EFSF As A First Loss Insurance Provider (reposted from Oct 21)
Based on the fast and furious headlines, it is hard to tell whether EFSF providing first loss “wraps” is still one of the myriad of possible outcomes, but it is worth looking at just in case.
The flying headlines obscure an important issue. While the different proposals sound like completely different solutions, they in fact fall into only a few groups based on assets to be acquired and capital structure. First the asset side, EFSF can make negotiated, conditional, bilateral loans (original EFSF), or buy other debt (primary or in the secondary markets), or equities and other assets. On the liability side, EFSF can be funded by the sponsor countries or provided with guarantees based on which funding is obtained in the markets. These are not really different – it’s a funding distinction, and there is no difference in terms of what obligations the sponsor countries take on. The other facet of the liability side is whether the backing is unlevered or levered. If levered, whether the leverage comes thru ECB funding, private funding, an insurance contract, an “AIG”-inspired supersenior CDS or any other form, it adds up to the same thing to EFSF from substance perspective – the differences are just form and possibly regulatory arbitrage to the provider of the leverage.
Eurozone Ratings
As a starting point, it makes sense to figure out what the rating of the “Eurozone” would be. Moody’s commonly uses a Weighted Average Rating Factor (WARF) model, that lets you figure out how they would rate a mix of differently rated assets. We can also apply the same concept to S&P ratings. For the Europe, lets take a look at the rating weighted by GDP, by Debt Outstanding, and by EFSF contribution rates. GDP-weighted ratings are produce best results because more influence is given to Germany and France. Debt-weighted ratings produce worst results because more influence is given to the weak countries. We also look at the ratings of all 17 members of the Eurozone, then with Greece, Ireland, and Portugal removed (“without GIP”), and finally also without Italy and Spain (“without PIIGS”).
The key to take away is that EFSF as a whole is NOT “AAA”. The only reason the original EFSF was rated AAA was because the original size was less than the amount backed by the AAA guarantors. As soon as the size of EFSF is greater than 451.5 billion (which is the max guarantee size of the AAA-rated countries), the entity on an un-leveraged basis is Aa3 even without the GIP countries. Without the PIIGS, EFSF can be as good as Aa1, but then the maximum guarantees only add up to 494 billion. As the potential for losses that might have to be repaid by non-AAA members increases, the rating has to decrease. If you are in any way relying on payment from a weaker member, then that member needs to be included in the rating.
The reality is, from a credit perspective, as the EFSF takes on more obligations, investors need to take into account how they are going to get paid back. So no matter what the rating agencies do, this is how investors need to think.
And guess what – that is exactly what the investors are doing. The EFSF 3.375% bonds of 2021 are trading with a yield of 3.43%. This is significantly higher than the 2.1% yield on German 10 year bonds, and even higher than the 3.22% yield on French 10 year bonds. It is important to remember, these are “good” EFSF bonds. They were done under the old documentation so should trade with a smaller risk premium than new bonds would.
Second Loss Greek Bonds
It is amazing that nothing has been done on the Greek situation. No update at all on any losses being forced on the banks. The original IIF proposal only had 1 scenario that would have forced a bank to take a write-down because it permanently reduced the principal. All the other solutions were designed so the banks would not have to book a loss. The talk about 21% haircuts is meaningless PR. The reality was that the deal was designed so banks would not take a write-down. Will that deal go ahead? If so, the EFSF is taking on massive amounts of Greek exposure.
We need some clarity on Greece as it does not fit into a wrapped model. All Greek debt with maturities of 9 months are longer are trading well below 50% of par. Without a restructuring, any 20% first loss Greek bonds would still have trouble getting placed in the real world. 20% first loss would be nice, but who would buy a new issue at even 80 when existing debt is trading at 40? Probably no one, so it wouldn’t work for Greece. If the EFSF was willing to provide 50% first loss insurance, it would attract private investors, but even the most optimistic investor would have to assume that 50% was gone from the EFSF funds.
Greece needs to be dealt with decisively and dramatically outside of the EFSF. So far, in spite of some rhetoric, that looks unlikely to happen and in one way or another the EFSF will be buying some Greek debt at wrong prices and banks will not recognize losses, but will have shifted some portion of their exposure directly to the EFSF. Those funds cannot be used to prop up Italy and Spain or Ireland or Portugal.
A simplified 2 country system
The EFSF has become almost too confusing to analyze. The EFSF can do so many things – buy bonds, guaranty bonds, invest in bank stocks, etc. It is extra confusing in that in many cases, the guarantors are the same risk as what is being guaranteed.
