FDJ United (FDJU) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for joining us for the FDJ conference call. I now hand over to Stephane Pallez, Chairwoman and CEO. Madam, please go ahead.

Thank you very much. So hello, everybody. Good evening, good morning, whatever. So I'm -- and as you know, I'm the Chairwoman and CEO of FDJ. And I'm with Pascal Chaffard, Executive Vice President, Finance, Performance and Strategy. So we're going to talk about our H1 results 2021. You may have read the press release that was published just after market close, and those -- and on the slide show that we will comment is available on our website on the Investor section. So we are going to start with a short presentation, of course, to allow you to have time for the Q&A session. I will start -- before turning to our first half 2021 results, I will start with the European Commission news. So as you have seen on Monday, European Commission has announced that they opened an in-depth investigation about the -- against the French state. So we are not part of at this stage of this process, but we will become soon. And this investigation is about the appropriateness of the EUR 380 million, that was paid by FDJ to the French state. As I'm quoting the terms of the European Commission, remuneration for executive rights related to FDJ for sports betting at point of sales and lottery. So as a lot of you have been present during the IPO, and after I think that you -- you have heard about this EUR 380 million payment. So you might remember that this amount represents the counterpart for the net value created by securing for 25 years the exclusive rights that FDJ Group previously held for an unlimited period. So this -- those rights were secured by the law, and the law and the text that followed the law said that in -- as a counterpart for this security over 25 -- on 25 years, FDJ had to pay what we call suit. So equalization payments. I think we call it in English during this process. And this EUR 380 million payment has been set after a process led with an independent body, the Holding and Transfer Commission, Commission des participations et des transferts, that has been publishing its methods for setting the sum that, by the way, we paid in 2020. So that's -- again, that's just a reminder of what -- of the basis of this sum and how it was fixed. And of course, we can come back to that later on. Just to also remind you that the risk of appeals against the legal texts enabling FDJ privatization as voted in Pacte Law and derived from, have been described in our documentation since the beginning in October 2019 registration document prior to the IPO, and reiterated since in our 2020 registration document. So this process has been open. Again, it's an investigation. It's an official investigation that has been opened by the commission. The next steps will be the release of the full notification in the European Commission official journal. Then, of course, we'll start the process itself. So it will be to the French state to answer. The delay is usually 1 month to answer but can be extended and it certainly will be extended. Hence, we will be a part of this procedure as an interested party. So that's basically what I can say at this point in terms of information that we -- that is known to us and virtually come back to that. But I want to come back now to the first half 2021 key highlights because we are quite happy to present those very good results. And I just want to highlight some elements before turning to Pascal. I think the first highlight is the good momentum despite restrictions related to the health crisis. Of course, those restrictions have been present a good part of this H1, but has been, of course, evolving, and I will -- I would come back to that. So it's interesting to note and [ spread ] that our business momentum has been and is quite good with stakes and revenue up high single digit compared to 2019. The pre-crisis comparison that we think is the most appropriate one. So I'm not going to take you back to the history of the various measures that have been taken by the French authorities since mid-March 2020. I think what's important is to state that now, this environment especially for our point of sale has been normalized between mid-May and the end of June with all our 30,000 for some sales opened from early June because of the last restrictions for bars and restaurants have been lifted. At the beginning of June, curfew also has been lifted. So of course, as we all know, we're not yet out of this. But I think it's quite interesting again to see that we have good momentum despite an environment that has been partly under some restrictions and that it's getting back to very good numbers when we come back to normal. So growth was already nice in Q1, which takes up 5 -- say, 6% versus Q1 2019, despite nearly 10% of our point of sales network being closed, and it has accelerated in Q2. It has accelerated mainly driven by point of sales lottery, instant games and especially Amigo. We have now also come back at the end of June, in line with June 2019 levels. And of course, with also a very good dynamic on sports betting, including the EURO 2020 as we -- as it was called championship. At the same time, of course, I want to stress because as you know, it's part of our business model that we have continued to deploy our commitments