An 8-Million-Vote Margin on 2.75 Billion Votes Cast
"How Could the Inspectors Possibly Confirm That The Numbers Were Right?
Talk about a photo-finish! At one of the shareholder meetings your editor and his business partner inspected this season, one of the six shareholder proposals received 49.85% of the votes cast while the company-favored votes against garnered 50.15% - a difference of just 8-millionodd votes out of 2.75 billion that were cast on the matter, or a mere three-tenths of one-percent of the votes cast.
Put the way a smart Inspector should put it, if just 4-million-odd votes were erroneously recorded as Against instead of For, the vote would go the other way. No wonder, we thought, that our article on “What, Exactly, Should Inspectors Be Doing to Inspect…and How Do We Know the Reported Numbers Are Right?” has been one of the most revisited and most carefully perused on our website, logging nine+ minutes per visit.
While we don’t want to give away all of our ‘trade secrets’ as Inspectors, here’s an overview of what we did to verify that the reported numbers were indeed correct. Amazingly, it did not take all that much time and clerical work to accomplish…IF one knows what one is doing, and if the systems themselves are readily “auditable.”
For starters, we were able to take the Broadridge numbers completely at face value, because (a) we had personally reviewed and observed their day-to-day quality-control procedures just a few months earlier...and (b) we had also reviewed the official reports of Broadridge’s outside auditors, stating that one can rely on their reports as having a 99.9% degree of accuracy - based partly on the Q-C procedures - and on their testing of them - and (c) as we knew, on the fact that the overwhelming majority of all the votes recorded were actually input by institutional voters themselves, or by their designated voting agents.
So we began with a review of all of the larger votes recorded by the tabulating agent, whose Q-C procedures we also knew very well - and which incorporated many reports and reporting features we’d helped them implement years earlier, when there were several other “squeakers” of similar closeness. This was much easier to accomplish than one might think - since there were relatively few votes of four million shares or more to look at that had not been input by voters themselves.
Then, we reviewed each of the methods by which the votes were recorded - and the number of votes in each category: Happy day for us all, the overwhelming majority of the votes were recorded automatically, with most of them via various kinds of actions taken by the voters themselves. For example, with respect to the registered-owner votes, 18 million votes were cast over the web - by voters themselves, typing away. Another 6.4 million votes were voted on the phone - with the voters themselves punching the buttons and getting a read-back, so no possible problems there either. The rest of the registered votes (10.275 million votes) were scanned and recorded directly into the system, and no “glitches” or “anomalies” were apparent: The recorded votes were consistent with what was expected to be a “fairly close vote” and with “all systems working consistently and accurately” as built.
As noted earlier, institutional investors enter their votes directly into the system themselves, or by using professional voting agents to do it, so no legitimate gripes can come from that source. And these votes constituted the overwhelming majority of all votes cast. The last two steps were the most important ones, given the overall closeness: (1) a review of each and every vote that was manually entered (where a “tally clerk” could easily transpose or over-state or understate a number - or worse, vote 4 million-odd shares the wrong way) and (2) a review of every initial “over-vote” - and how each such item was ultimately resolved and recorded.
The tabulating agent - Computershare - was ready and able to produce the full records for us on these items…And, happy day…the manually entered votes totaled-up to 2.6 million - a number that could not have changed the vote. So, although we could have done so had we needed to, we didn’t need to review the proxy cards that were manually entered, one-by-one. And happy day again, not one of the initially identified two million-share “over-voted positions” were over-voted on the proposal under special scrutiny.
We, and the company, were able to sign off on and release the final numbers with complete confidence in their accuracy, which made for a happy day indeed.
Readers, please remember that not every close vote can be resolved so easily - unless, of course, you have tabulating agents - and Inspectors - with procedures and controls in place like those described above. So choose your providers with special care - especially if you have items that you fear WILL be close come the end…
P.S, There was yet another important takeaway here: While there were 2.75 billion votes cast on the proposal in question, the quorum was 3.5 billion shares - thanks to Abstentions and to over three-quarters of a billion (!) Broker Non Votes. The BNVs amounted to 21% of the quorum! If only five percentage points of the BNVs had been VOTED - by individual investors who did not take the trouble to vote - the “squeaker” would have gone the company’s way with a much more comfortable margin…since, as noted earlier, the retail vote goes overwhelming to the management positions.