FINM7008 Chap.3 Forecasting Markets and Reading Return Distributions
Forecasting Markets and Reading Return Distributions
The forecasting record
Can anyone predict the market? The lecture reviews market timing, professional forecasts and the academic record of predictors, and explains why competition removes any public signal.
The historical average remains the forecast to beat.
Describing distributions
The chapter then scales mean and risk over time, defines skewness, kurtosis and the downside measures used when returns are not normal, and introduces the Sharpe ratio.
Where the tail measures fit
The course accepts a normal distribution as a working approximation for most portfolio questions, which is what makes mean and standard deviation enough.
This chapter marks the boundary of that assumption: once skewness or fat tails appear, value at risk, expected shortfall and the frequency of extreme losses describe the danger that a single standard deviation hides.
What this chapter covers
- 01
Market timing and the historical average
- 02
Why public predictors stop working
- 03
Scaling mean and risk with time
- 04
Sample standard deviation
- 05
Implied volatility indices
- 06
Skewness and kurtosis
- 07
Value at risk and expected shortfall
- 08
The Sharpe ratio
Worked example · free
Projecting monthly mean and risk to longer horizons
- 1The mean grows in proportion to time: 12 × 0.8 = 9.6% and 36 × 0.8 = 28.8%.
- 1The standard deviation grows with the square root of time: 4.5 × √12 = 15.59%.
- 1Over 36 months: 4.5 × √36 = 4.5 × 6 = 27.0%.
Key terms
- Market Timing
- A strategy of moving money between shares and safer assets on predictions of market moves.
- Kurtosis
- A measure of how fat the tails of a distribution are, zero for a normal distribution once 3 is subtracted.
- Value at Risk
- The loss at a chosen low percentile of the return distribution, such as the 5th or the 1st.
- Expected Shortfall
- The average loss in the worst-case tail of the return distribution, such as the worst 1% of outcomes.
Forecasting Markets and Reading Return Distributions FAQ
Why do most market predictors fail out of sample?
Once a public variable is known to predict high returns, traders buy as soon as it signals, lifting the price today and removing the high return. Welch and Goyal found that the proposed predictors forecast poorly and that the historical average usually did better.
Is standard deviation a complete measure of risk?
Only if returns are normally distributed. With skewed or fat-tailed returns, standard deviation understates the chance of extreme outcomes, and the course adds value at risk, expected shortfall and the lower partial standard deviation.
Why divide by n minus one for a sample standard deviation?
Using the sample mean in place of the true mean makes squared deviations slightly too small on average. Dividing by n − 1 instead of n corrects that bias, which is why the lecture calls it the unbiased estimate.
What do the VIX and the A-VIX measure?
They measure the volatility the options market expects over the next 30 days. The VIX is backed out of S&P 500 index option prices, the A-VIX out of ASX 200 option prices, and the VHSI by Hang Seng index options, giving a forward-looking alternative to a historical standard deviation.
What does negative skewness mean for an investor?
It means the left tail is longer, so large losses are more likely than equally large gains. A portfolio with negative skew can look comfortable on its mean and standard deviation while still carrying a meaningful chance of a severe loss.
Has anything improved on the historical average?
A 2024 follow-up by Goyal and Welch found only 4 of 29 newer predictors still performing well. Combining many public variables can help: Dong, Li, Rapach and Zhou used shrinkage on 100 predictors to build one signal with significant out-of-sample predictive power.
Exam move
Learn the forecasting argument as a three-line story: predictors are proposed, they fail out of sample, competition explains why. Then practise square-root-of-time scaling and one sample standard deviation by hand. Finish with one interpretation question: given skewness and kurtosis figures, say which extra risk measure you would report and why.
Keep the Welch and Goyal result and the 2024 follow-up as two dated facts you can cite in a short-answer question on market timing.
Working through Forecasting Markets and Reading Return Distributions in FINM7008? Sia is AskSia’s AI Finance tutor — ask any FINM7008 Forecasting Markets and Reading Return Distributions question and get a clear, step-by-step explanation grounded in how FINM7008 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
FINM7006 Applied Foundations of Finance · FINM8007 International Finance · STAT7055 Introductory Statistics for Business and Finance