in terms particularly of responsible gaming. One of our major commitments, and it was, of course, absolutely key during the -- and particularly during the EURO, is the fact that we dedicate 10% of our annual TV advertising to responsible gaming and particularly to the prevention of underage gambling in order to raise awareness about those people and their families concerned. It has been done, and this commitment has been particularly strong during this past year -- at past year up, with the EURO 2020 championship as you might have heard that it's, of course, a sensitive period in terms of responsible gaming. We have also continued our long-term support for our network of retailers. We are now in the process of establishing, materially establishing this dedicated funds with some other actors. We put EUR 15 million in this fund that is now in the process of being established and will be able to help approximately 500 businesses with fragile communities. So launch with the special commission to help them go out of this crisis in good shape. So to turn back more precisely to the figures, when we compare our figures to the first half of 2019, our revenues increased by 9% to EUR 1.1 billion based of an increase of 8% in stakes to EUR 9.2 billion. Stakes are up on all games except Amigo that, as I said, has been going back to a good level at the end of this -- of the semester, but was, of course, still affected at the beginning. What is, I think, particularly important that stakes have been going in all distribution channels, off-line and online, we're very satisfied that our point of sale activity increased by 2% with a recovery in Q2 linked to the reopening of the 10% of our network that was closed for most of the semester, of course, [indiscernible]. And online, of course, continued to be very dynamic, up 71% to more than EUR 1 billion. So on this basis, EBITDA for the first half of the year reached EUR 261 million and it's a margin of 24.1%. I want to say that this margin was exceptionally amplified by factors that are still a little bit specific to this first half of year 2021. In particularly, costs initially budgeted for H1 2021 were postponed to H2 2021 because of the health context, particularly, of course, costs linked to our network that we were not able to deploy due to sanitary conditions. But Pascal, of course, will come back to that. And of course, the online mix effect was positive. Online stakes represent now 12% of the total of our stakes in H1 2021. And as you see on the next slide, it's quite impressive to see where we come from on this with 6% in H1 2019, 9% in the first half of 2020 and 12% now. Of course, the lockdowns and curfews related to the crisis have been the catalyst for this boom. But it's quite impressive to see that we're still on a very, very dynamic growth, particularly compared to Q2 2020, which is a very demanding base of comparison. Of course, this performance is driven particularly by the ongoing very high growth of online lottery, more than 50% compared to H1 2020 and almost 130% compared to H1 2019. Although strong, we continue to believe that this growth is responsible because it's mainly linked to the growth in the number of online players. So that's also a good point for the future. On this basis and considering the fact that -- which is the good news that we have now more growth coming from the product sales, and that we will add more in the next month. We believe that growth in online industry stakes will exceed 30% in the whole year of 2021, reaching EUR 1.5 billion, and which means continuing in the midterm at an annual rate between 20% and 30%. Again, one of the factors being also the environment, but we believe we are on, however, on a very good dynamic midterm trend. So we believe that in this context, even though, of course, the sanitary environment is not completely stable. But as we have demonstrated our capacity to adapt to several phases in this crisis, we are now able to take a forward-looking view for the entire 2021 fiscal year. So we anticipate continued good momentum in the second half of the year, thanks to the positive impact of the reopening of point of sales, the resumption of Amigo and instant games, both in terms of activity and costs, which means that full year stake should be around EUR 18.8 billion, with growth in the second half, high single digits compared to full year 2019 and H1 2019 mid-single digits when you compare it to the second half of 2020, which was a very dynamic semester as you know. We expect also full year revenue of EUR 2.2 billion. As for the EBITDA margin, Pascal will also certainly come back to that. We believe it should be at least 22% or slightly higher based on the fact that we will have increase in the activity in the point of sale in H2 with a mix effect resulting in additional costs, shift in fixed costs between the first and second half of the year, again, to accompany the growth of the point of sales. And we also are definitely determined to continue to invest in our online business, online lottery business. In terms of OpEx in the second half of the year to continue to boost online midterm growth in the midterm to again continue this trends over the period. EBITDA to free cash flow conversion is expected to be over 80% as in the past. And I will now hand over to Pascal to comment in more details. Thank you very much.

