This is part one of a multi-post article discussing one means of using inter-market cross correlations to construct a statistical oscillator that may prospectively be used as either an effective filter or stand alone signal for index market timing purposes.
With the recent occurrence of the so called “Flash Crash,” it is now widely understood that extreme levels of cross correlation present a strong relationship to market risk and volatility. Being traders, we naturally look towards post-event turning points as opportunities to fade such moves.
Before we explore potential usage further, however, let’s get to the focus of part one: indicator construction. First, from the nightlyETF Rewind, I gathered a selection of twelve ETFs representing the major market sectors, as follows:
| Materials | XLB |
| Industrials | XLI |
| Energy | XLE |
| Utilities | XLU |
| Transports | IYT |
| Consumer Staples | XLP |
| Consumer Discretionary | XLY |
| Technology | XLK |
| Telecommunications | IYZ |
| Healthcare | XLV |
| Financials | XLF |
| Real Estate | IYR |
Using the historical daily time series of each ETF, I then measured each individual ETF’s one-month correlation with all other ETFs, and then took the sum of these results. After employing a small smoothing mechanism, I then took the 252-day percent rank of each daily summation inverted to match typical trader OS/OB levels. That’s it! This method produces the following oscillator relative to the S&P 500 proxy ETF SPY.
[Please Click to Enlarge]
In the next post, we’ll explore how the market performed during different deciles of oscillator ranking, as well as how the indicator may be used as a timing filter.
Lastly, as related reading, here is a recent article on current inter-market correlations going into the new quarter.
Related posts:
