Deutsche Pfandbriefbank AG

XETRA:PBB

Earnings Call Transcript 2020-Q3

[Operator]

Hello, ladies and gentlemen, and welcome to the Deutsche Pfandbrief AG conference call regarding the Q3 2020 results. [Operator Instructions] Let me now turn the floor over to your host, Walter Allwicher.

[Walter Allwicher, Head of Communications]

Good morning, and a warm welcome from Garching to our Q3 results call, and we'll jump right into the matter with Andreas Arndt presenting the results. And Andreas, the floor is yours.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Thank you very much. Good morning, and welcome to our analyst call regarding third quarter results. In the light of recently increased COVID infection numbers, I still hope that you and your families are all well and healthy. While the current fourth quarter sees infection numbers rising and lockdown measures having become more stringent again, third quarter results in some way, reflected a certain amount of stabilizing momentum, as seen during the summer. However, the last few weeks taught us to stay humble and cautious as infection rates are soaring and a second wave is underway. The news that vaccine has been found and tested might delight markets and people at large, of course. But in the short run, we have to face the fact, i.e., further negative implications from renewed lockdown such as shop closures, increasing joblessness, vacancies and so on. Third quarter, we show a good PBT of EUR 75 million, especially supported by lower risk provisions versus Q1 and Q2. Again, lower funding costs and positive carry from TLTRO since the end of June. As indicated, we remain watchful, although we have still no indications for any significant rise in individual loan losses, the current development of the pandemic and related measures give us enough reason to remain overall concerned. The crisis is far from mover, and it is different from last crisis, the global financial crisis which started with real estate and triggered the whole economic crisis thereafter. This time, the reaction chain works from the other end. The pandemic works on the economic development and the negative economic development causes vacancies, higher yield requirements, low cash flows and collection rates and lower valuations as a consequence for commercial real estate markets and with some delay. This and the fact that we are still 6 weeks before year-end and 3 months before we go to close our books for 2020, those are the reasons why we still abstained from full year guidance, but let's have a look at the third quarter figures now. As I said, PBT of EUR 75 million is slightly up against last year's third quarter, but represents a significant increase after the first 2 quarters of 2020 that were charged with corona-induced additional risk provisions. Subsequently, the 9-month figure of EUR 106 million reflects this as well and thus, remains below previous levels of previous year. The main drivers in Q3 were the following: NII is up 8% quarter-over-quarter and 13% year-over-year at EUR 126 million, driven by stable operating interest income and supported by lower funding costs; a positive contribution from the TLTRO 3; as well as improved floor income. 9-month figures are also up by 4% to EUR 354 million, plus EUR 13 million against last year. General admin expenses remained largely stable with the EUR 48 million for Q3 and EUR 145 million for 9 months, which only slightly higher year-on-year due to some higher cost for regulatory projects in 2020. In line with our expectations, risk provisions of EUR 14 million remained on moderate level due to 2 opposite effects. First one is we record EUR 70 million net release of model-based provisions in Stages 1 and 2, and we did take to our books another EUR 30 million additions in stage 3 from revaluations on already provisioned U.K. shopping centers, but no new stage 3 cases. New business reached EUR 1.6 billion in Q3, picking up again versus Q2, but continuing to reflect an overall lower investment activity in the market due to persisting uncertainties with regards to COVID-19 pandemic as well as our even more selective approach in this context. For the 9 months, new business volume adds up to EUR 4.4 billion, of which EUR 4.3 billion go to our core strategic proposition, i.e., real estate finance. The development segment of gross interest margins continue to show a positive trend quarter-over-quarter. Average margin was up a further 5 basis points in Q3 versus Q2, which brings us, on a quarterly basis, to more than 190 basis points and which brings up the 9 months figure to 180 basis points, which is more than 25 basis points up from the level of 2019. You may remember, we started early in 2019 with 125, 135 basis points, and you can see the way we have made in between. And we are well above the planned figures for margin development. With regards to financing volumes, the market mechanics on work as expected. The strategic Real Estate Finance volume continues to stay largely stable at EUR 26.8 billion, as low new business volume was counterbalanced by lower prepayments. Public investment finance and value portfolio are slightly down quarter-over-quarter. We adhere to our high-risk standards, staying selective and strongly focused on risk monitoring, not least because of this, NPLs remained low at a ratio of 0.8%. There's not much new on funding. As already mentioned last time, our precrisis activities as well as the participation in TLTRO 3 provides us with a capable liquidity position well into 2021. However, we take opportunities in the funding markets, especially via private placements, and did a very successful -- as you probably have seen and have noticed, a very successful pound mortgage Pfandbrief benchmark issuance at very attractive levels, the first SONIA-linked issuance at all for Pfandbrief. In total, we collected approximately EUR 1 billion in Q3, resulting in total new funding volume of EUR 3.4 billion for the first 9 months, excluding, of course, the effects from TLTRO. Capitalization remains solid with the CET1 ratio of 15.3%, only slightly down against last quarter and reflecting a precautionary PD downgrade to the U.K. and the U.S. portfolio. I'll come back to that later. With regards to dividends, there's no change. We follow the current recommendations of the ECB, and we'll reassess the situation beginning of next year and until preliminary full year results publication, which is some time early in March at the latest. The good operating performance in Q3 and operating resilience clearly supports our expectations to achieve a solid NII at stable general admin expenses for the full year 2020. Nevertheless, uncertainties on risk costs and valuation effects persist given the current environment and the current developments, i.e., increasing infection numbers and lockdown measures all of the place. So in short, as of today's assessment, we expect to achieve a solid full year result in 2020. But as mentioned, we continue to refrain from giving you full year -- full guidance for the full year. With that, I'll switch over to Page 5, make it very short because the numbers are there, the representation of the figures, which I just gave. 2 things I want to remark is what lead your attention to us. The cost-to-income ratio stays low despite crisis. And despite crisis and significant increase in risk provisioning, we are on sort of 3 quarters basis, on a positive PBT and an ROE, which is above 4%. Now let's, as usual, move to the outlook on markets. Overall investment activity, overall transaction activity continues to stay on reduced levels also in Q3 after having dropped significantly in Q2 due to COVID-19. The preliminary figures for Q3 show only a slight recovery in the U.S., while you saw further decline. The trends, which we discussed in the second quarter figures 3 months ago persist or did reinforce themselves. And they are -- first of all, commercial real estate markets are split markets. There are splits in between prime properties, which sell with high demand and high-value resilience, high liquidity and low LTVs, such as prime office, logistics and residential, and where excess equity is abundantly available. And the other one, the other