OLD ARTICLE – Original posted on June 30, 2015
In the context of the stock market and its underlying individual listed companies, what is the value of building, making and using forecasts?
By their very nature, forecasts for companies are an attempt to anticipate the future earnings of a company. They are not a promise of what the exact earnings are going to be, nor should they be a crude thumb-suck. At best they offer us an expected trajectory for these earnings: up, down, up-a-lot, or down-a-lot.
Forecasts are little more bias, logically-built arguments for where a specific person (or persons, in the case of broker consensus) think earnings are likely to be.
By their very nature, forecasts are often wrong.
Does that make them useless?
Like many things in life, it is not that simple. Not all forecasts are worth the same, nor are all forecasts equally inaccurate.
One analyst can be better than another analyst. Likewise, some sectors are easier to forecast than others.
A top rated analyst that is close to management and has an excellent grasp on a sector and its companies’ fundamentals will probably have a better quality forecast (or, at least, more justified argument for why their earnings forecast is where it is), than, say, an average man-on-the-street matchbox calculation. Likewise, the property sector with its stability in earnings, communicated contractual escalations is easier to forecast with greater accuracy (i.e. less moving parts) than a junior miner with mines ramping up, tonnage, cost and spot price variables all swinging results massively.
As a small cap analyst, my forecasts are almost always wrong. Smaller companies tend to have more variable and volatile earnings than their bigger, more stable blue chip counterparts.
Still, I find a lot of value in building these forecasts. I like to know where I think I am going and try to anticipate what speed bumps, sharp corners and other obstacles I might face along the road towards my destination.
Let me phrase it this way: After researching and understanding a business, after all the management meetings in the world and site visits, sitting down and putting together historical and forecast financial results for a company gives me a lot of insight into the key financial metrics and the interactions between then for a business.
I like to see how the qualitative information I gleamed about a business aligns to the quantitative financials of a business. This interaction is at the core of being an analyst.
If you want to only invest into “stories”, go buy a book. If you want to only invest into “numbers”, go be a quant or put your money into an algo. Stock picking is the intersection of both the story and the numbers.
No matter the complexity of any financial model I have ever built and its resulting forecasts and valuation, I have found that it often comes down to one or two variables. Perhaps it is the expected growth in sales volumes (real growth), sales inflation (pricing power or spot prices), cash generation and the balance between working capital, capex and growth… It might even come down to the estimated Cost of Equity or the Terminal Growth Rate in a DCF.
It is not always the same key assumption or two and it varies between company to company, but only after building the model, settling on my forecasts and arriving at my valuation and conclusion do I know what key metrics to look for in the company.
These key metrics I will then follow like a hawk and, if they change, then the company’s fortunes change, its fundamentals change and my opinion thereon changes.
I would never have this level of clarity if I did not sit down and work through the tedious intricacies of building a model and related forecasts for the said company.
So, perhaps the forecasts in and of themselves do not have a lot of value in them, but the process of building them up step-by-step, segment-by-segment and variable-by-variable have immense value.
Hence, I do believe that despite all odds, misses and inaccuracies that can be pegged to them, forecasts do have value. I would, though, be cautious of using the forecasts of others, especially where their key underlying assumptions are not disclosed.