In a DCF, what is terminal value?

I’ll start this post with a jarring statement (call it the TL;DR of this post) : o

In perpetuity, all companies in the same industry should be worth the same amount

Why do I say that?

Well, in valuation, we typically assume that all companies will grow at the same rate once they reach their normalized state. This rate is typically around 2.0% so it matches the targeted long-term inflation / targeted GDP growth for the United States. The reason is, that if you grew faster than the US economy forever, then mathematically, that company would take over the entire GDP. In other words, a company cannot outgrow the economy forever, and thus, it'll converge to this long-term growth rate. Of course, I'll caveat to say this discussion should all be $ adjusted (i.e. proportional value).

However, many times you will see 3.0% or other higher growth rates used to account for international opportunities or the expected industry growth above GDP for at least the longer middle term. Further, some people may not agree on long-term inflation and GDP expectations, and thus use a different estimate. But let's assume we could all agree on this growth.

So… If growth is the same for all these companies, then what else could affect the terminal value?

Margin and other cash-flow adjustments will also affect a model's terminal value

However, we tend to benchmark EBITDA margin, CAPEX, and working capital profiles to the industry averages, creating a long-term trend that matches long-term industry expectations. Yes, there are exceptions, but it’s important to realize that mature companies will tend to converge around the same metrics if they serve similar markets.

Finally, we have the discount rate. While we use a single discount rate in our DCF models, companies go through different stages of growth and risk profiles. Better models would break down these stages and adopt different discount rates for each stage. In practice, we don’t do that, so the discount rate will be a major differentiator and the reason why in reality not all terminal values are the same. However, if we did model more accurately, then all perpetuity calculations would get a very similar, if not the same discount rate.

Now the math:

A growing perpetuity is a stream of cash flows that lasts forever and grows forever

When we think about terminal value, we’re saying that a company is expected to return a growing dollar amount for the foreseeable future. Do you see the parallel? The model we use - the Gordon Growth Model - is the dividend discount model where [ X*(1+r) / (r-g) ] gives you the terminal value. As a final step, this value is discounted back to present value using mid-period discounting.

Another way to calculate terminal value is to use a terminal multiple [e.g. M * EBITDA ]. While the discounting is adjusted slightly (this is assumed to be an end-of-period cash flow), the multiple is a reflection of this perpetually growing return. You pay a multiple of revenue, (or a multiple of EBITDA), as a proxy for the cash you expect to receive forever, just like in a traditional market approach. These multiples can be benchmarked against public companies in mature stages, given that these companies are likely to match the growth and margin profile expected for your subject company at this stage.

While there are other more complicated models for situations in which the company has not reached a normalized state, the multiples method can take care of these situations.

Remember to choose a multiple that is commensurate with the company’s financial metrics at that point in time!

Back to practice:

The reason we are so interested in terminal value is that this tends to be the LARGEST component of our total DCF value. Despite having a major impact on value, we tend to plug and chug during this calculation and move on.

The reason?

In perpetuity, all companies with that same profile are worth exactly the same (*$ revenue adjusted).

If companies in an industry trend towards the same metrics, then the terminal value will be very similar in all your models - pre-discounting.

If we take this one step further, we can see that discount rates tend to coincide for companies in similar stages, with similar growth profiles, and similar maturing (i.e. similar number of years to maturity). Thus, our terminal values should in practice have low dispersion, and discount back to very similar present values, especially when using the same number of years in our discrete DCF period (e.g. 10 years).

What's your experience with terminal value? Do you spend a lot of time on these assumptions or is it an afterthought?

[email protected]

@JordiPujolCFA