Valuing Companies for Investment on the Stock Market

Same principles apply to publicly traded companies. Combine comps/precedent transactions and a DCF to get your target value. You can usually just get away with building a DCF model.

If you're investing, then you should be treating the purchase of a stock as fractional ownership in a business [this is one of the main principles of value investing], in which case the thought process and valuation of a piece of a business is largely the same as the valuation of the entire pie. There are times when this isn't true, however, as sometimes there's a clear benefit to owning the entire company and controlling its future. As a contrived example, perhaps a company is trading for $60, which is the right valuation for the operations, but you find that the company's assets on a standalone basis are worth $70. Ignoring transaction costs equity investors would be better off liquidating the business than keeping it running. If you own the whole company, you can easily make this decision and earn a $10 profit; if you own 1/1000th of a company and other shareholders disagree with your assessment, then you're out of luck. In an efficient market, however, others will realize this fact, drive up the value near $70, and push for liquidation [or a big investor would come in and do it] so you could still come out on top.

But yes, I think most good valuations take the viewpoint of valuing the whole pie and then afterwards breaking that down and valuing the various pieces. Determining an appropriate enterprise value, whether through cash flow analysis or multiples, is typically done rather than measuring equity value directly because EV is independent of capital structure and excess assets; after determining EV, then you can add in the excess assets of the firm you're valuing and take into account its capital structure in order to get the value of equity.

Multiples can be useful but I think you need to be very careful with them and approach the process with as much respect as if you were building a ground-up DCF model. It's very, very easy to pull up 4-5 comps mentioned in the 10-k, measure their EV/EBITDA ratios from last year and average the multiples, and then apply that to your target firm's EBITDA to get an EV number. You can do it in five minutes or a computer could easily do it in seconds. But it's also a useless process if you do it half assed - unless you get really lucky and just happened to get perfect comps, then it will takes a lot of work to do a proper valuation based on multiples. Probably as much work as a good DCF model. And then at the end of the day, with multiples you're really basing your valuation on the market being right about the value of other firms in the industry.

On which multiples - I'd say P/E is almost useless. It doesn't take into account capital structure, taxes can vary widely year to year, etc. EV/EBITDA is widely used and cap structure, tax independent but needs to be used with caution if the two companies have different capex needs. In general w/ multiples just make sure the numerator matches the denominator; i.e. if the numerator is equity value, the denominator should be earnings/cash flow/etc flowing to equity holders. And the other guideline is that as you move up the income statement (i.e., from EV/EBIT to EV/EBITDA to EV/EBITDAR to EV/Revenue, etc) you eliminate a lot of the "lumpiness" and one-time items, but your valuation starts to lose meaning. If you're using EV/Revenue to value two companies, that only makes sense if you expect them to have very similar cost structures going forward, which isn't likely.

In theory, cash flow multiples would be perfect because as an investor you really care about free cash flow, and earnings is just a proxy; but FCF is usually going to be very lumpy which makes it difficult to utilize as a base for multiples.