Monday 26 March 2018

Is Starhub's Dividend of 16 Cents Sustainable?

A few bloggers have written about whether Starhub's dividend of 16 cents per year is sustainable when Starhub released its results in Feb. This time last year, I reviewed the prospects of its various business segments and concluded that challenging times were ahead for Starhub's dividends. See Challenging Times Ahead for Starhub's Dividends for more info. In today's blog post, I will update the review and discuss whether Starhub's dividend of 16 cents is sustainable. Although mobile services is the largest business segment, I will discuss it last, for reasons explained later.

Pay TV

If you read Is Pay TV Still A Reliable Cash Cow?, you would know that the traditional advantage of Starhub over the other telcos is its cable TV network infrastructure, which is used to provide not only Pay TV but also cable broadband services. The cost structure for Pay TV is fixed capital cost for the network infrastructure, and mostly variable cost for the TV content. The network infrastructure has been in operations for many years and would have been mostly depreciated. Thus, the more subscribers Starhub can get on its Pay TV services, the more profits it can generate. Unfortunately, Pay TV has been on a decline in recent years due to competition from over-the-top service providers like Netflix and online privacy. Fig. 1 below shows the no. of Pay TV subscribers Starhub has since 2015.

Fig. 1: No. of Starhub Pay TV Subscribers

Although Starhub has been rationalising TV content to cut costs and investing into media companies like mm2 Asia to generate exclusive media content, it is difficult to see such measures being able to counter the decline in no. of subscribers. Furthermore, investments into media companies take away cashflows that could be paid out to shareholders. Thus, pay TV is no longer the cash cow it used to be.

Wired Broadband

The same story is repeated for the wired broadband business segment. Wired broadband can be categorised into cable broadband and fibre broadband. Although the total no. of broadband subscribers has remained steady, a breakdown of the no. of subscribers shows that the no. of cable broadband subscribers has continued to decline, in favour of fibre broadband. See Fig. 2 below for the no. of cable and fibre broadband subscribers.

Fig. 2: No. of Starhub Wired Broadband Subscribers

As mentioned above, cable broadband is served though the Pay TV network infrastructure, which is fixed capital cost and mostly depreciated. Thus, the less no. of cable broadband subscribers, the less profits Starhub can generate. On the other hand, fibre broadband is served though the Next Generation Nationwide Broadband Network run by Netlink Trust. Starhub has to pay Netlink a pre-determined amount of money for each subscriber. Thus, the profit margin for fibre broadband is lower than that for cable broadband.

Recently, Starhub has been running a promotion to get more customers on board its cable broadband services, but it is difficult to see Starhub being able to reverse the decline in the no. of cable broadband subscribers. Again, wired broadband is unlikely to become the cash cow it used to be.

Enterprise Fixed Network

This business segment is the rising star of Starhub, with increasing revenue annually. However, part of the revenue growth is from acquisition of companies. Exactly how much profits are generated from the revenue is unclear. Furthermore, as Starhub continues its acquisition path, this business segment will soak up more cashflow than it can generate. Last year, it spent $22.6M acquiring companies. This year, it has already spent $57.5M in doing so. So, do not expect this business segment to generate good cashflows to sustain its dividends.

Mobile Services

We have come to the most important business segment, which is the mobile services. Mobile services is under pressure from different sources, such as the SIM-only plans, data upsize plans, Mobile Virtual Network Operators (MVNOs) like Circles.Life, Zero Mobile and Zero1, plus TPG, the fourth telco. You can read more about the impact of SIM-only plans, data upsize plans and MVNOs at Will SIM-Only Plans Cannibalise Regular Telco Plans?Impact of Data Upsize Plans on Telcos and Will MVNOs Cannibalise Telcos' Business? respectively.

Thus, it looks like even mobile services is facing declining cashflows. However, Starhub recently did something that no other telcos did -- it silently raised the prices of its mobile services plans. It ditched the original regular plans which go by the names of 4G 3/ 4/ 5/ 6/ 12 (representing GB of data) and replaced them with plans which go by the names of XS/ S/ M/ L/ XL. These new plans come with unlimited data during weekends but also higher monthly fees. Generally, the fees are increased  across the board by $5.10 per month. As at end Dec 2017, Starhub has 1.37M post-paid subscribers. Assuming the higher fees do not drive them away, Starhub can expect to collect $83.8M (1.37M x $5.10 x 12 months) more in revenue and profits annually.

