She is about 31. She and her partner are looking at buying their first home together, and a child might not be in their plans at this point.

With where home prices are right now, and given they need space, but not too much space, even 2-room HDB flats are viable. This is possible if one turns 35 soon.

At some point as a single, I definitely find it daunting trying to get a HDB flat all by myself. The prudent part of me zoomed into 30 plus-year-old HDB flat back then that would cost $280,000 to $340,000 as someone turning 35. My friend and her partner may be okay with my choices back then but it will be a struggle to get these prices (or is it?)

I do understand the worry, and therefore their choice of choosing something smaller.

But I do think that if we put the number together, they should not let that worry affect their life so much.

There are choices to be make but they can get a decent outcome.

I don’t think a 2-room flat is the only answer and can be broader. I think there’s a sweet spot, a slightly bigger resale flat that still fits within what most of us would call prudent financial planning.

So this article is for her and her other half. It’s also for anyone in a similar spot who looks at today’s prices and quietly wonders whether they will ever be able to afford anything more than the smallest flat.

I’ve tried to keep it data focused. You can piece the numbers together yourself and decide if they fit your life.

Why 2-room flats are so popular now

There’s a reason why demand for 2-room flats has been so strong. It’s not the only reason, but it’s a big one.

A 2-room flat is cheaper in every way that matters up front:

• Lower price. Over the past 12 months, a 2-room flat that was 35 years old or younger sold for about $380,000 in the middle of the range.
• Smaller loan. On an HDB loan, the monthly instalment is about $1,300, compared with about $1,800 for a $520,000 flat.
• Less cash up front. The downpayment and stamp duty come to about $100,000, compared with about $140,000 for a $520,000 flat.

If you are nervous about money, those differences feel huge.

But there’s a trade-off that I think is obvious to many.

A typical 2-room flat is about 47 square metres. A 3-room is about 68 and a 4-room is about 93. A couple will really have to know their needs well. I always feel that we overrate how we expect our decisions not to change and because of that, many regret getting a smaller place.

Some couples who were sure about not having children change their minds. Some end up caring for a parent. Some just find out, a few years in, that two people working from home in 47 square metres is harder than they thought.

A Citizen-PR Couple Gets About $90,000 in Grants on a 3- or 4-Room Resale

My friend is a permanent resident (PR). Her partner is a Singapore citizen. That combination comes with a lot of confusion about grants, so let me lay it out.

The Enhanced CPF Housing Grant, or EHG for short, is the grant that depends on your income. The lower your household income, the more you get, up to $120,000. Once your household earns more than $9,000 a month, the EHG drops to zero.

Here’s what a first-time citizen and PR couple can get when buying a resale flat:

• Family Grant. $70,000 for a 2-room to 4-room flat, or $40,000 for a 5-room or bigger. A citizen couple gets $80,000 and $50,000.
• Proximity Housing Grant. $20,000 if you buy within 4 km of your parents, or $30,000 if you live with them. It doesn’t depend on your income, and there is no reduction for a PR.
• Enhanced CPF Housing Grant. Up to $120,000, depending on income. Zero if the household earns more than $9,000 a month.

I will use a ballpark $5,000 a month before CPF deduction each to estimate what my friend and partner are currently earning each. That’s $10,000 a month for the household. That’s below the $16,000 income ceiling for buying a flat with an HDB loan, so they qualify. But it’s above $9,000, so they get no EHG.

So if they buy a 3-room or 4-room resale flat near their parents, they could get about $90,000 in grants.

To put that in perspective, a $520,000 flat would effectively cost them about $430,000.

A few things to know about the grants:

• They go into your CPF. The grant is paid into your CPF Ordinary Account (OA). It reduces how much you need to borrow or pay up front.
• You return them when you sell. If you sell the flat, the grant goes back into your CPF with interest. You’re then treated as a second-timer, which affects grants for your next flat.
• Income is averaged. HDB looks at your average gross monthly income over the 12 months before you apply. But there are nuances so best to check with HDB or all these will likely be sorted out if you apply HFE.
• Marriage timing matters. If you buy as fiances, you must register your marriage before, or within 3 months of, the date the purchase completes.

There’s one more possibility. If the PR becomes a citizen while they still own the flat, they can apply for a $10,000 Citizen Top-Up.

I touch on how much grant that is available to them if they decide to go for it but in most of my numbers later, I remain conservative not to include any grants. I want to show that the math works even without them.

If my friend does get $70,000 to $90,000 in grants, that’s a bonus on top.

Why I use $520,000? 424 4-Room Flats Sold at $520,000 or Less in last 12 months

There is a reason why I use $520,000 in the example before.

I went through every HDB resale transaction in the past 12 months, from November 2025 to October 2026. That’s about 25,500 flats. I kept only flats that were 35 years old or younger.

Why do I consider only the flats that are 35 years or younger?

There’s a rule that the remaining lease on the flat must cover the youngest buyer until age 95. If it doesn’t, you can’t use your CPF and get the full HDB loan.

Here’s what that means in practice. My friend is 31, so she needs a flat with at least 64 years left on its 99-year lease. In other words, a flat that is no more than 35 years old.

So I kept to flats that would work for a young buyer.

Here’s what the data showed:

A quick word on these three numbers, because they help you read the tables later.

The middle price, or median, is the price where half the flats sold for less and half sold for more. The 25th percentile is where a quarter of the flats sold for less. The 75th percentile is where three quarters sold for less.

So if a street’s middle price is $520,000, its 25th percentile is $490,000 and its 75th percentile is $560,000, most flats on that street sold somewhere between $490,000 and $560,000. If you see one listed at $620,000, you know it’s at the pricey end and you should ask why.

Here’s why $520,000 is a useful number:

• For a 3-room, it buys a flat in the cheaper half. It sits just above the cheaper quarter of sales. It isn’t a bargain hunt, and it isn’t a stretch either.
• For a 4-room, it’s the cheaper end. About 1 in 5 of the 4-rooms sold under $600,000 went for $520,000 or less. That’s 424 flats in a year.
• For a 2-room, it’s well above the typical price. $520,000 is about 1.4 times what a typical 2-room sells for.

So $520,000 is roughly where a younger 3-room in a more affordable town and an older 4-room both start to show up.

Where do these flats sit?

1. For a 3-room, the middle price was close to $520,000 in Sembawang, Bukit Batok, Jurong West, Woodlands and Bukit Panjang. Sengkang and Punggol were a bit higher, at about $540,000 to $545,000.
2. For a 4-room at $520,000 or less, the most choices were in Jurong West, Bukit Panjang, Choa Chu Kang and Woodlands.

I’ve put the full lists at the end of this article, street by street, with the blocks that changed hands and their postal codes. If you are house hunting, that’s probably the most useful part.

A flat is pricey mostly if you cannot get a decent life servicing the flat. So we need to examine can my friend get a decent life.

Now for the part I care about most. If my friend and her partner buy a $520,000 flat, can they still live decently, save, and eventually stop working?

I ran this through my own cash flow planning tool Fenrir.

I made some assumptions about their life situations. I tried my best to keep things as simple and as conservative as I could make them:

• Age. Both are 31 today. I treat them as the same age to keep it simple.
• Income. Each earns $5,000 a month before CPF. That’s $4,000 a month take-home after their own CPF contribution.
• Bonus. Each gets a $12,000 yearly bonus before CPF, which is about $9,600 after CPF.
• Pay rises. Their pay goes up 3% a year. That’s about the same as inflation, so in real terms they never get a raise. That’s conservative.
• Spending. Each spends $3,000 a month, including tax. That’s $6,000 a month for the two of them. It also rises 3% a year, so they keep the same lifestyle all their lives.
• Investing. Each puts $1,000 a month into a combined investment portfolio, with their contribution rising 3% a year each.
• Starting savings. Just $5,000 in the portfolio between them. I wanted to start from almost nothing.
• Home. They eventually buy a $520,000 flat at 35 years old servicing a HDB loan.
• Retirement. They stop working at 65. The money needs to last until 100.
• Returns. The portfolio earns 7% a year. I’ll also show 5% a year. Cash in the bank earns almost nothing.

One important thing to notice. Every income number here is take-home pay, after CPF. The money going into CPF isn’t counted as spendable. That matters a lot for the mortgage, as you’ll see.

I tried to visualise the cash flow in a diagram below:

There are two moments in this plan where things change.

At 35, they buy the flat.

At 37, once things have settled down at home, each puts another $1,000 a month into the portfolio.

Where does that come from? Mostly from the bonuses, and from any pay rises they choose not to spend. Not everyone will be able to do this, and that’s fine. I’ll show what happens without it too.

How the flat gets paid for

Let me walk through the purchase at 35.

A $520,000 flat with an HDB loan needs:

• A downpayment of 25%. That’s $130,000.
• Stamp duty of about $10,200. That’s the tax you pay when you buy a property.
• A loan for the other 75%. That’s $390,000, at 2.6% a year over 25 years.

So they need about $140,000 up front, or about $70,000 each.

Here’s the part many young buyers don’t realise. Most of this can come from CPF.

The CPF Ordinary Account (OA) is the part of your CPF that can be used for housing. Every month, a slice of your salary, and of your employer’s contribution, goes in there.

By 35, after four years of working, his partner’s CPF OA is enough to cover his half of the downpayment and more.

My friend would see CPF OA contribution later but given her timeline, she should have $30,000 in her CPF OA by 35 years old to pay for part of the downpayment.

The rest, about $40,000, comes from cash they have saved.

The $40,000 doesn’t come out of the investment portfolio. It comes from the bonuses they kept as cash in the years before buying.

Then there’s the monthly mortgage.

On a $390,000 mortgage at 2.6% over 25 years, it’s about $1,800 a month.

That’s $900 each.

And here’s the good news. Each of them has more than $900 going into their CPF OA every month. So the whole mortgage can be paid from CPF, without touching their take-home pay.

In the plan, I show the mortgage as $0 per month out of their pocket. It’s not that there’s no mortgage. It’s that CPF pays it.

Take-Home of $8,000 a Month Is Almost All Spent or Invested

Here’s what their money looks like every year, from 31 to 100.

The dark bottom layer is spending. The green layer is what they invest each month. The light blue layer is the extra they invest from 37. The line on top is their take-home pay plus bonuses.

Notice how little space there is between the line and the stack. Almost everything they earn is either spent or invested. The bonuses are the only buffer.

That’s what I mean when I say this plan is prudent, but it isn’t loose. There’s not a lot of fat in it.

After 65, there’s no more pay. Everything they spend comes out of the investment portfolio.

The Portfolio Reaches About $529,000 at 40, Matching the Flat Price

This is the chart I most wanted my friend to see.

Here’s roughly what the investment portfolio looks like at different ages:

Look at age 40.

The money they have quietly built up is about the same as the price of the flat they bought five years earlier.

That’s what changes how the $520,000 looks.

At 31, half a million dollars feels like the biggest decision of your life.

At 40, it’s about what you have sitting in your investment account.

At 50, it’s a fraction of it.

At 65, $6,000 a Month in spending becomes $16,400. Does the Portfolio Last to 100?

When they retire at 65, their spending has grown with inflation from $6,000 a month to about $16,400 a month. That’s about $197,000 a year, and it keeps rising 3% a year after that. It sounds like a lot, but it buys the same lifestyle as $6,000 a month does today.

So can the portfolio support that until 100?

If we use straight-line cash flow mathematics it would. The following 3 chart laid out comes from Fenrir:

The top chart shows my friend’s asset growth and it is all the investment portfolio because the cash in the wallet is exhausted. The middle chart shows the income they earn, which ends at 65 years old with a black line showing their spending. You can see the spending extends beyond their work income years.

The bottom chart shows the spending going up with inflation.

Yet despite increase inflation-adjusted spending, the portfolio value… just gets stronger.

Now we might think that this is because we use 7% p.a. return so here is the top chart, but with 5% p.a. return:

We can then see the portfolio feeling more stress if the returns are lower, while still need to provide inflation-adjusted income.

I can go on and on about this area but that is not the point.

The point is that… you can have a decent outcome if prudent financial actions are done over time.

8 Things the Plan Leaves Out, From Grants to Children

There are a few common financial planning stuff that I did not factor into this cash flow assumptions.

Her actual plan can be much better than what I just projected.

• Grants. I left out the $70,000 to $90,000 they could get. If they get it, they have more room.
• The rest of their CPF. Apart from paying the mortgage, I ignored their CPF completely. That includes CPF LIFE, which will pay them a monthly income from 65. In real life, it would cut how much they need to take from the portfolio.
• Renovation. I left this out on purpose because it deserves its own article. For what it’s worth, they’d have about $60,000 to $80,000 in cash at 35 and 36 after the downpayment. A $40,000 to $50,000 renovation fits, but it uses up most of their buffer.
• Rising flat prices. I assumed the flat still costs $520,000 when they buy at 35. If prices rise 3% a year, it would cost about $585,000. Grants would more than cover that difference.
• An emergency fund. The only cash they hold is the leftover bonus money. If they spend it on renovation, they have nothing set aside for emergencies.
• Running costs of the flat. Service and conservancy charges, property tax and repairs aren’t counted separately. They would have to fit inside the $3,000 each.
• A career break. If one of them stops working for a few years, the plan would need to change, especially the mortgage from 56 to 60.
• Children. If they change their minds, everything changes.

I share these so you can see the gaps. A plan with gaps you know about is better than one that pretends to be perfect.

If you change your mind later

Say they buy a 2-room flat now, and in a few years decide they want something bigger. What does that involve?

First, there’s the minimum occupation period, or MOP. It’s the minimum time you must live in your flat before you can sell it. For a resale flat, it’s 5 years.

So if they buy at 35, the earliest they can sell is 40.

Then there are the costs of moving:

• Agent’s fees. If you use an agent to sell, expect to pay about 1% to 2% of the price [CHECK: typical rates].
• Legal fees. A few thousand dollars on each side [CHECK].
• Stamp duty on the next flat. About $15,600 on a $700,000 flat.
• CPF refund. Any CPF you used for the first flat, plus the interest it would have earned, goes back into your CPF when you sell. It doesn’t disappear, but that money is no longer free to spend.
• Grant refund. If you took a grant, that goes back too, and you lose first-timer status for the next flat. That can mean smaller grants, or none at all.
• Moving and a second renovation. You’ll likely renovate twice instead of once.

None of these are deal breakers. But they add up. Moving up later is possible, but it’s not free.

That’s why I think it’s worth asking now: is a little more space today cheaper than moving twice?

This is not supposed to be perfect

I know there are a lot of assumptions in here. Their income may not grow 3% every year. Their spending may not stay at $3,000 each. Life will throw things at them that no cash flow model can see coming.

But that was never the point.

The point is that a flat costing over half a million dollars looks scary. It looks even scarier when you hold it up against what our parents paid, or what my best friend paid for his flat.

So try looking at it another way.

If the two of them put away what they can, steadily and prudently, by 40 the money they have quietly built up is about the same as the price of the flat. By 50, it’s three to four times that.

When they are 45 and look back at that $520,000, I don’t think it will feel the same way it feels today. It will still be a lot of money. But it will no longer feel like the thing that decided their whole life.

And they are not carrying it alone. It’s two people committing to it together.

What I hope you take away

If there’s one thing I hope you take from this, it’s that there is hope. I don’t mean the hope you get from a perfect model. I mean hope for a situation very close to your own.

And if your picture looks different from the one here, I hope some of the lessons still carry over.

Take a look at the tables below. These are the streets where a 2-room, a 3-room or a 4-room has actually changed hands in the past year, and at what prices. Go and walk around a few of them. See if any feel like a place you could call home.

That might be a better starting point than the price tag alone.

The potential sweet spots of HDB Flats that my friend would want to pay attention to

I’ve put together a list of past resale transactions, and I’ve grouped them by the room types of HDB flats. I think these groups are the ones worth paying attention to if your friend and their partner are looking for a resale place to stay. I see them as sweet spots because they most likely won’t cost as much as some other options.

They’ll still have their flaws, but some of those flaws might be ones your friend and their partner can live with. Every couple’s situation is different, so what works for them is up to them. I don’t know how unique their situation is, but I feel there’s a good mix in the list, so there might be a few that suit them.

2-Room HDB Aged 35 or Younger: Medians of $352,000 to $426,000 in 12 Towns

The 2-room HDB are pretty young and most cannot bring themselves to pay these kind of prices. You be looking at 380k but perhaps there may be some 340k deals out there

Here are the flats by street

3-Room Resales Under $650,000: 91% in Choa Chu Kang Sold at or Below $520,000

They are also younger than 35 years old. This leaves out the pricier central towns like Queenstown, Toa Payoh, Bukit Merah, Kallang/Whampoa, Geylang, Clementi and Ang Mo Kio.

You won’t have the older 3-room HDB that Kyith used to stay at in this range. They are pretty new but also smaller.

I felt that this may be the sweet spot. There may be some areas that my friend would like.

4-Room Resales: 33% in Jurong West Sold at or Below $520,000

If we use a ball park $520,000 there may be some 4-room HDB that will fit the criteria as well.

You will actually be looking for the special deals that is in the lower range for Jurong West, Woodlands, Sembawang, Choa Chu Kang that are 20 plus year old.

My reader shared his intention to invest the $50,000 from his paid-up policy into perhaps VALU, the Vanguard FTSE Global All-Cap UCITS ETF. I reminded him that he should consider the role of his paid-up policy and whether he has protection needs.

We had a lengthy exchange and the conversation eventually pivot to how he visualise his current setup, with a view of a no-work-income lifestyle soon.

We started talking about his $165,000 mortgage on a condo he owns in Singapore.

I was trying to give him some useful pointers but I thought its good to let readers in on how I see this “having a mortgage in a no-work-income financial independent” situation.

This is where I landed.

