Yesterday Bangkok Post reported that Thailand’s largest mobile operator, Advanced Information Services, AIS, had agreed to participate in the re-auction of a single license of 10 MHz bandwidth in the 900 MHz spectrum. This is the license that JAS Mobile won in the auction last December and then forfeited by not making the first payment in March. The article says that AIS’ auction participation was “in exchange for an extension of the right to retain [AIS’] existing 400,000-strong 2G customer base”. This quid pro quo intrigued me, so let’s draft a business case together to understand the money.
The price AIS will pay is roughly ฿75Billion. Yesterday I wrote about a typical metric for valuing radio spectrum, cost per MHz-POP, or the price paid divided by the bandwidth in MegaHertz divided by the number of potential subscribers covered by the spectrum. At ฿56/MHz-POP, this spectrum is expensive by any standard. But for a country like Thailand with an ARPU of less than ฿220, this is a ludicrous price, one that will be difficult to bear with the constraints imposed by regulation and competition. A more typical price for Thailand, using global benchmarks and earlier Thai auctions, might be in the ฿15-20 per MHZ-POP range, about one third as much, or a total price of ฿25 Billion. In other words, at a ฿75 Billion price, AIS is over-paying by roughly ฿50 Billion. And by the way, True Move has already committed to that amount for their 10 MHz license, so they overpaid by฿50 Billion, too. So to wrap up this whole paragraph in a single sentence, AIS is paying an extra 50 Billion for the privilege of retaining their 400,000 2G subscribers.
So let’s assess the benefit of that. What are those 400,000 2G subscribers worth to AIS? As with any business case, let’s make some assumptions:
- Let’s begin with 400,000 subscribers. Let’s assume that every month, some of the subscriber go away, through churn, 3G/4G upgrade, or whatever means. Let’s assume that this “erosion” is 2% every month.
- On aggregate, Thai Mobile ARPU is around ฿220. But this group of subscribers is 100% 2G , which means it is more likely their ARPU is on the low end of average. So let’s assume ARPU of 100.
- Lastly, to calculate Net Present Value, let’s assume the cost of money is 6% annually.
- The timeframe for this business case will be 3 years, 36 months. Many arguments in both directions. Unit OpEx for the 2G network having only 400,000 subs will be high compared to the 3G and 4G networks, motivating AIS to shut it down as quickly as possible. On the other hand this may be a low-end cohort that simply cannot afford 3G/4G devices and data plans. AIS could continue to operate this network on a “best effort” basis, not replacing defective components, not upgrading, avoiding maintenance costs, to serve this cohort as cheaply as possible while retaining their loyalty. So on balance, let’s stick with 36 months.
- Let’s also assume that this whole business case is all guesswork and therefore the margin of error could be large. But we will also assess how much that matters.
Plugging these assumptions into a worksheet, the NPV is roughly ฿1 Billion over 3 years. So this back-of-the-envelope business case says the 2G subs are worth roughly ฿1 Billion. So why is AIS willing to pay ฿50 Billion for them? Somebody check my math, that doesn’t seem right.
I admit to ignoring many factors in this business case, factors that would both make the 2G subscribers more valuable and factors that make them less so. Factors such as
- the value of keeping these subscribers away from TRUE and DTAC;
- the additional revenue AIS will generate from this 4G spectrum (revenue which AIS were planning to collect even without the spectrum);
- eliminating the 2G network OpEx costs;
- the 4G OpEx costs of saved by a lower density 900MHz 4G network (compared to an 1800MHz or 2100MHz one).
But no amount of math, nor fiddling with the assumptions, will come anywhere close to closing that gap between the 2G benefits to AIS and the cost of the 4G spectrum.
Do you have any suggestions how to explain this disparity?