Tyler Cowen seems optimistic about algorithmic high-frequency trading on the grounds that it appears to increase liquidity and reduce bid-ask spreads. His co-blogger Alex Tabarrok sensibly counters that these benefits may be outweighed by increased volatility. In a bigger picture sense, I think someone needs to step back and explain what exactly the problem was as of ten years ago that’s been solved or somehow ameliorated by increasing the liquidity of US capital markets. This strikes me as the kind of thing where the diminishing returns curve is going to be pretty steep.
Then on the flipside, an awful lot of programming talent must be going into the algorithms arms race. Felix Salmon wrote yesterday about how VCs’ enthusiasm for funding engineer-helmed startups is making it difficult for large firms to assemble the teams of high-quality software developers they need to roll out massive cloud computing enterprises. That’s too bad, but I’d hate to restrict people’s ability to launch new firms just to create a pool of idle labor for Amazon, Apple, Google, and Microsoft to hire. Restricting banks’ ability to zoom around in circles trying to tweak their high-frequency algorithms, by contrast, seems like it would have very little in the way of costs.