Let me give you my model answer for myself, but also for most people, knowing what they really want.

Before I go on, one caveat.

The right answer depends on what the loan is for. A mortgage on the home you live in is a different problem from a mortgage on an investment property.

My reader’s loan is actually on an investment property. I’ll come back to that later.

But when most people ask me about this, they mean the home they live in.

So let me give you my model answer for that.

Say there is $165,000 left on the loan of the home you live in, and you are heading towards a life without a salary.

1. Set that $165,000 away from the money that provides you income. Keep it in its own pool, away from the money that pays your income. A separate account is fine. Don’t worry about the return on it. Its only job is to be there, so you could clear the loan on any day you choose.
2. Whether you really pay this $165,000 off is up to you. Knowing that you can is what gives you peace of mind.
3. For your income portfolio, calculate how much to Fund your must-pay costs carefully. These are the costs you can’t cut even in a terrible year. For these, I use a withdrawal rate of 3%. If you want income that can last forever, like I do, use 2.5%. Keep the money in a diversified mix, with more shares if it has to last a long time.
4. For your income portfolio, calculate how much to fund your flexible spending separately. Holidays, eating out, the nicer things. You can cut these when markets are bad, so 4% or 5% is reasonable.
5. Add up the money needed in #3 and #4 all up. If you have enough, you are financially independent in a careful, conservative way.

Here is a made up example, just to show how it works:

• Must-pay costs of $3,000 a month, or $36,000 a year. At 2.5%, you need $1.44 million.
• Flexible spending of $1,000 a month, or $12,000 a year. At 4%, you need $300,000.
• Plus the $165,000 set aside.
• Total: about $1.9 million.

If you’re comfortable using 3% for the must-pay part, that $1.44 million becomes $1.2 million, and the total drops to about $1.67 million.

So the more careful you are, the more you need. That’s the price of sleeping well.

If you have this, you could wake up tomorrow, clear the loan, and still have enough income. Or you could keep the loan. Either way, it’s a choice, not a burden.

That’s the plan I think everyone should aspire towards.

But I know it’s not how most of us think.

And I get why.

So let me walk through the choices people usually wrestle with, and why I ended up here.

You never end this game of money/return optimisation

Most of us have spent our whole lives optimizing. Every dollar had a job. The best choice was the one with the better number.

So the first thing many people ask is this.

Should I pay off the loan, or keep the cash and earn more than the loan costs me?

Say your loan costs 1.65%. A safe high yield deposit pays 2.5% for three years. On $165,000, you earn about $1,400 a year more than you pay.

“Nice. I’m ahead!”

Then three years pass.

“Okay, what now? The deposit is ending. The loan is resetting. Can I still beat it? Which bank? Should I just pay it down?”

A mortgage doesn’t last three years.

It can run 8 years, even 15. So you keep coming back to the same table, again and again.

Will you come out ahead over 15 years?

Honestly, I don’t know. I haven’t done enough research on something this niche, and an average return won’t tell you either. Some stretches will go your way. Some won’t.

What I do know is that the game doesn’t end. You pay for it with your attention and your worry.

I call that mental load.

And I think it’s the cost most of us forget to count.

2022: the year mortgages became a lottery

I have never taken a mortgage of my own. But for about 20 years of my working life, I have been paying off my parents’ mortgage.

My loan is fixed, and it’s smaller than what many of my friends carry. When the amount is small, you don’t feel the weight as much.

So in 2022, I barely felt a thing.

But I watched a lot of people go through it.

Did you manage to reprice or refinance a few months before rates shot up? (That just means moving your loan to a new rate when the old one ends.) If you did, lucky you. If you didn’t, you were suddenly paying a lot more.

You can be the same person taking the loan but the bill is different just by timing.

It felt like a lottery.

Picture someone paying $4,000 a month. Then it becomes $6,000. That’s $2,000 more, every single month.

Is $2,000 a lot? It depends on the person. But I think the real shock wasn’t the number. We had lived with low rates for so long that nobody could imagine high ones. Some part of us hoped they would never come.

Higher rates eventually came back.

And we actually see people paying off their loans. I had heard people say they might, if rates ever went up. But only in 2022 and 2023 did I see it happen, for real.

People who had borrowed against their insurance policies felt the same squeeze.

And everyone asked me the same thing. “Kyith, how long will rates stay this high? Will they come down?”

Rates are still much higher today than back then. And the honest answer is that nobody knows where they go next.

So I would never build a retirement plan that needs rates to stay low.

Why do I keep the loan out of the income pot

People often send me their plan. It usually sounds like this:

“Kyith, I’ve reached $3 million. At 3%, that’s $90,000 a year. That covers my living costs and my mortgage. I’m financially independent!”

Hmm.

I never quite know how to say this kindly and yet succinctly enough to them.

Here’s what I see.

The money that could have cleared the loan is sitting inside the income pot. A bigger pot feels good. The income looks higher. And the hope is that the income can keep pace with the mortgage payment.

But it’s two problems tied into one knot.

Go back to that $4,000 payment. That’s $48,000 a year, more than half of the $90,000. When it jumps to $6,000, it’s $72,000. That’s 80% of the income, gone to the bank.

They were financially independent in that snapshot but when one number variable change, their whole financial independent plan wobbled.

Here’s why mortgage matters so much.

Some costs you can bend. Groceries, you can trim. A holiday, you can skip. But a mortgage? You can’t shop for a cheaper one halfway. Loans from different banks look quite alike. You can choose fixed or floating, or pay part of it down, but the bill still comes every month.

That makes mortgage usually a must-pay cost.

And whatever strategy you use, dividends or a withdrawal rate, it all comes back to one question. What income can I safely take, even in the worst years? That’s what a safe withdrawal rate tries to answer.

So the more must-pay costs you have, the more money you need behind them.

• At 3%, a $4,000 monthly payment needs $1.6 million behind it.
• At $6,000, it needs $2.4 million.
• The rate move alone added $800,000 to what you need.

“But Kyith, a mortgage ends!”

True. It’s not forever, like food.

And that is the second reason why I prefer to consider the money meant for mortgage as a separate pool.

It takes the mortgage out of your must-pay costs. The money you need gets smaller, and so does the knot.

What if the loan is on a property you don’t live in?

Now, back to my reader.

His HDB flat is fully paid. His $165,000 loan is on a small condo he bought around 2015, one of those shoebox units.

I didn’t ask him if he rented it out. But I have a feeling he’ll be the first to tell me, “Kyith, managing tenants is not as easy as you think.”

I’ve always said tenants are not everyone’s cup of tea. It means a lot more when someone who has lived it says so.

The money side is simple enough. Rent, less the usual running costs, is your cash flow. If it covers the loan, nothing comes out of your pocket.

The harder questions are about the life you want.

• If something goes wrong, is your plan to sell the property and use the money to clear the loan?
• Do you even want to own it in that stage of life?

And to judge whether it was a good investment, put all the money that went in and came out together. It boils down to one yearly return. (It’s the same XIRR idea I used in my policy article.) Then compare it with what a deposit, a bond fund or a share fund would have given you. That’s the fair test.

People like to make their Financial Independence setup more and more complicated.

Some people have more than one property. As they get close to stopping work, they sometimes sell one and buy another. Or bring two into one.

That needs cash flow planning, and not just for this year. How much down payment will the next one need? How much will the old one bring in after costs? How many years of the new loan will you be paying in retirement?

Some people get an adviser to work this out. Some prefer to do it themselves. Both are fine.

But before the next move, I’d pause and ask two simple questions.

Am I creating problems for myself?

And does this still take me towards the life I want?

I don’t think we ask that enough.

So where does this leave us?

If I had to boil it down, I’d ask you three things.

• How much peace of mind do you want, and how much mental load are you willing to carry? Beating the loan is a game with no finish line.
• Is your mortgage tangled up with the money that pays you? If it is, try to untangle it now, while things are calm.
• What’s the ideal setup for the life you want? Even if you were retrenched tomorrow. Do you want an investment property? Do you want a mortgage? Do you want to be mortgage free? Start there, and the big decisions become a lot clearer.

You don’t have to pay off your mortgage.

You just need to be able to.

This is how I think about it.

Global Testing shares jumped about 110% in a day.

I didn’t even realise the price moved so much and only got to know about this when my broker Standard Chartered inform me about corporate action that needs to be done.

Some of my close investing friends was posting money bag GIFs and I got a shock its so much.

Global Testing is in Daedalus Income portfolio. So why wasn’t I able to feel the difference?

That’s because I own very little of it now. Just 50 shares. But Global Testing has been part of my portfolio on and off since 2015, and a lot of people have already shared their own stories about this stock.

So this post is simpler. I want to put my numbers and dates down in one place, so everyone can see them clearly.

Why I own what I own, and what happened

I first bought in May 2015, about S$6,900 worth, when the stock was priced at S$0.14.

When I first bought this, I liked it because it was a company with net cash, and I thought it was sitting at a decent value point. It was always trading at about a 10% dividend yield. When I looked at the business, I felt the earnings could cover the dividends.

That said, I can’t claim I knew a lot about it.

Back then, you really couldn’t find much information. All I knew was that it was a testing firm operating in Taiwan, and that’s all. Because of that, I never put a big chunk of my capital into it. I mainly wanted to watch it play out over time, and I figured I’d collect some dividends along the way. I also paid attention to how the company handled dividends and, at times, capital reductions.

In the end, I decided to sell. The main reason was that it’s a very illiquid counter. Most days, it felt like I was the only one buying, and if I wasn’t, there wouldn’t be much trading at all. So I always worried that if the company announced a poor dividend, the price would drop, and I wouldn’t be able to get out in time. Because of that, I decided I was done with this company.

What does Global Testing (Ticker AYN) do?

Global Testing tests computer chips for other companies.

A wafer is a thin slice of silicon with many chips printed on it. Before those chips end up in a phone or a car, someone has to check that each one works. That’s the job. Global Testing doesn’t design chips. It’s the quality check station.

Global Testing does wafer sorting and final testing, mostly for chips used in consumer electronics and communication devices, and more recently for cars.

The head office and factory are in the Hsinchu industrial park in Taiwan. The main customer markets are Japan, Taiwan and Singapore. It has been listed on the Singapore Exchange since 24 August 2005.

Global Testing… is a very Small firm

Just before the offer, the market valued the whole company at about S$34 million.

• Revenue: about US$46 million in 2025.
• Net profit: about US$3.9 million in 2025.
• People: one data site puts the headcount at 567.
• Shareholders: about 2,500, with just over half of the shares in public hands.

I would have reckon the subsidiary where some of you work at moved more money than this.

Who owns Global Testing

As of March 2024:

• Yageo group: about 29% (23.64% directly and 5.28% through Kuo Shin Investment).
• Chen Tie-Min: 10.87%.
• Chia Soon Loi: 7.25%.

Mr Chen is a Senior Executive Director of Global Testing. He is also the chairman of Yageo and of Tong Hsing, the company now making the offer.

16 years of Global Testing Earnings

A short history of its revenue and earnings from 2015 to 2026

• May 2015. Every 20 old shares became one. The company also paid S$0.075 per old share, which worked out to about S$1.50 per new share and about US$40 million in total.
• 2016 to 2017. A payout of about S$0.20 a share each year, a mix of capital reductions and dividends.
• 2018. Revenue fell 18.9% to US$22.8 million because customers placed fewer orders. A profit turned into a loss of about US$5.0 million. No dividend was recommended.
• 2019. Another loss, of US$3.3 million.
• 2020. Covid hit. The share price touched S$0.26 in March.
• 2021 to 2022. Revenue jumped 69% to US$38.7 million in 2021, helped by strong chip demand in Taiwan and higher testing prices. Net profit went from US$0.4 million to US$9.4 million. In 2022 it reached US$11.2 million.
• 2023. Revenue fell 12.5%. Net profit fell 69.7% to US$3.4 million.
• 2024 to 2026. Revenue recovered to US$46.4 million in 2025, but net profit fell 24.8% to US$3.9 million. In the first half of 2026, the company put surplus cash into investment-grade bonds.
• November 2026. Tong Hsing offered S$2.30 a share in cash.

The company reports in US dollars. The table below is in Singapore dollars, converted by TradingView, so the numbers will look a bit different from the company’s own announcements (for example, the 2018 loss of US$5.0 million shows here as S$6.7 million).

Actually, if we look at it, the earnings is not so consistent in the past, and perhaps that’s back in 2015 when I was thinking through that. I wonder whether the whether you have the cash flow to actually pay that pretty significant dividends itself.

And I eventually sort of with all of these things, you sort of like be on the fence that at a certain stage, most likely something that you have quite confidence that the cash flow will be there to pay your dividends, but you’re not sure whether it will drop off or not.

Usually that is a case of mindset of actually going in like this and seeing how it actually turns out.

My own transactions

I think with Fenrir, my application to do this kind of stock portfolio tracking I can actually have better detail to actually show people like kind of my past transaction so over here you are able to see the color-coded table of my transactions:

A few things to know about my records:

• My two May 2015 buys were at S$0.14 a share, before the 1-for-20 consolidation. That’s S$2.80 per new share.
• Bought: S$27,094, including fees.
• Sold: S$21,217, after fees.
• Dividends and capital reductions: S$9,086.
• Still held: 50 shares, worth S$108 at the S$2.16 market price, or S$115 at the S$2.30 offer.
• Net gain: about S$3,300, or 1.12 times my money.

The chart below shows when the money went in and out.

I sold 18,900 shares between August 2017 and April 2018, at prices between S$1.03 and S$1.27, an average of S$1.125.

I think that’s a nice shot to see prices and when we sell it. I think maybe next time I will kind of like do this more because it’s not something that I was able to do to actually help relate when did all these transactions happen about the prices. So sometimes some of these LLMs also teach me a thing or two about what are good illustrations to show relationships

How was My Overall Return?

Returns based on Money Weighted Return (XIRR)

A money-weighted return is a fancy way of saying the return that accounts for when I put money in and when I took it out.

Imagine I put S$1,000 into a stock that doubles in the first year, then top up another S$10,000 just before it falls 20%. A plain average of the two years makes it look fine. But most of my money was invested for the bad year. The money-weighted return captures that, because it gives more weight to the years when more of my money was at work.

Here are my annualised money-weighted returns, valuing the shares I still held at the time.

I left out 2015, because eight months is too short to mean anything. Today’s figure barely changes from 2018 because I only hold 50 shares. My real result was locked in when I sold.

Actually included a period where I’ve already sold off most of it, left only about 50 iShares.

So you’ll be able to see that if it’s kind of based on the last sale, the XIR is about 10.6%, which is meaning that majority of the holdings deciding actually earning 10.6%, and whatever that I am I have today, they account for significantly less than in terms of the returns

If I invested in an MSCI World Index instead

I compared my numbers with the iShares Core MSCI World ETF, which tracks developed-market shares around the world and reinvests dividends. I converted it into Singapore dollars.

• The index itself: from May 2015 to today, MSCI World returned about 10.7% p.a. in SGD.
• Same money, same dates: if I had put the same S$27,094 into MSCI World on my buy dates and left it, I’d have about S$83,500 today. That’s 3.1 times my money, or 11.3% a year.
• My actual result, reinvested: I received S$30,303 back in sales and payouts. If each dollar had gone into MSCI World the day it arrived, I’d have about S$84,650 today.

So I roughly matched the index. I did it with one small, illiquid stock instead of the whole world.

Here’s a simpler way to compare, with no timing of my own in it. Imagine S$100 put into each on 22 May 2015, with every payout reinvested.

For a single stock is a 10% return on a single stock is impressive.

This chart will let you see that actually there’s maybe an easier way of doing it. Of course, if I use the STI ETF instead of the MSCI World instead, perhaps it will be quite different.

What if I had never sold the bulk of my Global Testing shares?

At the peak I held 18,950 shares. I put in S$27,094 in total. If I had never sold, and kept every payout in cash:

Annualised returns can be pretty misleading if you don’t look at the actual dollar amount too. If I had kept that money compounding for another ten years, or started with a much bigger sum, the result would look very different.

In this case, you can see the profit was about $39,000. That’s probably small compared to my current portfolio, but I’d like to believe it was still a significant amount of money back then, maybe in 2015 or 2011.

The Returns if Someone buys after I sold off and held till today

I also looked at what a person would have made buying one day after my last sale (11 April 2018, at S$1.27) and holding to today.

• At S$2.16: 12.9% a year, 2.26 times the money.
• At the S$2.30 offer: 13.6% a year, 2.37 times the money.
• Payouts did a lot of the work. Price alone gives only 6.4% to 7.2% a year. The payouts added S$0.715 a share, which is 56% of that S$1.27 starting price.

Net Cash Was 93% of Market Cap Before the Offer

I think I should pay more attention to how much cash can build up on the balance sheet. In the global testing case, you can see how the cash builds up in the table below.

Just before the offer, the market cap was S$34 million and net cash was S$32 million. The business itself was priced at about S$2 million.

We can see why they would pay more than a 100% premium from the last traded price. Looking back, if I see a company trading at about 90% of its market cap in cash, I think I would have parked some money in it. I know it also pays a dividend, and that’s a decent return for waiting while the cash is sitting there.

The downside is that the company could turn out to be a fraud. If that happens, even after nine or ten years of holding it, I’d have nothing to say for myself.

I guess this is also a byproduct of what SGX is trying to do, which is to push most of our companies to increase shareholder value so that the exchange becomes more vibrant.

In this case, I’m not sure they actually want to enhance shareholder value. But if a company exits at a price 100% higher than the last traded price, that does enhance shareholder value in a way.

The Dividends and Capital Reductions

A capital reduction is the company handing back part of the money shareholders originally put in. A dividend is paid out of profit.