Thank you, Stephane, and good evening, everyone. I look forward to this presentation of the H1 2021 results. Please note first that 2019 data has been adjusted for the new tax regime effective since January 1, 2020, this is back to registration and the full year integration of [indiscernible] but are not projected for the exceptional luxury cycle, which has mainly impacted the second half or even the fourth quarter of 2019. Regarding our KPIs as of June 30, it is not relevant to consider the EBITDA to cash conversion rate because the half year analysis is impacted by calendar effects on this [ valuation ] in working capital. And it is also not relevant to extrapolate the net cash surplus position over the entire year as there is strong calendar impact at the end of the year with an advanced payment in December to the public levies due to -- for the month and the annual settlement of the -- for all the -- and claimed winning. So let's begin with the key figures for this half year, stake of EUR 9.2 billion, up 8.3% compared to June 30, 2019. The increase in sales is reflected in the increase in revenue, plus 8.7%. The EBITDA, I recall is defined in our case as the recurring operating income adjusted for depreciation and amortization, increased by 25.6% to EUR 261 million, hence margin of 24.1%. It was a margin of 20.5% in the first half of 2020 and 20.9% in the first half of 2019. As Stephane has mentioned before, this margin was exceptionally amplified by different things that I will come back, and I will come back later to this. If now we look at the increase of stake in detail. This 8% increase in stakes was driven by all activities with stakes of EUR 6.9 billion, the lottery grew by 4%. This growth was driven by instant games plus 7%, with launches during the first half of the year such as [indiscernible] in April or [indiscernible] in May were very well received while our best-selling games such as [ cash ] for example, continues to do well. Draw games are stable with 2 different trends. Amigo stakes are down by almost 50% due to the closure of about 10% of our point of sale network during most of the semester. It's worth mentioning that Amigo has recovered its full potential since the end of June. Excluding Amigo, draw game stakes are up 20%. Loto and EuroMillions continue to benefit from the success of their relaunches and have recorded long titles in Q1. Sport betting stakes at EUR 2.3 billion are up 25%, driven by a growing market and the EURO 2020 championship in June 2021. I won't go back to the progression by disruption channel that Stephane just mentioned. In terms now of revenues for our business units with broadly stable player payout rates, in both lottery and sports betting activities, the growth is in line with the growth in stakes that we have just commented on. Overall, the slightly higher increase in revenue compared to the stakes is explained by the development of adjacent activities or we can say also diversification, mainly payment in selected international B2B. Now let's detail the bridge from revenue to EBITDA. The EBITDA margin of H1 has been, as we said before, exceptionally amplified, as mentioned by Stephane and I. This is mainly due to 2 factors. The first remarkable point is the difference between the growth in the point of sale commission that you can see on the left-hand of the slide, plus 1.9% and the growth of the turnover, plus 8.7%. In H1, luxury growth comes exclusively from digital, more than 100% with a relative effect that has boosted the EBITDA margin of H1. This effect cannot be projected on a yearly basis as we expect a more balanced growth in H2 with point of sale back to a more normal situation. The second point to be noted is the decline in other cost of sale, down 2.5% compared to H1 2019, and on a small scale of the decline in G&A. This is in particular due to the fact that some expenses have been postponed to the second half of the year, mainly commercial expenses, given the health conditions in the first half of the year and the closure of 10% of our network. For example, POS marketing was very low on H1, and maintenance in POS was greatly reduced. We expect those costs to resume in H2. The other cost items increased in H1 2021 versus H1 2019. The continuation of our investments to develop our games and services offering led those costs to 32% growth compared to the first half of 2019, plus 19% versus H1 2020. This growth is driven by the group strategy, notably the continuous investment in the digital activity, which will be further increased in H2. It is also driven by development of the adjacent activity, which most of the costs are accounted for in this category. Advertisement and promotions have increased by nearly 12% compared to H1 2019. And this is a higher rate than the turnout growth, explained by the online activity that demands more [ A&P ] notably to recruit new players. And also explained by the new corporate campaign about the EURO 2020 that took place in 2021. As Stephane mentioned before, we expect a full year EBITDA margin greater than or equal to 22%, close to the EBITDA margin of 2020. The 200-basis point difference between the full year expected EBITDA margin and H1 can be explained by 3 factors, the catching up of costs in H2, commercial expenses, and as we already have mentioned, further investments, especially driven by digital. Second, the weight of POS activity with a more balanced growth between POS and online; and third, a part of the gap is also explained by traditional greater weight of cost in H1 versus -- in H2 versus H1. So now let's go to the breakdown of EBITDA. The group's EBITDA is made up of the contribution margin of our 2 business units, luxury and sport betting, and also the acceleration business units. And we deduct from this contribution margin, our holding cost, G&A and corporate communication. For lottery, the contribution margin was EUR 298 million, representing a margin rate close to 37% versus 33% in H1 2019. We already explained why this contribution margin is high, digital mix effects and also lower commercial investments. For sports betting, the contribution margin was EUR 61 million, representing a margin rate of 25% versus 24% in H1 2019. Finally, holding cost of EUR 98 million were stable compared to H1 2019. G&A are decreasing and corporate communication is increasing as we launched a new corporate campaign. If we look now at the evolution of EBITDA in absolute figures, we see that the increase in EBITDA between H1 2019 and H1 2021 is driven by clearly 80% by lottery, which are more than sports betting balance sheet from the postponement of expenses in H1, and also benefited from the digital mix effect that we have already registered. Let's now detail the shift from EBITDA of EUR 261 million to net income of EUR 146 million. The depreciation and amortization amounted EUR 63 million, which include about EUR 10 million related to 2 factors: first, an accelerated depreciation of digital projects over 3 years versus 5 years previously, with an impact of about EUR 7 million on H1 and EUR 10 million over the whole year. In effect, digital technology requires continuous reinvestment in development, which at the end have a shorter life span. Hence, the adjustment for -- of the depreciation period for technological development of digital projects. The second factor is the beginning of the amortization of the partnership with the organizing committee for Olympic Games, Paris 2024, which began in early 2021. The other significant figures is the tax charge of EUR 57 million, equivalent to a rate of 28%, which can be projected to the end of the year. One of the indicators of the group's cash position is the net cash surplus, which stood at EUR 712 million at June 30, 2021, compared with EUR 298 million at June 30, 2020. The bridge from cash and cash equivalents on the balance sheet to available cash of nearly EUR 1 billion is as follows: adding EUR 381 million of term deposits, which are available within 32 days and deducting EUR 104 million in sum, not available for other purposes. Then the bridge between available cash of nearly EUR 1 billion and the net cash surplus is as follows: adding EUR 296 million of other financial investments and deducting gross financial debt of EUR 555 million. Thank you for your attention. And now with Stephane, we are ready to answer your questions.

