Interview: Cosme de Montpellier & Jens Hallén, Fitch Ratings
Feb 21st, 2013
Fitch Ratings analysts toured Scandinavia earlier this month and Nordic FIs & Covered spoke to bank and covered bond analysts at the rating agency about their views on Nordic funding, house prices and more.
Fitch said last week that Nordic banks have high wholesale funding needs, but that this is not considered a significant risk. Could you explain this position?
Jens Hallén, director, financial institutions, Fitch Ratings: Looking at the Nordics in general, the wholesale funding reliance is high for all the major banks. It is slightly lower in Finland, but for Norway, Sweden and Denmark the loan-to-deposit ratio is around the 200% mark on average.
There is clearly a sensitivity to this high wholesale funding reliance. It means that they are reliant on the perception of investors, both domestic and international, which we have seen can change. That, of course, is the risk.
A key mitigating factor is that domestic investors make up a quite significant proportion of this, which effectively we view as almost a captive investor base, for a number of reasons. Firstly, if we are looking at covered bonds, for example, they are highly liquid, highly rated, and the insurance companies in the region and the banks need to invest in such assets in local currency — the latter also for their liquidity buffers — and the alternative triple-A rated assets would be government bonds, but they are relatively limited in this region. Take Denmark as an example: the mortgage bond market is about four times the size of the government bond market, and shown to be at least as liquid in times of stress. We see that domestic investor base in the region as being stable, and that is supportive of the banks.
I should mention that here we are talking about the largest banks, the rated banks. Of course, in Denmark, for example, there is clearly a difference between the three large banks that we rate and some of the smaller banks that have clearly found it very difficult to fund in the wholesale markets.
The banks’ large liquidity buffers are a further mitigating factor.
What do the big domestic markets mean for covered bonds?
Cosme de Montpellier, senior director, covered bonds, Fitch Ratings: Regarding Sweden and Denmark, I would highlight the reason why the covered bonds markets are so large: the vast majority of mortgages in these countries are refinanced on the market, especially in Denmark where you basically have a one to one relationship between the mortgage and the bond financing it. So it is quite natural that these markets are so large — this reflects the way the mortgage markets have been functioning.
On the positive side, it makes the covered bond market in these countries very liquid, and it also makes the mortgage market quite standardised in the sense that most mortgages are eligible for covered bond refinancing. This is a positive, as it makes it easier to refinance them.
The risk we see generally is not so much the size of the market, but more the risk of maturity concentration. In both of these countries mortgages have very long contractual maturities — for instance, in Sweden, some mortgages have a 60 year maturity — and the bonds that finance them have relatively short maturities. In Sweden, the weighted average life of covered bonds is about three to four years, which is relatively shorter than in most European countries.
In Denmark, there is a large issuance of one year bonds which are refinancing the one year interest reset mortgages, and they are typically only refinanced at only a few dates each year. Notably, very large refinancings take place in January — with the auctions in November-December and settlement in January, and these represent a sizeable amount of the GDP — the total volume of mortgage bonds to GDP in Denmark is about 140%. And you have up to 20%-30% of GDP being refinanced at this time. This concentration represents a risk for us.
We see that issuers are now tending to refinance on different auction periods. They are moving towards four auction periods — at the moment you tend to have two or three active periods depending on issuers, and the idea is to have four — effectively to spread out the concentration risk. The key risk for us is the combination of maturity concentration and large sizes, versus the size only, which is not necessarily negative.
Maybe I should mention one more thing on the Danish market where the size is not an issue. What used to happen is that the fixed rate mortgages were refinanced by fixed rate bonds, on a pass-through basis, so the size of these bonds has never been a problem. For these bonds, you have no refinancing risk for the issuer. This used to be the majority of the market in Denmark — now it’s much smaller, probably less than 20% of the market.
The Danish system is also potentially facing challenges under Basel III, such as through the NSFR. Do you see pressures on Danish banks to change further?
