Several people have mentioned recently that they would like to know how I go about determining a fair value for a stock. I use a fairly long, involved process to accomplish this but I thought I could share it in summary here and if people would like more detail they can email me privately on the topic.
Basically I use a mathematical method where I discount all future earnings and dividends back the the present to determine what a stock should be worth today. This method of course involves several assumptions, especially on the earnings growth rate going forward. I try to be as conservative, yet realistic as possible when valuing a stock. Around the internet you will find several versions of disounted cash flow calculators or dividend based discounted cash flow calculators. I use one version of each of these during my analysis.
Variables that I need to determine:
EPS (earnings per share) last 12 months
Estimated P/E (Price to Earnings) Ratio in 5 or 10 years - (to be conservative here I use the lowest P/E the company has ever traded at)
Discount Rate - I always use 11%
Current Dividend Rate
EPS Growth Rate going forward - This is the most important number in the calculations. I look at EPS growth long term as well as estimates going forward. I try to select a number that is conservative yet realistic given many factors such as the company's past, industry fundamentals, analyst estimates, company's performance in poor conditions, and future prospects.
All of these numbers get plugged into the two calculators mentioned above. Out of this I get two share price estimates. Let's say $40 and $55 per share are the estimates of fair value. I then average these two numbers ($48), and compare the current share price to the average. In order for the stock to be considered cheap, the company needs to be trading at a significant discount to the average price and in most cases the company must be trading below the lower estimate ($40). If a company is trading below the lower estimate then that is a good sign that the company is trading at relatively low levels compared with my earnings expectations and compared with what it has traded at in the past during similar market conditions. The reason that this happens is because as mentioned in my models I have used the lowest historical P/E, and I am conservative with my earnings estimates.
I will never use yield when determining a good entry point. The reason for this is because stocks always behave based on earnings growth and anticipated earnings growth. If a stock is yielding at a very high level relative to historic levels, this usually means that the earnings expectations are poor. Does this mean the stock is cheap? Not necessarily, it's only cheap if the market is pricing the stock as if it will grow earnings at 6% and it actually grows earnings at 10%. Perfect example of this currently are Pfizer (PFE), Rothmans (ROC), and Bank of Montreal (BMO).
Because stocks are always priced based on anticipated earnings growth, when modeling like this you always need to take a guess at future earnings growth in order to determine a price. If you can make an educated guess based on research, and be conservative then at the very least you will know when the stock is near fair value, and when it's expensive. For example recently I bought Sun Life Financial because the market was pricing the company to grow their EPS at much less than 9% per year based on my estimates where I used their lowest P/E ever as a baseline.
I certainly do not claim this process to be the 'be all and end all', however I truly think it gives me a much better idea of what is expensive and what is reasonable out there as far as valuations go. If a really great earnings growth company stumbles over and over, and their earnings continue to decline over time, then this model will not catch that, as it will probably indicate that the company is cheap each quarter as earnings are reported lower than estimates. What I find this model does catch is when a quality company is being discounted for fear of reduced earnings growth. Obviously the more stable and steady the earnings growth is the better the model works which allows it to work better with large established 'blue chip' non cyclical companies, which are typically the companies in which I invest.