices are depressed, but misses depressed prices in about 20 percent of the cases. On the other hand, a five-year moving average reference price triggers safeguard even in these cases, although the latter triggers safeguards too frequently overall.

1. INTRODUCTION

Agricultural markets are by nature cyclical and subject to wide fluctuations due, among other things, to weather variability. Other sources of instability include the subsidization of production (such as with counter-cyclical programmes) and exports, as well as anti-competitive behaviour of trading firms (both state-owned and private). All these affect the orderly development and flow of trade. As countries reduce tariffs and bind them at lower levels, they become increasingly vulnerable to external agricultural market instability and to import surges that could wipe out viable, well-established or nascent, agricultural production activities. Vulnerability to such external shocks is of particular concern to developing countries that are endeavouring to develop their agricultural potential and diversify production.

That import surges and depressed import prices pose threats to domestic market stability is no longer contested. There have been many reports of developing countries, particularly lower-income food-deficit countries, experiencing increasing numbers of import surges of various food products, notably since the mid-1990s. Often, these reports associate the surge with negative effects on local production and economy. Sharma (2005) documents 30 such reports and studies, all for the late 1990s and early 2000. Examples include the experience of Jamaica with respect to chicken, Kenya with respect to dairy products, Senegal with respect to tomato paste, and rice in Haiti. This review draws on several sources, notably various studies by FAO (FAO 2000; FAO 2003a; FAO 2003b; and Sharma et al., 2005) and national and international civil society organizations (e.g. Action Aid, 2002; APRODEV, 2004; Ceesay et al., 2005, Christian Aid, 2005; OXFAM, 2002, 2003 and 2004). There is widespread concern that these problems are likely to intensify in coming years as tariffs are lowered further, and before alternative forms of safeguards can be put in place. In several of these cases reported, imports increased by as much as 10-20 fold within a short period of 4-5 years with marked negative impact on domestic production, industry and employment. The phenomenon has been found to be relatively frequent for certain product groups, notably dairy products, poultry and some other meats, rice, sugar and vegetable oils.

Because of the negotiating history of the Uruguay Round (UR), the recourse to the Special Safeguard (SSG) of the Agreement on Agriculture (AoA) was limited to those countries undertaking tariffication. As a result, there is now the anomaly that some have the right to use the SSG to deal with import surges, whereas others, including many vulnerable developing countries, do not, despite the fact that all countries have embraced the tariff-only trade regime and are equally vulnerable to external shocks. In view of the above, a consensus has been reached in the WTO that there should be a simple to use special safeguard instrument accessible to the developing countries. This instrument is the Special Safeguard Mechanism or SSM.

The five key issues in the design of a SSM are: country eligibility; product eligibility; triggers; remedy; and duration. These are discussed at length in FAO (2005) and Sharma and Morrison (2005). Of these five, the trigger mechanism is of a relatively technical nature and is the subject of this paper.

2. DESIGNING EFFECTIVE TRIGGERS FOR THE SPECIAL SAFEGUARD MECHANISM

While almost all WTO negotiating proposals have viewed positively the idea of a special agricultural safeguard for use by the developing countries, hardly any concrete proposal was made on the form of that safeguard until recently. It was only in October 2005 that G-33 - the negotiating group that has been the main demander of such an instrument, the SSM - made a concrete proposal. Until then, most technical discussions had revolved around the mechanisms and formulas of the SSG of the AoA. In the meantime, various negotiating proposals and statements had stressed some or all of the following basic features as being desirable for the SSM, notably in relation to the triggers:

• Simple and transparent
• Relatively easy to invoke - not burdensome administratively
• Triggered in reaction to exceptional market conditions
• Remedy measures to be temporary in nature
• No requirement for proof of injury
• Should not lead to misuse - too frequent triggers, for example

As the SSM is expected to be similar to the SSG in terms of simplicity and effectiveness - and without the need for an injury proof - it was also generally expected that the SSM trigger mechanism would be similar to the SSG triggers: namely, automatic triggers for both depressed prices and import surges. With the new G-33 proposal, there are now two concrete schemes on the table and which are analysed here. The two trigger mechanisms are explained below, and evaluated primarily against the criterion of frequency of triggers.

2.1 The price trigger

As a safeguard, it is desirable that the SSM should be effective in responding to sharp, short-term price depressions. The key parameter that ensures this is the reference price used for triggering the safeguard. In addition to the two trigger mechanisms considered here, namely the SSG formula based on a fixed reference price and the one proposed by G-33 based on moving averages, other mechanisms have been tried by different countries as well as proposed in the literature. For example, alternative mechanisms have been used by some countries in Latin America for their price band schemes. In some cases, the triggers were based on moving average world market prices (as in Chile) while in others these were based on domestic prices.

