A brief introduction of the global foreign exchange market
As World War II (WWII) was approaching to its end in 1944, the US and its allies came together in the town of Bretton Woods in New Hampshire, USA, for an important conference, whose official name was the United Nations Monetary and Financial Conference. As this official name suggested, the 730 delegates representing all the 44 Allied nations in this gathering were trying to agree upon the global monetary and financial rules that were to be established right after the war. This important gathering would later come to be called the Bretton Woods Conference for the understandable reason, and the international financial system that it gave rise to in the aftermath of the war would be known as the Bretton Woods system.
As for the delegates, they made some highly significant decisions dining the conference. For example, they agreed upon the necessity of establishing two important international organizations after the war, both of which would play highly pivotal roles in the construction of the post-WWII global economic and financial architecture. These two institutions were the International Monetary Fund (IMF) and the International Bank for Reconstruction and Development (IBRD, also known as the World Bank).
If we focus on understanding the true meaning of the post-WWII economic and financial activities of the Allied nations (or you may choose to say the countries of the whole Western Bloc), it is clear that these nations were in search of an economically and financially intertwined world, where the economic benefits of countries would be interlinked. The main driver behind this motive of them for redesigning an economically and financially comiected world is very understandable from a classical liberal point of view: If the welfare of countries became mutually dependent, according to the classical thinkers, countries would never seek a war again in order not to jeopardize their own economic welfare.1 Based on this sort of a classical liberal perspective of ‘perpetual peace,’ boosting the free international trade was at the top of the to-do-list of the Western countries in the post-WWII world.
Once we approach the decisions taken at the Bretton Woods Conference from such an angle of peace and stability-seeking, we can see the complementarity between the decisions more easily. For example, the World Bank - as a bank for reconstruction and development, as its full official name suggested - was much needed because the war-tom infrastructure had to be fixed immediately in order to integrate the countries into the global production nets. Those countries would be able to trade with each other only if they have produced something tradable. If the logic behind the establishment of the World Bank is clear, let us now turn to the role of the IMF. First of all, it is not surprising to see that the IMF’s mandate was to facilitate the expansion and balanced growth of international trade.2 Obviously, the Western countries were tiying to minimize the risk of another episode of a destructive war in the future through weaving a robust cobweb of trade linkages among countries. One of the threats to the sustainability of international trade was the risk of financial collapse. Consider for a moment that countries A, B, and C are trading with each other and A is a net importer of B’s goods, while В is a net importer of C’s goods. What happens if A collapses for any reason, say due to the faulty decisions of the policymakers or simply due to an earthquake, etc.? A caimot pay for the goods it has purchased from В and so В may fail in honoring its payments to C. In a domino effect, the collapse of A spreads to other nations, and this may give world trade some bitter blows. The IMF, whose mandate was to make sure that the international trade would expand, was therefore established as a lender of last resort to countries like A in this example - i.e. to the countries that could spark a global meltdown of trade.
In addition to their decisions to establish the World Bank and the IMF, the Allied nations had also agreed upon an exchange rate standard whereby the value of the US Dollar would be fixed to gold and the values of the major currencies would be pegged to that of the US Dollar. This was a gold standard in disguise. Without any doubt, this sort of an internationally accepted fixed exchange rate regime had the capacity to serve as an effective tool to strip the exchange rate risk off the international transactions. Therefore, it should be seen probably as the most critical and essential step toward boosting free trade among Western countries.
Although the idea of achieving perpetual peace on earth using free international trade was an ingenious dream of the Western countries imminently following the end of WWII, some of the tools of this classical liberal perpetual peace idea lost their allure over the years. The gold standard of the Bretton Woods system, for instance, would stay in effect until 1971, when it was discarded by the US when the administration of President Richard M. Nixon, faced with economic and political hardships, decided to opt out and a set of leading nations, including the UK and France, hopped onto the same wagon with the US out of their desire to regain power in the design of their own domestic monetary policies.
After the dismissal of the Bretton Woods system of exchange rates, the world entered into a new phase. Exchange rates between an increasing number of currencies started to be determined freely by the interaction of the demand and supply forces in the market. Therefore, we can claim that this new era was, in fact, an era of floating exchange rates. In this new era, finding the tight specification of the economic/econometric models for better predicting the future exchange rates became an active field of research in the literature. From the early 1970s to 1983, economists had identified various macroeconomic variables that could be used to make exchange rate predictions until Rogoff and Meese showed in 1983 the uselessness of the various structural models - which were relying on these variables - against a driftless random walk.