side of the equation is retail, hotel and B locations, where the valuation is expected to come down, where the yields are going up, where there's very little liquidity, but where bottom fishers have not yet really started and where even low LTVs rarely sell. Second observation. Downward valuations have not yet started on a broad scale. Debt valuers start removing their uncertainty clause, meaning that valuations become binding, and banks and investors will have to look at revaluation needs on portfolios to reflect market and segmental trends. And on individual engagements with individual valuation driven by tenant situation, location, property type and so on. We will see probably more of that in the fourth quarter to come and the first half year 2020, when risk rates are expected to increase. PBB has already reflected and anticipated the expected repricing by removing the flattener from its stage 1 and 2 provisions, property price calibrations and applying additional RWA changes to the U.K. and U.S. real estate finance portfolio. Third point, more specifically on Continental Europe, the repricing, which we observed in so far focused on retail sector. Obviously, it appeared to be largely resilient throughout, but on the back of very low volumes. Yields are expected to increase slightly in most continental markets and logistics and residential assets are stable so far or see even increased pricing. On Germany specifically. I mean, one thing which is noteworthy, the German market has dethroned the London markets being the #1 destination for international investors now since a couple of quarters. The #1 investment place is Berlin as a destination and evidenced by the number and the size of transactions. Office properties in Germany and the prime locations are very expensive based on historic measures, on the historic prices with record low yields. Yields are expected to stay on low level, driven by continued low interest rates despite an increase in vacancy and deal activity investor sentiment towards retail property is and remains down, except food-based and before big box assets. Therefore, in that field, we will see yields to increase notably for shopping centers. And the last word on the United States. We see weaker trends for the office and retail sectors, counteracted by strength on the industrial logistics and apartment sector, while yields generally stay unchanged compared to last year with the exception of hotels. So to sum it up, in the course of COVID-19 pandemic, the commercial real estate market environment has clearly changed. Investors demand and take-up has overall decreased hotel and retail, except food retailing are particularly affected. However, the overall liquidity is still high, and we still do not observe significant reductions of market values all over the place, but we expect the broad weakening still to come until end of the year or early '21. And when economic impact will become more apparent. Now there's a bit more detail on the following 2 pages. I will go over that very briefly. That's -- I mean, the first one on that Slide 8 is more in-principle description of the interlinked nature of 2 forming events and developments, and the first one is corona crisis as a cause for severe economic downturn. We never had a 7% GDP downturn in Germany since ages. And the second one is corona as a catalyst or accelerator of structural change. Both together make it exceedingly difficult to arrive at a reliable prediction of a future picture of commercial real estate markets, so to speak, of the new normal, which we would expect in 3 or 4 years time to come. The impact of the -- that's sort of a natural conclusion, the impact of COVID-19, which will be predominantly depending on the strategy and duration of lockdown measures and state aid measures. There will be unemployment in solvencies, which will increase the effect subsequently on property values and cash flows. Moreover, COVID-19 will most likely be an accelerator of structural changes to the real estate sector, especially arising from the new hygiene and social distancing standards, but also an acceleration of digitalization, i.e., e-commercial online shopping and sustainability developments, which will clearly have far-reaching implications for the real estate sector. And the structural challenges, as just mentioned, and sort of go over that in the short cost manner, the state transformation of the retail sector towards e-commercial online shopping is very visible. The increasing trend towards working from home is a determinant factor. Working at office is more influenced by social distancing standards and by the idea that office becomes a team meeting place with special requirements and increasing environmental requirements, as I just mentioned. Against these developments, we continue to run a highly selective and conservative business approach with a strong risk monitoring. And as already said several times before, we go for quality and adequate risk return before we go for volume. As such -- and that's the repetition of earlier comments, as such, we focus on prime A locations, on top sponsors, professional real estate investors. We have high expertise and our experienced low leverage lending, long-term stable cash flows, tenant quality and lease rollover and solid covenant structures, which mitigate the risk of future volatility in property values provides for solid risk buffers, allows for early action and increases the commitment of investors and sponsors in the transaction, and thus, the willingness to inject more equity if that is needed. Currently, still no new loan commitments with regard to hotels and shopping centers being extended. Retail is only done on a highly selective basis with focus on neighborhood shopping high street retail, if at all. Now I leave the section on retail and hotel to your reading or perhaps to later discussion, and would move on to Slide 11. As always, NII, general admin expenses and risk provisioning are being discussed on the following pages. But a few comments are being allowed for the relative line items as follows. Fair value measurement, which is positive with a full EUR 1 million in Q3. After significant COVID-19 pandemic-related spread widening in Q1, we have seen some rebound in the last 2 quarters. Fair value position is predominantly -- as we have already discussed, predominantly related to German states, only small position goes to Italian sovereign and to supranationals, so the small rebound or some rebound from the impact, which we did see in the first quarter with minus 17. Net income from realization is, as expected, significantly lower year-over-year, with EUR 4 million versus EUR 15 million in Q3 and EUR 20 million versus EUR 31 million for the first 9 months, mainly reflecting significantly lower prepayments. Net income from hedge accounting is positive at EUR 6 million, which stems from a EUR 5 million one-off due to the conversion of reference rates into the euro short-term rates. It's a more technical aspect. Income taxes -- and that comes with the loan loss provisioning. Income taxes are at minus EUR 35 million, higher-than-expected tax rate, mainly driven by deferred taxes, which is caused by the nondeductibility of increased risk provisioning under German tax rules. The NII, and I'm on Page 12 now. NII remains robust with plus 8% quarter-over-quarter and 13% year-over-year in Q3. In my view, quite an achievement, even though including some supporting elements from the TLTRO. In addition, we continue to benefit from floors due to the low interest environment. All in all, this results in 4% increase in NII on a 9-month basis despite slightly lower average strategic real estate financing volume, a further rundown of the nonstrategic part of the portfolio low contribution from the equity book. It's profiting from the contribution on the liability side. The average cost of Pfandbrief are further down, senior unsecured. There's more preferred and less nonpreferred, which helps on the average funding cost. And there is, of course, the effect from the TLTRO 3. Now a few more details on risk provisioning, which is so far, in line with our expectations. Risk provisioning in the third quarter is influenced