An Interesting “Pairs Trade”

A very common strategy for “Risk Mitigation” is to hold both Stocks (often represented by the S&P 500 index) and Bonds (usually a long term bond fund – do realize that a fund has no maturity date, so you give up one of the protections of individual bonds. You can not just chose to hold to maturity to recover your principle; since other folks selling fund shares will cause fund sales of bonds prior to maturity and you will be allotted your share of that locked in loss of principle).

Also on this chart is TIP as a ‘safe haven’ in Treasuries with an inflation protection ‘kicker’.

Often there are opposite strategies possible for your choice of Risk Mitigation or Leverage Enhancement. Basically, you can “double down and double gain / loss” or you can cut it in half (as a gross generalization).

So instead of buying 1/2 of each to avoid the need to time things and mitigate the swings, you can alternate between the two, timing entries and exits to increase the risk, and also (PROVIDED you get the timing right…) increase the gains. You could even apply leverage techniques beyond that to increase the potential gain (or losses) on top of timing; so things like options and “double beta” funds that use options and futures contracts to give even greater volatility in the direction of the trade. (A direction you dearly hope you have correctly picked…)

So what does a chart of the S&P 500 vs the long term bonds look like? We’ll use SPY vs TLT as funds for those two types of investment. Here is a static image of a 10 year, weekly tick mark, live chart from Bigcharts.com:

SPY S&P 500 stocks vs TLT long duration bond fund 10yr Weekly with RSI MACD DMI

SPY S&P 500 stocks vs TLT long duration bond fund 10yr Weekly with RSI MACD DMI

Being a 10 year chart, there’s a lot to see here, and a couple of different market ‘natures’ on evidence. On the left side, we have a well developed bull market with a Fed holding fairly steady and modestly high interest rates. So the TLT bond line is mostly just laying there, flat. There are times this pairs trade will not work, during times when 1/2 of the trade, the bond portion, is essentially a ‘dead money’ market.

In the middle, we see a crash, and recovery. That’s an exceptional volatility time and IFF timed well, can make a lot of money. If timed wrong, it can lose a lot.

So the first lesson learned is that the strategy of putting 1/2 your money in each as a hedge, has a long period of under performance when stocks are hot and bonds are ‘dead money’. So why do it? Imagine a line 1/2 way between the bonds TLT line and the stocks SPY line. During that crash phase, it does lose some, but not nearly as much. It dampens volatility. That’s more of a ‘sleep well’ mix. Now if you could TIME the best moment to swap out of stocks and into bonds, that would be even better…

On the right side of the chart, after the crash, we have a very nervous recovery. The Euro zone is slowly imploding. The BOJ Bank Of Japan has decided to bugger the Yen (that had been a safe haven and that had been the ‘carry trade’ place to borrow at near zero interest rates). Now much of that has moved to the $US. So now our currency and our bonds move a bit more with money flow into and out of other markets and instruments. Here we see bond prices rising as The Fed kept cutting interest rates and there aren’t a whole lot of alternative safe havens. Notice how the peaks in TLT tends to touch the dips of SPY at bottoms of “dips” in stocks. That is in part an artifact of the starting point of the graph. Useful for this particular graph, but not a permanent feature. As the starting point changes, that tendency to ‘touch’ will evaporate. (The same shapes of the curves will remain, but since the ‘zero’ at the start point will shift, the curves will shift up and down relative to each other). The key point is to note how clearly they move in opposition. That makes for a decent pairs trade (as long as that relationship holds… so once The Fed and / or BOJ and / or EU Central bank etc. decide to shake up the global financial relationships again, it may well ‘fall apart’ – always be prepared to re-evaluate a trade strategy as conditions change. If (when?) the relationships break down, step out.

But for now, there is a nice “oscillator” going on, and fairly easy to read ‘indicators’ for when to swap sides. Notice, too, just how nicely a line 1/2 way between the two tickers nicely dampens things while still giving decent gains? Very different from on the ‘entry’ side of the crash phase. At some point that other behaviour will return. So if you do ‘split the two’, watch for a change of phase relationship between the tickers.

OK, with all that said, are there other things of use on this chart?

PSAR

On that upper price and Simple Moving Average chart, there are some little red dots. Those are PSAR Parabolic Stop and Reverse. This is often used to set ‘trailing stop loss’ orders. When prices then touch a PSAR dot, they swap to the other side and become a ‘buy if touched’ indicator. This works well in trending markets. It can cause “whipsaw” buy and sell activity in choppy sideways markets.

So look at the prior peak in 2007. Notice how the red dots are about equal above and below prices? There’s a sell, a buy, and a sell, all at about the same prices. That’s the cost of using PSAR. At major inflections, you take a few whipsaw trades. Notice how in the following strongly trending crash, and recovery, it keep you out during the fall, and gets you in for the rise (and keeps you in…) Notice, too, that sitting in TLT during “choppy” segments, and during “be out of SPY” segments, generally gives increased gains?

