(Source: Pixabay.com)

Over the last few years, my investing style has evolved to that of quantitative investing, founded in the fundamentals of value investing.

One of the strategies that I use in my portfolios is that of the low enterprise multiple, or low EV/EBIT. I have written about this strategy in a previous article. However, I would like to dig a little deeper into the strategy, particularly from a statistical point of view.

**The Cigar Butts**

Deep value investing is a somewhat vague, open subject in value investing. To me, it is essentially finding stocks selling at deep discounts to their intrinsic value. These types of stocks are often overlooked, discarded, forgotten, misunderstood, or going through a temporary rough patch. They are Buffett's "cigar butts". So we have identified the type of firms that may be deep value, but again, the $64,000 question: what is the intrinsic value?

In quantitative investing, we approach investing at a strategic, portfolio level. We also find strategies that have performed well in the past over long periods and apply the same cycles into the future. How do you go about finding deep value with quantitative investing?

We look no further than Benjamin Graham and Joel Greenblatt. Graham often used low P/E ratios to find cheap companies. As discussed in one of my earlier Magic Formula posts, P/E ratios can be misleading, as they often only use a portion of the financial information available, and thus, only paint part of the picture.

The price-to-earnings ratio is a great metric to get a quick idea of value, is easy to calculate, and is pre-calculated on nearly every finance website. Let's briefly break it down to see how it can be a misleading metric.

**The Price, P**

The price in the ratio refers to the current share price, which is reflecting the value of the market equity for common shares. As an equity investor, you may think that's all you should be concerned about. What is important is how much equity the firm has relative to its entire capital structure.

A firm with $10 billion in equity and $10 billion in debt is very different than a firm with $10 billion in equity and $100 billion in debt. The first firm needs much less debt to achieve the same amount of equity.

This is where the Enterprise Value, or EV, comes in handy, as it looks at the entire capital structure of a firm. The EV considers equity, but also total debt, minority interests in other firms, preferred equity/shares (note that the price component of P/E is the common equity only), and cash. The EV of a firm is considered the price that an acquiring firm would need to pay to acquire.

Cash can be an interesting component here. Some firms that do not require a great deal of assets to operate can be high-margin, cash-rich companies. In the sale of a company, the cash becomes the property of the acquirer, and essentially reduces the purchase price of the firm. This becomes clear when looking at the EV in equation form:

Enterprise Value = Market Equity + Total Debt + Preferred Equity + Minority Interests - Cash and Equity

Looking at EV instead of P, we can more easily compare firms.

**The Earnings, E**

The earnings component of the P/E ratio is the earnings per share, or EPS. This is simply the net income, taken directly from the income statement, divided by the total number of shares outstanding.

Net income can be deceiving, as it includes several non-cash charges, for better or for worse. This can include deprecation and amortization of goodwill.

As "cash flow is king" in business, it is not uncommon for firms to have positive cash flow but negative net income, or worse, positive net income but negative cash flow.

Net income can also include one-time items, both income and expense, which can skew the overall results. Investors should be investing based on the firm's normal operating scenario or conditions. Too many positive "specials" disappoint investors later once they run out. On the other hand, too many negative one-time items can lead to a lift after some positive news breaks.

Firms with significant long-term debt can also claim interest payments as tax-deductible expenses compared to firms with little to none.

Similarly, taxes can skew net income as well.

So what are the alternatives to net income, or EPS?

EBIT, EBITDA, and Operating Income are all variations of net income which do not include as many non-cash charges to net incomes. These also remove any references to interest or taxes, thus making firms in different industries more comparable.

Substituting EV for "Price" and EBIT for "Earnings", P/E becomes **EV/EBIT**.

You may recognize this as the "cheap" component of Joel Greenblatt's Magic Formula. The other half of Greenblatt's strategy is the quality component, or firms with high returns on capital, or ROC.

What happens if we just look at "cheap" companies?

**The Low Enterprise Multiple Strategy**

"Cheap" companies form the basis of the low Enterprise Multiple strategy. This strategy is "buy and hold" for a one-year holding period. At the end of the period, stocks are sold and new stocks passing the criteria are purchased.