by 3 factors. First of all, we have stage 1 and 2 releases according to model methodology. As time goes by and we move from the very low end of GDP projections to better GDP figures. The model starts releasing GLLP general loan loss provision and the release amounts to EUR 34 million for the third quarter. As a counterbalancing measure, we have removed the flattening effect of commercial real estate prices within the model, which were applied in Q1 and Q2, and applied a one-notch downgrade to all commercial real estate exposure in the U.K. and the U.S., both effects adding up to additional short-term -- to the additional stage 1 and 2 of minus 17. That, in total, gives a release of net EUR 17 million, to be said against further increased stage 3 provisioning for the U.K. shopping center by another EUR 31 million. That adds up in total to risk provisions of EUR 14 million, and there are a couple of important points to consider, which I want to share with you. First of all, we left the scenario assumptions unchanged. We believe that a minus 7 GDP movement in Germany for 2020 versus a 5.5% forecast by our government a couple of days ago is a sufficiently conservative calibration and caters for the also renewed lockdown, which is a partial and targeted lockdown in most countries. And in our view, does not come with the same impact on GDP as was assumed in the first wave. The single loan loss provision stage 3 pertain to already provisioned U.K. shopping centers, clearly a reflection of the structural weakness, which we have already discussed a couple of times. Structural weakness of the U.K. market in this asset class being over retailed. Overall, yield for U.K. retail properties presently has gone up or has been quoted at 7% plus. For shopping centers, probably a little bit higher. We have calibrated now our valuation on the basis of more than 10% yield, assuming that this will give us ample buffer for the future. We have no other new stage 3 cases with LLPs except for one hotel in London, which we did report already in the second quarter figures. The actual LTV of that engagement is just below 70% after actual revaluation a couple of days ago. The transaction shows a well filled cash deposit from cash traps and service the loan until the end of next year. This just as an example and a clarification. With the removal of the flattening effect on estimated commercial real estate price development for 2020 to '23 for our stage 1 and 2 model, we have no management overlays or other relief measures in place, i.e., negative GDP and price calibrations are fully reflected. On the contrary, given the uncertainties, especially on the Brexit discussions, we applied more conservative measures by downgrading PDs on the entire U.K. and U.S. real estate finance portfolio on a precautionary basis. The P&L effect was mentioned above. The RW effect is included in the 3 quarters -- quarter 3 computations. As before, we applied all necessary stage migrations as prescribed under normal circumstances and made no use of corona-induced counting easements. Still, this quarter, there were almost no migrations from stage 1 to stage 2 and no migrations to stage 3. With that, loan loss allowances increased from EUR 201 million to EUR 216 million, there, 55% or EUR 119 million attributable to model based general loan loss reserves in stage 1 and stage 2 and EUR 97 million allocated to single loan loss provisions. With that, we almost doubled the real estate finance coverage within the last 12 months to now 69 or almost 70 basis points. Given the overall uncertain state of affairs, we are likely to conservatively increase that figure to 75 basis points in Q4. That compares with approximately 80 basis points loan loss allowance, which is benchmarked in the market according to some IMF study recently on unsecured corporate loans. We do almost the same figure against the fully secured mortgage-backed exposure. Just for the sake of completeness, stage 3 coverage ratio now stands at 20%. Always worth to mention in this context, the ratio does not take into account the additional collateral. Page 14, I leave that for your reading. Most selling points we have already mentioned and move on to Page 15, operating costs. Operating costs remain under control, even though stable on a quarterly comparison, 9 months figure for GAE is slightly up for -- from 141 to 145. While admin costs are stable and stay stable, we had a couple of hefty charges for external consultancy or on regulatory and strategic projects and some extra expenses for substantial sanitary and technical corona related measures, which were balanced by substantial corona related savings marketing meetings and travels and so on. However, personnel expenses went up mostly due to internalization of some IT external consultants at lower average cost per capita, which were formally booked under admin and which should alleviate total GAE in the long run, while FTE increased moderately. The figure attributable to the general wage drift is very small and negligible. The write-downs of nonfinancial assets are mainly driven by scheduled depreciations, almost stable despite recognition of lease contracts as right-of-use assets according to IFRS 16 since mid of last year. I come here with -- to Slide 16 -- 17, apologize, Page 17 on the new business. After the peak of the lockdowns in April, May and significantly reduced market activity in Q2, while the European markets were down by 40% in transaction volumes, real Estate finance transactions in general picked up again in Q3 and so did we, with a new business volume that increased to EUR 1.6 billion, adding up to EUR 4.3 billion in 9 months. As I said, PBB sticks to its selective and conservative approach, i.e. continued low average LTV of 53% after 59% LTV on new business a year ago. No forced extensions, no new loan commitments and property types, hotel and real shopping centers. On the other hand, as I just mentioned, development of gross interest margin continued to show a positive trend towards larger than or greater than 180 basis points for 9 months average, which is a very pleasant and very good development. With regards to regions, we keep strong focus in Germany with 43% versus a 48% share in total portfolio. Why we strongly hold back in the U.K. with 7% versus 11% portfolio share. This compares, off the cuff, with EUR 400 million new business for the U.K. for 4 quarters to come or 4 quarters in 2020 as opposed to EUR 1.8 billion new business in the U.K. 3 years ago. CEE was rather successful, especially in Poland, predominantly in logistics and some mixed portfolio acquisitions and on some office prolongations. With regards to property types, main focus is on office with 50% portfolio share -- new business share into the portfolio counts for 47%, but significantly lower residential, 9% versus 18% portfolio share, which is not caused by over down-select by, but simply by the lack of investable opportunities. The process for those investments, for those properties are simply too high, the yields too low and the margins too low as well. As already mentioned, in retail in total, no new commitments since Q2, only prolongations of performing loans. Even though with less volume than precrisis, the pipeline remains good and supports solid new business volume for Q4 at continued elevated margin levels. Now Slide 18 reports on segment reporting. And again, that's a topic for your reading. And if you have questions on that, I'm more than happy to answer them later on. A few words on portfolio and portfolio quality. Portfolio quality remains remarkably stable despite COVID-19. Please bear in mind the following points, which serve as an explanation for the present state, but also for the developments, which we expect to come -- which we expect to see coming. Please bear in mind that, A, despite all negative outlook on commercial risk property prices, as reflected in the stage 1 and 2 models, actual price corrections across markets have not yet really taken place yet. And B, our valuations do not reflect actual market data of last, more than half of our portfolios with the bank since 3 years or more. But we do not revalue or increase market valuations just as prices go