Moody’s: Italians respond to pressures

The NPL clean-up underway in Italy is positive, but nuanced and differs across the banking industry, according to Moody’s Alain Laurin, associate managing director, and Edoardo Calandro, senior analyst. Here, they outline how this could affect the banks’ ratings, as well as the impact of the redemption of retail bonds and corporate depositor preference.

What is the best way for Italian banks to tackle NPLs, and a realistic timeline?

Alain Laurin: This is, needless to say, very topical — and Italy is not the only country in Europe having to do a clean-up. The European Banking Authority recently reported that in eight European countries NPLs constitute 10% or more of loan portfolios, which gives a sense of the scale of the problem. The EU and the ECB consider this issue to be of systemic importance. They believe there is a need to tackle it and, as Danièle Nouy put it, the work has to be done now — now, because the region’s economy is doing fine. There are variations in growth, of course, but certainly everybody wants to get the job done before the next real downturn. So there is some sense of urgency, all the more so since the role of intermediation played by banks is to a certain extent impeded by the high level of NPLs.

And of course it is critical in Italy, given the size of the Italian banking system and the size and importance of the Italian economy in Europe. There are therefore many different streams of measures being taken by the official sector to push this forward, and at the same time banks are very eager to do the job. The economy is better, so it is easier for banks to take action, and by the same token it is easier for the supervisory authority — the ECB — to put pressure on Italian banks to reduce their NPLs.

The ECB has clearly conveyed to banks that are under its supervision the message that they should reduce their NPL ratios, and we note that large Italian banks target below 10% NPL in the foreseeable future, if not yet achieved. Of course, it might be a challenge in different ways for different banks, but we believe that for many banks it will be done and can be done in one, two, or three years, depending on the case.However, this is not the only step — the first step is to go below 10%, and maybe later on the ECB will continue putting pressure on the banks to do more. We don’t know exactly when, but that is certainly the game-plan: after all, 10% is still about twice the average for the EU!

An important factor is the impact of IFRS9. The introduction of IFRS9 has propelled many Italian banks to accelerate the clean-up. This is because the mechanics of IFRS9 and the prudential framework prevailing here in Europe make it possible for banks to increase the level of provisions within the so-called stage three bucket — the worst category of loans — with very limited impact on Common Equity Tier 1 capital ratios, because there is a phase-in of five years. This provides an attractive context for banks in Italy to further write down their bad loans. Actually, we were not expecting great additional provisions in stage three buckets; we were expecting more provisions under stage two, which has been less the case. There are two different sides to the coin here: one would be to say, previously under IAS39, the bad loans were not sufficiently provisioned; the other would be to say under IFRS9 there is more flexibility to set aside provisions against NPLs and to increase their coverage — and that is exactly what Italian banks did, and we consider that positively.

Edoardo, perhaps you could elaborate on the different banks’ distance to 10%.

Edoardo Calandro: We see Italian banks going at different speeds, because the Italian banking system is made up of many banks. We have on one side of the spectrum banks such as UniCredit, which already started with a big clean-up more than a year ago, and it is continuing to dispose of problem loans, and we expect them to have a ratio below 8% by 2019. Or Intesa Sanpaolo, which took advantage of the adoption of IFRS9 and is accelerating its reduction of problem loans, and the bank expects to go below the threshold in a few quarters.

I also need to note that there are some smaller banks like Credem that always had a very prudent risk approach — it has a problem ratio of around 5%, similar to throughout the crisis.

So on the one side we have these banks that have always been below 10% or they are quickly getting there through disposals.

On the other side, we have banks that are still taking longer and have less aggressive plans. The first names that come to mind are MPS, which has its own story but still suffers from a large portfolio of problem loans, or Banco BPM, which also suffers from legacy issues, mostly coming from the former Banco Popolare.

So, different speeds, different starting points, and certainly different pressures and paths in the coming quarters.

