Some thoughts on averaging down & averaging up

John Hempton has a very interesting post on when to average down into a stock.

As a summary, one should not average down into a stock if

  • a company has a lot of financial leverage
  • a company has significant operating leverage
  • the company is in danger of becoming obsolete

I think this is already a pretty good advice, as a counter example he gives Coca Cola where one can average down “without much risk”. As this is a very interesting topic, I wanted to contribute my 5 cents to this:

Behavioural biases at work

In my experience, averaging down is often motivated by a couple of behavioural biases.

The major bias which “helps” investors and especially professional ones to average down in the wrong cases is in my experience the “over confidence” bias.

Often the case goes like this. An analyst/fund manager has really invested a lot of time into the analysis. He has read the reports, made his spreadsheets, he has spoken with management, has been “on site” and he is really sure that this stock is undervalued and starts buying. And the CEO is a really charismatic guy. The stock starts going down and the analyst/PM calls the CEO and the CEO says everything is great and the analyst becomes more confident and buys more  The stock goes further down and the analyst calls again. In some blogs questions about the management or the business come up. But the CEO is still super charismatic and tells the analyst some things he shouldn’t actually tell him under the promise not to tell anyone else (great new product coming etc.). Then it turns out that the CEO is either a crook or an idiot (or both) and the company collapses. I have seen this for instance in the infamous Globo Plc case but also in others.

Charismatic management is always a big issue, especially if there are questions about integrity. As I mentioned in the blog quite often, “Armchair investing” has some advantages in such cases.

Another quite frequent bias in that regard is Anchoring.

When your position is already at a loss, averaging down reduces your average percentage loss. I have to admit that I still sort my holdings according to percentage gains since I bought them. Averaging down suddenly makes a stock look much better in percentage terms. I could imagine that this could also be a (unconscious) motivation of fund managers to average down before reporting dates to show lower percentage losses.

This is clearly one form of Anchoring, where you behave differently as you have an “anchor” in the form of the initial purchase price.

So I guess one should verify if the urge of averaging down is influenced by behavioural aspects.

Integrity and Alignment of Management

And, following the Globo example my advice would be that you should never average down if the integrity of the management is questionable or if the interests are not aligned with (minority) shareholders. This is especially important if management and/or the major shareholders change.

I linked to Guy Spiers Horsehead example as a prime example for this. Even in my blog I had two “distressed” situations (IVG and Praktiker) where there was little or no alignment anymore between management and shareholders and selling at a loss turned out to be a good decision.

One could argue that you should never even think about investing into such companies, but sometimes you invest and it turns bad only afterwards.

On the flip side however I think you can invest (and average down) into more levered business model if management and shareholders are well aligned. Prime examples for this are Buffett’s long-term holdings Geico, Amex and Wells Fargo which all use leverage but are conservatively managed. The trick here is that those companies are often (a lot) less levered than their competitors. Berkshire itself is by the way also a good example for a very good long term levered investment where you could have averaged down in the past without much risk.

The initial assumptions were too optimistic

Another case where averaging down in my opinion is often a mistake is when your initial assumptions turn out to be  wrong.

So you thought for instance at a share price of 30 EURthat the stock should be worth 10 times next years earnings which you estimated to be 5 EUR per share.

Then it turns out that the company will only earn 2,50 EUR and the stock falls to 10 EUR. Your new fair value (10 times 2,5) would be 25 EUR and the stock would be highly undervalued at the current (much lower) share price. This would be a great argument to buy more.

However in my experience, if you are wrong once, there is a very high likelihood that you are wrong again. Of course sometimes you are lucky, but I would still advise against averaging down into companies who perform significantly worse than expected.

Interestingly, Prof. Damodaran did exactly this thing with Valeant some weeks ago.

That acceptance of feedback, though, has to be accompanied by an affirmation of faith; since it led me to buy the stock at $27, when my estimated value was $43 in May 2016, it should lead me to buy even more at $15, with my estimated value at $32.50. So, I doubled my Valeant holdings, well aware of the many dangers that I face: that the operating decline that you saw in the third quarter of 2016 will continue in the future years, that the debt load will become more painful if interest rates rise and that the recent indictments of executives will expose the firm to more legal jeopardy. If the essence of risk is best captured with the Chinese symbol for crisis, which is a combination of the symbols for danger and opportunity, Valeant would be a perfect illustration of how you cannot have one without the other!

As much as I respect him intellectually, I don’t think this is good portfolio management. So far he hasn’t lost money on his doubling up, but he could have made much more with buying the S&P 500 instead. On the other hand, he is a famous professor and I am an anonymous blogger…..

My biggest problem: Averaging up

In my own portfolio and since I run the blog, averaging down is not so much a problem. I have only done it in small amounts in and many cases not succesful which further reduces the risk of doing so in the future.

However my bigger problem is that I failed to “average up” when stocks did really well. I often start with relatively small positions (2 or 2,5%) and then try to observe the stock for a certain time in order to decide if to increase or not.

In many cases, the underlying business did as well as I expected and the stock went up quickly. And I failed to buy more. Some of my most succesful positions (IGA & XAO, G. Perrier) were small positions and I failed to increase them.

