Millicom International Cellular SA (TIGO) 2007 Q3 法說會逐字稿

Good afternoon, ladies and gentlemen, and welcome to today's Millicom 2007 third quarter results.

(OPERATOR INSTRUCTIONS)

At this time I would like to turn the call over to your hosts today, Marc Beuls, CEO, and David Sach, CFO. Please go ahead.

Thank you, Operator, and welcome to everyone who has joined us today. For this call both David and I will be using slides to run you through the results. It will be helpful to have the slides in front of you and you can find them on our home page at www.millicom.com. We will both be happy to answer any questions you have at the end, but first I would like to give you an overview of the results and run through the performance of each cluster.

So let's start with slide number two. Quarter three has been an excellent quarter for Millicom as we have seen broad based growth across all of our regions. During the quarter we substantially increased our rate of investment as we saw opportunities in a number of our markets to accelerate our build out and capture market share. Africa and Colombia have taken the lion's share of this investment as we want to build Tigo's market share quickly in these fast growing markets.

It is worth commenting, especially on Africa as there were a number of one-offs since quarter two, which resulted in slightly lower growth in these markets. But in quarter three our African operations regained their former vitality. Africa as a region grew by 52% and underlying this growth was an acceleration in subscriber acquisitions, growing at a strong 17% quarter on quarter.

So turning to the headline number on slide 2 of the pack, revenue for the third quarter grew 77% over the third quarter of 2006 and EBITDA grew by 60%. We added 2 million new subscribers in Q3, giving us a total of 20 million subscribers at the end of September 30, 2007. Even allowing for the acquisition of Colombia, underlying organic revenues grew by 46%, which is a very strong performance.

As I have already mentioned, Millicom's CapEx increased substantially during the quarter and was just under $350 million, which represents a 109% increase over a year ago and a 67% increase over the last quarter, and brings the CapEx for the year to date to $738 million. As you can see, CapEx is increasing significantly throughout the year and consequently we have increased our CapEx forecast for the full year from over $800 million to over $1 billion. We're expecting CapEx for 2008 to be at the similar level. This CapEx is a good indicator as to how positively we view the prospects for growth.

On slide three the results for accelerating investments can be seen clearly in terms of the rate of subscriber acquisition during the quarter. Millicom added 2 million total subscribers in Q3, as you can see on slide number three, bringing the total to 20 million at the end of September and representing a 77% increase over the third quarter of 2006. This result was considerably higher than our previous run rate in the first half of the year as we managed to invest faster than we had expected. If we are able to maintain our level of CapEx spending in Q4 then subscriber intake should be similarly strong.

In Q3 subscribers in Central America grew by 74% year on year. In South America, subscribers grew by 42% excluding the Colombian acquisition and by 170% including Colombia. In Colombia, Millicom added 211,000 subscribers in the quarter, showing that we are beginning to get traction in this market. Africa reported year-on-year subscriber growth of 44%, but even more presently a growth rate of 17% from the last quarter as we caught up from what had been a slower second quarter in terms of subscribers. Asia increased by 43%. Attributable cellular subscribers increased by 79% from the third quarter of 2006 to just under 17 million at the end of September 2007.

Millicom has one of the strictest definitions in the industry for what constitutes subscribers. Reported subscribers are those that have generated revenues within a 60-day period or, in the case of new subscribers, those that have started to generate revenues. We have introduced this tighter definition for new subscribers to count only those customers who have initiated a revenue generating activity so that we report only genuine subscribers to Tigo products and services.

We will stop reporting TDMA and CDMA customers in all Latin American operations except for Bolivia at the end of 2007. The phasing out of our legacy TDMA and CDMA networks is continuing and at the end of the third quarter less than 700,000 subscribers, 400,000 of which are in Bolivia, remained on these networks. During Q3 we took out 300,000 from our numbers and this process has resulted in a slightly higher than average churn but will be finished by the end of Q4 in Central America and Paraguay and probably by the end of 2008 in Bolivia.

While the higher ARPU customers have mostly already migrated to our GSM networks, it is likely that a number of the remaining TDMA and CDMA customers might not migrate. However, their loss from the networks will have little impact on revenues as they are very low ARPU customers with less than $1 per month. Phasing out these old TDMA and CDMA networks is important as it will release spectrum in the 850 and 1900 bands for 3G services, which we are planning to launch in our existing spectrum, plans based on current licenses in Latin America during 2008 and 2009.

Slide four shows that for Q3 we have recorded quarterly revenue of $686 million, an increase of 77% from the third quarter of 2006. And excluding Colombia, the underlying organic revenue growth was a healthy 46%. Organic revenue growth was once again highest in South America at 53% excluding Colombia, demonstrating the success of per-second billing and value-added services, which today constitute over 25% of revenues in Paraguay. In Central America, where the average penetration of our markets is 46%, revenue growth at 45% in Q3 is very strong. The third quarter revenues for Africa were also very strong, increasing by 52% year on year, regaining their momentum after a slower second quarter. The prospects in Africa continue to look good. For Asia the growth rate was 30% with strong performance in both Sri Lanka and Laos.

Referring to slide five, EBITDA for the three months ended September 30, 2007 was $296 million, a 60% increase from the third quarter of 2006 and showing an increase of $33 million or 13% from Q2 2007. EBITDA margin of 43%, which is very encouraging, particularly in this period of aggressive growth that we are currently experiencing.

So let's look at the results for each cluster, starting with Central America which today accounts for 44% of the group's revenue and 54% of the group EBITDA. On slide six you can see that the total subscribers reached 7.4 million, up 74% year on year and 10% over Q2. Nearly 700,000 new subscribers were added in the quarter, showing the effect of per-second billing on the cluster. Per-second billing was introduced across the region on February 7th and these results show the attraction of the Tigo offering today.

ARPU in Central America was stable at $20, which is above the Latin American average of around $17 reported by resurgent markets deserving a comment. When we reduced prices by some 25% in February 2007 to the moved per-second billing, it was logical that ARPU would fall. But due to the good price elasticity, ARPU fell by far less than the price cap. Since then we have seen the ARPU stabilize. We believe our higher than average ARPU is due to the fact that Tigo has attracted the higher proportion of the best customers in our market and because we maintain the most conservative definition of subscribers, understating our numbers compared to our competitors. This increases ARPU.

