Sunday, 20 September 2015

Why you should not buy an endowment plan.

To my fellow fresh graduates, congratulations on getting that degree scroll. 

By now, some of you would have secured your job and started bringing the dough home. As we accustom ourselves to the post-college life, many of us are entrusted with a newfound responsibility - managing our finances. How you control your spending and manage your savings will determine whether you can meet your financial goals in life. It is high time for us to maximize our dollars. 

Fortunately (or unfortunately), financial advice are aplenty both online and on the streets. We young tender looking fresh graduates in our office wear are probably one of the prime targets of the preying insurance agents. Buzzwords like 'returns', 'savings', 'beating inflation', and 'miserable banks' interest rate' resonate with your interest. Endowment plans, which are often marketed as a low-risk tool to help you meet financial goals while providing some insurance protections, may hit the sweet spot of many. The guaranteed cash value at maturity also entices the risk averse. 

Barring all the commissions and other fees and charges, endowment is inherently not a bad concept. This got me pondering whether I can build my own endowment using financial products that are highly accessible, while cutting out the unnecessary costs.

Structure of an Endowment

To do so, we need to understand the underlying structure of an endowment plan. An endowment can be essentially broken up into 3 parts, the guaranteed cash value, the non-guaranteed cash value, and the insurance protection (usually death and total and permanent disability). In a very simplified sense, we can think the premiums you paid as being channeled to each of the respective parts in an endowment. The premiums (after deducting all the commissions and fees) are allocated such that a significant portion of it will be used to buy low-risk-low-yield bonds to generate the guaranteed cash value. A small portion of the premiums will be used to pay for the cost of insurance. And finally, the remaining premiums will be invested in higher risk equities. 

The DIY-Endowment

In order to replicate an endowment, I will be combining 3 financial products that are highly accessible to the general public. They are: (1) the Singapore Saving Bonds (SSB); (2) the STI-ETF from POSB Invest-Saver; and (3) a direct term from NTUC income that covers death and TPD. Using this 3 products, I will try to replicate and compare it against the RevoSave (3-Pay-10). The RevoSave is a 10-year endowment plan with a $30k guaranteed cash value at maturity. Three $10k premiums are paid over the first 3 years of the policy.

To ensure the quality of comparison, I have included all the transaction fees for SSB and POSB Invest-Saver. STI-ETF are brought using the dollar cost averaging approach over 3 years at a monthly interval. The projected annualized return of STI-ETF is 9.2% (this takes guidance from the annualized return of STI between 2002 to 2013). The returns of SSB follows the interest schedule as shown on the SSB official page. The cost of insurance is $55 p.a. with a $50k coverage on death and TPD. 

Below is the benefit illustration of RevoSave for a 23 yo female non-smoker.

At 4.75% projected return, the RevoSave would generate a  projected return of $9689 at maturity with an average coverage of $39,158.40 over the 10-year period.

On the other hand, the DIY endowment yield superior performance compared to the RevoSave. The death and TPD is fixed at $50k throughout the 10 years. The non-guaranteed return is $11029.14, while the guaranteed return is $30000.03.

For those who are interested, the code for calculating the figures is pasted here.

Is the assumption of 9.2% annual return from STI-ETF a fair comparison against the projected return of 4.75% in the endowment? 
We must note that the 9.2% is the return from equities. On the other hand, the 4.75% from RevoSave is the return from an investment mix of both equities and other low yielding instruments. As of 31 Dec 2014, the participating fund of RevoSave is make up of 23% equities, 67% fixed income, and 10% of cash/loans/properties. Assuming the low yielding instruments yield an annualized returns of 3% to 3.5%, this means that RevoSave assumes their equities can generate a return of 8.9% to 10.6%. The higher equities exposure (18% in my diy-endowment vs 23% in RevoSave) also means higher risk exposure in RevoSave.

All in all, the diy-endowment is likely to yield better returns, offers higher protection for 9 out of 10 years, able to provide a guaranteed principle, lesser risk exposure, and greater flexibility to customize to your needs.