[Operator Instructions] We have a first question from James Ainley from Citi.

Yes. I've got 3, please. First of all, can you talk a bit more about the European Commission move? What do you think has motivated this? And I suppose related to that, how does that sort of overhang impact your views on that capital allocation, that is perhaps not politic to perhaps launch a big special dividend, for example, but how has that also impact to your thoughts on M&A? Second question is, how much do you think the COVID restrictions impacted revenue in the first half overall? And do you think you're now seeing an outsized bounce back now that the retailers back open? And I'll leave it there.

Okay. Thank you. So I will start with your first question about European Commission. And ask Pascal to come back to your question about the impact of COVID on our revenues in first half and therefore, expectation for H2. So it's rather difficult to comment on the motivation for European Commission move. Because, as I said, well, this -- we are not been part of this -- of the informal process that was before the official opening of this investigation. And again, it is and investigation that has been opened. So what we have -- what we know is that there have been several companies, several competitors that have been -- that has been asking European Commission to open an investigation. Again, not a surprise that they have tried this as we exposed those with from the beginning. And again, before the IPO, we knew that -- well, it is part of the business model of those companies to challenge exclusive rights, particularly. So again, everything that has been done to prepare the Pacte Law and the privatization and has been, I would say, thoroughly told before. And I also want to know that the French state has several contracts with the commission before proposing the Pacte Law, and voting the Pacte Law and so on. So it starts, again, something that is new and that is discovered. So again, no surprise to the fact that people are trying. Why is the commission opening the investigation now? I cannot comment because I don't have any clue about it and has not been part of the process before. Hence, what, again, I can state that this -- the legal framework that has been voted by Pacte on the -- on securing the exclusive rights and the fact that we are paying the net value of this security over 25 -- on 25 years for rights that we had before that were not secured within legal framework is something that was exposed and overexposed and discussed and went through company data and everything. So again, this is something that has been really very much studied and very publicly studied before it was enacted. So at this stage, I don't -- there is no consequence that I can draw from this. Again, the procedure of the investigation is starting. It's an investigation. And so I cannot -- I'm only going to be -- now I'm going to be an associated party. So of course, FDJ is going to be able offer to participate to the procedure besides the French state, which is, again, the direct party [indiscernible]. So I don't have -- I have no other comment than just say that, well, we're going to explain this situation. And I don't think there is any other consequence that I can draw at this point. Pascal?

Yes. So you wanted to talk about the impact of COVID on our revenue on H1 and what we can say about this on H2? On H1, if you want to summarize it very easily, the impact is Amigo, to be clear. Amigo was nearly half of it's potential on the H1. And this is the major impact. The rest of the impact is not so important. We learned to live with the COVID. We had a little bit more digital, a little bit less instant games maybe during the restrictions. But we cannot say that it has a major impact on our activity, except the closure of the bars. So if we talk about what we expect in H2 in our guidance, we don't expect a new COVID crisis coming up very hard. And what could happen, we don't know. But what could happen is if there is some restriction in point of sales, I don't talk about the closure of every bars, but some restriction on the point of sales. Maybe we will be able to mitigate it with the online activity. That was the case barely in first half. The just thing that we did not put into our guidance is the closure of bars, a new closure on a national basis of the bars. So we will see if this happens, but we hope that it will not happen in the second half of the year.

Could I just come back to my first question on [ BC ] movement. How does that impact your thoughts on capital allocation? And how you might use those service costs?

Well, there was no impact. First, for the reason that I've been stating, which is that we are very, very confident in what has been -- in the way this has been established and set by the state. And second, because as you know, we have, I would say, a large capacity given our balance sheet and our low leverage at this point. So frankly, at this stage, no impact. Sorry, so you're right, I should have said that. So thank you for asking the question again.

Next question from [ Ted John ] from Morgan Stanley.

The first one perhaps, on the size of these accelerated investments. If I look at the cost lines, they're pretty close to what you had in H2. And you said H2 is normally a bigger weighting for costs. So it doesn't look on the numbers like there's a huge amount of cost that's been deferred or pushed out. So could you give us an idea of how large the delayed POS advertising campaign, et cetera, looks like? Because your full year margin guidance for '20 doesn't look very conservative and those are quite large numbers. And then second of all, on the online development, I think, I heard 2 numbers, one of which was the lottery -- online lottery stakes were about 50%, you said. I think, importantly, overall online stakes above what at 30% for the year, was your expectation? And if I heard that right, that implies down 15% half-on-half in H2. Are you really expecting such a strong slowdown in online stakes in H2 to hit that? Or is that again conservative if I take the H1 number and run across flat, you'd end up with more than 40% growth. So some help or some color around those 2 would be useful.

Okay. Pascal, do you want to....

