This is a key instrument in our share selection via a set of filters (jointly agreed by club members). The top (best) selection is automatically bought
The systematic approach is a kind of automatic share selection & purchase. The members collectively agree a set of criteria that shares should pass. On a monthly basis, theses filters are run through a stock screener. The best (top) share that passes these filters is automatically bought. For example, some of these criteria could be as simple as:
- 0< PER (Price Earning Ratio) < 15
- 0< PSR (Price to Sales Ratio) < 2.5
- 0< PEG (Price Earning Growth) < 1.5
- Net Gearing < 50%
Once the club has agreed on the principle of automatic selection through a set of filters (entry criteria), it needs to establish the investment objectives: Three were identified: Growth, Value and Recovery. This was followed by debates on what is meant by these objectives and how to translate these into sets of specific & clear (entry) criteria.
The key stages to put in place a full systematic approach are:
- Discuss & agree the systematic concept
- Investment objectives: Growth, Value or Recovery.
- Select criteria, schedule, tools, portfolio parameters & policies (and exceptions)
- Monthly run
- Top share meeting criteria is selected (no consultation)
- Report to members
- Share bought (if appropriate)
- Add to systematic portfolio
- Monitor life cycle and exit criteria
- Review criteria, schedule & policies periodically
The diagram below illustrates the workflow and feedback loop of the systematic approach:
By far, growth was the most popular investment objective and had the most mutual agreement in the club. The growth filters (criteria) below are made up of two parts; one mandatory and one optional. Phase 2 is used if the output of the shortlist from Phase 1 has more than one entry:
Phase 1: Compulsory:
Phase 2: Optional criteria if more than one entry passes Phase 1
A shortlist from Phase 1 is obtained using a stock screener. If the shortlist has more than one entry, we move to phase 2 to narrow down, applying step 8 and then 9. If the shortlist still has more than one entry, we finalise in step 10 to obtain one entry. If the shortlist from Phase 1 is empty, then we skip the systematic results for that month. This has never happened in previous runs (2006-2008).
Table below is an example, output from the July 2007 execution. It is a truncated version (7 entries) of the full listing of the stock screener. In July 2007, 22 stocks passed Phase 1 and 2 stocks passed steps 8 and 9 of Phase 2: Delta and Victec Group. Given we already had Delta in the systematic portfolio, Victec Group was the selected share for July 2007. There was no need to go to step 10.
|London Scottish Bank||27.33%||9.71%||£139.18||407.87%||13.5||1||1.21|
|Ashley (Laura) Holding||36.36%||26.67%||£188.20||54.13%||24.3||0.3||0.86|
The main stock screener used was from Digital Look (now Sharecast). It was a comprehensive screener and widely used by UK private investors and investment clubs. It provided the conditions for the majority of filters (it was the default screener on the Proshare PIC portals too). A second screener from ADVFN was used to validate the results from the first screener. In reality, once a good screener is selected, some of the filters selected by the club would be adjusted or others selected if these are not available in the screener.
The output from the screener would then be fed into a Microsoft Excel spreadsheet for further manipulation (for example to sort the table, calculate the market’s deciles or identify outliers). We also need to use a spreadsheet because the stock screener did not have the specific facility to express our criteria to the granularity required.
In the example of July 2007 above, the stock screener output featured 188 stocks (with all qualifying attributes), of which 22 passed Phase 1. The MS spreadsheet was also used to compute the various deciles from the 188 entries for steps 1-10
Once a share has been bought and entered the systematic portfolio, additional steps are required to manage its lifecycle and exit from the portfolio. These exit criteria are:
A trailing Stop/Loss (S/L) of 15% of the peak share price (since purchase)
An investment will be sold after 12 months of purchase (unless it's share price passed the target price and has not dropped by the trailing S/L yet).
Target price is 150% of the purchase price (or 50% profit from purchase price)
In May 2006, we set up a pilot systematic portfolio. This would run along the club’s standard portfolio. A fund (£9k) was set aside, a portfolio with up to 6 entries for a period of one year. It proved very popular and the pilot ran 12 times from May 2006 to July 2007 and with a variant called “conditional-sys” to May 2008. All the occurrences were for the Growth objective. The systematic method provided interesting stocks and a good return on investment. The only exception was BETonSports (a total loss, company failed).