However, this $83.8M increase in revenue and profits will only materialise 2 years later, when all the existing 2-year telco contracts have expired and migrated to the new plans. For 2018, assuming the contract expiry dates are uniformly distributed, the expected increase in revenue and profits is about 25% of $83.8M or $21.0M.

Needless to say, the major caveat is Starhub's existing subscribers do not desert it in droves, considering that Singtel has mostly maintained prices (it raised price on its Combo 3 plan but dropped price on Combo 6 plan) while M1 has introduced big data plans, including an unlimited data plan at $118 per month. In the long run, this is likely to lead to subscribers switching to other telcos as they question the need to pay an extra $5.10 per month in exchange for unlimited data during weekends. But in the short run, there will be increased cashflow from the increase in prices.


Overall, we will see declining cashflows from Pay TV and wired broadband but increasing cashflows from mobile services. The Enterprise Fixed Network segment will continue to soak up cashflows through acquisitions. Thus, free cashflow will continue to decline in 2018. In 2017, free cashflow (based on Cashflow from Operations minus Cashflow from Investing) is only $190.1M. Dividing by 1,729.1M shares, free cashflow per share is only 11 cents. So, my conclusion is Starhub's dividend of 16 cents per share is not sustainable.

P.S. I am vested in M1, Netlink Trust and Singtel.

See related blog posts:

Sunday 18 March 2018

Will SIM-Only Plans Cannibalise Regular Telco Plans?

It has been a year since I last wrote about telcos. Looking back at what I wrote, it has been gratifying to see that I was right about M1's revenue/ profits bottoming out sometime in 2H2017 and Starhub facing challenges in its Pay TV and broadband businesses. See Challenging Times Ahead for Starhub's Dividends for more info.

Even so, there have been hits and misses in my analysis. One of the misses is the Impact of SIM-Only Plans on Telcos. There, I wrote that "for each subscriber who chooses to switch (from regular telco plans), SIM-only plans will result in a significant drop in revenue. The impact to profit is also negative but smaller than the decline in revenue. However, the no. of subscribers who choose to switch to SIM-only plans is likely to be small."

A year later, I have a better understanding of SIM-only plans. My analysis that SIM-only plans will result in a drop in revenue and profitability is still correct, as shown by the decline in Average Revenue Per User (ARPU) since 3Q2015, when M1 launched the first SIM-only plan. Given that existing subscribers who wish to downgrade from the regular telco plans to SIM-only plans have to wait until their contracts expire, and the typical contract period is 2 years, the effects of SIM-only plans will only disappear 2-3 years later, which is around now. As shown in Fig. 1 below, the year-on-year (YOY) decline in ARPU for M1 has moderated starting from 3Q2017.

Fig. 1: Changes in M1's Revenue & ARPU
The part that I got wrong is the popularity of SIM-only plans. It appears that SIM-only plans are quite popular, as shown by the rise in the no. of M1's post-paid customers, which has been growing by at least 3% YOY since 3Q2015 (note: the no. of post-paid customers reported by M1 also includes those of Circles.Life, the Mobile Virtual Network Operator that leases network capacity from M1. Thus, part of the rise in customer numbers is due to Circles.Life).

Fig. 2: Changes in M1's Post-Paid Customers

Thus, when you put the 2 parts together -- (1) SIM-only plans will result in a significant drop in revenue and more moderate drop in profits for each subscriber who chooses to switch from regular telco plans, and (2) SIM-only plans are quite popular, the inevitable outcome is that M1 will see declining revenue and profitability for 2-3 years from 3Q2015 till now. This is evident from the YOY decline in revenue from 3Q2015 to 2Q2017 as shown in Fig. 1 above. In other words, SIM-only plans will cannibalise the regular telco plans that are more profitable. 

Yet, although SIM-only plans will cannibalise the regular telco plans, they also create demand of their own. Given the low-cost nature of SIM-only plans, it will attract new customers who are not on the regular telco plans. In past earnings conference calls, M1 had mentioned that SIM-only plans are value accretive at the EBITDA level, because they attract new customers and eliminate the need for mobile phone subsidies. They also mentioned that approximately 30% of new customers are on the SIM-only plans.