So one simple Pairs Trade strategy here would be to pick a BIAS (bull market, bear / crashing market, flat-weaving) and then choose to be in stocks preferentially, bonds preferentially, or cash preferentially; but with ‘trades’ between the preferred side of the trade and the ‘alternate’. On the ‘lead in’ to 2007, it’s clearly a Bull Market Bias, so the desire is to be in stocks and alternate to cash mostly. Some segments give a little better gain in TLT, but the timing doesn’t always align well. (A very sleepy dull bond market). During the crash (when stock prices have crossed the SMA Simple Moving Average stack and PSAR is hollering BE OUT!) moving into TLT paid off very well. So during a time when prices are below the SMA stack, and the SMA stack has ‘rolled over’ (slowest on top, fastest on bottom) moving to bonds is a very good strategy.

BUT…

Notice, too, that the spike in bonds is a parabolic bubble in its own right. It is over and crashing FAST in Jan 2009. Don’t overstay that side of the trade. When taking that side of the trade, you need to look at the chart of TLT to decide when to exit that trade…

Notice too, that the drop in TLT after that spike comes before the exact bottom of SPY. It takes time for a load of money to flow out of bonds and into stocks to stop that strong a free fall. So sitting in cash between those two moments is ideal.

On the right side, PSAR is useful mostly when on the bottom side of prices. It tells you when to get out of stocks as a short term choppy ‘local top’ is happening. Getting back in to stocks again is best not done using PSAR. So ride the runs, and it says when to exit the slop, but then wait for a clear ‘dip’ to buy back in. For now, at least. the TLT line starts a run up, and riding it until it reverses gives a decent ‘swap to SPY’ call. The peak in TLT comes just a little after the bottom of the dip in SPY, so that “dip” can help remind when to swap horses out of TLT. You have a bit of time to exit TLT and get back into SPY. Time when both are running up together.

It also looks like TLT starts rises out of the dips in TLT just a bit before the top / rollover in SPY. ( I note in passing that right now, at the right hand edge, TLT is starting a rise…)

RSI and MACD

Now look at RSI. In the left side bull run bubble, it’s a weak oscillator between about 45 and 70. When it gets up near 80, there’s either a few months of drop, or at the top / crash, a series of ‘lower highs’ that ends down at 20 at the bottom of the crash. (That is the time to get ready to buy again…) On the right side, it’s also an oscillator, but a bit more reliable as we are not yet ‘bubbly’. (On the left side, the inflections of the dips below the mid-line do tell you when to BTFD – Buy the “friendly” dip). On the right side, we have the same behaviours. When ‘near 80’, we get stronger drops. Below 50 inflections are the BTFD moment. Inflections above 50, but not near 80, are not major drops, just minor dips. We have now got a ‘near 80’ inflection downward which implies a significant ‘dip’ coming.

Now look at MACD. Above zero is a bull market. Below zero a bear market (confirming the price / SMA stack rollover). “Crossover” to “red on top” is a ‘trade out’ and “blue on top” is a trade in, but when ‘below zero’ I usually leave any ‘long side trade in’ very fast. That is a ‘counter trend trade’ and it is often a lot easier to just sit in cash.

In particular, notice that “first 1/3 of 2008” blue on top inflection. Just a very short ‘counter trend rally’ back to the SMA stack and then a nauseating plunge. That’s fast trader time, not investor time. So once price is under an inverting SMA stack, it’s aggressive traders and NEGATIVE bias only. Right now, we have a positive SMA stack and prices above it. Most likely is just a return to near the lower slower SMA line and a BTFD moment. So a “flip” into TLT or cash, then a BTFD flip back. Until such time as the SMA lines ‘weave’ in a topping weave (like Jan 08).

If you look at MACD, it’s above zero in a wobbling sideways behavior. Bull market bias. Trade out on crossover to red on top, trade back on blue on top. RSI excursions give an earlier indicate of ‘change coming’. The blue line has ‘gone flat’, and looks like a crossover coming ‘soon’. So I’d put stop loss orders in place and / or buy puts. Notice that this is saying the same things as PSAR getting close to Price and as the rise of TLT.

Notice on MACD the ‘black blobs’ of “Divergence” on the zero line. They are the ‘blue minus red’. It’s dropped to very near zero. The momentum is leaving MACD. Clearly a tangent to the last of those black blobs is now tilted downward. I’d be out of stocks here, and into TLT, as a pairs trade; but ready to trade back on a resumption of trend (i.e. the BIAS stays ‘bull market’ and the preferred action is BTFD.)

ADX / DMI

This is a harder indicator to use. First, look at the black ADX line. It tells you how strong any given movement is. Below about 20 to 25 MACD doesn’t work as well and we swap to Slow Stochastic. MACD is for trending markets, Slow Stochastic for trading flatter untrending markets. Right now, ADX has been below 20 for about a year, and we’ve had a rising market. That’s suspect. Mostly I use that “rule” on 1 year daily charts, so I have to wonder if ADX is less useful for choosing between MACD and Slow Stochastic on a 10 year Weekly tick mark chart.

