A Rule of Thumb for Terminal Valuation Multiples

The terminal TEV/EBITDA multiple should be 12.0 + 0.6 x (current multiple - 12.0)
 

By: Greg Obenshain

One of the cruel realities of investing is that growth does not persist. Companies with above-market growth rates are no more likely than chance to have above-market growth rates in future years. The exceptions to this rule are easy to identify; they loom large in our consciousness simply because they become the largest companies. But any rigorous study of growth persistence will reveal that current growth is a terrible predictor of future growth, a topic we covered in detail in our article Persistence of Growth which replicated the findings of the 2001 paper The Level and Persistence of Growth Rates. 

This has important implications, not least of which is multiple mean reversion. In our article Fair Multiples, we showed that, while growth is the most important determinant of current multiples, it is not a significant predictor of future multiples. As a result, multiples mean revert. This multiple mean reversion happens both for high multiple companies, which see their multiples fall, and low multiple companies, which see their multiples rise.

The combination of overextrapolation of growth trends and the mean reversion of multiples is, in our estimation, the primary explanation for why value as a strategy works and growth does not. But if, in the back of your mind, you cannot shake the feeling that some companies deserve a higher multiple than others and that these companies have a moat that persists, you are correct. While growth does not persist, profitability metrics do, and these profitability metrics are statistically significant drivers of future multiples. As a result, while multiples mean revert because growth does not persist, they do not fully mean revert because profitability does persist. In our opinion, a high-quality company is likely to retain some sort of premium multiple in the future.

All of this has important implications for an analyst building discounted cash flow models that require an assumption about the terminal multiple. Using the current multiple is clearly wrong because multiples mean revert, but using an average market multiple is also wrong because multiples do not fully revert to the market average. So, to help, we present a simple rule of thumb for the terminal TEV/EBITDA multiple:

The terminal TEV / EBITDA multiple is 12.0 +  0.6 x (Current Multiple – 12.0)

In short, TEV multiples should revert toward 12.0x, and an analyst should assume that a company retains 60% of its premium or discount to that multiple.

So where did we come up with these numbers? We analyzed a monthly historical dataset of the US equities with over $10 billion in market cap and positive EBITDA starting in 1994. So this is for profitable large-cap companies. The monthly median TEV/EBITDA of the companies is shown below with the rolling cumulative average overlayed.

Figure 1: Median Monthly TEV/EBITDA and Cumulative Rolling Avg. TEV/EBITDA

Source: Capital IQ, Verdad Analysis

The current cumulative average TEV/EBITDA is 12.3x and has been in the 11.0x – 13.0x range for the past 20 years.  

For each company, each month, we calculated the simple difference between its TEV/EBITDA multiple and the cumulative average TEV/EBITDA multiple as of that month. We then ran a regression from December 1996 to September 2020 to predict the three-year forward TEV/EBITDA multiple (the last forward TEV/EBITDA observation is the current observation). The results of that regression give us the following estimated formula:

3Y FWD TEV/EBITDA Multiple = 11.72 +  0.61 x (TEV Multiple – Cumulative Avg. Median Multiple)

That’s it. We then called it 12.0 and 0.6 to make it simple. Nothing more sophisticated than that, but the results are intuitive and pass Stephen Colbert’s Truthiness test, which is nice. For those of you interested in the details, when regressions are run monthly (to avoid overlapping observations), the average monthly coefficient t-stat is 39.4, the average monthly independent variable t-stat is 11.8 and the average monthly r-squared is 0.34.

But why didn’t we include profitability in the equation if profitability is important for multiples and persists? The answer is that we could, but, in our opinion, it creates significant calculation complexity for the analyst while adding only modest benefit. The rule of thumb gets you most of the way. Below we show the equation with profitability metrics added.

3Y FWD TEV/EBITDA multiple =
               11.73 +
               0.58 x (TEV Multiple - Cumulative Avg. Median Multiple) +
               2.58 x (GP/Sales - Median GP/Sales for the Month) +
               0.51 x (FCF/EBITDA – Median FCF/EBITDA for the Month)

Note that, for this analysis, free cash flow is defined as cash from operations less capex. In practical terms, a company with GP/sales of 60% in a month when the median GP/sales is 40% would have a forward multiple 0.5x higher than the baseline estimate from our rule of thumb. A company with FCF/EBITDA of 60% in a month when the median FCF/EBITDA is 40% would have a forward multiple 0.1x higher than the baseline estimate. So it could be worth considering GP/sales as it has at least some impact. The median GP/sales in the dataset is 42%, and the median FCF/EBITDA is 49% for those who want to add some complexity, but the rule of thumb will get you most of the way to an answer.

Finally, we used three-year forward TEV/EBITDA in our regression. But what if an ambitious analyst is forecasting five years out? Does the terminal multiple change much at five years? Fortunately, the answer is no. The constant drops slightly to 11.6 from 11.7 and the coefficient on TEV/EBITDA less the cumulative average median multiple drops to 0.47 from 0.61. So slightly more mean reversion, but not significantly more. And we’d generally argue that you have already gone well into false precision at the five-year point anyway, so this change is the least of your problems.

And we’d also argue that using the exact numbers from our regression also represents false precision. The important point is that an analyst can use a consistent valuation approach that is grounded in what we know to be true: that multiples mean revert, as growth does not persist, but do not fully mean revert because company quality is sticky.