By how much do Italian banks need to clean up their portfolio to have investment grade BCA ratings?

Calandro: We don’t have specific targets or thresholds for NPLs that banks need to meet in order to get investment grade BCAs. We have of course an holistic approach to assessing asset risk, and the level of problem loans is one of those factors. It is one of the first factors that we mention, but also important to us are the level of capital, the level of profitability — and of course the funding and liquidity. In order to have a cleaner bank, the bank also needs to have sufficient capital and a good stream of earnings — the three things really come together for us.Having said that, I should also note that Italy is a country with a sovereign rating of Baa2, on negative outlook, which indicates a difficult operating environment in which a baa3 BCA is certainly challenging to attain. Amongst the larger names, we already have at that level Intesa Sanpaolo. It has not yet reduced its stock of problem loans below 10%, but it is on its way there — especially because of IFRS9 considerations, as alluded to earlier — while it is also a bank that has a good level of capital compared to the rest of the system, and a good stream of earnings — the bank has basically been profitable for its entire existence since being born out of Intesa and Sanpaolo in 2007. We also have Credem at Baa3 because of its exceptional problem loan ratio level — or rather, exceptional by Italian standards, even if it’s an average level for Europe. And just recently we assigned first time ratings to Mediobanca, including a BCA of baa3. That is because, thanks to its diversified nature and a different business model to most other banks, it has structurally a lower level of problem loans than a commercial bank suffering from legacy issues with SMEs, for example.

Finally, there is UniCredit at Ba1 — below investment grade, but we have a positive outlook on the ratings, indicating that the measures to clean up the balance sheet, to restore profitability and to improve capital are going in a direction that will improve the credit profile.

But, again, there is not a perfect correlation between the stock of NPLs and the BCAs; it is more seen in the context of capitalisation, earnings stability and earnings generation, and of course funding and liquidity considerations.

Which measure to clean up balance sheets will be the more effective: EC Pillar 1 or SSM Pillar 2?

Laurin: I don’t want to be too provocative, but I would dare say that neither of them is very effective, for very simple reasons.

The first one is that the so-called EC Pillar 1 is a regulation by nature, and for this regulation to be effective it has to be enacted, and it is still a work in progress, so we are going to have to wait until the EU agrees on a framework. Even if we assume they agree and they put this piece of legislation into the CRR2, the scope of the current EC legislation is about new loans only, which means it doesn’t touch the outstanding stock of NPLs. In this respect, this EC legislation makes no difference at all.

Does it mean it is useless? Certainly not. It will be useful for the treatment of NPLs in the future; it will work as a backstop measure if the accounting framework doesn’t do the trick, and that will possibly be an effective tool to avoid the accumulation of NPLs in the future.

Let’s now move to SSM’s Pillar 2 tool. Pillar 2 under the recently-published guidance of the ECB is certainly more effective in the sense that the scope is different. The scope of loans targeted by the ECB is new NPLs, meaning an NPL coming from the outstanding good, performing portfolio, when it turns sour. So these new ECB guidelines would apply to such NPLs — but with a catch, which is that it will only be applied in three years. The minimum provisioning requirement for these new NPLs (secured by collateral) is set at 40%, and will increase over time until full completion of the provisioning, i.e. in seven years. We now expect the ECB to publicly express its view as to how the NPL stock will be addressed.

But if we are now referring to the Pillar 2 measures that the ECB can deploy every day, certainly, that instrument is effective, because the ECB can say at any time: “The provisioning may be insufficient, I do not dispute the accounting behind that, but to be prudent I will impose a capital add-on.” And that’s the trick: the ECB has many instruments at its disposal to force banks to take action, and to do the clean-up.