There are always risks when trying to model a simplified system. You can overlook some details in the effort to simplify, but in this case, it is particularly useful, and most of what you give up, makes the system more optimistic than it would be in the real world.
So lets assume a system with only 2 countries, one country, let’s call it “Germany,” is providing the guaranty, and one country, let’s call it “Italy,” is using the guaranty. Normally a system that has multiple guarantors guaranteeing multiple things would have less risk as you would get a benefit for “correlation” or lack thereof. In a complex system not all the things that are being guaranteed would default at once. Not all the guarantors would have problems at the same time. In the EFSF structure, I think the argument that all the issuers receiving guarantees default at the same time is a far more persuasive argument than trying to say one issuer would massively underperform. The guarantors are all correlated to what they are guaranteeing as well. So with all the correlation in the real EFSF structure, I believe this simplified example actually understates the risk.
So let’s see how well this “insurance” works in the simplified and safer world of just 2 countries.
Let’s also simplify the debt profiles of these 2 countries. It makes it a lot easier to analyze.
Who will support the secondary market for Italian bonds?
As currently discussed, the EFSF will provide guarantees on new issues. So basically it will support countries as their bonds mature and need to be rolled over. The first obvious question is what happens to the bond market in the meantime? Since nothing is fixed right now, the situation in Italy remains the same (they may actually have even less incentive to change to improve) we could see some market weakness. Will the ECB step in and buy bonds to support the secondary market like they have been doing? Possibly, although they seem to be trying to get out of that role. It will be interesting to see how the market responds to an Italian head of the ECB buying Italian bonds in the secondary market. If there is not a potential conflict of interest here, then it is hard to understand the definition of conflict of interest. Could the EFSF buy bonds in the secondary market? It could but then it would need money to do so. So some of Germany’s guarantee would have to be turned into EFSF bonds and then that money used to buy Italian bonds in the secondary market. But that would reduce the amount of guarantee money available to provide first loss insurance. So perhaps the EFSF could leverage their secondary market purchase. In theory they could do that in the market, but in the real world, someone would want variation margin, thereby exposing the EFSF to mark to market risk. So the EFSF could go to their good friend at the ECB and get non-recourse financing which it can use to leverage the money to buy secondary bonds. The ECB is fine with non-recourse financing, because they expect no default ever, but if there is a default, they will call Germany to top up the ECB capital. Yes, the ECB is happy to lend to the EFSF on non-market terms, because if they have losses, they would make capital calls on the members of the EFSF. Ponzi just rolled over in his grave.
So, in any case, not all of the money can be allocated just to insuring new issues. Some amount of money and a plan to lever it, has to be allocated to supporting the secondary market.
How much of the EFSF will be used in the first year? Is the ECB at risk?
Italy has 360 billion coming due in the first year in our “hypothetical” example. For the moment, lets assume the 20% first loss guaranty is enough. Then Germany will use up 72 billion of its guaranty. If the ECB is trying to exit the secondary market purchase role, then Germany will have to use some of its guaranty to support the 1.2 Trillion of longer dated Italian bonds. That is actually a pretty big number. If they only had to buy up 20% of the market, and could get 5:1 leverage from the EFSF, that would be another 50 billion used up. So even on 5:1 leverage both supporting the new issue and the secondary market, Germany will have used up 121 billion of its guaranty commitment. Just how much did Germany commit to? In the real EFSF world, they committed to 211 and have other guarantors, so 121 might not be too bad, but in the real world there is Spain, Ireland, Portugal, and maybe Belgium to deal with.
Sticking to our hypothetical world, Germany would use 121 billion of guaranty commitment, partly to support the secondary market, and part as the first loss new issue insurance program. In this example the ECB (by financing the 80% of the German secondary market purchases) took on 200 billion of un-hedged Italian risk. Even in this simplistic world, someone might question how the ECB can do that? It can either print money or it can rely on the capital call to Germany. At some point that little bit of circularity goes from being cute and charming to what the heck is going on here! Maybe the EFSF could guaranty some ECB bonds? I’m not going to dwell on the subject here, but the secondary market purchases have to be a part of the program, and to be meaningful in terms of size, they will need to be leveraged, but that could put real pressure on people’s faith in the ECB.
