Observing external voting may be difficult as a result of the reduced transparency of the process and the costs involved. Collection of first-hand representative evidence from all steps of the process may be impossible. In principle the observation should take the following steps:

• an assessment of the political environment in which external voting is implemented. Is there a general agreement on the terms for external voting and is there general confidence in the process?;
• assessment of the legal framework and the checks and balances worked into the system;
• assessment of the pre-election campaign. The election laws of the country will not apply, but it would still be of interest to assess what is being done to help voters make an informed choice from abroad;
• if the voting is done in person, observing the voting, reconciliation, packing and sealing of the voting material on the basis of a reasonable sample. The assessment of how big the sample needs to be would depend on whether one wants to draw conclusions about the external voting as a separate process in its own right, which would require good coverage, or whether the external voting is only a small part of the overall assessment of the elections;
• if the voting is done by post, observing the verification of the lists of external electors, the distribution of voting material, the reception and verification of the same, and the count; and
• if the voting is electronic, assessment of the process of implementing the system, including the rules for verifying the identity of the voters, audit trails, and the validation performed by the EMB.

All external voting constitutes a challenge from an observation point of view. Domestic observation is often based on wide coverage rather than sampling, and local NGOs may therefore want to cover all polling places. However, domestic observer organizations may not have the resources needed for proper observation of voting in embassies and consular offices (although a worldwide network of Indonesians, mainly students, succeeded in achieving wide-ranging coverage of voting at Indonesian embassies in the 1999 transitional elections). International missions may have an easier task in terms of resources to observe external voting if it takes place in a controlled environment, not least because the sample may be more limited.

Postal and Internet voting would be more complicated to observe. The first problem is the identity of the voter. In Bosnia and Herzegovina a number of attempts to impersonate postal voters allegedly living abroad were exposed in 1998 and 2000. In the case of Internet voting some of the security problems may be solved in the future, but the basic paper trail—the ballot paper—will be missing. In paper-based voting it is always possible to order a recount, and the actual voting will have been followed by observers and party agents.

The most important condition for reliable voting in an uncontrolled environment is general trust in the EMB. Even external voting in a controlled environment requires such trust, since full observation is difficult for parties and NGOs.

The political environment

External voting in transitional democracies is often controversial, either for political or for technical reasons, as discussed above in relation to Croatia. In Zimbabwe, the opposition sought a more liberal ruling for external voting, possibly believing that President Robert Mugabe would have less support among those who have left the country. On the other hand, the same opposition criticized the implementation of the rather limited external voting for its lack of transparency and control.

The key issues for observers to assess are the political environment under which external voting is being conducted and the independence and integrity of the EMB. In particular the arrangements for external voting should be subject to a high degree of consensus, since the transparency of external voting will almost inevitably be less than that of in-country voting. Any dictate from a political majority is likely to attract charges of fraud and manipulation unless the administrators’ integrity is beyond doubt.

Personal voting in a controlled environment

External voting may be difficult to observe at all stages. If the voting is done in person in a controlled environment (such as embassies or consular offices), observation may be possible but will be expensive. If the resources are available, the observation of this kind of voting will be rather similar to the observation of in-country voting, where a sample is chosen for collection of first-hand evidence and the process is observed, from the opening to the close of voting to the reconciliation and sealing of the ballot material. However, there may be significant differences from the observation of in-country voting. Observers may stay in the same polling station throughout the voting, rather than moving from one polling station to another during election day. If the count does not take place in the polling station but in counting centres in the country itself, questions of the integrity of votes in transit arise. In such cases it becomes a task for observers to follow the processes of packing and sealing, reconciliation, and reception at the counting centre, as well as the count itself.

Postal voting

If voting is done by post, the actual voting may not be observed, but other processes such as the maintenance of the registers of external electors, the form of the letters sent out to the voters, the documentation required to accompany the ballot paper on its return, the checking of whether ballot papers are genuine and the count may be observed. However, the process may be spread over a period of time and may therefore be difficult to observe throughout.

Electronic voting

If the external voting is done electronically, for example over the Internet, it will be an even bigger challenge to observe all steps of the process. Confidence in the process will depend on the level of trust in the IT system, which in turn to a great extent is dependent on confidence in the EMB.

IT systems should meet a number of criteria, such as being reliable, user-friendly, secure (meaning that they can resist deliberate attacks from outside and inside at all times) and verifiable (meaning that they can be checked). IT systems may be manipulated from inside and from outside. Insiders may build in functions which may change the result in a certain direction. Such manipulation may in theory be conducted by the election administration itself. In transitional democracies, if the general trust in the election administration is low, an electronic voting system may add to the lack of confidence in the electoral process.

Full validation will be next to impossible for any observation mission to perform. One should therefore be very careful not to give the impression that an independent validation of the system has been carried out even though IT experts may have assessed parts of the process. Bringing in such expertise may give a false impression that the systems have been validated against international standards.

In particular, manipulation from inside is in practice almost impossible for an observation mission to assess. Even if the observation mission has IT expertise available, the experts should concentrate on the process which the EMB has carried out when acquiring the systems, rather that pretending to carry out an independent validation of the system. Such assessment of the process should include:

• the choice of supplier, including the requirements for certification against recognized standards for quality and security;
• the validation of the requirements specification and overall design;
• the use of quality auditing during implementation; and
• the strategies for testing the system.

In addition, the audit trails and all possibilities for recovering data in the event of a failure should be assessed. However, the main audit trail in traditional elections, namely the ballot paper, will be missing. It will therefore not be possible to reconstruct the results on the basis of a ‘paper trail’. This is why observers need to put much emphasis on the assessment of whether there is a general consensus on the introduction of Internet-based voting and on the integrity of the EMB. A prototype of the IT system may have been validated by the EMB, and it is only the EMB which can ensure that the system being used for the elections is actually the validated system, without any manipulation from election administrators.