Sunday, February 24, 2008

Spain's Balance of Payments, Bank Funding and Covered Bonds

This post constitutes a first attempt at making an assessment of the extent of the damage which has been sustained by the Spanish economy as a result of the global credit crunch.

The first thing to get clear (and very clear) in our heads at the present moment is that the economic correction which is currently taking place in Spain is very unusual one in terms of what we have become accustomed to in developed economies in modern times, since the transmission mechanism for Spain's problems does not run from a problem arising in the real economy (a correction in house prices for example), nor does it run from an attempt by a central bank to "burst" some sort of perceived asset bubble or other (Trichet and the ECB's tightening of interest rates), but rather the mechanism operates via a direct blow-out in the cylinder head gasket of the global financial system, a blow out which has produced an immediate and direct change in global credit and lending conditions, and in the level of risk appetite which prevails in the securitised morgatges/covered bonds sector of the wholesale money markets, and it is this change which is now making its impact felt on the real economy in Spain, with the actual and present danger that these negative consequences for the real economy may then be fed back into the financial sector, in the process creating some kind of ongoing lose-lose dynamic.

As I say, such a phenomenon is certainly unusual in a modern European context, although some may wish to point to parallels with what happened in Japan in the early 1990s, and the subsequent "lost decade". I wouldn't go so far at this point as to suggest that Spain is facing a lost decade, although the situation is very very serious (as I hope to show in the charts that follow), and at the very least Spain now faces several "lost years" and a massive macroeconomic structural adjustment.

Now a quick browse through the back posts on this blog should give any interested reader a reasonable resume of the current state of Spain's key macro economic indicators, and so I am not going to dwell on this part here. There are approximately 500,000 unsold homes standing on someones books in Spain, and the Spanish banks now attend the weekly liquidity auctions at the ECB to raise some 48 billion euros of funding (a figure which is double the 24 billion euro number they needed before the summer).

However, as Jean Cluade Trichet points out, "there is no question that the Spanish - or any other - banks are being bailed out" at this point. This, like so many of Trichet's statements is entirely true. But it is not the whole truth. That is, there is a bigger picture. And herein lies the problem.

It is, for example, also true that the ECB has not changed its rules to accept mortgage backed securities (like the US Fed had to under pressure last August), but this apparent constancy in rules also stands alongside the fact that a banking system which didn't need much recourse to the facility in question has now doubled its use of it, and in a very short space of time.

It is also true that Spanish banks were not allowed by the Bank of Spain to set up special purpose vehicles to finance their lending, but again, they set up the cedulas hipotecarias to find another way to do something which, at the end of the day, is not that disimmilar. The argument against SIVs is that off balance sheet lending is likely to be more risky, but in a certain sense (and as I try to argue and explain in my post here on the cedulas system) the cajas regionales have played the role of off balance sheet entities for the principal banks, and since we still don't know how far down the value of the entire Spanish mortgage pool is going to fall, we have no way at this point of making a valid assessment of the level of risk that has been actually taken on board.

It is also the case that the level of mortgage defaults in Spain is at this point comparatively small, but then again the whole process is only just starting, so it is far to early to say at this point whether or not cover will prove adequate in the longer term, but then again, the problem for the Spanish banking system may not originate in defaults in the first place, but rather from a perceived rise in their risk level if the value of the entire pool of mortgages on their books declines significantly.

Spain's External Balances

My main point of focus here is going to be on Spain's external imbalances. Essentially Spain has been living on a large external trade deficit in recent years. This deficit has produced a continuing deficit in the current account, and this deficit has effectively been balanced by attracting the funds from outside Spain to finance mortgage lending in the recent Spanish construction boom. It is this nexus of relations that I think it is necessary to have clear, otherwise the greater part of what now happens next will catch you completely unawares.

Now as I keep pointing out, another of the very specific characteristics of the way in which the property boom was financed here was the ability of the Spanish banking system to provide very low interest (variable) rate mortgages. Curiously, many commentators imagined that this (variable) component was the greatest rsik element in Spain. That is, they imagined that it was Spain's mortgage borrowers who were assuming the greatest part of the risk here. They were wrong. The risk at present has almost all been inadvertently assumed by Spain's banking system, and this decidedly odd situation has arisen (I'm sure this was never the intention) due to the recourse of Spain's banks to the widespreased use of securitised (or covered) bonds - the so called cedulas - and the provision of variable mortgages at very narrow margins (lets say around 0.75% or 1% over 1 year euribor). Now the banking system considered they were onto a sound bet here, since they in there turn could borrow (thanks to the investment AAA grade assigned to these bonds, making them virtually equivalent to government paper)at very fourable rates themselves (normally three month euribor), so they thought they were on a very safe bet. Again they were wrong, since it is just this very global repricing of risk appetite I mentioned earlier, and the reassessment of the AAA rating which was assigned to their mortgage bonds which goes with it, which has produced the problem.

Oh, yes, and there's a third little detail which makes all of this just that little bit worse. The different term structures of the lending and the borrowing. Basically the Spanish banks, and especially here the regional cajas who undoubtedly have the lions share of the problem, borrowed short and lent long. The majority of the mortgages issued between 2000 and 2007 were for between 20 and 30 years. Indeed during the last two years of the boom it even became fashionable to offer mortgages over 50 years, so sure did the banks feel of themselves.

