Nick dewhirst takes issue with ALAN GREENSPAN"S ANALYSIS OF MODELS PROJECTING MOVEMENTS IN EXCHANGE RATES
After falling a third against the euro over the last three years, the big question is, can the dollar fall much further? If so, can anyone predict when the collapse will end because, according to Alan Greenspan, to his knowledge no model projecting movements in exchange rates is superior to tossing a coin.
As a model builder, fortunately I can confirm that at least one set of models did expose the dollar"s over-valuation. In November 2001, I wrote that it was the most overvalued of 29 currencies researched in our Forex RouteMap and predicted that the dollar would be the next bubble to burst. After a couple of false starts, buy signals were generated for sterling in April, and for the euro and yen in July that year.
Based on calculations of real effective exchange rates the dollar had scope for a 20% decline on average before reaching the low end of its historic trading range. In practice, it has now reached that point by falling 12% against the yen, 25% against sterling and 32% against the euro.
Was that just luck? After all, given enough predictions, some are bound to be right and a few may even have generated the right answers for the right reasons. Even then, it is possible that the right reasons in the past may no longer be the right reasons in the future.
Unfortunately, speculation on this issue is unavoidable, because it is a vital ingredient in so many important investment decisions. Global asset allocation cannot ignore half of world stock market capitalisation. Projections of company profits elsewhere cannot ignore the world"s largest source of overseas earnings, nor pressures on export margins when competing against dollar-based rivals. Similarly, participants in bond markets cannot dismiss the effects on global supply of the world"s biggest issuer of government bonds.
Its effects extend into the most esoteric corners of the investment scene. Even the outlook for emerging market bonds is affected because the vast majority is denominated in US dollars. Competition between European and Japanese manufacturers in third markets is affected by the differing degrees to which their currencies rise against the dollar.
Hedging is a realistic possibility for sophisticated institutional investors but the size of futures contracts puts that beyond the reach of smaller investors and there is a serious regulatory risk to advisers dealing in futures for anyone other than expert investors, and that does not necessarily include pension fund trustees.
Predicting currency movements is like monitoring tectonic plates. Pressure builds up as a succession of minor factors accumulate. Eventually, just one tiny change is enough to make the plates move, causing an earthquake. Therefore countries can run current account deficits for years, provided there are offsetting capital account surpluses. Traditionally, this was the case for emerging markets, where capital goods imports could be funded on current account by portfolio and direct investment on capital account. In this case, imports of Asian consumer goods are funded by Asian savings in dollar-denominated financial instruments.
This can continue for so many years that persistent doom-mongers are discredited, so few listen to them when they are finally proved right. The US had run current account deficits over 4% of GDP for four years, before the dollar started to decline this time. In the previous dollar bubble of 1984, the US was similarly in deficit for a fourth year before the exchange rate began to fall. Elsewhere, Mexico had run deficits over 6% for three years before its crisis in 1994 and Thailand had operated deficits between 5% and 9% for seven years before its crisis in 1997.
However, once the break occurs, it is often so dramatic that it exceeds even the pessimist"s expectations. The Mexican peso and Thai baht both halved within a year. Last time round the dollar fell 30% on average in three years. So far it has only fallen 16% in two years.
There are therefore ample precedents to suggest that the dollar could fall much further. Indeed, as the starting point is a bigger deficit running for longer than last time round, there are good reasons to fear the worst. Furthermore, this time round, the world has an alternative reserve currency now waiting in the wings, namely the euro. It would not surprise me if the dollar fell even more than my extremely bearish prediction in 2001.
Then I wrote that the J-curve would come back to haunt the US. That is to say, the first effect on the balance of payments of a currency movement is perverse, because of the differential lag in price and volume responses for both imports and exports. As shown in the chart on the left, exports are now growing faster than imports but general expectations, according to Consensus Economics, are that the trade deficit will still be almost $600bn, or about 5% of GDP, next year. Such lags produce over-shooting for foreign exchange rates.
Step changes are difficult to model, so the best way of handling them is with a combination of valuation ratios and trend triggers. For reasons discussed below, our Forex RouteMap uses four of each. Signals have changed from bearish to mixed, but they are not yet bullish for the dollar.
Our experience is that when currency models are good, they are very, very good, but when they are bad, they are horrid. There are periods of many years where the models accurately describe the movements of the dollar, but these are interspersed with phases of a year or more, when the currency does precisely the opposite of what the model projects. There have been six such phases in the last quarter century. Last year was horrid.
Theoretically, the problem is that correlations fluctuate. Practically, the problem is that different players dominate markets at different times. In the case of currencies, three players interact in four different asset classes:
&149; Direct investment driven by corporations
&149; Short-term money markets driven by central banks and corporations
&149; Long-term bond markets driven by portfolio investors and central banks
&149; Equity markets driven by portfolio investors and corporations.
The chart on the right illustrates the varying significance of different variables by taking long-term bonds and equities as an example. The two series show the correlation between movements in the dollar"s effective exchange rate against a basket of currencies and the relative performance of US long-term government bonds on the one hand and US equities compared to a weighted index of world markets on the other hand.
Sometimes the dollar responds better to bond markets but at other times it reflects equity market pressures. Sometimes the two act together but at other times they act in opposite ways. Sometimes neither is dominant, while other factors take precedence, as different players dominate the market. In the present circumstances, that translates into a very weak dollar compared with European currencies, where portfolio flows are now dominant, but a more stable rate against Asian currencies, where central banks dominate.
The only stable relationship is an inverse correlation between the significance of equity markets and the effectiveness of our macro-economic model. Thus poor performance of our, and many other models in 2003, is directly related to the high significance of equity markets. What is therefore needed is a model of equity markets, which is where our Shares RouteMap comes in, but that is another subject.
Is Greenspan right? No, if he means currency movements are random, but yes if he means no single model works all the time. That is why the odds can best be improved by using a model of models approach to currency forecasting.
Despite improved risk appetite
FOS award limit increase
Relates to 136 million transaction reports
Ceremony will take place 13 November