Predatory pricing behaviour in network marketsAny attempts to explain the dramatic growth of the Internet in the last decade, or indeed the equally rapid spread of Internet related products such as e-mail or even peer to peer file sharing software, will need to recognise the importance of network externalities. Network effects mean that the value of the product to any user increases as the number of other users expands. As a result take-off of the product may be slow to begin with when there are relatively few users but, once a critical mass has been reached, more and more people will want to be part of the network. The network is subject to positive feedback, making participation in the market even more valuable as it grows. This feature of network markets was first recognised by Jeff Rohlfs back in the nineteen-seventies when he was working for an American telephone company. [Rohlfs, J (1974) A Theory of Interdependent Demand for a Communication Service, The Bell Journal of Economics and Management Science 5(1) 16-37.] More recently Rohlfs has expanded and updated his ideas in his book "Bandwagon Effects in High Technology Industries, MIT Press, 2000.
One implication of network externalities of this type is that a producer of a new network product, conscious that it might only be profitable to produce once a critical mass has been reached, may feel justified in keeping the price of the product very low at the beginning of the product's life in an attempt to stimulate sufficient demand quickly to reach this mass. Then, having established some degree of control over the market, the producer may attempt to raise the price so that profits can be earned and the revenue foregone during the launch phase can also be recovered. Of course, in this dynamic environment it may be that the technology develops and improvements in the production process mean that costs can be pushed down in turn enabling profits to be made without pushing up the price to much. But it is not uncommon for network goods to be offered very cheaply at an early stage of their development with prices being pushed up later. An extreme case might even see the product being given away for free early on to quickly build up the customer base; then later a fee is charged once customers have become hooked on the product. Web browsers and web-based news and information services come to mind as example.
The justification for this pattern of behaviour would be that (a) without the low price at the beginning the market would not be able to"take-off", (b) without being able to realise the expectation of higher prices later the company's development costs would not be covered - if there is no prospect of profits in the second phase there would be no incentive for the company to develop the market at all. This sounds OK when we are talking about completely new network goods, but what about when it is applied to a new brand of an exisiting product type? Here the new entrant is trying to displace an incumbent and grab sufficient market share to at least become viable. Alternatively an incumbent might keep the price lower a little longer than necessary to build up sales to the threshold level in order to see off any potential competitors attempting to enter the market.
Is this good for consumers? Well clearly it can be so long as the producer does not (a) eventually establish a monopoly position in the market and then also (b) exploit this monopoly power by pushing up prices so as to extract more than "normal profits". Whilst the sequence of low to high prices might be acceptable as a device for ensuring the initial viability of a market, it would be regarded rather differently if it essentially a predatory pricing strategy designed to capture a market from other suppliers.
A new paper by Joseph Farrell and Michael Katz [Farrell, J and Katz, M (2005) Competition or Predation? Consumer Coordination, Strategic Pricing and Price Floors in Network Markets, Journal of Industrial Economics LIII (2) June 203-231] revisits the issue. In considering a two period model to analyze potential policy intervention against predatory pricing behaviour in markets with network effects, they find that the anti-predation rules proposed by others may not always raise consumer welfare and improve efficiency. It all depends on the way that consumer expectations are formed and coordinated.