Qualcomm Inc (QCOM)

[Data_Rox]

Breakthrough Ideas (2-8-03 edition)

A Contrast Case of Ignoring Residual Uncertainty.

As a contrast case, you might read a recent report on the state of the 3G industry by Brian Modoff, Michael Thelander, and Daniel Kaplan of Deutsche Bank Securities, which is entitled, "The Rise of the 3G Empire." Please note that this is the same team that ravaged the 3GAmerica's white paper for its false claims. This synopsis is based on a 19-page summary of their major, 275-page, second report on 3G that was issued in May. Although I do not agree with its assumptions, much of its traditional analysis, and, particularly, with its crapshoot market forecasting that ignores residual uncertainties, it does contain useful technical information. Perhaps, because it is so traditional, it can both sharpen by contrast and either filter or offset my personal opinions by offering traditional stock analysts' interpretation of the state of the empire and its future.
http://www.base-earth.com/july-august2002/stateoftheindustry.html

Also, I do not know if Modoff and his team actually believe their own forecasts. However, Modoff has been skeptical that Qualcomm will meet its internal estimate of sales this year, 80-85 million chips. Modoff prefaced his market forecast for 3G with an apologetic reminder that such an endeavor "is akin to being ‘a visually impaired gambler taking part in a crapshoot.'" Thus, I conclude, before we begin, that you should only take this numerical analysis seriously if you are a blind investor who must gamble despite being handicapped by their hysterical blindness.

But, Modoff wants to keep his job; so his team plunges ahead, and sells the content of his impressively fact-filled and quantitative, 275-page, state-of-the-empire report; then, Deutsche Bank's sell-side analysts use its contents to make stock sales to a naïve public.
This second report downgrades the forecasts of their first report because the unpredictable economic world of telecom worsened. This is yet another demonstration of the reason post facto analysis is so successful and ex ante analysis remains a gamble.

The Deutsche Bank approach divided the world into eight regions. In each region, it began by using the region's 2001 installed base as its fundamental foundation to iterate yearly estimates until 2007. It iterates by taking each subsequent preceding year's estimated installed base and correcting it by estimated entries for subscriber additions, churn, and replacement rate before reiterating it anew. Fearing that a gray market might be developing for sale of used or reused handsets, they correct for this possibility reducing the total to take account of this new estimated parameter in report 2.

Point forecasting that chooses to ignore residual uncertainty is an exercise in garbage in/garbage out. First, all introduced figures are based on assumptions. "Assumptions" announce a plan to ignore residual uncertainties in favor of extending trend lines into the future after assuming certain rates of change. Given the relative difference in the sizes of the installed based of GSM and CDMA, given the procedure is to iterate "corrections" to the installed base, and given the further assumption of a negative growth rate in subscribers, this can only have one outcome, a trend that reflects the original lop-sided installed bases.

Modoff assumes that the installed base remains in place because "additions," "churn," and "replacement rates" are neither large nor systematically biased in favor of tipping to one flavor of 3G, and, thus, his set of assumptions make this outcome so. This second report is used to lower the earlier forecasts of the first report, reflecting the economic troubles existing within the industry today. After assuming that the number of subscriber additions follows a falling trend line in almost all regions, Deutsche Bank forecast a minus12.3% CAGR until 2007. This ignores the fact that half of the world's population has never made a phone call, that wireless is cheaper to roll out than wire line, and that Asia is poised for expansive growth.

Based on this assumption of a slowed rate of growth, Deutsche Bank forecast an installed based of 1.7 billion subscribers in 2007. They estimated the ratio of subscriber additions to replacement sales in 2002 to be 36% to 64%, with a subscriber/replacement ratio in 2007 of 14% to 86%. That is, because they assumed that the major growth in new mobile wireless subscribers is a phenomenon of the past, already behind the industry, it must rely predominantly on replacement/churn of handsets for sales, which they estimated to have a CAGR of 7% through 2007. Yet, Qualcomm is reporting up to 70% replacement sales this year. "By 2007, we believe that 6.3% of the installed base will be WCDMA subscribers, 68.3% GSM subscribers and 23.6% CDMA subscribers."

Although it is true that replacement sales are an important element in Qualcomm's strategy, this is so primarily because a consumer electronic model is being introduced that spurs rapid replacement to gain new and desirable features. It is not the case that Qualcomm believes its growth is behind them, or that replacement sales will be so anemic, or that GSM will long dominate the wireless world's installed base. For the latter, GSM1x is Qualcomm's reason why it will not.

Because Deutsche Bank Securities "assumes" that the yearly number of 3G additions is captive of a perpetual decreasing subscriber growth trend, this assumption determines that neither CDMA nor WCDMA can pass the old GSM/GPRS/EDGE installed base in their selected time frame. Because, Modoff's procedure is iterative, begins with the 2001 installed bases, and assumes no substantial crossover movement, just random churn from one technology to another, the present installed bases necessarily capture the 2007 outcomes. Thus, given Deutsche Bank's assumptions, the current lop-sided installed base of 2G GSM, under a scenario of slowly declining subscriber additions and a slow growth of replacements, dominates the future of 3G as well. Forgetabout China and India! Forgetabout Qualcomm's multimode strategy of harmonization.

Growth can inflect when network externalities magnify the magnitude of the scale of growth from arithmetic to geometric, even becoming off-scale. In high technology, Moore's law drives exponential changes in price/performance ratios that drive such growth. Thus, quantities are not always linear or normally distributed. In fact, some phenomena like networks exhibit unusual patterns of growth that can be best described by power laws.

Not only ignoring residual uncertainties, the lethal flaw inherent in forecasting continuing trends is that it necessarily ignores the drivers of explosive growth. In this case, Modoff assumed that only the rich-get-richer when, in fact, it is the fit-that-get-richest. Strategic architectural control trumps an installed base because of its great contribution to the fitness of enterprises.