The third quarter revenue for Central America grew by 45% from Q3 2006 and by 11% from the previous quarter to $300 million. EBITDA for Central America also increased by 45% to $161 million for the third quarter of 2007 and the EBITDA margin was strong at 54% due to our market-leading position, which means that we have a larger share of the more profitable on-net traffic. We've been able to sustain over 50% margin for several years and expect to be able to do so going forward.

Please turn to slide seven to look at the results for South America in more detail. Subscribers grew by 170% to 5.3 million including Colombia, which accounted for 2.5 million subscribers in the third quarter. In total, looking at the quarterly year on year basis, revenue grew by 245% to $215 million and EBITDA grew by 181% to $80 million. However, looking at the underlying growth excluding Colombia, the numbers were still strong with subscribers up 42%, revenue up by 53% and EBITDA up by 76%, showing clearly how we can achieve inverse growth rates by focusing on value-added services, which means we can achieve faster EBITDA growth than revenue growth, which exceeds subscriber growth. Some day we hope to see this dynamic in all of our markets.

The EBITDA margin for Q3 was 37%, affected by our efforts to grow our business in Colombia more aggressively. EBITDA margin in Colombia for the quarter was stable at 25%, despite substantial quarter-on-quarter revenue growth and as a consequence of our concerted efforts over recent months to build up our distribution networks. As you may recall, we were pleased with the results of a recent Nielsen survey that put us just fractionally below the market leader in terms of number of points of sale with inventory across the country.

Official figures released by the regulators for Q3 show that Tigo subscribers grew by 11% quarter on quarter versus 3.2% for the market as a whole. (inaudible) method of accounting subscribers Tigo added 211,000 net new subscribers in Colombia during the quarter, indicating that the operation is gathering momentum and we look forward to showcasing our Colombian operations to those of you who are joining us for our capital market trip next week.

Turning to slide number eight, the strong top-line performance and subscriber growth of our African cluster has been the most encouraging aspect of the third quarter results. Over 660,000 subscribers were added during the quarter, bringing the total at the end of September to 4.6 million, and revenue for the quarter increased by 52% year on year to $122 million. There was a cost to this growth in terms of margin, but this cost that we are willing to incur in the short term.

Despite the heavy cost of building out the new and extended networks in Africa, particularly in [VSE] in Chad where there is a lack of transportation and power infrastructure, we still reported a quarterly year-on-year increase in EBITDA of 8% to $34 million, resulting in an EBITDA margin of 28%. We took a very deliberate decision in the third quarter to chase growth rather than margin in the short term, but we expect that we can continue to invest in the future and will see margins return to the levels that we achieved historically as we achieve critical scale.

Slide nine shows that revenue for Asia was $50 million for the third quarter, up 30% from Q3 2006, and EBITDA was $21 million, up 41%, producing an EBITDA margin of 43%. Subscribers in the Asian cluster increased by 43% from Q3 2006 to 2.6 million at the end of Q3 2007. The stronger results for Asia reflect the substantial investments that have been made in Asia since 2006, prior to the launch of per-second billing in Cambodia and the launches of Tigo in Sri Lanka and Laos in Q1 this year. We've been very pleased with the development in Sri Lanka where, following substantial expenditure to the network, we are growing revenues in excess of 50% year on year while maintaining margins at over 50%. Laos continues to perform well and Cambodia still has much room for growth.

Now I would like to hand over to David to talk you through the financials.

Thank you, Marc. Please turn to slide ten where you will see the key financial ratios for the first nine months of 2007 compared to the first nine months of 2006. We continue to invest significantly in the Tigo brand in all our markets, which has resulted in higher sales and marketing costs as a percentage of revenues. We believe that these investments are helping drive the strong subscriber growth.

The G&A costs are also up, which again reflects the higher cost in new markets such as DRC and Colombia where there are significant rollouts in progress. Pleasingly though, G&A costs as a percentage of sales, excluding Colombia, have not risen. Cost of sales have fallen slightly as a percentage of revenues, reflecting the economies of scale that we are achieving as our market shares rise.

It is vital for Millicom's AAA model that we continue to keep a tight rein on costs, which enables us to reduce prices to our customers. Lower prices enable us to penetrate deeper into our markets and target lower ARPU customers. Growth will be driven by these lower ARPU subscribers going forward. In addition, we continue to expand our value-added services, particularly in Latin America. These services tend to have little incremental cost, which also positively impacts our gross margin.

Slide 11 shows that EBITDA margins have been stable at close to 43% in the past year, having fallen in Q4 of 2006, reflecting the expenditure of the rollout of Tigo in Colombia, DRC and Chad. Excluding Colombia, the EBITDA margin for the first nine months of 2007 was 47%, the same as last year. Over time we remain confident that we will see our margins trend back up to the high 40s as soon as we achieve critical mass and take further market share in these three markets.

On slide 12 you will see an overview of the profit and loss. The areas that we wish to bring to your attention are the increase in depreciation and amortization, reflecting the much higher level of CapEx, the decrease in the tax rate and the gains from discontinued operations which have a positive impact of $258 million due to the successful sale of PacTel in Q1.

Please turn to slide 13. The slide shows a strong growth in CapEx in Q3 to $347 million as against $208 million in Q2 and against $166 million in Q3 2006 with rises of 67% and 109% respectively. As Marc mentioned earlier, CapEx is a good indicator as to how we view our future potential. We demand internal rates of return of between 20% to 25% and paybacks of one to three years, which means we only invest where we see good growth opportunities.

The capacity expansion, which is where most of our Latin American CapEx is now spent, we expect to see returns within one to two years. We have needed to add substantially to our networks as per-second billing and market growth have driven the volume of minutes across all of Latin America. For network coverage, which is where the majority of our African and Asian CapEx is spent, we are targeting returns of two to three years.

