EC4103: I appreciate the intention behind this module, that is to expose undergraduates to the various aspects of policy making in Singapore. It covers some macroeconomic overview, financial institutions and regulatory bodies, productivity- and innovation-boosting initiatives in the recent years, inequality in Singapore, and various fiscal policies targeted at Singapore's main challenges: productivity and aging population. However, there are major flaws in this new modules. First, it IS merely an overview. The scope is too broad and thus the module is lacking in terms of depth and rigor in any particular aspect. Second, 100% CA is heavy, but to partitioned the 100% into five 20% segments with weekly submissions and presentations, it's a chore. The consensus is that there's no incentive to do rigorous, in-depth analysis in each component. Most people (including myself) would prefer to focus on other modules which allow us to focus on proper research in ONE specific topic/area for the whole semester. That tastes more like research. In summary, most stuff taught are good-to-know facts, the projects are like summarizing-existing-arguments-on-the internet-to-justify-your-stand, or use-STATA-to-get-some-numerical-answers. Maybe it should be renamed as An Introduction to the Singapore Economy, and be re-coded as a level 2000 module. Lol. For this module, I'd expect a B+. Finally, I honestly need to thank my group mates for their effort and accommodating with my hectic schedules.
EC4303: This module is less rigorous than I thought. Prof Oka focused more on the variety of econometric techniques than going super in-depth to a few of them. We covered OLS (duh), MLE, Nonparametric Regressions, Quantile Regressions, Time Series (ARMA/GARCH), (Dynamic) Panel Data, LASSO, CART, Potential Outcome Framework, RDD. That amounts to one new topic/method every week. Book presentation component was quite interesting but most of the time no one understands each other's part LOL. These chapters are also not tested in final exam unless it's covered in lecture notes (like Nonparametric and CART). The most memorable part was the term project, where my group took upon the challenge to do nonparametric GARCH (1-step ahead) forecasting on STI data. No amount of words can describe how much I've learn about econometrics in the process. Self-revising R to code the in-sample fitting and out-of-sample forecasting process, deciding between all the information criterion, learning about the L1 and L2 goodness-of-fit measure, and using RMSE vs Diebold-Mariano Test for forecasting superiority. It's an enjoyable process and when the final results are tabulated, there's so much pride in the work. Final was a little messy and I could probably do better with clearer thought process, nevertheless, I did my best at that point of time. I'd conservatively predict an A-.
EC4332: Took this module because the department DE-LISTED EC4331. Ultimate disappointment. I was looking forward to the New-Keynesian Models. In the end, I went to look for Prof Bodenstein to get the materials. Back to EC4332, it's not too bad although Dr. Park is pretty chill about his midterm and final exams. The questions weren't too difficult, the course content required only minimal effort on my part, the project was not too time-consuming. Prof Bodenstein told me early in the semester that the department wants a more qualitative EC4332, and Dr. Park delivered that well. The intuitions are conveyed across with intuitive examples and not overly daunting. For banking, he covered the Diamond-Dybvig model and extensions to bank runs and sunspot variables, IO-style symmetric banking equilibrium (both partial and complete). He did not touch on extended topics such as lender-borrow relationship, collaterals, and shadow banking. For those who plan to take EC4332 under Dr. Huang next sem, do note that he plans to cover this (and less on monetary economics). Hmm, since I'm quite confident about finals, I hope for an A for this module.
On the side note, I'm once again disappointed that the department does not offer EC4331 next semester due to the lack of manpower (Dr. Jo is switching to EC4343, new module taa-daa).
EC4333: This module overlaps greatly with MA3269 and the first half of MA4269. Took this module to clear my finance specialization requirements. First half goes through interest rates, bonds, risk aversion, stochastic dominance, portfolio hedging and some extra stuff on econometrics. The second half move towards Markowitz's portfolio selection theory, stochastic calculus and risk neutral valuation. This module is not too difficult under Prof Tsui. Anyone with sound understanding of probability theory and calculus would see the synergy of both in this course. My advice? Get the math right. First half you just need to know geometric series and calculus well. For the second half it's trickier, portfolio theory need some linear algebra, while for a good understanding of stochastic calculus one would need to know some rigorous math. Fortunately, you can survive the course with a just a good understanding of Ito's lemma, Wiener process, and the lognormal property. Speaking of which, I made a mistake representing lognormal-distributed stock price in normal form. Oops! Judging from the midterms and projects, I could still manage an A- if there's no standardization of scores in such a small class (of 20).
