This is a third contribution from Andrey S. of Russia (also responsible for the trading nuggets here and here). Needless to say, Andrey is this short-on-time blogger’s best friend at the moment.
Here Andrey shows that some hi-yield bond funds exhibit strong weekly follow-through. By “follow-through” I mean up weeks tend to be followed by up weeks and down weeks by down weeks.
This strategy needs a little TLC because it can’t be traded in its original form. Andrey ran his test on Vanguard’s mutual fund VWEHX (because more historical data is available), but Vanguard penalizes active trading. So in the second half of this post, I’ll show results applied to something a bit more trader friendly, the ETF HYG.
[logarithmically-scaled]
The graph above shows the result of Andrey’s strategy trading Vanguard’s VWEHX (red) vs buy and hold (blue) since inception in 08/1989. These results are frictionless (do not account for transaction costs), but like most of my tests of longer-term strategies, I’ve assumed a return on cash of half the nearest 3-month Treasury.
Strategy Rules: Go long at the close of the last trading day of the week if HY bonds close equal to or higher than the previous week. Close the position and move to cash if they close lower. This strategy is long-only.
And for the number-lovers:
Naturally, because VWEHX was on one long positive run over most of the test, a trend-following strategy is probably going to do pretty well, but note that it was able to outpace buy and hold with about a fourth less exposure to the market, perform significantly better on a risk-adjusted basis, and weather all major downturns well.
A Little TLC
As previously mentioned, VWEHX is not appropriate for active trading; however, the ETF HYG is.
Looking at the performance of HYG (red) since inception (04/11/2007) versus VWEHX (blue) in the graph below, we see that VWEHX has acted almost like a smoothed moving average of the ETF.
[logarithmically-scaled]
In the next graph I’ve applied Andrey’s strategy to the ETF (green) over this short period of time and compared it to both the original VWEHX results (red) and buy and hold (blue).
[logarithmically-scaled]
Clearly, in this very narrow window, the strategy has more or less worked as well on VWEHX as HYG (albeit with a lot more day-to-day volatility).
As I wrote about recently, we have to be very careful assuming something works as well on an intraday product like an ETF as on an index or EOD product like a mutual fund. But in this case, because of the longer-term nature of the strategy (i.e. it’s not exploiting some very short-term, super-sensitive opportunity), I think it might fly.
To keep the effusive praise rolling, another giant THANK YOU to Andrey for his idea. I hope you’ll be hearing more from him on this blog in the future.
Happy Trading,
ms
P.S. Most of the studies that I perform could be reproduced with leveraged mutual funds from Rydex, Profunds, or Direxion, so generally I exclude transaction costs and slippage. This one in particular could not, so a trader’s specific trading frictions would need to be considered.
P.S.S. I specifically chose HYG here because it has tracked VWEHX so well, but a number of other products have not. Don’t assume these results apply to all hi-yield mutual funds and ETFs.
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Trading Strategies, Treasuries & Interest Rates | 8 Comments
-
Subscribe to RSS Feed
Subscribe via Email
My email: michael @ marketsci.com (remove spaces)
Interesting stuff – might be easier to code with weekly data rather than daily data. In this scenario, go long the close of the weekly bar if the close is higher than the close of the prior week. Go to cash if the close is less than the prior close.
RE to d: I agree – the original data that Andrey sent me was on weekly bars (and if I was going to trade it, that’s how I would approach it), but I switched it to daily so that I could show accurate drawdown numbers and graphs. Good comment.
michael
I’ve never tried to trade mutual funds before and have a questions. I recall redeeming a Vanguard bond fund once and seeing the the price didn’t reflect all of the proceeds that I got for the trade. I think there was some kind of accrued interest in there.
Am I recalling this correctly? If not, I apologize.
Interesting. Whenever there are rigid rules about when a measurement is taken (in this case, only once at a certain time each week), I’m always curious about how the trading rule could be optimized to account for price action between these snapshots in time. As always, super-fitting the rule to the past data probably wouldn’t be smart, but I’m sure there are simple rules that can be instituted that incorporate intra-week prices.
Also, can any profitable rules be formed for down markets? Our national crisis is only getting worse, and banks still refuse to lend money to otherwise stable companies. This will lead to many defaults, and there’s a good chance HYG will be a lot lower than it is now, in just a short time from now (the only impediment to this outcome is even more massive federal cash infusions; this alternative scenario can be profitably hedged by shorting the USD and/or US treasuries). Any rules to maximize the profit on the way there?
RE to LyricalOne: I absolutely agree. I’ve often stress on the blog the idea of trading in multiple timeframes (short, intermediate, and long-term) and I think adding a shorter-term strategy that took advantage of intra-week prices would bear fruit (just not something that I’ve taken a look at).
RE “down markets” – the follow-through observation in this post holds to some degree to the downside as well. Shorting HY on down weeks has worked as well, just not as consistently as to the upside.
Appreciate the good comments.
michael
The use of VWEX as a surrogate for HYG needs some TLC too. VWEX is not particularly “high yield” since it is more than half investment grade. Sometimes it even has a little dusting of AAA. You might think of VWEX as better modeled by 0.6 HYG plus 0.4 LQD.
RE to Garch: I won’t argue the components of VWEHX because I’m not familiar with the product, but I be hard pressed to agree with the 0.6 HYG + 0.4 LQD logic. Graph a quick chart of the three together, and it should be pretty clear that HYG has been a reasonably good fit for VWEHX. LQD has been an entirely different animal. michael
Thanks Michael. My suggestion wasn’t to try to come up with a short-term strategy related to this, but simply to enter/exit similar time-frame trades at timepoints other than the close of the end of the week.