Barry Ritholtz posted an interest working paper today. What I found in the summary is quite interesting:

We find that HFT order flow is more correlated over time than that of the investment banks, both within and across stocks. This means that HFT firms tend more than their peer investment banks to buy or sell aggressively the same stock at the same time. Also, a typical HFT firm tends to simultaneously aggressively buy and sell multiple stocks at the same time to a larger extent than a typical investment bank.

What is a possible impact in term of market efficiency?

We find that instances of correlated trading by HFT firms are associated with a permanent price impact whereas correlated trading by investment banks is associated with only a temporary price impact. We interpret this as evidence that HFT correlated trading is information-based; in other words, HFT firms appear to be reacting simultaneously and quickly to new information as it arrives at the market place, which makes prices more efficient. This suggests that correlated trading by HFT firms does not appear to contribute to undue price pressure and price dislocations on a systematic basis in the UK equity market.

Thara – Quite an interesting defence of HFT. As the authors noted, this paper does not address the issue of HFT activities during the period of high stress such as the Flash Crash on May 6, 2010. What they show is that HFT firm can be quite effective in removing any price disparities.