Here’s one way to picture it. A dividend is a restaurant sharing out this year’s takings. A capital reduction is the owner giving back part of the money the partners put in to open it. Either way, you get cash. The difference is where it came from.

Here’s the Capital reduction and dividends paid per share, set against the 1 June share price.

In five of the seven years to 2022, shareholders were paid about 15% to 21% of the share price in cash. That’s the kind of yield that makes a stock look very juicy.

Added up since 2016, that’s S$0.625 per share from capital reductions and S$0.49 from dividends, so S$1.115 per share. And that’s on top of the 2015 payout of about S$1.50 per share.

Why the payouts shrank

Between 2016 and 2018, the company paid out more than it earned in free cash flow. Payouts were about S$7.1 million a year, against free cash flow of S$4.6 to 5.4 million. That’s why cash fell from US$9.3 million to US$5.0 million over those years.

After 2022, the payouts got much smaller (S$0.025 to S$0.10), while the cash pile kept growing. Part of the reason net cash reached S$32 million is that less was paid out.

Global Testing’s Poor Liquidity

Global Testing is very illiquid. That means few shares trade, so it can be hard to buy or sell a large amount without moving the price.

• Typical day: the median day had only about 5,000 to 14,000 shares traded, roughly S$5,000 to 18,000.
• Turnover: in a normal year only 3% to 8% of the whole company changed hands. The exceptions were 2017 (about 20%) and 2022 (about 29%).
• Quiet days: in 2020, only 84 days had a recorded trade.

In 2017 and early 2018 it was still possible to sell in size. By 2019 and 2020, turnover had dropped to 3.5% and 2.5%, and the price was S$0.26 by March 2020. Selling then would have been hard, and at a terrible price.

Tong Hsing’s S$2.30 Is a 129% Premium, or About S$45 Million for the Business

On 11 November 2026, Tong Hsing offered S$2.30 a share in cash, valuing the company at about S$76.7 million.

• The premium: about 129% above the one-month volume-weighted average price before trading was suspended. That’s the average price over the month, with more weight on days when more shares traded.
• How it works: a scheme of arrangement, a way for the buyer to take over all the shares. It would lead to delisting from the Singapore Exchange.
• What still has to happen: approval from shareholders, the Taiwanese investment authorities, the Singapore Exchange and the Singapore courts.
• The reaction: the shares rose about 110% to S$2.16 from S$1.03 on the day. They still trade below S$2.30 because the deal isn’t unconditional yet.
• The conflict: Mr Chen sits on both sides. Reports say he and the other interested directors stayed out of the board discussion and vote, and that independent advisers gave opinions on the price. I haven’t read that opinion myself.

Take the S$76.7 million offer and subtract about S$32 million of net cash. That’s roughly S$45 million for the business itself, against about S$2 million a day earlier. That’s still cheap against about S$15.6 million of EBITDA, but nowhere near the price the market was charging.

I bought my Tokio Marine Asia LifeSecure policy in November 2006, when I was 26. Around my birthday this year, the last premium went out. After 20 years and 240 monthly payments, the policy is now fully paid up.

I thought this was a good time to stop and look back honestly at it. There’s good reason for that.

For years, I’ve carried a quiet feeling that I made a mistake with this policy. I wanted to know if that feeling was real, or if I was just being hard on my younger self.

The best way to clarify if I made a mistake is to take a look at the numbers. Which is not new to me. I know how to read a policy projection. It just never felt like the best use of my evenings. There was always something more urgent than sitting with a spreadsheet to check a policy that I had to pay even if I made a mistake.

With some help from AI, what used to take an evening now takes a much shorter period.

This article is my notes to myself.

We got a bit to get into and if you are unsure what we will get into, here is the breakdown:

• I got my money back, and not a dollar more. After 20 years, my surrender value is almost exactly what I paid in.
• The costs come first and the growth comes later. The policy looks terrible in the early years, and only starts to earn a reasonable return once the costs are behind it.
• A fixed $50,000 doesn’t survive inflation. By 75, it buys about a third of what it bought in 2006, and only about a quarter if inflation runs at 3%.
• Investing the difference would have done better, but only if I stayed invested. A conservative portfolio roughly matched the policy. Anything with more equities beat it, with some frightening falls along the way.
• It was a mistake, but a small one. And it’s one that I don’t think I could have seen at 26.

If you own a policy like this, or you’re thinking about buying one, I hope my 20 years save you some of yours.

I Knew Term Was Cheaper, but I Wanted Cover That Would Last

I want to be clear about one thing first. I wasn’t that uneducated about whole life policies back then.

I understood that term insurance is the better deal compared to whole life. It’s much cheaper, and you can invest the difference yourself. Term is much more optimised if you wish to cover adequately was less prevalent advice but those who read personal finance books would be able to pick those up.

So why did I buy a limited whole life policy anyway?

I was thinking about securing a potential pot of money for medical crisis the whole of my life, not just the next few years.

Term insurance has an end date. Most of the critical illness plans I could buy stop at 65 or 70. If you want to keep them going past the age of 65-70, the premiums climb steeply, and you end up paying a large chunk of the payout yourself through the premiums.

But 65 and beyond is exactly when serious illness becomes more likely.

The textbook answer is that by then:

1. You should have done your financial planning properly.
2. You should have saved enough so that if you fall seriously ill, whatever your hospital insurance doesn’t cover, you can pay from your own savings.

That’s true, and it’s what I’m trying to do today.

Not everyone feels comfortable paying a big medical bill out of their own pocket, especially after they’ve stopped working.

I didn’t want my own savings to be the only plan if there is a way to optimise this.

I didn’t buy this policy for the return. I wasn’t chasing a high number.

What I wanted was a sum of money that would still be there, and still be worth something, when I needed it. If I fell seriously ill at 75 or 85, I wanted to be able to count on about $50,000. I hoped the policy’s value would grow enough to keep up with the years.

I chose to pay for 20 years because I would be young and working. I’d get the paying done early, and then the policy would just sit there for me.

I think don’t ask me why I think that $50,000 is enough for medical crisis without covering income replacement after financial independence.

At 26, I just felt that $50,000 feels like a lot of money.

Some people will say that isn’t much for a serious illness, and they’d be right. Having $200,000 of cover would be wonderful.

But I’ve seen up close what it’s like. Both my parents went through cancer. What I learned is this. Having $50,000 versus having nothing makes a real difference. It’s one less thing to worry about at the very worst time.

That was the thinking. $50,000 of cover, paid over 20 years, meant to be there for me when I’m old.

Limited Whole Life Means Paying for 20 Years, Then Being Covered for Life

The policy is Asia LifeSecure, from Asia Life, which is now part of Tokio Marine. It’s what’s called a limited whole life policy.

The name sounds complicated, but the idea is simple. You pay premiums for a limited time, which was 20 years for me. After that, you stop paying, you still own the policy, and it covers you for the rest of your life.

It also builds up a cash value. This is money inside the policy that belongs to you. You can take it out by cashing in the policy (the “surrender value”), or borrow against it.

How the cover changes over time

In the early years, almost everything the policy would pay out is pure insurance. Think of it like a term plan with a little bonus added on top.

Later on, that changes. The bonuses keep adding up and the cash value keeps growing. By age 65, more than half of what the policy would pay on death is my own accumulated money, not the insurer’s.

Projected figures from the my own 2006 policy benefit illustration.

We can see it in my own policy that as we get closer to 65 then 85 years old, much of the death benefit will come from the cash value accumulated over time.

After 20 Years of Premiums, the Surrender Value Is Almost Exactly What I Paid In

Here is what I paid:

• The basic plan: $78.85 a month.
• The Dread Disease Accelerator Rider: $16.90 a month. This pays $50,000 if I’m diagnosed with one of the covered serious illnesses.
• Total: $95.75 a month, or about $1,149 a year.
• Over 20 years: $22,980.

And here is what I would get if I cashed it in today.

The surrender value is $23,005.41. Of that, $16,610.69 is guaranteed, and $6,394.72 comes from bonuses.

That’s $25 more than I paid in. Over 20 years.

I sat with that number for a while.

Twenty years of faithful monthly payments, through job changes, through 2008, through Covid. And if I walked away today, I’d get my money back and not a dollar more.

To be fair to the policy, I wasn’t paying only for savings. Part of my premium bought 20 years of protection. If I had died, my family would have received $50,000 plus bonuses. If I had been diagnosed with a serious illness, the rider would have paid $50,000. I’m grateful I never needed either.

But that’s also the point.

The protection came at a cost, and now I can see exactly how much it cost.

One more thing surprised me. The 2006 illustration projected a cash value of $22,990 at year 20, and bonuses of $19,266 added to the policy. My actual figures today are $23,005 and $19,266.61. After 20 years, the projection was off by about $15.

If I gotten a Term Plan instead, I would have gotten the Same Cover but would have Cost Almost 4 Times Less

Around the same time when I bought the LifeSecure, I also asked for a quotation for a term plan. I wanted the same $50,000 if I died, and the same $50,000 if I was diagnosed with a serious illness.

The term plan would have cost $24.95 a month. My limited whole life policy cost $95.75 a month.

That’s almost 4 times as much.

Had I opted for the term plan, I will have to pay till age 70, which means I will still be paying another 24 years after today.

Today, my limited whole life policy would pay $69,267 if I died. The term plan would pay $50,000. That’s about $19,000 less.

I asked myself honestly whether my family would have felt that gap. I don’t think they would have noticed much given the magnitude.

Meanwhile, the money I would have saved, about $17,000 over 20 years, would have been mine to invest.

In a Limited Whole Life Policy Projection, the Costs Come First and the Growth Comes Later

Let me start with the number most people want: what return did I actually get?

If I take every premium I paid, month by month, and compare it with what I’d get if I cashed in today, the return is 0.01% a year.

I also calculated that if I leave out the rider, which is pure protection with no cash value, and count only the basic plan, it’s 1.9% a year.

We’re not supposed to judge insurance by its return. It’s protection, not an investment. But I don’t think we can run away from the number either. If you want enough cover for later in life, you need the money you put in to work as hard as it can. The return matters.

We calculate returns with money-weighted returns with the net cash flows using XIRR.

XIRR finds the one steady yearly rate that would turn all your payments, made at different times, into the amount you have today.

Imagine you put $95.75 into a savings account on the first of every month for 20 years. XIRR asks: what interest rate would that account need to pay so that today, it holds $23,005?

For me, the answer to that interest rate is almost zero.

You can do this yourself in Excel or Google Sheets. List each payment as a negative number with its date, put the surrender value as a positive number at today’s date, and use the XIRR function.

One thing we need to take note is that the return doesn’t move in a straight line

The return on a policy like this starts deeply negative, and then climbs very slowly.

Here’s what the projection table shows for the basic plan:

• Year 3: about -25% a year. Most of what I had paid went to costs.
• Year 10: about -1.6% a year.
• Around year 13: the cash value finally catches up with the premiums paid.
• Year 20: about +2.2% a year.
• Age 65: about +3.6% a year. The illustration itself says 3.58%, so the maths checks out.

You can take a look at how the return transitions as I age by reading the last column in the illustration table below:

So the policy looks terrible early, and better the longer you hold it. If you surrender in the early years, you lock in the worst of it. That’s something I wish more people knew before they buy, and before they cancel.

How to read the columns

Most people never read the projection table that comes with their policy. It’s a dense page of small numbers that nobody explains. Here’s what the main columns mean:

• Guaranteed. The insurer must pay this, no matter what happens to markets.
• Non-guaranteed. This comes from bonuses. The insurer projects it, but doesn’t promise it.
• Total cash value. The two added together. This is what you’d get if you cashed in.
• Death benefit. The sum assured plus the bonuses added so far.

I want folks to take note that the table above was a benefits illustration that was generated in 2006. Those bonus may not be what I eventually earn. It is actually a projection.

By year 20, the bonuses added to my policy came to $69,266. If I cashed in, I’d only get $22,990 of it. That’s about a third.

The bonus shows up in full when I die, but only partly when I cash in. By year 40, I’d get about 59% of it. The longer you wait, the more of the bonus becomes yours.

There are some value spikes at years 5, 10 and 15

If you look at how much bonus is added each year, there’s a pattern. For the first few years, it grows steadily by about $505, $510, $515. Then in year 5 it jumps to $1,158. In year 6 it falls back to $657, but now from a higher base.

The same thing happens at year 10, with a jump of $1,991, and at year 15, with $2,874. After that, it’s smooth.

This looks like a bonus step up. Some insurers add a larger bonus on certain policy anniversaries, or move your policy to a higher bonus rate once it has been in force long enough. Either way, you see a jump on that anniversary and a higher base afterwards.

It’s easy to miss unless you line the numbers up year by year.

I was curious what does this Limited Whole Life policy earn once I stop paying.

Once the premiums stop, the calculation is simpler because no money goes in, so I can use a plain growth rate between two values.

From year 20 to year 40, the cash value is projected to grow from $22,990 to $54,447.

If we calculate based on annualised return, that’s about 4.4% a year.

So I have two numbers for the same policy:

• 0.01% a year on everything I paid over 20 years, including the expensive early years and the cost of the rider.
• About 4.4% a year from here on, now that the costs are mostly behind me.

Both are true.

1. The first tells me what the whole 20 years earned.
2. The second tells me what waiting longer earns me. That growth slows gradually, from about 4.9% a year just after the premiums stop to about 3.9% a year by age 65.

Which brings up a question I’d never thought to ask before.

What did a 5.25% Assumed Return Mean in 2006?

The illustration assumes the insurer’s fund earns 5.25% a year, before costs. It states clearly that this isn’t guaranteed. After all the costs, what I see coming through to my policy is closer to 4% a year in the later years.

So I got curious. What was the starting point back then?

My policy is a participating policy. The insurer pools the premiums from many policyholders into one fund and invests it. When the fund does well, policyholders share in the profits through bonuses.

I don’t know exactly what Asia Life’s fund holds. My understanding is that funds like this lean heavily on fixed income, maybe 70% or so, partly because the insurer has to back the guarantees it makes.

If that’s roughly right, then what bonds were paying in 2006 matters a lot.

In 2006, Singapore government bonds yielded about 3% across almost every maturity, from 3 months to 15 years. The 10-year bond averaged 3.05%, and the 15-year 3.16%.

So 5.25% was about two percentage points above what government bonds paid. That extra return had to come from somewhere, such as shares, property or corporate bonds.

If 70% of the fund earned around 3% to 4%, the other 30% would need to earn roughly 8% to 10% a year to get the whole fund to 5.25%.

That’s not impossible. It is also likely that the assumed fixed income return is smoothed over the previous few years and so it will be higher (if you look at where the rates are at in 2006 and how it looks like a few years before)

I’m not drawing a conclusion here. I just found it useful to see where the starting line was.

An Opportunity Cost Analysis: A Conservative Portfolio Mix Would Have Roughly Matched the Policy. More Equities in Portfolio Would Have Beaten the Limited Whole Life Policy

Now back to the thought I asked you to hold.

What if I had bought the cheaper term plan in 2006 and invested the difference myself?

We got to decide what kind of portfolios I would have invested in. I would assume its a mixture of different allocation of fixed income and equities.

To keep it fair, I used simple model portfolios over exactly the same 20 years, November 2006 to October 2026:

• Equities: the MSCI All Country World IMI Index. This covers developed and emerging markets, in Singapore dollars, with dividends reinvested after withholding tax.
• Fixed income: the Bloomberg Global Aggregate Bond Index, hedged to Singapore dollars.
• We do 6 asset allocation mixes: from 100% fixed income to 100% equities, in steps of 20%.
• Costs: I took 0.5% a year off every portfolio, for fund fees and other costs.
• Rebalancing: back to the target mix every month.

This is a reasonable stand-in for a participating fund, with one difference. A participating fund likely holds more Singapore fixed income, and I don’t know how much of its equities are in emerging markets. So treat this as a comparison, not a copy.

How much money will go into this investment portfolio?

The policy cost me $95.75 a month. In the comparison, the same $95.75 goes out every month. $24.95 buys the term plan, and the remaining $70.80 is invested.

At age 46, when the policy is fully paid, I stop adding money to the portfolio too. From then on, each portfolio grows on its own at the same yearly return it earned over the past 20 years.

Table 1: the simple picture

First, let’s look at the portfolios as if they only need to grow the money I put in.

Each cell shows the value and the XIRR on all $95.75 a month paid. The policy’s value at 75 is my estimate, assuming its growth continues at about 4% a year after age 66.

In short:

• The policy: $23,005 at 46, $52,411 at 65, about $77,500 at 75.
• 60/40 mix: $27,688 at 46, $58,278 at 65, $86,221 at 75.
• 100% equities: $46,835 at 46, $152,098 at 65, $282,724 at 75.

If we invest in a portfolio with higher fixed income, the DIY portfolio value would not match the policy value based on projection. The portfolio mix that equals the policy is a 60% fixed income 40% equity.

That is pretty similar to the potential allocation mix of the participating funds that usually drive the returns of these cash value insurance policies.

But at 20 years mark, every mix from 80/20 upwards is already level with or ahead of the policy.

By 65 and 75, the gap in the value accumulated by DIY widens.

But in a way, I left something out.

After 20 years, my limited whole life policy won’t need me to pay more premiums. The term plan doesn’t stop. It keeps charging $24.95 a month until I’m 70, another 24 years.

That money has to come from somewhere.

To make the comparison fair, I take the $24.95 out of the investment pot every month until age 70.

Table 2: the fair comparison

Same as Table 1, but $24.95 a month is taken out of each portfolio from age 46 to 70.

If we factor in the premiums that I have to continue to pay the picture shifts:

What I take from it

• A conservative 60/40 allocation mix lands almost exactly where the policy does. 2.7% against 2.8% at 65. It’s a tie, give or take.
• The safest mixes lose. A portfolio that’s all fixed income earns close to nothing once the term premiums are paid.
• Mixes with 40% or more in equities win, and win clearly. At 75, a 40/60 portfolio is worth about $114,000 against the policy’s $77,500.