Another example is DOM Security where I started with 2%, my expectations materialized quickly, but I didn’t do anything because the price went up by more than 30% in a few weeks.

So I have been thinking on how to “cure” this and I came up with a few ideas to improve my investment process:

  • start in general with larger positions. In the past I would have started with a 2,5% position. In the future I want to start at least with 3 or 3,5% positions. This sounds small but over time should have some effect
  • increase based on the initial allocation. Up to now, I have restricted myself not to invest actively more into a position once it is above 5%. In the future, I will make this decision based on my initial allocation. So if I have invested 3% and the stock has doubled to 6% portfolio weight, I can still add 2%.
  • introduce a hard “timeline” to decide if I double up to a full position or sell for any “half” positions. In the past, I watched those stocks and then did …nothing. In the future I will set myself a specific time frame where I have to decide “hard” if an upgrade to a full position or sell. I will start with 12 month and apply this to my “active” 2016 new entries.

I think this doesn’t make the process a lot more complicated but should help me in giving good stocks a higher weight independent of their short-term price fluctuations.


Be carefull with averaging down. As John Hempton mentioned, there are many cases where averaging down is maybe not a good idea. To his list I would add the cases of “problematic management” and a “failed initial investment thesis”.

At least for me however, the failure to average up on good stocks is even more problematic. I hope the 3 changes mentioned will improve the investment process and hopefully also long-term results.



  • Hi mmi,
    did you consider having a flexible approach? If a stocks is in a down movement it could be better to go with smaller positions in the beginning. If the stock is in an uptrend or illiquid like DOM security starting with larger positions could be the better way.
    I my opinion averaging down is not bad by definition but if you have not checked if your thesis is still intact.
    I don’t understand why you consider introducing a timeline. If your stock moved up but hasn’t reached its full potential yet and you have other stocks that have; why would you sell or buy more of that stock. I think checking your thesis more often does the better job. If you reduce the number of stocks in the portfolio this is less time consuming anyway.

    • Well, the “flexible” approach is what I have now and it doesn’t work now. The timeline is a try to still keep it simple (one point in time) but hopefully improve ….and it should hopefully reduce the number of stocks.

      Checking the thesis more often in my opinion does nt help as I don’t have that much time…

      I have to think about the up- or downwards price trend but again this could be some kind of anchoring.


      • I agree, keeping it simple is key. However, I wouldn’t introduce too many new rules because they often require some more rules to work. How do you decide if you sell or buy on such date. It makes your life harder because the timing of that decision is random etc. Maybe a simple reduction of number of stocks is a good way to start. That’s just an idea.

  • Interesting topic and one I haven’t quite figured out for myself yet.

    Many investors have different takes on this – my impression is that this is mostly not based on sophisticated research but because of experience bias (in case this term doesn’t exist in that regard yet I invented it):

    They actually have an insufficient sample size but still draw conclusions and make rules for themselves based on their perception, which is mainly shaped by the relatively few outcomes they have seen so far.

    What makes the most sense to me personally is Damodaran’s approach to focus on my best estimate of fair value and the current price, and do not care too much if I just lowered my FV estimate (i.e. rather to try to factor the higher uncertainty regarding the risk of poor judgement on my side into the FV estimate or require a higher margin of safety).

    Excited to hear your opinion. Since this is my first comment but I have been reading your blog for quite a while, many thanks for your work here. Always a nice read.


    • pete,
      Damodaran’s approach strikes me as rather critical. As I asaid, you were already wrong once which in my opinion increases the risk that you are wrong the second time.


  • Very elaborated article! I refrain from averaging down, because of bad experiences. It hurt much more if it goes wrong, than it does good if it works. Still I have two position which are have been hit very hard. Their portfolio weight is so small that I’m forced to either sell them or to average down, otherwise these immaterial positions are taxing my 25 stock limit. Your article will help me to make a decision regarding these.

  • very interesting post, thank you for sharing your thoughts
    I think an alternative approach is to use time “time thresholds” building gradually the position in 2-3 steps after key events unfold the way you d expect them. For example, Buy 1% Majestic Wines now that I like it and wait for the next trading update, Buy 1% more after the profit warning as the market over-reacted and the thesis is still valid. Buy another 1% after the post-Xmas update that confirms the thesis. I think it might sound more tired-some but it can save a portfolio from blow ups.
    The downside is that such an approach requires large holdings in cash but it gives more time for thought and reflection.

    By the way, after reading your post on Globo, think Telit Communications (TCM LN) satisfy most of the criteria to become the next candidate..

  • How do you think about Admiral?
    Did they develop that good abroad that they could be worth averaging up? Or do you expect brexit and ongoing pound-weakness to overshine that effect? How to value chances and risks?
    IMHO Admiral could be a good object for discussing “exit or double up”.

  • You are starting to assume more risk…(which is good in my view).
    I would add (if I may) to consider a more concentrated portfolio (i.e. 20 stocks max).
    Sorry for being so blunt, I know this is quite personal.
    Take care and all the best wishes and luck for this new year!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.