Thinking about future returns and valuations again

Update on last years post and introduction of my IRR models

Last year, I wrote a post about my approach to valuing companies and using discount rates. A year has passed and I have developed my investment process. Let’s talk about future returns again.

In the previous article, I discussed using inverse DCF Models with a 10% growth rate for most investments. Inverse DCF Models return the required growth rate to achieve a return of the discount rate (i.e., a discount rate of 10% equates to 10% annualized returns). The advantage is that you don’t need a lot of assumptions, making it a bit more subjective than regular DCF models. However, the big problem with DCFs is the perpetual growth rate. After year 10 a company is expected to grow by the assumed GDP growth rate of 3% (2% inflation + a bit of real growth). This is unrealistic for many high-quality businesses that will most likely slow down their growth but probably have a longer time frame of above-average growth (say mid-single digits). That makes a big difference and makes great businesses seem overvalued in many cases. I continue to do inverse DCFs with 10-15% discount rates based on the cyclicality of the business, but I now also incorporate a 5-year IRR model in my valuation process.

IRR models

First, let's consider how a stock’s returns can be generated (below is an article I wrote with my friends at StockOpine, which discusses a similar topic).

Why does Margin expansion often lead to multiple expansion?

Over the long term, the main component is growth. That’s why sustainable revenue growth is very important. Revenue must translate into profits, so we must also see how the business's margins will develop. The company can also reduce its outstanding shares by repurchasing and paying cash dividends. Those are the four factors a company can control to generate returns for its shareholders.

There is one factor a company can’t control: valuation multiples. The market decides how much a company is valued, no matter its fundamental development. All of these factors can have a huge impact on a company's expected returns. Revenue growth should always be the starting point, with the other factors used as multipliers.

Below is a screenshot from my recent ASML update. My IRR model uses three cases with a pessimistic, standard and optimistic case. Of course, this takes more variables and guesstimates than an inverse DCF model, but I think it is also a better indicator of future returns. It’s tough to project out ten years from now, so reducing the time frame to five years and adding an exit multiple instead of a perpetual growth rate helps make the valuation more realistic. For margins and exit multiple, we also need the current multiple to calculate the difference between them and use them as multipliers for revenue growth. Share count changes are also used as a multiple. Dividends are included at 75% of the value to account for taxation (this should be adjusted based on your local dividend taxation and account for foreign dividend taxes).

The danger of this approach is to be too optimistic and overstate one of the variables. We shouldn’t lowball, especially in the optimistic case, but it’s a fine line. While margins and return of capital (buybacks and dividends) are somewhat predictable, the exit multiple isn’t. Here, it’s essential to be cautious and realistic. It wouldn’t make sense to expect an EV/EBITDA multiple of 5 or so for ASML, a monopoly in an industry with growing tailwinds. However, we shouldn’t extrapolate high valuations into the future. I like to stay conservative and not rely on (continued) high multiples and often model in a declining multiple. As long as the business continues to grow, that’s fine.