Here’s something that surprised me. Once the costs are behind it, the policy grows at about 4% a year. That’s almost exactly what the 60/40 portfolio earned over the past 20 years. In the later years, the policy behaves a lot like a conservative balanced portfolio.

Its problem is the first 20 years.

So mathematically, investing the difference wins.

But I’ve learned to be suspicious of a result that looks this clean.

Investing the Difference Only Works If You Stay Invested Through the Bad Years

Some of you might say I picked a bad starting point. Had I use a better starting point then the returns of the DIY portfolio would not have been so poor.

Well, I cannot choose the time period I lived through (and neither could you).

I just happen to lived through a negative sequence (living through the 2008 crisis) follow by a positive sequence of return.

I somehow feel that with my mindset back then, even if a 100% equity allocation DIY mix do better, I might not have earned it had we review it today.

I would not have the right temperament to keep dollar cost average that $70.80.

A drawdown is how far your portfolio falls from its last high point.

Imagine you have been investing faithfully for two years, and your portfolio has grown to $3,000. Then the crisis hits. Within months, it’s worth $1,500. You have done nothing wrong. You kept paying in every month. But half of it is gone, and you have no idea when, or if, it will come back.

That’s what the 100% equity portfolio went through.

At the bottom in February 2009, the all equity investor had 59 cents for every dollar they had put in. It took until May 2014, almost seven years from the previous high, to fully recover.

And the “safe” choice wasn’t safe from pain either. The all fixed income portfolio peaked in December 2020, fell 14% by late 2022, and still hasn’t got back to that high today.

What if I had started in 2011, after the crash?

Starting in 2011 skips the crash entirely. The equity heavy portfolios look much better. 100% equities would have earned 9.8% a year instead of 6.9%. But the all fixed income portfolio actually looks worse, 2.2% a year instead of 2.6%.

I would go back to an important point.

You touch your heart and ask yourself if it is easier to execute DIY over 2008 to 2013 without missing a single dollar cost average tranche or to just commit to a limited whole life insurance.

My limited whole policy’s cash value never showed me a crash. The bonuses inside a participating fund are smoothed over the years, so I never saw a 50% fall, and I never had a reason to panic.

I couldn’t see the drawdown, so I couldn’t run from it.

Many young people will look at the maths and think, “I’ll just put everything in equities and do much better.” That’s true on paper. But the return you actually get depends on what you do in the worst months, not the best ones.

A Fixed $50,000 Loses More Than Half Its Purchasing Power by Age 75

Now let me come back to the reason I bought this policy in the first place.

I wanted about $50,000 for a serious illness when I’m old, like 70, 75 or 85. For critical illness, the part of this policy that matters is the Dread Disease Accelerator Rider. It pays at least $50,000.

I was assuming it pays only that $50,000, and not the bonuses on top. If there is more, than it is a bonus.

The problem is inflation. I wonder if that $50,000 can retain its purchasing power.

I am hoping the returns over 40-50 years can do the heavy lifting to ensure that $50,000 at 26 retains its value at 86 years old.

Inflation means things cost more every year, so the same dollar buys less. If prices rise 2% a year, something that costs $100 today costs about $102 next year. It sounds tiny. Over 49 years, from 2006 to the year I turn 75, it adds up to something huge.

By 75, my $50,000 buys about 38% of what it bought in 2006. If inflation averages 3%, it’s about 23%.

To have the same buying power as $50,000 in 2006:

1. I’d need about $132,000 at 75 if inflation averages 2%,
2. and about $213,000 if it averages 3%.
3. And medical costs have often risen faster than general prices, so this may be understating the problem.

Can the policy, if it pays out the cash values on top of the $50,000 benefit meet these amount?

Unfortunately, they fall short if we assume that the rate of return is similar to the rate of growth after fully paying up the limited whole life (4% p.a.)

It will reach a nominal value of $105,000 instead of $132,000 or $213,000.

But doesn’t the policy beat inflation in the later years?

It does. Growing at about 4% a year beats 2% or 3% inflation. The problem is the first 20 years, when my money barely grew at all, and the rider, which doesn’t grow at all.

So my original idea was a bit naive. Not because whole life is evil, but because I sized it for the wrong number. I bought what $50,000 meant at 26, not what it would need to be at 75.

If I really need a coverage of $50,000 in 2006, I could have solved the problem by buying a significantly larger limited whole life.

By my calculation, I would need roughly 2.6 to 4.3 times the cover. That’s about $130,000 to $210,000 of rider cover, instead of $50,000.

That’s a sobering number.

And that will mean more out of pocket costs.

It Was a Small Mistake, and One That Is Hard to See at 26

When I owned this policy, I often felt I had made the wrong call. I hadn’t run the numbers. I just sensed that a fixed sum, inside a fund that is mostly fixed income, couldn’t hold its real value for 50 years. Fixed income gives you certainty. It doesn’t protect you well from inflation.

Now the numbers agree. I’ll chalk this one down as a mistake.

I take solace it does not hurt that much thankfully.

I got my money back in dollars. In buying power, I got back about 80% of what I put in. If I had invested the difference sensibly, I would be maybe $5,000 to $10,000 better off today. That’s a lower return, not a disaster.

In another universe where I didn’t buy this, I’d probably have bought $100,000 of term cover instead. At 65, that term plan would pay $100,000, more than this policy’s $91,485 death benefit, for about half the monthly premium. If something had happened to me in those years, my family would have been better off.

What surprised me was not the numbers. It was the feeling.

The amount was never large. It went out of my pay every month, and most months I didn’t notice it. But every time I thought about it, a quiet voice said, “You’re still paying for that decision.”

That’s the strange thing about a long commitment. A bad purchase is over once you’ve bought it. A policy you regret keeps showing up, month after month, year after year.

I think this is why so many people stay stuck with insurance products they regret. The amount isn’t big enough to make you act, and it’s too steady to forget. If you’re in that position, I understand how heavy that small, quiet feeling can get.

I’m curious how I’ll feel now that I no longer have to pay. I suspect it will feel lighter than I expect.

What I would do differently

Looking at it objectively, if the goal is to have money for a medical crisis after you’re financially independent, I think you’ll do better building your own medical sinking fund over time than relying on a limited whole life policy. If you know how to invest, and you can stay invested, you can do better than an insurance company’s fund.

That’s my approach today. My term policies cover critical illness until 65 or 70. After that, I’ll lean on a medical sinking fund I’m building for exactly this. I wrote about how I plan it [here: link to the medical sinking fund article]. That’s a topic for another day.

Insurance still has a job. A sinking fund can cover a $50,000 bill at 75, but not at 35, when it’s still small. That’s what insurance is for.

But I was 26.

I could only take this policy apart at 46. At 26, I don’t think I would have seen all of this, and I’m not sure I’d have listened if someone had told me.

A medical sinking fund sounds sensible. But in your twenties, there are always a hundred other priorities: a home, a wedding, a career, maybe children. The money for a fund that you won’t touch for 40 years quietly gets spent on something else. Nobody notices until it matters.

A policy at least forces the habit. I paid every month for 20 years without thinking. That counts for something, even if the return was poor.

So my advice isn’t “never buy limited whole life”. It’s this:

• Know what you are buying. Some go through the whole process, commit a sum that hurts them enough, into something they later regretted. Try not to be like that.
• The Math tells us that if you felt that you can hold an allocation mix more than 60% fixed income 40% equity, the cash value you accumulated would likely be more worth it if you don’t commit to a limited whole life. Provided you can really hold it over long periods (I highly doubt for many of you given how trigger happy you are to change your course of action over the years)
• I doubt a whole life plan’s investment return can help retain real purchasing power of the sum assured over time. You may have to buffer that by covering more.
• Start your own medical sinking fund early, even if it’s small. Time is the one thing you can’t buy back later.

The last lesson is… be forgiving to your younger self. As a younger person you should try your very best to understand these stuff but life will teach you that there are lessons that you can only comprehend as you see life unfolds.

You can check out my other personal notes in Managing Kyith.

1. The stock started 2025 at S$4.66. By the end of 2025, it closed at S$8.42. That’s an 81% gain in a single year, from a company most of us thought of as a steady dividend stock.
2. Then 2026 started even better. By 8 April, the stock closed at S$11.47, and the next day it touched an all-time high of S$11.63.

After that, things went quiet.

From March to October, the stock mostly moved sideways between S$10 and S$11.50. Seven months of going nowhere. Then in early November, it dropped hard, closing at S$9.26 on 6 November.

From the end of 2024 to the peak, the stock went up about 150%. Even after the recent fall, it’s still up about 99%.

That got me curious.

1. How much can a business really change in under two years for the market to reprice ST Engineering so significantly?
2. And is a gain like this backed by the fundamentals, or is it mostly the market getting excited?

I went through 16 years of numbers to find out.

A dividend stock that the market suddenly sees differently

Most of us know ST Engineering as a dividend stock.

For a long time, it was the kind of company you bought, tucked away, and almost forgot about. You collected your 15 to 16 cents a year, and the share price stayed between S$3 and S$4.50 for the better part of 15 years.

I should know that since I been collecting the dividends from 2005 to 2018 during my time working there.

What I find most telling is how steady its valuation was.

The price-to-earnings ratio (P/E) is how many dollars you pay for every dollar of yearly profit the company earns. Here’s ST Engineering’s P/E for every year since 2010:

I can also tally the data in charts:

For 15 years, the P/E stayed in a narrow band of about 17 to 22 times. The average PE was 20.5 times.

It didn’t matter much what profits did. In some years they fell, in others they grew a little.

The P/E barely moved.

I think that’s because people valued it for the dividend, not for growth.

At around S$3.50 to S$4, a 15 to 16 cent dividend gave you roughly 4%. Whenever the price fell, the dividend yield looked attractive and buyers came in. Whenever it rose too much, the yield looked thin and they stopped buying. The dividend kept the price in its lane.

Then in 2025, it broke out of that range.

The P/E went to 27.6 times in 2025 and about 31 times on average in 2026. Today, at S$9.26, it’s around 27.5 times my estimate of 2026 earnings.

That’s well above anything we saw in the previous 15 years.

What a higher ST Engineering P/E is really telling you

A higher P/E is not automatically a bad thing. Some people feel cautious about a company trading at high P/E and would salivate a company when the P/E fell to lower than normal.

When the market pays a higher P/E, it’s usually saying it believes the business has become better. Maybe it will grow faster, or its profits are more reliable, or it will keep growing for longer.

Imagine two companies, each earning S$1 a share this year.

The first one earns about S$1 every year. Some years a bit more, some years a bit less. Ten years from now, it’s still earning around S$1.

The second one is growing. Next year it earns S$1.12, the year after S$1.25, and ten years from now, close to S$3.

Would you pay the same price for both? Probably not. You’d happily pay more for the second one, because you’re buying a bigger stream of future profits. That’s what a higher P/E is. You pay the same price for today’s earnings, but you’re also paying for tomorrow’s.

So when ST Engineering’s P/E moved from 20 times to the high 20s, the market was basically saying:

This is no longer the steady, slow company we knew. It’s becoming a growth company.

The question is whether that’s true, and for how long.

If the growth lasts a year or two and then fades back to the old pace, today’s P/E is too high and the price will likely come down. If ST Engineering can keep growing faster for many years, today’s P/E could turn out to be reasonable, or even cheap.

The P/E alone can’t tell us which one it is. It only tells us what the market believes.

To find out whether that belief holds up, we have to look under the hood at the business itself.

ST Engineering’s revenue grew for years. It’s profit didn’t follow until recently.

If the business really got better, the first thing I’d expect to see is faster growth.

Revenue is the total money customers pay the company. Profit is what’s left for shareholders after all the costs, interest and taxes. You want both to grow, but it’s profit that ends up paying your dividends and driving the share price over time.

When I lined up the two over 16 years, I saw three very different periods:

2010 to 2018: Revenue and profits were just poor. Revenue went from S$5.98 billion to S$6.70 billion. Profit went from S$491 million to S$494 million. In eight years, profit grew by S$3 million. No wonder the market treated it like a bond.

2018 to 2023: Revenue picked up, but profit didn’t. This is the part I think a lot of investors would have missed.

Revenue grew 8.6% a year, helped by organic growth plus buying MRAS in 2019 and TransCore in 2022. By 2023, the company was selling 50% more than in 2018.

Yet profit only grew 3.5% a year.

The company was getting bigger, but it wasn’t keeping more of what it earned. All that extra revenue got eaten up somewhere along the way. I’ll explain where in the next section.

2023 to 2026: Profit finally stepped up. Profit grew about 21.5% a year, twice as fast as revenue. This is the growth the market got excited about, and it’s when the P/E broke out of its 15-year range.

From 2010 to 2023, revenue grew 69% while profit grew only 19%. You can observe the two lines drifted apart for years. Then from 2023, profit shot up and closed most of the gap.

Here’s what made me stop and think.

Over the full 16 years, revenue grew about 2.3 times and profit grew about 2.1 times. After all that excitement, profit has only just caught up with revenue.

So the big jump in profit over the past three years might not mean ST Engineering has become a fundamentally more profitable business.

A good part of it could be profit making up for lost time.

That matters a lot for the P/E.

If you look only at the last three years, you see a company growing profits at 20% a year, and a P/E of 27 or 30 times looks reasonable. If you believe part of that 20% was a one-time catch-up, then the real growth rate is closer to the 10% pace of revenue, and the same P/E looks a lot more expensive.

To find out which one is right, we need to understand why profit lagged for so long, and whether the things holding it back are truly gone.

Is ST Engineering, in its current form, a more profitable business than before?

If ST Engineering really has become a better business, I will expect it to keep more of every dollar it earns than it used to. That’s what profit margins tell us.

The easiest way to think about margins is to follow S$100 of revenue down to shareholders. Here’s what happened to every S$100 ST Engineering earned in the first half of 2026:

• S$19.50 left after the direct cost of doing the work, like materials, parts and labour on the job. This is the gross margin.
• S$15.00 left after running the company, like office staff, sales and R&D. This is the EBITDA margin. It’s the best measure of how profitable the day-to-day operations are.
• S$7.80 left for shareholders, after depreciation of equipment, amortisation of acquired businesses, interest on debt, tax and the share belonging to minority partners. This is the net margin.

Now compare that to the past:

When I first looked at this, I had to read it twice. The three levels tell three different stories.

The gross margin is lower than before. ST Engineering keeps about 19.5 cents per dollar after direct costs, versus about 21 to 22 cents in the early 2010s. I think this is mainly the mix of work. Engine repair passes a lot of expensive parts through to customers at low markups, and big rail and tolling projects carry thinner margins.

The EBITDA margin is better than before. Operations now keep about 15 cents per dollar, versus about 13 cents in the 2010s. The company runs leaner. Overheads fell from 12.1% of revenue in 2022 to 10.2% in 2025. This is a real improvement and management deserves credit for it.

The net margin is only back to where it was. At 7.8%, shareholders keep about the same as in the late 2010s, and still less than the 8.7% of the early 2010s.

So how can the operations be more profitable while shareholders end up with the same, or less?

The answer is in the middle. Look at how much gets lost between EBITDA and net profit:

In 2010 to 2013, about 4 cents of every dollar disappeared between EBITDA and net profit. By 2022 and 2023, after buying TransCore, it was 8 to 9.5 cents.

That shaded area grew because of the acquisitions. When ST Engineering bought TransCore for US$2.68 billion with borrowed money, two things came with it:

• Interest. Total borrowings jumped from S$2.1 billion to S$6.5 billion in 2022. Every year, the interest on that debt comes out before shareholders get anything.
• Amortisation. When you buy a company for more than its assets are worth on paper, part of the price gets written off slowly over many years. It’s a non-cash cost, but it still reduces reported profit.

On top of that, the satellite communications business, bought over the years for about S$0.6 billion, kept losing money as players like Starlink changed the industry. In November 2025, ST Engineering wrote down S$667 million of its value and is now looking to sell it.

So the operations got better, but the cost of the acquisitions ate up all of that improvement and more.

Since 2023, that gap has been narrowing. Debt came down from S$6.5 billion to S$4.7 billion. Amortisation fell 22% in the first half of 2026. Satcom losses got smaller. The gap is now about 7 cents.

That’s where most of the recent profit jump came from. Profit is going up faster than revenue because less is being lost in the middle.

Here’s why this matters so much. ST Engineering will do around S$13.6 billion of revenue in 2026. Every 1 cent per dollar of net margin is worth about S$136 million of profit, which is about 13% of the total. A recovery from 5.8% to 7.8% adds roughly S$270 million, and that’s without selling a single extra contract.

So, is the business more profitable than before?

The operations are, a little. The shareholders’ share isn’t.

The bigger question is how much of this recovery is left. I don’t think the gap will go back to 4 cents. TransCore’s debt won’t fully disappear, and the amortisation runs for years. If the gap settles at around 6 to 7 cents, most of the margin catch-up has already happened.

And once the catch-up is done, profit can only grow about as fast as revenue, plus whatever extra efficiency management can find.

Are we paying a fair price for the growth?

What you pay for a stock matters. But so does what you get for it.

A P/E of 27 times can be expensive for one company and cheap for another. It depends on how fast the profits are growing. This is where the PEG ratio helps.

The PEG ratio takes the P/E and divides it by how fast earnings are growing. It asks a simple question: am I paying a fair price for the growth I’m getting?

Imagine two stocks, both with a P/E of 20.

The first one grows its profit 20% a year. Its PEG is 20 ÷ 20 = 1.

The second one grows its profit 5% a year. Its PEG is 20 ÷ 5 = 4.

