金融时间序列分析

I enrolled in this course fully aware that it would be challenging and not designed for easy grades as a graduate course. After completing it, I can say with confidence that I have no regret at all — the knowledge, perspective, and practical understanding I gained far outweigh the difficulty.

Who is this course for and NOT for

• Pick this course if:
  • You want to go for or has interest in quantitative finance or applied statistics in finance related goal should consider this course.
• DON'T pick this course if:
  • Your interest in finance is more into traditional or market research (like corporate finance, financial social research and etc), there are many other finance courses that can help your goal.
  • Statistics is not your thing.
  • You choose this course to get “Easy GPA”, including for those who has studied MA309 Time Series Analysis in Fall 2024 and 2025 even though Prof. Zhang already taught GARCH Model. Remember, there is no such thing as free lunch!

My background

Although I am a Statistics major and had Time Series Analysis as my Major Core Course, I found out that most of the contents which are from https://www.stat.berkeley.edu/~bartlett/courses/153-fall2010/index.html are barely applicable in finance which I planned to go to.

During the summer break 2024, I found out that there is Time Series Analysis for Finance while I was preparing and estimating which elective courses from Finance Department I want to pick for my Major Elective Course credit. So, I visited Prof. Sun's office, the theoretical abstract mathematical information in MA to SARIMA model and the frequency based domain time series are useless. In fact, I would say around only 15% to 20% of my Time Series Analysis courses overlaps. The new knowledge I learnt will be discussed below.

As you can see from the course code, this is a graduate-level course. However, undergraduate students (which was my profile at that time) can also take it. Also, since I was the only international student at that time, she asked me if it was fine for her to use Chinese for lecture speaking and I answered “Yes” since I already got used to it in my major classes.

Grading Criteria

20% Assignments+ 30% Quiz (2 times) + 30% Midterm Exam + 20% Final Project & Presentation + 3% Bonus Point from Optional Bonus Homework + a few or even some bonus point from Final Project & Presentation.

Note: actually she planned to have 3 quizzes at that time but it turned out to be 2 quizzes still.

Course contents

Time Series is a crucial knowledge if you plan to do any quantitative field in finance. So, the outline of the course is given below.

In your first five weeks, you will learn the following before your first quiz:

• What kind of time series pattern exists as an introduction. For example, seasonal.
• Definition of stationary which is fundamental in any time series course. This is important for long term forecast.
• Review of probability and statistics function involved in the course which are mean, variance, covariance, correlation. Because in time series, you will learn autocovariance and autocorrelation.
• The definition of white noise (or known as the residual from linear regression in time series version).
• Moving Average (MA) Model which is essentially a linear combination of past white noises
• Autoregressive (AR) Model
• Stationary and invertible, their importance and how to determine if a model is stationary or invertible.
• ARMA model and Seasonal Model

Next, the remaining contents before the midterm are:

• Non-stationary time series model
• Multivariate Time Series and VAR, VMA and VARMA models (If you have a solid foundation of AR and MA models and Applied Linear Algebra, then it is not hard as essentially it is repeating the AR and MA models but in matrix form)
• Impulse Reaction Function (IRF) and Variance Decomposition
• Co-integration and Error Correction Model (including the Vector version, which is VECM)

After you have done your midterm, you will focus on risk side for Quiz 2. That is:

• Concept of financial volatility
• Standard volatility model: ARCH and GARCH Model along with their forecast mathematical result
• Asymmetrical volatility model. This is important because in standard volatility model, it is symmetrical which is problematic as in real life stock price movement move different degree in each of positive and negative (especially in crisis) condition.
• Value at Risk, Expected Shortfall, and Extreme Value Theory: this is essential for financial risk management. Requires a strong grasp of Confidence Interval formula and Variance formula of at least two random variable.

For my case, the Quiz 2 is open book but you have to print it or make your own cheat sheet. there was no limit on how many pages but please do not print too much pages.

My past expectation and experience

Since I only cared about learning necessary stuffs, I aimed for at least A-. In fact, this was crucial for my undergraduate thesis. In my opinion, Prof. Sun is an understanding professor since her body language and communication style with students is not stiff in her teaching as someone who has been in the university for at least 10 years so she knows the history of Finance Department and even Mathematics Department (Can even understand that some mathematics knowledge is not useful in finance). A bit humorous too. Also, if it turned out that students generally did not do well, she considers to adjust the score (which she did for a couple of points, please check below).

I would say her grading style is very fair though it might be a bit grade killer for others. For example, in the second major problem of two major problems in GARCH Model homework, I only had some number calculation mistakes even though my algebraic mathematical derivation process is perfectly fine. So I got 45/50 for the homework. So, be careful!

Even after the end of that semester, I still keep in touch with her since she has the same research interest with me which is Financial Time Series Analysis and other quantitative topics.

For project, she commends the whole class for having a generally better project performance compared to the previous experiences. In fact, when I asked where did the 25 bonus grading come from from my project grade 125/100, she gave higher bonus in project part than last year because the performance on quiz parts in Fall 2024 semester is relatively poor.

Somewhere in the Final Exam Week period, she personally congratulates me in QQ for achieving number 1 ranking in her course and getting 100 (A+) for my 总评成绩. As an international student, I never cared about ranking. But given the fact that this is graduate course and all 2021 students in Department of Statistics and Data Science have MA309 Time Series Analysis from Fall 2023 as their biggest nightmare (because of average score of C+ and super weird teaching style at that time), it is my proudest achievement given that I am not considered to be a top student in my major. I am just a student who was trying to maximize my statistics major learning in financial application (especially with the 2021 curriculum which is a bit too obsolete for current market need). Don't trust me? here.

General student performance

Finally here are some statistics about the performance:

• Total grade (总评成绩）:
  • Average: 83 (B)
  • Median: 85 (B)
• Midterm grade（期中考试成绩）which already includes the adjustment of +6:
  • Best score recorded (最高分): 100分 (Like damn bro… this is not me by the way)
  • Median（中位数）：74分
  • Mean（平均分）：67分。
  • Mine（我的）：89分
• Quiz 2 grade:
  • Average: 35 (not sure if it's 35/65 or 35%, but either way it is a failing grade…)
  • Even the highest performance does not exceed 90%. How did I know that? I am assuming I had the best performance since the professor said: “你的quiz 2比班上其他学生都答得好”which translates to “you did better on Quiz 2 than any other student in the class.”. I got 51/60. Besides, The question format about any GARCH model and asymmetrical GARCH Model topics in Quiz 2 is different compared to previous one.
• For bonus point from bonus homework, I get 1.5/3.

Conclusion

To sum up everything, I did not regret taking this course and in fact learn so much even though I already learnt Time Series Analysis from my own department as my Major Core Course. Prof. Sun is a compassionate professor even though her grading might be a bit too harsh for some people. If you want to pursue a career in quantitative finance (or others related), not enrolling in this course would be a big mistake. If anyone wants my notes, feel free to reply this comment! Anyways, good luck with your study in this course if you choose this one!

Regards,
an average SUSTech international student in Statistics Major
Vincent