Perhaps the best way of thinking about regular telco plans and SIM-only plans is to borrow an analogy from the airline industry: full-service airlines and budget airlines. Regular telco plans are like full-service airlines; you get a mobile phone, telco services (voice, SMS & data) and maybe a few Value-Added Services (VAS). Whereas SIM-only plans are like budget airlines; you get only telco services. If you need a mobile phone or some VAS, you need to top-up cash. Although budget airlines do take away customers from full-service airlines, they also create demand of their own. Likewise, SIM-only plans will create demand of their own. 

Nowadays, it is not uncommon to have airlines operating full-service brands and budget-service brands under one roof. Likewise, all 3 local telcos have regular and SIM-only plans. Investing Wolf recently compared the regular and SIM-only plans of the 3 telcos (see link). One of them is actually not very competitive in the SIM-only plans, perhaps worried about the cannibalisation of its regular telco plans! It is as good as not having SIM-only plans. We all know what happens to airlines that insist on flying full-service only.

P.S. I am vested in M1, Netlink Trust and Singtel.

See related blog posts:

Sunday 11 March 2018

How To Select a Housing Loan Package

After selecting our desired condominium, the next step is to choose a housing loan package. There is a variety of loan packages, such as fixed or floating interest rates. For fixed interest rate packages, you could choose whether to fix the interest rates for 1, 2 or 3 years, after which the loan reverts to floating interest rates. For floating interest rate packages, you could select whether to peg the interest rates to the Singapore Interbank Offered Rate (SIBOR), Swap Offer Rate (SOR) or fixed deposit rates. Assuming that you choose a SIBOR or SOR package, you have to further decide whether to peg to the 1-month or 3-month SIBOR/ SOR. Likewise, for fixed deposit rate packages, you can choose between 9-month, 15-month or 36-month fixed deposit rates. The choices can be fairly overwhelming.

Step 1: Fixed or Floating

The first step to decide is whether to go for a fixed or floating interest rate package. Given that US Federal Reserve (Fed) has increased interest rates 5 times since Dec 2015 and plans to increase them another 4 times this year, we decided it is prudent to choose a fixed interest rate package instead of a floating interest rate package.

Step 2: Fix for How Many Years

This is a tricky question. Thankfully, there are some hints. Besides setting the current interest rates, the US Fed governors also provide their projections of future interest rates. By plotting the interest rate projections, we can see how US interest rates are likely to move in the coming years. Fig. 1 below shows the current Fed dot plot, which indicates that the median interest rate projections will rise from 1.375% in 2017 to 2.125% in 2018, 2.6875% in 2019 before finally peaking at 3.0625% in 2020. Having said that, do note that these are just projections by individual US Fed governors. The actual interest rates may differ from the projections depending on how strong the economy is in the coming years.

Fig. 1: US Fed Dot Plot

Given the rise in interest rates over the next 3 years, it makes sense to fix the interest rates for 3 years. However, on the other hand, a 3-year fixed interest rate package is naturally more expensive than a 2-year package, since banks bear the risks of interest rates rising rapidly. Using the bank that we chose as an example, the interest rates for a 2-year and 3-year package are as follows:

2-Year 3-Year
Year 1 1.48% 1.68%
Year 2 1.48% 1.68%
Year 3 15M FD+1.43% 1.68%
Afterwards 15M FD+1.43% 15M FD+1.55%

As shown above, not only is the fixed portion of the interest rates higher, the margin for the floating portion of the interest rates is also higher (note: not all loan packages are as such). For a $650,000 loan over a 25-year tenure, we will end up paying $11,867 more in total interest if we were to select the 3-year loan package. That is equivalent to an extra interest of 1.83% on the $650,000 loan. For the 2-year loan package to be more expensive than the 3-year loan package, the 15-month Fixed Deposit (15M FD) rate, which is currently 0.25%, has to reach close to 2.0%.

We decided on the 2-year loan package. In the event interest rates continue to rise, we can choose to re-finance and fix the interest rates when the lock-in period expires. If conditions permit at that time, we might also choose to pay down some of the loan.

Step 3: Peg to Which Base Interest Rate

As discussed above, we selected a loan package that is pegged to the fixed deposit rates. Fixed deposit rates are actually board rates set by individual banks and are therefore less transparent compared to SIBOR and SOR, which are set collectively by a group of banks. The conventional wisdom is that if banks were to raise their fixed deposit rates to earn more interest on the loans, they would also have to pay more interest on the fixed deposits. Hence, banks are less likely to raise fixed deposit interest rates. However, the reality is that 98% of local banks' deposits have maturity of less than 1 year. Raising the interest rates on fixed deposits of more than 1 year maturity will not hurt them. See Behind Fixed Deposit Home Loan Rates for more info.

