Another day, another half-baked over-long comment. In this case it’s a comment at Le Colonel Chabert about software and the implications for Marx’s theory of value. I’m posting this here because 1) commenting elsewhere at length uses up my blogging time for the day; and 2) it kind of relates to my post yesterday on value – I want to keep these together because otherwise they’re scattered all over the place. Keynes returns soon, I promise!
Competition is a crucial element in Marx’s theory of value and price of production. Where the forces of competition are suspended (or, rather, transformed since there is always some possibility of competition) there will be exceptions to the general law of value. Land rent is the classic case.
We need to be clear about ‘competition’, though, because Marx’s use is different from (and superior to) the neoclassical sense of competition (‘perfect’ or ‘imperfect’) many are more familiar with. For the neoclassicals it is solely competition between sellers in the commodity market. For Marx it crucially also includes the ability of capital to move between different branches of production. Also, for Marx there is no need for ‘perfect’ mobility of capital or a large number of competitors in the commodity market. The mere potential for new entrants into the industry is enough to discipline production processes. Even a monopolist in an industry is disciplined by the need to chase the average rate of return on capital or it will find it harder to obtain funds on the capital market. The law of value assumes a tendency towards the equalisation of profit rates. Firms with long-run higher-than-average profit rates are a good indicator that the normal disciplines of competition have been evaded to some extent – though in capitalist history higher-than-average profit rates have inevitably been eventually eroded.
The question to ask about software is, then, what are the unusual features of competition in this industry, and what is normal? First, in the commodity market: Intellectual property laws protect specific articles of software. But they do not protect them from close substitutes. There are use-value specific things about software that in combination with IP laws give more protection – the famous ‘lock-in’ and network effects, for example, that have made Windows so dominant in the operating system market. There is little doubt that Microsoft has benefited for a long time from these effects, accordingly winning very high rates of return on capital. But it has not been immune from competition – from Apple, from Linux.
The second thing about the software market is the extremely low marginal cost: that is, it costs next to nothing to produce additional items, though the initial outlay (design and programming) is expensive. This makes software firms extremely vulnerable from new entrants and piracy. The latter explains intellectual property law – without it capitalism is unable to allocate labour and capital properly to do the initial outlay. (Exceptions – state subsidy a la university research; cross subsidisation from advertising revenue a la the media; or cross-subsidisation from service/support/consulting work, which has been the strategy of many open-source oriented firms, including large ones like IBM.) Piracy is not the only threat here, though – there is also the risk of perfectly legal competition from a close substitute with an alternative lower-cost business model. Netscape vs IE is the classic case, also witness the tantrums Microsoft was throwing about open source not so long ago.
The third thing is that these firms are subject to competition on the capital markets just like any other capitalist firm. If the rate-of-return is high, capital will drift into the industry in some way and get in the incumbents’ faces (unless there is a true barrier to entry, in which case capital may flow to the incumbents). If it is low, capital will flow out: the incumbents will face higher costs to borrow and issue equity.
So to sum up, software firms are subject to fairly intense competition. To some extent it has followed the pattern of many previous new industries before their maturity, and as with them we would expect some time to pass before competitive conditions (including labour processes) settle down. But the low marginal cost issue (i.e. the ease of duplicating the product) will always haunt it and I expect make competition quite chaotic and unpredictable. It is interesting in that the high initial costs are skilled labour costs rather than fixed capital costs as in airlines (which also have high initial costs and low marginal costs and are therefore prone to debilitating bouts of price warfare). But this underscores the point that it is labour at the bottom of it all, and what makes it seem so weird is the disjuncture between the labour-time necessary to do the initial creative work (lots) and the labour-time necessary to reproduce the product (very little).