Coverage CapEx is today more weighted to the cost of building sites, including towers, buildings and fences, rather than the cost of imported equipment. In recent periods, the cost of steel and concrete have been rising whereas the prices of network equipment have been falling. As such, the rates of return and paybacks for coverage CapEx are lower than capacity expansion. Overall, the current blended rates of return are consistent with our historical averages.

It is fair to say that we were surprised in Q3 at the rate in which we have been able to invest. Africa presents some particularly difficult challenges in building out the networks with inadequate roads and infrastructure, limited electricity, uncertain regulatory environments and a lack of highly qualified experienced personnel.

It is encouraging that we progressed well this quarter with our expansion but stress that it is difficult to accurately predict the timing of our CapEx spending going forward. The higher level of CapEx in Q3 should help sustain subscriber growth in Q4, which in turn will help sustain revenue growth in 2008. Due to the difficulties in accurately predicting the timing of CapEx spending, we remain cautious in our forecasting and predict full-year CapEx at just over $1 billion. However we do expect that the CapEx-to-sales ratio will continue to trend downwards.

Slide 14 shows that depreciation has risen considerably from the level of roughly $50 million a quarter in the first nine months of 2006 to roughly $85 million a quarter on average during the past year. With CapEx rising substantially in quarter three and forecast to stay above $250 million on average over the next five quarters, we can expect to see depreciation continuing to rise for several years. Depreciation will converge to the average CapEx spending over the past six- to seven-year period. As a guide, depreciation as a percentage of revenues has stayed fairly constant and we expect this to remain the case, even when investment in 3G technologies begins in 2008.

On slide 15 you will see the debt has stayed steady from quarter two to quarter three with debt at the end of quarter three of $1.6 billion. We have continued to increase debt in the operating companies but took the opportunity to repurchase $45 million of the senior notes due in 2013. Moving debt from corporate to the operating companies is tax beneficial to us. Today, Millicom is in an under-levered position with net debt to EBITDA of 0.5 to 1 and we recognize the need to increase gearing over time.

Our target level is a ratio of 2-to-1 in terms of net debt to EBITDA and we are looking at investment opportunities that could justify an increase in debt to give us a more efficient balance sheet and the ability to utilize our cash in the best interest of shareholders. Our priorities to maximize shareholder value are; firstly, to grow our existing businesses; secondly, to buy out minority partners where possible; thirdly to enter new markets; and finally, to return capital to shareholders. Given that the buyout of minority partners and external acquisitions are opportunistic and not fully in our control, we have obtained approval at the last AGM to buy back shares up to 5% of the share capital.

Please turn to slide 16. You will be able to see that interest expense has stabilized in the last four quarters and the simple effective rate has been constant at approximately 10%. We are still looking at ways of reducing our interest cost and, importantly, we have been exploring how we can borrow more at the local level in our subsidiaries where the interest payments on local debt are tax deductible against local profits.

We will repurchase additional senior notes if the opportunity arises, because the interest on these notes is not fully tax deductible since there is insufficient income at the corporate holding company. As mentioned in previous calls, the notes are redeemable in December 2008 so we'll very likely redeem any outstanding notes then. Repaying these tax inefficient notes with local operating company debt improves our overall group tax rate.

Our overall group tax position is summarized on slide 17 where you can see that our operations effective tax rate has risen owing to the operating losses in Colombia and DRC. We expect this trend to reverse in future as these operations become profitable. You can also see that we have been able to reduce the difference between the operations effective tax rate and the overall group rate by reducing the nondeductible net corporate expenses. We have done this by increasing the management and brand fees that we charge our operating companies and very recently by reducing our corporate interest by repurchasing some of the senior notes. Reducing this difference lowers our overall group effective tax rate.

In this quarter we benefited from some one-off items that help reduce our overall group tax rate to below 30%. Bolivia has traditionally paid taxes on revenues, but now that it's becoming profitable it will likely start paying taxes on profits. As such, we have recorded a deferred tax asset in anticipation of this change. We anticipate that the quarter four tax rate will be slightly higher and that the full-year rate will be closer to 30%. Going forward it should be possible to keep our tax rate at or below 30% with a more balanced geographical mix of profits, further repayment of our corporate 10% notes and our ability to transfer more income to corporate to fully offset the remaining non-tax deductible costs.

Please turn to slide 18 which shows that, today, Millicom can finance its increased level of investment from internal cash generation. The closing balance of $1.1 billion in cash means that we have net debt of $584 million and are well positioned to continue with our accelerated investment program and to explore the other options that we have already discussed.

Slide 19 illustrates how the investment breaks down across our businesses. CapEx was $738 million in the nine months of 2007 and is trending towards a full-year CapEx figure of over $1 billion.

Thank you. Now I will hand you back to Marc.

Thanks, David. To summarize the past quarter and looking forward, I can say that Q3 has been an excellent quarter and we expect 2007 to be another record year. In Q3 we have managed to increase the quarterly rate of investment and this has led to higher subscriber numbers than we had anticipated. We have raised our CapEx forecast to $1 billion for 2007 and expect a similar level of CapEx in 2008.

In Central America, we have slightly improved our market position despite the competitive environment that we operate within and, with stable ARPUs of $20 and EBITDA margins of 54%, we are performing strongly. The introduction of per-second billing has improved the affordability of our services, leading to a high subscriber intake. Across Latin America value-added services will become increasingly important and this is why we are planning gradually to introduce 3G alongside our 2.5G services within our current frequencies and licenses during 2008 and 2009. We are closing down our TDMA and CDMA networks to ensure that we have enough spectrum available for 3G.

South America has performed strongly and the underlying revenue growth, ex-Colombia, was an encouraging 53%. Our operation in Colombia is gaining traction and we are excited about showing many of you the operations during our IR visit next week, especially how we are building the foundations for our future growth on the back of much improved accessibility of the Tigo prepaid services.

In Africa, along with all operators, we are facing challenges owing to the lack of basic infrastructure, which at times slows development and leads to higher OpEx in the near term. Today Tigo has an aggressive expansion in Africa and, as you have seen from our numbers in Q3, we are making good progress. We continue to see great opportunities in Africa as the mobile penetration is low and the growth prospects are exciting. Consistent implementation of our AAA strategy will allow us to take advantage of those opportunities and we expect that Africa will become our fastest-growing region.

