Exchange Rate Forecasting

1
Exchange Rate Forecasting

• No haruspex could look at his fellow without
  laughing.
• Marcus Porcius Cato (234-149 B.C.)

2
Roman Fortune Tellers

• haruspex look at organs (haru organs, spex
  look at)
• augur, augury, auspices (look at birds).
• inauguration, exauguration
• Pontifex Maximus (P.M.)

3
Three puzzles in exchange rate economics.

• 1. The determination puzzle.
• 2. The excess volatility puzzle.
• 3. The clustering puzzle.

4
1. The determination puzzle

• Exchange rate movements are virtually unrelated
  to economic fundamentals (at least over periods
  of less than two years).

5
The Dollar/DM Exchange Rate and Relative
U.S./German Price Levels, 1964-1997
6
2. The excess volatility puzzle

• Exchange rates are excessively volatile
  relative to economic fundamentals.

7
Month-to-Month Variablility of the Dollar/DM
Exchange Rate and of the U.S./German Price level
Ratio, 1974-1998
8
3. The clustering puzzle

• Large swings in exchange rates are normally
  followed by large swings in either direction.

9
I.   How to forecast exchange rate with
fundamental models.

• Procedure
• a. Specify exchange rate model
• b. Fit the model with historical
• data.
• c. Evaluate the performance of the
• model with out of sample
• forecast

10
I.   How to forecast exchange rate with
fundamental models.(?)

• e.g. Monetary Approach
• ??M s?M s ?i? i ?Y?Y ?????????(????)??????
  ????????
• ?? Monetary Approach
• Et
• ??,Et f ( Mt s?Mt s ?it? it ?Yt?Yt)
• () (-)
  () (-) (-) ()

11
I.   How to forecast exchange rate with
fundamental models.(?)

• ??????????
• ln Et a b ( lnMt s lnMt s ) c ( it
  it ) d ( lnYt lnYt )
• Monthly Data
• 1 t
  T
• Estimate with these data,
• ???????
• a 0.8 b 0.3 c 0.2 d -0.4 R2 0.95

12
I.   How to forecast exchange rate with
fundamental models.(?)

• ??????? (1)?????(2)?
• lnEt 0.8 0.3 ( lnMt s lnMt s)
• 0.2 ( it it ) 0.4(lnYt lnYt )

13
I.   How to forecast exchange rate with
fundamental models.(?)

• Evaluate
• ? t1 ? M s?M s ?i? i ?Y?Y ??(2)?
• ??
• Et1 (forecasted exchange rate of t1 )
• ??,
• ? t2 ?M s?M s?i? i ?Y?Y ??(2)?
• ??
• Et2 (forecasted exchange rate of t2 )
• ???

14
I.   How to forecast exchange rate with
fundamental models.(?)

• ????
• ?????(Mean Squared Errors, MSE)
• MSE

?
n
1
(Et1- Et1)
2
n
i 1
Actual Exchange Rate
Forecasted Exchange Rate
15
I.   How to forecast exchange rate with
fundamental models.(?)??

• ???(Ms?M s ?i) ??. ??????????,???????????????
• ????.?????????????(??????).
• ??????????.
• ???? a. Exchange rates are excessively volatile
  relative to economic fundamentals. b. Large
  swings in exchange rates are usually followed by
  large swings in either direction.

16
Forecast future spot rate using forward exchange
rate

• e

• tF t1 Et1
• ?

• e

Forward Premium of Discount
Expected Rate of Depreciation
17
Forecast future spot rate by current spot rate

• Random Walk(????. Exchange rates do not have
  memory!)
• Et1 E t et1. et1 are identical, and
  independent distributions with mean 0
  (i.i.d.). Therefore, the best forecast of future
  spot rate is simply current spot rate, i.e.
  Et1e E t
• Comment on non-linear exchange rate models
  random walk is very, very difficult to beat.

18
Practices

• Exchange-Rate Forecasting (1)
• Model Success and Failure
• ??????,??????????????????,??????????????????
  ?,????????(exchange rate models)?????????????????,
  ???,??????????????????????????

19
Practices(?)

• Exchange-Rate Forecasting (2)
• Predictions During the Estimation Period vs.
  Predictions Outside the Estimation Period.
  
• Richard Meese and Kenneth Rogoff
  (Empirical Exchange Rate Models of the
  Seventies Do They Fit Out of Sample? Journal of
  International Economics, 1983), Frankel, et al.
  1996, Bergin 2001

20
Practices(?)

• Predictions During the Estimation Period vs.
  Predictions Outside the Estimation Period(?)
• Meese and Rogoff???????????????????????????(??
  ??????????????)?????,???? ???????????????????????
  ?????,??????????????????????,????????????????????
  (???????????????????????)?

21
Practices(?)

• Exchange-Rate Forecasting (3)
• Whos the better predictor?
• Current Spot Rate ??
• Forward Rate
• ??
• Professionals!

22
Practices(?)

• Exchange-Rate Forecasting (4)
• Gregory P Hopper, What determines the exchange
  rate  Economic factors or market sentiment?
  Economic Review, Federal Reserve Bank of
  Philadelphia, 1997.
• Economic models and indeed fundamental
  economic quantities are not very useful in
  explaining the history of the exchange rate or in
  forecasting its value over the next year or so.
  This fact has important implications for market
  participants.

23
Practices(?)

• The Record of Forecasting Services
• Econometric-Oriented Services vs.
  
• Subjective-Evaluation Services vs.
• Technically Oriented Services
  

24
Practices(?)

• The Record of Forecasting Services (1)
• Stephen Goodman (No Better than the Toss of a
  Coin, Euromoney, 1978)?????
• Stephen Goodman???Econometric Techniques
  (???? ????????????????????????????????????,??????
  ?)?Subjective Evaluations?Technical
  Analysis(????)?????????????????

25
Practices(?)

• The Record of Forecasting Services (1)
• Stephen Goodman????
• Econometric forecasters (??????????)
  ???????????????????????
• ?????????????(Must be interpreted with great
  precaution.)

26
Practices(?)

• The Record of Forecasting Services(2)
• Richard Levich(Analyzing the Accuracy of Foreign
  Exchange Advisory Services Theory and Evidence,
  1980)
• ????,subjective?technical
  forecasts?econometric forecasts???? ?????,?????

27
Practices(?)

• The Record of Forecasting Services(3)
• Easton, Stephen A and Lalor, Paul A. The
  accuracy and timeliness of survey forecasts of
  six-month and twelve-month, Applied Financial
  Economics, 1995.
• (1). The experts' forecasts, both individually
  and collectively, failed to predict the direction
  of movement in exchange rates as accurately as
  did the simple forecast implicit in the forward
  rate margin. 

28
Practices(?)

• The Record of Forecasting Services(3)
• (2). Collectively, the experts' forecasts were
  also less accurate in predicting the magnitude of
  exchange rates than were predictions provided by
  either naive spot rate or forward rate
  forecasting models. 
• (3). Individually, of 14 institutions which
  contributed forecasts to 15 or more surveys, only
  one provided forecasts which outperformed a
  simple no-change model.

29
V. ??

• It is simply futile to forecast
  exchange rates. Save your energy and hedge your
  exchange rate exposure.