Bloomberg reports that Greece’s four systemic banks (Alpha Bank, Piraeus Bank, National Bank of Greece and Eurobank) have passed the ECB stress test and all lenders have kept a common equity Tier 1 capital ratio above the legal minimum of 4.5% in the adverse scenario.
Piraeus Bank CET1 stood at 5.9% in adverse scenario, 14.52% in baseline scenario
National Bank of Greece CET1 stood at 6.92% in adverse scenario, 15.99% in baseline scenario
Eurobank Ergasias CET1 stood at 6.75% in adverse scenario, 16.56% in baseline scenario
Alpha Bank CET1 stood at 9.69% in adverse scenario, 20.37% in baseline scenario
Greek lenders are up around 10% since our recommendation in December, and should have much more room to run, now that the forced capital raise is no longer a risk.
Greece has been in the news for several years in a rather negative context: bailouts, austerity, “grexit” and debt haircuts are too often associated with the Greek saga. But each downturn eventually ends and gives way to a recovery. We are indeed at the stage of the cycle, where various important reforms have been put in place, the economy is showing signs of life while banks have been generously recapitalized. The dust is slowly settling and after the depression-like collapse in growth, there are signs of stabilization and even green shoots.
After three bailouts and a near 30% drop in nominal GDP, Greece is a poorer country but many of the underlying problems are recognized and are being addressed. The collapsing economy created a hole in the balance sheets of Greek banks, as non-performing loans (NPLs) rose to 35% of outstanding loans, wiping out equity capital in the banking system. However banks have undergone three separate recapitalizations (equity injections) and now – even accounting for non performing assets – have a solid capital buffer. They are in all likelihood over-capitalized in an environment of improving economy, and favorable bad debt dynamics. Crucially, the legal and judicial framework for dealing with NPLs, mortgages and collateral has been revised, allowing banks to recover significant amounts and reduce NPLs in line with ECB/SSM targets.
After the tremendous pain of the recent years, the stage is set for gradual improvement, which judging by economic data has just begun. The biggest worry might be the unsustainable government debt, however there is a silver lining here too. Only a relatively small share of the debt is owed to the private sector. The majority is owed to the troika (European Commission, European Central Bank and the IMF) and debt relief of that portion can reasonably be expected. Such debt relief would have no negative implications for Greek banks, on the contrary it would ensure sovereign debt sustainability. Greek bank equities are cheap and are discounting a dire future. It’s time to buy as the risk-return ratio is favorable.
1. A short history of the Greek Sovereign Crisis
The Greek debt crises is almost entirely due to the country’s government finances. The private sector hasn’t overborrowed, and banks were collateral damage – not the main culprits – of the sovereign’s problems. The origins of the crisis go back to the late 90’s when Greece was preparing to join the European Monetary Union (EMU). The EMU required joining member countries to fulfill certain debt and fiscal criteria, ie to maintain government debt to GDP below 60% and budget deficits <3% in a given year (the original 1992 Maastricht treaty rules). Unfortunately the Greek government at the time decided to blatantly lie, and through a series of financial transactions concealed the true level of debt and budget deficits. Surprisingly, the cheat was not discovered until much later and the country joined the European Union (EU) and the Eurozone in January 2001.
These lies went largely unnoticed until after the Global Financial Crises (GFC) when risk premia exploded and the cost of borrowing for peripheral European issuers such as Greece rose considerably. In late 2009, a newly-elected government announced that the deficit for that year was estimated to be 12.7% instead of 6%. The European Commission was outraged, but it was just the beginning as it really took a few more years to clean up the Greek statistics agency and establish a true picture of the sovereign’s liabilities (including debt issued by state-owned companies). In the light of these revelations the price of Greek bonds fell sharply (yields and spreads vs Germany exploded), and soon put at risk the government’s refinancing (or rolling over) of its debt.
As the Greek government was not able to borrow new funds, in March 2010 the first bailout package of €110bn was agreed with the troika, in exchange for a demanding austerity program and deficit reduction in Greece. Tax evasion, uncovering of further state liabilities as well as collapse of growth due to austerity have made it quickly clear that a another bailout would be required. In February 2012 the second bailout for €130bn was agreed between the troika and the Greek government. At this stage, a haircut on Greek bonds was negotiated (this is when Greece effectively defaulted). Certain privately held bonds were swapped for much longer maturities, effectively reducing the face value of these bonds by 50%, while the real value (due to maturity extension) was reduced by three quarters. However the official sector (in particular bonds held by the ECB and IMF or 44% of the total debt) were excluded from this exchange. As a result the impact on the overall debt load, was much more limited than the 50% reduction in face value would imply. Furthermore, the Greek banking system, which participated in the bond swap, took large losses on their holdings, and needed recapitalization to the tune of €22bn. As a result, a nominal haircut of 50% on €200bn of bonds, only effectively reduced the outstanding debt by €85bn, or 23.9% of the outstanding debt.  Not quite enough as it turned out.
