Some thoughts from a former analyst:
I believe that the inherent limitations of our intellectual capacity require us to reduce even the most complex analyses to no more than three or four variables. As a result, companies that produce millions of products a year for billions of dollars of revenues have their valuations calculated by investors using literally less than a handful of factors. Nuance and subtlety play distinct second fiddles to grotesque simplification. One can be dead right in identifying a development and assessing its potential implications and have to wait years for its manifestation in the stock market. Case in point - AMD's decision to embrace systems architecture as the key determinant in its processor design. Even then, there's no guarantee that sheer innovative brilliance results in business success; after all, if that were the case, we'd all be dazzled by the ongoing triumphs of the Digital Equipment Corporation (RIP).
But as all these things work their way through investors' collective consciousness, the easiest way to deal with them (in my experience) is the awareness of three things: (1) that investors use this simplification to bin companies and their equities into four classes (A,B,C and D, with the bottom class reflecting those companies with dubious on-going potential); (2) that real investing money gets made and lost with the market (Benjamin Graham's "Mr. Market", if you will)deciding to rebin a company; and (3) most investors, through laziness and/or stubbornness, will retain their perceptions of a company long after the rationale/s for those perceptions are appropriate.
In light of those beliefs, consider the following:
a) An analyst's comment on the unique nature of the AMD/Intel "competitive motion shift" conveniently overlooks several similar situations in the past. In fact, the killer tech stocks of the late 1990s (e.g., Dell, Cisco, Oracle, EMC, Applied Materials) all were companies that used some fundamental business focus to overcome a competitor or competitors with a previously superior position. The gains by AMD are merely a current manifestation of an aspect of technology that will continue to occur. Pay attention, Joe, there's a wonderful world out there that teaches its lessons for free.
b) An analyst's comment about AMD's potential for execution missteps misses three things: (1) AMD's execution problems have taken on a distinctly historical tone; (2) the processor execution problems seem to have bedeviled Intel for the past several years; and (3) all technology, by definition, has execution risk - if leading edge product development were easy, "wafer fabs" would be called "convenience stores." Here's a thought, Chris - if you're going to use cut-and-paste from old reports to recycle "analysis", restrict the usage to items of merit.
Now comes the hard part - gauging the extent to which the processor market will grow and how much of that growth will be taken by the smaller competitor with the pronounced superior momentum. Part of the reason that I felt that AMD's stock was a superior investment to Intel's was that relatively small shifts in share would have an outsized benefit to AMD at relatively small cost to Intel. Furthermore, with an enormous market share advantage, Intel was restrained in its ability to wage a pricing war since every dollar taken from AMD would take about six from Intel. As AMD's dollar share increases (and at a greater rate than its unit share due to enhanced ASPs), that restraint for Intel gets reduced. To a certain extent, Intel has been in a trap of its own devising; that ceases to be the case if AMD continues to accrue share.
So how much are these franchises worth? Intel still makes a whole lot more money than AMD, but on a per share basis the disparity has nearly disappeared (when you adjust Intel's EPS for the corporate tendency to use most of the earnings to buy stock back in the market, Intel shareholders have made very little money over the past year [net income of $8.7 billion in 2005 was used to buy stock back so that s/holders' equity actually fell $500 million and book value only went from $5.83 to $6.02 on the slightly reduced share count - thus only $0.19 of the $1.40 reported EPS actually accrued to ongoing shareholders]; AMD s/holders' equity went up $340 million and book value went from $8.02 to $8.13).
My take is that Intel has gone from an A company to a A- over the past several years, while AMD has gone from a C- to a C+ (with a great deal of the current disparity reflective of historic trends and their concomitant financial effects). Should, as I believe, the current trends continue over the next several quarters and the lags typical of most investors work as usual, I expect that these perceptions are likely to stabilize with Intel as a B+ and AMD as a B- as the awareness of the altered technological status of the two companies becomes more widespread and as the incrementally greater financial benefits accrue to AMD (even if Intel continues to make more money than AMD, the positive delta to AMD remains more pronounced).
What to do with the stocks? The valuation technique that I moved to while covering the group at Wells Fargo aggregated what you get now (tangible book) plus what the company is estimated to make for shareholders over the next ten years (defined as EPS reduced by the negative effects of share buybacks).
Goodwill reduces Intel's book value by about 11%, so we're talking about $5.35 per share. Therefore, the current price assumes about $17.35 in potential earnings over the next ten years. That seems about right - if you exclude the discounted value of those earnings and assume that the use of buybacks abates.
For AMD, assuming that Spansion is worth its carrying value (big if), the $37.30 stock price less a book of $8.13 values the earnings stream at $29.17. Again sidestepping the time value of that stream, that would require an average of nearly $3 per share every year for ten years. Assuming 450 million shares ($1.35 billion net income), a 30% tax rate and a 15% pretax margin, this average requires revenues of $12.85 billion; at a 20% rate and 20% pretax, it's still $8.4 billion. The former revenue scenario appears unlikely. While the latter is hardly a slam dunk, it's quite doable.
What's so magical about ten years? Nothing, actually. But we only have visibility into a company's technology for about 3-5 years and while we should pay more for a company's prospects than what we can visualize, there's no justification for paying for perpetuity when the annals of tech companies are replete with examples of just how difficult it is to succeed on an ongoing basis.
So for the first time in my memory, we have a period where both companies have their shares trading at what appear to this observer to be realistic representations of their intrinsic values. Maybe that's a sign that the Apocalypse is nigh upon us, but it's more likely that it's just an interesting coincidence.
Tad LaFountain |