As we discussed in the beginning, deep value firms are often the cigar butts and are not considered quality firms, nor do they typically achieve high returns on capital. But that said, they still have "one last puff" of value, and this strategy attempts to find those puffs. Either these firms are at or near the bottom of their cycle and we are betting on their recovery, or they are attractive to other larger firms and it is the acquisition premium we are after.

**The Modifiers**

Once you take out the quality component, you are left with some unknown, obscure companies. When using quality as a component of a strategy, you often screen out the value traps, or those stocks that are cheap for a reason. When the high ROC component is removed, another mechanism to screen out potential value traps should be added.

In quantitative investing, there are several tools we can employ to help us steer clear of these minefields. The factors below are some of the techniques used to screen out potential value traps.

*Altman Z-Score*

We would want to screen out any firms that are at risk of filing for bankruptcy. While there are some single metrics that are sometimes used to check for this such as interest coverage ratio, etc., these are isolated ratios that only get part of the picture.

The Altman Z-Score is a statistical model that tests how probable a firm is to file for bankruptcy, based on the reported financials. This test weighs several factors differently and combines subtests to a single score. Generally, firms scoring greater than 1.81 are less likely to going bankrupt in the near future. The higher the score, the better.

I have written about practical application of the Altman Z-Score in previous articles (here and here).

*Beneish M-Score*

Checks for manipulation of earnings. The threshold is a negative value. Typically, less than -2.22 shows signs of an honest accounting team. Students at Cornell had reportedly used this test, among others, to flag Enron as an earnings manipulator, while analysts at the time were touting the stock.

**The Universe**

Similar to the Magic Formula, EV/EBIT is not well suited to utility companies and financial firms due to their financial reporting requirements, and they are thus excluded from the screen.

Over-the-counter stocks (OTC) with limited liquidity are excluded from the strategy.

Non-US firms trading on the US market (or ADRs, American Depository Receipts) are included in this strategy. These firms will appear foreign to most domestic investors. ADRs are often obscure and less followed; however, in our case, this is a good thing. Less analyst coverage often results in more inefficient pricing of stocks, allowing for more upswing.

We will be looking at two universes: the Large Cap (largest 20% companies by market cap) and Large & Mid Cap (largest 50% of companies by market cap).

**Typical Stocks**

As mentioned, this is a cigar butt strategy. You will see stocks that have been beaten down, are going through a public scandal, have had an extended period of underperformance, have questionable management, industries at the bottom of their cycles, etc. Some investors may not feel comfortable investing in these types of firms for this reason. Recall, however, that this strategy looks at picking up firms at or near the bottom of their trough, and which are preparing to return to mean or their average performance level (both at a business and stock level).

As we will see, there is always volatility in quantitative strategies. As an example, let's look at a few stocks that were screened as candidates on January 1 of this year. (I will be following up with a full review of the strategy since January 1 of this year in a future article.)

Since purchase, a handful of stocks in the Large & Mid Cap universe have soared.

- Argan (NYSEMKT:AGX), Construction & Engineering, +102%
- Alliance Resource Partners (NASDAQ:ARLP), Coal & Consumables, + 106%

Mind you, there have also been some losers to date:

(Source: Portfolio 123 Data)

Being able to look at the big picture is the key to holding on in quantitative investing.

**Variations of Low EV/EBIT**

I will take this opportunity to mention that not all Low EV/EBIT strategies are the same. Any difference in the universe(s) or the "modifiers" used (if any), or differences in market cap cutoffs in the respective universes can make a considerable difference to which stocks are screened and may result in varying performance. This would also apply to any strategy, for that matter. When comparing performance, be sure it is an "apples to apples" comparison.

**The Performance**

To assess this strategy, the sections below contain summaries of various backtest data since January 1999.

**Large Cap US**

Let's take a look to see how this cigar butt investing strategy has done since 1999 for Large Cap companies.

*Summary Statistics*

(Source: Portfolio123 data and author calculations & table)

(Note: I will be writing a future article on the various statistical terms above.)