One reason why a further third bailout was necessary, were the regrettable events of 2015. In January Syriza won 149 out of 300 seats in the parliament on an anti-austerity program, heavily criticizing previous bailouts and the troika. Investors were becoming increasingly worried about a upcoming confrontation between the government of Alexis Tsipras and the troika, and deposit outflows accelerated in the first half of 2015 after a relative stable period in 2013-2014. This further increased the dependency of the banking system on financing from the European Central Bank (ECB). The government was still running deficits and was going to struggle to refinance upcoming debt maturities in this environment. After a lot of truly Greek drama in June-July and a referendum, the third Greek bailout of €86bn was finally backed by the Greek parliament on August 14th 2015. It was another severe austerity package and required a number of reforms: streamlining VAT, safeguarding independence of the statistical agency, automatic public spending cuts, reform the labor market, privatize state assets and (yet again, for the third time) recapitalize the banks. Alexis Tsiparas ad Syriza are still in power today and have mostly been overseeing the implementation of this bailout. While the troika continues to oversee the reform program, the Greek government has largely aligned itself with the European agenda, as can be seen in a recent quote by the prime minister on state television: “We inherited unemployment at 27 percent, we will hand it over at 17 percent. We inherited a country that was on the edge of a cliff, we will deliver a country that is at the core of Europe… and an economy that will be growing, with social justice, which we did not even see when the economy was growing” (ERT-3 on the week of 11-Dec-2017). Note the reference to Greece being at the “core of Europe”. Admittedly parts of the ruling Syriza party are not wholeheartedly supporting some of the reforms, but as far as this author is aware no mainstream party advocates a break with Europe anymore.
2. Greek bank fundamentals
NB: A lot of our analysis will focus on the four largest and systemically important Greek banks (Alpha Bank, Piraeus Bank, Eurobank Ergasias and National Bank of Greece), the “Greek4”. These four institutions represent two-thirds of the assets of the banking system (€260bn vs €391bn) and are all listed on the Athens Stock Exchange (with some also listed in the US). We will occasionally also use data reported by the Bank of Greece for the entire banking system or Monetary Financial Institutions (MFIs). For further information on these stocks, you may follow Jan Willem Barnhoorn on Seeking Alpha, who provides updates and great commentary.
The banking system got hit by the crisis in two ways: directly, by the one-off losses from haircuts on sovereign bonds (the blip in 2012 on the chart above is the result of the “haircut” or Greek debt restructuring), and more importantly, indirectly by an explosion in non-performing loans due to the collapse of the economy as GDP cratered, while fiscal expenditure was cut. NPLs account for around 36% of outstanding loans as of Q2 2017.
Non-performing Loans are the elephant in the room
There is no doubt, that it is the huge pile of NPLs that will represent the biggest challenge to the Greek banks in the coming years. This problem does have the attention it deserves and progress is being made in this space. Since 2016 the ECB via its so called Single Supervisory Mechanism (SSM) has been working with the Bank of Greece and the Hellenic Bank Association on a plan to reduce non-performing loans and non-performing exposure – NPE being a slightly broader measure of impaired loans. A detailed NPL reduction trajectory has been agreed, and appears to be broadly on track as can be seen in the table below.
As of September 2017 the Greek banks were on track with NPEs falling to €99.1bn (vs €99.9bn target), while NPLs fell to €70.2bn. And these are still early days, as the targets are back-loaded, with the biggest reductions to occur in 2018 and 2019 bringing down the NPL ratio to 21.1% from the current 35.1%. There are three main ways for banks to reduce their bad loans: liquidations, sale of loans and write-offs (each contributing about €10bn reduction over the next 2-3 years according to the ECB/SSM targets).
There has been encouraging newsflow on this front at a micro level. In October Eurobank sold a pool of €1.5bn unsecured consumer loans to Sweden’s Intrum. Admittedly the loans were sold at only 3% of face value (+fees) but unsecured consumer loans are the most troubled category (residential and corporate fairing relatively better). During recent Q3 result conference calls, Alpha Bank and Piraeus Bank both mentioned accelerated NPL reductions in the near term, while Eurobank assured investors about their ability to meet targets set with supervising authorities. Furthermore, over the past 12 months a number of important legal changes have been made while more are planned, to improve non-performing loan recovery. These legislative efforts show that the government (admittedly with some pressure from the creditors) is making a significant effort to tackle the NPL problem. Some of these are difficult politically (repossessing properties to pay the banks does not get many votes in Greece apparently), but considerable progress has been made regardless.
