"I just did a chart correlation between the canadian dollar and gold going back 21 years, the extent of my data on the $CD."
That's good, I bet you have a very nice & expensive piece of software and data supplier for doing that.
Now DoubleD and Jim, what was your baseline? Gold is price in US dollar, or if you want Canadian dollars, but that would be a currency X-rate between Canada and USA. So if you correlate GOLD with the Canadian dollar you are using tanted data. Example:
POG, USD versus GBP, and CAD versus GBP all go up. But the CAD/USD drops. Is the correlation lost? Not necessarily. You are dealing with variables of more then 1 dimensions.
CAD := f(USD, GBP, YEN, MRK, FFR, CHF .... )
This is why I said:"I'm not going to give you a partial differentials course on Special Drawing Rights and Cross Currency components."
You need to analyze EACH of the components as reference to a baseline. You can't use a straightline correlation formula for a range of 21 years, as that neglects monetary policy. You'd get no meanful correlation on anything over that range. Once you find the appropiate index for GOLD as a currency, and SDR aint it, then you can run the correlation. I'll give you a hint:
A) [GBP, YEN, USD, CAD, SDR, MRK-EUR, DXY0]
B) USA[M2, M3, 30&1_YR_BONDS]
Using from both A & B you can create a formula that will equal the XAU, and the POG. Both of which correlate to the real values. And then you can look forward. That's why you can look for time to short both Gold an XAU stocks. But the exact formula.... you'll never get from me. :p
Jim: Run an integrating window over your data, with a .8 alpha range. Then rerun the correlation test. I'm off to read about deconvolutions theory |