Same P/E, but with the second one you pay four times as much for each unit of growth. That’s a lot of money for a slow business.

As a rough guide, here’s how I’d read the PEG:

• Below 1.4: cheap for the growth. You’re paying less than the growth rate. These are rare in good businesses.
• 1.4 to 2: fair. A reasonable price for a decent business that’s growing.
• Above 2: expensive for the growth. You’re paying a lot for each unit of growth. Either the market expects growth to speed up, or you’re paying for something other than growth.

This is just a rule of thumb so context is important. High-quality companies with very reliable earnings often trade above 2 for years.

Here’s ST Engineering’s PEG over time. For growth, I used how fast earnings per share grew each year over the previous three years.

If you went only by PEG, ST Engineering looked expensive for most of the last 15 years.

In most years the PEG was well above 2, sometimes 4 or 5. In 2020 it was almost 18. And in six of those years, there was no growth at all. You were paying 20 times earnings for a company whose profit had gone backwards over the previous three years.

So were investors overpaying all that time?

The investors were just weren’t paying for growth.

People bought ST Engineering because it was a steady, government-linked defence and engineering company that paid you about 4% a year, every year. It was closer to owning a bond than a growth stock. Nobody looked at their fixed deposit and complained that it wasn’t growing.

That’s why the P/E stayed at 20 times regardless of whether profit went up or down. The PEG was never the right way to value it.

Now look at 2025 and 2026.

For the first time, the PEG dropped into the fair range: 1.6 in 2025 and 1.4 in 2026. At today’s price of S$9.26, it’s about 1.3.

On this measure, ST Engineering has never looked cheaper. Even after the price nearly doubled, the growth seems to justify it.

But there’s a catch, and it goes back to the margins.

The 21.7% growth in that PEG includes the one-off margin recovery. Profit grew fast partly because the cost of the acquisitions was shrinking. That part can’t repeat.

If I use 11.5% growth, which I think is more sustainable, the PEG today is about 2.4. That puts it back in the expensive range, about where it sat for most of the 2010s.

A PEG is only as good as the growth number you put into it. Use the last three years and the stock looks cheap. Use what’s likely to come next and it looks fully priced.

Doubling of share price in about a year: how much was earned?

Let me be specific about the run.

At the end of 2024, ST Engineering closed at S$4.66. The first real push came on 27 February 2025, when it reported record revenue, profit and order book for 2024. Then European defence stocks surged in early March, and on 18 March management laid out its 2029 targets and a new dividend policy. The stock went from around S$5 to almost S$7 in three weeks.

It never really looked back.

On 14 January 2026, it closed at S$9.39, double the price at the end of 2024. That’s 100% in about a year, from a stock that had gone nowhere for 15 years. By 9 April 2026 it touched S$11.63, up 150% in just over 15 months.

So how much of that was the business, and how much was the market getting excited?

Here’s how I worked it out. You don’t need to be a finance person to follow it.

A share price is basically two things multiplied together:

Price = earnings per share × P/E

So a price can go up for only two reasons. The company earns more, or the market pays more for each dollar of earnings. To separate them, I asked one simple question:

What would the share price be today if the business had grown, but the market had kept paying the same P/E as before?

At the end of 2024:

• Earnings per share was 22.4 cents.
• The price was S$4.66.
• So the P/E was about 20.8 times.

Today:

• Earnings per share is about 33.7 cents (my 2026 estimate).
• The price is S$9.26.
• So the P/E is about 27.5 times.

If the market had kept paying 20.8 times, the price today would be 33.7 cents × 20.8 = S$7.01.

That means:

• From S$4.66 to S$7.01 (+S$2.35) came from earnings growth. This is the part the business earned.
• From S$7.01 to S$9.26 (+S$2.25) came from the higher P/E. This is the part that came from the market’s change in mood.

About half and half.

At the peak of S$11.63, the same maths gives S$2.35 from earnings and S$4.62 from the higher P/E. At the top, about one-third of the gain was the business and two-thirds was the market’s excitement.

(If you want to be more precise, the answer shifts a bit depending on where you start and which year’s earnings you use. Starting from mid-2024, or using a slightly different method, you get anywhere from 40% to 60% from earnings. Whichever way you cut it, roughly half the gain is earnings and half is a higher P/E.)

What does this mean from here?

The P/E going from about 20 times to 27 times is the market saying: “This is a better business now, and it deserves a higher price.” That re-rating has already happened. It’s already in the price you pay today.

From here, I don’t think you can count on the P/E going much higher. If anything, history says it’s more likely to drift down than up. So if you buy today, your return mostly depends on one thing:

Can ST Engineering keep growing its earnings, and for how long?

There are really two ways a company can justify a high P/E:

• The sprint. Earnings grow very fast for a few years, then slow down to a normal pace.
• The marathon. Earnings grow at a decent but not spectacular rate, and keep doing it for a long time.

Both can add up to the same value today. But they’re very different bets.

Three possible futures for ST Engineering

I worked out what ST Engineering would be worth today under three different futures. The valuation will be based on discounted cash flow(DCF).

Scenario 1: The new growth fades, and ST Engineering goes back to its old self

Say the past few years turn out to be a cycle, not a new normal. Growth falls back to the 2.5% a year we saw in the 2010s.

• If it fades soon (one more good year in 2027, then back to 2.5%): about S$6.40 a share, roughly 31% below today.
• If it fades after 2029 (management hits its targets, then back to 2.5%): about S$7.41 a share, roughly 20% below today.

This is the scenario I think investors should take seriously. ST Engineering grew this slowly for almost a decade, so it isn’t a far-fetched outcome. Even if management delivers everything it has promised up to 2029, the stock would be worth about 20% less than today if growth stalls after that.

Scenario 2: The new growth rate lasts a few more years

Say ST Engineering keeps growing profit at 12% a year for five years, to 2031, then settles into normal growth of 3% a year, roughly in line with the economy.

• Value: about S$9.42 a share, roughly the same as today’s price.

So today’s price is essentially paying for about five more years of strong growth.

Scenario 3: The new growth rate lasts 10 years

Say the defence cycle, engine repair demand and smart city projects keep ST Engineering growing at 12% a year all the way to 2036, then it settles at 3%.

• Value: about S$12.95 a share, roughly 40% above today.

This is the bull case. It needs a decade of strong growth from a company that, until recently, had a decade of almost none.

The chart shows all of this in one picture. Each bar is what the stock is worth depending on how many years the 12% growth lasts. Each extra year of strong growth adds about 50 to 75 cents to the value.

At S$9.26, the market is pricing in about 4.7 years of 12% growth. That’s roughly management’s 2029 plan, plus another year or two after it.

Depending on the growth rate and the return you want, it’s somewhere between 4 and 7 years:

At the peak of S$11.63, it was a different story. The market was pricing in about 8 years of 12% growth, roughly to 2034 or 2035. That’s well beyond anything management has guided for. In effect, investors were betting that ST Engineering would ride the whole European defence cycle at full speed.

We can also flip around by looking at if we are less demanding of return (6.5% a year) or more demanding (7.5% a year), what price are we okay to own.

The table seem to point that we are either expecting this profit growth to be longer, or that we are okay with a lower 6.5% per year return by paying the current price.

What about the dividends?

Most people bought ST Engineering for the dividend. So I wanted to see how that dividend has actually grown, and what it has been worth as a yield along the way.

Here’s the dividend for each financial year, compared with the share price on 1 June of that year. I picked 1 June because it’s roughly the middle of the year. The prevailing dividend yield is what you would have earned that year if you bought on that day.

The payout ratio is the share of each year’s profit paid out as dividends. The 2010 to 2014 dividends include special dividends.

Over 15 years, from FY2010 to FY2025, the dividend grew from 14.55 cents to 18 cents. That’s about 1.4% a year.

If you include my estimate for 2026, it’s about 2.2% a year.

That’s slower than inflation for most of the period. I think it’s worth sitting with that for a moment. The stock many of us thought of as a reliable income stock barely raised its dividend for over a decade.

When I look closer, the dividend history falls into three periods:

• 2010 to 2012: rising. The dividend went from 14.55 to 16.8 cents, helped by special dividends on top of the ordinary ones.
• 2013 to 2021: frozen at 15 cents for nine years. In 2013 the company cut its payout ratio from 90% to 80%, and the dividend dropped. After that it stayed at exactly 15 cents every year, through good years and Covid alike.
• 2022 onwards: growing again. 16 cents, then 17, then 18. From 2021 to 2025, the dividend grew about 4.7% a year, and 2026 looks set for a bigger jump.

The prevailing dividend yield tells an even more interesting story.

From 2010 to 2024, the prevailing dividend yield was roughly 4 to 5% almost every year. That’s very steady, and I think that steadiness is exactly what held the share price in its S$3 to S$4.50 range. When the yield went above 5%, buyers came in. When it dropped near 3.5%, they stopped.

Then in 2025, the prevailing dividend yield fell to 2.3%. In 2026, it’s about 1.8%.

The dividend didn’t get cut. It actually grew. The share price just rose much faster than the dividend.

That’s the clearest sign that the people buying ST Engineering today are not dividend investors. Nobody buys a stock at 1.8% for income when a Singapore Savings Bond pays more. [Edit if you want to compare with the current SSB or T-bill rate.]

There’s one more thing in the table worth noticing: the payout ratio.

It used to be 80 to 95%. ST Engineering paid out almost everything it earned, which left little to reinvest and explains why the dividend couldn’t grow. Now it’s down to about 60 to 66%. The company is keeping more profit to fund growth, and the dividend grows more slowly than earnings.

That’s by design. Under the new policy from 2025, ST Engineering pays out about one-third of each year’s increase in net profit as extra dividends.

Here’s what that means in practice. Say profit grows 13.5% next year, in line with management’s 2029 target. Earnings per share would go up by about 3.7 cents. One-third of that is about 1.2 cents, so the dividend would rise from about 20.5 cents to about 21.7 cents. That’s a dividend growth of about 6%, roughly half the speed of profit growth.

So the dividend will likely keep growing, and faster than it did in the 2010s. But if you buy at today’s prices, it will take a long time before the dividend alone gives you a yield like the old 4%.

At S$9.26 and 20.5 cents, you start at about 2.2%. Even if the dividend grows 6% a year, you’d need about 11 to 12 years for the yield on your purchase price to reach 4%.

So, what are you really buying at S$9.26?

I think I’ll leave the final call to you. But here’s how I’d sum up what I found.

Part of the climb was earned, and part was market giving ST Engineering a better valuation. From the end of 2024, about half of the gain came from real earnings growth. The other half came from the market deciding ST Engineering deserved a P/E of 27 instead of 20.

I wouldn’t count on the P/E going higher. For 15 years, this stock traded at about 20 times earnings. The re-rating to the high 20s has already happened. From here, you’d better hope it isn’t a P/E contraction. If the market loses faith and goes back to paying 20 times, even good earnings growth would struggle to lift the share price.

The earnings part is priced for 4 to 7 more years of strong growth. At S$9.26, the market expects ST Engineering to keep growing profit about 12% a year until roughly 2031 before tapering off. Some of that is backed by a big order book and management’s 2029 targets. The years after that are a bet on the defence cycle, engine repair demand and smart city projects lasting into the next decade.

Do you agree that it can keep that up? That’s for you to decide.

But here’s the part I think is worth sitting with, even if you do agree.

All the values in my scenarios use a 7% required return. That means if ST Engineering does exactly what today’s price expects, growing about 12% a year for about five years, you would earn about 7% a year from here, including dividends. In investor terms, your XIRR would be around 7% a year.

That’s a decent return. But it’s a long way from the 100% in a year that recent shareholders enjoyed.

To do much better than 7%, you need the growth to last longer than five years, or the market to pay an even higher P/E. To do worse, the growth only needs to fade a bit earlier, or the P/E only needs to slip back toward its old average of 20 times.

So if you’re looking at ST Engineering today because of how well it has done, I think it’s worth moderating your expectations. The business has genuinely become better. But a good business and a good investment are not always the same thing. It depends on the price you pay.

You can review my past spending logs by going to the personal notes section of investment moats here in the future.

How I Group My Spending.

One of the reasons for publishing the spending is to show people just how spending should be. Some would wonder if it is consistent or erratic, high or low. If you are interested, you see it, and you form your own tale about it.

You will be able to tune in to a spending profile of someone who is single, older, a fully paid up home. If you spend more than this, then you can ponder why is that and what you think about it. If you spend less than this, then you can ponder why and what do you think about it.

I group my spending based around a few technical grouping:

1. Flexible or Inflexible: There are some spending that we can be more flexible with. The spending tends to fluctuate over time. There are some spending that is more inflexible. The impact of this is felt more if you are retrench from work, wish to take a hard pivot in your life or career, planning for financial security or future retirement. A more inflexible spending would require your planned income stream to be more conservative while you can take some more risk if you have flexibility in your spending.
2. Finite or Ongoing: There are some spending that will stop at some point but there are some spending that we don’t see it stopping objectively. Finite spending are insurance premiums, mortgage, allowance for kids, allowance for parents etc. Ongoing spending is a certain kind of transport spending.
3. Role or Responsibility: What am I currently? Am I a worker? A husband? Or a Son? Some of the spending are group this way so that we are able to see just how much we are spending on something. Some of these responsibility will go away. For example, you spend on some travels, clothes, pay income tax because you are a worker. But if you are planning for a non-working phase of life, would you wish to know how much that you spend today can be peeled away.

In a way, this works for me because I always have an eye from the Financial Independence planning perspective. You might not, and you have your own reasons.

I hope that your way of grouping have some sensibility to it and helps you in your own way.

This is the spending for the month of July. Illustrations can be found on my Instagram here.

This month’s spending is pretty even, though it’s a bit higher than normal. I think that’s mostly down to a few one-off items.

I will break how I spend in the subsequent sections.

The Essential Inflexible Spend

I wrote about my essential inflexible spend here in this note: What kind of lifestyle would I need to buy for myself?

I group and track this set of rather inflexible spending to be reflective about how realistic is the numbers as part of the notes written above. The income stream from my Daedalus Income Portfolio is meant to pay for this spending as well.

The food spending is mainly the groceries that I use in my meal prep. I would eat 1 meal mainly a day, but think 15 days in a month, I would usually eat 2 meals. So what you see is just that.

Groceries and food bill is at $86, which is much lower than the 2 x $140, $124, $190, $167 and $161 last 6 months. This is still lower than the 2023’s essential food planning of $360 per month. Transport rose above $100 for the first month of the year. I think CDC help a fair bit in keeping the food spending to this amount.

Utilities include the town council fee, Zero1 $7 plan for 200GB that I use as my home broadband and Giga mobile plan.

Overall, the essential inflexible spend is lower than the $933 monthly plan for Daedalus Income portfolio. The base annual inflexible essential spending budget was $854 in Nov 2022 and $933 is adjusted for 3 years of US inflation.

Here is the US CPI for the past few years while we are on this:

The Basic Inflexible Spending

The difference between the basic and the essential is that.. basic inflexible are less essential. You can probably think of something that you can at least be slightly more flexible… but actually not too flexible about that would make life more sane.

You want to renovate your place every 10 years or replace something right? You want to try and make life easier on an ongoing basis right? You want to take better care of your health right? Some of you will have strong opinions about some areas, more than others. To you the spending can be rather inflexible, but if you touch your heart, you know you can cut if the conditions are really bad.

If it fits those it ends up here.

This is also eventually what Daedalus Income Portfolio also need to provide.

I describe this spending in this note: Aside from my most essential spending needs, I need $5,160 yearly for my basic needs. I would set aside $174,000 to provide income for it.

I mentioned that I got a new mattress last month. This month I’ll pay off the rest of the down payment for the bed, which costs up to $375 once I receive it.

The other expense, which works out to $260, is actually my routine dental check. I know I put it off for far too long, and I’m sorry it took this long. But we finally got one checkup in place.

The Flexible Spends

I would group most of my spending that are flexible into one big pot. Realistically, this is the spending that make life interesting.

Realistically, we can be more flexible with these spends as well.

I find this spending to be less important for long term retirement planning but you might hold a different opinion. I do think that if you wish to enjoy something, maybe you would also want to work for it. But there are some sub-accounts that allow me to tune in to them.

I attended two funeral.

My colleague Elyse’s mom passed away.

I also got a shock when I got to know my ex-colleague passed away.

The $300 is the Pek Kim contribution to the respective funeral.

I bought a few things under Hobbies & Entertainment.

I saw someone giving away this Ikea Trotten. The Trotten is a desk that you have to manually turn the lever to change its height. Sounds very troublesome but I don’t actually change height so much. However, the lever is immensely useful to adjust the height a little to see if it is the right height.

I was shocked that I got the size wrong but its not too bad. I paid $60 for a transport and man. I went down with the transport to help carry the table back.

I think my back isn’t in great shape, and my piriformis experience taught me that strengthening it works better than just stretching it.

After watching a few videos, I think back extensions are the way to go. They’re hard to do well without the right equipment, so I went to Lazada and bought a back extension machine. It took a while to arrive because it shipped from China by sea, probably about three weeks. All in, it cost me about $80.

I also bought a bunch of stuff:

• PowerPac 6-gang power extension — $18.00
• Dell/Lenovo/Acer power adapter converter — $17.35
• Laptop holder — $14.50
• Leobog Graywood v4 switches (50 pcs) — $13.38
• EDC pocket mini pouch — $10.52
• Leobog Ice Soul switches (30 pcs) — $8.10

I spent $45 on coffee for myself this month:

This month’s treat for colleagues at Providend: $103

The break down:

The people I treated:

Spending Funded from Other Sources

These are spending that comes from sinking fund/saving groups that we capitalized.

This would be mainly for my real insurance protection needs. You can read more about them here: Cutting My F.I. Capital Needs for Insurance Premiums from $131,366 to $58,132 by Prepaying for It.