EC5104R: THIS IS ONE CRAZY MODULE. NUFF SAID. LOL. Okay I'll elaborate. There's 3 undergraduates, 3 master's and 10 doctoral students in this class. Suddenly, you feel a bit dumb lol. Okay, the only thing that might help is real analysis background. MA2108, MA3110, MA3209, MA4211, MA4266. Maybe EC3314. Main properties of functions and correspondences covered are continuity, convexity, completeness and compactness. These lead to major results such as Mean Value Theorem, Intermediate Value Theorem, Extreme Value Theorem, Maximum Theorem, Fixed Point Theorems, Benveniste-Scheinkman Theorem, Inverse Function Theorem, and Implicit Function Theorem. Building on these theorems, we derive the conditions in which optimal solutions exist and how these solutions can be computed using the Lagrangian method and Karush-Kuhn-Tucker conditions for static problems, and the equivalence of the unique, optimal solution of the Bellman Equation to the solution of the Sequence Problem in dynamic programming. In summary, HARD, VERY HARD. Not even exaggerating. Getting a B+ would be great!
EC5332R: I audited this module under the permission of Dr. Bodenstein. However, I've only managed to attend two seminars as it clashes with MA3209 lectures (or when lunch takes precedence lol). The course follows the same textbook as the banking part of EC4332, and also covers the extended topics with greater detail. I must say that I did learn a lot more about banking in this class, albeit my lack of attendance. The access to core materials also further my understanding of banking, but I didn't have time to look through everything yet.
MA3209: It's like a more organized version of EC5104R. Probably because EC5104R is geared towards application, thus only focus on specific theorems. This module focused on real analysis in metric space, particularly the Euclidean n-spaces, bounded and/or continuous function spaces. Most of the things are generalization of theorems in MA2108 and MA3110. Concepts of open and closed sets are studied in detail. Continuity concept is generalized to metric spaces. Together with completeness, connectedness, and compactness properties of metric subspaces, these topics dominated the first half of the course. The second half focused on differentiability of multivariate functions, which is a rigorous synthesis of MA1104 and MA3110. Inverse Function Theorem, Implicit Function Theorem, and the Lagrangian Theorem are the main results of this segment of the course. Damn, I was invoking Implicit Function Theorem in the final exam, but I wrote "by Inverse Function Theorem". I hope I don't get penalized to badly. Maybe a B+? Any lower would be an S, but I don't wish to S/U this module as it is supposed to be help my postgrad application.
MA4269: Ahh this module was fun but I kinda screwed the midterm (it's only 20% though). This module is really exotic. Exotic options. European, American, Asian, Barrier, Lookback, Exchange. Wow. It's really interesting to learn beyond the vanilla options and discrete-time n-step binomial pricing. The course focus on continuous-time finance, with the martingale approach being the main risk-neutral valuation method. PDE formulation governing the asset value is also a huge topic, but we are not required to solve the PDEs in this course. However, to use the martingale pricing approach, the Martingale Representation Theorem, Radon-Nikodyn Theorem, and Girsanov Theorem are indispensable. So, we have to learn some basic concepts regarding probability measures and how to transform between different numeraires. Well, actually, just between money and stock numeraires. It's really a good introduction to exotic options, and you'll see why quantitative finance/financial engineering positions are filled with mathematicians, physicists and engineers instead of business school graduates. Final exam was a race against time, I did my best within the time limit. Hopefully the bell curve is on my side. Most probably a B+.
Yep, that's all for now about the modules. I might do a proper one within the week considering it's MPE period but first and foremost, I shall settle my thesis proposal.