The progress on the legal front is so important, because the old framework was particularly inefficient at recovering bad debts. As a result many borrowers in Greece decided to default on their debts even though they would able to continue servicing them. This group has been branded “strategic defaulters”. A Bank of Greece study found that as many as one in six companies were not servicing their debts despite having the ability or selectively defaulting, hoping to reach a settlement or having their debts forgiven. This certainly raises hope that recovery rates on at least part of the outstanding NPLs will be high.
Legislative changes to unclog the NPL pipeline
There are many examples of how the pre-2016 legal framework was inadequate and made recovering bad loans slow and costly. The General Bankruptcy Code allowed the liquidation process to last for up to ten years, which is obviously hugely inefficient for the creditors and the whole system. Over-indebted individuals could game the system by almost continuously reapplying for debt relief leaving limited time for creditors to enforce their rights, between applications. Repossessions of properties have been extremely limited in recent years. The legal framework was inefficient in the best of times and when the tsunami of NPLs hit the banks in 2010-2015, it was in clear need of overhaul. A number of changes, recommended by Greece’s creditors and the Hellenic Financial Stability Fund (HFSF), have already been implemented and several more are forthcoming. Many more steps are taken to improve the operating environment for lenders – let us highlight four main initiatives below.
Out-of-Court Debt Workout Out-of-Court Debt Workout has gone into effect on the 3rd of August 2017 and will allow voluntary debt restructurings. 300’000 companies, who have outstanding bank loans or arrears to the state, are eligible. The framework allows viable companies who have a hard time meeting their financial obligations to settle their debts based on their real financial situation and after conducting a full assessment of their assets. 7’550 initiated the process as of end of October.
Code of Civil Procedure amendments The Code of Civil Procedure has been amended in 2015 to streamline procedures before civil courts as shown above. Further changes planned for early 2018 will allow banks to claim 100% of owed loans from bankrupt businesses (as opposed to 65% currently) after property taxes and six months of backpay to employees. A clear step to boosting the recoveries (and value) of NPLs.
E-auctions: Since the crisis began repossessions have been kept to a minimum – mostly by politicians, but more recently due to disruptions of public auctions. Seeing little risk and urgency of foreclosure, strategic defaulters who had the capacity to continue paying back their loan, decided to suspend payments. A non-payment culture fostered. This dynamic contributed to the very high level of NPLs today as the backlog hasn’t been cleared in years. The good news is that, change is already happening. Legislative changes in Q2 2017 lay the ground for banks to sell property off via online electronic auctions, a step that would finally start clearing the backlog. After delays of several months, the website was launched and first e-auctions have taken place in early December and it looks like they will accelerate in 2018, with the target being about 1’000 properties/month in mid-2018.
Auctions could also trigger another positive side-effect for banks, and the banking system in general by encouraging strategic defaulters to come forward and start meeting their loan repayments to prevent their property being sold at well below its market value.
Politically, the government appears to be walking a fine line though, as Alexis Tsipras originally campaigned using the slogan “no homes in the hands of bankers”. However as long as the Greek government is committed to working with its European partners, it may have little choice but to follow their recommendations. Properties of low-income families are protected from repossessions and being auctioned off, if the owed amount is below €200k or the property is valued at less than €280k.
Credit Bureau: An independent agency will be created in early 2018. It will track tax, social security and bank loan arrears for individuals and companies. It will surely be helpful in identifying people who are “gaming” the system and surveying credit quality of borrowers in general.
These reforms go a long way in fixing the inadequate legal framework and address the “non-payment culture”. It should allow banks to effectively deal with the huge NPL overhang and make the NPL/NPE reduction targets agreed with the ECB achievable.
The balance sheets of the banks are in rude health. As can be seen below, the equity capital has been replenished (after a long series of write-offs and losses) and today at over €33bn is higher than the pre-crisis peak (while assets are somewhat lower than previous peaks). Commo Equity Tier 1 ratio (CET1) is at 17.1%, while the Capital Adequacy Ratio (CAR) is at 17.2% (both higher than year ago levels).
Liquidity has similarly been improving. For several years, and particularly during and after 2015, the banking system became dependent on emergency liquidity assistance (ELA) from the ECB/Bank of Greece, which is an expensive type of funding. Use of ELA funding has been declining. Eurobank, National Bank of Greece and Piraeus Bank all raised funds through covered bonds over the past three months. Finally deposits have been rising modestly over the course of the year. Confidence is building, the crisis is over – this is slowly being reflected in the cost of credit.
