More ETFs time series have been extended beyond their inception

Dear All,

Some of you may not be aware of our powerful ETF price data extensions. Using a variety of methods we have extended over 60 ETFs so they are more useful in simulations. You can find the complete list in the document “Portfolio123 ETF Price Extensions” in SUPPORT->Library

Let us know how you like this and if you need more extensions.

Thanks

NOTES:
-In the extended data only closing prices are created. OHL are set to the Close and volumes are set to 0
-If an ETF has been extended a message appears in the Technical Chart and Prices page indicating the extended dates.

Marco, can you please add, extend and offer this ETF as a benchmark?

XIC:CN Available since 2001 : Replicates S&P TSX Total Return.
Extend to 1995-1999 with S&P TSX Total Return Index data.

Canadian Portfolios do not have a proper benchmark yet and we are not able to pull good statistics because of it unless we download the data from Bloomberg or other financial data providers.

Thanks!

This is great, Marco. But when can we do custom series using ETFs and CEFs? I’m especially interested in the latter, and don’t see why they’re excluded from custom series. It seems that’s the only major part of P123 that excludes ETFs and CEFs. And now that we have all this great data for these 60 ETFs, it would be wonderful to be able to use them in custom series too.

Or maybe this is already possible and I just don’t know how to do it . . .

Thanks, - Yuval

Thank you.

Marco, all of these are potential benchmarks as well.

Just like BND is a great new bench for bond strategies, I have requested SH be made a bench for short strategies. See:

https://www.portfolio123.com/mvnforum/viewthread_thread,10712

In addition, VIG would be a great bench for dividend strategies, and you could construct a blend of 40% BND & 60% SPY (rebalanced monthly or quarterly) to represent a “Balanced” benchmark for ETF strategies that blend asset classes.

Thank you, Marco. This is really great. Is it possible to add SPXL & TMF?

Marco, very much appreciated.
Could you also please provide extended data for:
VelocityShares Daily 2x VIX ST ETN (TVIX) for which data is available from 11/30/2010,
and
iPath S&P 500 VIX ST Futures ETN (VXX) for which data is available from 1/30/2009.

Both follow the S&P 500 VIX Short-Term Futures™ Index

Thanks a lot Marco, that’s great
I would add to Georg’s request: XIV, VXZ, ZIV. There are historical data on VIX futures indexes at least from July 2007 and likely older (VIX futures started in 2004):
for VXX, XIV: https://us.spindices.com/indices/strategy/sp-500-vix-short-term-index-mcap
for VXZ,ZIV: https://us.spindices.com/indices/strategy/sp-500-vix-mid-term-futures-index
I would also add TMV to the request of TMF

This is great. Will P123 be implementing ETF fundamental data such as P/B, P/S, P/E, P/CF? Thank you for providing the extra dates!

Hi,

One caveat that I would like to share to save time for others (took me a little while to understand what was happening)

Since the volume is set to 0 for the extended ETFs period, do not use “variable slippage” in your simulations when using an ETF over its “extended” period. At 0 volume, It will assume that this is a very thinly traded stock and give you a massive slippage punishment.

Obvious when thinking it through, but I could not at first understand why IEI was showing a catastrophic collapse in a sim with a market timer for 2005-2007.
See the public sim here https://www.portfolio123.com/port_summary.jsp?portid=1500866
(change slippage to variable to see the difference)

Regards,

Jerome

Jerome, thanks for pointing this out. As our extended data project is attracting attention we need to come up with a better variable slippage algorithm for ETFs. For now a quick solution is to just cap the slippage to 0.1% for ETFs.

borjapiano, there’s are no plans to add ETF fundamentals. One of the main obstacles is obtaining historical constituent data.

We’ll review all the other feedback as well . Thanks

Marco, these extended data series are very useful, thanks a lot.
May I ask you to extend TMF? You have already extended its inverse TMV, so it makes a lot of sense.
Thanks again.

We are happy to announce that the following ETFs were extended:

SPXL
TMV
TMF
TVIX
VXX
XIV
VXZ
ZIV
MTUM
IWF
IWV
HYG
XIC CN

Regarding XIC CN and VIG, a full review of the current benchmarks is underway. By the end of next week they will be available as benchmarks.

Great, thanks Riki.
Edit: TMF is still unavailable before 2009, unlike TMV

Good to see a total return Canadian benchmark.

Not sure if people here realize that XIC CN did not always follow the S&P®/TSX® Capped Composite Index.
From the iShares Fact Sheet:

IIRC, the only difference between XIC CN and XIU CN up to November, 2005 was XIC CN used a capped version of the S&P®/TSX® 60 Index. This was to avoid a repeat of the Nortel problem. At its peak, Nortel was over 30% of the index.

If XIC CN is used as a benchmark, the benchmark will be following different indexes at different times. Probably better to use XIU CN. Or extend XIC CN back from November 2005.

It is more important to have the S&P TSX Total Return as a benchmark from 2005-2017 with XIC:CN than to have the more continuous XIU:CN because no portfolio manager benches against the S&P TSX 60. If I present a Candian Equity strategy benched against the S&P TSX 60 to prospects they will tell me to swap the benchmark and recalculate all my stats. The S&P TSX contains about 250 names and is more representative of Canadian Equities.

@Riki37

I will send you the S&P TSX Total Return data in daily from 1999-2017 Monday so you can extend it correctly until November 15, 2005 so it is more seamless.

Thanks for the valuable inputs.

We already have the S&P TSX Total Return data (actually from 1994). We are currently working on some programmatic issues for CAD$ etfs as benchmarks and the data overwrite from 2001 and 2005, but the whole benchmarks revision should be completed by the end of the week.

Marco, thank you for extending the ETF data. The inverse equity ETFs and inverse fixed income ETFs are of particular interest. It is now possible to build short models which actually work.
• For the same asset classes and short holding periods, short selling inverse ETFs should produce similar returns as being long the corresponding long ETFs.
• Therefore, trading systems that work with long ETFs should also be profitable when shorting inverse ETFs.

Here is a short model which:
• During up-equity markets typically shorts two inverse equity ETFs of the four specified to be available: SH, SDS, SPXU, and TWM.
• During down-equity markets it shorts two inverse fixed income ETFs of the three specified to be available: TBT, TBX, or PST.
• A simulation of this trading system from 2000 to 2017 showed a high 38% annualized return with a maximum drawdown of only -19%.

Unfortunately short models are not allowed as DMs otherwise I would launch it. With a nominal Leverage Ratio of 1.5 this model with DM 3ETF-Trader plus shows spectacular returns of 67% with max D/D of -25%. (Maximum leverage goes to 1.99.)
$10,000 invested on 12/31/1999 would have had a market value of $94,434,989 as of 9/19/2017.



I suggest using caution while deploying leverage and/or short selling on highly leveraged or short ETFs: borrowing costs are usually punitive (for example, currently 13% on IB for PST) and during heavy market stress rates skyrocket to prohibitive levels. Part of those very high returns could be actual investment opportunities, but for sure a part of it can’t be harvested because of liquidity issues. Our current margin system is fixed, and it is less effective for backtesting periods of big fluctuations in financing rates.

If I had to trade that type of models on my own, I would use more liquid futures/swaps and the ETF simulation to calibrate duration/beta exposure.

Thank you for this comment.
The system does not allow Margin Carry Cost to be set higher than 10%. It should be easy for P123 to increase this.

The short model with 10% margin carry cost still shows a high 30% annualized return with a max D/D= -20%.

Thank you P123 for all the new benchmarks!! Very nice.