up, and they did go up over the last 3 years. We'll say market valuation of, say, 80 3 years ago, when the loan was acquired, might have moved up to 100 last year. The corona-based valuation may ask for a reduction of 15% against the corona levels. This leaves still a 5% plus over the issue valuation under which we have the loan on our books. And C, we regularly review the valuations on our entire book on a monthly basis. However, there is reason to bring down the valuation, be it loss of tenants, reduced cash flows, the state of the building and whatsoever, we adjust, and we will see more adjustments as values start observing more accountable price movements, which will move yields, presumably upwards, and thus affect the transaction discount rate and thus, the individual valuation. As you may imagine, we made it through the portfolio more than once to spot weaknesses and revaluation needs, and they have been duly taken into account. With all that in mind, the average LTV of 52% in our real estate portfolio needs to be read and needs to be understand and stable quarterly -- quarter-over-quarter. In all that, we believe that this provides a solid risk buffer with almost 50% equity or other money in front of us and provides a strong involvement and commitment of the investor or sponsor in the transaction. And that's one of the huge structural differences to the financial crisis 12 years ago. The negative change in expected loss classification is a technical reflection of a precaution PD downgrade of the U.K. and U.S. real estate finance portfolio, taking into account the uncertainties related to Brexit and more volatile U.S. markets. So far, the current picture looks quite okay. However, as I mentioned, some of the effects are still to come. Although every transaction has been recently revisited, not every property, as described, has recently been revalued. This comes either by specific incident, general review of sub portfolios as part of the regular annual review of the transaction. As a consequence of what I presented on Page 21, NPL remained on low level with no additions in Q3 apart from marginal technical effects. Workout loans are unchanged, only EUR 14 million. And at this point, please allow some further comments corona related -- on corona related measures. We are focusing on finding individual solutions, helping our clients over the current situations but still adamant to adhere to our strict risk standards. Thus, agreements in most cases require and include support elements from sponsor side, i.e. additional cash from the transaction or new equity. The measures that we agree to mainly relate to changes in covenant structures and extensions of amortization. All loans, where corona related measures were applied are performing. Now a couple of comments on Slides 23 and 24. I keep it rather short. On the funding side, funding volume is lower than previous year's levels. However, due to strong pre-crisis funding and continued use of market opportunities, we after all collected EUR 3.4 billion in new funding, which is still a sizable figure. In Q3, we have done EUR 1 billion, of which EUR 400 million account for private placements and one issuance, which I already mentioned in GBP 500 million. Inaugural SONIA-linked 3 years mortgage frankly, that's a difficult word, in September. This issuance was not quite successful, it's very successful in 3 ways. And I don't use very successful very often, so it was very successful. It is the first benchmark, frankly since January, issued in a very difficult market environment. Second point is it's been 3x oversubscribed with an order book of GBP 1.7 billion. And it did attract an attractive spread of SONIA, plus 38 basis points, which were tightened versus the initial guidance of 43 basis points and still includes the currency component. To say it in short, liquidity remains comfortable. As you know, in Q2, we also participated in TLTRO 3 with a volume of EUR 7.5 billion, of which EUR 1.9 billion were used to replace the former version of the TLTRO. The 3 key messages on Page 25. We have strong liquidity buffers in place. We are reaching out far into '21 with the liquidity, which we roll forward even under internal stress tests. And we are not exposed to corporate drawings down -- drawdowns on liquidity. We have attracting funding sources available. The Pfandbrief is a resilient funding source. The market is open and is able to deliver Pfandbrief as ECB keeps costs down. We see continued strong demand for private placements, both for Pfandbrief and senior unsecured. And the retail deposit funding channel was sort of reactivated 3 months ago and did react to that and is scalable. It's still expensive, but it's still a very strategic component to our funding mix. And needless to say, TLTRO 3 provides an attractive -- currently as low as minus 1% funding alleviation and flexible source of funding. Even though we have no imminent funding needs, we continue to take opportunities depending on the market conditions, and we are fully prepared to go-to-market for a green bond framework issuance if the market opportunity should be there. I would always say, the best thing comes at last, capitalization remains strong. A short overview on capitalization, was 15.3%. I think we show a strong picture. We do that against risk weights of more than 50% on our strategic portfolio and the real estate finance portfolio. It's a reduction of EUR 400 million. So the reduction in CET1 against end of June figure stems from a EUR 400 million increase in risk-weighted assets, mainly resulting from the reclassification effects of the precautionary PD downgrade of all business partners in the U.K. and the United States real estate finance portfolio, in the light of the already mentioned uncertainties. The fact that the increase looks modest is a reflection of our risk profile and a reflection of the risk calibration towards May 4 levels, which we have undertaken by year-end 2019 and which generally result into a higher stability of risk-weighted assets. Please bear in mind that risk weighting has increased to more than 50%, as I just mentioned, now again -- 50% now against a starting point more than 1 year ago of 25% in 2019. But also to be clear, we do expect some more risk-weighted assets increases to come in '21. Now I'm coming to the end of my presentation. And to sum it up briefly, the operating performance continues to be solid and with good results in Q3. NII remains robust, supported by positive effects from the TLTRO. Operating costs are kept well under control and in line with the overall economic improvement. Risk provisions stayed below levels seen in Q1 and Q2. Second point to make on this basis, we remain confident to achieve solid positive full year result in 2020. However, uncertainty with regards to risk provisioning persists, especially in the light of currently increasing infection numbers and lockdown measures. What do we expect for the fourth quarter? We do expect -- and we know that we have a good deal pipeline, which supports the expectation of solid new business volume in Q4 at continued high-margin levels, a bit under pressure though in the fourth quarter, but still very acceptable, as prepayments are expected to remain low, the strategic refinance -- real estate financing volume should slightly increase, and NII is expected to stay robust, prepayment fees to stay comparatively low. As usual, for the fourth quarter, we expect -- general admin expenses are expected to come out higher and risk provisioning, as I said at the beginning of my presentation, risk provisioning is the moving factor. It remains uncertain given the current development of COVID-19 pandemic. And easily forgotten, we continue to work on our investments, notably on digitalization such as client portal and digitalization of our credit process, which will hopefully be a larger topic to report on with the fourth quarter results. And finally, on dividend, we stick to our in-principle guidance on dividend and return to you with more new -- early next year alongside with the forthcoming ECB review. With that, I conclude my presentation, and thank you very much for your patience. I'm now happy to take your questions. Thank you.