Let us articulate what Meese and Rogoff did in 1983 and how they arrived at their gloomy finding, which is a finding that is still empirically valid after all those years, and which is a finding that also serves as the very cause for the existence of this book. In a seminal paper they wrote in the Journal of International Economics, Meese and Rogoff compared the out-of-sample forecasting powers of the six popular structural and time series models of exchange rates using the data for Dollar/Pound, Dollar/Gennan Mark, Dollar/Yen, and trade-weighted Dollar exchange rates. The structural models they employed were the flexible-price (Frenkel-Bilson) and sticky-price (Dombusch-Frankel) monetary models along with a sticky-price model augmented with the current account (Hooper-Morton).
The most general specification, covering all these three structural models, was as follows:
where the regressand, s, stood for the log value of the Dollar price of the foreign currency. As for the regressors, m - in' stood for the log differences in the money supplies between the US and the foreign country. By the way, the asterisk signs on top of some variables were to indicate that these were the foreign country observations of that variable. Similarly, у - у stood for the log value differences in the growth rates of the US and the foreign country, г - r* and 7re - ne were the differences in the non-log-transformed values of the short-run interest rates and the_expected inflation rates in the US and the foreign country, respectively. ТВ showed cumulated trade balance of the US and ТВ was its counterpart for the foreign country. The disturbance term of this stochastic model was u.
Eq. (1) was the most general specification combining all the three structural models since the model in (1) was boiling down to the Frenkel-Bilson model once the researcher set a4, a5, and a6 equal to zero. By the way, we will reproduce this model from scratch in the following chapter as we talk about the monetary model of exchange rates. Similarly, the researcher could set only a5 and a6 equal to zero and then Eq. (1) would become the Dombusch-Frankel model. Eq. (1) itself, with none of the coefficients constrained to be zero, is the Hooper-Morton model.
Apart from these three structural models, Meese and Rogoff tested univariate and multivariate time series models as well. They employed a long AR model where the longest lag (M) they considered was based on a function of sample size, i.e. M = N/logN. As the multivariate model, they employed the AR terms along with lagged values of the structural variables in Eq. (1). Using the mean squared prediction error (MSPE) and mean absolute error (MAE) from the out-of-sample forecasts as the model selection criteria, Meese and Rogoff chose a driftless random walk model as the best predictor above all the structural and time seiies models they tried. That was a shocking finding simply because random walk models caimot be predicted. In order to understand why it is impossible to predict the random walk processes, let us look at the driftless random walk model that was used in the paper of Meese and Rogoff:
where sf+1 and s( are the exchange rates at time /+ 1 and t, respectively, and el+ x is nothing but a white noise error term with a central value of zero and time-invariant variance (i.e. st+1~ [0, o}). If we make a prediction at time t for the exchange rate that is likely to occur at tune t + 1, we encounter the following absurd result.
In the preceding notation, Е,(г,+1) becomes equal to zero and goes away since we assumed et + x ~ (0, o), i.e. economic agents - either because they form their expectations rationally or adaptively - do not make systematic mistakes. Er (st) is equal to st because sr is known at time t and therefore we do not need to expect a value for it - it is given. In sum, E,(s,) + E,(e(+]) = s,+0=E,(s,+1) and that is why the expected value of tomorrow’s exchange rate is equal to the current exchange rate. In sum, the best prediction is to take the last closing value of the exchange rate and use it as our best estimate for the future. As if this is not disappointing enough, there are other problems of using a random walk model for predictions. Remember that we said Et (г(+1) becomes equal to zero and goes away. But this is the expected value of the error. In reality, however, the error can be a negative or a positive value, as well (after all it is centrally distributed around zero with the non-zero variation of a in both directions); the random walk only tells us that the most likely value of the exchange rate tomorrow is the current value of the exchange rate, but there is some non-negligible chance that the realized value of the exchange rate tomorrow can exceed the current value and there is also some non-negligible chance that the realized value of the exchange rate tomorrow can be lower than the current value. There is no useful information in this prediction, as you see. Anyone could tell the same outcome without bothering herself with mathematical models. There is simply (and sadly) no new information in saying that tomorrow the exchange rate is either going to be higher than today’s exchange rate or be equal to it or be lower than it; these are the all possibilities, anyway.
To wrap up the discussion so far, the global foreign exchange (FX) market is huge. The daily cross-border turnover on average in this market is estimated to linger around $6.6 trillion by the Bank for International Settlements (BIS). That figure is more than enough to give currencies the top spot as the most widely traded asset class in the global financial markets. Yet there exists no known economic/econometric model that one could rely on in order to predict the future movements of exchange rates. As you can tell, this is analogous to saying that we do not have any reliable asset-pricing model for the most popular financial asset in the world.
- 1 Democracy and free trade, according to classical liberals, would reduce the incidence of wars (see O’neal et al. (1996) for further discussion and testing of that view), www.jstor.org/stable/42513l?seq=l
- 2 For better insight about the objectives of the International Monetary Fund, see the following link: www.imf.org/en/About/Factsheets/The-IMF-and-the-World- Trade-Organization