[yummyfajitas]

Nice article.

I'd also point out Alex Tabarrok, who makes a good counterpoint about HF trading.

His argument is that high frequency trading has only minimal societal value (positive or negative), but strongly influences the distribution of gains and thus may cause an overinvestment of resources. Read it here:

http://www.marginalrevolution.com/marginalrevolution/2009/08...

[huntse]

Yeah, I read that. As a matter of fact I think he's a very smart guy but somewhat wrong about this. Marketmaking is of positive value to society so they get paid for that. The marketmaker is being paid to take some risk and when you hit the offer to buy a share you're paying a little bit to the MM for taking some market risk for you.

Any other trading functions can be thought of as picking up alpha that other, less nimble, market participants are leaving on the table. That's the redistributionist part he's talking about there, but actually this is positive for society also because it increases liquidity also and makes the market more efficient. I also think that in general it's positive for society for the smartest and best people in a particular endeavour to be well-rewarded, because that gives a strong incentive to other smart people to step into the field.

Most of the carping about HFT comes from people who are losing business because they can't compete. Instead of moaning they should hire a roomful of brilliant programmers and write their own winning algos.

[yummyfajitas]

But regarding Tabarrok's point, I think it's this. It's important for the markets to move much faster than the physical goods they convey information about. I.e., house prices should move much faster than house construction.

Maybe a better analogy is this: there is pirate treasure at GPS coordinates (X,Y). It's good for society if one guy sends a boat to retrieve that treasure. It's less good for society if 5 people get into a race for the treasure. Maybe we get the treasure in 5 days rather than 7, but we also sent 5 boat trips out rather than 1.

Similarly, picking up alpha in 12 seconds rather than 15 is probably not useful to the rest of the world, although picking up alpha is better than leaving it on the table.

I do agree with you that most of the carping about HFT is stupid, however.