[Walter Allwicher, Head of Communications]

[Operator Instructions] We have 3 questions registered right now. And that be Tobias Lukesch from Kepler; Mengxian Sun from Deutsche Bank; Johannes Thormann from HSBC. And we will work the questions in that order, but we would also require a patience, we will be back in a sec. Sorry for the delay. And so we will start with Tobias Lukesch from Kepler.

[Tobias Lukesch, Equity Research Analyst]

Yes. Three questions from my side, please. One is on risk provisioning, one is on capital and the third one on dividend. Starting with provisioning. So you mentioned, again, like in the last quarter's U.K. shopping centers with specific provisions. Could you please remind us again of the size of your U.K. thing center portfolio, how much has been touched in terms of specific provisions and what kind of coverage ratio do we see on that? And then on the remainder of that U.K. shopping center portfolio, how much of that is at risk to see also potential A rating migration and/or specific risk provisions potentially down the line? On capital, you mentioned the PD revisions, especially for the U.K., U.S. My understanding is Germany is quite a safe place. If we adjust for that NPL part, which will be healed soon -- you have literally no NPLs in Germany. So I was wondering, A, do we have to expect some PT revisions also for Germany? And secondly, you just mentioned RWA rating migration impact in '21. Could you outline a bit more what kind of migration you might see from that market valuation changes and what kind of magnitude this would have either in terms of RWA and/or the CET1 ratio. And lastly then, on the dividend, you have included it already at half year in the quarter 1 ratio. Nobody knows when the ECB dividend ban will be lifted. Let's assume the ECB will not do that in December, but only in March and April and allow for dividend distribution for -- in '21, basically. Given that the quarter 1 ratio is lower, given that you are highlighting for more pressure on the ratio, is it fair to assume that we should then rather think of a [ gun ] 2019 dividend and rather focus on the normal [ pour ], which is 50 plus 25%, i.e. 75% pour on the 2020 results?

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Thank you very much. Now on the risk side, the U.K. shopping center. Now the -- so the total magnitude of retail shopping centers in the U.K. resembles very much the NPL figure, which you see in the presentation, which is a EUR 500 million and take off non-U.K. exposure, which is about EUR 100 million, leaves you with EUR 400 million, which more or less is the size of the U.K. shopping center exposure, which we carry. And that's been significantly and substantially provided for. Now the other information, which gives you an idea about the coverage ratio, which we have on that, just on that, is that most of the single loan loss provisions, which we keep in stock of the loan loss allowances, which amount to almost EUR 100 million, 90% of that is related to the U.K. shopping center. I reiterate 2 things, which are important in sort of giving an overall picture to the U.K. retail business. U.K. shopping center is a special thing. It's to be seen differently from retail parks, big box centers, high street retail or whatever, the different yields and different values on these different subsegments. In any case, we monitor that closely. And in all cases, and that includes the NPL exposure, the cash flows are still there to serve debt service coverage. So even for those which we have provisioned, we still receive our interest and our amortization according to schedule. The structural weakness is mostly with the U.K. shopping centers. And of course, we also expect some elements of correction and value changes -- well, the changes for the rest of the segment or the subsegment, but not to the same tune. So presently, we feel okay with the situation which we have there. On capital, you were asking how the market valuation will work on risk-weighted assets for 2021. Now as you may safely assume, we do some planning, and we also do some risk-weighted asset planning, and we will comment on that a little bit further when we come to fourth quarter results. And up until now, all I can say is basically the same, what I said about the outlook or the guidance. There will be risks. There will be changes in PD and LGD, and there will be also some movements in risk-weighted assets. I do not only say that on a precautionary basis or forewarning basis, but I do say that, knowing that the sensitivity of risk-weighted assets, which we see on our book is very much determined by the already reached level of risk-weighted -- risk weighting on our strategic portfolio. The -- and small detour if you look at the total portfolio and note the fact that a substantial part of the total portfolio, i.e., value portfolio and public investment finance. So the entire public sector engagement is under standard approach and therefore, resilient for further significant changes that leaves us with a strategic portfolio. And if you look across the other banks, there are probably not so many who can dispose of risk weighting on their strategic real estate finance portfolio to the tune of 50% or so. So we believe we are well cushioned. We still will see some movement. How much of that depends very much on the further development of the crisis. Now on dividend, that's always a sensitive issue. Yes, we sort of capitalized the 2019 dividend. Technically, from an accounting perspective and a regulator perspective, that's disposable from where it is. But as I said before, and will continue to say we look out for ECB to give new guidance by the end of this year. There's a lot of discussion out. There are different streaks and movements within ECB and the national competent authorities on their views on what the banks should be. I remind you of the latest statement of Mr. Mersch, member of the directorate of the ECB, saying that if nothing goes wrong further or the crisis is not sharpening up. You would expect that banks will have sort of individual access to application of dividends on an individual basis as time goes by. So -- and everything else would be speculation. That's the timing, and that's the course of action which we take in the context of the fourth quarter results.