As big as the EFSF is, its guarantees will be used up fairly quickly. It is likely that in the real world, the guarantees will be used up in about 1 year, possibly less. That is scary. But what is worse is that the next round of problems and panic summits will occur long before the EFSF guarantees are used up. As soon as the market realizes that the EFSF will run out of guarantees before the real problem is fixed, we will be back in sell-off mode, with the difference that this time Germany will be in much worse shape because of everything it has guaranteed already and because it will have to contribute to the ECB for their purchases as well.
The mechanics of the first “First Loss Insured” bonds
In a year, in the simplified example, we will have the first big issuance. Germany has 330 billion of debt maturing that it has to roll, and Italy has 360 billion. Both countries probably need to issue a bit more debt than that, because they are both running deficits that need to be funded. So Germany will be guaranteeing a bit more than 72 billion of Italian debt. The big questions are, where will German debt price, where will new “guaranteed” Italian debt price, and what happens to non-guaranteed Italian debt.
German yields will increase. Germany will be rolling their existing debt, plus a little bit extra for that year’s deficit, and will now have created a contingent liability. If Italy defaults, Germany will need to come up with 72 billion. The fact that Italy knows it has Germany on the hook does little to make Italy more prudent. If anything it increases the probability of default, because they can afford to be more reckless.
It is hard to say how much this will impact German yields. It will depend on the state of the German economy and the state of the Italian economy. To the extent both are improving, all is good, but then all is good without the EFSF. If both are deteriorating, then the problem is magnified and the cost to them is large. People can pretend a guarantee is not the same as funded debt, but that is accounting rules. At a very real level, it is the same – it is a real, binding obligation. If Credit Default Swaps only had value in the event of default they would not trade daily. Their value changes based on the probability of default (probability of a Credit Event actually). As a Credit Event appears more likely, the value increases, and as credits improve the value of a CDS declines. The guarantee works roughly the same. If Italy looks okay, the guarantee will not impact the market’s assessment of Germany’s credit much. As Italy declines, investors will start to view that guarantee as being part of the German debt. That will impact prices as the size is material even for a country as strong as Germany. For the first time, Germany will not have control over the riskiness of their debt. No policy they institute would affect the likelihood of that guarantee being called on.
Furthermore, leverage has a price. If Germany had just guaranteed 75 billion of Italian debt, then in default, it might expect to lose about 40 billion (with recovery around 45%). With the first loss risk, they will now lose the full 75 billion if recovery is less than 80%. Have any of the EFSF actions done anything to reduce the probability of default for Italian debt? No. The probability of default of Italy is the same as it was before the program because their obligations haven’t materially changed and this has done nothing for their economy. All it does is shift who bears the losses if there is a default.
New Prices of 3 Year Debt
To try and keep this as simple as possible, since even this simplified analysis is already incredibly confusing, let’s assume that they are going to issue all the bonds with 3-year maturity. If the yield curve was unchanged, straight 3-year German debt should price at 0.8% and straight Italian 3-year debt should be at 5%.
The pricing is all a bit circular, but what isn’t at this point. Let’s assume the cost to Germany is 20 bps and their 3-year prices at 1%. Is that a reasonable assumption? If Germany was really trying to issue an additional 75 billion of debt, it would push out their yields some amount, and I would guess by more than the 20 bps. The cost would be more if Germany had been issuing bonds throughout the year to make secondary market purchases. Those purchases too would impact the credit quality and scarcity (or lack thereof) of German debt instruments outstanding. So it is just a guess, but the right number is some increase in Germany’s cost of funds relative to not having provided the guarantees or participated in secondary market purchases. On the other hand, this cost to Germany at this stage is negligible. The weaker either country was, the higher the cost to Germany. A decent guess for where the new Italian debt should trade is 80% of Italy’s “normal” yield and 20% of Germany’s yield. That ignores that the German risk is second loss, but accounts for it being less liquid. In any case it is a decent ballpark figure. Italian yields should be 5%, but has Italy become any safer? No, since nothing about the plan fixes Italy’s debt load or its current deficit. Will some investors want to sell the old bonds to buy the new “better” bonds? Yes, I think you need to assume that some risk-averse investors will shift out of the old bonds into the new bonds. I think the move in old Italian bonds would be reasonably significant. I could easily see them going from 5% to 6%, but for now let’s assume they only go to 5.5% (this will be another case where the worse Italy is doing, the bigger that drop will be since more investors will want to hold only new insured bonds).
Then by our simple math the new Italian bonds will yield 4.6% (80% * 5.5% + 20% * 1%). That is an improvement of 0.4% from their standalone yield. I have to say that is kind of paltry. Maybe the first loss will give a bigger benefit, but in real world not all the guarantees are coming from Germany.
