PERCENTAGE of STOCKS ABOVE 200-DAY MA

Research
1

I’m working on a composite of 8 Breadth indicators, with one indicator being the number of stocks in the S&P 500 Universe (IVV) trading above their x-day moving average. I calculate this composite score using multiple Custom Formulas and Custom Series that reference one another. However, it starts with a formula that counts the number of S&P 500 stocks trading above (for example) their 500-day moving average:

(UnivCnt(“close(0) > SMA(500,0)”)/UnivCnt(“True”)) * 100

This looks like it will be a solid signal to include in my Breadth Composite. Here is the Raw signal plotted:

Stocks Above Their 500-day MA

Here is an example of the SIGNALS that might be generated from this series:

Stocks Above 500-day MA - SIGNALS

The signals aren’t perfect, but remember that this is going into a Breadth Score (“Internal Risk Composite” or “IRC”) of eight different breadth indicators that together provide very accurate signals. Moreover, the Breath Signal Composite is only one component of a multi-part system that includes composites for Stock Fundamental Signals, Macroeconomic Signals, Technical Signals, and Sentiment Signals. A composite system like this generates amazingly accurate signals.

However, it just occurred to me that this series may not be showing what I initially thought it showed. I may have been the victim of faulty thinking about this series because the constituents of ETF universes may not be historically accurate at any given point in time (since 1999).

I am not sure if P123 maintains a historical constituent list for all of their configured universes (S&P 500, Prussell 2000, S&P 1500 Composite, Nasdaq 100, NYSE, etc.). For example, the top chart above shows a reading of 59.6 today, and if we go back to June 30, 2014, it shows a high for the series at 94.34.

However, I’m wondering if that reading in 2014 was for the constituents of the S&P 500 ETF (IVV) on that day, or if it’s showing me 2014’s reading for TODAY’s IVV constituents? If the later is the case, wouldn’t that be applicable to all signals that require a count/percentage of the S&P 500 (or any other universe that’s used)? While other technical signals would be accurate, I’m not sure if any series that requires a historical count of the number of stocks in an ETF meeting certain criteria would be accurate.

What about historical Net Highs, Advancing vs. Declining stocks in a universe, and other similar readings that require a count of stocks meeting certain criteria. Thanks for the feedback!

Chris

2

The charts aren’t appearing on this page for some reason. Does anyone know why?

However, if you click the underlined links, it will open the charts.

3

All our universes are composed of their historical constituents. There is no look-ahead bias there.

4

Thanks for the quick response, Yuval.

1) Just to confirm, let’s say I run the Custom Series above showing the number of stocks above their 500-day moving average on June 30, 2014. The reading is 94.34%, which indicates that 472 S&P 500 stocks were above their 500-day moving average on that date. That figure is from the S&P 500 (IVV) list of constituents on June 30, 2014 — and not that date’s count of stocks above their 200-day MA using the current list of IVV constituents?

(The reason I want to confirm is that you said “yes” in your earlier answer, but I believe your response goes against information provided to me by Paul in a related conversation. I just want to make certain about this, as it’s critical to my models.)

2) If “yes,” I assume your answer also applies to the 29 other pre-configured Universes currently listed by Portfolio123 and is true for Jan. 2, 1999 to today? Since the iShares Core S&P 500 ETF (IVV) was launched in May 2000, from where does the list of constituents come prior to IVV’s launch date (eg, 1999 and first-half of 2000)?

3) What about Custom Universes I have created? Are their constituents historically accurate also?

4) What about P123’s pre-configured Benchmarks? Does the same answer apply to Benchmarks also?

Thank you.

Chris

P.S. - Is there some reason that my charts don’t show in the above post?

5

Yes, all the universes and benchmarks reflect the historical constituents at the time of the data you’re looking at. We have extended the historical constituents of certain newer ETFs using reliable historical data.