Hallén: There are two things in relation to this. The LCR is also affected, because they used to be having the auctions within the last 30 days of the refinancing, which created a significant outflow that would not be covered by liquid assets. That is being dealt with by putting the auctions slightly earlier. It doesn’t change the system, but it changes the immediate refinancing of those bonds.
When we are looking at the NSFR, there are two things. I think the banks are looking at pushing out the mortgages to more than the current one year. When we are looking at the pricing structure for mortgages, we see banks trying to look at two to three years rather than just the one year.
It is difficult when we are looking at the new regulations, because of course with the LCR you also have the eligibility of covered bonds as a Level 2 asset, although also limited in terms of amount. What we are hearing from the regulators is that they are working on finding a compromise. It is probably too early to say whether they will reach a compromise or not, but when we analyse the banks we see the investment in Danish covered bonds as being stable and solid.
Do any of the mortgage markets stand out in terms of risks?
De Montpellier: There is one covered bond market that has increased more than any other among the Nordic countries, and that’s Norway, where the law was only passed in 2007.
However, you have certain things which are common to these countries’ mortgage markets. I would say that in general households are highly indebted. But you also have tax deductibility of interest, which means that it makes sense to see higher debt than the euro-zone average. You also have social security benefits that are typically higher and for longer than in most EU countries. Both aspects should be taken into account.
The second thing that needs to be taken into account is that the net financial assets position is quite different if you compare Norway and Denmark, which are the two extremes in the Nordic countries. You have a much larger net financial assets position in Denmark than in Norway. In Norway you see that the majority, 80%-85% of the wealth of households generally is in their houses.
The third aspect of these mortgage markets is that you have a lot of floating rates or interest resets in these countries. You have quite a high percentage of interest only, in Denmark especially, but also in Sweden, and a little less in Norway. All of this means that households are quite sensitive to changes in interest rates. Denmark is quite sensitive because of the high indebtedness of people and the high percentage of mortgages that are interest only.
At the same time, in Norway you have 100% of the market being on the standard variable rate that can be changed with a few weeks’ notice. And because most of the households’ wealth is in their houses, a change in house prices or a change in interest rates or both will impact consumption. So from that point of view consumption is more sensitive to mortgage interest rates in Norway than in the other countries. The worst case scenario we could have is that any impact on consumption would impact unemployment, which could impact mortgage arrears, which could mean that more properties would go on the market, and that could impact house prices. Although a remote risk, this vicious circle would mean house price declines would result in house prices declining further because of this sensitivity to consumption.
The regulators have been making various moves that appear directed towards dampening house price growth, by limiting LTVs, for example. Are their policy decisions sensible?
Hallén: Sweden was the first country to go out with an LTV cap, and that was also followed by a pretty deep discussion in Sweden about house prices, etc. We have since seen house prices stabilising, from 2010. This action and discussion was also followed by the banks tightening their behaviour in the mortgage market, looking less at gaining market share and more to profitability. That limited credit growth and has also impacted the house price increases.
I should probably add that when we look at Norway, at the house prices, this is clearly an area that we focus a lot of attention on. But I should also stress that our base case when we are looking at the banks is not for a significant correction to happen. Our base case is for a flattening of house prices. And we also think that serviceability of households remains quite strong. So when we look at the banks’ ratings, we probably see more of what Cosme was talking about in terms of the impact on consumption, and then an impact on Norwegian corporates, so an impact on banks’ corporates exposures will probably be the most significant effect.
De Montpellier: Regarding affordability for Norwegian households, borrowers in Norway are stressed with a 5% increase in interest rates, which is higher than in Denmark, for instance, where the borrower is stressed in terms of affordability with the highest currently available mortgage coupon, and assuming the loan is amortising (while Sweden is to a certain degree similar to Norway). So in Norway you do have a cushion.