Other possibilities have been discussed in the literature. For example, Valdes and Foster (2004) discuss the use of price trends, moving averages of various lengths, preceding year’s price and also minimum average cost of production of the world’s most efficient exporter as potential reference prices. The main concern and consideration here is with avoiding excessive or too fewer triggers which result when current import prices bear no relationship to the trigger prices.

2.1.1 The SSG price trigger of the Uruguay Round

The formula for the price trigger of the SSG is shown in Box 1. In this formula, a safeguard is triggered when the current import price exceeds the trigger or reference price, which is fixed as the average import price for 1986 to 1988. Figure 1 illustrates how the formula works. Essentially, once the safeguard is triggered, the additional duty levied varies depending on the depth of the price depression. The figure also shows that the additional duty does not completely offset the fall in the import price. For example, additional duties would amount to 4 percent, 28 percent and 170 percent when the import price falls below the trigger price by 20 percent, 50 percent and 80 percent respectively.

2.1.2 Trigger based on moving-average reference price - the G-33 proposal

In contrast to the fixed reference price scheme of the SSG, the G-33 in its October 2005 proposal suggested the use of three-year moving average prices as the reference price. The price trigger will be equal to the average monthly price for the most recent three-year period, preceding the year when a safeguard is triggered, and for which data are available. A safeguard is triggered when the current import price exceeds the trigger price.

2.1.3 Assessing the alternative proposals

The two important criteria for assessing the desirability of a reference price for the purpose of the safeguard would be simplicity and effectiveness. The instrument should be effective in the sense that it triggers the safeguard when needed, but at the same time it is also important that safeguards are not triggered too frequently.

From the standpoint of simplicity, a fixed reference price has a distinct advantage over the moving-average (MA) price in that once fixed for the implementation period, the trigger price is known in advance. Moreover, the SSM will require a substantial amount of data where the instrument is being extended to all tariff lines. In addition to assembling the price statistics for so many tariff lines, countries will need to update these on a continuous basis for computing the MA prices. The fixed reference price, on the other hand, does not require updating.

The main disadvantage of a fixed reference price is that it does not incorporate information on price trends, unless updated periodically. As a result, when current prices deviate significantly from longer-term trends, safeguards may be triggered inappropriately. The base period chosen also becomes very important. For example, if this happens to correspond to a cyclical high in world markets, the SSM will be triggered more frequently than is desirable, while if it corresponds to a cyclical low, SSMs will be triggered too infrequently.

By contrast, reference prices based on MAs, or other similar trend-based references, incorporate information on the long-term movement of commodity prices. Where the length of the MA is chosen appropriately, the reference prices may reflect more accurately the opportunity costs of domestic production. The shorter the memory, the more sensitive is the trend to sharp but short deviations in prices which may not be representative of the long-run opportunity costs.

One potential flaw with MAs, despite their attractiveness as a means of smoothing price fluctuations, is that the reference price based on MAs can produce outcomes that are inconsistent with the objective of protecting against exceptionally low prices. In the case of a downturn in import prices, the moving average price follows the actual price gradually and after a delay depending upon the number of periods in the moving average. As a result, the moving average price may fail to trigger a safeguard in the face of a persistent fall in import prices. When import prices are rising, the reference prices remain below the rising actual prices and so do not trigger a safeguard - which is a desirable property.

In order to illustrate these points, world market prices (unit import prices) for 10 primary and processed agricultural products over 21 years were analysed. Figure 2 plots actual prices against two reference prices - a three-year moving average (MA-3) price as in the G-33 proposal, and 1986-88 average price (the SSG reference price).

TABLE 1
Total number and percent of safeguard triggers during 1983-2003 for various reference prices

Note: Total number of potential triggers for a product is 21 (i.e. 21 years covered). The percent of triggers on the right side of the table is potential percent of triggers, the potential number of triggers being 21 (21 years), and for the "all products" total, potential number of triggers is 210 (21 years times 10 products). A safeguard is triggered when current prices are below 90 percent of the reference price (i.e. 10 percent de minimis level is assumed).

Source: Author.

The problem with the MA-3 formula is apparent from the graphs. Taking raw sugar as an example, the MA-3 does not trigger a safeguard in 1985, a depressed year, nor in 2001 and 2003.^[13] This is because current prices in 1999 and 2000 were already very depressed and as a result the MA-3 prices for the next two years were too low to trigger a safeguard. For wheat, the same thing happens in 2001 and 2002. Many cases like this can be found for the 10 commodities. The case of refined sugar is similar for the three years, 2001 to 2003, when the MA-3 does not trigger a safeguard despite these being years of depressed prices. Overall the MA-3 did not trigger a safeguard in about 20 percent of the cases when prices were depressed. This is the main limitation of the MA-3 reference price.

