# Enterprise Value to EBITDA Ratio Backtest

The EBITDA to enterprise value (EV) ratio is a widely used valuation multiple to assess the relative value of companies. It is calculated by simply taking earnings before interest, taxes, depreciation and amortization (EBITDA) and dividing by enterprise value (EV). Of course, you need to know the definitions of both of those terms to really be able to understand this stock fundamental.

EBITDA is equal to net income with interest, taxes, depreciation, and amortization added back into it. EBITDA is somewhat useful for analyzing and comparing profitability between companies and industries because it removes differences due to tax rates and eliminates the effects of financing and accounting decisions. However, many value investors shun this fundamental due to its past abuse during the tech bubble. Charlie Munger is quoted as saying, “I think that, every time you saw the word EBITDA [earnings], you should substitute the word “bullshit” earnings.”

Enterprise Value takes into account the entire value of a company. It is equal to market capitalization plus debt minority interest and preferred shares, minus total cash and cash equivalents. However, for this backtest I simplified the formula similarly to the way Richard Tortoriello did in Quantitative Strategies for Achieving Alpha and just used market cap plus debt minus cash.

An advantage of the EV/EBITDA multiple is that it is capital structure-neutral, and, therefore, this multiple can be used for direct cross-companies application. The reciprocate multiple EBITDA/EV is used as a measure of cash return on investment. Given that EBITDA ignores capital structure by excluding interest and EV doesn’t discriminate between equity holders or debt holders, also accounting for the entire firm, this ratio provides a complete snapshot on the relative value of entire companies. You can use EV/EBITDA when you want to see the cash-generating power of the entire firm, and you don’t care whether it’s equity or debt financing this cash-generating operation. That’s why EV/EBITDA is used for pure valuation.

I used Portfolio123 to conduct a 10-year backtest of the EBITDA/EV ratio. I filtered out ADRs, non-US companies, companies in the miscellaneous financial services industry category (to mainly filter out closed-end funds), stocks trading below $2, market caps less than $433 million (approximately matching the average cut-off Tortoriello used), and companies that did not have a EBITDA/EV ratio due to missing data. The results are as follows:

Review the full Enterprise Value to EBITDA ratio backtest spreadsheet on Google Docs.

### EV to EBITDA : Rolling 3-Yr Periods Excess Returns vs. Universe

As you can see from the table and charts above, stocks ranked in the highest 20% (1st quintile) of EBITDA/EV produced a CAGR of 11.37% from January 1, 2002 to December 31, 2011. Averaging the annual excess returns versus the universe using four different quarterly start dates, the average excess returns for the first quintile is 4.59%. That is very similar to the 5.3% Richard Tortoriello reported in Quantitative Strategies for Achieving Alpha for the period 1988-2007. Tortoriello found EV/EBITDA to be one of the most consistent (in terms of Sharpe Ratio) factors he tested. In the recent paper, Analyzing Valuation Measures: A Performance Horse-Race Over the Past 40 Years, Wesley Gray and Jack Vogel also found the EBITDA/EV ratio to have been the best performing metric historically. They found that it outperforms many investor favorites such as price-to-earnings, free-cash-flow to total enterprise value, and book-to-market.

Do you use the EV/EBITDA ratio? If not, why not given these results? What other fundamentals would you like to see backtest here? Please leave you responses in the comments section below.

What criterias were exhibited by those in the 1st Quintile?

I’ve been reading around and the consensus is that anywhere from EV/EBITDA of 6-8 is considered “cheap”.. does that ring true?

Good question Tom. The average for the 1st Quintile with a 1st quarter start from 2002 to 2011 is 5.9 and the median is 6.23. The minimum is 0.23 and the maximum is 8.5. It sounds like that consensus of between 6-8 falls right into this 1st quintile.

Thanks for the response!

And are the EV/EBITDA numbers coming from LTM or Forward estimates?

The numbers are TTM.

Great post; glad to read someone uses this strategy in this day of “passive management”. This kind of valuation of stocks needs much more attention!

Personally I prefer Net Cash Flow / Enterprise Value which would include interest expense. I have written an article titled “Best Stock Valuation Calculation to Value Company Shares is ROEV” for anyone interested at:

http://blog.arborinvestmentplanner.com/2011/05/best-stock-valuation-calculation-to-value-company-shares-is-roev

Hi Ken. I’m glad you liked the post. I just did a quick test of your Net Cash Flow/EV.

~~The top quintile underperformed EBITDA/EV from Jan 1, 2002 to 12/31/2011. Net Cash Flow / Enterprise Value return 10.41% annualized versus the 11.37% annualized return for EBITDA/EV over the same period.~~[Edit Jan 18, 2012: I made a mistake formulating Net Cash Flow when I first ran the quick test. The revised test revealed 11.48% annualized return over the past 10 years for Net Cash Flow/Enterprise Value.] Please let me know if you are interested in seeing the full backtest results for Net Cash Flow / Enterprise Value.Awesome. Thank you George. Yes, that would be great! I hope you will submit this post to the Self-Directed Investing For Retirement Carnival on the AAAMP Blog.

Hi Ken,

I goofed the first time I ran the Net Cash Flow / Enterprise Value. I accidentally subtracted depreciation and amortization from net income instead of adding it back to net income to determine “net cash flow”. Now that I’ve corrected the formula for Net Cash Flow / Enterprise Value the top quintile outperforms EBITDA/EV from Jan 1, 2002 to 12/31/2011. Net Cash Flow / Enterprise Value return

11.48%annualized versus the 11.37% annualized return for EBITDA/EV over the same period. Those numbers are very similar, so I will definitely be doing a complete backtest on this factor. The complete backtest will look at 4 different quarterly start dates and it will also generate Sharpe and Sortino ratios.Your awesome George! I appreciate your diligence and desire to get it right and provide your readers some great information. I look forward to your coming research!

How exactly do you guys backtest?

I would like to back test a simple strategy: EV/EBITDA 20%

Is this possible?

Tom – I used StockScreen123 to do the backtest and then I put the results into a spreadsheet to do some additional calculations. It would be very possible to do a back test of EV/EBITDA 20%.

Much appreciated your article, I specifically conduct EV/EBITDA analysis on all potential investments because I find it better than EPS and other more conventional metrics. I prefer to (as best as I can) work off next 12 months projected EBITDA rather than ttm. EBITDA is such a tricky term that can be so different for investments in the same industry. I wrote a brief piece on EBITDA that you and your readers may find interesting at http://www.valuation-rx.com/ebitda.html

Thanks and keep up the good work!

what EV/ebitda criteria were used for the 2nd and 3r quintile?