UPDATE (8 Nov 2015)
Sunday Times published an article titled Make (full) sense of insurance policies on 8 Nov 2015 that highlights important things to look out for before committing to  an endowment policy. The article provides a useful table which compares the investment returns on insurer's most representative participating funds for the past 7 years. Let's take a look at how they compare against my DIY endowment. (Note: I am assuming a risk-free product with similar return profile as SSB exists for the period of comparison)

A quick glance shows that the performance of all funds are comparable. But it is important to note that the returns of the insurer's participating funds are NOT the effective or net returns that policyholders will get. The returns have yet to consider they hefty distribution cost, management expenses, commissions, and cost of insurance. All these cost can easily shave off more than 1.5% of the returns.

Clearly, my DIY portfolio significantly outperformed many of the endowment plans based on past 7 years result. 

Wednesday, 3 June 2015

Last Academic Update AY2014/15 Sem 02

As the morning sun rose and crept into my bedroom, I sat in front of the desk, where I had spent countless hours grinding for the past 4 years, anxiously awaiting for the moment my phone vibrates and reveals the outcome. 

A blend of anticipation and trepidation engulfed me. The same cocktail of emotions which overcame me when I was still a freshie waiting for my Yr1 Sem1 result. But this time, more is at stake. I needed at least an A- for FYP in order to achieve the First Class Hons (FCH) I worked so hard for, and in turn, a FCH means quite a significant difference (+8.2%) in my starting pay.

The phone buzzed and there it was, ST4199 A-. Exuberance awash me as my efforts have bore fruit. It has been a long journey. When I first set foot in NUS, I worried about my ability to cope; then I started worrying about not being able to get an internship; I was quite depressed when my job applications were rejected repeatedly by Dept of Stats (Singstat); And my FYP was quite a mess, hence I was very worried about my grade. Luckily, things always turn out fine.

ST4199 Honours Project in Statistics A-
ST4232 Nonparametric Statistics A-
ST4242 Analysis of Longitudinal Data B+
ST4245 Statistical Methods for Finance  A


The content is very light. The entire syllabus consists of approximatedly 70-80 pages worth of content. The materials taught are also quite easy. A/P Yu Tao likes to read off lecture notes, but his explanation generally helps to clear doubts. Assignments and final exam were manageable.

The Content is heavy, and the module requires a lot of SAS coding which I am not very well-versed with. Spent a lot of time doing trial and error to get some output from SAS for my tutorials and I have no idea what's going for the most of the semester. Overall, I did not have a good experience with this module. Module was taught by A/P Li Jialiang.

For people who have taken Financial Economics (EC3333), Applied Time Series Analysis (ST3233), and Linear Model (ST4233), there were not many new concepts introduced. As usual, Prof Xia Yingcun's midterm and final exams follow his tutorials closely. As long as you understand the tutorials, the exams will be manageable.

If anyone would like to get a softcopy of the lecture notes, tutorials for any of the modules I have taken, just leave a comment below, or email me at Do note that, I don't have any secret study or access to pyp solutions. The files I share will contain only lecture notes, tutorials (+ solutions), assignments (+ solutions), and perhaps some pyp obtained from the library portal. Hence, asking me for materials towards the end of the semester is kind of pointless. The materials are meant for people who are keen to read ahead, or have a glimpse of the contents before they bid for the modules.

That's all for my last academic update. I am not sure if I will still be posting anymore academically related post in the future. I hope the series of module reviews and freshmen guides had help some of you. Feel free to contact me if you have any doubts related to the course, I will try to answer them to the best of my ability. Bye.

Wednesday, 27 May 2015

The End of a Chapter

How time flies.. 4 years had zoomed past, and my journey as a student has officially ended. When I was in JC, I have always had a knack for Statistics despite my distaste for Mathematics. Upon my JC graduation, I wanted a uni course which interest me, leverages on my strengths, and promise a certain level of career prospect. Statistics naturally fulfilled the first 2 criteria, but I was uncertain about the career prospect for a statistician or a statistics graduates, especially in Singapore's context. I did some research online then, but the relevant information were scarce. In the end, I took a leap of faith and accepted NUS Science with a very murky idea of the potential career opportunities. On hindsight, I am glad that I have chosen this course. 

The Programme
The statistics course in NUS is not going to be easy. Getting a First Class Hons will not be a walk in the park. With a significant number of foreigners who are brainy and hardworking, the competition is steep.

Comparing with economics modules, the workload of statistics mods tends to be heavier and requires more effort to understand. It is perfectly normal to leave a statistics lecture utterly confused, especially if  you did not skimmed through the reading materials before the lecture.  

