In Maclean’s, Stephen Gordon assures you that there is no mastermind at work, determining what happens to the Canadian dollar:
The Canadian dollar fell from 97 cents US to below 89 cents US in the weeks following the Bank of Canada’s decision to shift its monetary policy stance away from a tightening bias. (It has recently rebounded to hold steady at around 91 cents as I write.) These developments have provided additional fodder for those pundits who are in the habit of offering their views about where the dollar should go and/or where it will go (the two are separate issues). These views fill up media space, but they shouldn’t be taken too seriously. The foreign exchange market is one where the “semi-strong“ form of the Efficient Market Hypothesis holds: movements in exchange rates cannot be predicted using publicly-available information.
If everyone really believed that the Canadian dollar will end up at (say) 85 US cents, then everyone would sell CAD at its current price to buy USD, wait for the price of USD to increase – which is the same thing as waiting for the CAD to depreciate – and then sell at the higher price. But if everyone does that, the CAD would be bid down to the point where it is no longer profitable: 85 cents. This is why you should take predictions about foreign exchange movements with a grain of salt: if you could actually predict them, the last thing you’d do is tell anyone.
This doesn’t mean that exchange rate movements are completely random: some of the fluctuations can be ascribed to variations in the ‘fundamentals’. But what really drives these movements are the unexpected changes in the fundamentals. And unexpected changes are, by definition, unpredictable. The most reliable forecasting model is a random walk: the exchange rate next period is the current exchange rate plus a white noise error term. The best prediction for where the exchange rate is going is where it is now.