I believe it is because during times of market stress it is difficult to price the ETF relative to the NAV of the underlying. So when they hedge (by selling the underlying when they are long the ETF) they have a risk that the price of the underlying settles at a worse than expected price.

Thanks, Hans-Peter Smit Wouldn't that just be a matter of adjusting the risk premium? Penny for dollar is still a price in a stressed market, what else may be missing?

Hi Kelvin, I think your right, but by how much? Ideally they want to be out of the market, but when they are obliged to make the market they can't. Perhaps their pricing algo isn't working adequately at those times and are not willing to take the risks that may occur. They don't want risk, just a lot of volume where they can almost riskless earn the spread.

As the article noted: " Shares of the funds are often easier to trade than their underlying bonds, potentially posing a risk if there’s a sudden rush for the exit."

Thanks, Billy Ethridge and Hans-Peter Smit. I also come across this: http://www.bloomberg.com/news/articles/2016-09-23/dutch-speed-trader-is-taking-on-citadel-in-the-u-s-etf-market I wonder what the Dutch - FlowTraders can do, and Virtu and/or Citadel cannot do in here? Beyond risk appetite or n/a hedging product, is there anything to do with allocation, tying the hedge and the ETF together electronically?

Virtu are just more sensitive to the ultimate impossibility of completely mitigating the blow out risk - seems like others are willing to ignore or take the risk to chase the nickels

Thanks PAUL GAMBLES. I noticed earnings for most HFT firms have been down significantly and Virtu may be seeking to sell shares in an IPO. I wonder if it's an overall market issue, or is it a matter of finding the winning algo? Instead of a lose-lose situation, would it be possible to have at least a win-lose in this market environment?

I'd agree that we're seeing crowding out in the sector which is likely to result in lower margins or increased risk-taking (or both) for participants. I'd see the HY business exposure more as a win small until/unless one day you lose big, possibly terminally and I think that seems to be Virtu's view.

I work in a stat arb shop. Here's my opinion.

It looks like Virtu wants to capitalize on the pricing mismatch between the bond ETFs and the bonds that those ETFs own. In order to do that they need to buy whichever is cheap (either bonds or ETFs) and sell what is expensive. They make their money on the pricing error between the two. If the bonds are illiquid then they can't buy/sell them and can't lock in the money to be made from the pricing error.

Thanks Greg Carleu. If Virtu only wants to make quick money on the pricing error through speed trading, then how are they supposed to be a market maker? I don’t know if such arbitrage may turn out to be phantom liquidity, but I assume a typical market maker should stand ready to provide market liquidity by committing/ sponsoring some of these bonds. If a market maker is not willing to have skin in the game/ hold reasonable inventory, why grant them a market maker status? Wouldn’t that be unfair to other stat arb?

OK I got it, Virtu is a passive market maker, they don't do statistical arbitrage and they don't hedge with some other basket of instruments. They indeed use speed and market structure advantages to gain edges over others. So stat arb folks should indeed dislike them.