Okay. I can take it. If I well understood your first question, you want to know where is this acceleration of expenses that you don't see in our figures. You don't see it if you look at the total expenses because first half, we had some items very low like we said on the cost of sales, et cetera. The acceleration we talk about is about develop -- offers and service development. This item is 32% in excess compared to first half of 2019 is, as I said, 19% in excess when we compare it to the first half of 2020. And inside those costs, there is our road map of development of -- digital development. And this is one point where we will accelerate. This acceleration is not EUR 20 million to be clear, it's some million, it's some million, and it will be seen at the end of the year in this item called gaming development and operation. So this -- and the reason why in the first half, you don't see a bigger expenses in other items is because of the cost of sales and G&A were down. And your second question is what is the growth of the online lottery in second part of the year? We expect it to be around more 20% than 15%, to be clear. And this has to be put -- it has to be compared to what we expect on the -- also on the retail network. We expect high H2 in the retail network. This is also why we expect 20% growth in second half of the year on the lottery.

Okay. I'll just come back a bit. That's useful. On the online stakes overall, I think, you definitely spoke to 30% growth for the year, but could be 71% in H1. Did I hear that number correctly? [indiscernible] H2.

I understand. The 71% in H1 is a global online stakes including horse betting. If you look only at online lottery, the correct figure is 51%, 51% of growth in H1. This will go better with the 20% I talked about in H2, to go to the 30% overall on the full year.

Next question from a Grewal from Bank of America.

Just a couple of questions. Firstly, what have you seen with online stakes since reopening, given it's been of months now? Have you seen a huge pullback? Is that how you're getting to the number? And you've also mentioned Amigo's recovering and it's open everywhere. Are you seeing those volumes return back to normal? Or does it remain suppressed? And last question. Apologies if you've gone through this already, but -- what's the likely time line for the European Commission investigation? What should we be watching out for?

I will start maybe with your third question. On the time frame, quite difficult to answer. What is known is that there will be probably in the next weeks, but sometimes it takes longer, but I would say probably in the next weeks, a release of the full notification of the decision of the commission to open this investigation. We -- as I said, the French State, at first, has 1 month to reply to this opening, but we'll ask 1 month delay as it is very usual. So this leads us to reply at the end of September. This can be extended actually and is often extended as I understand it. All the interested parties can also contribute. So this -- there is a first phase. I think the commission has 18 months in principle to -- well, to conclude this investigation one way or another. But actually, it can also extend this way. So -- and on top of it, of course, if at some point the commission is coming with a decision at the end of this investigation, which can take quite a time, you can have, of course, you can appeal. All the parties can appeal this decision. So -- my experience from those procedures, not my direct experience in FDJ but the experience that I heard of, is that is something that is -- that can be fairly long. And that you sort of have to live it because it's quite -- again, it's not completely -- it's not entwined in a very strict delay as, again, as I've seen it. So we'll see. When we know more, we'll tell you. But at this point, quite difficult to be more precise. Pascal, may be...

Maybe I can take the answers on the online stakes and the recovering of the POS network. First, on the online stakes, you are right. We had a 90% growth on H1 and the growth on the -- solely on Q1. And the growth on Q2 is less important than one in the Q1. But I will not talk about pullback, it's wrong to say that, for 3 reasons. First reason, on the Q2 2020, the comparison basis was very high because we doubled, almost doubled, the online lottery during the first curfew, or the first -- sorry, the first lockdown. So the comparison basis is quite high on Q2. Second thing, we had a recovery of point of sales at the end of the semester, which has also an impact on the growth of the online. But it's good news globally for FDJ. And the third thing is more conjunctural. We had a lot of long cycles in the lottery as we talked about in Q1. For example, a EUR 210 million jackpot on EuroMillions and this leads a lot of people to the online lottery. So we had not this opportunity in Q2, but we hope that the luck will be back with us in Q3, and Q4 above that. So the results of online lottery in Q2 is not a bad result. We have consolidated what we had before. The player that has now being recruited to the online lottery are still playing on the online lottery and they are playing more and more. And we are working very hard to make them stay and make them also play a little bit more because you know that our players on the online play on average a little bit less players on the point of sales. So we have a lot of things to do to make it grow in the future. The second thing, if I well understood your question, is are we back to normal on the point of sale? Definitely, yes. Totally back to normal. What is really very, very interesting and we are very happy with that is the number of point of sale opened. We talked about 20% of the point of sales closed the first lockdown and 10%, and we added some -- we were a little bit [indiscernible] to say what will happen with those point of sales closed for a long period. Will they reopen really or not? And we are very happy to say that in June, we have, again, more than 330,000 point of sales. This is good, very good news. And we are also very happy to see Amigo, at the end of June, back to normal with exactly the same kind of turnover that we had before the crisis. And we were a little bit worried also about Amigo because this game has been almost sleeping during 8 months, and it has restarted as if he had never slept.

Next question from Jaafar Mestari from Exane BNP Paribas.

