How do you pick a multiple in DCF terminal value calculation (multiples method)

Thank you!

I've never seen levered free cash flow DCF in any valuation model, not sure why you would take this approach versus traditional unlevered fcf and take into account the impact of leverage in the discount rate. If you want to piece meal the equity and the debt, you could do the APV valuation DCF methodology which I think is more accepted on the street, well at least taught in b-school. And yes, you would have to use a p/e ratio that is sensitive to the cap structure of your comps, further diluting your analysis unless you had a sold p/e. You can also use the GGM for equity since it's essentially what was used to value equity to begin with in perpetuity.

You can do a levered DCF and still use an EBITDA multiple at terminus, you just have to reduce net debt at terminus and then discount that at your COE for PV of TV.

Typically if your comps don't have similar enough (enough being the key word here) fundamental business risk, then you must not have selected the right set of comps. One expects companies in the same industry with a similar product mix to be affected by similar risk factors and thus have an unlevered (asset) beta in a certain range. What kind of company are you valuing? Does it have a highly unique, niche product? If so you might have to look a little harder (or get more creative) in your comp set selection.

But to answer your original question, yes I think it would be a bit odd theoretically speaking to use the average EBITDA multiple for terminal value if your claim is that your company has fundamentally different business risk from its comps. Remember that multiples are really factors of 3 things - growth, profitability and risk, and so ideally speaking a good comp set should be able to tell you something about both valuation (multiples) and risk (betas).

Now note that business risk is not the same as overall risk of the company. Overall risk is a function of the business risk AND the capital structure (that is why you must relever the target company after you unlever its comps)

Which ever makes the deal look best.

Comps?

Whats your question? The terminal value should just the perpetuity of cash flows in the stable growth period.

Ideally, you'd be looking at how market is valuing comparable companies and/or any precedent transactions in the space

In actuality, as an analyst you'll stick in a number your VP or MD gives you

Everything above is correct. Typically we'll put together comps (prev. trans and market) for the targets peer group, calc the mean, median, max and min to arrive a multiples (usually just use the mean, but always defer to what your senior guys want to use or just use your brain / common sense in a pinch). Usually whatever multiple you take away from that could be you're base multiple for your sensitivities.

The multiples are for the 5th or in some cases the 10th year, so do we take the median EV/EBITDA multiple of comparables as calculated from today figures? Or some expected multiple in 5 years?

At my bank, we use Capital IQ to develop a list of precedent transactions and public company comparables. From there, we use excel to spread key financials (ie EV/Revenue, EV/EBITDA, EV/EBIT, P/E, etc.). We also set up a Min, Max, Median, Mean box that is calculated based all of the multiples the provided. Typically, you then "put a finger in the air" based on the Median/Mean #s and the most comparable companies/transactions and say "OK, the mean is 1.5x EBITDA multiple, but the best comparables are at around 2.0xE EBITDA multiple...OK, lets put down 1.8x!) -- Then you go to the client, show them your analysis and, since they know the industry, will say WOW those #s are HUGE and you say "OK, we will use lower multiples" and they say good! "then you say OK, lets use 1.2X instead! WOOHOO!

net net working capital X 5

You both forgot about the core driver of terminal value in DCF methodology: note from MD with magic number.

You're still getting to the same place (EV), so...

Maybe I'm missing something here, but the EBITDA multiple gives you a value estimate that factors in that the profitability metric is before taxes. In other words, if the average/median of other market deals is 8.0x, that means the EV was 8.0x pre-tax, pre-interest, pre-depreciation, and pre-amortization earnings for other companies.

I suppose you could take out taxes, but then the starting point for comparable multiples you would use as a baseline would technically be EBIDA (never heard of using that though).

What you are referring to is a "foward EV/EBITDA" multiple which is pretty much never used because it depends entirely on your own inputs. So no, don't use your projected EBITDA for your EV/EBITDA multiple...Use the LTM EBITDA for comparability's sake when you do spread your comps.