Do a Double Major. 
For those who have the capacity to do more, consider a double major. 

The Statistics curriculum consists of one of the highest number of unrestricted electives (36MCs) compared to many other majors. This gives statistics students considerable flexibly to take on a second major without being heavily overloaded. For example, with careful planning, one can fit a second major in Economics into the programme requirement without additional coursework.

A double major do matters to the employers, and will enhance your employability. Apart from the intangibles like demonstrating your diverse interest and knowledge, your willingness and ability to do more than the minimum, certain combinations of majors are highly valued by employers. Combinations such as Statistics + Economics or Statistics + Computing are highly complementary, and for the latter, I foresee a rising demand for such graduates. Needless to say, a second major will also broaden your career options. Personally, a double major did opened up more opportunities for me, and I would not have secured my relatively-well-paying job offer without my second major. 

Career Paths
Majority of the statistics grads should not face much difficulties securing a job, especially for students who are graduating with honors. Many of my course mates managed to secure job offer(s) before the final exams. Some of the possible career paths available to statistics grads include:

Healthcare (Clinical Analytics) 
Biostatisticians with the various healthcare/research facilities and healthcare authorities 
Statisticians in the private and public sectors. 
Banking (risk, analytics, ops and tech) 
Data Analytics
Market Research

However, some of these jobs have high barrier of entry, and so I would advise you to start preparing for your career as early as possible. Join networking session to find out what are the entry requirements for the job (pro. qualifications/exams, relevant internships, etc.). Get relevant experience through internships. Attend some career workshop conducted by NUS career center.

Let me end this section with a quote from Google's Chief Economist, Hal Varian, on statistics and data.

“I keep saying the sexy job in the next ten years will be statisticians. People think I’m joking, but who would’ve guessed that computer engineers would’ve been the sexy job of the 1990s? The ability to take data—to be able to understand it, to process it, to extract value from it, to visualize it, to communicate it—that’s going to be a hugely important skill in the next decades, not only at the professional level but even at the educational level for elementary school kids, for high school kids, for college kids. Because now we really do have essentially free and ubiquitous data. So the complimentary scarce factor is the ability to understand that data and extract value from it.” – Hal Varian, Google’s chief economist

Modules Ranking
Finally, here's my personal ranking for the modules I have taken over the past 4 years. They reflect my interest, preference, strengths, and weaknesses. Therefore, my experience with these modules may be wholly different from yours. 

Most interesting mods
1. GEK1508 Einstein's Universe and Quantum Weirdness by Prof Phil Chan
2. EC3312 Game theory and Application to economy by Luo Xiao
3. EC3333 Financial Economics I by Lim Boon Tiong

For those who are intrigued by the movie Interstellar (2014), I highly recommend GEK1508. The module explores the intricacies of Relativity, Quantum Mechanics and strings. Prof Phil Chan is also an extremely passionate lecturer who makes the lectures an enjoyable experience. Easily the best module I have taken in NUS.  

Most Difficult ST mods
1. MA2108 Mathematical Analysis I by A/P Lee Soo Teck
2. ST3236 Stochastic Processes I by Sun RongFeng
3. ST3246 Statistical Models for Actuarial Science by A/P Lim Tiong Wee

Most Useful mods
1. ST2137 Computer Aided Data Analysis by Dr. David Chew
2. ST4231 Computer Intensive Statistical Methods by Alexandre Hoang THIERY
3. ST3239 Survey Methodology  by Dr. Chua Tin Chiu

Apart from the Most Useless modules listed below, many of the ST/MA lvl1000-2000 mods are useful for building up our foundation in understanding and applying statistics. The top 3 listed here are special mentions which I think are very useful in both industrial and academia. 

Mose Useless mods
1. MA2108 Mathematical Analysis I by A/P Lee Soo Teck
2.  ST3236 Stochastic Processes I by Sun RongFeng

I don't recall much from these two modules. Basically, they are just learn-and-dump mods as I have not applied anything from these two modules for any of my higher level modules.  