In Asia, we have improved our service offering by introducing per-second billing in Cambodia and launching Tigo in Sri Lanka and Laos. Growth rates in that region have improved, particularly in Sri Lanka, and we expect to see Asia performing in line with the two other regions.

That concludes my comments and we will now be happy to take your questions. So Operator, can I please have the first question?

Thank you. (OPERATOR INSTRUCTIONS) David Kestenbaum from Morgan Joseph is on the phone with a question.

Okay, thanks. Nice quarter guys.

Thank you, David.

Can you give us more color on the CapEx, exactly where you're going to be spending the incremental $200 million? Is that all going to be in Africa? Or what's the split between Africa, say, and Central America or South America?

I think today Africa and South America, specifically Colombia, are the regions where we are spending the biggest amounts of CapEx, and that's how I can see this going forward. Let's not forget that it's in Africa that we see the lowest mobile penetration rates and that's where the need for additional infrastructure is the highest.

Okay. And then ARPU, I have calculating it went up a lot in South America on a sequential basis. Can you talk about the reasons for that? Was that based on your success with the value-added services?

Definitely. As you know, Paraguay is our champion when it comes to value-added services with 25% of their recurring revenues coming from those services. But we also have seen continued positive impact of the per-second which we introduced about two years ago in Paraguay. And then, of course, with Colombia, where we're improving our position in the market as we are growing faster than the market, with all the changes in our tariff packages we have been doing over the last couple of quarters, that gives us better clients and better revenues from those clients. And that explains the high ARPU.

Okay. And then what do you think -- I mean your view on the effect of per-second billing on ARPU has kind of changed over time, where do you think that stands going forward? I mean do you think ARPUs will come back to the point where they were at at one point? Or do you think that they will continue to kind of trend down over time?

Well, you know my view in terms of ARPUs and that has not changed. The ARPUs over time will come down. What we see now in Central America is that we see stable ARPUs. This, I think, is driven by a number of factors, increased usage as a result of the effective price reductions, secondly also having value-added services start becoming a bigger part of revenues in Central America.

And thirdly, I think we're tapping into new markets, or new market segments, rather, in Central America, people who were not able to afford a phone before can now, with per-second billing and the ePIN, afford a phone and buy minutes on a daily basis. I think those are the drivers of ARPUs. But again, the value-added services will become also a much bigger part of revenues in Central America going forward.

Okay, thank you.

Thank you, David.

Adrian Dawes from Hartwell is on the phone with a question.

Yes, terrific quarter, I just can't believe how good it was. I'm delighted, however. Two questions, one short term and one longer term. What are your expectations for getting the margins in Colombia up to the average of the other contiguous countries? Do you think that's farfetched? Do you think it will happen? If so, when?

And secondly the longer-term question is, your last statement about we're under-leveraged, as David pointed out in his discussion of CapEx, and we expect to be able to generate exciting opportunities. Can you talk a little bit about some of those and how soon you think they're going to happen and what the barriers to those happening are or how you feel about that in general? Thank you.

So as far as Colombia is concerned, you know that at this point in time our top priority is growing the top line. And if that were to cost a little bit of the EBITDA level that's fine with us, because we know that we need to grow our market share and we need to get to better economies of scale in order to grow that margin going forward. We think we're going to need a couple of years to get to the, call it, Millicom average EBITDA margin in Colombia.

In terms of the balance sheet structure and what we're going to do with the cash we have and the ability to raise debt is, as we said before, our top priority is to buy out local partners. But at times the difference in price expectations are very high and we only want to buy out minority shareholders if we think it's accretive to our business and to our share prices. So that kind of sets a level as to how much we can pay.

Secondly, in terms of new markets, we are continuously looking at new markets, primarily focusing on African and Latin America where we see new opportunities. New licenses in Central America are just around the corner. And we might be looking at some other new markets in order to repeat the success stories that we have seen so far in Colombia and also, for instance in Africa, to take advantage of the good growth outlook we see in our existing businesses. So that's how we would prefer to spend our money and, as David said, we also have the option now to buy back stock if we think that that is the best way to spend our money.

Marc, in terms of the improvement in the Colombian operations, do you think it'll be a gradual improvement over those two years? Or do you think we'll see a step function? And you didn't include any Asian activities in your discussion of potential acquisitions, new licenses et cetera, can you comment on both those elements please?

Yes, I think it's going to be a gradual improvement in Colombia. As you know, we have a market share of, say, around 10% now and we want to grow that market share to like the mid 20s and that will take time. And as we grow with our very good distribution network there, which we will show to participants to our Capital Markets Day next week in Colombia, we will get there. But it's going to be gradually, I don't think there are going to be any big steps that will be taken, it's a gradual process.

In terms of Asia, yes at this point in time we're not actively looking at new assets in Asia. We don't really see any attractive opportunities at this point in time other than, of course, we continue to follow the events in Vietnam. But I don't think anything will happen there in the course of 2007.

Well, since 2007 is almost over, can you see any visibility at all there? Or is it just more bureaucratic stonewalling.

Well, we don't know, Bill, I think this process is going to go into 2008 and latest rumor is that it could be like by the middle of 2008 before anything is going to happen there, given that a number of privatizations apparently have to happen before privatizations in the telecommunications sector would be started.

Okay thanks, just an absolutely marvelous quarter. Congratulations.

Thank you, Bill.

Thanks.

Anders Wennberg from RAM is on the phone with a question.

Congratulations for a great quarter. That really shows the strength of the Tigo brand and your concept.

Thank you.

Still just to check and be a little bit paranoiac, can you remind us where competition could be increasing? I've heard that MTC Kuwait bought themselves into Ghana recently and I wonder also if you could confirm that and if you have any new competitors coming in places like Guatemala or Honduras or any other important countries.

