When Richard K. Lyons was working at Columbia University more than 10 years ago, a friend who was a foreign-currency trader invited him to spend a day on the floor at Merrill Lynch in New York. Lyons spent hours sitting next to his friend, who was trading millions of dollars at a time. At one point Lyons' friend asked, "Do you think I should be long or short on this one?" Lyons, a PhD in economics from the Massachusetts Institute of Technology, didn't have a clue.
The question gnawed at him. Lyons began exploring the topic and ended up writing a book on it, the result of his research into his friend's question. Now a 45-year-old professor of finance and economics at the Haas School of Business at UC-Berkeley, Lyons published The Microstructure Approach to Exchange Rates in March, 2006. The book makes a strong case for using trading activity to understand what will happen to foreign currency.
Previous currency theorists, in contrast, have looked mainly at big-picture economic indicators such as interest rates, inflation, or the trade balance. They forecast currencies by assuming the investors participating all have similar goals and the same access to information. When they're making forecasts, many strategists don't care whether a big order came from a hedge fund on a Tuesday. Instead, they look at factors such as the economic strength of the country using a given currency.
BUFFET'S BOO-BOO. Warren Buffet is a case in point. The lauded investor and CEO of Berkshire Hathaway warned about the U.S. trade balance and began investing billions against the dollar in recent years. But then the dollar's value rose to 90.96 on Dec. 30, 2005, from 81 on Dec. 31 2004, according to the New York Board of Trade's U.S. Dollar Index (DX). Buffet noted in his 2005 Chairman's Letter that his view on America's long-term trade imbalance remains unchanged, but his conviction cost Berkshire $955 million pre-tax that year. (Buffet did point out in his letter that Berkshire is $2 billion in the black since first entering into currency contracts.)
Buffet ended up reducing his direct position in currencies "somewhat" in 2005 by buying foreign-currency-denominated stocks instead. Now the dollar index has dropped to around 85.47 as of May 4, from 88.81 on Jan. 31, 2006. (Warren Buffet's spokesman didn't respond to a request for an interview by press time.)
Lyons thinks that enough speculators bet on rising U.S. interest rates to boost the value of the dollar last year. "You couldn't know the importance of that interest-rate effect unless you were watching these big institutions" trade, Lyons says. "That's something you, I, retail investors, and Warren Buffet don't know. But the big banks get a glimpse." Bank dealers see some of those trades when they do them with their clients.
Meanwhile, retail investors are paying increasing attention to foreign-currency markets in recent years, as concern grows about the U.S. economy. BusinessWeek Online reporter Sonja Ryst interviewed Lyons to learn more about how people should think about these currencies.
What is the traditional method of predicting currency values?
The traditional way of predicting value is based on macroeconomic variables, like the growth rate of the whole economy, the inflation rate, interest rates, or the trade balance (imports and exports). Those variables, in most of our standard textbook models, are supposed to explain exchange rates, and they're the sort that people use to predict currency values. But they don't do well. They do abysmally poorly.... The textbook models don't really open the door for trading itself to have an important role on the exchange rate.
How about for longer-term forecasts?
[Variables like interest rates or inflation] can help explain what's going on two or three or five years out. But not for trying to forecast or explain what happens in the next month or so.
What new strategy do you suggest for predicting currency values?
A very important part of the information in the market is the amount of money coming from hedge funds, nonfinancial corporations, and mutual funds. Their activity is important for understanding not just how prices move today, but also how they might move in the future. But information about these transactions is not publicly available.
Since the information about currency-trading activity isn't readily available, how does one get hold of it?
One of the fun things is that electronic trading has created an archive of data for us to analyze. Fifteen or 20 years ago, when trades were done over the phone, we had no record about what was happening. Now we can see who traded with whom and whether it moved the market.... If you're a bank, then you get to see the trades because your clients come to you and do them through you. When you put that information together at the end of the day, you learn something about where the market might be going.
How much do the financial institutions actually know? Aren't they also making guesses in the dark?
Everybody is getting pieces of information about trading activity and doing the best they can. The information gets valuable when you're one of the big players in the market, like UBS or Citibank (C) or JP Morgan Chase. Even then, they're only seeing about 10% or 15% of what happens in the foreign-exchange market. But when you see that large a fraction of what nobody else can see, how valuable is that?
Why doesn't the general public have access to this information?
It's public in stock markets because there are regulatory requirements for disclosing the price and quantity whenever you do a trade. But in the foreign-exchange market, there's no regulatory requirement in any country. If the U.S. said, "Okay, anyone who does a trade in New York has to report information about it," then any market participant that doesn't want to report can simply switch and trade in Toronto, Mexico, London, or anywhere else instead. Trading would migrate to places that don't have restrictions, so no single country is willing to step out and say to the market, "We'll force you to do something you don't want to do."
Do you have any advice for the average individual investor?
You have a safer portfolio if you have some of it invested internationally, because then you're doing a better job of managing your risk and spreading out your exposure to different parts of the world. But if you want to make money on predicting where the markets will go, invest outside the U.S. now and hold the investment for a long period of time -- like two or three years. It's a sensible idea to take advantage of what is likely to be a decline in the dollar over a long-term horizon.