Examples of share selected & bought via the systematic method (in mm/yyyy):
Example of shares sold, triggered by the exit criteria (extract from the July 2007 report): Format below includes information on P/L so far, scheduled annual date and exit policy.
|Name||Profit/Loss||Scheduled end date||Exit criteria|
|Lavendon Group||42%||12 Months = 10/03/2008||Exit: S/L 15% off peak.|
|LogicaCMG||-18%||12 Months = 10/01/2008||Exit: S/L 15% off peak.|
|Robert Walters||53%||12 Months = 10/07/2007||Exit: 50% (not in policy)|
|Delta||8%||12 Months = 10/05/2007||Exit: S/L 15% off peak.|
|LandofLeather||15%||12 Months = 10/08/2007||Exit: S/L 15% off peak.|
|Hyder Consulting||31%||12 Months = 10/09/2007||Exit: S/L 15% off peak.|
|BetonSPort||-100%||12 Months = 10/06/2007||Exit: LSE liquidation.|
One report is issued for every systematic run (typically monthly basis), it summarises the output for that month’s stock selection, a snapshot of the current systematic portfolio, previous transactions and instructions for the dealing team. It is worth looking at a couple of examples to have a feel. Here is the report for July 2007, and another monthly report for the conditional sys (May 2008)
After the first year’s pilot, members also wanted to have a final say (go/no-go) on the stock selected. So a variant was created called “conditional sys”:
After a share has passed all the filters and selected, it is not automatically bought. Club members are given two working days notice before the selected share is bought. Members can cancel out (using the voting system in place) the decision but it needs to be completed within the two working days notice period (otherwise the share is bought).This allows for exceptional occasions where members feel the share selected is really at odds with the objective of the filters or the club (e.g. certain sectors or companies, examples included ethical ones such as gambling).
The filters have been adjusted from time to time: Some of the filters could also be relaxed if for instance no shares have passed the filters for a few months. For example, a basic rule for PER (Price Earning ratio) could be: 5<PE<12 but then relaxed by incrementally increasing the PER’s upper bound until one stock passes the filters.
The systematic approach is important to keep the club’s momentum going when there are few investment proposals; and that is certainly a good way to fill-up the portfolio by (default and) common agreement. When the club received very few submissions, it was equally reluctant to drop the quality of submissions (e.g. members just bringing light proposals on the night of the meetings aka the Shouts & tips approach). The template for proposals (in place at the time) was comprehensive and took a long time to prepare.
More importantly, the systematic shortlist provided impetus for members to research the other candidates in that list. Within a year, the appetite for investment proposals started to emerge: Hence, the request for a variant with collective participation, the “conditional sys”. Within a year of the systematic pilot, yet another solution called semi-systematic was devised: This scheme is a half-way house between a systematic and a light touch proposal approach.
The systematic is an important strategy in the club’s investment toolset, in particular when the conditions are such that it would at least allow for continuity in building a portfolio and a good way forward (until such a point where proposals start to materialise). The systematic method was run over 20 times. Since mid-2008, it has not been necessary to recall the systematic way. The semi-systematic variant also proved very popular as collective participation was buoyant.
The systematic performance is mainly driven by (1) the quality and performance of its key attributes, and (2) a diligent execution. The following attributes impact most the performance (by priority order)
- Entry criteria (set of filters)
- Exit criteria (set of conditions)
- Frequency of run
- Portfolio size
The systematic method does not prescribe a specific investment objective, nor any of the four attributes above. It is a generic “shell” or “framework” and multiple combinations of the four attributes above are possible, as are variants. What we shared here is our experience in selecting and operating a few instances of a systematic portfolio
Ease of use
We found the systematic approach simple to operate and maintain once the setup was in place. The part that needed more effort was the agreement in principle, investment objectives and definitions of the criteria. Identifying an owner for the systematic portfolio is also important to oversee the end-to-end operation (life-cycle) and success of the systematic portfolio.
The systematic approach was heavily inspired by two main sources:
1. Primarily, the mechanical portfolio from the Rolling Stocks investment club (RSIC): In 2002 and 2004, Mobius invited Dave Gaskell from the RSIC a couple of times (for discussions, exchanges and presentations). These meetings proved valuable and extremely useful. We identified the mechanical portfolio as a good experiment to pursue and an interesting concept. In addition, it was a timely opportunity to add to our portfolio given the lack of quality investment proposals submitted to the club.
2. The second source was the book by JP O'Shaughnessy. It had a strong influence on the formulation of the filters, in particular the PSR: "What Works in Wall Street, A Guide to the Best-Performing Investment Strategies of All Time", James P O'Shaughnessy, McGraw-Hill, 3rd edition.
Proshare PIC seminars had also contributed to our screener criteria: These seminars were heavily geared towards long term fundamental strategies. This included the Zulu principle (that contributed to the Hemscott platform, now part of Morningstar). The Proshare PIC toolkit carried strong elements of fundamental principles to identify quality companies. Mobius used many of these principles for the selection of criteria. For example these were the ten steps (Zulu principle):
- Earning history
- Earning prospects
- P/E ratio (a general health indicator)
- Cash flow (best indicator)
- balance sheet (debt/gearing)
- Net asset value/book value
- Directors’ dealings (buying good)
- Stop loss
- Be patient (avoid tips, rumours) follow quality press, brokers
The systematic approach is also so commonly called mechanical (and also known as programmatic or automated).