At present DMI+ is on top “Blue on top”. It has not yet inflected downward. That implies ‘not quite yet’ the top. AT the red / blue crossover, it’s a clear ‘be out’; but DMI is often just a little bit late. So watch it to confirm what you are planning / doing while RSI gives more of an early ‘heads up’.

During the bull market phase on the left, a ‘red on top’ blip is the “BTFD” indication. Some times the ‘blue black’ crossover says a local top and dip coming. During the crash, it’s clearly screaming “be out”. That giant ADX peak inflects just before the clear bottom, and the “blue on top” crossover is a tiny bit after the dead bottom, but not too bad an entry. On the right side, a ‘color / black’ crossover is calling a ‘change positions’ pretty well.

Overall, it looks to me like “now” we’ve gone to more of a ‘red peak BTFD’ style of reading the indication, ADX black line less ‘chatty’, and mostly swapping on inflections of DMI+/- lines.

So less useful (but we’ve got a lot of working indicators to use here already…)

TLT View

Does using a chart of TLT give any added ‘advice’?

Live chart:

TLT 10 year weekly live chart

TLT 10 year weekly live chart

Here it looks like TLT is more clearly saying “Buy TLT now”.

MACD is “blue on top” crossover. DMI has blue crossover of both ADX black line and the DMI- red line. We do have MACD below zero, so negative bias. We’ve also got RSI from more of a middle oscillator position and we’ve got DMI with recent ‘red on top’. So bonds are a bit weak at the moment, but likely to have a short term run upwards.

We also note that price has gone through the PSAR dots saying “buy in”.

So, all in all, TLT is saying ‘get out of stocks for a while and trade into bonds’.

Depending on personal preferences, that could be just going to “1/2 and 1/2” as a hedge / “rebalance”; or it could be ‘flip to bonds’ as a higher risk trade.

Looking at all the indicators in detail is left as an exercise for the reader (as I’ve given an example above on SPY)

A One Year View

It is also possible to look at those same tickers on a one year basis. Does a shorter time frame, closer view, given any better timing or any more precise information?

These will be live charts too, so what I say now will not match them if looked at 6 months or a year from now.

TLT one year live chart

TLT one year live chart

Here we can see it called an ‘entry’ to TLT a good month ago. MACD crossover. RSI with “higher lows”. MACD above zero (on this faster time scale). DMI clearly “blue on top” (and has been for a while) with reasonably strong trend.

Overall, it looks like a decent faster “call” on what to do.

How about the SPY chart?

SPY vs TLT one year live chart

SPY vs TLT one year live chart

It looks like the “get in” for TLT calls an exit on SPY just a bit earlier than SPY itself. It also looks a bit like an “entry” call for SPY comes a bit before the top on TLT. Nice, if it holds up.

In Conclusion

Hopefully this will help folks think about “pairs trades” and how you can use them. In hedge funds, a “pairs trade” is usually a “long / short” strategy. Long retail short SPY for example, if you expect retail to out perform the general market. It is also possible to do things like “long long” and even alternating which side of the pair you ‘go long’ at any one time.

There are as many potential ‘pairs trades’ are they are things to compare to each other. $US vs Yen. $US vs EU Stocks. You name it. All it takes is comparing two tickers and seeing some kind of strong correlation over time. For example, copper usually rises prior to retail sales. Somebody needs to buy the copper to make the stuff before all those phones, computers, cars, and even houses, make it to market and get sold. So early in a rise you could do “long copper short retail”. If you are “too early” both still drop but you don’t get hurt. If you “called the bottom” exactly, copper rises while retail falls a bit more, or goes flat. You win. After about 2 to 3 months, retail starts to rise and either you take the trade off, close out the short retail side, or even ‘double down’ by swapping to ‘long retail long copper’ in a strong “Risk On” bet.

Looking at TLT and SPY, it looks to me like a ‘swap between them’ is working well in this market status, with the entry to each side of the trade calling the swap fairly well, but with many other indicators to confirm (or raise doubts…) about the reasonableness of the trade or the risk being taken.

We will be able to watch these live charts over the next few months and see how well it works. Right now, they are saying leave stocks and be in bonds (and in fact TLT said to do the swap a month ago). During that month, stocks has a ‘peak’, but mostly was flat both sides. Unless you traded out at that peak, being in bonds was more productive.

Overall, it looks to me like a safer, and better, strategy than “buy and hold and pray”. For those who don’t want to do the trade / swapping, these charts also help to indicate best times to “re-balance”. So if you are 100% in stocks, this is saying it’s likely to be a good time to ‘re-balance’ into 1/2 stocks and 1/2 bonds at this point. Also, you don’t need to go 100% stocks or 100% bonds, you can use a 60-40 bias swap if desired. So now would be a “60% bonds / 40% stocks” balance and when the indicators swap to be in stocks, you could have 60% stocks and 40% bonds. You can chose any range from 100% – 0% trades to 50%-50% static non-swapping to any ratio in between. Basically, it can be a modest risk strategy to use in small amounts to get comfortable with it.

Hopefully this is interesting to folks as a more advanced strategy examination.

Subscribe to feed

Advertisements

About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in Economics - Trading - and Money and tagged , , , , . Bookmark the permalink.