If I capitalized this spending, I take it out from my recurring spend. A few line items of my insurance are really finite. Term insurance is. My $50,000 Limited Whole Life is. My health insurance is in my inflexible essential spend above, which I have plan as part of my expenses even if I am not spending it today (so as to shore it up for the later years).

1. Singlife Mindef: $30.35
2. TM Limited Whole Life: $95.75
3. Singlife Disability Income: $36.05
4. Singlife Term Protector: $585

This month is greater due to annual Singlife Term Protector premiums.

Spending on Work

I don’t have a lot of responsibilities like a lot of other people. Spending on work is a way to track those spending that is directly attributed to work. If I stop working this goes away.

Not much spending at work this month.

Check out my past spending reports.

You can find my past spending reports in my Personal Notes section. It is under my spending logs.

If you want to trade these stocks I mentioned, you can open an account with Interactive Brokers. Interactive Brokers is the leading low-cost and efficient broker I use and trust to invest & trade my holdings in Singapore, the United States, London Stock Exchange and Hong Kong Stock Exchange. They allow you to trade stocks, ETFs, options, futures, forex, bonds and funds worldwide from a single integrated account.

You can read more about my thoughts about Interactive Brokers in this Interactive Brokers Deep Dive Series, starting with how to create & fund your Interactive Brokers account easily.

I was struggling to stay asleep for a couple of days so my hope is to come home and don’t think about much, to go and sleep.

Basket, I woke up in the middle of the night, decide to check TradingView and its down like 3%. But I am glad that I slept much, much better than 2 nights this week. Its damn weird to lose probably $30,000 in a night but still feel glad about things.

We are probably 6% down from the all time highs on the S&P 500. Here’s where we are at:

We are probably back to May 2026 prices.

The orange line shows the percentage of S&P 500 stocks above their 200 day moving average and its about 45% still. Weak but in real weakness, they can get really low. At the bottom in 2025, its at 16%.

The second line shows the 3-month VIX divide by spot VIX. This shows the relationship between longer term VIX and shorter term. Longer term VIX prices in how much volatility the market participants expect in the longer term. If they feel its bad, then the 3-month VIX will not be low. The spot VIX shows where we are at now.

If you divide longer term with shorter term, the smaller this number gets the more it means currently its so so bad relative to the future which people think is better.

If I zoom out, you can see where we are at:

Its where most major bottoms are but that doesn’t mean no more additional drawdown.

The bottom breadth is the RSI, and objectively, its not even near the traditional oversold region.We don’t even need to talk about negative divergence yet.

After the 2018 Midterms, the S&P 500 Fell 16% in a Seasonally Strong Period

I think the about turn after the market did so well yesterday would have shock a few people. This was after the Democrats took both the House and the Senate.

There is a lot of similarities potentially between this mid-term and the mid-term of 2018.

In 2018 we have a late market draw down. A late market draw down in November and December is or should I say was… less common because those are seasonally strong periods.

The 2018 mid-term results was announced around 6 Nov:

The market puked like 16%.

There are definitely some similarities and some differences. I’m going to walk through some of the differences, since some of the earlier things should help you guys understand them better.

Firstly, was the election outcome.

2018 produced a split Congress.

Democrats took the House and Republicans kept the Senate, and markets mostly read that as gridlock. In 2026, Democrats won control of both the Senate and the House. They are projected to hold at least 54 Senate seats, with the House majority estimated at 251 to 268. Washington is still divided, though, because President Trump keeps his veto over any federal legislation.

Secondly, this sell off at this point, may have a clearer policy target while in 2018, people think the cause wasn’t so much the election.

This time the selling is tied directly to the results, through state-level data-center policy that doesn’t require Congress.

1. Goldman Sachs estimates that from January, Democratic-controlled states backing data-center moratoriums will hold 45% of planned US data center capacity.
2. Texas governor-elect Gina Hinojosa has proposed a statewide moratorium on new data-center permits, reportedly exempting projects already under construction.
3. Wells Fargo says Texas holds 27% of planned capacity. There are offsetting points: much of the capacity due in 2027 is already being built, so the bigger headwind may not arrive until 2028. Alphabet, Meta and Microsoft also largely held steady after the election, while the neoclouds took the hit.

Thirdly, the environment with oil prices are pretty different.

In late 2018, crude fell roughly 40% from its October peak, which eased inflation and later gave the Fed room to pivot. Now oil is near $96, and Brent rose 4.1% on Thursday, with the Middle East conflict keeping energy prices up. That makes a quick Fed pivot like January 2019 harder.

That said the similarities is:

1. It’s generally a blue wave taking place.
2. The first day of the market after the result was good.
3. Distance from S&P 500 all-time high was about 6%
4. The leaders in 2018 were the FAANG, semis this time its AI chips data centers and neoclouds.

Could we see another 16% drop toward the end of the year, just like in 2018?

I think the lesson here is that end-of-year seasonality is usually a strong period, but a year like 2018 shows it can go the other way.

Whether that was a coincidence or not, it reminds us that it can happen. I might be wrong, but I think the best lesson is to stay open-minded.

I don’t think it’s a case of “end of the year, so put the money in and don’t miss the November and December run.” I think it’s usually more about the November run.

I’m pretty sure that’s especially true for someone like me who’s in small caps. US small caps, I think, have their strongest run around the end of one year and the start of the next. Still, some years are different. So what we’re doing here is a little bit of market timing, and I know that’s not something to take lightly.

I have tracked almost every issue of Singapore Savings Bond since its issue in Oct 2015.

The December 2026 SSB bonds (SBDEC26 GX26120A) yield an interest rate of 2.39%/yr for the next ten years. You can apply through ATM or Internet Banking via the three banks (UOB, OCBC, DBS)

However, if you only hold the SSB bonds for one year, with two semi-annual payments, your interest rate is 1.83%/yr.

The one-year SSB yield seems to be heading down, showing a less flat curve.

$10,000 will grow to $12,407 in 10 years.

The Singapore Government backs this bond, which you can invest in if you have a CDP or SRS account (this includes Singapore Permanent Residents and Foreigners).

A single person can own not more than SG$200,000 worth of Singapore Savings Bonds. You can also use your Supplementary Retirement Scheme (SRS) account to make a purchase.

You can find out more information about the SSB here.

Note that every month, there will be a new issue you can subscribe to via ATM. The 1 to 10-year yield you will get will differ from this month’s ladder, as shown above.

Last month’s bond yields 2.45%/yr for ten years and 1.67%/yr for one year.

Here is the current historical SSB 10-Year Yield Curve with the 1-Year Yield Curve since Oct 2015, when SSB was started (Click on the chart, and move over the line to see the actual yield for that month):

How to Apply for the Singapore Savings Bond – Application and Redemption Schedule

You will apply for the bonds throughout the month. At the end of the month, you will know how many of the bonds you applied were successful.

Here is the schedule for application and redemption if you wish to sell:

You have from the second day of the month to about the 25th of the month (technically the 4th day from the last working day) to apply or decide to redeem the SSB you wish to redeem.

Your bond will be in your CDP on the 1st of the following month. You will see your cash in your bank account linked to your CDP account on the 1st of next month.

You May Not Get All the Singapore Savings Bonds That You Apply For

Do note that when you apply for the Singapore Savings Bonds, you may not get all that you apply for. Think of this as you are bidding for an amount which is determined by the demand and supply of Singapore Savings Bonds.

When the interest rate is low, the demand tends to be lower relative to history, and you can get a more significant amount. Still, if the interest rate is very high, demand can be so overwhelming that you may get a small portion you apply for.

For example, in the August 2022 issue, you can apply for $100,000, but the maximum allotted amount per person was $9,000 only. If you applied for $8,000, you would get your total $8,000 allocation.

To review the past allotment trend, you can take a look at SSB Allotment Results here.

How do the Singapore Savings Bonds Compare to SGS Bonds or Singapore Treasury Bills?

Singapore savings bonds are like a “unit trust” or a “fund” of SGS Bonds.

But what is the difference between buying SGS Bonds and its sister, the T-Bills, directly?

The Government also issues the SGS Bonds and T-Bills, which are AAA rated.

Here is a MAS detailed comparison of the three:

The main advantage of the 1-year SGS Bonds and Six-month Singapore Treasury Bills is that you can get a more significant allocation currently compared to the Singapore Savings Bonds. This means that if you need to earn a good interest yield of $400,000, you get a better chance to fulfil that with 1-year SGS Bonds and Six-month Treasury Bills.

The short-term interest rates are getting rather exciting, and short-term SGS bonds and treasury bills may be applicable to supplement your Singapore Savings Bonds allocation.

I wrote a guide to show how you can easily buy the Singapore Treasury Bill and SGS Bonds here. You can read How to Buy Singapore 6-Month Treasury Bills (T-Bills) or 1-Year SGS Bonds.

My Past Value Add Articles Regarding the Singapore Savings Bonds

Read my past write-ups:

1. This Singapore Savings Bonds: Liquidity, Higher Returns and Government Backing. Dream?
2. More details of the Singapore Savings Bond. Looks like my Emergency Funds now
3. Singapore Savings Bonds Max Holding Limit is $200,000 for now. Apply via DBS, OCBC, UOB ATM
4. Singapore Savings Bonds’ Inflation Protection Abilities
5. Some instructions on how to apply for the Singapore Savings Bonds

Past Issues of SSB and their Rates:

• 2015 Oct, Nov, Dec
• 2016 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2017 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov
• 2018 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2019 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2020 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Oct, Nov, Dec
• 2021 Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2022 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2023 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2024 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2025 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
• 2026 Jan, Feb, Mar, Apr, May, Jun, Jul, Aug

Here are your other Higher Return, Safe and Short-Term Savings & Investment Options for Singaporeans in 2026

You may be wondering whether other savings & investment options give you higher returns but are still relatively safe and liquid enough.

Here are different other categories of securities to consider:

[supsystic-tables id=1]

This table is updated as of 17th November 2022.

There are other securities or products that may fail to meet the criteria to give back your principal, high liquidity and good returns. Structured deposits contain derivatives that increase the degree of risk. Many cash management portfolios of Robo-advisers and banks contain short-duration bond funds. Their values may fluctuate in the short term and may not be ideal if you require a 100% return of your principal amount.

The returns provided are not cast in stone and will fluctuate based on the current short-term interest rates. You should adopt more goal-based planning and use the most suitable instruments/securities to help you accumulate or spend down your wealth instead of having all your money in short-term savings & investment options.

If you want to trade these stocks I mentioned, you can open an account with Interactive Brokers. Interactive Brokers is the leading low-cost and efficient broker I use and trust to invest & trade my holdings in Singapore, the United States, London Stock Exchange and Hong Kong Stock Exchange. They allow you to trade stocks, ETFs, options, futures, forex, bonds and funds worldwide from a single integrated account.

You can read more about my thoughts about Interactive Brokers in this Interactive Brokers Deep Dive Series, starting with how to create & fund your Interactive Brokers account easily.

If you want to trade these stocks I mentioned, you can open an account with Interactive Brokers. Interactive Brokers is the leading low-cost and efficient broker I use and trust to invest & trade my holdings in Singapore, the United States, London Stock Exchange and Hong Kong Stock Exchange. They allow you to trade stocks, ETFs, options, futures, forex, bonds and funds worldwide from a single integrated account.

You can read more about my thoughts about Interactive Brokers in this Interactive Brokers Deep Dive Series, starting with how to create & fund your Interactive Brokers account easily.

I explained how I constructed this portfolio in Deconstructing Daedalus Income Portfolio and Why I Currently Invest in These Funds for Daedalus. You might not understand what I wrote below if you haven’t read this post.

All my personal planning notes such as income planning, insurance planning, investment & portfolio construction will be under my personal notes section of this blog.

You can also find the past updates similar to this in the personal planning notes section.

Portfolio Change Since Last Update

The portfolio was valued at $1.833 million at the end of September and is at $1.851 million at the end of September.

We reported a portfolio change of +0.98% or +$18,000 for October 2026.

The portfolio is valued in SGD because that is the currency that I would most likely be spending on.

As of 3rd November 2026, the portfolio is valued at $1.864 million.

Portfolio Attribution – Why did the portfolio do better/worse compare to last month [or a year ago if this is a December update]?

We all want to know what cause the portfolio to do better or worse. If you have just one fund that covers the MSCI World, or you have a bunch of funds, would you know if it did better or worse?

In this section, I try my best to explain the portfolio performance in my way.

Here are the primary security holding returns for the month-to-date and year-to-date for the funds that I own [the first table] and reference benchmark ETFs [the second table]:

This table shows the fund holdings denoting the month-to-date and year-to-date performance of the funds that I own.

The Major Index ETFs is present to compare the performance. Just to be clear, I do not own the major index ETFs and you should see the Daedalus Income Portfolio Holdings table.

The returns of all funds are in USD. This includes the performance of the Dimensional funds, which I use the returns of the USD share class so that the returns are comparable. I have also listed the major index ETF performance for comparison.

a. General Equity Performance

October was the month the AI trade finally cracked. The S&P 500 hit a record on 6 October and then slid about 3% to end the month down roughly 1%. The damage was concentrated in technology. The US tech sector fund XLK fell 8.3%, its worst month since September 2022. Weak results and soft guidance from Google, Amazon and Intel made investors ask a hard question: will all this AI spending actually pay off? On the last trading day Intel fell 28% and the US chip index fell 7% in one session.

What is interesting is what happened underneath. Health care, financials, consumer staples and energy all rose, and this cushioned most of the tech losses. Energy stocks gained about 14% as oil stayed high. Brent crude oil ended the month above $113 after a tanker hit mines in the Strait of Hormuz. So the average stock did better than the index this month, which is the opposite of what we saw in September. Money did not leave the market. It moved from companies that are spending cash today to companies that are making cash today.

Interest rates gave investors some relief. The Fed kept its rate at 3.75% to 4.00% on 28 October, but two officials voted for another hike, and the market now puts roughly a three in four chance on a December hike. Even so, bond yields fell. The US 10 year yield touched 5.31% on 5 October, the highest since 2002, then dropped to 4.94% by month end as investors looked for safety. Lower yields helped smaller companies, and US small caps recovered in the second half of the month.

Emerging markets were hit the hardest. The MSCI Emerging Markets index fell about 8%. Korea fell about 17% and Taiwan about 6%, because these markets are full of chip companies and foreign investors sold in record amounts. Japan’s Nikkei was down about 7% by the third week of the month for the same reason. Indonesia had its worst month in 18 years. Brazil fell about 12% after President Lula won another term, as investors worry about more government spending. India was the odd one out, roughly flat, because it has very little AI exposure.

Europe was a quiet winner. The STOXX 600 had its best month since June, helped by having far fewer tech companies than the US. Gold fell to a one year low, which surprised many people given the war and inflation, but with US bonds paying close to 5%, gold has a lot of competition.

For a Singapore investor, the US dollar strengthened a little, about 0.4% against the Singapore dollar, which gave US dollar funds a small lift. A plain global index fund lost around 1% to 2% in October. Portfolios that tilt toward smaller and cheaper companies, like mine, mostly did better, as long as they did not hold too much of Asia.

b. Developed Equity Performance

There were a few multifactor funds targeting the developed equities region in Daedalus:

I hope you look at them as a diversified group of equities that gives exposure to developed large cap and mid cap global equities. They also systematically gives exposure to cheaper and more profitable companies with a little bit of short term momentum.

My developed world funds had a mixed month, but the order was the reverse of September. IWDA, which tracks the MSCI World, lost 1.34%. JPGL was my best fund, up 3.24%, and GGRA gained 0.47%. AVGC lost 0.79% and IFSW was the weakest, down 2.40%. Dimensional Global Core Equity, which I do not own, gained 0.61%.

This makes sense when tech falls hard. JPGL spreads its money more evenly across value, quality and momentum stocks and holds far less of the mega cap tech names, so it did well when money rotated to health care, financials and energy. IFSW likely held more of the chip and tech names that had done well this year, since momentum tends to follow the winners, and those were the stocks that fell. In September it was the other way around.

Year to date, JPGL (14.61%), IFSW (14.02%) and AVGC (13.77%) are all well ahead of IWDA (10.37%), by about 3.5% to 4%. GGRA at 6.32% is still the laggard of the group. It had a decent month, but it remains the fund I keep an eye on.

c. Developed + Emerging Markets Equity Performance

The Dimensional World Equity sits as part of my SRS account. It is a single fund that gives exposure to the developed and emerging markets large cap and mid cap stocks.

You should compare this against the MSCI All Country World and MSCI All Country World IMI.

Dimensional World Equity was roughly flat in October, down 0.18%. ACWD, the large and mid cap all country index, lost 2.24%, and IMID, which adds small caps, lost 2.00%.

Dimensional tilts toward smaller and cheaper companies and holds less of the very large tech names. In a month when big tech and Asian chip stocks fell, that tilt protected it. It still has emerging markets inside, which is why it did a little worse than Dimensional Global Core Equity (up 0.61%), which only owns developed markets.

Year to date, Dimensional World Equity is up 13.56% against 10.52% for ACWD and 11.03% for IMID. Last month the lead was less than 1%. Now it is about 3%.

d. Emerging Markets Equity Performance

AVEM and EMSD is my emerging market exposure. One is a large, mid and small cap fund that should tilt towards value and profitability. The other is a pure emerging market small cap with no factor tilts.

This was a very poor month for emerging markets. AVEM lost 8.33%, slightly better than EIMI which lost 8.87%. EMSD, my emerging market small cap fund, lost 8.28%.

Korea and Taiwan make up a large part of these funds, and both were hit hard by the chip selloff and record foreign selling. AVEM did a little better than EIMI because it holds less of the very largest chip stocks. EMSD did not escape either. Smaller companies in Korea and Taiwan are often suppliers in the same chip chain, and EMSD also holds Brazil and Indonesia, which both had a bad month.