On the other hand, although SIBOR and SOR are more transparent, they are more volatile compared to fixed deposit rates. The shorter the SIBOR/ SOR tenure (i.e. 1-month vs 3-month SIBOR/ SOR), the more volatile the rates are. Also, between SIBOR and SOR, SOR is affected by the USD/SGD exchange rate and therefore fluctuates more than SIBOR. See Why Singapore Interest Rates Might Rise Faster than Expected for more info.

Comparing between the transparency of SIBOR/ SOR loan packages and the stability of fixed deposit loan packages, we went for stability as they provide greater visibility on the amount we have to pay every month, which helps us in planning other expenses.


There is a large variety of housing loan packages. It can get overwhelming at times, especially since you have to decide on a loan package quickly after you sign the option to purchase the property. Choosing the right loan package can save you some money and offer greater visibility in later years.

See related blog posts:

Sunday 4 March 2018

How Much is Proximity to a MRT Station Worth?

In my last 2 blog posts, I discussed the price differential of different ages of condominiums, size of units, as well as how long is the leasehold of the condos. See Areas Where We Saved for Our House Purchase and Could We Afford a Freehold Property? for more info. In both posts, I mentioned my preference for proximity to a MRT station, as there are usually more amenities such as shopping centre, market, hawker centre, etc. Staying close to a MRT station also means that we save on a car, which could potentially cost $200K or more in future expenses. However, I also note that the closer a condo is to a MRT station, the more expensive it is. So how much more expensive is it?

Let us compare 2 condos that are still being built. One is within walking distance of a MRT station while the other one is far away from it. Using the same convention as the last 2 blog posts, the one that is near to a MRT station is Condo C (the same Condo C in Areas Where We Saved for Our House Purchase), while the one that is far away is known as Condo D (Condo D is in the same area as Condos E & F in Could We Afford a Freehold Property?) Both are 99-year leaseholds. Both will be completed in 2020. So, the main difference between Condos C and D is the proximity to a MRT station.

The average transaction price of the condo units for the 2 condos from Jan 2017 to Jan 2018, by no. of rooms and floor area respectively, are shown below.

Rooms Condo C Condo D

2 $1,027 $845
2+Study $1,091 -
3 $1,226 $1,138
4 $1,532 $1,385
5 $1,717 -

Area Condo C Condo D

601-700 - $845
701-800 $1,027 -
801-900 $1,091 -
901-1000 $1,226 $1,138
1001-1100 - -
1101-1200 $1,532 $1,385
1201-1300 - -
1301-1400 $1,717 -

Do note the limitations for the data above, which are described in greater details in Areas Where We Saved for Our House Purchase. However, because both are newly launched condos, there are a lot more transactions for them. Hence, their prices are more reflective of the current investor sentiments. On average, Condo C has 12.2 transactions per month while Condo D has 63.9 transactions per month.

From the data above, it is natural that Condo C, which is close to a MRT station, is more expensive than Condo D. For a 2-room unit, Condo C is more expensive than Condo D by $182K, but that is also partly because Condo C has a larger floor area. The average floor area of a 2-room unit in Condo C is 771 sq ft, while that in Condo D is 660 sq ft. The price difference for a 3-room unit and 4-room unit is $88K and $147K respectively.

Using a 3-room unit as an example, would I be willing to pay $88K more for proximity to a MRT station? Definitely! In terms of finance, we save on a car (or 2 cars to last us until we retire). Each car can easily cost $100K, not to mention the running costs. So while we pay $88K more for the convenience of a MRT station, we save at least $200K in future expenses from not having to own a car. The net saving is at least $112K for staying close to a MRT station. In terms of travelling time, assuming we do not own a car, we save at least 15 minutes travelling from the MRT station to Condo D as compared to Condo C. Multiply that by 2 for the journey to-and-fro, and multiply the resultant figure by the no. of journeys per year, it adds up to very significant time savings. In terms of convenience, there are usually a lot more amenities near to a MRT station. If you are hungry, you could go to some nearby coffeeshops which are open 24 hours a day. If you forget to buy something, you could walk to a 7-eleven store to get it. In contrast, if you stay in an area with no MRT station, even getting something can mean a 20-minute bus ride to the nearest town!

In conclusion, while a condo within walking distance of a MRT station is more expensive, the higher price will usually pay off by itself in terms of finance, travelling time and convenience.