Interestingly enough, I think it would be a safe bet that one could find professors teaching undergraduate economics who will argue that currency values will follow simple "all else equal" rules in one lecture, and then go to another lecture and state that some version of market efficiency holds. Consistency is the hobgoblin of small minds, &c.
One could attempt to salvage "all else equal" arguments by stating that they are the sensitivities of some currency valuation model. The problem with that argument is straightforward: this model would have to do a convincing job of forecasting currency values. Someone in possession of said model would be able to generate a rather impressive track record doing directional currency trades. I do not claim to be in close contact with the forex community, but I saw little evidence of this happening.
Why are Currencies Hard to Model?A foreign exchange rate is a relative price: at what ratio will entities be willing to exchange one currency for another? At the minimum, we need to keep an eye on what is happening in those two currency areas, but we might also need to worry about events in other currency areas. For example, an emerging market crisis is often associated with a rise in the U.S. dollar, which might rise versus the British pound -- even though the U.K. and U.S. economies might not be measurably effected by the turmoil elsewhere.
The following entities trade foreign exchange.
- Firms and households that engage in foreign trade in goods, services, and tourism.
- Investors with international portfolios.
- Foreign exchange speculators.
In principle, the last group (speculators) are there to balance the flows among the previous two groups, but speculation can become the tail that wags the dog (for some time at least; speculator balance sheets and risk limits are finite). And as countries like the U.K. discovered during the Bretton Woods era, trading firms can engage in covert currency speculation just through the mechanisms of trade finance ("leads and lags").
If we assume that speculators cannot drive currency values too far away from fair value -- a staple of analysis in bond, equity, and derivatives markets -- we are stuck with the reality that there are two candidate values for fair value. The first is the level of the currency that balances trade flows. The second is the value that balances portfolio flows. The reality is that the market value of international portfolios dwarfs net trade flows -- so portfolio flows are dominant. This changes the currency valuation problem into the problem of the relative valuation of all financial assets. This is extremely complex, and so instead currency strategists do things like run carry portfolios or fool around with volatility strategies.
Post-Script: Bond Valuations?The rates market (e.g., not looking at credit spreads) is much easier to deal with than the foreign exchange market. Bond valuation is driven by the cost of financing a long position, and for a government bond, that financing cost is typically close to the policy rate. Meanwhile, central bankers follow the economic cycle with a lag, in a fairly predictable fashion (e.g., the Taylor Rule).
The problem is that forecasting the economy is hard (the underlying message of Recessions: Volume I). As such, it is hard to outperform the market. Roughly speaking, the only easy money in directional rate trading is finding a critical mass of bond investors who do not believe that rate expectations matter for valuation, and take the other side of the trade. However, natural selection eventually closes that avenue for outperformance.
(c) Brian Romanchuk 2020