Year to date, emerging markets went from the strongest part of my portfolio to the middle of the pack in one month. AVEM is up 10.07% against EIMI at 10.68%. EMSD is only up 2.15%, and is the weakest fund I own this year. It is a small position, so the impact on the whole portfolio is limited.

e. Small Cap Equity Performance.

About 32% of the portfolio or 36.5% of the equity allocation is invested in Global Small Cap Value or Value-weighted funds. You should look at Dimensional Global Targeted Value, AVGS. USSC used to be a significant allocation but no more. I still leave the performance in for this year just so you can contextualize its performance.

Small caps had a good month, a nice change after September. Dimensional Global Targeted Value gained 4.21% and AVGS gained 1.75%. EMSD was the exception, down 8.28% for the reasons I covered in the emerging markets section.

International small cap value dragged as the small cap materials company did not do as well as the US companies.

US small value led the way. USSC, which I do not own, gained 4.22% and the S&P 600 (USML) gained 2.48%. The Russell 2000 (R2US) was roughly flat at minus 0.23%, because it holds more growth and loss making companies than the S&P 600, which only admits profitable firms. Global Targeted Value has a large weight in US small value, so it moved like USSC. AVGS held up less well because it also owns small caps in Japan and Asia, which fell. It still beat the world small cap index WSML, which lost 0.18%.

Year to date, AVGS is the standout of my whole portfolio at 20.04%, well ahead of WSML (12.12%), R2US (13.13%) and USML (16.19%). Global Targeted Value has recovered to 14.43%. Mid caps (SPY4) are up 10.25%.

f. Global Aggregate Bond Performance

11.5% of the portfolio is in iShares Core Global Aggregate Bond UCITS ETF (AGGU).

The chart below is the US government yield curve at 30th September (light blue) and 30th October (dark blue):

After a poor September, bonds bounced back. AGGU, my global aggregate bond fund, gained 1.52% and is now almost back to break even for the year at minus 0.14%. ERNA, the ultrashort US dollar bond fund, gained 0.48% and is up 3.18% year to date.

The US yield curve moved down by a large amount in one month, but only from the 1 year mark onward.

• Yields from 2 to 30 years fell by 0.32% to 0.39%.
• The 2 year fell 0.32% to 4.56%, the 5 year fell 0.38% to 4.71%,
• the 10 year fell 0.35% to 4.94%
• and the 30 year fell 0.33% to 5.31%.
• Bills of 3 months and below barely moved and still pay around 4.1% to 4.2%, because the Fed did not change rates.

The shape of the curve did not change much. The gap between the 10 year and the 2 year is 0.38%, close to last month’s 0.41%, so the curve still slopes upward. The 10 year went above 5.3% early in the month before falling back. When yields fall, bond prices rise, which is why AGGU recovered. With a December hike still possible, I expect bonds to stay volatile.

g. Currency Effect

The US dollar strengthened slightly against the Singapore dollar in October. USD/SGD started the month around 1.273 and ended around 1.278, a rise of about 0.4%.

Most of the move came in the last week. The US dollar was strongest around 1.281 on 28 October, when the Fed held rates but two officials voted for a hike, before easing slightly by month end.

Since most of my funds are in US dollars and my portfolio is measured in Singapore dollars, a stronger US dollar gave my returns a small lift of about 0.4%. It is not a big factor this month. Different sources show month end rates that differ by a few pips, so the exact figure may be a little different at your broker.

Role of Portfolio

I am building Daedalus to provide inflation-adjusted income that last: from today (age 45) for the rest of your life — potentially forever.

Daedalus is to provide for my essential and basic living expenses.

I am currently not drawing down on Daedalus.

I sized Daedalus versus the starting income that I can withdraw with the Safe Withdrawal Rate (SWR) framework. The SWR framework is the best way to create a perpetual inflation adjusted income stream that survives some of the most challenging 30-year, 50-year, 70-year income sequences.

You can read more about the spending I design Daedalus for and the SWR:

1. My notes regarding my essential spending.
2. My notes regarding my basic spending.
3. My elaboration of the Safe Withdrawal Rate: Article | YouTube Video

I will explain more about the income plan below.

Based on current portfolio value, the amount of monthly passive income that can be conservatively generated from the portfolio is

The goal of the portfolio is to generate steady, inflation-adjusted income to cover my essential living expenses. It’s built using a conservative initial withdrawal rate of 2.0–2.5%, which is designed to hold up even under extremely tough market conditions — including scenarios like the Great Depression, prolonged periods of high inflation (averaging 5.5–6% over 30 years), or major global conflicts. In other words, it’s stress-tested to withstand some of the worst financial environments in history.

Many did not realize inflation is not 3%, 3%, 3%, 3% every year but it will look like a sequence like such on top. Some of the toughest sequences that kills an income portfolio is not a 30% portfolio fall but persistently high inflation.

You will lose purchasing power, not just run out of money with a poorly design income portfolio.

If I want to preserve my purchasing power, then I need an income strategy to provide me an income that adjust for inflation in this way AND for the portfolio to last for a long time. This is why I rely on the Safe Withdrawal Rate (SWR) framework, which is a framework to size our portfolio capital to conservatively provide income even in high inflation sequences.

Here is a visual illustration of how the income stream will be based on the current portfolio value:

If I start withdrawing income based on a 2% SWR, or 2.5% SWR, my income progression will adjust based on last-year’s inflation rate. This allows me to preserve my purchasing power.

This hypothetical inflation sequence is challenging, and in year 15 the income is vastly higher than the start.

If your portfolio is lesser, or more, your starting income is different. You do not have to follow how conservative my initial withdrawal rate is. Even a 3% initial safe withdrawal rate is pretty conservative.

Amount of Cash Flow/Income Withdrawn/Extracted from Daedalus Income Portfolio

I wish to be fully transparent about the schedule of withdrawals from the portfolio because if the goal of the portfolio is eventually or currently provide income for spending, you would be interested to know how much is taken out from the portfolio.

There have not been any withdrawals or cash flow extraction for spending since the publication of the portfolio. I will update as and when it happens.

Capital Injected Into Daedalus

My goal for Daedalus is not to inject capital into the portfolio since its first report in May 2024. One of the reason is to show a portfolio that is buy-and-hold, instead of some weird “war chest market timing” strategy.

Usually the injection is if I have SRS commitments to reduce the tax expense.

Here are my historical Capital Injections to the portfolio:

Investment Strategy & Philosophy

After trying my best to learn how to invest for a while, the portfolio expresses my thoughts about investing at this point.

The portfolio is run in a

1. Strategic: allocation doesn’t change by short-term events.
2. Systematic: rules/decision-tree-based implemented either myself or an external manager.
3. Low-cost: investment implementation cost is kept reasonably low both on the fund level and also on the custodian level.
4. Passive: I spend relatively little effort mentally considering investments and also action-wise.

You can read more in this note article: Deconstructing Daedalus My Passive Income Investment Portfolio for My Essential & Basic Spending.

Portfolio Change Since Last Update (Usually Last Month)

There are no portfolio changes for the month.

Current Holdings – By Dollar Value and Percentages

The following table shows more details about the securities that I currently held.

The securities are grouped based on general strategy, whether they are:

1. Systematic Passive Fixed Income to reduce volatility.
2. Systematic Passive, which are equities that tries to capture the market risk in a systematic manner.
3. Systematic Active, which are equities that tries to capture various proven risk premiums such as value, momentum, quality, high profitability, and size in a systematic manner.
4. Hard Assets, securities that are more commodities related, that tends to do well in commodities related inflation situations.
5. Legacy Tracking Positions, securities that I still keep around, minimized to track their historical performance.

Portfolio Grouped by Account Money Source Location

Generally, you won’t have just one view about the securities in your portfolio.

In the following sections, I show my portfolio when viewed from different angles.

The first is the portfolio based on location.

1. Cash means held in accounts that we can make independent choices of which platform that we choose to invest in, when we decide to buy, when we decide to sell without any liquidity, tax, or locking considerations.
2. SRS Account is a Singapore-related retirement account. There are tax advantages on your ordinary working income if a person contributes to it. You can defer the income tax until after your retirement, where only 50% of your withdrawal then will be tax, at the prevailing tax bracket then.

This view does nothing much but some might be curious whether it makes up my money in CPF, here or there and so basically these are basically my cash monies and SRS.

Portfolio Grouped by Geographical Region Exposure of Securities

The second view groups the securities based on its geographical exposure.

Returns comes potentially from taking systematic risks and risks comes partly from the macro, interest and inflation exposure in different geographical regions.

The general regions:

1. Developed Markets – Strategies that systematically considers the large-cap and mid-cap equities in developed countries. You can view the countries, and sector composition at this MSCI World Index page.
2. Global – Strategies that systematically considers the large-cap, mid-cap, small-cap equities in both developed and emerging market countries. You can view the countries, and sector composition at this MSCI Emerging Markets IMI Index page.
3. US – Strategies that mainly tap small-cap US equities.

Portfolio Grouped by Fund, Cash or Individual Security

The third view groups the securities based on whether they are fund, cash or individual securities.

Almost 100% of the portfolio is implemented with funds. Funds can be:

1. Singapore Unit Trusts domiciled in Ireland.
2. London Stock Exchange listed exchange traded funds (ETFs) domiciled in Ireland.

Portfolio Grouped by Investment Strategy.

The last view groups the securities based on commonly known high level strategy names.

What Systematic Active Means: Funds that help me execute passively very specific, repeatable underlying securities selection on an ongoing basis. Here are some examples of the systematic active strategies in my portfolio:

1. Global Multifactor: From a basket of 1,600 developed market large and mid-cap stocks, rank the stocks by their value, by their 12-month momentum, by their degree of ROE and debt to asset, and then own the top 300. Do this every half-yearly or quarterly. You end up with a strategy that consistently owns 300 companies that are cheaper, quality and have greater momentum relative to a market cap weighted index.
2. Small Cap Value: From a basket of 3,000 developed market small cap stocks, rank the stocks based on price-to-book value (include intangibles in the book value). Also rank the stocks by operating earnings minus interest divide by book value. Eliminate the companies with low profitability. What we end up is two group of small cap stocks: The more profitable small caps but not too expensive, and the small caps stocks that are at least profitable but are very cheap. Own the top 30-35% of this cohort consistently. Have a manager that consistently helps me execute this.

In contrast, Systematic Passive are funds that help me track certain benchmark indexes. These indexes can be market-cap weighted, or equal-weighted, and reconstituted periodically so that they mirror the performance of benchmark indexes.

Hard Assets are the funds that provide exposure to energy-related securities.

Systematic Passive Fixed Income main helps damp the volatility of the portfolio. They are maintain based on the historical research that it is better to be less than 100% in equities if your portfolio is meant for income.

The fixed income/cash should not be viewed as a war chest to rebalanced to equity or take profit from equity. This is a strategic long term allocation whose main purpose is to optimized negative sequence of return risks.

The Main Custodians for the Securities in this Portfolio

The current custodians are:

1. Cash: Interactive Brokers LLC (not SG)
2. SRS: Philips FAME

If you want to trade these stocks I mentioned, you can open an account with Interactive Brokers. Interactive Brokers is the leading low-cost and efficient broker I use and trust to invest & trade my holdings in Singapore, the United States, London Stock Exchange and Hong Kong Stock Exchange. They allow you to trade stocks, ETFs, options, futures, forex, bonds and funds worldwide from a single integrated account.

You can read more about my thoughts about Interactive Brokers in this Interactive Brokers Deep Dive Series, starting with how to create & fund your Interactive Brokers account easily.

A lot of Singaporeans think this way too. The local banks feel like the last thing that would ever fail. And if something did go wrong, the government would step in.

I agree that the government is unlikely to let our major banks fail. But that protects the banking system, not the price you paid for the shares. Banking is still a cyclical business. If you buy at the wrong time, you may have to sit through years of falling prices and dividend cuts. That is normal for banks.

What got me thinking more deeply is everything that has happened over the past six years: blockchain, decentralized finance (DeFi), and now AI agents. Most of the noise has been about crypto prices.

This post is about something quieter: how these developments could change traditional banking and finance in ways that might surprise a lot of people.

I’m not writing this as a prediction. Roughly 20 to 25% of my portfolio is in financials, mostly through ETFs. So I want to think through what could happen, and what it means for how I invest.

How a bank actually makes money (traditionally)

Before we talk about what could change, it helps to understand the basic engine of a bank.

A bank borrows money from you and lends it to someone else. When you put money in a savings account or fixed deposit, you are effectively lending to the bank. The bank then lends that money out as mortgages, car loans and business loans.

The bank wants to pay you as little interest as possible, and charge borrowers as much as it reasonably can. The difference is its profit. This gap, measured as a percentage of what the bank lends and invests, is called the net interest margin, or NIM.

There is another layer.

Deposits are short-term money. You can withdraw your savings anytime, and fixed deposits usually last only months. Loans tend to be long-term. A mortgage can run 25 or 30 years. So the bank borrows short and lends long.

Usually, long-term interest rates are higher than short-term rates. If you plot interest rates against how long the money is borrowed for, you get what is called the yield curve.

When the curve is steep, the bank earns a fat spread. When it flattens, margins get squeezed. When it inverts, with short-term rates above long-term rates, the business becomes very hard to run profitably. Imagine paying 4% on deposits and lending out at 3%.

A nuance for Singapore banks

In practice, our local banks don’t fit this textbook picture neatly.

Many of their loans, such as mortgages and corporate loans, are floating-rate. They reset quickly when interest rates change. Meanwhile, a lot of money sits in current and savings accounts that pay very little.

So for DBS, OCBC and UOB, profits tend to move with the overall level of interest rates as much as with the slope of the curve. When rates are high, loan income rises quickly while deposit costs lag behind. When rates fall, margins shrink.

DBS shows this clearly.

In 2021, when rates were near zero, it guided for a net interest margin of 1.45% to 1.50%. By 2023, after the rate hikes, its margin had risen to 2.15% and its net profit crossed S$10 billion for the first time. Same bank, same business, very different earnings. That is the interest rate cycle at work.

Margins aren’t everything: volume, funding and bad loans

A thin margin doesn’t automatically mean a bad year for a bank.

Central banks raise rates when the economy is running hot. Short-term rates go up, and long-term rates may not rise as much, so the margin narrows. But a hot economy also means more people and businesses want to borrow. A bank can make up for thinner margins by lending more. Volume matters.

How much a bank can lend is limited by how much funding it has, and deposits are the main source.

When a bank wants to grow faster than its deposits allow, it can borrow from other places, such as short-term loans from other banks and institutions. This is often called wholesale funding. It usually costs more and is less stable than ordinary deposits.

A bank that is hungry for profits, perhaps because shareholders demand growth, may lean on this. It works well in good times. It is also what hurts the most when the downturn comes.

The real test is the down cycle

Banks are tested in recessions, not booms. The question then is: who did you lend to, and can they still pay you back when times are tough?

A good loan isn’t one where the borrower never struggles. It is one where the borrower struggles but still repays.

Many Singapore investors today have not lived through a proper credit cycle. We did have a recession in 2020, but bad loans never really spiked, because loan moratoriums and government support absorbed much of the damage. It would be unusual to go 20 years without a downturn that actually bites.

If you look at banks that have survived 100 years or more, including many US regional banks, they usually have one thing in common. They have a disciplined way of deciding who to lend to, built up over decades, almost like a standard operating procedure. That conservatism costs them growth in good times. It is also why they are still around.

So as an investor, you have to ask yourself: do you prefer a bank that plays the long game? Or one that believes it is good enough to play the long game and the short game at the same time?

The real moat: lazy money

I think this is where most would underestimate the part for some disruption.

The big three local banks have a huge advantage that has little to do with clever lending. People trust them, and people don’t move their money.

Think about your own savings. Most of it probably sits where it has always been. Not because that is the best rate available, but because moving money is a chore. You would need to open a new account, verify your identity, set up transfers, and maybe change where your salary goes. Most people don’t bother.

I think of this as lazy money, and banks earn a lot from it. Because their deposits don’t move, the big banks can pay less interest than they would otherwise need to. Smaller players such as Sing Investments & Finance or Singapura Finance have to offer more attractive rates to win deposits.

Friction is part of the moat.

When my fellow executive directors at Bigscribe, Alvin and Thomas, and I went to close our company bank account, parts of the process still required a trip to the branch. Even today, some of banking is surprisingly old-school.

What happened when rates went up

The 2022 to 2023 rate hike cycle gives us a preview of what happens when people do start moving their money.

• September 2022: UOB raised its 12-month fixed deposit rate to 2.6% and OCBC to 2.3%, while DBS kept its top rate at 1.3%. People queued at bank branches, some for hours.
• December 2022: The 6-month Singapore T-bill cut-off yield hit 4.4%, with S$9.3 billion of applications for a S$4.6 billion issue.
• Early 2023: OCBC offered 4.08% on an 8-month fixed deposit.
• Mid-2023: OCBC and UOB cut their rates to 2.70% once deposit outflows had stabilised and they no longer needed cash as urgently.

Two things stand out.

First, DBS largely stayed out of the promotional race and still posted record profits. It had the stickiest deposits, so it didn’t need to pay up. That is the moat working.

Second, the promotions were short and targeted. Banks raised rates when they needed money and cut them as soon as they had enough. This only works because most people don’t watch rates constantly. Keep that in mind. It becomes important later.

What are “rails”?

People in finance and tech talk about rails a lot. It is a useful idea, so let me explain it simply.

Rails are the systems that move something from one place to another. Think of a railway track. The track itself doesn’t do much. But once it exists, trains run on it, towns grow around the stations, and a lot of activity happens because things can now move easily.