Additional investigation was made with a five year moving average (MA-5). The longer the period of the MA, the smoother the reference price is and so the less likely it is that years of depressed prices will fail to trigger a safeguard. Table 1 compares the number of triggers with MA-3 and MA-5. It shows that in the 21 years covered, taking into account all 10 products, the MA-5 triggers the most, 40 percent of the time, followed by MA-3 (32 percent) and SSG trigger (16 percent). Not only is the number of triggers more with MA-5, it was also found that the MA-5 reference price triggers safeguards in periods of depressed prices that the MA-3 missed, as noted above. This is the major advantage of the MA-5 over MA-3. On the other hand, it appears that the MA-5 seems to trigger safeguards too frequently, which may not be a desirable property.

A scheme based on a fixed reference price obviously triggers safeguard in periods of depressed prices where the reference price (i.e. the base period) is carefully chosen. A reference price fixed for a high-price cycle in world markets will trigger the safeguard too frequently while one based on a low-price cycle may not be effective. Figure 2 shows several cases where the fixed reference prices are relatively high (e.g. paddy, raw sugar) as well as relatively low (e.g. wheat, husked rice and shelled groundnuts). The choice of base period is crucial. This is illustrated in Figure 3. It shows the frequency of the SSG triggers for various reference prices, based on three-year averages since 1995.^[14] For almost all the 16 products analysed, world market prices were high initially (i.e. during 1995-97), depressed during 1999-2001 and higher again during the last 2-3 years. As a result, the number of the SSG triggers was large with the reference price for 1995-97 (a total of 83 triggers out of the 160 potential triggers (16 products times 10 years covered), but very low (only 17 triggers) when 1999-01 was used for the reference price.

The challenge obviously lies in deciding on a particular period that is appropriate for a safeguard. Based on Figure 3, if the SSM is to cover exceptional cases only, then reference periods such as 1998-2000 or 2001-03 would appear appropriate because these reference prices cover the "exceptionally" depressed years observed around 1999-2001. On the other hand, while the reference periods of 1999-2001 and 2000-02 provide very little safeguard for the low years, reference periods such as 1995-97 or 1996-98 would trigger the SMM too frequently.

2.3 The volume trigger

As with the price trigger, this sub-section assesses the two prominent triggers - the SSG formula and the MA-3 formula proposed in the latest G-33 proposal.

2.3.1 The SSG volume trigger

The trigger volume in this case is derived from: i) actual imports averaged over the preceding three years; ii) the share of imports in domestic consumption over the same period; and iii) the absolute volume change in consumption over the most recent year for which data are available (Box 2). The trigger level is higher (and the probability of using the trigger less), the greater the three-year average level of imports, the lower the share of imports in domestic consumption, and the faster the growth in domestic consumption. Additionally, the relevant AoA provision states that the maximum extra duty may not exceed 30 percent of the ordinary level of duty in effect during the year in which the SSG is invoked; it may not be levied beyond the end of the year in which it has been imposed; and it cannot be applied to imports taking place within tariff quotas.

2.3.2 Trigger volume based on moving averages of imports

The October 2005 G-33 text proposed, similar to that for the price trigger, a reference or trigger import level based on a moving average of imports. It is proposed that the trigger level of import should be equal to the average annual volume of imports for the most recent three-year period preceding the year of importation for which data are available. A safeguard is triggered when current import volume exceeds the trigger level, subject to some additional provisions - no trigger for a de minimis level of import, for example.

2.3.3 Assessing the volume triggers

The key issue in assessing the two alternative volume triggers is their relative effectiveness in responding to import surges. Figure 4 compares the performance of the SSG formula and the MA-3 formula by simulating the triggers for rice for four sample countries selected randomly.

The figures show that the SSG formula would have triggered much less frequently than the MA-3 formula. For the four countries and 10 years (40 observations in all), the SSG formula would have triggered a safeguard only five times (13 percent of cases) while the MA-3 formula would have triggered a safeguard 17 times (43 percent of cases). Arguments can be made in favour of both formulae. Since a SSM is part of the Special and Differential Treatment meant to be for use by the developing countries, with more vulnerable agriculture, more frequent triggers might be considered desirable. On the other hand, too frequent triggers might be regarded as undesirably disruptive of trade.