Banks do not typically have a “revenue” line. Instead, the top line in the income statement is called Net Interest Income or NIM. It represents the spread between interest received (on loans and other assets) and interests paid (or cost of funds). The NIM of Greek banks is a healthy 2.6%, whereas many European banks only have a NIM of about half that amount.
Let us briefly consider bank accounting for expected losses on lending. Provisions are a charge against anticipated losses from lending in the income statement of the bank. They are taken on a quarterly basis. Provisions have moderated over the last few quarters after a final spike after the referendum and panic in 2015. Given the improved economic prospects (discussed below), provisions are likely to remain contained. As net interest income is consistently strong, low provisions will directly feed through to higher profits and over time build up higher levels of (book) equity.
Accumulated provisions become (loan loss) reserves (on the balance sheet). In recent years, significant reserves have been set aside by the banks to cover losses from NPLs. Indeed, the NPL coverage ratio (reserves divided by NPLs) amount to around 60% and have been on a rising trend.
One would be correct to point out that 40% of the NPL book is not covered by reserves. However the other missing element in this analysis is loan collateral (and ultimately NPL recoveries). The vast majority of loans given by banks has collateral, and as explained above, this collateral will be increasingly recoverable thanks to the legislative changes. According to calculations by Jan Barnhoorn, the coverage ratio including collateral values for the Greek4 banks was between 111% and 125% as of December 2016. Thus, the NPL stock is currently well covered by reserves and collateral. Going forward, reserves may not have to rise any further, if, as we anticipate, NPLs stop rising. This will be helped by an economic recovery.
How much upside is there?
The share prices of Greek banks are down over 90% (in some cases 99%) from the highs in 2007. It has lead some investors to assume that in a recovery scenario the upside might be in the 10-100 times range. That is in all likelihood wrong. The market value of the equity of the Greek4 banks has fallen from €50bn to €10bn, while their market value as a share of GDP has fallen from 25% to 4%. Thus a more realistic upside target in a full recovery scenario would be 6 times or 600% – an exciting prospect. The difference between the drop in the share prices of over 90% and the drop in market value of 80% is of course accounted for by the discounted rights issues and three capital increases at Greek lenders. There is upside in the Greek bank trade, but it’s in the 3x-6x range (not 10x-100x – it is not bitcoin after all).
Another approach would be to consider the multiple of market value (or price) to tangible book value (per share) for banks in Greece. This measure is currently very depressed (below 0.4x) and could potentially rise towards the Eurostoxx Bank average (which is itself very depressed historically) of almost 1.0x in a recovery scenario. That would suggest expected returns of around 3x for Greek lenders.
3. The Greek Economy
The Greek economy is by no means in a boom, and neither are we forecasting one. The positive view on bank equities is predicated on economic stabilization with modest growth going forward. Bank of Greece expects growth of 2.4%-2.5% in 2018 and 2019 – it might not seem like much, but it would in fact be the highest rate of growth since early 2008. For almost ten years companies and consumers have been extremely thrifty and cautious. The return of confidence will make a huge difference.
Apart from being a financial crisis, the Greek saga has been a real social tragedy for the Greeks. Thus, the most important economic indicator for the Greek people and politicians must be the unemployment rate. Although still high, it is down over seven percentage points from the peak, which translates to almost 400’000 less unemployed.
The number of people quitting their jobs – usually in order to take up a better job – was frequently cited as Janet Yellen’s favorite indicator of market tightness (the famous “quits rate”) – there appears to be good news in Greece as the number of voluntary quits has risen.
Greece has not benefited from a currency devaluation (as it remains in the Eurozone), and there was no real export boom. Although imports did collapse during the crisis, bringing the current account into balance.
Tourism has contributed to balancing the current account and has been one important source of growth in recent years as tourist arrivals are set to breach 23 million this year. Greece is an attractive destination with its 6’000 islands, thousands of kilometers of coastline, Mediterranean climate and good food. In addition, by most accounts it remains an affordable destination. As per a recent BoG release, travel receipts in January-October 2017 rose 14.2% year-on-year.
After years of living thriftily, there is probably some pent-up demand for cars. Before 2008 an average of 40 thousand cars would get registered on average per month, and this measure fell to below 15 thousand between 2011 and 2017. Surely the age of the average vehicle has risen and the replacement cycle will kick in, once confidence returns.
In terms of housing prices, we can only speak of stabilization after a fall of 40% peak to trough, however construction permits are showing timid signs of life.
Further “green shoots” in the construction sector can be seen in demand for cement.