In terms of return, we enter 20%+ territory in terms of average 1-year arithmetic return. We also see the same 20%+ value in volatility (standard deviation), meaning these higher returns came with higher volatility. The return per unit of volatility (Sharpe & Sortino ratios), however, are considerably higher than the Russell 1000TR benchmark.

Maximum drawdown is more than 4% greater than the benchmark. Still not too bad.

An interesting statistic here is the seasonal standard deviation. It is important to realize that all quantitative strategies have some seasonal element to them, meaning if you buy and hold in January, your results will be different than if you buy and hold in June of every year. Comparing to the benchmark, returns from our strategy can vary wildly, depending on the time of the year you invest. Since 1999, the highest returns have been while investing in January (CAGR of 23.5%), and drop off to a low nearly 50% of January investments in August (CAGR of 13.6%). Something to be aware of, but also note that the seasonal maximum and minimums can change over time. Rebalancing in January may show the highest returns for a few years, but this could change in years to come.

*Base Rates*

Average results over the entire period since 1999 are helpful, but to really understand a strategy, it is important to look at all the rolling periods within the time span. We then calculate the odds of beating the benchmark over these rolling periods, also known as "base rates". This is a key concept as discussed in quantitative investing pioneer James P. O'Shaughnessy's book, "*What Works on Wall Street*".

(BM = Benchmark, Russell 1000 TR)

(Source: Portfolio123 data and author calculations & table)

These base rates are fairly good. In nearly 91% of the 10-year rolling periods, an investor would have beaten the benchmark (Russell 1000TR). Over a shorter 1-year period, an investor would have had nearly a 73% chance of doing so.

When outperformance did occur, it did so with returns in excess of 22% for all of the periods.

*Rolling 5-year Return Trend*

Returns for all strategies ebb and flow over long periods. It is important to understand this, particularly when your strategy seems to be floundering. Chances are the strategy is in a lull, but is likely to return to positive and/or market-beating territory in the future.

The graph below shows the excess return in terms of 5-year CAGR over all the rolling periods studied, from the benchmark.

(Source: Portfolio123 data and author calculations & graph)

After a peak of 65% 5-year CAGR above the benchmark ending period 2005, 5-year returns steadily decreased until the end of 2012. It managed to stay in positive territory until early 2012 to a minimum loss of about 10%, but then one year later, managed to peak at 50%. This is another illustration of how strategies can follow up with higher returns after a longer period of underperformance.

*Maximum and Minimum Compounded Annual Rates*

The table below shows the best and worst rolling period of each of the time spans for the low EV/EBIT strategy for Large & Mid Cap stocks in the US.

(Source: Portfolio123 data and author calculations & table)

This strategy has hit negative territory at minimum returns for 1-, 3-, and 5-year periods, but did enjoy positive minimum returns with longer holding periods, i.e., 7 and 10 years.

While the rates are important, it is also as important to look at an investment in absolute dollars. The CAGR is a rate that, when compounded, will provide a given return. When you look at the dollar value of an investment at a given CAGR over a time period, it can paint a more illustrative picture. For example, comparing the 10-year minimum of our strategy to the 10-year maximum of the benchmark, CAGR is 6.7% and 9.5% respectively (or a 42% difference). In absolute terms, the dollar value of the investments is $19,044 and $24,778 respectively - only a 30% difference. In other words, the **maximum** 10-year performance of our benchmark is only 30% greater than our **minimum** 10-year strategy return.

**The Performance, ****Large &** **Mid ****Cap US**

In many cases (but not all), performance of strategies can be enhanced when looking at more stocks than your garden variety large caps. The theory is that the smaller the stock, the less followed they are and the more inefficiently priced. Let's see if expanding our universe of cigar butts has been of benefit to our performance.

*Summary Statistics*

(Source: Portfolio123 Data and author calculations & table)

Like its Large Cap cousin, the Large & Mid Cap strategy has shown very high returns compared to its benchmark, the Russell 3000TR. The arithmetic average annual return is slightly higher than the large cap version; however, the geometric return (CAGR) is lower, suggesting more volatility. This is apparent when you see the standard deviation of 24%. It is considered a highly volatile strategy. Also with the higher return, the Sharpe and Sortino ratios are lower than in the Large Cap strategy, indicating that an investor is taking on more volatility per unit of return. Is the slightly higher average return worth the added volatility? This is for each investor to decide for himself or herself.