[Walter Allwicher, Head of Communications]

So next question comes from Mengxian Sun.

[Mengxian Sun, Research Analyst]

So 2 questions on my side. One on the revaluation on your property portfolio and the second one on loan loss provision. So to the property revaluation, you mentioned that the banks need to conduct some certain reprice or the revaluation on the property portfolio in general, NPV has already started the process. So my question is how much of the portfolio have you already finished with the revaluation? And how much left do you still need to look into it? And the second question is the loan loss provision. So given the uncertainty of the pandemic development and the further pressures on the commercial real estate market, I understand that it is difficult to plan how much to book in the risk positioning However, you have released $13 million risk provision in this quarter in stage 1 and 2. Is it not too early for the relief given the headwinds upcoming? And I will be highly appreciate if you can share some thoughts on that.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Now I'll start with the second point. We did not release EUR 34 million. We did release EUR 17 million. So we had consciously and in a very much intended way, counterbalanced the release, which otherwise would have been shown. But we're transparent on the parts and components of it. So it's visible on how we arrive at the figures. The lease, that's a technical or methodological matter. Basically, if I may try to explain that to you, the -- say, minus 7% in GDP is an average figure, which composed of quarterly GDP movements. And you have a first quarter of, say, minus 3%, you have a second quarter of minus 10%, you have a third quarter of some minus 5%, and you have a fourth quarter of plus 2%. And I don't do the math now because I just picked the figures, but that might average out at 7% for the year. Now if you move that figure one quarter along -- along the line, we'll take the second, third and fourth quarter and the first quarter of '21, you arrive at a different average figure. And that might not be minus 7, that might be minus 3 or minus 2. Now that positive difference works on the model and basically, from a technical point of view, releases single [indiscernible], the general loan loss provision, stage 1 and stage 2. Now we will apply a further, say, methodological change in the fourth quarter, and I would expect the stage 1 and 2 provisions to go back to the levels which we have seen already, so to balance out the release, which we have seen. But what we have done in the third quarter is entirely in line with the methodology and the mechanics how general loan loss provisioning works. So at the end of the day, by end of the year or by today, I would expect some 55% to 60% loan loss provisions on the general loan loss provision sides, on the stage 1 and 2 side, and that should sort of carry also into fourth quarter. So we -- very clearly, to be stated, we want to keep these kind of precautionary levels of provisioning. So I hope that was not too confusing. Now on the revaluation of portfolio, that's -- in a way, it's more difficult to explain. I've tried that earlier on in saying, we go through the portfolio on a monthly basis. And we do revaluations where revaluations are needed. So you can certainly assume that since onset of the corona crisis, the whole portfolio has been turned over not only once. So -- but you can't take from that on a one-to-one relationship, whether we have downgrades or upgrades because what we don't do is if we look at the signal exposure, and that was my example of a property, which values at 80, say, 3 years ago. What we don't do is revaluations to the positive. The -- so we may have a situation where we look at deterioration due to corona reasons of some 10%, 15% or 20% or whatever, and still may end up with the same value which we had before. And simply because of the fact that the revaluation measure is a pre-corona figure, which does not take into account the revaluation or the appreciation with the valuation has seen in the meantime since we booked the loan. So we will see cases where we booked the loan last year at 100 and now the discount will be 20. There will be cases of that. Still, does that matter for the risk calibration or for the RWA computation? Not necessarily. It depends on the overall situation of the loan. It depends on what the LTV covenants are. And if the LTV was 50% before, then a 20% discount on the value still leaves us with a comfortable LTV probably in the vicinity between 70 to -- 60% to 70%. So it's a very individual thing, and we do this case by case. There's no landslide revaluation measure that we say all retail, all hotel or Germany or U.K. has been downgraded by XYZ, and that's the new risk-weighted asset figure or that's the new LLP figure.

[Walter Allwicher, Head of Communications]

Next question comes from Johannes Thormann.

