My Rambling Thoughts

Tenant-owner duality

News: Are HDB residents tenants or owners?

Date: 24 Jun 2017. Source: ST.

The report on the homes in Geylang Lorong 3 reaching the end of their leases highlights the issues surrounding leasehold properties (Govt taking back 191 homes in Geylang when lease ends; June 21).

While applying to rent out my HDB flat, I noticed that the Housing Board referred to me as a "tenant", and my potential tenant as a "sub-tenant".

He has the audacity to ask this question? :lol:

News: HDB buyers are home owners, not tenants

Date: 19 Jul 2017. Source: ST.

We thank Mr Larry Leong for his feedback (Are HDB residents tenants or owners?; June 24).

Purchasers of HDB flats are owners of their property.

Flat owners enjoy rights to exclusive possession of the flat during the tenure of the flat lease.

They can sell, let out, and renovate their flats, within the guidelines specified in the Lease and Housing and Development Act.

Legally, you are a tenant. But HDB can call you whatever they want, including a "owner". It has no legal basis and does not confer you the rights of an owner. Except when it comes to property tax, yes, you are an owner. :lol:

A look at housing loan

As of 1st July 2017.

Outstanding$258,379.57
Years left19 years
Installment$1,376
Interest (last month)$465.10
Effective interest rate2.16%
Interest-to-installment ratio33.8%
Used CPF amount$124,905.78
Accrued interest$15,152.43

Questions:

  • Should I refinance?
  • Should I pay down this debt?
  • Should I use cash or CPF to make principal payment?

Refinance

To 19 years. Never ever refinance longer than existing tenure.

RateInstallmentSavings
1.3%$1,277.69$22,415.62
1.5%$1,301.18$17,058.89
1.7%$1,324.94$11,640.69
1.9%$1,348.98$6,161.17

There is significant savings (in interest) if interest rate is 1.5% and below. Will it hold long term? If so, it is worth refinancing.

Principal payment

This table shows the balance at the start of the year, the payment for that year and the principal/interest breakdown.

YearBalancePaymentInterestPrincipal
5279,66016,5775,93610,641
7258,14516,5775,46611,110
9235,68216,5774,97611,600
11212,22816,5774,46512,112
13187,74016,5773,93012,646
15162,17116,5773,37313,204

It is better to make principal payment in the early years as the interest component is higher. Every ~$20k+ cuts two years from the tenure.

Cash or CPF

The issue with using CPF is accrued interest. Allowing the use of CPF for housing loan is a stealth-of-hand to make you pay interest for your retirement savings. (What a scam. :lol:)

The accrued interest goes away only after you sell the flat. You need to repay the CPF loan + accrued interest. It is fine if you are not able to repay the interest in full; it is always zero'ed out.

If you don't sell your flat, the accrued interest accumulates at 2.5% compounded. In my case, I'm adding $3.5k of interest yearly. That is $3.5k CPF Board should be paying me.

Previously, I thought of using CPF to make principal payment. Now I think it is a very bad idea.

Takeway

As always, the key is to minimize the interest.

Right now, I'm paying interest of ~$5.6k to the bank vs $3.5k to CPF. It makes sense to reduce the bank loan now. At some point, I will pay more interest to CPF, then it is better to pay down the CPF loan.

You can repay your CPF loan and accrued interest using form HSD/VR (Application to Make Cash Refund of CPF Savings Used for Property).

Claymore, Battle of the North

Claymore is Berserk-lite. If you like one, you'll probably like the other.

Claymore manga is complete at 27 volumes. According to the Wiki, the first English volume was released in Apr 2006 and the last volume Oct 2015! (Original Japanese run was May 2001 to Dec 2014.)

Claymore is a Shonen fighting manga, hence it has many of the genre's traits: barebone story, many long fight scenes, lots of cannon fodder, use of deus ex machina to get out of impossible situations, but also surprising revelations and plot twists to keep the story going. The last point is mostly the reason I still read this type of manga; I don't like it as a rule.

(If the fight scenes are condensed, I'll say there are just 12 volumes worth of story. In comparison, Berserk feels like 80% fighting and 20% story.)

The question as always is, is it done well enough?

For Claymore, the answer is yes. It is light on gore, the story is not too bad, the bad guys have convincing motivations, and the plot twists are plausible. I rate it B-. Go read it now! :lol:

The Battle of the North is the mid-way point of the series. It spans from Vol 9 to 11. 24 out of 47 warriors of the Organization were deployed to the North. They were: #6 (Miria), #8 (Flora), #9 (Jean), #11 (Undine), #13 (Veronica), #14 (Cynthia), #15 (Deneve), #17 (Eliza), #18 (Lily), #20 (Queenie), #22 (Helen), #24 (Zelda), #27 (Emelia), #30 (Wendy), #31 (Tabitha), 35 (Pamela), #36 (Claudia), #37 (Natalia), #39 (Carla), #40 (Yuma), #41 (Matilda), #43 (Juliana), #44 (Diana), #47 (Clare).

