I've been bullish on the dollar for some time now.
My view was based on the idea that, with the Federal Reserve getting ready to taper off its quantitative easing, while other central banks are still determined to keep their extraordinary policies in place, the dollar was going to benefit from the divergence in monetary policy.
Unfortunately, it hasn't worked out quite like I had expected. On the contrary, the dollar has been weakening against its G10 counterparts as investors revised down their forecasts for future U.S. rates.
This was happening even before Laurence Summers surprised the market by withdrawing from the race for the Fed Chairmanship, leaving the dovish Janet Yellen in first place. That's just made investors think that U.S. rates will remain low for even longer than they would have under a Summers Fed.
At the same time, tightening expectations for several other countries have been revised up, and I'd have to say that's where I've been most surprised.
The divergence of monetary policy that I had expected is turning into convergence, which is eroding the dollar's relative advantage and causing it to weaken. Perhaps the best investment strategy in such times is to avoid the dollar entirely and look for trades in the cross rates where policy divergence still seems likely.
(Read more: Questions over Fed chief a bigger issue than taper)
To begin with, the introduction of "forward guidance" in the UK has failed to change expectations so far.
While the Bank of England tried to convince investors that interest rates would stay low for longer than the market was discounting, they wound up persuading them of just the opposite.
The graph below shows that the forecast for three-month rates in two years is now higher than the peak before forward guidance was introduced. That's largely due to a lot of better-than-expected economic news from the UK, as shown by the sharp rise in the Citi economic surprise index.
The Bank of England made its "forward guidance" conditional on the state of the economy, particularly the unemployment rate coming down to 7.0 percent.
The sudden improvement in the data has made people think that the economy may meet the Bank's unemployment target sooner than they had thought and thus interest rates will rise sooner than they had expected, too. The clincher on this was the surprising fall in the U.K. unemployment rate in July.
Following Norway's higher-than-expected inflation figures for August, there was a change in view there, too.
Norges Bank had hinted at further easing this year, but the data turned expectations around; now they could even adopt a tightening bias or at least start talking about higher rates at their Sep. 19th policy meeting.
That possibility has sent the Norwegian krone sharply higher against both the dollar and the euro (although reports of a slowdown from Norges Bank's regional survey on Friday dampened rate hike expectations.)
Finally, just last Thursday, the Reserve Bank of New Zealand actually did change its bias. While the Bank left the Official Cash Rate (OCR) unchanged at 2.5 percent, as was generally expected, and repeated its prediction that rates would remain unchanged for the remainder of this year, it added that "OCR increases will likely be required next year."
This is the first G10 central bank to warn explicitly of rate increases. The market now expects a rate hike from New Zealand sometime in the second quarter, whereas previously it was expected sometime in the second half of the year.
(Read more: Fed may drag bank stocks down in 2013)
Meanwhile, over the last few weeks the market has actually been reducing its U.S. interest rate forecasts due to weaker-than-expected economic data, particularly the August non-farm payroll figure.
Changing expectations for interest rates in other countries are what's driving the FX market now. Using the June 2016 futures contracts for three-month funds, we can calculate what the market expects for interest rates two years from now.
This graph shows the change in expectations for each currency's 3m rates in June 2016 on the Y axis (not the actual rates, but rather the change in the market's forecast for the rates) over the last two weeks.
Against that, on the X axis, is the change in the dollar's value against each currency.
As you can see, the two line up almost in a straight line: the change in interest rate expectations explains almost all of the change in the dollar's value against the G10 currencies over the last two weeks.
I'm hesitant to come out with any conclusion about the dollar the day before the FOMC meeting starts. As this meeting will probably decide to begin "tapering off" the Fed's quantitative easing, it's a major event that can blow up anyone's forecasts.
Because of that, perhaps it'd be better to avoid the dollar entirely until this decision is out of the way and the impact of the Fed's decision, whatever it is, is clearer.
Instead, this graph gives us some guidance on trades that don't involve the dollar. It suggests going buying the New Zealand dollar or the Australian dollar – between the two, I prefer the kiwi dollar because, despite the sharp rise in interest rate expectations in Australia, the weak August labor market report has left investors undecided about a future rate hike (the probability priced in for the next 12 months is just 58 percent, vs 90 percent for New Zealand, and it's still assuming that another rate cut is indeed possible in Australia.)
Short AUD/NZD would be the way to play that expected divergence in monetary policy.
(Read more: Summers for Fed chief report moves dollar)
Note that the British pound did not gain as much against the dollar as the change in interest rate expectations would have implied; that may indicate the rally in the pound is starting to lose momentum. Nonetheless, we would probably have to see a turn in the economic indicators before we see a turn in the currency.
While the Norwegian krone (not shown on this graph) was a poor performer last week because of Norges Bank's regional survey, from my point of view that just presents a better entry point as the above-target level of inflation will make it difficult for Norges Bank not to change its interest rate expectations at this week's meeting.
(Read more: Best bond play for taper angst?)
Against these longs, I believe the Japanese yen continues to be the most appropriate funding currency, as the Bank of Japan is likely to be the last major central bank to change policy.
The European Central Bank is also continuing to discuss ways of easing and euro zone growth is likely to disappoint, in my view, so I would also consider the euro on the short side.