Note that DIA did also spike down in tandem. The programs play the futures & DIA back and forth. Program risk arbs weren't big enough to soak up all the volume....but sure shoulda been by the next day! Liquidity and speed were also an issue to nail them both down.
In regards to the DOW futures drop without instant arb correction, there's been multiple times worse in the mini DOW futures, and >50x worse in similar contracts. As to why spreaders didn't step in on that occasion, it's too damned dangerous.. and not for the reason you're thinking. The exchanges have a nasty habit of completely busting the trades executed to correct their markets, often based upon the pit contract's high/low. The computer can't know in an instant what's going to busted and what's not, so the smart thing to do is often to stay out, otherwise you'll end up legged in a market that's almost always going against you. There are still pure arbs out lurking in the shadows. However, it's a tough game since other forces typically keep the markets in line well before arbs could ever make money executing. A quick study will show you that the spreads vary less than the minimum possible transaction cost 99.9999% of the time. As a simple example, consider MSFT on ARCA vs MSFT on INET. The two markets are entirely independent, so they can get out of line (and do - once in a blue moon, like when ARCA's match engine falls over, again). There's a simple force keeping the two in line much tighter than could ever do though: smart order routers are everywhere. If ARCA is higher than INET, it'll get pounded back into line as the orders hitting bids are routed to ARCA instead of INET. That's sometimes referred to as "1-sided arbitrage", and occurs in many much more complex markets situations as well.
Landboy, You are thinking like an average investor (sorry if I am blunt)! There are ways to prevent that from happening you know! And, NAV can be calculated every 15 seconds, or even less if necessary! _______________________________________________ Ok, even if 2% arb. doesn't exist, hell, I'd settle for 1% or even .3-.5%. However, you got to admit it, facts don't lie. At the close of market sometime last week, the QQQQ was trading at around .73 +-.1 premium to its NAV. I didn't believe it at first, but fact doesn't lie! What I don't understand is, if there are all these programs out there aiming to arb. the market, how come QQQQ was trading at such high premium? Why didn't people or program arbitrage it immediately at that time to bring the premium down? http://www.nasdaq.com/AspxContent/Nasdaq100PremiumDiscount.aspx The above link to the nasdaq website showing that there were five days in the 3rd quarter of '04, in which the QQQQ traded at .25-.5% discount and four days of premium of .25-.5%. Not mentioning a day of discount at .5-1%. Are you guys sure about all the arb. programs out there, or are you just guessing? People could make a pretty good living with .25% premium/discount! That's around $.1 per share (38X.0025)! Let's minus $.02 per share for the sake of transaction costs, and $.03 interests cost for 7 days at 4% (.04 X 38/365 X 7). $.05 profit is pretty darn good!
We know the programs are out there. Half of them are right here, ET'rs with automated programs. Futures will often trade at a premium/discount to cash as well. I'd venture that QQQQ's risk arb are gonna be tied to futures, and that what you saw was due to futures premium or liquidity. I've got programs that have NAV's down to sub 1 second--a pure necessity for arbitrage. The world wasn't invented yesterday. If you've thought of it.....we've got it covered already.
I"ll have to be blunt too, guess who calculates these NAVS, ME! That's right, my 9 to 5 job is to calculate NAVs for ETFs, and guess how many times a day I calculate it? every 15 seconds you say? NIce try, the answer is ONCE, one time, and that's at 5 o'clock EST like I said before. That is the ONLY time you can call up Barclays or State Street to exchange at NAV. No I'm not an avg investor, I actually speak from hands on experience
I misread your original post. The error you made was much like the error beginning futures traders make, so I thought that was what you were talking about. In any case, I would have said there is no such disparity in the NAV, nor could there be. Only a trading error would cause even a 25 cent swing.
So what if you can only exchange at NAV with banks only once per day, you can duplicate it any time you want.
Well, you were wrong before and you're wrong now. AMEX datafeed contains NAV estimates (Intraday Indicative Values) that are updated every 15 seconds throughout the day. http://www.amex.com/?href=/etf/Glossary/Gloss.htm BTW, another cost related to ETF arbitrage is the creation unit cost. It is on the order of one or two cents per share but that's enough to cut into arbitrage margins. Martin