Analysis of MFI Results By Country

Analysis of MFI Results By Country

It was pointed out the the actual book on magic formula investing says to exclude foreign stocks.  However, if you look at the screens, it is not that unusual to see Canadian, British, Irish, Israeli and Chinese stocks.  I am not absolutely sure why that is, if I were to guess it is they include companies that file a 10-K and exclude those that file a 20-F.

I know that certain Chinese companies during the era of rampant reverse mergers and ensuing fraud really torpedoed some MFI tracking results and frankly cost me some real money as well.  That was one of the reasons I quit in 2011.

I was asked if I could analyze MFI tracking results by country.  That is not so easy to do accurately or quickly.  One issue is I have have data back to 2006 and many of those tickers are not active any longer (bought or changed). A second issue is even an automated approach would not get what we wanted - like CBEH.  This company actually has China in it's name, but if you look it up in SEC.gov it says it is registered in Delaware and has a Florida address.  So is it Chinese or USA?

So I used automated results as much as I could, but then tried to override those that I knew were wrong, like CBEH.  I also made educated guesses on the many companies that no longer had a valid ticker.  I am sure I have quite a few wrong.  But I also believe that directionally, my results should be valid.  I am sure I have 95% in the right bucket (and USA is far and away the most common bucket as you'll see).

Overall Tables

First let us look at overall results to remind us where we are.  Here are MFI Results by purchase year:

Year	Average Change
2006	13.9%
2007	-20.5%
2008	-8.2%
2009	39.3%
2010	8.0%
2011	-0.8%
2012	27.2%
2013	24.5%
2014	3.8%
2015	-3.8%
2016	17.1%
2017	7.6%
Grand Total	8.9%

Now let us look at US versus non US.

Year	Non USA	USA
2006	12.6%	14.0%
2007	-7.9%	-22.5%
2008	15.8%	-11.7%
2009	31.9%	40.5%
2010	-19.3%	11.7%
2011	-24.5%	-0.1%
2012	9.3%	28.3%
2013	14.6%	25.0%
2014	-4.4%	4.2%
2015	3.0%	-4.2%
2016	-7.4%	17.7%
2017	2.2%	7.7%
Grand Total	5.3%	9.2%

So you can see non-USA has clearly underperformed.  Of course you can see it doesn't move the overall needle that much (8.9% to 9.2%) as 93% of the stocks (by my quick work) are USA.  You can see what I believe to be the reverse merger issue, rampant in 2010 and 2011 as non USA are -19.3% and -24.5%.

By Country

Year	Average Change	Count
Bermuda	56.4%	1
Canada	1.8%	192
Cayman Islands	32.7%	48
China	-28.5%	96
Hong Kong	9.7%	23
Ireland	21.7%	46
Israel	20.1%	36
Netherlands	60.0%	12
Russia	152.5%	9
Singapore	-17.2%	17
South Africa	-14.1%	14
Switzerland	16.6%	18
Taiwan	-52.5%	7
United Kingdom	8.8%	7
USA	9.2%	6,641
Grand Total	8.9%	7,167

So 96 stock years from China averaging -28.5%.  All 7 of the Taiwan stock years are the same company (GIGM).  Canada has also been pretty mediocre (a lot of mining stocks).

Here is what results look like if we just exclude China:

Year	Average Change
2006	13.9%
2007	-20.0%
2008	-8.4%
2009	41.9%
2010	11.5%
2011	-0.5%
2012	27.6%
2013	24.7%
2014	3.8%
2015	-3.8%
2016	17.1%
2017	7.6%
Grand Total	9.5%

So it takes the results from 8.9% to 9.5%. A pretty good gain.  But I do worry that we're fighting yesterday's war.  Here are China results by purchase year:

Year	Average Change	Count
2007	-71.7%	6
2008	-1.6%	15
2009	-16.3%	26
2010	-39.1%	38
2011	-47.1%	4
2012	-17.3%	5
2013	-48.6%	2
Grand Total	-28.5%	96

So while Chinese results suck, they have not really been on the screen recently.

Final Thought

One interesting takeaway is that by removing China, MFI is much closer to the Russell 3000 benchmark.  For purchase years 2017 and prior, average return per stock year was 8.9% (MFI) to 9.8% (Russell 3000), By removing China (via an inexact approach) it tightens significantly to 9.5% to 9.8%.  But US only is just 9.2%.

I don't know.  I think including/excluding countries is not critical.  I do think it is critical to try and identify companies that have some sort of Fraud.  That was the real problem here.