Yes, that's correct, MTC bought themselves into an existing operator in Ghana, so that is not a new operator, it is not an increase in the number of operators. But that news is fresh so we don't know what is going to happen. What we will be doing is just focus on growth and profitability in Ghana. And as you've seen, the subscriber intake has -- it continues to be very -- or has been very strong again in Africa, so I think that is very, very encouraging.

In terms of other markets, I think Honduras everybody is familiar with that, there is a process ongoing about the fourth operator and I'm sure you've read the names of the companies interested in that fourth license and I think they are the usual suspects. In Guatemala, there are all kinds of rumors as to whether that fourth operator is going to launch service or not. The, the latest rumor I heard was that it was not going to happen. But we don't know.

As a matter of fact, we're not really sitting there and waiting until this happens, in all markets we just continue to take advantage of the growth opportunities and we'll see how we deal with competition. So far we have been dealing very successfully with competition in our 16 markets.

Okay, thanks a lot.

Thanks.

Soomit Datta from New Street Research is on the phone with a question.

Yes, hi guys. Just a quick question on Colombia. First of all, the ARPU number there, if I just divide revenues by average subs, that seems to be ticking up nicely on a quarterly basis. Is there any particular reason for that? And you talk about a $20 ARPU number across Central America, is that an ARPU number you will be targeting for the Colombian business? That's the first question.

Okay, for Colombia, yes, I think it's a combination of first of all improving the quality of our subscribers. We had a number of very low-quality subscribers in the subscriber base when we entered the business a year ago. Through some changes in the tariff packages and some customers leaving us and better customers joining us, we have been able to increase the quality of our subscriber base thanks to our much improved distribution networks and the accessibility there, but also the availability our network is really getting a lot better.

And needless to say, we have taken a number of initiatives in terms of the affordability. That I think is driving that ARPU in Colombia. I can't give you any guidance as to where we think that ARPU is going to end up, but we think that we should be able to bring, going forward, good customers on board in Colombia, as we've done in other markets where we've followed a similar strategy.

And just as a quick follow up on Colombia, can you give us a sense of what percentage of the revenues at the moment are actually handset sales?

It's a fairly small percentage but I don't have the number for you here right here. But it's not an important number.

Okay, fine. And then, sorry, the second question was just on Africa, obviously you're taking a slight hit to OpEx in the short term, can you give us a sense of whether you sort of hope to get back to historic margins of perhaps over 40%, is that going to come in the next couple of quarters? Or are we talking maybe a couple of years away?

No, I don't think it's going to be a couple of years away. I think in Africa we're facing specific issues which we've spoken about before, but at the same time we have fantastic opportunities and we want to take advantage of those opportunities. That means gaining market share, getting more subscribers on our network now that I think we've gotten into a certain momentum. But I do think that our margins will get back to the historical levels and, when that will be we will report to the market. But I remain very positive about Africa both in terms of subscriber intake as well as margin development.

Okay, thanks very much.

Thanks.

Sven Skold from Swedbank is on the phone with a question.

Thank you. Just a follow-up question on Latin America and excluding Colombia then, only the old Latin American businesses. Looking at Q2 you grew much faster this year than last year, or you added more subscribers in Q2 than in Q2 last year, and the same happened now in Q3. Is there anything that hinders a record intake now in the fourth quarter in Latin America, excluding Colombia? And I know that you're going to adjust some subscribers, but --.

No, I think that with all the initiatives we've been taking, per-second billing, ePIN, value-added services, that makes Tigo a very attractive brand in the market. And that is what is driving that increased subscriber growth. And as you know, the fourth quarter historically is always a very strong one, so we'll see. We haven't launched our Christmas campaigns yet, it's too early, it's only October, but typically fourth quarter is very good. So I think we broke ground in Latin America in general by introducing this per-second billing and ePIN and all this focus on value-added services and I think that's paying off now.

So, growth in the market is there, competition is there, but also still quite fair opening up for a record, hopefully, Q4 now?

Well, we'll report that in February.

And just on Congo, do you feel that you have a kind of critical mass now in Congo with more base stations on the ground?

Yes, but you know it's not a critical mass yet if you were to put that into another country, given that Congo is so big. So we need to continue spending money on infrastructure going forward in order to cover more parts of the country, but also start coping with the increased usage of our network we have been seeing over the last couple of months. So I don't think we've reached that real critical mass, I think we want to go a lot higher and that's when we will see the economies of scale and we will see the improvements at the level of the EBITDA margin and stuff like that. Although we have seen some improvements already, but I think there is still much more to do.

But am I wrong? Don't you already have more base stations in Congo than you had ever in Pakistan?

I can't remember how many we had in Pakistan, so it's no longer in our statistics and in our memories. But we are building a big network. I think rarely have we spent that much money in a new market in the beginning. So I think this is unique, yes, we are spending a lot compared to other markets when we enter them.

Okay, good, thanks.

Thanks, Sven.

Drake Johnstone from Davenport is on the phone with a question.

Hey, guys.

Hi, Drake.

This might be coming at it a different way, I know you've been talking about various African countries, but do you expect your margin in Africa to remain near Q3 levels for a couple quarters? Or do you actually start seeing some slight improvement over a few quarters? What's your view there?

I think we see, we expect improvements because in some markets we've seen a positive trend in terms of market share gains, you've seen yourself the subscriber numbers, and that should create better economies. By definition that should lead to higher margins going forward. But again, it's just like in Colombia, then Africa is the same, I think at this point in time our focus is on top line growth. We think there are great opportunities in those markets, and then we want to go for those.

So what I hear you saying is, perhaps near term, if there is a margin expansion, it may be somewhat incremental with -- sometime next year you might start getting more operating leverage as you get the subscriber levels you desire?

This operating leverage you get from month to month, we have been building networks and now we start filling up the capacity in those networks and that will bring down the G&A per subscriber and will, by definition, have a positive impact on EBITDA going forward. So this is not something that is light years away, this is something that happens gradually as we build bigger networks and as we get more economies of scale and we fill our networks up more and more.

Okay. Great. Thanks very much.

Thank you, Drake.

Stephen Mead with Anchor Capital is on the phone with a question.

Yes, Marc?

