There has been a thought I've been mulling over the past few years with regard to the calls I've made during my investment career, and the performance subsequent to the formation of these views. Why did I call some companies so badly? And why did I miss some stonking ten-baggers? It's easy to just fall back on "information-set was incomplete" as a convenient excuse, but I think some reflection wouldn't be amiss.
Let's start with the assumption that I was always a value investor. That is, apart from what my tactical trading opinions may have been at the time, I believed that the stock had an intrinsic value that could be estimated within an error range of approximately 30%. The reason for the wide range would be lack of non-public information, acts of god, macroeconomic events etc. However, when I see a stock become a ten-bagger, and it doesn't seem at the point in time in the future that it is considerably overvalued, I ask myself how could I have been SO wrong! What analytical flaws plague me?
I have since hypothesized that my problem is that I have been schooled in a value investors school that is very static. I think it suffers this problem due to two main reasons: most value investor analytical approaches use backward looking data, such as formulas, charts, reversions to the mean, etc. Secondly, the investment process in many value investment houses were primarily formulated in environments that were slow moving, mature, declining in growth rate. Such as mature Europe and America.
I shall attempt to explain why these two factors pose problems for value investing in fast changing environments. Firstly, being backward looking, mean reverting etc will totally ignore structural changes. In a mature environment, such structural changes which have such massive impact on a stock do not happen as frequently. Furthermore, fast changing markets tend to be younger markets. Hence the financial time-series data is shorter, resulting in judgements based on past data less meaningful. On the second point, investment process being cultivated in slow moving markets. The problem with this is that there is a widely held belief that a stock's intrinsic value should be well known after doing thorough research, and unless anything particularly unforeseen occurs, the stock's intrinsic value should appreciate at it's cost of capital from the time that the initial valuation is conducted. This creates two problems. The habit of not revisiting stock valuations as frequently as one should, resulting in outdated and therefore incorrect valuations. And far more worrisome, creating the culture that looks down on those who have frequent valuations which veer violently over time, as this apparently suggests that the analyst doesn't have a firm grip on value and is being swayed by the market and sentiment.
Paradoxically, this creates a culture sometimes of value investors applauding those who have valuations far from intrinsic value. Even though a 'value' analyst may have had his number 80% below NAV, of a company, which is clearly incorrect if the book consists of healthy assets, it is seen favourably that he stuck with his valuation through thick and thin. This indicates good process, having conviction in his valuation.
But sometimes, realities do change. And in fast changing economies, faster than one would think. The clear example would be how Nokia and Blackberry, the former at one stage being the largest phone manufacturer in the world, the latter the largest smartphone manufacturer, so quickly fell from grace.
So what is the correct process to adopt? I propose that value investing should move away from the shackles of being the step-child of bond investing, discounting future dividends into perpetuity, and align itself more closely with the real options approach.
Companies, and the economy, exist in a tree structure, such as in the real options approach. At each time step, the states have changed. Valuation for the prior period was conducted by averaging out the future states that could have occurred, such as Expected Value in statistics. However, upon progressing to the next time step in the tree structure, the forward-looking probability structure has changed, and certain branches which initially formed part of the averaging calculation have now been eliminated, and therefore a whole new valuation needs to be done with this new statistical set. This explains why the 'unwinding of discount' attitude which is prevalent amongst value investors is unhelpful.
Essentially, what I'm calling for is the suggestion that a value analyst should focus on doing a total reevaluation of all the stocks in his portfolio during each 'time-step'. This should perhaps be every 6 or 12 months. And it's okay if the value changes wildly after a single time-step. Irreversible things happen to companies in short periods of time, both good and bad.
A value investor who is stale isn't really a value investor at all.
