The neglected art of selling

By low, sell high!  That oft-used 80s catch cry is what all investors should aim to do.  In most cases the first part – buy low – can often be the easiest part to do.  I could have picked 10 random stocks from the ASX200 in March 2009 and I’d be on a winner because buying anything after a crash is a pretty solid strategy.   There are also countless volumes of material written on when to buy stocks, from charting to value investing and every weird combination in between.

The irony is that the most difficult part of investing – the selling – is least-discussed.  Once you’ve snared yourself a bargain and watch gleefully as the share price soars, your thoughts turn to selling.  “Should I sell now? I’ve got a made profit but it could go higher.  Or it could go lower.”  Maybe that junk stock stayed junk and you’re wondering whether to cut your losses.  “Should I hold out for some recovery?  What if it goes lower?”

At Empire we apply a system to selling shares that is just as rigorous as the system we use to buy them.  In fact, they’re both based on the same thing – value.

Value – what’s the share worth?

First things first – we need to know what the share’s value is.  Note the emphasis – value is what you get, price is what you pay.  To determine when to sell a share based on fundamentals, you need to know what it’s intrinsic value*.

Unfortunately this means our approach isn’t applicable to traders.   Terms like “price action”,”trading range” and “moving average” don’t enter the value investing lexicon, so I’m afraid this won’t help the chartists and speculators.

* For more details, read anything about Buffet, Munger and Graham and eventually you’ll get the skinny on value investing.  I won’t go into it here – it’s a topic that’d require several more posts.

Share Value vs Cash in hand

So you know the value.  Hopefully you’ve bought below value and the share is now at or above value – giving you a nice paper profit.  You are now faced with two choices: Hold onto the share or sell it, making a profit and paying capital gains tax.

In both instances you will be left with a certain measurable value.  In the first option (holding), the value will be the value of the share.  In the second option (selling) you’ll be left with value in the form of cash minus your capital gains tax liability (and brokerage).  In other words your “sale cash”, which is your initial capital plus + net profit.


So when do I sell Q?

Easy – when your sale cash exceeds your share valuation.  If you think a share will give you $10.00 in value and your after-tax sale cash is $9.00, why would you sell?  Same in reverse – why hold onto a $10 value share when you’ll get $12 after selling and paying tax?

If you do your valuations right you’ll have taken into account business risk, tax and inflation so the two outcomes (share value and holding cash after sale) can be directly compared.

The Sell Formula

Your blogger happens to be handy with formulas, so I’ve come up with this one to help you figure out when to sell:

S > (V – B*t) / (1 – t)

Where S is sell price, B is buy price, V is value and t is your tax rate on the capital gains (this will change depending upon your tax bracket and/or any capital gains discount).

Recreate this in an excel file and you’ve got a ready-made sell calculator.

What if the Value Changes?

Well, let’s hope that happens.  Good companies should increase in value over time as they add to their equity base.  In which case your sell price will become higher.

If your company decreases in value the same approach still holds.  Why hold onto a company you now think is worth $9 when you’d get $11 if you sold them?  The fact you bought them at $15 might hurt your pride, but it doesn’t change the facts – it is better to have the cash in your hot little hand than hold onto a losing stock where your only strategy is hope

It Can’t be That Simple

Actually, from a mathematical point of view it is that simple.  The only thing this approach doesn’t factor in is opportunity cost.  You may have a share worth $10, and the cash value on selling will be $9.50 – so you should hold according to the math.  However, if you identify another under-valued share elsewhere that will give you a superior return, you may be better off cashing in and reinvesting the profits into the new share.  If you don’t, and that under-valued share doubles whilst you wait for a measly 50c price rise in the share you held, you’d be upset.

That is what we call opportunity cost.  It’s something you’ll have to subjectively manage yourself with every sell decision.

So there you have it – a nice, simple and systematic approach that gives a pretty good sell price signal.  Using this approach and some subjective cost-of-opportunity assessment, will aid you in selling at about the right time from a value point of view.  The share price may soar beyond value, but then you’ve entered the realm of speculation as to where it’ll finish. I much prefer to book solid profits at a rational price then pray to Mr Market for another 5%, when the end result could be a 10% drop.