This is important because we want to know who did not participate, why and if they turned up later: #1 (Alicia), #2 (Beth), #3 (Galatea), #4 (Ophelia), #5 (Rafaela), #7 (Eva), #10, #12, #16, #19, #21, #22, #25, #26, #28, #29, #32 (Katea), #33, #34, #38, #42, #45, #46.

#1 and #2 were the Organization's secret weapons. They showed up right after the battle to mop up the remaining enemy forces.

#3 was "the eyes" and was observing the battle from afar with #1 and #2.

#4 had died earlier. Apparently she had not been replaced.

#5 was looking for her sister.

#7 just bought it in the initial invasion, together with Lucia, Kate and an unnamed warrior. They should be ranked in the 10s and 20s.

Raquel and an unnamed warrior died earlier around the same time as #32. Raquel should be ranked in the 20s and the unnamed warrior in the 10s.

Out of the remaining 23 warriors not sent to the front, only 15 were alive. We saw 4 of them (#1, #2, #3, #5), the rest did not appear.

There was a time skip of seven years after the battle. How many unnamed warriors carried over? It was not mentioned. We were simply introduced to a new generation of warriors.

Standard Chartered Bonus$aver, worth it?

Here's what you get:

  • Base: 0.1% below $200,000 balance
  • Spend: 1.78% on first $100,000 if $2,000, 0.78% if $500
  • Salary: 1% on $3,000
  • Invest: 0.75%; $12k annual insurance premium or $30k unit trust
  • Payment: 0.25% if three $50 bill payments

It is easy to aim for 2.13% or 3.13%. Assume $100k, is it worth to spend $2k each month to get the latter?

$100k @ 2.13%, $2,130 interest, minimum spending $6,000 ($500 per month).

$100k @ 3.13%, $3,130 interest, minimum spending $24,000 ($2,000 per month).

Worth it? You be the judge.

OCBC 360

OCBC 360 has been revised down a couple of times.

  • Base: 0.05%
  • Spend: 0.3% on $500
  • Salary: 1.2% on $2,000
  • Invest: 0.6% or 1.2%
  • Payment: 0.3% if three bill payments total $150

Most probable case is 1.85% up to $70k. Interest is $1,295.

If spend $600 on OCBC 365 CC, will get 3% rebate ($18) on eligible transactions, 0.3% otherwise. Assuming 50% rate, that's $108 a year.

UOB One

UOB One has the simplest criteria, but is tiered and up to $50k. The effective interest rate is 2.432%. Interest is $1,216.

Caveat

All three requires you to spend at least $500 per month, or $6,000 a year. There is no free lunch — you will be spending more than you earn in interest. The break-even point is around $400 per month. If you spend less than that, it makes more sense to forgo these tiered accounts.

Raspberry Pi 3 I/O achilles heel

The Raspberry Pi 3 is a fantastic board. It has quad-core 1.2 GHz Cortex-A53 CPU, 1 GB RAM, HDMI, 802.11n WiFi (2.4 GHz), Bluetooth, Fast Ethernet and 4 USB 2.0 ports. All for US$35.

It is perfectly usable as a low-end PC. What's the catch?

Weak I/O. It only has one USB 2.0 bus to its Ethernet port and 4 USB ports, for a total bandwidth of around 30 MB/s.

It is unsuitable as a network file server. It can only serve files at 10 MB/s due to its Fast Ethernet port. If we use a USB-to-Gigabit Ethernet adapter, it goes up to 20 - 30 MB/s, still too slow.

An ideal low-cost headless file server:

  • Dual-core 1.0 GHz ARMv8
  • 2 GB RAM
  • Gigabit Ethernet
  • 2x USB 3.0
  • 2x SATA

Raspberry PI 3 has another weakness out-of-box: it uses SD card by default. SD cards have extremely poor write speed and simple wear-leveling algorithm. But it can be remedied to some extent.

Some people switch to using USB flash drive, but I prefer to stick to SD card if possible. It is easy to find Raspberry PI 3 SD card benchmarks. From there, we can buy the SD card with the best 4 kB random write speed.

Unfortunately, there is no public information on wear leveling. You have to tease the information from the marketing material. All SD cards should have wear leveling to some extent, but I suspect only the best SD cards use static wear leveling. Static wear leveling is extremely useful because it spreads the wear evenly across all blocks, whether used or unused. Dynamic wear leveling uses only unused blocks.

For cheaper SD cards, simply use a larger capacity SD card and leave plenty of free space, e.g. if you use 4 GB, pick a 16 GB SD card instead.