Hi there.

Hi. Just going back to Central America and the ARPU, I was trying to get a sense of, on the value-added services, what does the subscriber count look like? And in terms of ARPU for the value-added service, what does that ARPU look like? And what are the services that people are paying for at this point?

Well, really, they're paying for everything they use as value-added services, so we don't give the house away for free. And about 60% to 70%, depending on one country to the other, is SMS, peer-to-peer SMS, could be premium SMS but also ring tone downloads, music downloads, all of those things. And we see the revenue coming out of value-added services at this point in time growing at a faster rate than we see voice revenue growing, so it will become a bigger part of our business going forward.

And again, some countries like in Paraguay, I can't remember the exact percentage, but it's a very high percentage of our overall customer base, postpaid and prepaid, that are using value-added services because they're focused on those services very early on. We're now doing the same, that means we are educating, we are teaching people how to use value-added services, so we expect that also in other Latin American markets we're going to see more and more of our subscriber base using those services. But in Paraguay I can't remember the numbers, it's a very high number today already.

In terms of the ARPU for a typical value-added service customer like in Central America, what's that ARPU look like?

The ARPU is, like I said, our top for value-added services market in Paraguay was 25% of recurring revenue. I think in Millicom as an average today, I think we're around 10%. I think Central America is probably slightly higher than that, a few points higher than that I would think, so you can calculate yourself as to what kind of recurring revenues -- sorry, ARPU is coming out of value-added services in Central America.

Okay. And then can you talk a little bit more about sort of what the conditions on the ground are like in Congo relative to what we read about in the Economist and stuff like that?

What I read in the Economist is that they're all convinced that mobile telephony has a long way to go in Africa and then that's why we're there. And, yes, conditions on the ground are challenging, but we've been focusing on the larger cities and, of course, as you build your network, we will be going more into the rural areas where the challenges will be bigger. I think the country is relatively stable at this point in time, except for a certain area in the eastern part of the country around [Kehu] where we have a small network. And overall, we see that more of our traffic now is coming from cities other than Kinshasa. Where in the beginning it was primarily focused on Kinshasa, we now see that traffic in other cities is growing faster than in Kinshasa, which is good, which confirms that we are rolling out our network which remains challenging that roll out.

And just going back to Africa, in general as you look at the countries you're either two or three in those markets in terms of operator, and going forward what's the mix in terms of growth, taking market share versus the incremental growth in those markets? And just can you comment on that just generally?

Yes, focus is on market share, not at any price. I've said historically that we're happy to be a profitable number two so we don't necessarily need to be number one at any price. We know that number two in those African markets can be very attractive, it can give us a lot of traffic and traffic with good profitability. As you've seen in our third quarter results, we're now focusing a little bit more on that top-line growth.

Why? Because I think we're now better prepared infrastructure wise, people wise with some changes we did in the management teams, for instance in Tanzania where we see excellent results being produced. And that makes us focus a little bit more on the top line growth. But yes, we do want to be a number two at least, that's the lowest position we want to have in the African market so that's what we're targeting for.

And then one other question on the share buyback, can you talk about the specifics of the authorization? And where would the cash flow come from for repurchases?

Well, I can't really say much about that, Steve, at this point in time because it's just one of the ways to use our cash. And you know that where that cash is, it's about $1 billion or over $1 billion on our bank accounts, and as you know we have a very low leverage at this point in time. But I can't really say more about that.

So you don't have an actual authorization, you talked about the 5%.

Well, we have an authorization from the shareholders, they gave us an authorization in the month of May at our AGM, and we can buy up to 5% of our outstanding shares provided we have distributable reserves, which we have today.

And have you bought back any shares?

No, otherwise we would have informed the market.

Okay. Thanks, Marc.

Thanks, Steve.

Kevin Roe from Roe Equity Research is on the phone with a question.

Thank you. Terrific quarter, gentlemen.

Thank you, Kevin.

Thank you.

A couple questions. You've spent a lot of time on margins, you've told us you expect Central American margins to be at 50% or better, so sustainable there, Africa returning to historical levels, upside in margins in Colombia. You didn't mention anything about Asia, the 43% margin that we saw in the quarter, is there upside there, is that level sustainable?

I think that's a level that is sustainable, although I don't see it growing because of, for instance, the revenue share agreement we have with the government in Colombia and also because of some new taxes that were recently introduced in Sri Lanka. So I don't really see any growth in that EBITDA margin.

You mean Cambodia not Colombia, right?

Sorry, Cambodia.

That's helpful. Switching to new license opportunities, can you give us any update on the new license opportunity in Panama?

As I said before, we are following that very closely, we have a team of people working on that out of Central America and we will participate. And I'm sure you guys have seen our name on the list of interested parties so we will be going into that process, the same as we intend to do so for Costa Rica and Nicaragua and maybe other markets in South America and in Africa.

And lastly, Colombia and Congo, the two big engines of growth for the near term and beyond, we've seen nice sequential net add growth the last several quarters and particularly in the third quarter both countries ramped nicely. Do you expect through into the fourth quarter and to 2008 that the absolute number of net adds will continue to grow from these levels? Is there incremental growth upside?

I think in Colombia and in Congo, I don't think we have gone all the way yet, I think there is still upside there. But as far as Congo is concerned, it will be driven by our network buildout so we need to continue investing, which as you know is challenging. And if we continue to do so, I would expect that we're going to continue to see very, very strong subscriber intake.

The same in Colombia, I think there I think it's a matter of growing that network and also growing our points of distribution and maybe, who knows, taking the one or the other initiative to improve the affordability of our services, combined with, of course, also value-added services. A percentage of value-added services of recurring revenues in Colombia is amongst the lowest within Millicom at this point in time, despite Colombia being the country with probably the highest GDP per capita. So these are the things we will be focusing on, which will make Tigo a very aspirational brand and hopefully that's going to attract more customers.

Terrific. See you all in Bogota.

See you there.

Okay, see you there.

Peter Nielsen from Cheuvreux is on the phone with a question.