A alternative way to assess the two formula could be to determine whether or not they would trigger a safeguard when there is an import surge. In the WTO trade remedy measures, import surges are defined generally: there is no objective measure of a surge as there is in the SSG.^[15]

In disputes involving the Safeguards Agreement, panels have generally decided whether a surge exists or not by first looking at the data on import trends. For example, the 56 000 tonnes increase in imports in Costa Rica between 1999 and 2002 would probably have been regarded as a surge by the dispute panel. Similarly for Guatemala between 1998 to 2001. Yet, in both cases, the SSG formula did not trigger a safeguard because the trigger import volume exceeded the current import volume.^[16]

There are many such cases where the SSG formula would not have triggered safeguard in the face of apparent import surges. In some cases, the reason why the trigger import volume exceeded the current import level (and hence a safeguard was not triggered) was because the change in consumption in the previous years was negative. This in turn was the result of increased production in those years that pushed higher the (apparent) consumption. For example, a much higher apparent consumption in the previous year due to a good harvest prevented a trigger later in 2001 for Guatemala. However, based on the import trends (Figure 4), 2001 should have qualified as a surge year, as any WTO Panel in a Safeguards dispute would probably have concluded.

A key problem with the SSG formula is that the current conditions for trigger are influenced by past events - up to three years previously and which have little to do with any current surges a country may be experiencing. It is for this reason that a SSG is inappropriately triggered for rice in Honduras in 1999 when imports were actually lower than in 1998. This is an awkward outcome and should not have happened.

Comparing the left and right-side graphs in Figure 4, it is possible to assess some other features of the two formula. One was noted earlier - that the MA-3 formula leads to more frequent triggers than the SSG formula. Other than that, some of the problems noted above with respect to the SSG formula are also found with the MA formula. This follows because both the formulae rely to a large extent on moving averages for trigger volumes. As a result, one finds that the MA-3 formula would trigger a safeguard in 2003 for Costa Rica when rice imports actually declined. This is also the case for Guatemala in 2002 and 2003, for Honduras in 1999 and for Nicaragua in 2001. At the same time, it might be questioned whether some of the occasions when the SSM was triggered could actually be characterized as surge years.

In summary, as was also noted in the discussion on price safeguard above, a major source of anomalies in outcomes are historical imports that are embedded in the MA approach, and so applied to both the SSG and MA-3 cases. In the case of the SSG formula, past consumption and production also play a role. It is difficult to understand why these past events - some of them taking place three years back - should influence the decision on a trigger now. The WTO general trade remedy measures do not assign such weights to historical parameters.

There is a considerable room for simplifying the SSG volume trigger by assigning little or no weights to the past events and basing the decision about the trigger on the most recent developments in imports (over the past 4-6 months, for example), in relation to some benchmark of what constitutes excessive imports that disrupt, or threat to disrupt, domestic markets to the extent that producers are injured. One such benchmark would be the gap between normal import needs and actual imports. The former would be determined by trend consumption and current production while the latter would be actual imports in the most recent months. A safeguard would be triggered when actual imports exceed that benchmark level. Such a formula would not only be very simple but would also have much more intuitive appeal than the current SSG formula.

Further to the discussion above, previous papers on this subject (FAO 2005; Sharma and Morrison 2005) discuss other features of the SSG volume trigger that need to be revisited for the purpose of the SSM. Those papers argue that the SSG formula may have some built-in biases against agricultural economies as exist in the least-developed and other lower-income developing countries, for the following reasons:

• the formula is biased against countries with lower degrees of openness (i.e. lower import to production or consumption ratios, which is more common for poorer countries);

• the formula is biased where consumption is rising (biased in the sense that the chance of a trigger is reduced and food consumption growth is typically higher or stronger in lower-income countries);

• the formula raises the level of the trigger imports where consumption statistics are missing, and this reduces the probability of triggering a safeguard (consumption statistics are generally lacking in lower-income countries).

3. CONCLUSIONS

That import surges and periods of prolonged depressed world market prices can be significantly disruptive to agricultural development and livelihoods in the developing countries is hardly questioned. There is also a consensus that these countries will require a simple-to-use safeguard until such time as they develop their capability for alternative forms of safeguards, including the use of general WTO trade remedy measures. Moreover, it is recognised that because of the negotiating history of the UR, there is now the anomaly that some countries have the right to use the SSG while others - notably many vulnerable developing countries - do not. The response to these concerns has been the agreement that a SSM will be established for the developing countries.