It’s early days, but construction activity should progressively normalize as banks become healthier. Construction activity is running at very low levels and the trend is higher.
Finally, leading economic indicators are encouraging. The Purchasing Managers’ Index recently came out at levels rarely seen since 2008. The leading New Orders component is also on the way up.
To summarize the economic situation we will borrow a paragraph from a recent Bank of Greece report.
The positive course of the economy is reflected not only in GDP figures, but also in several key indicators of economic activity, such as industrial production, retail sales, private sector employment flows, exports of goods and services, and foreign direct investment, as well as soft data such as the manufacturing PMI and the economic sentiment indicator. Improvements are also visible in the financial sector: bank deposits of the non-financial private sector have increased, the decline in bank credit to non-financial corporations has slowed, and banks’ dependence on central bank financing has gradually decreased.
After going through hell over the last eight years, Greece is finally turning the corner and Greek banks, which are cheap and over-capitalized will benefit from the recovery. The legal framework is being put in place to deal with the NPL overhang, while better growth will also be favorable for NPL formations and collateral values. The top four lenders in Greece will be reporting healthy profits, as provisions remain contained. Equity values will continue to grow and before the end of the decade some will start paying dividends again – of course, at that stage share prices will be much higher.
A word of caution
To be sure, many problems remain in Greece: the government still has too much debt, demographics and long-term growth prospects are uninspiring, non-payment culture is alive and well, while unions have too much power. Greek politics will remain volatile and although odds favor a recovery bank stocks, there is never such a thing as certainty. This trade is for seasoned investors who are able to take a loss should one occur – widows and orphans stay away.
The main risk of this trade is related to Greek politics. Although no major political party advocates a break with Europe, that could change. Recently, forces to the left of Syriza organized campaigns and demonstrations during repossessions of properties and in front of courthouses during auctions. A split in the ruling Syriza can also not be ruled out. But as long as Greece remains in the Euro and owes considerable sums to the troika, it will have to follow the path of reforms. It would be a huge waste to abandon it now, that so much has been achieved.
M1 released Q3 2017 results on the 16th of October and I am now ready to close a long-standing negative position on the sector. Singtel remains the top long-term pick.
M1 Results Takeaways
This is not to say that the results were excellent, but at this stage the negatives are largely recognized and increasingly priced in, while signs of stabilization or dare I say improvement are becoming visible. Let’s briefly review the results (all growth numbers expressed as change vs previous year’s 3rd quarter):
number of mobile subscribers broadly stable (+0.7%) with postpaid rising 3.2% and prepaid falling 3.4%
voice minutes down -10% (postpaid) and -26% (prepaid)
acquisition cost per customer S$378 vs S$357 (although in-line with previous years)
ARPUs generally negative
Data usage up massively to 4.2GB/month from 3.4GB a year ago
As a result, mobile data contributed 55.9% of revenues, up 1.7% points
Mobile Telecommunications revenue +3.4% to S$160.5m
Fibre customers up 30k to 182k, while ARPU declines modestly
EBITDA broadly stable (+1.3%)
Operating cashflow is broadly stable ex-working capital changes
Free Cashflow generation was stable ex-working capital changes and ex-spectrum rights payments
Net Debt/EBITDA remained stable at at 1.3x (up 0.1x)
As noted earlier, dividends have been cut to 11.1 cents in 2017 (from 15.3 in 2016 and 21.2 in 2014) and further steep cuts should not be required.
Clearly, data usage is exploding and further substituting traditional voice calls – a trend likely to continue. 33% of data plans were exceeded – ie customers went over limit and had to pay extra. There will surely be further pressure on the market in the coming years as TPG enters the game, however for now 1) data usage is growing and will help the whole sector 2) dividends and expectations have been reduced sufficiently and 3) current results are stable to positive. A dividend yield of 5% or more (6.1% on trailing 12-month dividends paid) is likely to be maintained from the current entry price (S$1.80). I am thus closing the long-held short position on M1 and Starhub.
Singtel remains top long-term pick
Singtel remains the top-pick as it is best placed to defend its turf in Singapore via triple and quadruple plays – and important advantage in an environment of growing data volumes – as well as high market share at 52.2%. The positive view on Singtel is, however, not driven by a view on the domestic market, which represents less than 10% of its value, but rather its associate companies. Bharti Airtel (about 25% of Singtel’s value) is the clear winner from India’s mobile sector consolidation. It has recently announced the acquisition of Tata’s consumer mobile service in a favorable transaction. The share price of Bharti has responded appropriately, rising over 30%. Globe Telecom is performing well, gaining market share in the Philippines, while Telkomsel has reported strongest YoY growth numbers among operators in Indonesia. Data volumes at AIS in Thailand have been exploding. Even Optus in Australia is experiencing a period of recovery after network upgrades have been completed.