Most Difficult EC mods
1. EC3312 Game theory and Application to economy by Luo Xiao
2. EC3333 Financial Economics I by Lim Boon Tiong

Easiest EC mods
1. EC2101 Microeconomic Analysis I by Zhang Yang
2. EC3101 Microeconomic Analysis II by SNG Tuan Hwee
3. EC3361 Labor Economics I by Peter James McGee

Thursday, 25 December 2014

Academic Update AY2014/15 Sem 01

Initially I planned to skip Academic Update for this sem as I am really busy with my internship, but then, I saw a spike in blog visit so I will just do a short update for the past semester to satisfy your curiosity. =P

Did relatively badly compared to the previous semesters. The heck-care attitude is really getting onto me as my CAP is unlikely to drop below 4.5 now (Assuming my FYP can get an A-, I just need an average grade of 2.0 for my remaining 3 mods).

Computer Intensive Statistical Methods
Linear Model
Bayesian Statistics
Macroeconomic Analysis II


The module content have some overlap with ST3247 and ST4234, I would recommend people to take ST3247 first and then do both ST4231 and ST4234 concurrently. For  the first half of the semester, the content is manageable, however, things get more difficult in the second half. Compared to Prof Vik who taught me ST3247, Alexandre Hoang THIERY is weaker in explaining difficult concepts. However, he does upload very detailed tutorial solutions and R codes, studying them will greatly enhance your understanding.  

Midterm mean-median: 46.16-46 (/70)

ST4233 A compulsory module for stats students. Workload is extremely light with only weekly tutorial which consist of a few questions, one graded assignment and finals. The assignment is basically free marks for all, so everything will depends on the finals (which is not difficult, but I screwed it up). Although Zhou Wang spent most of his lecture proving theorems and lemmas, there are not much proving in his Finals.

Assignment mean-median: 96-96 (/100)

This module has an extremely light content, you will probably spend most of the time manipulating some pdf/pmf to get the desired form and identify the distribution. Catching Zhang Jin-Ting's accent can be quite challenging initially, but you will get used to it over time. Overall, he is a decent lecturer. 

 Midterm mean-median: 70-72 (/100)

It was a long time since i last touched macro as I took EC2102 during y1s2, and macro was not my strength to begin with, hence I did not do well for it. It dosnt help that VU Thanhhai is quite week in his explanation and his mid term was really terrible with lots of grammar error and there was a question with double negative.

Midterm mean-median: 21.83-23 (/100)

As usual, if anyone would like to have materials such as lecture notes or tutorials for any of the modules that I have taken before, just leave a comment below or email me ( with your nus email. Do NOTE that, I don't any secret study guide or access to pyp solutions. The file I share will contain only lecture notes, tutorials/ assignment and its solution and maybe some pyp found from the library portal. Hence asking me for the materials towards the end of the semester is kind of pointless. They are meant to be shared with people who would like to read ahead or have a glimpse before they bid for the module.

Friday, 11 July 2014

Review for year 1 Modules

Some of you may be wondering why I did not post any module review for modules I have taken in my first year. That’s because I blur blur and accidentally deleted them some time back. As the semester is about to start, I thought it might be helpful to post my review for some of these year 1 modules. Note that I have taken these modules 3 years ago so it might not accurately reflect the current syllabus or assessment criteria, but I doubt that there were much changes.


A lot of theorems which can be quite intimidating at first, so it is necessary to be familiar with all of them in order to use them during the time-packed examination.
Dr Ng Kah Loon is the best lecturer I have encountered so far and I voted him as the best lecturer for this semester. Great in delivering concepts. Pace is consistent and manageable if you are prepared. Helps to keep the lecture quiet so that it is more conducive for those that are keen to learn. And a sense of humour that keep the lecture engaging.

Assessment and workload
20% midterm. 10% lab quiz. 70% final. Lab quiz is for you to score so do prepare for it. His Mid-term is challenging, expect the mean to be around 22-24 out of 40. Finals on the other hand is much easier.


More in depth compared to A level calculus. Some proofing questions that JC students might be unfamiliar with, and will take some effort to get the hand of it.
Lecturer: Wang Fei
Passionate and particular in teaching the right concepts.  He also like to spend some times during the lecture to share how students should adopt the right attitude when doing math. Some might think that he is wasting time talking about it, but I personally find them quite useful. Pace is quite fast, hence it is advisable to do pre- and post-lecture reading up on your own.  
Assessment and workload
Homework - 15%
Lab assignment - 5%
Midterm - 20%
Finals - 60%

Personal experience
I made a comeback for this module and got an A. Basically I screwed up my mid-term due to some personal reason, scoring slightly below the mean score. And partly because I don’t know many people in the course, I am among the few that could not get full marks for all the homework and lab assignment as I have to do them on my own while most of the others just copy or discuss with each other. However, I put in a lot of effort after I got back my mid-term results and breezed through the finals while many others complain that the paper was ‘imba’. So with a lot of hard work, you can score for this module.