Various negotiating proposals and texts have commented on the desirable features of such a safeguard. These included simplicity, transparency, administratively straightforward to invoke, and effective in responding to exceptional market conditions. Yet, with the exception of the G-33 proposal of late October 2005, none of the proposals including the 2004 Framework Agreement has been specific on the technical "design" elements of the SSM. It was generally held that the SSM would be technically similar to the SSG.

As a contribution to the discussion of the technical aspects of the SSM, this paper evaluated two formulae for price and import volume triggers - the SSG formula and one based on three year moving averages (MA) as proposed by G-33. There are advantages and disadvantages with both schemes.

The three-year moving average reference price does a good job most of the time in triggering a safeguard when prices are depressed, but misses out depressed prices in about 20 percent of the cases. Specifically, it misses out in those periods when depressed prices are persistent. Several such cases were noted in Section 2. This happens because where current prices are depressed in the previous 2-3 years, the MA-3 prices for the next two years become correspondingly depressed and thus fail to trigger a safeguard. On the other hand, a five-year moving average reference price triggers safeguard even in these cases, although it might be considered to trigger safeguards too frequently overall.

In the case of the volume trigger, it was shown that the current SSG formula also suffers from some of the problems as above, notably that it does not always trigger a safeguard when the data on import trends clearly indicate that a surge - generally defined as in the WTO Safeguard Agreement - is actually occurring. The problem was traced to the weight that some of the developments that take place in the past, namely changes in production and consumption levels two or three years back, carry in the current decision on whether or not to trigger a safeguard. As a result, in a number of cases a safeguard is not triggered when it should have been. It was also noted that the MA-3 formula triggers safeguards more frequently than the SSG formula. While this may or may not be taken as a desirable feature, the MA-3 formula also suffers to some extent from the same problems as was noted in the case of the MA-3 price trigger: because of the weight given to past developments, a safeguard is triggered even when imports are actually falling. Thus, there are problems with both formula and so more experiments with alternative formulae are needed before a decision is made on the SSM triggers.

REFERENCES

Action Aid. 200). Farmgate: the developmental impact of agricultural subsidies, available at www.actionaid.org

APRODEV. 2004. The effects of massive dumping of frozen poultry imports from the EU into Cameroon and West Africa.

Ceesay, M., Njie, M. & Jagne, M. 2005. The effects of importation of poultry meat and eggs on small-scale poultry producers in the Gambia. Study commissioned by Action Aid (The Gambia) and Oxfam International (The Gambia).

Christian Aid. 2005. For richer or poorer - transforming economic partnership agreements between Europe and Africa. www.christianaid.org.uk

FAO. 2000. Synthesis of the Country Case Studies, in Agriculture, trade and food security: issues and options in the WTO negotiations from the perspective of developing countries, Volume II, country case studies. FAO, Rome. Available on-line at www.fao.org/trade.

FAO. 2003a. WTO Agreement on Agriculture implementation experience: developing country studies. FAO, Rome. Available at www.fao.org/trade.

FAO. 2003b. Some trade policy issues relating to trends in agricultural imports in the context of food security, Committee on Commodity Problems, CCP 03/10. http://www.fao.org/unfao/bodies/ccp/ccp64/ccp64-e.htm

FAO. 2005. A special safeguard mechanism for developing countries. FAO Trade Policy Technical Note Number 9. FAO, Rome.

OXFAM. 2002. Milking the CAP - how Europe’s dairy regime is devastating livelihoods in the developing world, OXFAM Briefing Paper number 34.

OXFAM. 2003. Dumping without borders: how US agricultural policies are destroying the livelihoods of Mexican maize farmers, Oxfam Briefing Paper No. 50.

OXFAM. 2004. A raw deal for rice under DR-CAFTA: How the Free Trade Agreement threatens the livelihoods of Central American farmers, OXFAM Briefing Paper No. 68.

Sharma, R. 2005. Overview of reported cases of import surges from the standpoint of the analytical content, FAO Import Surge Project Working Paper No. 1.

Sharma, R. & Morrison, J. 2005. Considerations for the design of a Special Safeguard Mechanism. Paper presented at the IATRC summer symposium, Seville, Spain, 19-21 June, 2005.

Sharma, R., Nyange, D., Duteutre, G. & Morgan, N. 2005. The impact of import surges: country case study results for Senegal and Tanzania, FAO Commodity and Trade Research Working Paper No. 11. FAO, Rome. Available at http://www.fao.org/es/ESC/en/41470/index.html

Valdes, A. & Foster, W. 2004. Special safeguards for developing country agriculture in WTO negotiations. In G. Anania, M. Bohman, C. Carter and A. McCalla (eds.) Agricultural policy reform and the WTO: where are we heading? Edward Elgar, London.