Singtel has a 4.7% dividend yield, which is safe, an excellent diversified portfolio and great management. It makes it the top long-term pick in the sector – risk-on!
This is a short update of share price performance, following my piece from January calling for continued weakness in M1 and Starhub due to the entry of a fourth mobile operator (TPG).
As can be seen on the screenshot, total return performance (including paid out dividends) on both M1 and Starhub has been weak (at -9% and -5.5%), clearly underperforming Singtel (~flat) and the broader market (+11%).
Both telcos have largely recognized the weakening operational environment by cutting dividends. M1 has paid out 11.1 cents in 2017 (vs 15.3 in 2016 and 21.2 in 2014), while Starhub has lowered the quarterly pay out by 20% (from 5 cents to 4 cents). These dividend reductions leave yields at 6% and 6.2% respectively.
These operators face continued pricing pressure, while TPG hasn’t started marketing yet. In the meanwhile, unlimited data plans have are making inroads into the Singapore mobile market, with great benefit to the consumers and the broader Smart Nation objectives, while at the expense of shareholders.
The Singapore telecom regulator has opened the market to a fourth operator (TPG), thereby increasing pressure on revenues and profits of the three incumbents. We have explained how this may force M1 and Starhub to cut dividends in a post in January. In this piece we would like to present the main takeaways from a research piece published by Credit Suisse on the second of March. Thanks to Chee Tiong Lim for kindly sharing the piece. The main takeaways are presented below.
Mobile pricing to fall by 15-25% over the next 3 years
Pricing for mobile data and calls will drop significantly as TPG rolls out its service. This is great news for consumers who will pay lower bills, but it will be painful for the incumbents – in particular M1 and Starhub who have significant exposure to the Singapore mobile market. According to the forecast, overall revenues from mobile in Singapore will trend lower for the next 6 years (-7%), in addition, these revenues will now be divided among 4 operators resulting in mobile (revenues at M1 and Starhub falling by 16% and 13% respectively.
Key TPG Forecasts
The company will spend S$ 500m in CAPEX by 2021.
It will start the mobile operation in 2H 2018 with 775 sites (base stations) and reach 1’000 sites by 2021
The company has a good mix of low-band and high-band spectrum (2x10Mhz in 900Mhz and 2x40Mhz in 2300Mhz). This spectrum could support a market share up to 28%.
TPG is expected to have 150 employees by 2021 (vs 1’500 at M1 today) – a lean operation indeed
TPG will enter the broadband market (possibly before launching the mobile service) in order to compete more effectively in mobile.
Credit Suisse doesn’t expect TPG to become massively profitable over the forecast 2017-2022 period (although the break-even crucially depends on how much market share the company manages to take and how far ARPU will fall). Indeed, the profitability of the whole sector will fall due to new competition brought about by the entrance of TPG. Surely the incumbents (Singtel, Starhub and M1) will be cutting costs as revenues decline. The real winners will be the consumers who will get more data for less – the ultimate goal of the regulator.
Conclusion: Remain underperform on M1 and StarHub
The inevitable conclusion is to remain underweight the sector and in particular M1 and Starhub. Singtel is in a different situation as Singapore revenues represent a small portion of the company’s enterprise value – the bulk consists of its interests in Telekomsel (Indonesia), Bharti (India), AIS (Thailand), Intouch (Thailand), Globe (Philippines) as well as Optus (Australia). Shareprices of M1 and Starhub are down substantially over the past 2 years, however as per out analysis dividends are not sustainable – especially in this deteriorating environment.
As can be seen in the table below, valuation of M1 and Starhub (8x-9x EV/EBITDA) still appears to be on the high side within the sector. Given that earnings will be under pressure in the coming years, there will be further pressure on the share prices of these companies. As an aside, the company this analyst likes most from this list is SmarTone (315 HK) – but that’s an idea for another post!
On December 14th The Infocomm Media Development Authority (IMDA) announced that TPG Telecom won the “New Entrant Spectrum Auction” with a bid of S$105 million and will become the fourth mobile operator in Singapore. TPG has acquired spectrum in both the 900MHz and 2.3GHz band and the company will have 18 months from the start of the spectrum rights to provide street level coverage and an additional year to provide network is coverage in buildings. For close to 20 years mobile voice-and-data was a three-player market in Singapore, however the sector will soon be dividing the pie between four players – the three incumbents and the new entrant. Although the speed at which the new entrant takes market shares is a subject of debate, there is no doubt that it will put pressure on pricing and revenues in the sector – a trend that was already visible in recent quarters.