Thank you, Peter Kurt Nielsen from Cheuvreux. Just two questions at this stage please. Firstly, you obviously significantly beat your own guidance for (inaudible) in the quarter, given the level of CapEx you're investing at the moment, is the current level something which you would anticipate in your future quarters as well? And secondly just to clarify, did I understand, David, correct that the tax rate going forward would be around 30%? Thank you.

Yes, we've now reached the 2 million subscriber mark. We expect that, of course, the fourth quarter is going to be a very strong quarter. And given that our focus will be a lot on Africa where there is still a lot of subscriber growth ahead of us, we expect to continue producing very, very strong subscriber numbers going forward. David?

Yes, I'm slightly more optimistic on the tax rate. As you can see on the tax slide the difference between the operations effective rate and the group rate comes down due to the initiatives that I've talked about on several calls that we're implementing. And we've been successful there, which I think is good news. I've talked about the operating losses in Colombia and the Congo and the impact that's had, you know that's been quite positive, Colombia in particular, the progress we're making there on our net losses has been good. So yes, I do expect the tax rate to be at 30% or possibly even below going forward.

Thank you very much.

You're welcome.

Thank you, Peter.

Alexander Vassiouk from Morgan Stanley is on the phone with a question. Yes, hello. I just wanted to clarify two points on the African margins. One is do you expect any further dilution of the EBITDA margin in this region? Or do you feel the margins have now bottomed out?

I think given that we had a very strong quarter, record subscriber intake in Africa, and that of course impacted the margin, I don't think we will see a lower margin than what we saw this quarter because this is a combination of high growth and lower margin. I think we'll see more margin going forward, especially that the subscribers we got into the third quarter, they will start generating revenue and EBITDA in the fourth quarter and then the subsequent quarters.

Right, thanks. And also would it be possible for you to clarify to what extent the margin pressure in Africa currently is coming from DRC as opposed to other established markets? So would you say the pressure is mainly concentrated within DRC and other countries are holding more or less stable?

No, I think the pressure and the growth, let's put it that way, the growth is coming from all of our markets. Of course, some of the markets were a little bit slow because of Ramadan at the end of the third quarter, but I would say that, yes, the impact has been kind of across Africa. There wasn't really one market that disappointed us during the third quarter different from the second quarter where there were a couple of markets that disappointed, that wasn't the case. So it's a well spread effort here and also well spread cost across the seven markets in Africa.

Yes, because correct me if I'm wrong, but I think a few months ago in the investor meetings you were indicating that you would expect DRC to break even at the EBITDA level in 2008. How has that expectation changed? Are you still confident about 2008 as a breakeven point?

I think we can see that happening in 2008, it all depends on how hard we will push in terms of network build out and subscriber intake. And as you've seen from this quarter, or from the third quarter, we at times want to push a little bit harder because we see those opportunities in those markets. And with us being better prepared to take advantage of the opportunities in Africa, largely also because of some great guys we have out there running our businesses at times in very difficult circumstances, I think that is making the difference at this point in time. We might not be as well prepared a year ago or two years ago when we really start looking at investing more money in Africa.

Okay, thank you very much.

Thank you.

Edward Field from M&G is on the phone with a question.

Hi, Marc, congratulations, there's a lot of happy shareholders out there.

Thank you.

I have a question just regarding new value-added services. Can you just talk a little bit about what you see in terms of opportunities for music, gambling and payment or banking services?

I'm sure the music and the gambling, I'm sure our customers are using those services, not that they are offered by ourselves. Probably music downloads some of them are, but not the gambling that I guess they can do themselves. But I think the big push for us going forward will be the introduction of 3G where, especially for the higher users, call them the post-paid users and the high-end prepaid users, we will allow them to get access to higher speeds when it comes to data downloads.

And I think that will allow us to take on some new traffic on our network which we don't have today. But at the same time we have some very creative guys sitting in Paraguay who come up every month almost with some new things nowadays, they've come up with a service that allows people to produce their own ring tone and they can share it with others. But every time there's something new.

Banking services as such I don't really see as a big revenue generator in the short term or medium term, I think banking penetration is still very low in most of the markets where we are today. What I can see maybe are some money transfer services that we might introduce and others might introduce over time. But the web banking, as we know it here in Europe for instance, I don't really see that as a major revenue generator going forward or a big value-added service going forward.

Thanks. And what markets are going to be your first to be upgraded to 3G?

The first one will be Paraguay, the orders were placed, I think, a month or so ago, and we expect to start launching our first services in second quarter of next year and then the other markets will follow. As you know, Paraguay is our best market when it comes to value-added services so that's where we're going to try out some things first and then we will copy them into the other markets, as we've done that before with other value-added services.

Thanks.

Thank you.

(inaudible) from Handelsbanken is on the phone with a question.

Yes. Thank you very much. Congratulations, a very strong report there. I'm just curious on the ARPU, making the calculation here, could you say anything on how big of a share are corporate uses in the network there in South and Central America?

You can make the assumption that most of our postpaid subscribers are coming out of Latin America, exception made for, I think, some in Mauritius and Cambodia, I think. But all other postpaid subscribers are primarily based out of Latin America and they do represent a relatively small percentage of the overall subscriber base but of course do represent a much bigger part of our revenue. And I don't have the exact number at this point in time, but I think our ARPUs in postpaids are probably about three times, four times the level of our prepaid subscribers. So you can make the calculation yourself more or less.

And then when it comes to interconnect, if you would make a calculation on average bill premiums, sir, can you say anything on that? I mean revenues coming from other sources than the end users that are supporting ARPU?

Yes, but ARPU represents all of that, so ARPU represents all kinds of traffic, incoming, outgoing, voice, value-added services all in one and the same.

Can you give any idea of how big of a share is the interconnect in value terms or share in percentage of the ARPU that is coming --?

Again, it very much varies from one country to another because it all depends on what the cost is of interconnect. We have countries with very low interconnect and we have countries with much higher interconnect, so I don't really have an average for you there.

Okay. Then there comes the CapEx, how important is the backbone and transmission to the base stations in your planning horizon now, going towards 3G, increasing capacity, et cetera.

