Jan 10 (Reuters) - A controversial method of calculating compensatory duties on unfairly priced exports has roiled the World Trade Organisation (WTO) and isolated the United States from the rest of the WTO membership.

Under WTO rules, goods exported at a price below that in the home market are said to be dumped if competitors in the export market are shown to have been hurt as a result.

The authorities in the importing country can compensate for this by imposing an anti-dumping duty reflecting the difference between the home and export price.

Calculating this anti-dumping margin often involves comparing several products, some of which may turn out to be priced more cheaply at home than in the export market.

Controversially, the United States ignores these cases when it calculates the margin, a practice known as zeroing. Other countries charge that this unfairly drives up the size of the compensatory duty.

This is how zeroing works:

A. CALCULATION METHODS

There are three methods to calculate anti-dumping margins.

1. The first is "weighted average to weighted average".

You take the average price of some goods in the home market and compare them with the average price in the export market.

Imagine a shipmaker selling ships at home and abroad. In its domestic market, country A, it sells one ship for $50 and one ship for $100, an average price of $75.

In the export market, country B, it sells one ship for $40 and one ship for $80, an average of $60.

The difference between the two average prices is $15. So the authorities in country B take this difference, divide it by the average price of $60 in their country and calculate an anti-dumping margin of 25 percent, which they can impose on imports.

In practice it is a little more complicated. Imagine two consignments of DVD players, one of 20,000 units and one of 50,000 units. The export price would be weighted to reflect the different size of the consignments, hence the term "weighted average to weighted average".

Zeroing cannot be applied with this type of calculation.

2. Transaction to transaction.

Here the authorities do not take an average of the products in question, but look at the home market, and identify the transaction which is most comparable with one in the export market.

There is an element of judgement as the transactions have to be comparable over time, and the products may not be identical.

Taking the example above, they may compare the ship sold for $50 in country A with the one sold for $80 in country B. Here there is no dumping, because the ship was sold for less at home than the similar ship was exported for.

Technically there is a negative dumping margin of $30.

But the other two ships are also comparable, and here one has been sold at home for $100 and the other exported for only $40 -- a dumping margin of $60.

So for the entire series of transactions, the total anti-dumping margin is $60 minus $30 divided by the total cost in the export market of $40 plus $80, or $30 divided by $120 or 25 percent.

But in zeroing, the negative dumping margin on the first ship would be ignored, on the principle that there is no dumping so the dumping margin is zero. Then the calculation would be $60 divided by $40 plus $80, or 50 percent -- much higher than the anti-dumping duty when zeroing is not used.

3. There is a third possibility -- weighted average to transaction.

Here the average price in the home market is compared with the individual price of each transaction in the export market.

This is only used in special cases that have to be justified.

B. SIMPLE VERSUS MODEL

The zeroing used in method (2), transaction to transaction, is known as "simple zeroing".

There is also a way of using zeroing in method (1), weighted average to weighted average, where there are differences in the various shipments being compared. This is known as "model zeroing", because different types or models of a product are compared.

For instance imagine that country B is investigating the dumping of batteries by firms in country A.

It would do a weighted average to weighted average calculation for double-A batteries and another for triple-A batteries.

Assume the double-A batteries are sold for more in the export market than at home, in other words they are not being dumped or technically there is a negative dumping margin. But assume the triple-A batteries are sold for less, so they are dumped.

If zeroing is applied to this kind of model calculation, the negative margin on the double-A batteries would be ignored, or treated as zero, instead of mitigating the margin on the triple-A batteries.

C. ORIGINAL VERSUS REVIEW

Anti-dumping margins are calculated during an "original investigation" when dumping is first suspected.

But once anti-dumping measures are in place they can be reviewed, requiring a further investigation.

For instance WTO rules specify that anti-dumping measures must be withdrawn after 5 years, unless dumping is still taking place, in which case they can be renewed. The investigation into whether dumping is continuing is known as a "sunset review" in the United States and an "expiry review" in Europe.

A number of transactions over the 5-year period would be examined. Where there is a negative dumping margin there is scope to use zeroing.

Or a new exporting company may ask the authorities in the importing country to assess whether they are dumping.

D. AS SUCH VERSUS AS APPLIED

Anti-dumping investigations can target either a particular case of alleged dumping, or "as applied" dumping, or the law or legal framework in a country allowing dumping to take place, or "as such" dumping. (Writing by Jonathan Lynn, Editing by Matthew Jones)