Hi everyone,

I think that the article is spot on with the errors and if anyone has built a valuation model from Copeland's Valuation book should be aware of most those issues. Errors 1, 2, & 4 are joined at the hip. I would argue that another egregious error not mentioned but deserves top honors is the error of not normalizing earnings. I think Philip does a good job of making sure the cyclical nature of the business is taken into account. This is also one of Graham's big warnings.

Anyway, I put together a spreadsheet to try and show this with Jim's FORM analysis. I wish the analysis was put on the open boards so everyone could take advantage of it :((Or link to one that is available like the Garmin analysis - but I digress).

So here we go with some data points:

Year                    2006     2007     2008     2009     2010            Terminal     Terminal (ALT)
NOPAT                   51,715   69,297   92,859   124,431  166,737         171,739      171,739
FCF                     29,705   43,332   57,318   76,806   102,921         106,009      161,926
IC Investment           22,010   25,965   35,541   47,625   63,816          65,730       9,814
                        
Beginning IC            132,487  154,497  180,462  216,003  263,628         327,444      327,444
Ending IC               154,497  180,462  216,003  263,628  327,444         393,174      337,258
                        
ROIC (average beg+end)  36.04%   41.38%   46.84%   51.89%   56.42%          43.68%       50.92%
ROIC (beginning)        39.03%   44.85%   51.46%   57.61%   63.25%          52.45%       52.45%
                        
IC Growth               16.61%   16.81%   19.69%   22.05%   24.21%          20.07%       3.00%
IC Reinvestment         42.56%   37.47%   38.27%   38.27%   38.27%          38.27%       5.71%
 
                                                            Terminal Value  1,247,160    1,905,007

And the calculation of invested capital (my calculation) which is needed to generate the ROIC numbers:

Beginning IC Calculation
 
Operating Cash @ 5% Rev          11,874
Excess Cash                      199,734
 
Non-Cash Working Capital
Current Assets                   80,936
Current Liabilities              60,434
Non-Cash Working Capital         20,502
 
Net PP&E                         81,588
Other Assets net of Liabilities  1,841
 
Invested Capital                 
Operating Cash                   11,874
Non-Cash Working Capital         20,502
Net PP&E                         81,588
Other Assets net of Liabilities  1,841
Invested Capital                 115,805
 
Operating Leases                 16682
Invested Capital adj             132,487
 
Research & Development           ???

I provided an alternative Terminal Value to compensate for error 2. However the best way to deal with not generating an error 2 is by extending the forecast enough to stabilize the earnings and returns on capital (error 1). Its hard to rationalize generating a terminal value based on stable earnings off a year when revenues were growing at 30%. You can see the problem by looking at the assume IC growth in the terminal value which is too high.

If you look at the calculated ROIC's, based on my invested capital calculations, they are increasing through the forecasted years and result in a terminal value calculation implying FORM will earn 52% on invested capital forever. This actually could happen if FORM refuses to lower margins to continue growth. That would be the rational behind my terminal value alternative calculation. That example can be found today with many high ROIC companies facing future low growth prospects.

So that's how I would review the potential errors found in the FORM analysis as discussed by Mauboussin. Glad you asked huh? Sheesh. Hey but I gave an alternative TV which would make it less overvalued :)))

**********

While I am at it, there is one paragraph at the end of the error 2 discussion that I disagree with, at least in how I am interpreting Mauboussin.

Maouboussin writes:

Most DCF models fail the economically sound and transparent test because of poor structure: the explicit forecast periods are too short and the continuing value estimates carry too much value. An investor should have a clear handle on the economic assumptions or implications behind whatever continuing value approach they choose.

A single stage DCF is in effect one big terminal value calculation. Therefore technically a calculation isn't "wrong" if its terminal value is larger than the explicit cash flows because we know that a single stage DCF is perfectly fine structure for a mature slow grower. Further to correct error 2 most of the time the terminal value will increase as growth expenditures are reduced which I showed in the alternative terminal value calculation. So its not so much the time horizon itself but the implications in growth expenditures and operating margins built into the terminal value that fail the "economically sound" test.

Geez - what an opinion.

Matthew