Of special note is the seasonal volatility of a whopping 441%. The time of year at which this strategy is rebalanced can have a dramatic effect on returns. January has shown a CAGR of nearly 26% since 1999, while investing and re-balancing in September shows a CAGR of less than half of that, or 12.1%. This is not to stay an investor should have invested and rebalanced only in January; the months that show the maximum returns over time can change.

*Base Rates*

Once again we take a look at base rates to get a comprehensive picture of the returns over all possible rolling periods.

(Source: Portfolio123 Data and author calculations & table)

The volatility of this strategy is evident in the lower odds of beating the benchmark over shorter time periods. The Large Cap version of this strategy has beaten its benchmark nearly 73% of the time over 1-year periods, however the Large & Mid Cap version lags by 7% to just under 66%.

Similarly the rolling 3-, 5-, and 7-year time periods lagged the Large Cap version of the strategy in terms of beating the benchmark. What is interesting, however, is that if any investor held for 10 years, he/she would have had nearly a 93% chance of beating the benchmark of the period - a few percentage points higher than the Large Cap strategy.

*Maximum and Minimum Compounded Annual Rates*

The table below shows the best and worst rolling periods for each of the time spans for the low EV/EBIT strategy, Large Cap & Mid Cap US.

(Source: Portfolio123 and author calculations & table)

From the other statistics above, we have seen that this strategy has relatively more volatility than its Large Cap cousin (and other strategies as well, for that matter). This is particularly clear now when you look at the minimum 7- and 10-year annual compounded returns: they are negative! While base rates for the 10-year period are quite good (93%, see above), if you did not beat the benchmark, there was a chance that you would have actually lost money over the full 10-year period (and 7-year period). Of particular note is that the ending dates for the negative 7- and 10-year periods are fairly recent: June 2014 and May 2016 respectively. On the other hand, if you were on the opposite ends of the spectrum, the 7- and 10-year periods ending December 2006 and November 2009 would have resulted in the highest CAGR over the periods.

*Rolling 5-year Return Trend*

Many quantitative value strategies saw peak 5-year CAGR difference fron their benchmark in the early 2000s. The low EV/EBIT strategy for the Large & Mid Cap universe has actually shown 5-year peaks getting higher since the early 2000s, peaking near the end of 2013. Typically after a peak, the performance has dropped off, as is the case here, showing underperformance right up until today to a rather humbling level of 30% less than the benchmark ending May 2016.

(Source: Portfolio123 data and author calculations & graph)

**Small Caps**

One of the areas we like to look at for added returns is the small cap universe. Oftentimes, these are the most inefficiently priced and can reward investors with high returns, with or without added volatility. The drawback is the liquidity consideration; many investors with large portfolio values cannot invest a meaningful amount of money into these types of stocks, as it would send the security price soaring.

We have researched the low EV/EBIT strategy for stocks greater than 25% of the smallest market cap. We generally found that the strategy had slightly less returns than the Large and Large/Mid cap versions; however, the volatility was proportionately higher. For that return, we would practically see the Large and Large & Mid Cap versions as being the best reference point for these strategies.

**To Conclude**

There is no doubt that the Earnings Yield portion of Greenblatt's original Magic Formula (with some important modifications) has earned high returns. The Large Cap strategy has done so with some level of volatility, whereas the Large & Mid Cap version takes on more volatility for some added return. While return is the end goal for investors, it is critical that quantitative investors be aware of the volatility. If you cannot stomach the ups and downs and end up throwing in the towel, you may lose precious investing time.

Remember that low EV/EBIT type stocks are cigar butts. Obscure, unknown, and possibly originating in countries outside of the US. If that doesn't bother you, then this strategy could be a fit. If holding on to household names is what is important to you, then this may not be the strategy for you.

A final note: As we are using statistical data to help us make informed decisions, there is never any guarantee that past performance is indicative of future performance.

Follow me on Seeking Alpha for more articles on Quantitative Value Investing!

**Disclosure:** I am/we are long ARLP, AGX.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.