Exchange Rate Forecasting

1
Exchange Rate Forecasting
6 ?
2
Overview

• Resolving Controversies in Exchange Rate
  Forecasting
• The Forecasting Approach and the Market Setting
• Forecast Performance Evaluation Accurate versus
  Useful Forecasts
• Assessing the Economic Value of Currency Forecasts

3
Overview

• Forecasting Methods Some Specific Examples
• Short-Run Forecasts Trends versus Random Walk
• Long-Run Forecasts Reversion to the Mean?
• Composite Forecasts Theory and Examples
• Policy Issues and Special Forecasting Problems
• Consumers of Exchange Rate Forecasts
• Producers of Exchange Rate Forecasts
• Special Problems in Exchange Rate Forecasting

4
A Framework forForecasting Exchange Rates
5
The Forecasting Approachand the Market Setting

• Under a pegged exchange rate regime, the exchange
  rate between two countries is allowed to vary
  only within narrow bands.
• Once the exchange rate becomes misaligned,
  economic models may assist with the direction and
  magnitude of a likely rate change.
• However, these models offer little guidance about
  the timing of change since the decision to adjust
  the peg is typically a political decision.

6
The Forecasting Approachand the Market Setting

• Under a floating exchange rate system, the
  exchange rate is free to adjust in response to
  changing relative macroeconomic conditions.
• Forecasting in such an environment is easier when
  the market is efficient, but forecasting for
  profit can be more difficult for the same reason.
• Forecasting under hybrid exchange rate systems
  can reflect elements of both the pegged and
  floating process.

7
The Forecasting Approachand the Market Setting

• Another important element in selecting a
  forecasting approach is the forecast horizon.
• Empirical evidence indicates that market
  participants place greater reliance on technical
  models for their very short-run forecasts.
• As the forecast horizon lengthens, they turn to
  fundamentals.
• In the middle range of horizons, composite
  forecasts may be used.

8
The Forecasting Approachand the Market Setting

• Forecasters must also make a distinction between
  real and nominal exchange rates.
• With relatively constant price levels in the
  short run, changes in the two rates are similar.
  However, in the longer run, they may take
  substantially different courses.
• Unlike the real exchange rate, the nominal rate
  has no affinity for a mean value. However, the
  real rate may be difficult to predict in the
  short run.

9
Forecast Performance EvaluationAccurate versus
Useful Forecasts

• The traditional econometric approach begins with
  the forecast error made at time t

10
Forecast Performance EvaluationAccurate versus
Useful Forecasts

• It is important to distinguish between accurate
  forecasts and useful forecasts.
• Accurate forecasts have small forecasting errors
  gauged by traditional statistical measures, while
  useful forecasts are those on the right side of
  the market, leading to profitable speculative
  positions and correct hedging decisions.
• A forecast may be relatively more accurate, yet
  not useful (since it leads to incorrect
  decisions).

11
Forecast Performance EvaluationAccurate versus
Useful Forecasts

• In the absence of a currency risk premium, the
  right side of the market implies the right
  side of the forward rate.

12
Forecast Performance EvaluationAccurate versus
Useful Forecasts

• To measure usefulness, let

• Then, the test for usefulness is

• According to the binomial distribution

13
Assessing the Economic Value ofCurrency Forecasts

• The links among accurate forecasts, useful
  forecasts, and economic value are ambiguous.
• An accurate forecast on the wrong side of the
  forward rate has no economic value.
• But even a statistically significant track record
  of correct forecasts may be insufficient to
  produce profitable results.
• The users of a forecast may also be more
  interested in the forecasts marginal value,
  relative to the other forecasting models used.

14
Forecasting MethodsSome Specific Examples

• Short-Run Forecasts Trends versus Random Walk
• Many economists have concluded that a random-walk
  forecast is the best short-term formulation.
• However, trend-following models appear to earn
  profits from exploiting patterns in spot rates.
• This may be because the underlying relationships
  are nonlinear. Technical models do not impose a
  linear relationship, which is assumed by linear
  regression models.

15
Forecasting MethodsSome Specific Examples

• Long-Run Forecasts Reversion to the Mean?
• The likelihood that long-term foreign exchange
  rate forecasting will be successful is enhanced
  when an economic model acts as an anchor to reign
  in unexpected short-term changes.
• Overall, empirical evidence supports a reversion
  to the equilibrium exchange rate implied by
  macroeconomic fundamentals.

16
Forecasting MethodsSome Specific Examples

• Composite Forecasts Theory and Examples
• A composite forecast brings together the
  information in alternative forecasting models so
  as to outperform the individual forecasts.
• To draw information from the available pool of n
  forecasts, a variety of alternative weighting
  systems are possible
• weight w 1/n arithmetic
  average
• heavier
  weight for more
  accurate forecasts

17
Forecasting MethodsSome Specific Examples

• choose wi to minimize the average forecast error,
  conditional on the standard deviation, using
  linear regression.

18
Policy Issues andSpecial Forecasting Problems

• The consumers of exchange rate forecasts face a
  number of challenging policy issues.
• They need to decide which forecasting method to
  follow or which foreign exchange advisory service
  forecasts to purchase.
• Then, they face the problem of integrating the
  currency forecasts into their business decisions.

19
Policy Issues andSpecial Forecasting Problems

• Like the consumers, the producers of exchange
  rate forecasts need to first settle on the
  desired forecasting horizon before selecting a
  forecasting approach.
• Then, they face the problem of operating in a
  competitive industry.
• They also need to solve various econometric
  problems, such as revising their forecasting
  model as time passes.

20
Policy Issues andSpecial Forecasting Problems

• Special problems can also arise when structural
  breaks interrupt the stability and stationarity
  of an exchange rate series.
• Examples of such structural breaks include
• exchange rate system changes
• monetary policy changes
• the unification of East and West Germany in 1992
• changes in consumer spending patterns