Wal-Mart in 1988
Rayneman at Unreasonable Returns made a good point that it looks like it was more likely that Warren Buffett was purchasing Wal-Mart (NYSE: WMT) in late 1987 based on the U.S. News article that Whopper referenced in the deliberate practice challenge. I tried to ignore the stock price of WMT in 1990 and I didn’t remember the $23 limit price Buffett was quoted as using in the article. So I was either approaching the challenge with intellectual honesty or ignorantly oblivious to prices referred to in the information given. My intrinsic value estimate was $21-1/8. That was far from the range WMT was trading in 1990 but close to Buffett’s $23 limit price. However, my 1990 intrinsic value estimate is pretty meaningless put up against Buffett’s 1987 limit price. I am going to roll back my projections to 1987, make a couple of general assumption changes and see how the valuation would look in 1987.
The last time WMT traded under $23 was between October 1987 and January 1988. The 1988 annual report was for the fiscal year ending January 31, 1988. I rolled back my projections and made some assumption changes based on data provided in the annual report. Store openings was changed to 125 Wal-Mart stores and 18 Sam’s clubs from 165 Wal-Mart stores and 25 Sam’s clubs over the next 10 years. The average cost of a new store in fiscal year 1988 was approximately $4 million compared to $6 million in 1990. I assumed construction inflation of 3% per year to project capital expenditures (capex). I didn’t make any changes to the cost structure.
These changes resulted in the following forecast:
The shares outstanding in 1988 were slightly less than in 1990. The discount rate used remained the same in the valuations.
At first this looks like a dramatic discrepency for just a two year difference but this is consistent with a 21% annual growth rate.
Using a 5-year time horizon yields an intrinsic value estimate of $18.70 compared to the $26.30 estimate I had in 1990.
This is consistent with a 19% annual growth rate in intrinsic value.
Using the 10-year time horizon, Buffett would have been purchasing WMT with close to a 40% discount to intrinsic value when he was using a $23 price limit. I’m pretty sure Buffett wouldn’t have anchored so much on a $23 limit with that size of a discount.
On the other hand, using the 5-year time horizon, Buffett would have been purchasing WMT at a little more than a 20% premium to intrinsic value. I’m having a difficult time reconciling these differences.
However, I think a clue to Buffett’s thinking is nicely illustrated at Student of Value’s graphs of the 45 year history of WMT. The amazing consistency of operating profitability that has only declined to 6% from 7% since the company’s early stage and has ranged between 5.4% and 8.6% during the time frame points to a company with a strong franchise and sustainable competitive advantage. Additionally, Student identifies 1990-1996 as a transitional period where returns fall to a lower and more sustainable level. Buffett has an uncanny ability to get ahead of inflection points and I have to think he had some sense that WMT was going to transition into international expansion.
Geoff Gannon at Guru Focus makes a good point about viewing growth as a qualitative factor instead of a quantitative factor. Gannon makes the point about Coca-Cola (NYSE:KO) that Coke is a cola and has no taste memory. People do not get tired of drinking it. There is no taste fatigue. That makes it a franchise where bottling Coke can be repeated day after day after day and each day will be profitable. It is repeatable. Like KO, WMT is a franchise because it is repeatable. Worldwide, people do not grow tired of buying products at the cheapest price possible. Inexpensive goods never go out of style. That is one of the reasons that you can count on growth in WMT’s revenues as it expands into new markets.
Wal-Mart in 1990
Going back to my 1990 valuation, I think it is of interest to compare how my assumptions compare to the actual results of the company in 2000. On the top line, I projected revenues to be $89.9 billion per year which was only half of the $165 billion the company achieved in 2000. I clearly missed international expansion. The 1990 annual report had no information regarding international expansion. However, there may have been channels of information in 1990 that might have given an investor an indication that international expansion was a possibility.
I projected total stores open to be 3,125 versus the 3,989 the company had open in 2000. Again, this is attributable to international stores. My projection looks better when compared to the 2,985 U.S. stores open in 2000.
I projected net income to be $3.2 billion in 2000 while the company reported $5.4 billion. Another miss on my part. I projected 2000 EPS of $5.64. Adjusting for both of the two for one stock splits that occurred between 1990 and 2000, the split adjusted projected EPS is $1.41. This compares to the company’s 2000 reported EPS of $1.20. I did not project anywhere enough stock issuance during the 10-years as the shares outstanding basically doubled. That is why my projected EPS was 17% higher than the actual.
This a good example of the problems with projecting the future and why conservative estimates should be used. Even with my projections being off as much as they are, a purchase of WMT based on my estimates would have worked out to be a five bagger. Adjusting the January 31, 1990 stock price of $42.62 for both splits gives a split adjusted price of $10.66 which would have appreciated to $54.75 by January 31, 2000. Granted, the P/E (Price to Earnings) expansion that occurred was about to start deflating a couple of months later. However, even if the 45 P/E that WMT was selling at on January 31, 2000 were to contract back to the 22 P/E WMT was selling at on January 31, 1990, a holder of the stock would still be showing a profit after the tech bubble burst. As Buffett is known to say, “I’d rather be approximately right than precisely wrong.”