The Voting system
A three-point score system 5-3-1 with a minimum 60% pass threshold. At decision time, members are invited to rate the proposal..
When the discussion and Q&As for a proposal have ended, the chairman invites each member in turn to respond with a score. The chairman then totals up the points and announces the decision.
For example, if we have 4 members present, the maximum a share can get is 20 points (4 x 5points); Any share getting more than 13 points (>60%) would get the decision in its favour. Suppose the scoring for these 4 members is 5,5,3,1; hence a total of 14 points (greater than 13 points), then the vote is carried. A voting sequence from the 4 members of 5,5,1,1 (12 points) would not.
Some possible interpretations of the 5-3-1 point scheme:
- Five points may mean this is a stock you would strongly support or you would have prepared a case to propose to the club.
- Three points suggest that you would be happy to support the share.
- One point would imply that you reject this stock.
For sell decisions, a total score greater than 60% is sufficient for the sale to take place. For buy decisions, a tiered system is in place, using the ratio of total participants score over the total possible points. This ratio (in %) defines the maximum amount to be invested in the buy decisions.
The Key components of the voting system
- A three-point score system 5-3-1
- A minimum threshold 60%
- A tiered system for buy decisions (that sets the limit on the amount invested)
- One member one vote
The amount invested per buy decision is guided by a four-tier level (see table 1 below). The total number of points scored has a direct bearing on how much gets invested. This in a way shows the level of collective or joint confidence in the decision. The ratio of total participants score over the total possible points (in %) is used: There are four levels in the tiered system; less than or equal to 60%, between 60-70%, 70-80% and greater than 80%.
The club has a rule that sets the maximum amount invested per holding in the portfolio, (let’s call it MAX): Simply put, MAX is the club’s Net Asset Value (NAV) divided by 12 (which is the maximum number of slots in a portfolio). The four-tier level stipulates the upper limit for each investment relative to MAX and the ratio. The higher the ratio, the higher the upper limit (as % of MAX)
|Ratio (%)||Decision (% of MAX)||Decision (£) - maximum|
|≤ 60%||do not proceed||do not proceed|
|> 60%||Up to 50%||£5,000|
|> 70%||Up to 75%||£7,500|
|> 80%||Up to 100%||£10,000|
|Table 1: Upper limit on amount invested|
The third column in the table above is an example with MAX = £10k. For instance, if the portfolio NAV stands around £120K and the maximum portfolio size is 12 entries, then MAX per holding (including any further top-ups) is £10K (£120k/12). This limit is revised annually at the AGM.
One of Mobius tenets has been to buy shares it has strong confidence in, and as the club gets to know the share/company over time and also builds more confidence, it accumulates (tops up), up to the MAX value. The Mobius portfolio has accumulated many shares in this fashion. The long term average time length of a stock in the Mobius portfolio is 16 months. The longest share held in the portfolio is Persimmon (10 years & 3 months with a 25% AER investment return).
How it works in practice
To keep it simple, a second (lookup) table was prepared (table 2 below). It contains the minimum number of points for each limit at 60%, 70% and 80%. For example with 5 participants, 16 points are required for a pass and 19, 21 to pass the 70% and 80%. This helps the chairman save time by not having to compute percentages every time. So the chairman has two tables at hand; (1) the table that defines the maximum amount invested for that total score (table 1) and (2) the lookup table for ease of reference (table 2).
He starts by a simple row-lookup matching the number of participants and cross referencing the column best fit (which interval) for the ratio of total participants' score. Once he has the interval, then the maximum amount for that purchase is found.
- Suppose we have 7 participants voting with (5,5,5,5,3,1,1),
- the total points counted is 25,
- a lookup on the row with 7 participants indicates that this share got more than 60% and less than 70%; 22 < 25 points < 26,
- Table 1 provides a guideline of up to £5,000, lookup at (> 60%).
If the number of points for the seven members were say 27 (5,5,5,5,3,3,1), then the recommended amount would be up to £7,500
|Participating members||max points||> 60%||> 70%||> 80%|
|Table 2: Lookup number participants vs total score|
The previous system - Why change it?
Up to 2012, the voting scheme was the widely used binary (yes, no) system with a simple majority. Over time, it was felt that it was not allowing members to better express their rating & confidence in a proposal. In addition, too many shares were bought or sold either hastily or with too fine a majority, somehow losing the principle “if in doubt, best leave it, better opportunities elsewhere”. This led to buying and selling the same stocks shortly afterwards, or indeed selling quality holdings. To make things worse, there were other issues related to late voting, typically after 10:30pm which meant members hurriedly voted to end the meeting (other rules have been put in place to mitigate this).
So to address these issues the new voting system
- proposed a three point score system to allow more granularity in the decision making scale.
- raised the bar to 60% so that there is a clear cut decision.
Voting per member
The principle of one member one vote has been long standing in the club. However, this has been challenged in the club in recent AGMs. Alternative methods have been discussed, based on a system weighted to some degree on the members units owned (proportional voting):
The reason is simple: A combination of proportional ownership and risk tolerance. With a club of 23+ years standing, you are bound to have members with large holdings compared to new members or those with smaller holdings. So when new members join in (with small funds), they will want to experiment and learn. This can sometimes bring frictions with long-standing members when it comes to the level of risk taken. Normally a good debate and the voting system settles this (and this has been an accepted practice). There have been attempts to bring changes to the “one member one vote” to a system weighted to some degree on the members units owned, but so far it remains “one member one vote”.
The multi-point score system
There are many variants of a multi-point scoring system (with 3,4,5- point scales). We opted for three to keep it simple. Some literature research also seems to suggest a 3 or 4 point scale would be sufficient. As for the actual scores of 5-3-1, they seem somehow subjective. The aim here is to keep the bar (threshold) high enough to absolutely make sure there is enough strength in the participants mutual decision: Hence the strictly greater than 60% threshold. So if all participants went for 3s, then the decision would fail. Equally, 5s and 1s in equal numbers would fail. The 5-3-1 is a heuristic model that worked for Mobius (so far). There could be better models out there if we refined the aim and specified more quantitatively the point system.
Room for improvement
One of the areas for further consideration is the way the score is delivered: Open (public) versus blind (secret). As we are dealing with people collectively making decisions, there is room for tactical or group voting. It is a natural phenomenon; either because we are friends, or new members helping each other or tactical, “he voted for mine so I feel I should vote for his proposal….”. There is nothing particularly wrong, it is just human nature. One way to hope for a more objective result is secret voting: This will allow members to freely & objectively vote without the conscious awareness of being judged by others on the way they have voted.