IDMA is in charge of spectrum licenses in the city-state (before 2016 the predecessor organization – IDA – held this responsibility). It’s mission is not strictly regulatory, indeed it is to foster the development of the information & communication sectors, considering interests of end-users and nurturing a healthy environment for innovation. The organization has first announced that a new player would enter the market in July 2015. It underscored that previous episodes of more competition in the mobile communication sector (in France & Spain) have shown more competitive price plans and infrastructure upgrades by incumbents. It also pointed out that mobile traffic is expected to grow significantly in the coming years providing enough growth for the sector even with a new competitor. This last statement is true in data terms (gigabytes of data transferred), but as we will see is less certain in terms of dollar value o revenues. In any case, IDMA would like to see lower prices. Cheap access to the internet is one of the requirements to succeed in some of the top-level objectives of the authorities such as the Smart Nation national project – better living and advancement through technology, info-comm technologies and big data.
TPG Telecom paid S$105 million – three times the auction reserve price – for the right to become Singapore’s fourth mobile operator. In a press release TPG said it expects to invest a further S$200-300 million to build the network. Earlier estimates for the roll out of the fourth mobile network were higher and it is likely that the total bill will be higher. However TPG should be able to reduce the number of towers by optimizing its locations using inputs such as current data traffic at various hotspot locations, data quality and building floor plans across the island-nation.
TPG Telecom is 34%-owned by David Teoh. Originally from Malaysia, he founded TPG’s predecessor – a desktop computer distributor – and transformed it into Australia’s second-largest broadband provider. With this achievement the entrepreneur and investor has gained tremendous respect in the industry and also made it into the billionaires’ club. TPG Telecom has a market capitalization of $5.8 billion AUD, with another $1.4bn of debt and generated an underlying EBITDA (earnings before interest depreciation and tax) of $775m online AUD in the year ending July-2016. In this context it is clear, that the commitment to invest S$400m+ in the network roll out in Singapore will not be a stretch for the company. TPG commands a 27% market share in broadband access in Australia with 1.87m subscribers and has 475k mobile subscribers in a MVNO (mobile virtual network operator) network operated through Vodafone infrastructure. Thus, the new network in Singapore, will be the company’s first mobile network in which it owns and controls the infrastructure. According to the press release, operations will become profitable once a market share of 5-6% is reached.
Analysts at broker CLSA have made a simple model for TPG’s new operation for the first 10 years. They assume an annual rate of growth of 2% in the total number of subscribers from the current 8.6m (penetration is already at almost 150%) to 10.5m in 2028. TPG’s share would reach 10% by 2028 on these assumptions and revenues would amount to over S$350m. The model uses a constant ARPU (average revenue per user) of S$30 (per user per month) – ie a mobile bill of S$30 for the average user. The EBITDA margin is expected to start at 35% in 2018 and rise to 45% over the next two years and remain there. This analyst would disagree with these rosy assumptions in the early years and would expect higher customer acquisition costs. For example the third mobile operator in Singapore – M1 – calculates the cost of acquiring a postpaid customer at S$363. At this rate acquiring 300k subscribers would cost over 100m SGD – a huge amount even compared to the initial CAPEX spend on network roll out. It is not clear that the offered packages will be so much more competitive than the incumbents’ at the time, allowing to attract tens of thousands of subscribers easily. The market will be adjusting the pricing of packages in the in expectation of the launch of TPG’s offer over the next 18 months and that is precisely why the existing players in the Singapore mobile market will remain under pressure.
Singapore Mobile Telecom market
Almost since the outset in the late 90’s the telecom market has had three players: Singtel, Starhub and M1 with current market shares of 50%, 27% and 23% respectively. The penetration rate is high at 148.8%, which means that there is more than one connected device per person. The market is fully deregulated since 2001 and is dominated by postpaid subscribers. Over two-thirds of contracts have a defined duration of 1-2 years and come bundled with a handset. This allows operators to lock subscribers in and prevents them from switching during the duration of the contract. What makes these deals appealing to consumers is the discount (or effectively subsidy) that the operators give on handsets in these packages. Instead of paying S$1000 for an iPhone 6, the consumer will pay S$200 and a share of his bill over the subsequent 24 months (S$25-30) will effectively go towards paying down the handset. This is still the most common way of subscribing to a mobile service, it is however under pressure since the appearance of SIM-only (or Line-only) contracts first introduced by M1 in July 2015 and quickly taken up by others. SIM-only offered for the first time a cost-effective alternative for users who did not wish to purchase a handset.