Latest posts by __ADAM__ (see all)


  1. Nice, straightforward, and easy to understand. The real secret sauce, though, is coming up with a solid and useful valuation that can be used in the formula…

  2. Very true Womble, but we’re a little too jealous of our IP to put our valuation method out there holus bolus. We do talk about our basic valuation philosophy on the Empire website. Sufficed to say, we take a lot of cues from Buffet/Munger and Roger Montgomery.

  3. The problem is that value is unobservable and unknowable. What you are actually talking about is perceived value which at the end of the day is your opinion. While it may be, and probably is, an informed opinion, it is still all just guesswork. In other words attempts by finance to turn it into some kind of science are faulty. It is an art.

    • I’d agree to a point Lionel. Value will change based on the perspective of the investor because their investing requirements and assumptions will differ. However, these variables can still be used to derive an intrinsic value (to the investor). We base our valuations on a discounted cash flow with a required return of 15%. The subjective nature of the valuation comes into forecasting return on equity (ROE) and the assumed dividend/reinvestment split. That subjectivity means value investing will never be a formulaic science, but I believe it’s better than astrology by candlestick 🙂

      • We base our valuations on a discounted cash flow with a required return of 15%. The subjective nature of the valuation comes into forecasting return on equity (ROE) and the assumed dividend/reinvestment split.

        expected future cashflows are subjective. the perceived value of something is a guess. With some people the guess is more informed than with others but it is a still a guess — something that people seem to lose sight of, particularly analysts who often talk about value as if it is observable. (maybe their self confidence or self delusion is why you rarely get confidence intervals on valuations)

      • Funny you mention confidence intervals Lionel. It’s something I’ve often talked about with my business partner The Prince and something I’d love to introduce to Empire. Showing our valuations as a probability distribution rather than a single number output would be a far better representation of a company’s value. It would allow the uncertainty around the forecasted variables to be shown graphically – the wider the distribution the less certain the valuation. Alas products like “@ Risk” aren’t cheap; however it is definitely on the to-do list.

        • Who needs @Risk? (what does it cost BTW)

          You could e.g. sample from a company’s current returns distribution and so on. Or alternatively sample from company’s earnings distribution and then do a valuation. MC methods like that are really only appropriate for well established companies. Bear in mind that it is never possible to eliminate the subjectivity from the equation — by using distributions you are pushing the subjectivity into different areas — but nevertheless it is a better approach IMO and would separate you from rote learnt methods used by other analysts.

        • Distributions beat point values anyday! Yey for pushing the cause! 😉

          In the event of not actually producing a distribution chart, even just including a few other variables such as standard deviation and standard error give the interpreter a MUCH better idea of what is actually going on.

          My 2c

      • Good points Lionel. Very good points actually. Thought about giving up law? 😀

        Its’ why we are skeptical of products/pundits stating, absolutely, that stock XYZ has an intrinsic value of $X.

        In reality, there is no such thing (which is why I think many prefer to just say the price is right, when the market price is just a temporary agreement between a minute number of participants – its the non-participants that move market value)

        Going to confidence intervals and a range of valuations is more correct (but will never be exact – nothing in finance or economics can be exact because they are human sciences), but not very “sellable” to the average investor.

        • Yeah it bugs the hell out of me to read reports in which people talk about “value” as if it was something observable/knowable like the current share price or a cricket score.

          As far as not being sellable to an average investor, as an investment advisor IMO you should be cautioning people away from investing in the stock market if they didn’t “get” the idea that there is a wide distribution of outcomes.

          The key is to frame it in everyday laymans terms but every investor needs to grasp this conceptually even if they don’t know the underlying stats etc.

        • We are actually investment managers (as opposed to advisers) who like to talk about our methods. None the less, your point is still valid and I am sure the concept can be easily shown with a simple graph.
          With regards to distributions, you wouldn’t get enough sample data from past earnings history. It would also violate the whole “past performance is not a reliable indicator of future earnings” blah blah blah principle. However, assigning distributions to future ROE/reinvestment splits based on your subjective analysis (it’ll never be a science!) would provide a distribution output for your valuation. The wider the input distributions (i.e. the less certain you are) the wider your valuation output, depending on the impact of the variable in your calcs.
          Definite food for though Lionel – maybe I’ll tackle the discrete vs “confidence interval” valuation issue in a future post.

  4. Q, I think it’s fair to say that your sell approach is Graham-esqe value investing, but I don’t think that you could call it Buffet-esqe. Buffet preferred holding time is “forever”.