That's becoming more and more important and that's one of the reasons why we've been laying ourselves some fiber optic in our networks in the countries, starting with Central America. And also South America and we have started doing the same in Africa, places like Ghana, for instance, we started doing that. So given that the volume of traffic is increasing, we are moving in some parts of the network away from microwave to fiber optic. It also increases the, of course, the availability of the network going forward, given that we work with rings. So yes that is important for our growth going forward.

Are you also as part of your expansion plans considering to go into fixed within the portfolio not only mobile by the wireless and the IP communication and the broader offering to the customers in different markets?

Our core business will be mobile going forward but at the same time we're doing some WiMAX in some of the markets, and so that works. But it's a niche product at this point in time.

Okay. And then finally, if I may, I just wondered when it comes to subscriber counting here, have you done any changes to the metrics or the standard procedure you had used previously in the measuring the number of subscribers?

As you know, we took a few provisions against the TDMA/CDMA subscribers in Latin America, so we're obviously churning those off and we're going to churn them off all before year end. And then, in terms of tightening up the definition, we've tightened up when we are counting new subscribers. So we want to make sure that new subscribers actually create a revenue-generating activity before they become a subscriber and don't, in the case of giving a SIM away, that we don't count that as a subscriber unless they actually generate revenue for us. So, slightly tightened up the definition there as well. So we want to make sure -- go ahead.

Is it different applied during the third quarter compared to the second quarter like, for example, in Colombia? Have you done a different metric or a different definition?

A slightly different definition was introduced in the third quarter so that, yes, there was a one-time only catch-up impact from changing that new subscriber definition slightly in Q3.

So explaining a part of the 200 plus charge in customers is also that you have expanded or changed the definition of what an active customer is?

Well, yes, we've changed it but it has not increased the number of subscribers, if that's what you're trying to say.

Well, I just needed to understand the very strong intake in the quarter, if that's part of an explanation where the regulator has had a higher number than what you have presented previously --.

Yes, but regulators is a different definition.

That's in Colombia and that's on a 90-day rather than a 60-day. But no, the subscriber intake is very genuine so, again, the only thing we've done is we've taken some provisions against all the technology subscribers in Latin America, so that would actually reduce slightly the intake. And then, obviously, we've changed the definition of new subscribers slightly, which again would slightly reduce the intake.

All right, I'm just curious is there any change that could have supported the very strong intake in the quarter. That was --.

No, let me put it very clearly so that there is no misunderstanding there. Had we not changed the definition of our subscribers, we would have had a higher subscriber intake in the third quarter.

Okay, thank you very much.

Andreas Ekstrom from Carnegie is on the phone with a question.

Yes, hi. My question has already been answered, thank you.

Okay.

Richard Prentiss from Raymond James Financial is on the phone with a question.

Yes, good afternoon. I wanted to probe a little further on the 3G that you mentioned. You mentioned you're going to be launching it in Paraguay I think second quarter. Can you talk to us a little bit about the CapEx cost as you look at deploying that into Paraguay and other markets, how much you're looking to spend? And also, does that mete back to that same IRR hurdle of the 20% to 25% that you mentioned earlier in the call?

Yes, although we look at 3G as a new business so there we apply the same criteria as we do for a new business, that means we're looking an IRR of that level over a period up to five years, and given that we will be tapping into completely new revenue, you know, it's the kind of revenues which we don't have today. Yes, we will be spending an important part of our CapEx next year on 3G, although the way technology works nowadays, when you buy equipment for what they call low traffic areas, when you buy equipment you basically buy 2G and 3G in one and the same box. So going forward it's going to be very difficult to basically split on those things, it's only in the high-traffic areas where you will buy specifically 3G equipment and put it on top of the 2G equipment.

Would we expect you'll be putting the 3G equipment predominantly in the high-traffic areas? Or will it be kind of throughout? How do you prioritize where to put 3G?

You will, by definition, go for the cities and the high-traffic areas. But if we were to buy our new networks or go into new cities, there we probably will rely on the new equipment with 2G and 3G in one and the same box. But for the big cities like Asuncion in Paraguay, yes, that will be 3G on top of 2G.

Great. Thanks a lot.

Adrian Dawes from Hartwell is on the phone with a question.

Marc, I just wanted to make sure I've understood this correctly (inaudible) again. The value-added services that you're offering obviously have higher margins. So they would, if they're adopted more widely than they are now and only 10% have been adopted across the company, actually increase your margins. That's in the phrase of a question. Therefore shouldn't the ARPU margins go up? That's the first part.

Secondly, inasmuch as you've already said, and we knew I think, that the per capita income in the various countries does relate to ARPU and does relate to penetration et cetera. And since Colombia, as you've said, has a higher per capita income than the other countries, shouldn't they, both because they're going to adapt the higher margin services and because they have the wherewithal to obviously pay for them due to the higher per capita income, shouldn't they actually eventually have higher margins than the rest of the neighboring countries?

In terms of the margins on value-added services there are a number of services you offer like SMS where typically it's sender keeps it all. So there it's like an on-net call so you don't have to share any of the revenue your customer generates, different from if you have a cross-net call where you have to kind of share the revenue with another operator. Yes, typically the countries with the higher GDP per capita should be able to spend more money on value-added services because they have more disposable income and, yes, they might be adopters of new technologies, new value-added services.

However, as I've said before, it very much is in our own control and our own hands because we need to make sure that we inform, we train our customers, we develop the services so that they can use more and more of those. But, yes, Colombia as a country with the highest GDP per capita should do that. But it will also require a lot of effort from our side.

Okay. Thanks very much. I just wanted to make sure because if in two years you're up to industry standards or your standards, you should be able to go higher after that just given the per capita income et cetera. Thank you.

Okay, thank you. Okay, any more questions?

No, there are no further questions.

Okay. So thank you, everybody, for being on today's call. For those who will join us on the Capital Markets Day in Colombia, in Bogota, I hope to see you all on Monday morning at 8 o'clock for our presentation. And to all the others, have a good day and we'll speak I guess in February for our Q4 and 2007 results. Thank you, goodbye.