Financial Modelling: Assumptions Built on Assumptions

Wouldn’t it be great to know the future? Sure it might take the spontaneity out of life. But you could become the best investor ever.

There is no such thing as a perfect investment record. Everyone picks up some losers along the way. But if you had an idea of what might happen, you could avoid the ‘bad eggs’ as they say.

This is what financial models try to do. They predict what might happen in the future. Not just profits, but revenues, debt, cash flows and a bunch of other business fundamentals. Judging by the calibre of investors that use financial models, from investment bankers to private equity firms, you’d think they’re the best way to value businesses.

And maybe they are…for the short-term. But I’d argue financial models are a lot less useful than most investors believe.

No Informational Edge

In the early 1900s, investors that could get their hands on annual reports had a huge advantage. Because information was not widely disseminated, having financial data on a company gave you a far better understanding of that business than many other investors.

But fast forward to today, this informational edge is almost non-existent.  All the important information relating to a business can be found at the click of a button. And for those who really want to dig deep, they can always call up management.

Investors aren’t just looking at the same figures, they’re also running the same models. While they might differ slightly with growth assumptions or debt accumulation, they are very similar as an end result.

Therefore it doesn’t seem that running financial models give you a massive informational edge. So why do company’s model financial statements into the future?

I would think it’s for security. Having a number that says, company A will have X dollars in earnings and will produce Y dollars in free cash flow. Think about it, some firms are investing hundreds of millions into businesses each year.

They want to be dam sure they knew everything about the company and what it might look like in the future to try and prevent potentially losing investors money.

But not only does financial modelling not give you an informational edge. It can also give you misleading information.

What’s worse it can give you misleading information.

Assumptions on Assumptions

Say you were going to forecast a three statement model (income statement, balance sheet and cash flow statement) for Apple Inc. (NYSE:APPL).

The first thing you’d do is to collect a bunch of historical data, say the last five annual reports.

Once you’ve put it into an excel sheet, you would then start creating you assumptions. You might take the average growth of revenues over the last four years and project it out. Or you could take a top down approach and keep Apple’s market share constant while growing the population of smartphone, laptop and table users.

Then, to make assumptions on things like COGS, accounts receivables and a bunch of other element, you’d likely link back as a percentage of revenues. It’s a simple but common way you’d build out a three statement model.

But if we step back and look at the model, there will only be two or three key variables driving everything else. Therefore if your assumption is wrong for one key variable, the whole model will be littered with errors.

So what’s the alternative?

Simple Yet Conservative

You could always just forecast free cash flows. You’d still be open to make forecasting errors on growth and your discount rate.  But it’s a far simpler, quicker way to work out more or less the same thing.

Let’s use the example of when Warren Buffett first bought The Coca-Cola Company (NYSE:KO).

The year was 1988. Coke was worth $14.8 billion at the time, trading at five times its book value and had a dividend yield of 6.6%. A dividend yield of 6.6% might sounds great, but the 30-year treasury yield was 9% at the time.

Buffett saw something in the company others didn’t. They were a behemoth, growing sales each year with a durable competitive advantage and able managers.

Oh yeah, and the business was cheap.

In 1988, Coke reported around $828 million in free cash flow – the cash a business can either reinvest, pay out special dividends, buy back shares, etc. From 1981-88, Coke was growing free cash flow by an average of 17.8% each year.

Therefore, if Coke could grow free cash flow at similar rates into the future, it would be a business worth buying.

Let’s assume free cash flow will grow at 15% in the next five years. Four years after that, it will grow by 10%. And on the tenth year it will grow by 5% in perpetuity.

Based on these assumptions, Coke was worth $21.8 billion in 1988. Meaning the stock was trading at almost a 50% discount when Buffett bought in.

Of course if you want to be more conservative you can. For example, let’s say free cash flow will grow by 10% in the first nine years and then at 5% from thereafter. Based on these more conservative assumptions, Coke was still trading at a 20% discount.

The point is you don’t need to create financial models to value businesses. The real is in your ability to analyse a business, understand its potential future and the downside risk you are taking on. Yet, modelling isn’t all that bad.

If you have some time and are willing, building financial models with give you an intimate understanding of financial statements. You’ll know how they interact with each other and how companies sometimes try to hide behind accounting figures.

Leave a Reply

Your email address will not be published. Required fields are marked *