In addition to pressure on operator revenues from SIM-only packages, share of traditional services such as voice, SMS and roaming have been on the decline as users prefer to use messaging platforms such as Line and Whatsapp, which use the data connection. These “legacy” revenues have been declining for some time and were offset by subscriber growth and other business lines, but in essence the operators have been running only to stand still.
Company valuations and Stock Implications
As can be seen below, the shares prices of M1 and Starhub have had rather lousy two years (total returns in SGD assuming reinvested dividends are shown). Singtel stock has not suffered, but then the vast majority of Singtel’s revenues comes from abroad – Optus (Australia), Bharti Airtel (India), AIS (Thailand). Indeed only 12% of its revenues come from the Singaporean consumer business – pure mobile revenues (ex-Pay TV and ex-broadband) represent and even lower percentage. However M1 and Starhub – companies whose businesses are entirely reliant on Singapore are down 23% and 39% since the beginning of 2015.
The apparent reason for the poor performance is the expectation of more competition from the new entrant. However the underlying business has been showing signs of weakness at both Starhub and M1 before the IDMA announced the new licence would be issued. Over the period 2013-2015 EBITDA was flat (M1) to down (Starhub) while revenues were trending lower – weakness more clearly visible in recent quarters. The drop in legacy revenues (SMS, voice and international calls) and increase in the share of SIM-only plans and declinging ARPUs are all putting downward pressure on revenues.
However the most important and overlooked reason behind the sell-off is the realisation that dividends will need to be cut. This is important because many investors bought Singapore telco shares for the dividend in a yield-starved environment. The proposition of a high dividend yield combined with a household name appeared very attractive. Yet during 2013-2015 both companies did not cover their dividends with cashflows. During this period M1 generated S$125m FCF (free cashflow or operating cashflow after capital expenditure), while it paid out $S190m in dividends per year (using 3-year averages in both cases), thus creating a total cash shortfall of close to S$200m. Indeed M1’s net debt increased S$150m during this period, suggesting that about a third of the dividend was debt-financed. Similarly Starhub generated on average S$280m SGD of FCF in 2013-15, while it paid out S$345 in dividends thereby financing about 20% of the dividends paid through new debt. The company’s net debt has increased by just under S$100m thus confirming the above observation. Both companies would have needed (and maybe initially planned for) a strong growth environment in order to eventually cover dividends from internal funds, but have been in a flat-growth environment at best while pressure on revenues and margins is now intensifying.
What is priced in?
Using last 12 months’ trailing dividends M1 and Starhub yield 7.7% and 7.0%, which is on the face of it a juicy yield, but these dividends will not be maintained. The timing of the dividend cut, was not easy to predict as leverage at both companies is under control (ND/EBITDA at 1.0x (M1) and 0.7x (Starhub) as of dec-2015) and the shenanigans could go on for a while longer. However the sell-off makes it clear that the market recognizes the insufficient cashflow generation and the managements will have an incentive to put a floor in and reset the dividend at a lower level.
How much lower do payouts need to go? At the very least payouts can not exceed free cashflow, but in reality given a deteriorating environment, I would expect them to fall comfortably (ie 25%) below recent FCF figures in order to ensure reasonable coverage and leave some leeway. That would take dividends at M1 and Starhub 50% and 40% lower respectively and using current share prices (of S$2.0 and S$2.85) would imply dividend yields of 3.8% and 4.2%. Given that Singtel with its excellent assets and geographic diversification currently yields 4.7 (and is not at risk of reducing their payout), I would view M1 and Starhub as relatively expensive. For dividend yields to become comparable or slightly superior to Singtel (say 5%) share prices would have to fall by a further 20-25% for M1 and 15-20% for Starhub. Dividend yields are not necessarily the holy grail of valuation, but the objective here is simply to demonstrate that the current price still doesn’t discount the new normal in terms of profitability and cashflow generation, and the dividend cut announcement effect is still ahead of us. At those levels (20% lower for M1 at $S1.6 and 15% lower for Starhub at S$2.40) the share prices wouldn’t be discounting the worst case scenario with a terminal decline in revenues but would be pricing in the upcoming dividend cut and some further revenue/margin weakness in the coming 3-4 years.
TPG Telecom appears to be an aggressive and well-financed new entrant in a sector which has already been experiencing revenue pressures and in terms of dividend payout was “living above its means”. We are consuming ever more data and in its 2015 new mobile entrant announcement the IDA acknowledged as much, expecting 40%-60% traffic growth in the coming years. The incumbents will in all likelihood see traffic growth, but that will not translate to rising revenues with falling prices for each gigabyte of data transferred. The operators (in particular M1 and Starhub) will have to adjust to declining revenues by containing costs, but dividend reductions are unlikely to be avoided.