I've got 2 questions, please, and maybe I'll ask them separately, if that's okay. So firstly, just on H2 margin. So H1 margin was 24% and your full year guidance suggests something like only 20% for H2. Qualitatively, I understand all the points you're making but the numbers don't seem that big, so you're flagging marketing investments, but then you're saying it's less than EUR 20 million. And you're either flagging online mix as an H1 one-off, but then you just said that it will continue to outgrow the rest of the business in online lottery in H2. So I'm just not quite sure how we can bridge 24% with 20% on a very, very similar revenue base. Is there a headwind that I'm missing here?

Pascal?

Okay. Okay. Stephane told me that I will have this question. And I have this question, I think, it's a...

I already asked these questions.

These are normal questions. If I can help you by bridging the margin of H1 to the global year, I said 3 items. The first item is the low expenses in H1 and more important expenses in H2. And you can assume that it's something like 1 point of EBITDA from 24% to 22% to the whole year. So you can assume also that it is 2 points when you compare it from H1 to H2. If I make it clear, if you have EUR 10 million, EUR 10 million means 1 point of EBITDA on H1. If you have EUR 10 million that are missing on H1 and that will be there on H2, the impact is not EUR 10 million. It's EUR 20 million on a yearly -- on H2, from H1 to H2. So it's 2 points when you said that. So we are talking about quite small numbers at the end, but it's quite huge impact on the margin. So the first thing is the expenses were low in H1 and more important in H2. And in surplus of this -- in addition of this effect, we will invest more in H2, but it's just in addition to this effect. So one point is explained by -- from 24% to 22% is explained by what I said. You can assume that out point is explained by the -- what we said about the strength of the point of sales in H2 more important than H1. And the rest is due to this effect that we see every year except last year, because the year is very different from the others, where we have always more expenses in H2 than in H1. So if I can summarize, the bridge from 24% to 22% is globally 1 point low expensive H1 and more expensive H2, out point the digital mix -- mix digital point of effects. And [ off ] point due to sort of traditional more expensive in H2. Is that more clear?

Yes, that seems, just like us, to close to 400 basis points.

To be clear, there is nothing else there. No headwinds not communicated here.

Okay. So 400 bps from all these things. And I guess, secondly, I just think it would be great to have a sort of public finances 101 refresher here. Because we're starting to see some dangerous almost misinformation. Some newspapers are calling the EUR 380 million, they're calling it license costs, which obviously would be very cheap license costs, where as, as you said correctly...

[indiscernible]

Yes. No, I know you said correctly, it's a...

Yes, absolutely. No, no, you're right. I'm sort of chasing -- I'm chasing this. Because it's not a license cost. I think it's absolutely clear from the beginning. It has been explained all over the place, but some people still call it license cost, which is absolutely, which is a big, big mistake. And so we are chasing that. And because, again, when you look at what we have explained from the beginning, 2019, it has never been a license cost. So the comparison with license cost is just stupid.

That's why we collect equalization payments. [indiscernible]

But to be fair, neither you or the states or the CPT have ever discussed the gross numbers. So roughly if you drew a [indiscernible] in the mind of civil servants today, what would you say was the implied cost of the license, just so we can get a better sense of how realistic it was?

No, there was no cost of the license. I think there was -- the CPT has published, I don't know, I don't remember what is the exact name of what was published, but has published its decision or its advice to the state, stating that this equalization payment was set according to 3 different methods that...

I can summarize.

Maybe you can summarize [indiscernible] The method was actually -- they actually indicated in their advice the range that resulted from any of those 3 methods. And on this basis, they actually set this net value, but there was no license costs at any point.

The 3 methods, to be clear, and I think it's quite important, and it has been made public. You can find it...

Yes, yes. I've got it with me. I've got it with me. It's in there. As you said, it [indiscernible]. And I guess if you want to assess how fair it was, you have to decide what the growth was. So maybe let's ask it completely separately. If I'm an investor with a 25-year or 26-year investment horizon, I need to embed a license cost in how I value the FDJ shares because you're going to have to pay this and next time you're not going to have [indiscernible].

[indiscernible]

This is another question. This is another question. This is another question, which is what would we have to pay at the end of the 25 years period, to extend this -- those [indiscernible].

And there, you're right, it will not be an equalization payment at this time. We don't know, but it will not be called an equalization payment because the rights will have ended, and it will be new rights.

But I think [indiscernible] asked this question during the IPO process. And clearly, we have not been -- we've not been able to give you this price because it's absolutely impossible to give this price now. So we know that -- it's fair to say that at the end of 25 years, there will be something to pay. But I don't think we have today the possibility to state what it would be.

That is true.

Okay. Is it -- I'll leave it here. Just to get an order of magnitude, realistically, 25 years to operate this business, which is going to do EUR 418 million of EBITDA this year. Are we talking hundreds of millions of euros or are we talking billion of euros? What was implied in how the state priced the equalization payment?

