Portfolio allocation

  1. 1,158 Posts.
    lightbulb Created with Sketch. 593
    Serious longer-term investors only on this thread please - tag others as you see fit. It was getting hard to discuss this on twitter..
    @Klogg @MarsC @travelightor @Just_a_guy @fire_bull 
    I am keen on seeing how other investors go about their portfolio allocation, to try to optimise my own decision making. It would be great to see if others are willing to share their experiences as well....
    Key thinkers that have influenced my thoughts on allocation are Howard Marks (his memos on how he decides when to sell/exit positions), Lee Freeman-Shor, Ian Cassel as well as experiential.
    Whilst I don't have specific rules in principle I:
    - weight positions according to 1) downside potential, 2) risk/reward and 3) conviction in the business prospects.
    - I have internal competition for funds- new positions must be better than adding to ANY existing position AND
    - lower conviction positions need to justify continuation in the portfolio. New stocks that are down over 3-6 months are culled except in exceptional cases.
    I try to keep to 10-15 stocks and generally don't have a cost base of any position of more than 20-25%. My portfolio is very out of whack now as for a large chunk I had MND, CWP, XRO and TCH/APT as my 4 biggest positions which were very uncorrelated in practice. However due to market performance and my feeling that MND cannot generate the same kind of returns I want moving forward, I have a huge proportion of my equities (which are approx 2/3 of liquid assets, the other 1/3 is cash) in APT and XRO and have taught myself how to stomach huge volatility...

    Below is a list of all stocks I have bought over the last 4 years with absolute and % returns (from late January I think)...there are some pretty bad decisions in here (especially at the start).  I made a profit on 61% of positions, with an average return of 76% and loss of 19% on losses during what was a very favourable period for my investment style.  A lot of the profit was/is from TCH/APT and XRO.

    I made 31 buys in 2015 (55% profitable, reward/risk= 2.84), 24 in 2016 (75% profitable, reward/risk= 3.08), 11 in 2017 (73% profitable, reward/risk=5.54), 17 in 2018 (47% profitable, reward/risk 1.77). A lot of my returns are in year 2- so I won't count out 2018 just yet. Average hold duration for positions overall is 322 days (this is skewed down by recent purchases in 2018 and a period of high trading activity at start of 2015), 334, 213, 356, 158 days respectively for positions bought in 2015,2016,2017 and 2018 respectively...

    In retrospect the biggest mistakes by far were not buying big enough stakes in what were clearly sure things due to lack of confidence (RSG at circa 30c and BSE at 2-10c) & I will add ANO <60c, ASH at 17c, selling too early (TPW at 40c, RHP at 90c). Bad decisions also seemed to cluster around periods of over-performance, which is a constant battle.
    https://hotcopper.com.au/data/attachments/1430/1430424-e3b5a63347846b913947471bc432632e.jpg
    https://hotcopper.com.au/data/attachments/1430/1430426-dc33b3e79d80e640a982be924584989f.jpg

    My current portfolio at the time of this benchmarking activity was:
    APT 36.2%
    XRO 25.7%
    CWP 8.3%
    KME 6.5%
    ASH 6.4%
    LVE 4.9%
    FRX 3.4%
    ANO 3.3%
    SRH 2.6%
    SSG 1.8%
    SES 0.8
    %

    It is clearly inadequately diversified, and I will clearly have to re-adjust at some stage this year, but I really don't want to sell and XRO (definite no) or APT (could sell some of my more recently acquired position if it keeps rising out of sync with rest of portfolio), so I am a bit stuck. The weekest links are definitely the last two, and also possibly FRX, which I am happy to take a chance on given the very skewed upside potential.

    I have been separately estimating expected returns based on probability weighting possible outcomes to get a blended target- and I have found that after 6 months of doing it, my estimates are woeful. Part of this is due to the way I am analysing the data (essentially assuming 3 month return continues for 12 months, which is actually the opposite of what happens with mean reversion- which is what we can see with the 6m return graphs)...Also the first 6 months of FY 19 were unusually brutal.

    https://hotcopper.com.au/data/attachments/1430/1430444-cb320139a43c6ee9f9655c1ce89c537c.jpg
    https://hotcopper.com.au/data/attachments/1430/1430446-0dcac01b403d5d08df772d2f2f8dfa47.jpg

    I would be interested in others' comments or thoughts on how they do this.  Although not much attention is spent upon this aspect of investing, I am increasingly convinced it is the most important determinant of outcome.  There is a lot of literature on asset allocation being the primary determinant of investment return across asset classes, and I see no reason why it should be any different at an individual stock level...
 
arrow-down-2 Created with Sketch. arrow-down-2 Created with Sketch.