In finance, rails move money and information. They have evolved over time:

1. Paper and post: cheques and letters, sent and processed by hand.
2. Telephone and telex: instructions given over the phone or wire.
3. Digital networks: SWIFT for international transfers, and GIRO, FAST and PayNow for local ones.

Each step made moving money faster and cheaper.

Another way to picture it: rails are like roads.

Today, the banks own most of the roads and the toll booths. Switching banks is like leaving the expressway, finding a new route and paying a new toll. It is a hassle, so most people stay on the road they are on.

Rails are more than pipes

Good rails don’t just move things. They also handle what goes wrong.

For an online travel agency, the hard part isn’t showing you a list of hotels. It is dealing with disputes. The hotel overbooked. Your flight was delayed and you need to cancel. You want a refund.

For banking, the hard part is trust and safety. Will all my money arrive? Is it going to the right person? If something goes wrong, who fixes it? Is the place receiving my money safe?

So when people talk about building better rails, they mean building a better supply chain for money. Faster and cheaper, but still trustworthy when things go wrong.

What’s quietly happening on the rails

Blockchain and DeFi were hottest around 2020 and 2021.

Five years later, it might look like not much has happened, at least at our local banks. The most visible change is that you can now buy and hold some digital currencies through bank platforms.

But quite a lot has happened below the surface. It just isn’t aimed at you and me. Most of it is about rebuilding the business-to-business (B2B) rails: how banks, companies and institutions move money and assets between each other.

A few examples:

• DBS Token Services. Launched in 2024 for institutional clients. Its core product, Treasury Tokens, are essentially tokenized deposits that allow round-the-clock cross-border payments between DBS branches. For a company’s finance team, this means moving cash between, say, its Singapore and Hong Kong subsidiaries on a Sunday night, and building rules into how payments happen.
• Partior. A blockchain payments network formed by JPMorgan, DBS and Temasek, which grew out of MAS’s Project Ubin. It lets banks settle cross-border payments in multiple currencies in real time, with both sides of a currency exchange settling at the same moment.
• Project Guardian. MAS’s collaboration with the industry to test tokenized assets such as bonds, funds and foreign exchange. It is a testing ground, not a regulation.

What does “tokenized” mean? It simply means recording who owns something on a blockchain instead of a traditional database.

Take a company like ST Engineering.

It has a fixed number of shares, and each share is a record of who owns a slice of the company. Long ago, that record was kept on paper. Today, it sits in a digital database.

A blockchain is just another kind of database, one that many parties can share and trust without needing a single central record-keeper. So moving share ownership onto a blockchain is not as wild as it sounds.

It’s not just Singapore

The same thing is happening overseas, at banks big and small.

The best-known example is JPMorgan, the largest US bank.

It runs a blockchain unit called Kinexys (formerly Onyx). Its deposit token, JPM Coin (ticker JPMD), is a digital version of US dollar deposits held at JPMorgan. It went live for institutional clients in November 2025 on Base, a public blockchain network built by Coinbase, and settles almost instantly, around the clock.

In October 2026, it was also made available on Ethereum. Only clients that JPMorgan has approved can hold it.

Citi, BNY and others have similar products. But this creates a new problem.

Why banks need shared rails

A JPMorgan deposit token is a claim on JPMorgan. It can only move between JPMorgan’s own clients. A Citi token can only move between Citi’s clients.

Think of it like store vouchers. A voucher from one shopping mall is useful inside that mall, but you can’t spend it next door. If a company keeps money with five banks, it now holds five kinds of vouchers that can’t talk to each other.

Say a company wants to move US$100 million from its Citi tokens to JPMorgan tokens. Today, it would have to cash out the Citi tokens, wire the dollars to JPMorgan the old way, and then ask JPMorgan to issue new tokens. That defeats much of the point.

So banks are now building shared networks, so that tokens from different banks can be swapped in a single step. In the US, there are two to know about: one for the giants, and one for the regional banks.

The big banks’ network: The Clearing House

The Clearing House is a payments company owned by the large US banks. It already runs two of America’s core payment systems, RTP for instant payments and CHIPS for large transfers, which together clear and settle more than US$2 trillion a day.

In June 2026, it announced a shared network for tokenised deposits, backed by banks including JPMorgan, Bank of America, Citi and Wells Fargo.

Around 17 large banks are now behind it, and it is expected to open to banks of all sizes in the first half of 2027.

In September 2026, it picked a technology firm called Quant to build it.

How it is meant to work, in simple terms:

1. Each bank still issues its own tokens and looks after its own customers’ wallets.
2. A shared, permissioned ledger connects all the banks. “Permissioned” means only approved institutions can join, unlike public blockchains where anyone can take part.
3. When a token moves from one bank to another, the actual dollars settle over RTP or CHIPS at the same time.

So what is the result?

That US$100 million move becomes one instruction instead of three steps, any time of day.

Payments can also be programmed to happen automatically when certain conditions are met, such as paying a supplier the moment goods are delivered.

The regional banks’ network: Cari Network

This is the one I find more interesting, because it shows what smaller banks are doing to avoid being left behind.

Cari Network is being built by five US regional banks:

Together, they hold roughly US$780 billion in assets. That sounds big, but each is a fraction of JPMorgan’s size.

Cari is led by Gene Ludwig, a former US Comptroller of the Currency, which is the top regulator of America’s national banks. So it has serious regulatory credibility behind it.

A few things make it different:

• One shared token for all the banks. Instead of five sets of vouchers, there is one token that works the same at every member bank. You get it through your own bank, but a token from Huntington is identical to one from M&T. The network controls how many tokens exist on behalf of all members.
• Private, but checkable. Only approved participants can use it, and transactions are kept private, while regulators can still audit activity when needed. It runs on Prividium, a permissioned network built on ZKsync technology and anchored to Ethereum.
• Starting small. At first, money can only move between customers of the five banks. The plan was a basic working version by March 2026, a pilot with customers in the third quarter of 2026, and full commercial launch by December 2026.

Why banks are doing this

The short answer: to protect their deposits.

Stablecoins such as USDT and USDC are digital dollars issued by companies, not banks. When you hold them, your money is no longer a deposit in a bank.

A January 2026 Standard Chartered report estimated that US banks could lose US$500 billion in deposits to stablecoins by 2028, with regional banks hit hardest.

The banks think tokenized deposits is the answer.

1. You get the speed of a stablecoin, but your money stays a regular bank deposit.
2. It stays on the bank’s balance sheet, it can still be covered by deposit insurance, and the bank can still lend it out.
3. With a stablecoin, the backing money usually sits in government bonds instead, and the bank earns nothing from it.

You can notice two things.

First, the giants can build their own tokens, while the regional banks have chosen to team up. That is the “plug in rather than build alone” pattern I mentioned earlier.

Second, there is an irony here.

These networks are meant to keep money inside banks, but if moving a deposit from one bank to another becomes a single instruction, it also becomes easier for money to leave your bank for a competitor. The banks may be protecting the banking system’s deposits while weakening each individual bank’s lazy-money moat.

How big is this today?

I think its not that big yet.

DBS doesn’t publish volumes for Token Services. The best public benchmark comes from Citi, whose similar token service handles about US$1 billion in transactions, though Citi didn’t say over what period. That is tiny compared with the trillions banks move through traditional channels.

Most Project Guardian pilots have been limited to accredited and institutional investors, and tokenised money-market funds are the most common live use case. MAS’s proposed shared ledger for tokenised assets, Global Layer One, is still in development, with first production use expected around 2026 to 2027.

Slow markets do change eventually, though.

Bond trading stayed old-school, with deals done over phone and chat, long after stock trading went electronic. That has been shifting in recent years.

The rails often change quietly first, and ordinary customers like us only notice it years later.

It’s a bit like the ATM. Before it came along, you had to visit a branch and queue to take out cash. Then one day, you didn’t.

A lesson from the travel industry

To see how AI could change finance, it may help to look at an industry where it is already shaking things up.

For years, if you wanted a hotel, you went to Booking.com, Agoda or Expedia. These sites, called online travel agencies (OTAs), give you one place to compare thousands of hotels. They earn a commission, called a take rate, on each booking. Much of their value came from the website or app itself: the screen you were used to searching on.

AI agents may change that. An AI agent is an assistant that doesn’t just answer questions but takes action for you. You tell it what you want, and it searches, compares and books the hotel.

In September 2026, Meta’s AI agent Muse began letting users plan and book hotels. On 23 September, Expedia fell 7.7%, Airbnb 7.6% and Booking Holdings 5.1% in a single day. Investors feared that if an agent can understand what a traveller wants, compare options across many sources and complete the booking, the travel agency gets cut out.

There is a weird part to the sell off LOL.

The selloff was triggered by Expedia joining Muse, so that users could book Expedia’s hotels without ever visiting Expedia’s site. And Meta reportedly plans to charge a commission on purchases made through Muse.

Going with an agent may not remove the middleman. The new middleman may be Muse.

Is it the shop or the service?

The real question for any platform is this: strip away the interface, and what value is left?

Imagine hotels connecting directly to AI agents through a common standard, paying a lower commission than the OTAs charge. If an OTA’s only value was the listing and the search box, its business is in trouble.

But OTAs also do things that are hard to replace. If the hotel messes up your booking, someone has to handle the complaint. If your flight is delayed and you need to cancel, someone has to process the refund. If you booked directly and the hotel doesn’t respond, was saving a few dollars still worth it? For many people, that service is what the commission really pays for.

The same logic applies elsewhere. A property portal whose only value is its listings becomes less important if the government opens up more property data that AI agents can access directly.

So the test is simple. Are you the shop, or are you the service?

Now apply this to your bank

Let’s bring this back to deposits.

Remember how banks offered short, targeted promotions in 2022 and 2023, then cut rates once they had enough cash? That only works because most savers don’t check rates every week, and moving money is a hassle.

Now imagine the next time rates go up.

Instead of reading blog posts comparing fixed deposit rates, I simply ask my AI agent:

find me the highest deposit rate among the banks and finance companies I’m comfortable with, check that they are financially sound, and move S$50,000 there.

The agent searches, compares and starts the transfer. When a better rate appears three months later, it moves the money again.

What used to be a high switching cost becomes a low one. Lazy money becomes alert money.

The banks’ dilemma

What can the banks do?

They can pay higher rates to keep deposits. But that eats directly into the net interest margin that made them so profitable in the first place.

The short-burst promotion also stops working. If agents are watching rates all the time, a bank can’t pay up for a few months and then quietly cut. It has to stay competitive across the board.

We have already seen how fast money moves once the friction is gone. In March 2023, customers pulled US$42 billion out of Silicon Valley Bank in a single day, in what has been described as the first US bank run in a world of online withdrawals. That was panic, not rate-chasing. But it showed that deposits are only as sticky as the effort it takes to move them.

The flip side

This isn’t only bad news for the big banks. A large bank that builds these new rails early could capture volume it never could before, from customers locally and overseas. Lower friction cuts both ways. It also makes it easier for money to flow in.

Why it’s not that simple

Before we can conclude that the big banks are going to get their lunch eaten, there may be some important frictions that won’t disappear easily.

Trust still matters

Moving money is always a trade-off between how much you earn and how much you trust where it goes. Many people will accept a slightly lower rate to keep their savings with a bank they know.

But how much trust do you actually need?

In Singapore, the Deposit Insurance Scheme covers Singapore-dollar deposits up to S$100,000 per depositor at each full bank or finance company. The limit was raised from S$75,000 in April 2024, and all full banks and finance companies are members unless MAS exempts them.

So for most retail savers, the safety gap between DBS and a smaller finance company is narrower than many people assume. A good AI agent would likely point this out.

The agent is the front door, not the rail

Let’s walk through how this would actually work.

Say my agent finds that Singapura Finance offers the best rate. It pulls up the company’s track record, how well capitalized it is, and whether my deposit would be insured. I approve, and the transfer happens.

The money still travels over existing rails, such as PayNow or FAST, from my DBS account to Singapura Finance.

The chatbot is just a new front door. Behind it, the banks still do the heavy lifting: checking where the money came from (source of funds), how the person built their wealth (source of wealth), and screening for fraud. That responsibility, and the risk that comes with it, stays with the institutions.

More volume means more cost

If agents send a flood of new deposits to a small player like Singapura Finance, can it cope?

More customers also means more exposure to fraud and cyberattacks. To capture the opportunity, smaller institutions will need to invest in systems, compliance and security. That eats into their advantage.

Fast money can also leave fast. In March 2025, Chocolate Finance suspended instant withdrawals after a rush of requests, triggered by a poorly communicated change to its rewards programme Customers’ money was not lost, and the firm processed over S$500 million in withdrawals within a week. But it showed how quickly sentiment can turn when moving money is easy.

The middle still costs something

I think many people are too optimistic about how cheap things will become. They imagine the middleman disappears and fees fall to almost nothing. But there are real costs in the middle of every supply chain: handling disputes, making sure things work, meeting regulations.

It’s like a warranty on your phone. Some people happily skip it. But when the outcome is uncertain and costly, many of us want that protection, and that is part of what we pay for.

Wise is a good example of what real disruption looks like. Sending money overseas through a bank used to cost around 3% or more once you include fees and exchange rate markups. Wise brought that down dramatically. Its average cross-border fee fell to 0.50% in the quarter to June 2026, its lowest ever (Wise). Much lower, but still not zero.

What happened next is the interesting part.

Banks started using Wise as their rails. Standard Chartered now runs its SC Remit service on Wise’s infrastructure, and other partners include Morgan Stanley, Google Pay, Monzo and Nubank.

Two details are worth noting:

• Wise doesn’t use blockchain or stablecoins. It built a cheaper, compliant network on conventional rails. Disruption is about better rails, not necessarily crypto.
• Wise now holds US$41.2 billion in customer balances. That is money people would once have kept in a bank. It is a small but real sign of deposits quietly moving away from traditional banks.

Who builds the new rails?

When people hear “decentralised finance”, they often picture the extreme version.

Here is what DeFi looks like in practice. You take S$5,000 from your DBS account and convert it into a digital currency. You move it onto a blockchain such as Ethereum or Solana, and possibly onto a faster “layer 2” network built on top of them. There, you put it into a pool that lends to others or helps people trade, and you earn a yield.

Many people lost a lot of money doing this. Some losses came from token prices collapsing, a bit like holding a currency that loses half its value and never recovers. Others came from hacks, bugs in the code, and outright collapses such as Terra/Luna in 2022. Unsurprisingly, many people don’t trust DeFi.

But the change I’m describing doesn’t require everyone to move to DeFi. It is an evolution, not a leap.

A more likely path

A more realistic scenario looks like this. A company with a profitable core business builds better rails on the side. Other institutions, especially smaller ones, see the advantage of plugging in.

The players most likely to do this are challengers such as SoFi in the US and Nubank in Latin America. They have always tried to take business from the incumbents. They are big enough to set standards, but not so dominant that others are afraid to join.

Imagine one of them, or someone like them, creates an open standard: a common way for banks, finance companies and AI agents to connect. Other institutions see little risk in trying it. As more join, the network becomes more useful, which attracts even more members. That is a network effect. Then one day, someone at a big bank realises: we have a real problem.

It doesn’t even need to be a giant.

Picture a local firm like Chocolate Finance, which already has a tech team. Or even Providend.

Suppose I had an idea for a better way to move and manage money.

Building it isn’t as hard as it used to be. We bring in a few close partners. We explain it well and educate people on the safety and integrity questions they care about. The reviews are good, more people join, and it slowly chips away at someone else’s margins.

Evolution, not disruption

To be fair, I don’t think “disruption” is quite the right word. Finance has always evolved. We used to need more bank branches. Now we need fewer. Some of today’s experiments will become tomorrow’s standard rails, and we will stop noticing them.

History also suggests that incumbents often adapt and benefit. Better rails create new products to sell and new fees to earn. Big banks in the US, Singapore and elsewhere may move more volume and protect their profits that way. So it is genuinely hard to say who wins and who loses.

What I would say is this: don’t be surprised if, once these rails exist, deposits at the big banks start to shrink.

What this means for my portfolio

So why does any of this matter to me?

Even though most of my money is in exchange-traded funds (ETFs), around 20 to 25% of my portfolio ends up in financials with the ETFs.

Financials is a broad bucket. It includes traditional banks, small regional banks, newer players like SoFi and Nubank in tiny amounts, and insurers and specialty insurers, which are a whole different business that I know less about.

One specific worry: my Avantis Global Small Cap Value fund holds a number of smaller US regional banks. If these changes play out, some of them may need to merge, and a few could be badly hurt.

I have spent the past couple of years building tools with AI and seeing the friction in wealth and asset management up close.

I can see the potential.

But I honestly can’t tell whether this makes the overall pie for financial firms bigger or smaller.

Build a strategy that doesn’t need you to be right

After many years of investing, here is what I have learned: I can have opinions, but they are often wrong.

So I would rather not build a portfolio whose fate depends on whether I call this correctly. I expect some parts of my portfolio to be disrupted, including parts I currently feel confident about. That’s fine, as long as the overall plan still works.

If disruption never reaches traditional finance, a concentrated bank portfolio will do fine. But what if it does? Do you want sleepless nights because 80% of your portfolio sits in three banks?

Singapore REIT investors have been through something similar. Many were confident that blue-chip REITs would keep doing a splendid job. Then interest rates rose sharply from 2022, and many REITs fell a lot, even those with good management. You can debate some of those management teams, but the lesson is the same: good management can’t fight large forces.

That is why a broad ETF makes sense to me here. I don’t have to pick the winners. If the incumbents adapt, I own them. If the challengers win, I own some of them too.

The goal isn’t to predict the future. It is to build a wealth or income strategy that turns out reasonably well whether my opinions are right or wrong.