The following is the result of the GA action item A08-06 which says: "Executive to develop a formula to determine the boundaries between the TERENA membership categories".

When TERENA was established, National Members were put into categories of voting rights and membership fees according to the Gross National Product (GNP) of their countries. As to possible future changes of category, the only thing decided was the following: "Members that believe to be in the wrong category can apply for a change of category, to be decided by the GA. This will mostly be the case when the GNP of a given country overtakes (in positive or negative sense) the GNP of a country in a neighbouring category."

The GA accepted in its meeting in Madrid in October 1997 that TERENA will be using the GNP data as published by the United Nations, even though these are generally (like other statistics) a few years old. It was also decided that the TERENA Secretariat should check the GNPs of the countries of the TERENA National Members once a year.

The implication of this is that changes in membership categories should not just happen when members propose them (as in the quote above) but that the TEC should each year propose to the GA changes of members' categories in the case the GNP figures indicate that this could be justified.

A remaining problem is the following. Suppose this year country A is in category (n+1) and country B is in category (n), but according to more recent UN figures country B's GNP has become higher than that of country A. Will then country B be promoted to category (n+1) because it has overtaken country A in the positive sense, or will country A be demoted to category (n), because it has overtaken country B in the negative sense?

Any formula or algorithm to determine the boundaries between membership categories must fulfil the following requirements:

Taking into account these requirements it is not possible to develop a mathematical formula to calculate the boundary between categories. Any formula that has as its output a boundary expressed in billions of USD will need to have as its input the combined GNPs of countries expressed in billions of USD. Since GNPs change a lot and any average of GNPs will also be changing in unpredictable ways, such a formula would never be stable. For instance, such a formula could result in a country having to go to a higher or lower category because a completely different country of a completely different size has seen a large fall or growth of its GNP.

Or, as another example, such a formula could result in a country going to another category because a completely different country has left TERENA or has joined TERENA. Clearly such artefacts are undesirable.

It is therefore proposed to abandon the idea of a formula and to agree an algorithm to formulate proposals to the GA for changes of membership category of certain members. Staying as closely as possible to the rules used thus far (and accepted by the GA), the Executive proposes the following.

Each year in the preparations for the October GA meeting the latest available UN statistics will be used to find the GNPs of the countries of TERENA's National Members, expressed in billions of US dollars. The amounts will be rounded off to the nearest integer (amount of billion US dollars) and those figures will be called the "current GNP" of each country. The same procedure having been applied one year earlier has produced the "previous GNP" of each country.

A change of membership category will be proposed to the GA only if the following holds:

There are two countries A and B, last year country A was put in category (n+1) and country B was put in category (n), and the current GNP for country B is higher than or equal to the current GNP for country A.

Then there are three possibilities:

- The current GNP for country A is higher than or equal to the previous GNP for country A (and consequently the current GNP for country B is higher than the previous GNP for country B).
*In this case it will be proposed to promote country B to category (n+1)* - The current GNP for country B is lower than or equal to the previous GNP
for country B (and consequently the current GNP for country A is lower
than the previous GNP for country A).
*In this case it will be proposed to demote country A to category (n).* - The current GNP for country A is lower than the previous GNP for country A while the current GNP for country B is higher than the previous GNP for country B. In this case a boundary point needs to be set to determine promotion/demotion. P is defined as
the average of the highest previous GNP in category (n) and the lowest
previous GNP in category (n+1).
- If the current GNP for country A is lower than P, it will be proposed to demote country A to category (n).
*If the current GNP for country B is higher than or equal to P, it will be proposed to promote country B to category (n+1).*

All promotions and demotions will have effect from January 1st following the October GA meeting in which these decisions have been taken.

As a phase-in, during the first year after having been promoted a National Member will pay a membership fee which is the average of the membership fee of its current category and the membership fee of its former category.

In addition, the overall picture of all GNPs and categories of countries should be revisited once every four or five years, to check that the repeated application of the algorithm has not introduced any unfairness.

Revisiting the figures as presented to the General Assembly in the Madrid meeting, the decisions the GA took there would have been exactly the same if the algorithm described above would have been applied.

An update on the GNPs has been published by the United Nations during February 1998. See attached spreadsheet GA(98)002, in which the current GNPs are also compared to the previous GNPs.

The net result of the use of the proposed algorithm is that there will be no changes in membership category for 1999 - with one exception, as described under GA agenda item 5, Ukraine. This illustrates the stableness of the proposed algorithm.