The Bloodhound System Exchange

As we have noted in previous blog posts, Apple may have lost its shine, but it still remains the largest capitalization stock on the U.S. exchanges – retaking the lead from Exxon earlier this month. It also didn’t prevent it from grabbing the top spot in last week’s 15th annual Barron’s 500 – a ranking of the 500 largest publicly traded companies in the U.S. and Canada. It was 2nd last year when the stock was near $580, so I’m not too sure what it says about the predictive power of the list (although it diid make a run to $700 shortly thereafter). Laggard J.C. Penny took 491st last year, and nabbed place 500 this year – so it might say somethin.’

“The Barron’s 500 is a unique ranking of the largest publicly traded companies in the U.S. and Canada, as measured by total sales in the latest fiscal year. It is prepared by HOLT, a unit of Credit Suisse, which compares companies on the basis of three equally weighted metrics: median three-year cash-flow-based return on investment, the one-year change in that measure relative to the three-year median, and adjusted sales growth in the latest fiscal year.”

The entire current list can be found on their site.

Although nice to be named a “good company,” we are focused on the predictive power. The top 10 last year were:

with the bottom:

Considering three year cash flow is one third of the grade, I’m a little surprised the list bounces around as much as it does. Western Digital ranked 447th last year, only to rebound to third this year.

Barron’s utilized the Credit Suisse HOLT system of CFROI (Cash Flow Return on Investment) as is basis of rank. CFROI is a modified measure of ROA (Return on Assets). Return on assets can be mathematically split into profit margin times asset turnover. The HOLT system substitutes accounting profit with operating cash flow, and gross assets with gross investment. However, a modest flaw in the HOLT system is it uses a forward looking estimates and a number of assumptions. Most notably, HOLT factors in an “objective” mean reversion into what it calls a company’s corporate life cycle. Therefore, one needs to accurately predict future cash flows and objectively acknowledge where a company sits of the life cycle chart (growth, fade, mature…). Results are comparable intra-industry, but become more difficult to compare across industries.

My favorite Barron’s editor, Jackie Doherty, takes a look if there is an ability to find the best bargain stocks among those 500. They rank the stocks based on P/E and track the performance of the cheapest 30. “The results were so impressive that we’re repeating the exercise this year. On average, the 30 stocks rose by 42% in the 12 months ended April 26, far outpacing the Standard & Poor’s 500, which returned 15.6%.” However, that’s where I get a little confused. On one hand, they laud themselves for placing J.C. Penny in the bottom 1% last year (and dead last this), yet on the other hand, create a strategy in which “companies with strong financial performance” no matter where they are on the list. Three of the five cheapest stocks are in the bottom third of the list, and one of the names on the list, Continental United (UAL) is 23 places out of last. By their definition, UAL isn’t a strong financial performer, in fact it got grades of “D” and “F”s for financial performance. Rather, it’s just a big, “cheap” company. Thirteen names on this year’s list (just shy of half) were on last year’s list too – meaning they stayed cheap.

We took a look at it a different way. First we used Bloodhound’s screening tool to identify the cut off level for the 500 largest capitalization names listed on the NYSE, NASDAQ and OTC exchanges. The current Mendoza line is approximately $6.3 billion, equal to where is was in 2007. However, the in-between years, and the years prior to 2007, the cut off was significantly lower. As such, our current candidate pool will be larger than 500 names as we will determine our view with a little cushion by identifying companies with market caps greater than $5bn. As of today, 593 companies fit that bill. In comparison, there were 468 such names in 2005.

We ranked the universe on our own calculation of cash flow return, “CFROA.” We looked at EBITDA less interest, thereby ignoring changes in working capital which can work against a business when it’s growing and improve the cash flow situation when business is actually deteriorating. To get a return figure, we divide the cash flows by total assets. Since Bloodhound has the ability to screen historically, we set the screen date to closely match that of the Barron’s 500 – May 1, 2012. Of the largest 518 market caps, Lorillad (LO) topped the list with a 56% CFROA. Barron’s 500 top contender at that time, CF Industries, ranked 18th. Apple placed 25th. Interestingly, all four of the top ranked companies lost money in the subsequent year.

It immediately begs the question, is cash flow return on investment a predictor of “good companies?” On first glance, the most obvious missing piece is valuation. All factors are accounting-oriented; there is no measurement of market value or price. Fortunately, The Bloodhound System allows users to formulate a strategy and simulate it over 26 years of real-life data, and allows us to study the question.

The average returns of the top 10 and top 50 holdings over the last 26 years is approximately 13%. However, it should be noted that prior to 2004, our buy rule of $5 billion market cap reduces our candidates, capturing a bigger percentage of the largest cap names. When one compares the top 100 names over the last 10 years, the results are not materially different from that of the S&P 500.

The top 10 names have underperformed in each of the last four years, although substantially outperformed the 2008 downturn.

Looking at valuation, as one would expect, P/E ratios are all over the place. P/Es range from 3.4x to 79x. On the other hand, CFROA ranges from 56% by Lorillad to negative by Tesla (TSLA). Rather than ranking the whole universe by P/E as the article does, we looked at only the top CFROA companies, and ranked them by valuation. The median CFROA of the group is 14%. Therefore we ranked the universe by P/E, but created a buy rule that set a hurdle of 15%, to capture the best.

In the last four and a half years, the strategy of picking the ten lowest P/E names among the highest CFROA companies did significantly better than just picking the highest CFROA companies alone. That shouldn’t be too shocking. Amidst a four year bull run, low P/E names are almost by definition poised to outperform. However, that strategy got destroyed in 2008 – down 61.5%. Outside of 2008, the strategy performs quite admirably, but that’s like saying, “besides hitting the iceberg, the Titanic passengers quite enjoyed themselves.”

A strategy of 50-names performs in-line, but also underperforms the S&P in the last three years, and realistically underperforms the strategy ranked solely by CFROA.

I looked at reversing the process: taking the highest CFROA names that had a hurdle of a certain P/E, but none performed materially different. As such, it’s probably nice to be mentioned in the Barron’s 500, but it likely means little for future performance.