Re: How Much Has Bill Overcharged America?

["tom hull" <thull@netway.com>]

  Charles Mueller wrote:
  
  
  >I haven't seen any figures here yet as to how many dollars we're
  >actually talking about--nothing on the industry's total sales, Microsoft's
  >revenues (hardware, software), it's reported total/unit costs, net profits,
  >and so on.  The Heritage Foundation--in its defense of Bill--gives us these
  >beginning numbers:
  >
  >1.  1996 'Computer Industry' Sales...............................$570 billion
  >
  >2.  Microsoft's 1996 'Computer Industry' Sales..........$ 10 billion (2%)
  >
  >3.  Industry's 1996 'Software Revenues'......................$ 250 billion
  >
  >4.  Microsoft's 1996 'Software Revenues'.....................$9.4 billion (4%)
  
  
  HF's point is obviously that Microsoft can't be much of a monopoly with
  only a 2-4% market share. Of course, that's ridiculous, on two counts:
  
  1) Most of the industry has no real relationship with Microsoft, so shouldn't
  be figured into the totals. Microsoft's only hardware business areas are
  mouses and keyboards: a tiny slice where, even with their retail presence
  and OEM leverage, they're insignificant. I don't know what the software
  breakdown is, but I'd expect that most of the money spent on software is
  for custom-written software, consulting, and integration services: not
  products, not business competitive to or even built on Microsoft. Most
  of the rest of the money goes to expensive software packages that
  businesses use to run their business: manufacturing, engineering, 
  process control, MIS. Microsoft would like to get their mits on some of
  that money, but they're not a player there. Throw out the areas where
  Microsoft isn't a player, and you're basically left with OS, apps, and
  development tools for desktop systems and small servers for white
  collar workers, small businesses, and home use. The pct. goes up
  quite a bit then, but still doesn't tell enough of the story. The software
  world is split into thousands of niche markets, and each niche supports
  a small number of products that are largely irrelevant to every other
  niche. So even with the narrowed market definition, many companies
  and products manage to stay clear of Microsoft.
  
  2) Some niches are much more critical than others, and some niches
  can be combined in ways that allow a company to artificially consolidate
  them, thereby extending its power. Microsoft is not just a big company
  that owns and sells a lot of stuff (like, say, Computer Associates); it is
  an interlocked web of strategic pieces that it has been able to leverage
  from its initial IBM PC-DOS deal.
  
  >        If Bill Gates really has serious monopoly power in the computer
  >industry, it is likely that the 'costs' he has imposed on it extend far
  >beyond the 'paltry' $10 billion or so he reels into the Microsoft coffers
  >each year.  On the basis of our century of U.S. antitrust experience, we
  >would expect that a removal of his monopoly would (a) drastically deflate
  >his own $10 billion 'take' in 2 ways, by collapsing his unit prices and
  >squeezing the waste out of his unit costs, plus (b) dramatically shrink the
  >overall (and unit) costs of the REST of the computer industry, including
  >both the hardware and software sides.  In other words, ending a monopoly in
  >a 'small' but key segment of an industry almost invariably brings a vast
  >reduction in its OVERALL costs, prices, and total revenue.  The lack of
  >competition generates bloat and unavoidable inefficiencies that pyramid
  >throughout the choked industry.  
  
  'Tis true that monopoly manufacturers tend to be less efficient producers
  than they would be if they had to face competition; however, that's a side
  effect of complacency, not a necessary attribute. Microsoft is anything but
  complacent, and software is different. Consider, for instance, how a
  monopoly software company might reduce unit costs:
  
  1) Reduce the physical package costs to a minimum, e.g. by omitting
  printed documentation. (If you own a publishing company, you can always
  sell the books separately, later, at a premium.)
  
  2) Don't bundle any support or customer service. Let them buy that after
  the sale, separately, when you can charge more. The less you tell them
  the more they're at your mercy.
  
  3) Slash your QA expenses. Customers will live with the bugs, or pay you
  to tell them about workarounds, maybe even pay you to teach them how
  to use the product to avoid the bugs. (Only thing to worry about here is
  returns, but intimidation helps keep them down.) If you keep this up, really
  desperate customers will even pay you to fix the bugs.
  
  4) Sell through OEMs and direct through corporate IS directors. Make
  them copy and install and support the stuff, just paying you license fees.
  Let them take the heat.
  
  Now, given what Microsoft has already done, how much more in the way
  of unit costs can real competition force Microsoft to squeeze out?
  
  Competition would cause Microsoft's prices to drop, but Microsoft's
  pockets are deeper, and their revenue streams and distribution channels
  are broader, than any foreseeable competition.
  
  The effects of this on the unit costs of the rest of the industry are harder
  to gauge, particularly if you consider the argument that Microsoft prices
  their products too low. Low prices have two major effects: they broaden
  the market, and they deter competition. (Remember that prices have no
  relation to costs as long as the business remains viable: in the absence
  of competition, prices are based on what the market will bear, a figure
  that to some extent varies with the size of the market. If you plot how
  many units you can sell at any given margin, the ideal price is the one
  that gives you the most profit.) Throughout most of its history, Microsoft
  has been hell bent on broadening the market, so they have generally
  been willing to sell things cheaper than they could have gotten away
  with had they merely been profit-maximizers.
  
  >        Historically, successful monopoly cases have caused industry
  >revenues (units sold times price, or total cost to the consumer) to drop by
  >some 10% to as much as 75%--most frequently something in the roughly 30%
  >range.  A conservative estimate here, for example, would suggest that Bill's
  >monopoly is costing the consumer at least 10% of the overall computer
  >industry's $570 billion in yearly sales, e.g., $57 billion per annum.
  
  The logic here is all backwards: the industry is not configured right; the
  historical record is not necessarily applicable; the answer is bogus (not
  necessarily wrong, but not established by logic). Let me try an alternative
  formulation (not backed by a lot of facts, but with some logic):
  
  I doubt that the overall market for Microsoft-influenced/competitive software
  comes to more than $60B; probably less. I'm not sure how important that
  figure is, but by your formula that would set the overcharge somewhere
  between $6B (10%) and $45B (75%) per year.
  
  Let's say that Microsoft has $10B of this, and that it's all overpriced due to
  monopoly. The direct overcharge would be the premium that Microsoft is
  able to charge over their competitors. If their competitors were commercial
  software companies, playing on a level, fair field, it seems unlikely that
  prices would drop more than 50% of Microsoft levels. (The big expense
  factor in commercial software is always marketing.) That would give you
  $5B in direct savings.
  
  The increase in competition should also decrease the prices of other
  software that currently competes with Microsoft. There isn't much of this
  (after all, Microsoft is a monopoly), and there's less margin to reduce
  there, but maybe we get another $1B (total so far $6B).
  
  These software cost savings should then be reflected in the cost basis
  of the entire economy that uses the products. For any given company
  such savings would be negligible, and much of the savings would be
  lost, but let's say we salvaged 50%, or $3B (total so far $9B).
  
  You can't really talk about saved money being re-spent, because the
  savings are already being re-spent by the monopolists, so that's a wash.
  The only other things you might get are:
  
  1) Damages, because the monopolist can get away with shoddy quality,
  which costs customers in time, lost productivity, and possible liability.
  
  2) Lack of innovation, because monopolies eliminate and usurp 
  opportunities, hide information, restrict and hamper know how.
  
  These are harder to estimate. I wouldn't be surprised if, against evenly
  balanced commercial software competition, these come to $3B each
  (total $15B). That would put my estimate pretty close to the 30% norm
  that Charles Mueller expects, but nowhere near his $57B calculation.
  
  Of course, if you replaced "evenly competitive commercial software"
  with free software, these numbers start to shoot way up. Free software
  can be built and distributed for about 10% of the cost of commercial
  software. (The development share of commercial software revenues
  and capital investment is less than 20% of sales, industry wide. Some
  of that cost is in restrictive and unnecessary "productization", and
  more costs are added in premature obsolescence. Free software
  has negligible copy/distribution costs, no sales costs, no profit. Service
  is separate, but freely competitive.)
  
  So with free software you would get: $9B from Microsoft; a much bigger
  percentage from the rest of the industry (say 70% of the remaining $50B,
  or $35B; that gives us $44B); comparable cost savings throughout the
  economy (another 50% gives us $66B); more reliable software, with
  less turnover due to planned obsolescence and retraining (say, another
  $12B, up to $78B); and who knows how much savings in innovation,
  now that anybody can see the code.
  
  >        As that good ol' boy from Illinois,  Senator Ev' Dirksen, once
  >reminded us, "A billion here, a billion there, pretty soon you're talking
  >about real money."
  
  Indeed. Antitrust is good and necessary. Free software is even better.
  The costs of monopoly are great, but small compared to the costs of
  running the industry for profit instead of for service.
  
  ----------------------------------------
  Tom Hull / thull@netway.com / http://www.contex.com/ftwalk