Building a Financial Model

A lot of investors rely on analysts' forecasts when they are assessing the potential value of a stock. That's not necessarily a bad idea; analysts have privileged access to companies' directors and often have substantial experience and expertise in their sector. Still, if you really want to do your own research, you're going to need to start building your own financial models. For some stocks, there aren't any forecasts available; for others, the single forecast out there might be out of date, or it's a forecast by the broker, which you can expect to be somewhat optimistic in many cases. Besides, even if there are other forecasts out there, building your own model will give you an in-depth understanding of the company and its business, far more than just reading the annual report. First, you'll need to assess the business model of the company. Is there a convenient unit of volume? For instance, it's houses with housebuilders, kilowatt hours with electricity companies, and so on. That might apply on the cost side too; for retailers, square metres of retail space is an important figure. Given these units, you'll often be able to estimates revenues and at least some of the costs - this is extremely useful, as you can then analyse where growth is coming on - is it coming from increased volume of sales, or just increased prices?

You'll also need to look at whether gross margin or operating margin is the key ratio. For retailers, it's gross margin - effectively that measures what mark-up they're making on their goods. For a software company, on the other hand, the gross margin is usually 90% of more - there's practically no cost of sales - so it's operating margin that is more important.

If a company has mainly staff costs, you can estimate the number of staff and what they're likely to be paid - obviously a caterer or construction company will tend to have lower costs per employee than a computer consultancy or investment manager.

Start off with the last couple of years' real figures, and then simply build up next year in the same format. By employing unit-based forecasts, or by looking at what margin you might expect, and using a chosen growth rate for revenue, you can build up next year's profit and loss account. For instance, with a computing firm I might look at what other firms in its area are making as operating profit margins - and then forecast, say, it will have margins a couple of percent lower, because it has some duplication of costs while it's setting up an Indian outsourcing arm.

You might also adjust the margins if you know that cost inputs are increasing - for instance, in the food industry or in forex master levels review, where malt and hops shot up in price in 2008-9.

Once you've created the model, you need to check it. A good way to do this is to work out the other ratios - for instance you might work out contractor day rates for a computer company by working out the daily cost per employee, and doubling it. That's very rule of thumb but it should give you an idea of whether the model is working.

A huge advantage of having this kind of model is that you can flex it. You can say 'what if' the price of fuel went up again? What impact would it have on British Airways? You can ask 'what if' a company hired more staff, 'what if' Marston's closed a few pubs, 'what if' the rate of bad debt at Lloyds increased.

Broker forecasts all make assumptions, and you don't really know what those assumptions are - you can make a range, and you'll be picking assumptions that you think are viable. If you're a contrarian investor, you may end up with a rather different result from many of the analysts, simply because you have made different assumptions about the economy, or the oil price.

And when the company reports results, you'll be able to check your model - and your assumptions. You can really hold the company to account, because you have a financial model that will show you exactly what is going on.