Frankly, today, I'm not able to answer this question. And since I must say that, again, I can't tell you what we've been stating on and what has been published in terms of the first equalization payment. In the context that we are today, I'm not going to say anything else than what is public information on that. And I don't have any other [indiscernible] public information on this. And so frankly speaking, I cannot answer your question.

Next question from Sabrina Blanc from Societe Generale.

Sabrina Blanc from Societe Generale speaking. I have 3 questions, if I may. 2 are very simple. And the first one is regarding the working capital. And you mentioned that we have nonrecurring element in H1 and on a full year basis. And could you come back on those effects? And the second one is a question about CapEx. How should we see at the end of the year? And the third one, sorry, I come back on the equalization. Just could you remind us that -- if I understood correctly, on top of the equalization payment, I know that you had to pay some other elements like the gather and collected player winnings or -- sorry, I don't know how this is acting in this, but [indiscernible]. Could you give us a clear view on that?

Okay. So on the last one, you're absolutely right. We did not only pay EUR 380 million. We actually paid over EUR 400 million. My recollection that we paid EUR 420 million. Because in the context of the -- as a consequence of the privatization, we gave back to the state all those different elements, the one that you mentioned, [Foreign Language]. That's, of course, were part of the scheme when the state was our majority shareholder, but were not compatible with the privatization. So not only we paid EUR 380 million, but we also paid back over EUR 400 million. And in the current framework of relationship with the state, we are not able to keep the [Foreign Language].

The unclaimed winnings.

Unclaimed winnings, thank you, Pascal. So we have to give them back every year. So it's, of course, in addition to what we take the first year as a kind of setup of the new relations between the state and us.

It makes it very clear that the newer framework has some elements that benefited to FDJ, some other elements that were negative. And this is why this equalization payment is some plus, some negative things. And this cannot be only plus, it's only -- it's plus and minus. Maybe to continue with the working capital because you will see that we will also talk about unclaimed winnings. For the working capital, the reason why it is not relevant in -- early in H1 is that public levies are paid to the state every month. But in December, we have 2 months, we have -- we pay the public levy of November. And we pay also a different payment for December. So when you see those working capital in June, it's not normalized. You have more working capital in June that we have at the end of the year. The second thing that is not normalized is the unclaimed winnings. We collect the unclaimed -- we don't collect. We keep the unclaimed winnings for all our activity or luxury and [indiscernible] all the year. And at the end of the year, we give it back to the state. So we have also a second effect -- negative effect on the working capital at the end of the year. This is why it's not interesting to look at it just at the end of June. If I continue with your question on CapEx, you have probably seen that our CapEx was quite low on the first H1. They will be higher on H2. So please don't only double the H1 number. It will be more. It will be more in line with what we said during the IPO and after something like EUR 100 million CapEx a year, more in line with that than EUR [ 33 ] million multiplied by 2.

Next question from Simon Davies from Deutsche Bank.

Three for me, please. Firstly, just on digital. Very, very strong growth in digital performance in the first half. What should we think in terms of the likely or the possibility of scaling down the point-of-sale network over the next year or so given the extent of that growth? And are you going to reset your targets in terms of digital penetration for lottery given that strong performance? Secondly, I may have missed this, but in terms of the repayments of unclaimed winnings, can you give us a sense of what the normal range would be at the year-end? And lastly, in slightly left field, but obviously, we've got regulation coming through in Germany and Holland, both legalizing online casino. Is that something that you think might happen in France over the next sort of 2 to 3 years in terms of legalizing iGaming?

Okay. So on digital. I understand that your question is, do we have plan to scale down our point-of-sale networks? It's really -- it's really not at all our view because we are quite convinced that the good strategy is actually to build digital in addition to the point of sale, which is as you've seen, for instance, in the figures that we have this year, a very strong element of customer relations and growth. So the question I think for us is more how we combine our 2 channels, as we call it, we call it an omnichannel strategy. The question is more how we combine the 2 rather than scale down our point-of-sale network. And again, we are quite satisfied with the idea that we can actually combine low growth, but on a large basis of customers, with high growth in the digital. In addition to that, I just want to remind you that all [indiscernible] for sports bettings are only in point of sale. It's actually quite a good business today. It's not growing as fast as the online markets, but still growing close to 20% in the last numbers with a very -- with a very nice level of margin. So it will really be a pity, I think, to shrink that part of our business today. So in terms of digital objectives, I think we said that -- we believe that online lottery can grow between 20% and 30%. We have not, at this point set midterm targets for online lottery share of our business. We will come back to that in the fall as we are -- we have the project of refreshing our midterm objectives with probably in the Capital Markets Day in the fall, if, of course, the environment is enabling us to do that. So that, I think for your first question. Your second question was on the level of unclaimed winnings. Pascal, you want to...

