The current system of ISO codes provides three codes for each country, a two-letter country code, a three-letter country code, and a three-digit numeric code. The alphabetic codes were introduced in ISO-3166 in 1974, and the three-digit numeric codes were introduced in 1981. Some of the ISO codes do not refer to countries, per se, but to geographic territories. For example, there is a code for the Heard and McDonald Islands, which are both uninhabited parts of Australia.
The two-letter code for the United States is US, the three-letter code is USA, and the numeric code is 840. Letter codes are used for political changes and numerical codes are used for territorial changes. Germany, Yugoslavia and Ethiopia have all received changes in their numeric codes because they have gone through geographic changes, adding territory in the case of Germany, and losing territory in the case of Yugoslavia and Ethiopia, but the alphabetic codes have not changed.
On the other hand, Burma (renamed Myanmar), and the French Territory of Afars and Issas (renamed Djibouti) underwent alphabetic code changes, but not numeric changes because there were political, but not geographic changes. When the Soviet Union was transformed into Russia, both its alphabetic and numeric codes were changed.
Because our system focuses on both current and historic codes, there is an important difference between our system and that of the ISO. Since the ISO's primary interest is in providing current codes for countries, they allow country codes to be reused, but only after a five-year period has passed since the old code fell out of use. The two-letter code GE used to apply to the Gilbert and Ellice Islands, but now is used for Georgia. In my opinion, a code should never be reused. The reason for this is obvious. If you want a single set of codes for all countries over time, you cannot replicate codes because of the confusion this creates.
When a country has gone through a political change, the exising codes are changed to reflect this. "Dead" countries are given a four-letter country code by the ISO to indicate political or name changes that have occurred since the codes were introduced. The new four-letter code is formed by combining the old and the new two-letter codes for each country. Burma now carries the code BUMM to show that Burma (BU) became Myanmar (MM). Countries that had no clear successor state have an HH added to their code, so the new code for the Soviet Union (SU) is SUHH.
For reasons stated below, we do not believe that this is the best solution for dealing with "dead" countries. Our system never reassigns a two-letter or three-letter code to avoid historical conflicts. Instead we use three-letter codes to differentiate political changes over time. Four-letter codes are reserved for subnational governments, such as states and provinces.
We have used the ISO codes wherever possible to avoid confusion; however, because the focus of our history of currencies is on the past as well as the present, we have found it necessary to make a number of changes in the methodology the ISO uses for producing country and currency codes.
First, there is no difference between the two-letter codes the ISO uses and the codes we use. The only difference is that we have added several two-letter codes for some historical territories and countries not covered by the ISO.
It is necessary to provide two-letter codes for each country or territory because the two-letter codes are used to determine the three-letter currency codes. For example, the symbol for the United States Dollar is USD. This combines the two-letter code for the United States (US) with the letter "D" for Dollar. Without a two-letter country code, you cannot create a three-letter currency code.
Nevertheless, not all countries have two-letter codes. There are several reasons for this. First, a territory within a country that has monetary autonomy may not have applied for a code. The Isle of Man has no ISO code, but issues its own currency and has its own Internet code (IM). Second, the country may not be internationally recognized, and consequently may not have applied for a code. Transdiniestra and Somaliland have not received an ISO code because they have not applied for one, but the Palestinian Territories have received a code because they did apply. Third, a country may have existed before the codes were introduced. For example, we provide the code VR for the Republic of (South) Vietnam. We differentiate between the two-letter codes we have created and those assigned by the ISO by placing an asterisk next to our codes.
The problem with the two-letter codes is that even though there are 626 possible two-letter combinations, most of the more common letters combinations have already been used. Thus, any logical two-letter code for a country such as Memel has already been used.
Once you begin to use three-letter codes, you have 17,576 possible combinations, which provides a sufficient number of codes to cover all countries that have existed in the past. Three-letter codes are preferable to three-digit codes for two reasons. First, letters provide a greater number of possible combinations - 17,576 letter codes as opposed to 1000 numerical codes. Second, logical letter codes can be created for each historical geopolitical entitity. No similar logic could be provided for three-digit numeric codes.
For this reason, we have added a large number of three-letter codes, not four-letter codes, to describe historical changes in countries. We do this in several ways.
First, we use the two-letter code as a basis for different three-letter codes that are used to indicate political changes that have happened over time. To create new codes, we take the two-letter prefix for each country, and add a code that would correspond to the change in sovereignty. Hence, we can differentiate between the Tsarist Russian Empire (RUE), the Russian Socialist Federated Soviet Republic (RUF) that existed from 1918 to 1922, and the current Russian Federation (RUS).
The ISO does not attempt to replicate the two-letter codes in the first two letters of the three-letter code. For example, the two-letter code for the Cayman Islands is KY, but the three-letter code is CYM. The two-letter code for Antarctica is AQ, but the three-letter code is ATA.
Three-letter codes also allow us to indicate how the sovereign status of the country has changed over time. This allows us to differentiate between Indonesia while it was the Netherlands Indies (IDD), while it was under Japanese occupation (IDJ), and when it became independent Indonesia (IDN).
Second, three-letter codes are also provided for "dead" countries. In most cases, we do not provide a two-letter code for "dead" countries. By providing only a three-letter code, we can indicate that the country no longer exists. Hence, Memel, which existed as an independent Allied protectorate before it was incorporated into Lithuania in 1923, has its own three-letter code (MML), but has no two-letter code.
Currently, the ISO only provides four-letter codes for "dead" countries, such as the Soviet Union (SUHH), using the system described above. However, we think a better system would be to assign three-letter codes to "dead" countries, and reserve four-letter codes for smaller political entities within each country.
The ISO has already established local codes for subnational political entitites-states, provinces, cities, counties, etc. These could be used in combination with the two-letter codes to create regional four-letter codes and local five-letter codes. In the case of regional codes, the first two letters would indicate the country, and the second two letters would indicate the province or state. Local codes would take the two letters from the country codes and the three letters from the local code. Once regional codes have been created, the national and regional codes can easily be combined.
For example, Texas could be given the code USTX, US for the United States and TX for Texas. Dallas, Texas could be given the code USDAL, US for the United States and DAL for Dallas, Texas. On the other hand, Dallas, Pennsylvania has been given the code DSP, so USDSP could indicate Dallas in Pennsylvania (USPA) rather than Dallas in Texas.
Numeric codes were originally assigned by putting all of the countries in alphabetic order, then assigning every fourth number to a country. Consequently, Afghanistan has the code 004, Albania 008, etc. Some numeric codes now appear to be out of order because the country has changed its name. Benin has the code 204 because it was originally known as Dahomey. Numbers above 900 were not used for countries.
Over time, the ISO has begun to use up the remaining numbers either because of geographic changes within countries, or because of the birth of new countries. Ethiopia's code (230) was changed (to 231) when Eritrea seceded from Ethiopia. The Czech Republic received a new code when Czechoslovakia (200) was split into the Czech Republic (203) and the Slovak Republic (703).
We do not use any numeric codes for historic countries, but choose to use three-letter codes, primarily because there are a greater number of possible three-letter codes than three-number codes, and because there is no real logic in the system for assigning numeric codes for "new" countries.