Project - Exchange rate forecasting
Part 1.Forecast the exchange rate for the following currencies relative to the U.S. dollar for one monthand for two monthsfrom February 20, 2018. In other words, forecast what the exchange rate will be for each currencyonMarch 20,and April 20, 2018. The currencies are:
Euro, €
British pound, £
Canadian dollar, C$
Brazilian real, R$
Japanese yen, ¥
Kenyan shilling, KSh
Values for these currencies on February 21 are:(to be completed next week)
Euro, € EUR/USD
British pound, £ GBP/USD
Canadian dollar, C$ USD/CAD
Japanese yen, ¥ USD/JPY
Brazilian real, R$ USD/BRL
Kenyan shilling, KSh USD/KES
Use the following forecasting methods to estimate the two future spot exchange rate in one month and in two months for each currency as indicated:
random walk - use for all six currencies
moving average - use for three currencies
auto-regression - use for the three currencies that do not use moving average
purchasing power parity - use for all currencies
international fisher effect - use for two currencies for the 1-month forecast only
unbiased forward rate/forward expectation parity - use for two currencies that do not use the international fisher effect for the 1-month forecast only
structural model - use for the two currencies that do not use the international fisher effect of the unbiased forward rate for the 1-month forecast only
weighted average - use for all six currencies
expert forecast - use for one or more currencies of your choice
Part 2. For each currency and each horizon, using the weighted average forecasts and the February 20 spotexchange rate, recommend whether an investor should buy or sell the foreign currency todayin order to profit from the forecastchange in the exchange rate.
(That is, if according to the weighted average forecast, the currency is expected to strengthen relative to the dollar, recommend that the investor purchase the currency, and if it expected to weaken relative to the dollar, recommend the sale of the currency.)
Each group should submit
(1) forecasts for each currency using each method(in a easy to follow table, see sample below) with all the backup information (calculations, equations and sources) to determine how you obtain and calculate each forecast,
(2) a brief written explanation if necessary to support the quantitative forecasts (For example, you should articulate the reasons for your choice of variables in the structural model, and for choice of weights in the weighted average.),
(3) an explanation of reasoning behind your"expert" forecast, and
(4) the buy/sellrecommendationsfor each foreign currency for each investment horizon (three recommendations for each currency presented in an easy to read summary table, see sample below).
Data sources (partial list):
1. Pacific Exchange Rate Service, UBC, fx.sauder (historical exchange rates - daily, weekly, monthly averages; Canadian dollar forward rates)
2.US Federal Reserve, Board of Governors databases, federalreserve(historical exchange rate and US data - mostly banking data)
3. Federal Reserve Economic Databases (FRED), 8th Federal Reserve District, research.stlouisfed (historical US data in a variety of areas including, trade (some country specific), GDP, inflation, employment, etc.)
4. European central bank statistics,ecb (interest rates and other statistics for the Euro area), and other country's central banks, links available at bis (statistical data for the county, e.g., interest rates, GDP, reserves and etc.)
5.IMF's World Economic Outlook database,
link available on Canvas (historical and forecasted economic data for IMF members, e.g., inflation rates)
6.IMF member economic data summaries, SDDS, SDDS Plus, and GDDS data
link available on Canvas(current economic data for IMF members and links to original sources)
7. investing/currencies spot, forward, and historical exchange rates
8.Commercial banks in a country may have information about interest rates and forward rates.
forecasts, uses "artificial intelligence" techniques (could mean statistical methods or non-statistical methods such as "neural nets") to generate forecasts for some currencies and some horizons. Feel free to look at these but the purpose of this assignment is to determine forecasts using the techniques indicated above, not to find some other person's forecast.
All group members will receive the same grade on this project. It will be up to you to allocate work among the members of the group. Grades reflect correct application of the forecasting techniques and correct recommendation given your group's expert forecast. (Grades do not depend on howaccuratethe forecasts are or whether the trading recommendations are profitable.
Please submit a summary table of forecasts and recommendations in the form below together with spreadsheet(s) that contain all the calculations for each currency:
|
currencies
|
EUR/USD
|
GBP/USD
|
USD/CAD
|
USD/JPY
|
USD/BRL
|
USD/KES
|
|
method
|
|
|
1-month forecast
|
|
RW
|
|
|
|
|
|
|
|
PPP
|
|
|
|
|
|
|
|
IFE
|
|
|
|
|
|
|
|
UFR
|
|
|
|
|
|
|
|
MA
|
|
|
|
|
|
|
|
AR
|
|
|
|
|
|
|
|
SM
|
|
|
|
|
|
|
|
WA
|
|
|
|
|
|
|
|
EXPERT
|
|
|
|
|
|
|
|
2-month forecast
|
|
RW
|
|
|
|
|
|
|
|
PPP
|
|
|
|
|
|
|
|
MA
|
|
|
|
|
|
|
|
AR
|
|
|
|
|
|
|
|
WA
|
|
|
|
|
|
|
|
buy/sell recommendations
|
|
1 month
|
|
|
|
|
|
|
|
2 months
|
|
|
|
|
|
|
[On this project, grades are not everything; reputation and rewards are also at stake. I will post each team's forecasts and recommendations and we will track them to determine which forecasters are most accurate and whose recommendations are most profitable.
For purpose of comparison, accuracy of forecasts depends on mean percentage deviation and mean absolute percentage deviation of each forecast from actual exchange rate on the forecast dates.
Actual exchange rates are those reported by the Pacific Exchange Rate Service (at the University of British Columbia, fx.sauder.ubc.ca/) subject to verification when data from the US NY Federal Reserve is released about one week later. Profitability of trading recommendations is based on the difference between the spot market rate on February 20and the rate on March 20, and April 20. For purposes of determining profitability, we will ignore interest costs and earnings and trading costs.
There will be some type of recognition at the end of the class for the teams with the most accurate forecasts and/or profitable recommendations. We may also provide some appropriate acknowledgement of those teams with less accurate forecasts or less profitable recommendations.]