Almost similar to H2 Econs but more in depth and some new stuff. Not recommended for those without Economics background as the content is really heavy.

Teaching staff:
Dr Cornie and another male lecturer. Dr Cornie's lecture just reap off everything from the textbook. Sometimes she seems a bit confuse over the slides. The male lecturer was much better, teaching stuffs that are beyond the textbook. 

Assessment and workload
Online tutorial (10%) - some technical screw up of the platform at the initial stage but was quickly rectified.
Mid-Term (30%) - MCQ. Easy. Everyone scored very high. If you are aiming for A+, u need to score full mark for this.
Finals (60%) - MCQ. Slightly harder but still quite simple. You still need to score close to full mark to get A+

Personal experience
Bell curve is extremely steep, majority of the students come in with econs background. Careless mistakes is not tolerated if you are aiming for A or A+.


Almost identical to JC stats with a bit new stuff towards the end. But the concept is still the same. Very dry module that makes me sleepy very lecture.

Teaching staff:
Wong Yean Ling speaks really really fast. Always on sick leave as she seems to be really ill most of the time... long winded but give good summary of the chapters

Assessment and workload
Mid-term - 35%, Finals - 60%, Assignment - 5%. Assignment is a free 5%, anyone could do it easily. Midterm is easy. Finals is tricky

Personal experience
Got an A-. Very dry and easy module, which is why I did not put in enough effort and didn’t get an A. Bell curve was terribly skewed as most people scored really high for it. Those with H2 math will be at an advantage.


General/special relativity. Quantum Mechanics. Particle Physics. No difficult math involved at all as it is only an exposure module. Abstract ideas that you can either accept it or you can’t.  Which is probably the reason why no matter how badly u do, the lecturer won’t fail you so that you can S/U it. Highly recommended for those that like to think a lot and has an interest in modern science and abstract ideas.

Teaching staff:
Prof Phil Chan is a passionate lecturer. However, he like to say 'please take note of it' at basically everything.
Tutors are very clear in clarifying doubts and explain concepts in another way to complement what was being taught in the lecture.

Assessment and workload
30%-Group Project (in groups of 2-3) Write a few thousand words report (either a story you made up or a book review). Choose your teammates carefully, avoid those who are intending to S/U if possible

20%x2 -2 tests one held during midterms (40 mcqs,1hr) and the other (45mcqs,1hr) before reading week. Both open book. Even though it is open book, you still need to be extremely familiar with the material as you will not have the luxury to flip through the lecture notes slowly to find your answers.
5%- star gazing (just go for 2 sessions. It's on certain Fri 7-10pm) Just need to go and sign attendance if you are not interested.

5% - forum discussions
10%- E-learning. Does not always apply. Not every semester will have e-learning.
5%- In class assignment. Free frag, TA essentially give you the answers.
5%- Tutorial Attendance


I did not have any background in computing, but was still able to follow the syllabus fine. The module teach a lot of basics C programming which are easy to learn. However the challenging part is applying these basics programming techniques to write programs which can be quite complicated and not straight forward. However, if you get yourself familiar with the programming language and is able to think logically, you will do fine in this module.

Assessment and workload
5 Lab Assignments - 5%
Mark is awarded by attempting. As long as you submit your assignment with reasonable effort, you will be awarded the 1% for each assignment. NOTE: Do not copy your program from your friends. They are able to detect copied work and I can assure you that you won’t like the outcome if you are caught.

Discussion Attendance - 5%
Free 5%. Just make sure you attend every tutorial and present some solutions to the problems and you will get all 5%

Practical Exam 1. - 10%
Easy. Median mark was around 8.8/10. Try to score well for this.

