Inefficient markets

November 20, 2002

Technology and path dependency

Does the best technology always win out? Some economists say no. They have argued in recent years that chance often plays a decisive role in the adoption of technologies and that advantage normally goes to the first-movers in a new market. During the bull market, such views were widely accepted and served to hype the tech bubble. In more sober times, the same claims are receiving the scepticism they deserve.

It was Stanford economist W Brian Arthur who put forward the hypothesis that historical accident often determines the success or failure of competing new technologies. According to Arthur, new technologies are characterised by "network effects": the more people who use them, the more attractive they become. As a result, consumers get "locked-in" to the particular technology which gains initial market share, even when it is inferior to competitors.

Arthur found "path dependency," as he called it, in particular historical episodes when "inferior" technologies had seen off their "superior" competitors. He cited the adoption of the Qwerty keyboard for typewriters in the 19th century, which prevented later improved keyboards (such as the Dvorak keyboard, developed in the 1930s) from making inroads. Arthur also cited the victory of the VHS video system over Sony's Betamax, which some claimed was technologically superior. VHS was said to have won this battle by chance, getting more shelf space in rental stores early on, until the market tipped decisively in its favour.

The theory of "path dependency" has important implications for investment. If a single business could dominate its entire market and the marginal cost of acquiring new customers was next to zero, then the value of the business would be limited only by the size of the market it served. This was part of the investment case for the new economy: start-up companies were told to go for broke, losing vast amounts of money in the hope of infinite riches.

In fact, the theory of path dependency is highly dubious. Neither of Arthur's historical examples stands up to close scrutiny. For instance, all personal computers now offer users the option of switching to the Dvorak keyboard, but very few choose to do so because it does not actually increase typing speed substantially. Claims that people would find it difficult to learn a new keyboard are overblown. After all, over the last decade millions of people have switched from their old word processing and spreadsheet applications. Some economists also dispute the claim that VHS is inferior to Betamax, since the former has greater recording capacity. And even if we accept that VHS was the worse system, the arrival of DVDs is now rendering all videos obsolete.

This is not to say that "network effects" do not exist. The most successful internet companies are those that have created peer-to-peer communities, such as auctioneer eBay, where the value of the network increases exponentially with the number of users (in accordance with "Metcalfe's Law"). Similar networks are now evolving in a variety of fields, including rare books, genealogy, gambling, and even re-uniting old school friends. It would be difficult, however, to argue that these networking firms are inferior to those they have displaced. Nor can we be certain their competitive advantage will endure. In a world of rapid technological change and low switching costs, path dependency does not hold for ever.

Should short-selling be banned?

The Financial Services Authority is considering what to do with the plague of bears rampaging through the stock market. The banning of short-selling is one idea in the air. Short-selling is when someone borrows a stock they do not own and sells it for, say, ?5 on the assumption that next week it will cost only ?4, at which point they buy it, hand it back to the person they borrowed it from and pocket the ?1 difference. If banning is not feasible, the enemies of short-selling want to force short-sellers to show their hands. They claim that as investors are forced to reveal significant shareholdings, so bears should likewise be required to reveal big short positions.

In addition, the bear-baiters would like to increase the penalties for short-sellers who fail to return borrowed stock on time. Perhaps they have in mind the 19th-century Wall Street ditty:

He who sells what isn't his'n,

Must buy it back or go to prison.

Hedge funds fear that if they were obliged to reveal their shorts positions, then other players in the market would gang up on them. They also argue that increasing the penalties for failing to return borrowed stock would simply discourage short-selling in general. Since market efficiency depends on bulls and bears engaging on a level playing field, neither of these suggestions would bring any benefit.

There is, however, the possibility of increasing the transparency of short-selling without hindering the activities of the bears. Two reforms suggest themselves. First, the aggregate size of short positions in each stock could be published, as it is in the US, without revealing the identities of the individual short-sellers. Secondly, the cost of borrowing stock, which varies with supply and demand, could also be made public. Currently, this price sensitive information resides within the investment banks which act as brokers to the hedge funds. There is no reason why it should not be made more widely available.