[Johannes Thormann, Global Head of Exchanges and Analyst]

Johannes Thormann, HSBC. I have some questions, please. First of all, regarding just the statements you made on certain property types. I'm a bit puzzled. You say now, no hotels anymore, but when the crisis erupted earlier this year, you still did some financings when the prices were higher. It sounds a bit countercyclical. What is driving the assumption? Because you would normally say, with current knowledge and lower prices, you should have better clarity on financing of some objects. Secondly, still struggling -- what has changed in your model-driven releases versus Q2 of this year? If you could prove -- because you say GDP assumptions are unchanged and other things are unchanged. So what has really changed in the model to release it? And the next thing is on dividend, talking -- you're, I guess, also talking to the regulator like other bank CEOs. And I'm a bit surprised that you think they will differentiate a -- or they will take a different -- differentiated dividend approach. Do you also think that they will do a dividend on the last 12 months' earnings? Have they signaled anything that they rather will be comfortable if you just do your base dividend and so on? And last but not least, probably a short update on your tax rate outlook for this year.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Now tax rate is sort of a bit of a disappointment because the German tax man does not acknowledge the tax deductibility of loan loss provisions. So the more we do on the conservative side to write up loan loss provisions, the more detriment we see or the more downside we see on the tax rate. So the tax rate, and I'm not sort of running ahead of our accounting and tax people now, but the tax rate will be probably closer to 30% than it will be to 20%. But that's the doing of the German tax regime. Now on dividend, the differentiation. I mean, first of all, I'll repeat myself, I would say something more definite when the fourth quarter results come out, first point. The second point is I quoted Mr. Mersch that in his opinion, and that doesn't necessarily need to be the opinion of the ECB, when it comes to it, the ECB should take a differentiated approach and should rely on the individual application of each bank to pay or not to pay dividends. Now to which extent that will be the course of action the ECB will take by the end of the year is completely open and is very hard to predict. I do participate in some of the discussions of Bundesbank, ECB and whatever, where the heads of the national competent authorities and the directors of ECB try to position themselves. And there are 2 different, diverging opinions, one on the more conservative side and saying that the crisis is just about to begin. And I will say, we can't take away from banks, as an asset class, the ability to pay a dividend as a major part of the value proposition, yes. So it is very difficult to predict. But as I said before, you may rely on the fact that the management Board of the bank is very much cognizant of the fact that shareholders like dividends. And that if the results are okay, the dividend should be part of the game. So on the model releases, I probably wasn't clear enough on that, on the previous question. It's -- we do not change anything in the model. The model assumptions are the same. The model assumptions are driven by 2 things. One is by GDP assumptions for the entire portfolio around minus 6.8%, 6.9%; and for German part of it at 7%. And that's something which we also put into the presentation or can give you more details on if you want to. That is very much in line or even on the more conservative side of the economic forecasting institutions, which regularly issued their forecasts. So -- and that was the same figure in Q2 as we had -- did employ or did use for the third quarter. The other element, which also is unchanged is our assumptions about commercial real estate prices. We have a very elaborate and detailed system in segmenting the different property classes and make predictions about 2021, '22, '23 price development for offices in the various countries for retail in the various countries and so on. And those assumptions, again, are unchanged. What changes is simply the time as time goes by. The average is moving. The average of minus 7%, as I said, consists of 4 quarters with different figures. And the more you move the quarterly average to the right-hand side and look into '21 figures with a plus of, I can't remember exactly, 4-point something, you come into more positive territory. And if you have a provisioning requirement at the mid of the year, at the time when the -- when you're sort of right in the middle of this minus 7%, and you come out with a provisioning requirement of EUR 84 million in this case. Then if you move to the better side of it, just due to tight [indiscernible], due to the timing side of the whole equation, you basically have less requirements to build provisions, i.e., you release provisions. Now that's a technicality against which we can't do much because otherwise, KPMG would come along and say you're not acting consistently on your models. And that is the reason why that target blurs a little bit the picture in terms of methodology. That's why we said, but look, we want to be more cautious on U.K. and U.S. exposure, and we take out the only softener, which we did employ so far, the flattening of the commercial real estate price development in the first year 2020. And that was the sort of counterbalancing factor of EUR 17 million. So it's actually a more conservative approach to how shall I get to proceeding to stage 1, stage 2 loan loss provisions. So in terms of methodology, we stick to the model which we have. It's a conservatively calibrated model. And we did certain things in order to counterbalance that. So last point, which was your first point on hotels. I'm not 100% sure whether I got the question right. We did have some -- we did write or did underwrite some hotel new business in the first quarter. That was pre-corona. But since then, we have refrained from any new engagement in hotels. If I understand your question right, you may say or you might have said that it would actually be a good time to underwrite hotels business because now you see what the resilience of the hotels is in the time of crisis. Now that presently doesn't go down well with our Chief Risk Officer, and we probably would refrain from that for some more time to come, that we see how the fallout of the entire thing will be. So we are actually quite happy that we said we have a very limited exposure of our portfolio to hotels business, and we want to keep it this way for at least the next months to come. Johannes, does that help?

[Johannes Thormann, Global Head of Exchanges and Analyst]

Yes. Understood.

[Walter Allwicher, Head of Communications]

We move on to Philipp Häßler from Pareto.

[Philipp Häßler, Analyst]

Yes. Philipp Häßler from Pareto. 2 questions are left from my side. Firstly, on the net income from fair value measurement. In Q2 and Q3, you have only recovered part of the negative effect from Q1. Maybe you could explain again what are the parameters that you would recover more of those fair value losses from Q1, on what parameters this actually depends? And the second question, only a short number question. The positive effect from TLTRO 2 in the net interest income in Q3?

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Sorry, the last one, can you say that again?

[Philipp Häßler, Analyst]

The positive impact from the TLTRO 2 in the net interest income...