In this respect Pillar 2 is a very powerful instrument — again with the caveat that it is case-by-case, which means the ECB cannot tell every bank to do certain things in the same way, because it would be akin to imposing regulation. That is why when the ECB first published its addendum many people complained, saying: “No you cannot do that. You are trying to impose a Pillar 1 measure on us.” And the ECB had to take a step back and clarify the point, explaining that the so called addendum is a Pillar 2 measure. But although the addendum as a Pillar 1 instrument was rebuffed, what was not rebuffed at all is the ability of the ECB to impose certain things. That is why some banks who were claiming, “we won’t do this, we won’t do that”, have sometimes changed their mind. I suspect they changed their mind for a reason, that reason being pressure from the supervisor.

How do you estimate the adequate level of capital required against NPLs?

Laurin: To be precise, capital is not supposed to directly cover NPLs; it is supposed to cover unexpected losses. The expected losses, i.e. the losses in the portfolio, are to be addressed by means of specific provisions under IFRS9. If there is a gap — if the supervisor believes the provisioning is not sufficient — the supervisor will impose a deduction from capital. At Moody’s, we may express some doubt as to whether or not the bank is adequately provisioned against these NPLs.

That being said, if we do consider the broader picture — that there might be variations around the amount of losses in the NPL portfolio — banks should account for this risk. This is why the NPL portfolio not only requires provisions against identified losses but also capital to account for the unexpected losses that may arise. This issue, which is addressed in the EU regulatory framework, is currently being pursued by the so-called EBA guidelines, which are not yet finalized.

UniCredit recently said the cost in capital of these measures for it is estimated to be 90bp of CET1, which is a big number. This leads me to the conclusion there are certainly variations between banks in the manner in which they construct or conceived their internal models, and that is certainly the objective of the ECB review project (TRIM), to fix that problem. You can expect the ECB to impose more capital on banks through modelling changes, without waiting for Basel IV.

How will retail senior reimbursement impact LGF and your ratings?

Calandro: To clarify: we are talking about those actual retail bonds that in the past pure retail clients underwrote in branches, which were perceived by clients as savings products rather than investment products — so we exclude everything that goes into the private banking portfolio, or for affluent or more sophisticated clients.

As these instruments mature and are recycled into deposits, or are put into wealth management products, we see that as neutral for the BCA and the assessment of the funding profile of the bank. That is because we always considered them a stable source of funding, so we never had specific concerns about the rollover. Now, they are not rolling over, but they will remain with the bank as deposits that we believe are sticky, or they are being directed by the bank mainly into wealth management products, which is actually positive for banks because it increases commissions and increases profitability — which, as I said earlier, is another point of concern for us. So in terms of the standalone assessment, it is broadly neutral.

Nevertheless, regarding the impact on ratings, as assessed under our loss given failure analysis, this is negative, because these bonds are supposed to be bail-in-able, but as they mature they are recycled into more senior retail deposits, hence the volume of bail-in-able debt falls, reducing protection for senior bondholders in a resolution scenario, and reducing protection for wholesale depositors, too. As a matter of fact, in 2017 we had a few downgrades for this specific reason, because, especially for smaller institutions that were not tapping the wholesale market, the only stock of bail-in-able debt was these retail bonds that were quickly decreasing, and that had a negative impact on Italian banks’ ratings. And we still have several negative outlooks on Italian banks’ ratings also for this specific reason.

What is the impact on your Italian bank ratings of the upcoming corporate deposit preference?

Calandro: We have already incorporated it into our ratings. There was a clear framework in place a couple of years ago, agreed by Parliament in December 2015, and we knew that the cut-off date would be January 2019. So between these dates we took two large rating actions on Italian banks, to already embed the corporate deposit preference in the ratings: we upgraded 16 deposit ratings by one notch and two by two notches, and we downgraded five senior unsecured ratings by one notch. With the introduction of the preference, in a resolution scenario the senior unsecured will no longer benefit from sharing the losses with junior deposits, hence the five downgrades. But in the same resolution scenario, symmetrically the junior depositors or corporate depositors benefit from a clear protection provided by senior debt, hence the upgrades.