But they borrowed short. Not the very short liquidity type borrowing we are increasingly seeing Spanish banks resorting to at the ECBs weekly money auctions, but rather the lions share of the borrowing, which was done using cedulas, and normally over a ten year term. That is to say, that while the vast majority of the mortgages issued will still be outsanding come 2025, almost all of the bonds which go with them will need to be refinanced, between 2012 and 2017. And here is where we hit the snag, since the money markets which the Spanish banks need to do the refinancing are currently closed to them. These money markets can of course be reopened, but at a price (ie the price of a risk premium), and that is really the bigger half of the snag, since the debtors are on "variable" mortgages which are effectively fixed (at 1 year euribor plus something, euribor can of course go up, but it can also go down, as I think we are about to see in the next moves at the ECB), so the borrowers, it turns out, really do have a yardstick that lets them know what they are into. This is not the case with the lenders though, since while we do not know what eventual risk premium will be built in to fund Spanish banks (this in part depends on how far and how fast property prices fall, and how much difficulty they have maintaining their mortgage pools), we do know that the euribor 3 month days are over, and we could even contemplate the possibility that if bad goes to worse, and even worse, and then worse again, then they could be asked to pay even more than they are recieving from their mortgage paying clients!

And the amounts of money are not small. One good recent estimate put the total quantity the banks will need to "roll over" in the space of about 5 years at some 300thousand million euros.

OK. Now lets look at some charts. First off we could take a look at the level of incoming bond purchasing funds into Spain over the last several years (thanks to the monthly balance of payments data made available by the Bank of Spain).

Now as we can see on a rough and ready basis the bonds boom really took off sometime in 2002, and it came to a sharp and rather unfortunate end in the middle of 2007 (I would say on or around the 9 August). In any event the height of the boom was in 2005, 2006 and the first half of 2007. Now before going any further with my line of argument, lets just look at one more piece of rather important evidence, the refi interest rate set by the ECB (and which of course regulated 1 year euribor) and the rate of inflation in Spain.

And as we can see here, the take of date for the bonds boom is hardly coincidental, since it more or less coincides with the arrival over a sustained period of time of negative interest rates in Spain, a fact which to a large extent explains the intensity of the housing boom, since under these conditions borrowing money does not seem especially onerous, and indeed it might even be considered illogical not to do so. Certainly Spain's very low level of personal savings in recent years can only be properly understood in these terms.

Now if we return to the task in hand, lets take a close-up look at the most recent years for those bond inflows.

As we can see, the last few months is not the first time that these flows have "wobbled", but this time the wobble is much more sustained and, since the wholesale money markets are at this point effectively closed to the Spanish banks to raise money in this way, there are good reasons for imagining that this time the change will be a more permanent one.

Now, why, apart from the implications for the banking system, is all this so important for Spain? Well lets now take a look at another chart, the one for the current account deficit. And as we can see, the growing consumer boom associated with the construction one had a long term and pretty disastrous impact on Spain's trade balance (and in particular its energy component, all those extra houses use energy remember). And this deterioration in the external trade position has remained, and arguably even gotten slightly worse, even as the economy has started to slow down.

Now normally this deficit has been offset by the sum total balance of funds coming in as part of the financial account, but as we can see in the chart below this has also dropped off somewhat since August last year.

The consequence of this is that Spain's banks are increasingly having to find the money via the eurosystem to make the books balance, and this is basically the significance of that increase in borrowing that the Spanish banks have had to undertake at the weekly ECB auctions (referred to in this post - also please note that I have changed this post at this point thanks to some clarification from a knowledgeable commenter, see comments). If we look at the chart below, which shows the net asset and liability position of the Spanish banks vis-a-vis the eurosystem, those months when the bars are above the zero line indicate there has been an increase in the amount of money borrowed by the Spanish banks from the eurosystem (presumeably largely or exclusively at the ECB weekly auctions), and this money is essentially needed to settle the monthly deficit in the balance of payments account. As we can see, prior to August 2007 these quantities showed no particular trend, but since August, and for every month for which we have data, the Spanish banks have needed to raise additional money at the ECB. This is really a direct result of the fact that the banks have been unable to sell their cedulas for cash in the global markets. What the Spanish banking system lacks right now is a way to generate cash on a stable basis to meet the needs of the current account deficit.

And cash may well not be that easy to find since, while money may be attracted by offering higher interest rates, it is not clear what the funds would be needed or used for (rolling over existing liabilities should be more or less neautral here) and the rate of increase in bank lending has been slowing steadily all through 2007.

A country with a large and sustained current account deficit - as we can see in the case of the United States - can do one of two things. It can tighten money and credit conditions in order to try and use an indirect method to slow internal demand, or it can allow the value of the currency to slide (as we have seen with the dollar) in an attempt to reduce the deficit. Well Spain has no autoctonous currency to let slide, so the only real alternative is to squeeze internal lending to try and reduce the deficit, and in the meantime lean on the ECB for money. The contraction in lending needed to reduce a deficit of this magnitude might be very large indeed, and the consequences for the real economy would be substantial, so it is to be anticipated that in the short term some other alternatives will be sought. However it is interesting to note that in December 2007 the month on month increase in lending came to almost a dead stop. At this point it is too early to know exactly what significance to attach to this. But clearly the position does need watching carefully.

Basically this kind of sudden stop in lending, coupled with a growing sense of systemic crisis in the banks would be the worst of all worlds. To be monitored.

No comments: