Perennial Gale of the Software Industry

When the antitrust divisions in the US and the EU had filed suits against Microsoft for "bundling" its software with Internet Explorer, the courts said Microsoft's dominance at the time being was enough for it to see unmatched profits for years to come.

Something missing from almost every simple antitrust debate and all-too-common in an experienced antitrust debate is what has been referred to as dynamic efficiency. A half-century ago, Joseph Schumpeter had offered an analysis of capitalist industry as being one where monopolies are common and frequently swept aside by a "perennial gale of creative destruction". This dynamic competition cannot be analyzed - the way it commonly is - by price-competitions in short-term comparative statics. This is competition that is analyzed solely through current prices and short-run profit-maximizing equilibria.

Economist at MIT, Richard Schmalensee, had testified on behalf of Microsoft in the US antitrust cases. He points out in one of his articles four Schumpeterian characterstics of the software industry in general. The first is that software companies compete dynamically with each other. After the fixed costs of designing, developing, and testing new software, the marginal cost of creating copies of each product is "trivial". The cost of producing software is almost entirely fixed, given that advertising is also a fixed cost.

The second is that the value of software increases depending on its compatibility with other products, which Schmalensee calls "systems" and "network" effects. So the quality of a software platform depends on which programs and interfaces can be used with it. For example, few people would want an XBox 360 if there weren't any good games for it, or an HDDVD player if there are fewer DVDs for it. These "format wars" become contested grounds for dynamism and creative destruction.

Third, despite the network industry models, major innovators in the software market occur repeatedly and market power can quickly be displaced or complemented by rival firms in software or hardware. Schmalensee cites Excel software as evidence of rapid technological change, which was initially only available for the Macintosh and then incorporated into other platforms later.

Forth, the increase in functionality through dynamic competition involves increasing the durability of components and product features broadens the scope of competition. For example, spell-checkers, which were originally not bundled into word-processing products, shifts the boundaries between the categories of software markets in competition.

Something I find interesting in this framework is the distinction made between "Schumpeterian industries" and the possibility of, presumably, "non-Schumpeterian industries". Are there such industries if we take Schumpeter's thesis seriously?

Just as the Attorney General at the US Department of Justice Antitrust Division, Thomas Barnett, has used Schumpeter's dynamic efficiency framework to make the case for monopolies (and also the confusing case against cartels), it is evidence that the influence of Schumpeter's thought is quite widespread today, yet it is questionable whether many economists or antitrust authorities understand the claims he is making. An industry like Portland Cement may be non-Schumpeterian today, yet in the long-run ultimately displaced by the perennial gale of creative destruction. Portland Cement has used the same production processes for over a hundred years and nothing has changed, given that it is a very standardized product, yet cost-saving production methods over time have indeed been dynamic. Software, on the other hand, changes rapidly every six months and is so obviously dynamic we grant it Schumpeterian status. By the same Moorean-Schumpeterian token, with every six months that go by perhaps Microsoft itself matters less and less.

My single-most irksome contention I hold against the authorities on antitrust is that they appear to cherrypick from high-sounding theorists like Schumpeter to prove a point they are much less committed to than they would admit. Perhaps next they will be quoting from Thomas Malthus in order to explain why they ought to impose export-controls on rice, or John Maynard Keynes in order to explain why in the short-run we should actually fuck with the economy, and eventually (as they often do) quote the eugenic words of Thomas Carlyle in order to explain why everything they do is so dismally scientific.