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Okay. The positive impact of the TLTRO #3 is -- well, it's easily calculated. If you take the gross amount and supply, the benefit which you would describe us to realize through the participation TLTRO on a quarterly basis. That gives a, by and large, good impression. Another way to approach that is to say, between second quarter and third quarter, there's a pickup or uplift in NII, which to a large extent, relies on or is caused by the TLTRO, not all. So we have underlying trend in NII, which is positive. But the larger portion of that is attributable to the ECB funding. On net income fair value, the minus EUR 17 million in the first quarter. That's a part of the portfolio, which is classified as fair value to P&L and is one of the remnants and leftovers of the IFRS 9 reclassifications after introduction of IFRS 9, where a very limited portfolio was classified as not fit for at cost and needs to be run on a fair value basis and does not go through capital, but goes through the P&L. And that basically consists of 3 categories: one is German lender, selected titles, which we -- or the selected bonds, which we bought some long time ago. The second is a small residual on Italian state bonds. And the third one is supranational entity, which is accounted under that. Now -- and the way to recoup that is if the situation stays what it is, then with maturity of the bonds, we expect to get 100% back. But as that still is some years away, the rest needs to be done by market movements. And market movements have helped us over the last 2 quarters on the positive side. And as spreads are tightening again, we would expect to see some of the EUR 17 million coming back, and we actually saw EUR 4 million of that in the last quarter. Now if the spreads stay stable, we won't see a change. If spreads are tightening in further, we will see some positive effects. If it goes the other way, we see negative implications.

[Walter Allwicher, Head of Communications]

Next on our list is Jochen Schmitt from Metzler.

[Jochen Schmitt, Research Analyst]

I have 1 question on Slide 40 of your presentation. On your retail loan portfolio, could you provide a split into what you would classify as shopping center or shopping mall, outlet center, retail park, neighborhood center or high street? That's my question.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

In very broad terms over the entire portfolio of the 16%, half of that goes into shopping. And the rest -- the remaining part is high street, is big box or is retail parks. And the 8% of the 16%, and that's probably the background of your question, the -- I have to think about that. .Out of the shopping center, 20% originate from the U.K. and approximately half of that is in Germany. So that should give you, broadly a picture where we are.

[Walter Allwicher, Head of Communications]

That brings us to Christoph Blieffert from Commerzbank.

[Christoph Blieffert, Equity Analyst of Financials]

The first question is, what is the loan volume for which you have granted forbearance measures to your customers? If I remember correctly, this was some 1% of the portfolio in the second quarter.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

So -- I didn't get that. The...

[Christoph Blieffert, Equity Analyst of Financials]

The loan volume for which you have granted forbearance measures to your clients.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Okay.

[Christoph Blieffert, Equity Analyst of Financials]

And the second question is on new business margins. And here, any indications whether beneficial conditions you have seen in the third quarter will prevail in the fourth quarter and also in 2021? And the last question, could you give us some indication whether you have faced any additional NPLs in the first weeks of the fourth quarter?

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

No. The last one is the easy one, no. The second one, new business, can you repeat that again because I might not have fully...

[Christoph Blieffert, Equity Analyst of Financials]

I mean your new business margin was above 180 basis points in the...

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Okay. Margin developments. I apologize. Yes. Margin development in fourth quarter is, as I cautiously indicated, is a bit under pressure, and that is simply due to the fact that where we do business, everybody wants to do business now. We are going for prime locations, prime assets. And if there are transactions or where there are transactions, the bulk of the market will be found there. And therefore, we have tight competition again, which somehow is observed right in the middle of the crisis, but that's what it is. If you can manage to get your hands around the transaction for [indiscernible] in Paris, an office building with, I don't know, EUR 150 million in value, you will expect that all the big French banks are there, that large U.K. banks are there, the German Landesbanken are there, and everybody thinks it's a transaction to do. So if you want to go for the high-quality side of the business, you have to be flexible on the pricing. We try to avoid as much as possible, but we have to see that also, other banks are interested and keen to do some transactions. And as usual, it's been decided alongside with the price. So on a yearly basis, I don't think that we see much of a negative movement, but the fourth quarter in itself, on an isolated basis, margin-wise, will be a bit under pressure. Forbearance measures, I should answer this in 2 ways. First of all, we have no forbearance nonperforming unless we have classified this as NPL, and this is what is visible and is recognizable in our presentation. What we do is the kind of -- working with the client on any kinds of measures which help the client to bridge liquidity, if that is needed, under basically 3 conditions. One is we must be sure and we must be sufficiently convinced that there's a long-term business case, that the likelihood of the probability to survive is there. The second thing is we want to see engagement by the sponsors and investors with own money. And the third one is related both to regulatory as well as to IFRS standards. We'd like to do it -- or not like to, we'll do a restructuring in a way that the present value of the transaction is not affected. As long as the present value of the transaction is not changing, we have no provisioning needs and we have basically no forbearance needs. So does that answer your question?

[Christoph Blieffert, Equity Analyst of Financials]

More or less.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Okay. That's an honest answer.

[Walter Allwicher, Head of Communications]

More, more? Or more less? Thank you, Christoph. And one follow-up question from Tobias Lukesch.

[Tobias Lukesch, Equity Research Analyst]

Yes. One follow-up on costs, please. Kindly excuse me if I missed that, but I think it's fair to assume higher costs in Q4, but could you, again, please quickly elaborate whether or not you expect -- or how do you expect costs to evolve compared to 2019. So is it feasible in your view to be 2% to 3% lower or 5% lower? Will it be flattish? A quick answer on that.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

Let me answer in relative terms. Usually, as I said, Q4 shows more cost than the average run rate of the preceding 3 quarters. And we will have the same phenomenon in 2020. But we have 2 aspects, which also come into play. First of all, the [indiscernible], the value-add tax restrains -- the 16% restrains somewhat the possibility, the ability to prepone expenses and costs into the current year. That gives a bit of a cap on that side. And you should not take the uplift in between third quarter, fourth quarter 2019 as a blueprint or as an indicator for the uplift, which we would expect 2020, third to fourth quarter.

[Tobias Lukesch, Equity Research Analyst]

Okay. So I would read a low-cost development actually into it.

[Andreas Arndt, Chairman of Management Board, CEO & CFO]

It's a less significant cost uptick than we have seen last year.

[Walter Allwicher, Head of Communications]

We've worked through the list. And thanks, everybody, for joining us today. We appreciate your continued interest in PBB. It's very certain that we live in difficult times, and we hope that you all keep safe and well. And that going forward, we'll have a more prosperous environment for everybody to work in. Thanks again. Take care. Bye.