The returns of some popular ETF pairs diverge more frequently than traders assume in all timeframes.
After the close of yesterday, Monday, January 27, 2020, a trader was astonished in Twitter by the fact that while GLD ETF rose on the day, GDX ETF fell.
Usually traders are astonished about some price action patterns because they have not done homework. Daily price action divergences are random events and occur more frequently than assumed even between highly correlated ETFs and securities.
Below is a chart of GDX and GLD that shows the sum of their daily returns on the bottom only when there is divergence (return of GDX > return of GLD, or the other way around.)
There are a total of 3444 trading days on the chart above, or 3443 returns, and 710 of them diverge as shown above.
Therefore, divergence has occurred 20.6% of the time, or much more frequently than some traders believe it has or it should have.
Maybe what confuses traders is a high correlation, as shown in the chart below.
The average 0 -lag, 252-day correlation of daily returns is nearly 0.79%. However, correlation of market returns is non-linear and stochastic. Probably a correlation of more than 0.95 is needed to guarantee low divergence rate between the returns of the two price series. Although 0.8 correlation sounds high, in markets it may not be significant.
Note that GDX and GLD have diverged about 16.5% of the time also in monthly timeframe.
In yearly timeframe, the divergence is even more pronounced at 28.6% although the sample is small.
The conclusion is that the returns of seemingly correlated security pairs may diverge more often than traders think and that can occur in all timeframes.
Charting and backtesting program: Amibroker
Data provider: Norgate Data
Technical and quantitative analysis of major stock indexes and 34 popular ETFs are included in our Weekly Premium Reports. Market signals for position traders are offered by our premium Market Signals service