Topic 10: Expectations, Speculation and Forecasting in the FX Market

Question 1

What is the efficient market hypothesis?

Answer 1

That asset prices fully reflect all relevent information as soon as it becomes avaliable.

Question 2

How would we model the stock market if we expect rational expectations to hold?

Answer 2

As a random walk:

A_t+1 = A_t + ε

Question 3

Define Speculation

Answer 3

Buying or selling a commodity (inc. f.x.) in anticipation of making a profit from a future price change.

Question 4

Is speculation good or bad?

Answer 4

• There can be good and bad speculation.
• Good speculation stabilizes changes in prices.
• Bad speculation makes price more volatile.

Question 5

Explain how speculation can be stabilizing with an example.

Answer 5

Consider that you expect a depreciation of the Australian dollar due to lower iron ore prices.

1. You purchase f.c. in anticipation of it’s relative price rise. This drives up the price, devaluing the dollar.
2. At the end of the period, demand for d.c. falls due to the lower export prices (Q_S=Q_XP_X).
3. As the dollar depreciates, you begin to sell the f.c, increasing demand for the dollar. This appreciates it.

The effects of speculation works against the effects of the initial price change, smoothing the value of the dollar over time.

Question 6

What is destabilizing speculation?

Answer 6

Speculation which is not based on fundemental economics. This can occur due to mob psychology / irrational exhuberance or pessimism - i.e. expectations are not inlign with future values.

This causes changes in the exchange rate that push it away from it’s fundemental value, which is must eventually fall to.

Question 7

What is a forecast?

Answer 7

A forecast is a statement about the value or range of values that a variable will take on at some specific future date. (When the future value of this variable is unknown or uncertain).

Equivelent to the defition of prediction.

Question 8

Why forcast?

(one liner)

Answer 8

We need to make a decision now based on the value of variables in the future.

Question 9

Why do we bother with models if the efficient market hypothesis is true?

Answer 9

Because prediction and explination is not the same.

The traders who use arbitrage to eliminate prices are using predictions to determine the appropriate price.

The models explain why the price is where it is.

Question 10

What kinds of activities motivate us to forecast?

Answer 10

Moosa breaks them up into acitivities that involve:

• Speculation
• Firms financial decisions
• Macro decisions

Question 11

What speculating activites require forecasting?

Answer 11

• Spot and uncovered interest speculation
• Spot forward speculation
• Option speculation

Question 12

What firm financial decision related activities involve forecasting?

Answer 12

• Hedging.
• Investment and capital budgeting (f.x. predictions specifically).
• Financing decisions.
• Pricing decisions.
• Strategic planning.

Question 13

What does Nick say we can use to forecast?

Answer 13

• Fundemental Models
• Time Series Data
• Technical Analysis / Trading Rules
• Judgemental Forecasts.

Question 14

What kind of fundemental models do we look at?

Answer 14

• Single Equation.
• Multi-equation.

Question 15

What single equation fundemental models have we studied?

Answer 15

• APPP: S = ^P*/_P
• CIP: S = F⁽¹⁺ⁱ⁾/_(1+i*)
• MM: S = ^M/_kPY

Note: No clear exogenous / endogenous variables.

Question 16

The construction of a single equation forecasting model.

Answer 16

1. S_t = a₀ + a₁X₁ … a_nX_nt + ɛ
2. Estimate: *put hats on the estimators*

Question 17

What are the steps to construct a AR model. How would we categorize this?

Answer 17

• It is a single equation time series model.

1. S_t = a₀ + a₁S_t-1 … a_nX_t-n + ɛ - for AR(n).
2. Estimate coefficients a₁.
3. Typically we only use AR 1 or 2.

Question 18

Explain what Technical Analysis and Trading Rules are.

Answer 18

TA: Identifying patterns in data. Try and see shapes (triangles, flags, etc.)

TR: Not technically forecasts. Mechanical rules on when to buy / sell.

Economists are not generally impressed with these techniques.

Question 19

What is judgemental forecasting?

Answer 19

A bunch of people, sitting around, looking at data and making a nice old consensus reckon.

Nick calls this an informal fundemental view, looking at all sorts of variables they reckon might be important.

Question 20

What is the indicator of a good forecasting model?

Answer 20

• Zero mean error
• Lowest variation/spread to other zero mean forecastors.

Question 21

What are some indicators of variation?

Answer 21

• MAE - Mean absoluted error
• MSE - Mean square error
• RMSE - Square root mean square error
• MAD - mean absolute deviation
• DSY - downside semi-variance
• V - Variance

Question 22

What is the formula for MAD?

Answer 22

= 1/n Σ ( X_t - X_{AVE t} )

Question 23

What is the formula for DSV?

Answer 23

= ¹/_(n-1) Σ X_i²

Where Xi ₌ (Xi - mean(X)) where X < mean(X) and 0 otherwise.

Question 24

What is the percentage forecasting error?