Practical Exam 2 - 25%
Challenging. Medium was around 60-70%. Still do-able. Practice on recursive

Mid-Term - 15%

Final - 40%   

Personal experience
I scored an A- for this module which is kind of disappointing. For some reasons, I could not do some of the not-so-difficult questions during the exam and some fatal careless mistakes which caused me my A. The good thing about this module is that it is not bell-curve based, so you do not need to worry if your course mate is extremely good in the module. As long as u manage to get certain marks, you will get the grades u deserved. Another thing is that the grading of this module is extremely transparent, you will more or less know your overall result and hence predict ur grade the minute u finish your final paper.


A more in depth version of the first part of MA1102R (no differentiation and integration). The content of the module is very light with only about 5 topics. However, there are a lot of proving in this module. Textbook is not a must-have, but will come in handy at times.

Assessment and workload
Homework Assignment - 10%
Like all other maths modules, everyone will be scoring near full mark for homework assignments, so make sure you do too.

Tutorial participation - 5%
Present the tutorial questions 2 times in class.

Mid-term - 20%
Relatively easy paper. Be careful of careless mistakes. Median score is 71 for my semester. The format of the paper is exactly the same as the previous sem, even the questions are similar. I could have done much better had I known this fact.

Final - 65%
A lot of proving questions. Personally I feel that the paper is not do-able at all, I do not know how to do at least half the paper and wrote rubbish in it, praying that it might earn me some sympathy marks.   

Personal experience
The most terrible experience I had ever had for any math modules. For the first time in NUS I could not do half the paper and left the exam hall feeling really terrible. But in the end I think the bell-curve saved me and I still manage to get a B+.


Content was not very heavy. Focus a lot on understanding and manipulating some economic framework graphically. Textbook is a MUST. If you are unwilling to buy the textbook, there is a free eBook ver. online available for download (google it yourself).

Assessment and workload

Tutorial Participation - 10%
No idea how is the point awarded. But just be as active as you possibly can. If you are really shy, attempt to present your solution at least twice.

Mid-Term - 30%
MCQ. Slightly challenging. Median was 21 and the 75th percentile was 24 out of 30. Know your graphs well.

Finals - 60%
Closed book. Know your graphs well such as how to derive some graph from another and practice on solving simultaneous equations will be useful.

Personal experience
Got an A for this module. Know your graphs well and understand them and you will do fine in this module. If you know some math, it will be an added advantage.


Astronomical objects and events. Pretty interesting most of the time, particularly if you like astronomy. Prof Cindy had a funny accent, and she can be quite funny sometimes.  Textbook not required.

Assessment and workload
Tutorial Attendance + Assignments - 20%
Almost free 20%. Attend every tutorial and do the assignments (simple calculations, just be careful of careless mistakes) and you should expect more than 18%

Term test 1 - 20%
MCQ - closed book. Beware of stupid questions such as the date of next astronomical events which does not question your understanding of the concepts you have learned. For the semester I took, the mid-term was harder than usual as the highest mark was only 28/30. The lecturer claim that this was because she had run out of easy questions. Focus on memorization skills

Term Test 2 - 20%
MCQ - Open book. Test more on concepts. Some questions requires you to know where to find the answer in the lecture notes unless you memorized every single little details which is kind of impossible. Term test 2 might be more challenging to some, but because I suck in memorization, I think I fare better for this test compared to the first one.

Group project + individual Report (20% + 20%)
Do an experiment related to astronomy with a group of around 4 people (google for the experiment). The prof will show you a lot of past project done by students from the previous semesters. It is okay to do experiments that other people had done before, or will be doing, because trust me, after so many semesters, I believe the prof have seen every possible experiment out there. The IMPORTANT thing is that your experiment must have a lot of variables (at least 3) for you to play around with. For the individual report, you need to demonstrate your understanding of the experimental results and explain why certain trend or anomaly exist. As long as your explanation make sense from physics point of view, then it will be fine.

Personal experience
Got an A+ for this module. Read up and understand the concepts from the lecture notes. Even though it is an open book test for term test 2, prepare a cheat sheet with important dates or information so that you can find them quickly during the test. I spent a lot of effort doing the individual report, (more than a week including reading up and researching on the topic I did) while I have some team mate who had not even started the report less than a week before the deadline, I even heard a random stranger boasting that he completed his report within 5 hrs. Spend some effort on the report and it will definitely help u with your grades especially when a lot of people are taking this module with a heck-care altitude.