26 September 2005

Valuation of WMT - Look at What the Rookie Did

A very helpful reader, John Coumarianos
figured out that I favored Wal-Mart under my Discount Shopping Thesis. John then put forth the challenge: "How about a valuation analysis now?"

I didn't have a valuation method, or know anything about developing one. So over the weekend, I accepted the challenge to come up with a valuation for WMT.

So let's look at what the rookie did:

I worked through valuation materials from Prof. Aswath Damodaran of NYU/Stern at his resource-rich site.

I also pulled numbers for WMT off of Yahoo! Finance, SmartMoney.com, and MSN Money.

There was a tremendous amount of trial and error with the various FCFE models I found. Invariably, I would be left with one or two variables that I just could not figure out. I settled on the following last night at around 4 am:

Value Per Share = (Free Cash Flow to Equity(1 + Expected Growth Rate of Earnings)) / (Cost of Equity - Expected Growth Rate of Earnings)

Here's how I figured the inputs to the above equation:

For the numerator, I tried operating cash flow minus cash expenses for growth equals FCFE. Operating cash flow is net income ($10.267B) plus depreciation and amortization ($4.405B) minus (capital expenditures ($12.893B) + change in working capital ($- 1.4B)) which equaled $3.179B.

Now for the denominator, I had to artfully fudge some numbers. The Cost of Equity I devised was the approximate 10-year interest rate for U.S. bonds, or 4.25%, plus a 5% premium for the U.S. market times Beta (.52), minus the Expected Growth Rate of Earnings for WMT of 5.8%

Solving for VPS, I divided $3.179B by the outstanding number of shares (4.16B) to get .764. I then multiplied .764 by (1 + expected growth rate of earnings per share, or 1.058) to get .808. I then divided .808 by ((.0425) + .52(.05) - (.058)) to get $76.98.

$76.98 is significantly higher than the $43.22 share price that WMT traded at after Friday's close. But that's nothing compared to some of the inflated numbers I came up with in earlier attempts.

Assuming my model isn't flawed, and that's assuming quite a lot, WMT is looking really cheap.

Trying to put a value on WMT has been a frustrating but educational experience. Any suggestions for tweaking my model/figures, or does anyone have a completely different valuation model that you would suggest I use?


John Coumarianos said...

Nice job. It's probably a bit richer than I would have come up with, but I'm no equity analyst. Typically, if you're discounting future cash flows, you'll estmate future free cash flow (operating cash flow minus CapEx) for a few years (maybe seven years or more if you think the company has serious competitive advantages). I suppose something like 10% for the discount rate is fine too for WMT. Then, you typically do a "terminal" or "perpretuity" value to account for the continuing cash flows (when you're not confident in estimating them anymore, and when you expect the competitive advantages to erode). That seems like what you've done here -- the terminal or perpetuity part of a DCF (discounted cash flow) model. Most people choose 3% (the long-term growth rate -- after competitive advantages erode -- which is the rate of inflation for the terminal value formula). They figure they're estmating growth in the first few years of cash flows (and, remember, competitive advantages warranting more growth tend to erode over time thanks to the competitiveness of capitalism).

Other investors like Bruce Greenwald (you should check out his books) think DCF models are too much of a crapshoot. Greenwald is also suspicous of earnings growth, because he thinks investors forget too easily the capital and costs necessary to generate growth, which winds up being of no economic value in most cases. So he prefers combining a "net asset value" (assets minus liabilities on the balance sheet) and an "earnings power value" (the most recent earnings divided by the cost of capital. That usually winds up being more conservative than a straight DCF model, I think

But, nice job overall. If your analysis is right, you've got yourself the proverbial 50-cent dollar. In any case, I think it's a useful exercise to go through a valuation model or to try to hang a number on the company, if for no other reason than to remind yourself that you're buying a piece of a living, breathing business and to force yourself to think about what it's worth.

Try playing with the numbers in your model. For example, if you change your long-term growth rate asumption, you'll see the value fluctuate significantly. That always gives me pause. But this is a nice beginning.

Anonymous said...

Here's a suggestion. Go back multiple years, running your VPS model simulation against annual data points to learn of the values estimated, while matching those values against the relevent market prices, creating a time series comparison. This should give you a feel for the robustness of your model versus market prices.

Good luck.