Unclaimed winnings. Yes. I can take it. Unclaimed winnings. So unclaimed winnings are globally on -- if we talk about instant games, you can assume that it is something like 1% of the stake. Of the stake on point of sale because you don't have any unclaimed winnings online as we know every customer and we pay the customer automatically. So no unclaimed winnings. So you can assume 1% of the instant games on offline, and it is less than 1% on sports betting and draw games -- less -- between 0.5% and 1% of the stakes. Globally, if we talk about figures, the average unclaimed winnings for [ some ] games should be something like EUR 80 million, and something like EUR [ 50 ] million for sports betting and draw games. Last year, we had a lower level for unclaimed winnings because, as you know, our activity was globally down. And as a lot of point of sales were closed, we did not cut off this unclaimed winnings at the speed that we do it usually. So the figures for the 2020 year was EUR 70 million globally.

Can you do this question on the [indiscernible] ...

Online casino. [indiscernible]

Yes, yes. Online casino. Online casino has been [indiscernible] in the context of the [indiscernible], so in 2019. So clearly, some people advocated for taking the opportunity of the new regulation framework to open online casino. It was clearly and massively rejected, frankly speaking, in our parliament. So I don't think this is something that is likely to happen in the short term. I know that some actors are lobbying for it, but I don't think it's something that is visible, I would say, again, in the short term. So it's difficult to comment further.

Next question from Alexandre Gérard from CIC.

Firstly, on sports betting. I'd love to have some more information to be able to better appreciate your performance over the first half of the year. Is it that 25% you mentioned compared to 2019 coming more from new players that you've been able to reflect? Or is it a higher average spending per player, and in terms of market share, consequently over the first half of the year compared to your competitors? Where do you -- where are you? And also to get to appreciate your performance by distribution channel off-line versus online, and also in terms of contribution margin between the 2 segments, where are we? And second question, just on M&A opportunities. Are there new upcoming M&A opportunities that might be interesting for you to penetrate to other segments of the market like poker?

Maybe to start with your last point. On poker, I think we've been quite clear by the fact that we think that to be a profitable and I would say, significant actor on sport betting, you actually need to other online activities. So that poker in that regard is definitely for us a potential complement that we would like to get in addition to our sports betting activity. And we have again, in this -- as a consequence of this statement, we have engaged discussion to have access to some portfolio offers in -- that you would get from specialized actors, so [indiscernible] So we are not yet there, but it's part of plans. I think we didn't make any mystery about it. Again, not because poker is a fabulous activity, but it's because it's a way to keep your customers with you rather than have them being chased by others and taken by others, including the sports betting. So that's for poker. On sports betting, I am embarrassed by your question because you -- basically you ask us to communicate data that would separate online sports betting and off-line sports betting, and we do not do it. We communicate on the whole segment as a whole. So quite difficult to answer precisely to your questions. What we can say, however, in terms of performance on online sports betting is that it seems that we've been quite good at -- in the context of the market and particularly during this semester. And during the euro, we've been able not only to keep our market share, but probably to slightly increase it. However, as you, I think, as you know, certainly, we are not amongst the 3 leaders of the market. So good -- I would say, good position and good defense and probably a little bit grow of our market share, but I would say no major change regarding our market share on online sports betting. The whole of our activity, again, has been going at a good level. But again, different level of growth between the point of sale and the markets. No surprise regarding that. And again, do you want to...

Just maybe you asked also -- this growth has been has been driven by recruitment or upsell. Clearly, what we aim on -- for betting online is to recruit new players. That's why we got quite a lot to do so, and we recruited a lot of new players during this first half. And also, we try to do some upsell from existing players with the 2 that we had a bigger recruitment plan on this H1. Your third question was about the difference of -- on contribution margin between online and offline. I will answer for lottery because for betting, there is no relative effect offline because the revenue is not the same. On the lottery side, the figures have not changed. It is always 1.5x to 2x the contribution margin of the point of sales online.

I think your last question was on the M&A or potential M&A and we see no -- no news today that can be shared.

Next question from Johanna Jourdain from ODDO.

Just one question for me regarding the -- follow-up regarding your last comment on the poker and potential discussions to have access to poker offer. [indiscernible], I would like to better understand how does it work in the economics that would be related to that kind of project? Or would you have to pay some fees to use these technology capabilities or the brand for that? So could you give us more details? And how long should we expect in terms of having some announcements for this kind of project?

Yes. Well, of course, yes, we would have to pay a regulatory fee to access -- to the platform in terms of revenue sharing. That's the type of agreements that I would say, all the players that actually do have access to those platforms pay. So it's very common. Nothing special. I think we will come back to that when and if we have an agreement to explain, I already said. So it's a little bit early to give you more details.

[indiscernible] revenue sharing, it's low investment and low risk.

Yes.

Thank you. That was the last question. I give back the floor to your conclusion.

Okay. Well, thank you very much for your questions. I hope that we've been covering a fair amount of subjects and of course, don't hesitate to -- if you have any follow-up, to contact us. We'd be happy if we have the answers to answer you. So bye-bye. Thank you.

Goodbye. Thank you very much.

Thank you, ladies and gentlemen. This concludes today's conference call. Thank you for your participation. You may now disconnect your lines.