Malkiel in FT.com a while ago, hat tip to financialrounds.blogspot.com:
Read the whole thing here.
As de facto market makers, high-frequency traders can exploit pricing anomalies and pick up pennies at the expense of other traders. Such activities are not sinister. The paradox of the efficient market hypothesis is that the people whose trades help make the market efficient must be compensated for their efforts. As former SEC Chairman Arthur Levitt has written: "We should not set a speed limit to slow everyone down to the pace set by those unwilling or unable to compete." High-frequency trading networks let large and small investors enjoy a more efficient and less costly trading environment.
The logic behind Malkiel's argument is horrible. High frequency traders don't make anything more efficient for anybody, because they don't provide actual liquidity - they take shares, jack up the price by a few cents and re-provide those shares back to the marketplace. I couldn't get any more shares than I could a second ago, I just have to pay more for them, and the same goes for the selling side.
The argument that "High-frequency trading networks let large and small investors enjoy a more efficient and less costly trading environment" is thus idiotic. Yes, spreads are smaller, but only because you have middlemen taking the spread from a longer term buyer and a longer term seller, so the efficiency isn't a good thing - it raises the cost of capital for businesses because all of us long-term investors have to demand extra pre-HFT return to compensate for the fact that we're going to get fleeced on both the buy and sell end by high-frequency guys. Similarly, it cannot be a less costly environment for everyone other than a high-frequency trader because there is a limited sum of money in the spread, and the high-frequency trader is taking some of it. By definition it must be more costly - in this case, if I'm a buyer, and without high frequency trading I could have been at the top of the order book to buy at 75 and sell at 80, but because there are high frequency traders I have to price a buy at 77 and sell at 78, that means I have lost a potential 2 dollars that hasn't gone to the long term investor on the opposite side of the trade - it's gone to a HFT.
The only way you could believe this was true is if you presuppose that "efficiency" is best measured by the size of a spread, not the cost of capital. That should be basic finance, except everyone in the Malkiel camp ignores it.