Do you use a single valuation ratio (only price to earnings for example) when selecting companies for your portfolio?

If so you may want to reconsider.

The problem if you use only one valuation ratio is that it can, depending on market situation, lead to you significantly underperforming the market in spite of the tested long-term outperformance of the ratio.

 

The solution
To overcome this problem what you want to consider using is a valuation factor that combines a few valuation measures.

That is what James O’Shaughnessy tested in the fourth edition of his book What works on Wall Street.

 

Better 10 year performance every time
In the book he shows that stocks selected based on a combination of valuation measures outperformed stocks scoring highest on any single valuation ratio 82% of the time in all 10-year rolling periods he tested between 1964 and 2009.

James called his combination of valuations measures a Value Composite One (VC1) factor and used the following five valuation ratios to calculate it:

  • Price to book
  • Price to sales
  • Earnings before interest, taxes, depreciation and amortization (EBITDA) to Enterprise value (EV)
  • Price to free cash flow
  • Price to earnings

How is it calculated?
To calculate the VC1 factor you have to assign a percentile ranking (1 to 100) to each of the five valuation ratios for each company.

For example if a company has a price to earnings (PE) ratio that is in the lowest 1% of all the companies in the market, it gets a rank of 1 (lower is more undervalued) and if a company has a PE ratio in the highest 1% (it is overvalued) it gets a rank of 100.

Once you have done this for all five valuation ratios you add up all the values for each company and (using this value) rank all the companies in percentiles (from 1 to 100).
Companies that are the most undervalued get VC1 score of 1 those with the worst (most expensive) score get a score of 100.

 

Does it work?
Like you, before I use any strategy to select investment, I want to make sure it works.

So I asked my good friend Philip Vanstraceele to help me and we tested the VC1 investment strategy on companies in Europe for the 12 years between June 2001 and June 2012.

Over this period if you bought only non-financial companies (in Europe, the UK, Switzerland and the Nordic countries) with a VC1 value of less than 20 you would have earned a respectable 12,4% per year or 383,8% in total.

This compares very favourably to the 4,9% per year or 73% of all the companies in the database returned over the same period.

We also tested what would have happened if you only bought companies with a VC1 value of more than 80 (in other words overvalued companies).

Over the same 12 year period from June 2001 to June 2012 you would have lost 4.4% per year for a total loss of 40,7% over 12 years.

Here are the results in table form:

Value Composite One > 80 (Worst stocks) Value Composite One < 20 (Best stocks)
cumul. cumul.
Rebalance Date Return No. Comp. 100.0% Rebalance Date Return No. Comp. 100.0%
7/6/2001-7/6/2002 -39.12% 520 60.9% 7/6/2001-7/6/2002 2.44% 513 102.4%
7/6/2002-7/6/2003 -14.67% 500 51.9% 7/6/2002-7/6/2003 -0.59% 494 101.8%
7/6/2003-7/6/2004 32.41% 489 68.8% 7/6/2003-7/6/2004 41.47% 475 144.1%
7/6/2004-7/6/2005 17.44% 539 80.8% 7/6/2004-7/6/2005 35.56% 513 195.3%
7/6/2005-7/6/2006 23.38% 560 99.7% 7/6/2005-7/6/2006 43.91% 551 281.0%
7/6/2006-7/6/2007 23.52% 640 123.1% 7/6/2006-7/6/2007 43.01% 627 401.9%
7/6/2007-7/6/2008 -29.17% 740 87.2% 7/6/2007-7/6/2008 -21.35% 722 316.1%
7/6/2008-7/6/2009 -31.17% 680 60.0% 7/6/2008-7/6/2009 -25.16% 665 236.6%
7/6/2009-7/6/2010 6.14% 580 63.7% 7/6/2009-7/6/2010 35.43% 570 320.4%
7/6/2010-7/6/2011 10.66% 600 70.5% 7/6/2010-7/6/2011 21.73% 589 390.1%
7/6/2011-7/6/2012 -18.77% 620 57.3% 7/6/2011-7/6/2012 -17.28% 599 322.7%
7/6/2012-1/18/2013 3.50% 558 59.3% 7/6/2012-1/18/2013 18.95% 532 383.8%
CAGR -4.43% CAGR 12.36%

 

So you can see the VC1 value can help your investment returns substantially.

 

It’s all done for you
As you have seen it’s quite a lot of work to calculate a VC1 score. I also don’t do it myself and use a screener service I have built.

You can find out about the stock screener here: www.quant-investing.com/screener

 

You don’t have to buy 500 companies
What I am sure you also saw in the table above is that the tested portfolios each year contained an average of about 500 companies.

It is of course not feasible for you to do and that is why I suggest you combine the VC1 score with another indicator such as momentum, F-score or shareholder yield.

Not only will this give you a lot less companies to invest in but your returns are also very likely to be substantially higher as we discovered in the 2012 research study titled:

QuanInv-3D_smallQuantitative Value Investing In Europe: What Works for Achieving Alpha

Profitable investing