In this study, I'm looking at the performance of gold, in terms of percentage gains/loss compared to the performance of the US dollar, tracked by the dollar index (DXY). As gold moves inversely with the US dollar, I inversed the DXY to set a comparative benchmark, hence 1/DXY.
Why this approach is better than just looking at the anti-correlation of XAUUSD and DXY?
luke1827
⋅
Sorry dp, I'm not sure what you mean by anti-correlation. Do you mean charting XAUUSD against DXY directly and seeing the divergences?
dp
⋅
Anti-correlation means negative correlation, sorry for that lingo -- its just negative value of correlation coefficient.
The problem is, if you draw correlation, say, XAUUSD and DXY, it's hard to say that most of the time they are negatively correlated -- like people often say "...dollar up, gold down; dollar down, gold up...' -- the correlation changes the sign over time.
What you are doing through looking at 1/DXY and XAUUSD correlation -- you are basically inverting dollar index. Because Dollar index is mostly dominated by EURO, one can use EURUSD in place of 1/DXY, so that inversion mechanism is more or less obvious. But as I previously wrote, I'm unable to find persistent (anti) correlation of currencies and commodities.
Basically my point is that inverting DXY doesn't give any new information in terms of correlation. Without inversion you get one sign, with the inversion -- another.
More important point is -- is it possible to say that in certain situations correlation works for some reason, and doesn't work in others?