The good thing about correlated markets that open and close at different times is the potential for gap trades. The opening gap will occur when one market makes a decent move while the other is closed. Once that second market opens, you’ll most likely see a gap in the direction that the first market has moved.
This article will show you how to project an open and therefore determine if a trading opportunity exists.
The key to trading gap opens it to be able to estimate where the market is going to open and then trading any reasonable discrepancy.
Let’s say we have Market A rally by 1% while market B is closed. If these markets are highly correlated, you would then expect market B to open around 1% higher. If it opens 0.5% higher, there may be a buying opportunity as it plays catch up to Market A. If it opens 1.5% higher, it would most likely look like a good sell and it returns to normality or equilibrium.
So let’s look at some real examples. Trading the Aussie market versus the US market can offer these ‘mispricings’ several times per month. Within a normal day there is at least at hour where the US trades while the Au markets are closed. That goes for both bonds and SPI.
Assuming there is no market specific news, a move in the US Treasuries or E-mini while the SPI or Au bonds are closed will most likely cause a gap. Depending on the size of the movement, it may also present a trading opportunity.
Naturally, the key to figuring out if there is an opportunity is calculating where a market should open. We sometimes call that calculation a fair value, but more accurately, it’s a projected value.
The simple way is to take the percentage move in the open market and apply it to the closed market. It assumes the markets move hand in hand, but for the shorter-term it can often be a fair assumption to make.
On the Monday of the Queen’s Birthday long weekend the US Treasuries were open while the Au bonds were closed. Likewise the e-mini was open while the SPI was closed. Both Aussie markets were to re-open at 05.00 in the afternoon.
SPI versus E-mini
The E-mini came back from the weekend and opened with a gap higher. About an hour before the SPI re-open, the E-mini was up 14.25pts or 1.08%.
If we add that 1.08% to where the SPI closed, we have a projected open of 4154 (4110 *1.08%).
Now in practice, a gap of 44pts on open is pretty big. It’s not unprecedented, but it’s pretty big. To be a buyer, you would look for an open far below here. You’d be buying an open at 4110, 4120, 4130 in points higher in anticipation of a strong start.
On the sell side, if we watch the pre-open and see a potential open above 4154, the best trade to make would be looking for a quick profit taking correction. In fact you’d think anywhere near 4154 would be ideal given that is a big gap to start with.
In normal times, when the projected open is not so great, you’d stay flat on an open anywhere near the projected value. It is only the discrepancies, perhaps we can call them mispricings, that offer the trading opportunity.
For example, let’s say we had a close at 4110 and a projected open slightly higher at 4120. An open above say 4130 would definitely look high and be a good sell. An open below 4110 would definitely be a good buy. An open within that range probably does not mean a heck of a lot.
Trading the Au 10yr bond versus the US 10yr Tnote can offer some great opportunities. Just like the SPI-E-mini, these markets are influenced by their own fundamentals but we also have the Au market following the US based on global economic influences. It means the spread is related, but not perfectly so. That can make for good gap trade using our projected value method.
Leading up to the Au re-open after on the Queen’s birthday, we had the Tnote trading lower by around 14/32nds.
For us to project an open for the Aussie 10s, we have to convert the Tnote move to basis points. We have to do this calculation because of the different ways US Treasuries are quoted versus the Aussie bonds. US markets are quoted as price per $100 par value whereas the Aussie bonds are quoted 100 minus yield, similar to that of the Eurodollar or other STIRs.
At 133-01, we have the Tnote 14/32nds lower. The trick to applying this to the bonds is to use DV01s (see our Guide to Pricing Yield Curve Spreads).
The current DV01 for the Tnote according to the CME site* is $79.12. This is approximately 2.53 32nds (79.12 / 31.25) – with $31.25 being the dollar value of one 32nd.
This says that a move of 2.53 32nds in the Tnote is approximately one basis point.
If we have the market trading 14/32nds lower, then that is a move of -5.5bps (-14/2.53).
From this we can expect a 5.5bp lower opening in the Au 10s given the move in the Treasuries. The Au bond closed at 96.985, so this projects at open at 96.930. Again, this is a decent sized gap, but it shows even a slightly lower open will be a good sell or a massively lower open, say in the high 96.80’s, could be a good buy.
The SPI opened at 4144 and the Bonds at 96.940.
While the SPI opened 10pts below the projected value, it was a large enough gap higher from the previous close to make a long trade a risky one. If anything, it was a ‘when in doubt stay out’ scenario.
As for the bonds, at open at 96.940 was too close to our projected value at 96.93 to be getting short (although the trade would have worked in hindsight).
Plan the Trade, Trade the Plan
Trading an open on a projected value is all about that old adage “Plan the trade, trade the plan”. You can use the projected value calculation to estimate where the market might open if ‘all goes well’. The trading opportunity comes when the market opens a reasonable amount above or below. In practice, there are at least a few times per month when these trades exist.
The example above was one where the Au markets were closed for a holiday. Day to day however, the US markets continue to trade after both the day time and night time Au sessions end. It therefore makes sense to be watching for these gap trades every day.
* DV01s for the Treasuries are shown at: http://www.cmegroup.com/trading/interest-rates/duration.html