Methodology

This page explains how I arrive at a number. The number on its own is worth very little — what matters is the reasoning behind it, and you can’t judge that unless I show you how it’s produced.

What I publish

Three kinds of valuation, and each one says which it is:

Intrinsic. What a business is worth based on the cash it can generate over its life, discounted at a rate that reflects its risk. This is most of what I do.

Relative. What the market is paying for comparable businesses, and what that implies for this one.

Contingent claim. Where the value is dominated by an option rather than by expected cash flows — an undeveloped resource, a single drug in trials, a company whose equity is effectively an option on a recovery.

Every valuation ends in an estimated value per share, shown next to the market price on the day. It never ends in an instruction. I don’t tell you to buy or sell anything, and I don’t publish views on where a share price goes over the next weeks or months, because I don’t have any.

The framework

The intrinsic valuations are discounted cash flow models. I forecast what a business earns over the next decade and what it has to reinvest to keep earning it, then discount those cash flows at a risk-adjusted discount rate. Growth is never free here: anything I forecast has to be funded out of reinvestment, and nothing grows faster than the economy forever.

The method is Aswath Damodaran’s, taught in his public courses and his books, with his spreadsheets freely available on his website. I highly recommend you check out his work. I work from his framework and his published data, in my own implementation of it — adapted to the way I work and to the companies I cover, most of which are small enough to have no credit rating and no analyst following. Any errors are mine, not his.

The model

I value companies in my own spreadsheet rather than in a borrowed one, and subscribers get the populated file for every valuation. It is one input sheet and one valuation sheet:

The input sheet holds the base year — revenue, operating income, tax, debt, cash, share count — each with a note saying which filing it came from. Then the assumptions: revenue growth in year one and over years two to five, the operating margin I think the business ends up at and the year it gets there, how much capital it has to spend to buy that growth, the tax rate, and the return it earns on capital in perpetuity. Then the cost of capital: risk-free rate, equity risk premium weighted by where the revenue actually comes from, beta, cost of debt, debt ratio. And the share price, with the date and time I took it.

The valuation sheet runs the ten-year forecast off those inputs, computes a terminal value on the assumption the business eventually grows no faster than the economy, discounts it all back, subtracts debt, adds cash, and divides by the share count. It also carries a grid showing what the value per share becomes under different assumptions about the two drivers that move it most.

Before I trusted this model I valued five companies in both it and Damodaran’s own workbook and reconciled the two answers to the cent, and I re-check it against his whenever he reissues it.

Time horizon

Unless a valuation says otherwise, I am working on a horizon of at least 1 year, and normally much longer — most positions I expect to hold 5 to 10 years, and some I intend to hold for 20 or more. That is how long I think it takes for a gap between price and value to close, and it is how I invest my own money.

I state a minimum rather than a range on purpose: I can tell you this is not a short-term call, but I can’t honestly tell you today which year I’ll sell. Occasionally a company’s situation sets its own clock — a takeover, a drug trial, a refinancing — and where that’s true, the valuation says so and gives a shorter, specific horizon.

When a valuation stops being my view

A valuation is a view on a date, and it says so. It is superseded the next time I value the same company. Published valuations are never edited — if I change my mind, that is a new valuation, and the old one stays up.

You can see every valuation I’ve published in the last twelve months on everything I’ve published, and what I actually own on what I own.

What this is not

It is not a measurement. A valuation is an estimate produced from assumptions I chose, and people choosing different assumptions will get different answers — which is why I publish the assumptions and the model rather than only the conclusion.

It is not advice. I am not licensed or regulated, I don’t know your circumstances, and nothing I publish is a recommendation that you personally buy or sell anything. For more, see disclosure.