    This relates back to the tightening of your valuation model. If you only buy the very best companies at the very best prices – there shouldn’t be a need to sell (based on price alone).

    • Early Buffett (when he ran a series of partnerships) was effectively Grahamite style value investing – he sold quite a lot during that period. In fact, for a large period of time he was a trader. (mainly options, but did a lot of arbitrage). He also tinkered with “cigar-butt” plays – something that works in sideways markets, but not the boom market of the 1960’s that he later experienced, so he changed tack.

      He found “buy and hold forever” later in the 1970’s when he realised he can take advantage of huge equity positions in off-market transactions, i.e buying up whole companies. This removed the problem of opportunity cost during the woeful sideways market of 1969 to 1982.

      Search/research any of Buffett’s quotes about being a small (i.e under $100 million) investor: you should be buying and selling, taking advantage of an inefficient market that over and underprices businesses. The bigger you get, the harder this gets, unless you start to take advantage of arbitrage opportunities (ala Seth Klarman).

      People make the mistake of either A. following Buffet religiously, thinking they can emulate someone who is bordering on Aspergers/OCD and has been that way for 70+ years or B. that he always was a “buy and hold” investor. Even his partner, Charlie Munger, started out as a property speculator.

      We have a few companies that we don’t want to sell – because they are truly wonderful. But if the market goes crazy and we can find better bargains, then you have to sell.

      This is where the concept of risk is mistaken – sometimes holding is a risk, because of opportunity cost. Using a selling technique as Q described is actually a risk management technique.

      Opportunity cost goes both ways and markets always overshoot both ways.

    • If you buy the very best companies, then their value should increase over time, in which case your sell price should also increase. So if the market is rational, you should never have to sell a truly wonderful company as the sale price will never outpace its value.

      However, if the market is truly manic (pre GFC Aussie share market is a classic) then not selling would be a mistake in my opinion. If the market reverts to value in the long run, it must do from both and under-valued and over-valued perspective. So if COH was trading right now for $100, I’d be selling up and either waiting for it to fall back below its value or for it’s value to catch up to it’s sky-high price in the future. Otherwise you have cash tied up I a share which is very over valued and will most likely return to its value in the long term (i.e. decrease in price). That cash could be better utilised elsewhere in the mean time.

  5. ceteris paribus

    I strongly disagree that buying low is the easy part. People buying the All Ords at 3150 in March 2009 did so against widespread screams that the index would retreat to below 1500.

    Hindsight glosses over the anxiety prevalent in the market at that time.

    In fact, while the business cycle and equity cycle remains the great opportunity, there is no such thing as precise “buying in” and “buying out”. It is all about percentages- “averaging in”, progressively and systematically, at cheaper times and “averaging out” in expensive times.

    Anyone who tells you otherwise is playing to be a Delphic seer.

    • Maybe I was too cavalier with my opening paragraph Ceteris. However, my point was to highlight how the buying of shares is so often analysed at the expense of how to sell them.

      I agree that no one can consistently predict the top or bottom of a market, however I believe using a value-based approach goes a long way to avoiding the risk of loss by purchasing too high (or selling at the wrong time, as discussed in the article).

      Edit: Sorry – that last sentence sounded sarcastic, even though it wasn’t my intention. I am actually curious as to how you know when to switch from “buying in” to “buying out” with your averaging approach.

      • ceteris paribus

        Yes, I agree that value-readings are essential to gauge where makets are moving towards extremes, based upon positive and negative sentiment.

        My point was that it is not easy to pick where the cycle turns, at the bottom or the top.

        I now see your comment about easy “buying-in” was a way of emphasizing the point that many people, illogically, pay extraordinary attention to buying price but little attention to strategic selling.

        As for my method, I never hit winners nor losers in equities- a very conservative and very patient percentage index game in a super fund, which has “no switching fees”, minimal buy/sell spread and 48 hour switching time. Very boring.

    • It is all about percentages- “averaging in”, progressively and systematically, at cheaper times and “averaging out” in expensive times.

      At the risk of flying completely solo, I disagree. I think that there is an “optimal extreme”, at which point you should transition from no-position (or not change) to all-in (or out.

  6. I think this is where pure value investors really lose out. Yes, use a valuation as a trigger for selling. I use it to place a tighter trailing stop. That way you still get to ride the big bubbles.