Last Thursday I was kindly allowed by my local Deputy Returning Officer to observe the process for opening postal ballots. I saw the postal ballots for the Edinburgh constituencies being opened, the personal identifiers (signatures and dates of birth) verified and the disallowed votes removed. I can happily confirm that postal vote agents, safely corralled behind a chain barrier, would not be able to see the votes on any of the ballot papers. At the session there were two SNP postal vote agents, who came concerned about postal vote sampling by other parties, but quickly realised there was nothing to see and wandered off.
Having read the procedure for opening ballots, I had naively assumed that the sampling of ballots took place when ballot papers were removed from the inner envelope. Having witnessed the process in operation it is clear that tallying of votes could potentially take place when the reverse sides of all the ballot papers are scanned to find the ballot papers matching personal statements that have failed the verification step. This was impossible in Edinburgh because the scanners had covers to prevent observers seeing the pile of ballot papers coming out of the scanner face-up, but the process does not follow the Electoral Commission guidance so there is nothing in the guidance to recommend that when ballot papers are scanned the scanner must be covered.
Postal vote opening procedure
6.23 The processes to be followed when opening postal ballot packs are
provided for in legislation:
Stage 1: opening of the postal voters’ ballot box
Count and record the number of returned postal ballot packs (i.e., the
number of envelopes ‘B’ in the postal voters’ ballot box).
Open covering envelope ‘B’ and remove the postal voting statement and
ballot paper envelope.
Check the number on the postal voting statement matches the number
on the ballot paper envelope (envelope ‘A’).
Place a mark in the postal voters’ list or postal proxy voters’ list as
appropriate to show that a postal voting statement has been returned.
Stage 2: checking the personal identifiers
Check that the elector has signed the statement and given a date of
Check the signature and date of birth on the postal voting statement
matches those on the personal identifiers record.
If you reject a postal voting statement, you must mark the statement
‘rejected’, attach to it the ballot paper envelope (if there is no such
envelope you must attach it to the ballot paper) and place it in the
receptacle for rejected votes. Before placing it in the receptacle, you
must show it to the agents and, if any of them object to your decision,
add the words “rejection objected to”. You should also record the reason
for the rejection.
Stage 3: opening of postal ballot paper envelopes
Open the ballot paper envelope (envelope ‘A’) and remove the ballot
paper ensuring the ballot paper is kept face down at all times.
Check the number on the ballot paper envelope (envelope ‘A’) matches
the number on the back of the ballot paper.
Place the ballot paper in the postal ballot box.
Stage 4: sealing the postal ballot boxes
Count and record the number of postal ballot papers to be sealed in
each postal ballot box.
Seal and securely store the postal ballot boxes.
In this process the ballot paper is not opened until the personal identifier has been checked. The key stage in this process is Stage 2: checking the signature and date-of-birth returned with the ballot with the signature and date of birth on the application for a postal vote. The law requires 20% to be checked, but to prevent postal ballot fraud Returning Officers are encouraged to check 100%. Thousands of postal voting statements have to be matched with thousands of personal identifier records. It is hard to imagine how this would be done manually, so unsurprisingly the postal voting statements are scanned and matched with the pre-scanned personal identifier records using a computer system, supported by manual checking of all rejects. However, to scan them the personal voting statements have had to be separated from the ballot papers. To find the ballot papers associated with the personal identifiers that have not passed validation, due to missing dates, missing signatures, signatures not matching etc, it is then necessary to scan the reverse side of the actual ballot paper to match up the unique codes. Using standard scanners the pile of ballot papers are fed in reverse side up, but come out front side up. In Edinburgh the scanners were covered, but if they were not covered a postal vote agent would be able to see every vote being spit out of the scanner at a steady rate of one every two seconds. It is very difficult to imagine an agent doing this without it being obvious to anyone nearby that that is what they are doing.
I understand that that software incorporating signature comparison is now widely used to check identifiers supplied with postal votes, but it appears that this means that the processes used do not follow the Electoral Commission template and also creates a risk that the secrecy of voting will be compromised.