Since about the Spring of 1988, the divergences among the major currencies have been at least as striking as their movements against the dollar. In fact in the case of the Yen, the movement against the European currencies was much more pronounced than against the dollar. Why? What’s going on?

In a word, trend-following. In the first place, the dollar in the late 1980s was perceived to have reached a sort of equilibrium level, where observers by and large did not feel it to be notably overvalued or undervalued. The US trade gap was more or less stable; and real US interest rates seemed to be roughly in line with those elsewhere. Anyway, few observers any longer

had any confidence in their ability to predict the course of the dollar on the basis of the “fundamentals”*; and without the benefit of hindsight, its trending had become unreliable. So they were happy to throw in the towel and try a different game. The new game was spotting trends in the cross rates. For instance, when

the rise in sterling against the D-Mark got overdone and it began to fall back from its peak of DM 3.27, Lo! you had a trend: for sterling had been falling against the DM ever since the War, after all. And when the DM moved up out of its decade-long trading range of 8O-85c against the Swiss franc, you had another trend, as it moved first to 90c and then to 92c. The same thing happened when the Yen began giving up some of its long-run gains against the D-Mark, and the DM rose with little respite from near 70 Yen to 95 Yen.

These trend movements, with the possible exception of the Yen, have been defined with hindsight. In real time, what you see is a price movement, giving way to a fluctuation, which in turn gives way to either a continuation in the price movement (a trend?), or a reversal to the opposite direction (a new trend). Brokers and bank dealing desks of course love clients to deal in the crosses*: they generate more income in commissions and spreads. But is there any way of forecasting them?

We have two bare decades of continuous experience of floating currencies. Until recently, practically all currency trading was conducted through the dollar. That meant that if you wanted to buy D-Marks for pounds –or for SF or whatever –the rate was calculated via the dollar (and so were buying– and-selling spreads). Recently this has been changing, as more banks have begun to deal the major cross-rates direct, without reference to the buck. The trend got further impetus when the IMM decided to trade the D-Mark crosses (Ј in DM and DM in Yen & SF) starting 29 May 1991. The change has probably contributed to recent volatility in the cross rates. This volatility could be here to stay, so there may be a greater need to forecast the crosses. Is it possible?

The answer is, I suspect, that good price-based traders have had, and may still have, a window of opportunity with the cross-rates, because they’re relatively untrodden ground – though maybe not for long. For others, the opportunities for forecasting significant cross-rate moves are scarce.

The dollar parities offer more reliable pickings, I think. Over the years, we have been able to monitor the relationships between the dollar and such things as interest rates, inflation, trade, politics, money supply, economic growth and the other fundamental data. We have also been able to watch the interaction between currency values and other data which are more concerned with the psychology of the participants –the indicators of sentiment.

Watching these relationships over the years, one has seen the false prophets among them hog the limelight; observed the quiet workings of

the valid relationships; noted currency experts being hoodwinked; seen the well-meaning, plodding efforts of the media (I was once a newspaper columnist, so I can empathise). Over time I have come to believe that probabilities can be found in the dollar parities which cannot be found in stock markets or bond markets. I think the currency markets are in fact easier to forecast.

So which are the valid relationships, the ones that have worked reliably over time? And which are the frauds? As with all such questions, you need to throw out the preconceptions and start with a fresh slate. “Go back to before ABC”, as Levi Strauss or someone said.

At any moment, the price of a currency represents equilibrium between buying and selling pressure. What causes a price to change is a change in pressure. Change is the first factor to isolate. But of course pressure can change daily – even hourly or every minute –and weekly and monthly and yearly. We’re lost unless we know what time frame we’re dealing with. To get a line on that we can look at a few price histories and remind ourselves that there have certainly been price swings that have been measured in years, punctuated by counter-moves measured in months. In other words, we must be on the look-out for relationships that are valid over multi-year spans and others which are valid for multi-month spans –and maybe others valid for weeks or just days.

In real life, these divisions are very elusive if not illusory –existing as they do only with hindsight. In real life, we live in a continuous present where days turn into weeks, and weeks turn into months and so on. The trick which helps us cope with this situation is to ask “what is the underlying trend?”. If you can diagnose such a thing and find a valid underlying rationale for it, then you can divide daily and weekly and monthly fluctuations into l) those that are in line with the trend and 2) those that are counter-trend moves. This simple ‘with-or-against’ distinction helps keep you on your time-frame track.

So what constitutes a “valid underlying rationale” for a price move? In two words, expected return: or more specifically expected total return* over the foreseeable time-horizon. If you go right back to before ABC, there can be no other driving force behind a price move than a change in expectations of total return. So the factors that enter into total return will be the forces that drive supply and demand for currencies via changing expectations. The quantifiable ones are interest yields* and inflation: just those two, simple as that. The unquantifiable one, the future exchange rate,

sellers, what else is new? Well, there’s nothing new under the sun, as the prophet said. But here goes. Following from the plodding analysis above, there are a couple of positive conclusions; quite a lot of grey neutral territory; and a whole lot of cumbersome baggage we can assign to the trash can. Let’s start with the baggage.

What about trade and money supply and budget deficits and employment and economic growth? And what about intervention and political events like German monetary union? And what about capital flows? Well where possible, I have tracked these factors and found out, empirically, whether they do have any reliable relationships with currencies. Some are charted in this chapter and others elsewhere in the book, and you can see with your own eyes that most of the supposed relationships are myths. And remember that a supposed relationship which

reliable is absolutely useless – except that it can sometimes be used as a contrarian indicator: if you know the crowd is following a false trail, it can give you the confidence to go against the crowd at the right moment. The point about the above factors is that they have no direct bearing on actual total returns, yet they are reflected or discounted* in currency rates. They produce deviations from the norm – excesses or counter-trend moves – which must be ignored or countered, if we are not to be blown off course. They are in fact just noise.