House prices in Norway have increased by around 40% since the end of 2006, so are quite high. It’s clearly an outlier in the Nordics.
There are plenty of good reasons why the prices have increased: low interest rates; the high increase in wages, generally; tax on imputed rents was abolished back in 2005 and that alone could have an impact of perhaps 5% on house price increases; and you’ve got the interest rate tax deductibility. You’ve got plenty of things which explain why house prices have had to increase over the last few years in Norway.
What we are saying is that it’s probably overpriced today, but it’s probably not a bubble. There are very good reasons why prices have increased. We think that it is a bit overpriced, although maybe not as much as what the IMF is saying — they published a figure of 20% last year.
One of the examples you could give is if you are a first time buyer and are less than 34 years of age, you can get a mortgage of above 85% of the property value — which is the new LTV limit — if you have additional security — that’s a product proposed by most of the large banks in Norway. But if the 85% LTV is a constraint for first time buyers, then this is a sign that the market is overpriced.
We are revising our triple-A assumptions for house price decline in Norway. Although these are not the base case that we expect, but are stressed assumptions used for covered bonds rating purposes, we have increased them for Norway to take into account the risk that prices go down significantly.
Are the Norwegian authorities right to have concerns about high levels of assets being transferred to covered bond issuers? And what impact might any new regulations have on covered bonds?
Hallén: If I can start from the bank side, there are probably two areas that might be worthwhile mentioning. One is this transfer of assets from the banks to the covered bond vehicles, and whether that has a rating implication. When we are analysing the banks we look at the legal entity and if all of the good assets are effectively being transferred out, clearly that has an effect on what is left on the balance sheet. But that is not what we are seeing today.
Secondly, what we expect to see is an end to the growth that we have seen in assets being transferred and covered bonds funding. We don’t expect that to increase. We have probably reached a limit now where the banks will continue to fund a proportion of the new mortgages with covered bonds, but not for the percentage to increase that much further. From a bank rating point of view we don’t see the levels where we are today being a negative driver.
But clearly for any market there comes a point when so many of the good assets have been encumbered that that has an effect on the senior unsecured creditors.
De Montpellier: In these countries — in Norway but also in Sweden — you have an LTV cap for mortgages in the cover pools that is lower than the cap for CRD-compliant covered bonds — it is 75% and not 80% for residential mortgages. So if you combine this 75% with a house price decline you need to have a cushion to be able to maintain overcollateralisation levels. At the moment issuers have the cushions and it is not a key concern.
The Norwegian authorities are also looking at giving the specialist issuers limited banking licences to be able to access central bank funding directly. Is that something that would be positive?
De Montpellier: We need to look into that in more detail, but if that means that they could, for instance, repo their own covered bonds under their own name without the assistance of the parent bank then that could be positive regarding liquidity mitigants.
Are the recent changes to Swedish covered bond regulations significant?
De Montpellier: Our general impression is that they were not key changes, but more the implementation into regulation of rules that were already applied in practice. One of the examples would be that issuers have to stress the impact of house price declines on the cover pool. Obviously they did it before; what they have to do now is to check it for pre-defined levels, so 5%, 10%, 15%, etc down to 30% declines. So that’s now in the regulation, but it’s something they were checking anyway, although they were not necessarily reporting on this. That’s something which is positive.
We understand the Swedish regulator has increased the minimum requirement for swap counterparties within the regulation. Although this is not a reflection of our criteria, we would stress that, generally, we certainly wouldn’t want regulations to put in place our rating criteria for the very simple reason that we can change our criteria. Also, it could make it easier for a counterparty to be in breach of the regulations if the criteria are too conservative. For instance, we could end up with a situation where the minimum criteria are too high and no counterparty is eligible. Our position is that it is important that regulators look into hedging and counterparty risk generally, which is something we look at very closely as part of our Discontinuity Cap. From that point of view, the change is a positive. However, putting in place hard criteria in the regulation could sometimes be impractical.