This seems a rather simple question: if US rates increase, short gold, otherwise, buy gold.
This is however too simplistic. Traditionally, gold rises when the US$ decreases, but also when US interest rates increase. Gold also moves higher in times of fear or instability.
Below is a figure whose upper panel shows in green the difference between the 10Y rates and the 10Y TIPS (which are the 10Y inflation indexed rates.) This difference is the expected inflation.
It is interesting to see that there is a 90% inverse correlation between the 10Y TIPS and gold. The push in gold prices started late 2008, when the Fed embarked on its QE program. This program forces the 10Y rates down, but the 10Y TIPS continued to push deeper because the QE program was generating monetary inflation.
Gold started to weaken by late 2012, when the Fed started talking about "tapering." However, the BOJ's NIRP decision in February 2016 was the reason for the second push in gold. In itself, NIRP does not generate liquidity, but it forced Japanese investors into 0% yielding assets: cash and gold.
When the Fed increased rates on December 14, 2015, gold hardly moved for the next few weeks. However, the real move higher started with the BOJ NIRP. This tells me that what matters is not really the Fed's actions on Wednesday, but mostly how the BOJ reacts. If the Fed raises rates, the BOJ will happily do nothing and gold ought to weaken. (Japanese market rates will also move higher and will turn positive.)
However, if the Fed leaves rates unchanged, which is the most probable case, then all eyes will be on the BOJ. The BOJ will escalate NIRP, probably in conjunction with a US$ buying push. This will unleash much liquidity into safe haven assets.
We can see below that traders are already selling the Yen.
We can also see that gold futures are no longer attracting sellers.
The US$ is being pushed higher mostly because of weakness in the Euro and the GBP. European leaders are discussing a new investment program, which will need fresh money, while England prefers to renounce the European single market instead of allowing free circulation of persons. This has weakened both the Euro and the GBP.
The Gold ETFs
It is interesting to note the difference in behavior between the Total Effective Volume of GLD, which seems to be under accumulation, while the TEV pattern of PHYS points to selling.
The difference lies in the fact that PHYS is a closed end fund. This means that when demand is high, PHYS will sell at a premium to its Net Asset Value (basically the value of the gold bars in the fund.)
We can see below that the BOJ NIRP attracted buyers in PHYS.
This is the time when the PHYS/GLD ratio increased, indicating that PHYS started to sell at a premium. Around April, that premium was strong enough to attract arbitrageurs between PHYS and GLD (Funds shorting PHYS to buy GLD and make the risk free 1% premium difference.) This is why the Effective Volume patterns of both PHYS and CEF are not reliable: they measure a combination of real buying and selling combined with arbitrage transactions.
Note below that the Money Flow on the Precious Metals sector is very positive. This also indicates that large investors started buying gold miners ahead of the Fed and the BOJ's decisions.
Conclusions:
I believe there is a roughly 80% probability that gold will be higher by the end of next week.