If cost to Germany is only 0.2% and gain to Italy is 0.4%, although small, should we be cheering?
Not even close. If you were going to lend to an Italian bank, would you lend to them based on the “insured bond” rate or the full exposure rate? Since the EFSF has jumped to 10 trillion, I am assuming banks are going to have to fund without the benefit of any insurance. If your choice is straight Italy at 5.5% or an Italian bank, you have to charge the bank more. The bank in turn has to charge more to its customers. So the funding for banks will have increased by 0.5%. In fact, rates throughout the entire country will increase by the amount straight debt yields increase. The ECB could just buy more bonds, right? Ponzi is now kicking himself for thinking too small.
Couldn’t Straight Italian Bond Yields Improve?
Nothing about the program does anything to change the state of the Italian economy. Their risk of not being able to pay back debt has not changed. They will receive some benefit from reduced funding costs on their new issues, but it only impacts the new issues, and most of the bonds they have maturing in the next year have fairly low coupons anyways since they were issued when Italy was less of a mess. The yields will move, but largely as a function of the Italian economy. If it starts improving and they get annual deficits under control expect yields and spreads to improve. If the economy weakens or deficits get worse, expect spreads and yields to increase. Expecting yields to improve because of a bunch of artificial shenanigans is unrealistic. Over the short term, sure, anything is possible, but over the course of weeks/months the real economy and the real appetite for risk will win out.
How can Germany protect itself?
If Germany is worried about the moral hazard risk and has any concerns about creating debt problems of its own, then it can impose conditions on purchases. It can put constraints on the timing or amount of guarantees available. But if they do that, then we are back to monthly or quarterly TROIKA-watching. I know it is hard to remember but Greece has been “solved” multiple times already. The problem is so far Greece has never quite managed to live up to their obligations to get the next round of money. If the deal puts a lot of constraints on the ability of weak countries to tap the facility, then they can go ahead and start scheduling the next round of summits.
What am I missing?
Maybe I am missing something. In the end I just don’t see how much this really helps or that it buys much time. It links everyone together, but aside from Germany, no one is in great shape and the risk to Germany and France hardly seem worth the benefits to the periphery. Maybe their analysis is much better and really fixes everything, but I struggle to see how that could be the case. They seem to constantly ignore one cost or another. They constantly seem to ignore the risk already sitting at the ECB, which until the ECB decides to just print money, is real.
Maybe the banking model is a better idea? Somehow I don’t see that situation working that much better, and ultimately leads to the same conclusion that either the economies have to get a lot better, or the ECB has to print money, or everyone will get dragged down together to protect the weak.
What would I do?
I would arrange for Greece to default. Announce that they are not paying. Have as many institutions lined up in advance to accept a reasonable restructuring package – 60% principal reduction, and have DIP financing in place so Greece can function. Yes, this will scare the market, and we will have a sharp decline. The fact that a Credit Event has occurred will scare the market.
The ECB and the FED have to have emergency financing lines ready to go. It should come at a reasonable price. At a price that lets the ECB and FED book a lot of money at the expense of the banks that needed to tap the lines. This money can reduce Sovereign debt or current budget deficits. No bank should have to sell something because they lose financing. Forced selling into a cascading market would be a disaster. That financing cannot be permanent as banks still need to deleverage, but the timing of that can wait.
The EU and Treasury should make capital available to banks. It should have terms similar to what Buffett got with BAC. That seems like a fair deal for the citizens and for the banks. No bank should be forced to take the capital. If they think they can go it alone or raise private money at a better prices, that is fine. The offer can only remain outstanding for a period of 3 months, after which time, no capital can be provided to the banks. It is not a free option, it is meant to be a fair deal for those who need it. I would like to see restrictions so that the banks in the worst shape, just have to go under rather than taking money to stay alive for a little longer, but so much money can be made from the equity investments that work out over time (under the Buffet deal) that I can live with a couple of bad investments.
I would encourage the formation of new boutique Investment Banks, and finding ways to allow existing boutiques to grow. Make it easier, rather than more difficult for new banks to be started. The playing field should change to let those who are ready and prepared take over. This would be really bad for existing weak banks, but not at all bad for the well-managed strong banks and great for the economy in the long run.
Also every bank should be encouraged to collapse their Greek CDS trades ASAP to clean up administrative complexity of the situation. The regulators can make the dealers collapse CDS trades by stepping out and assigning/unwinding and set a deadline of year end to have every single CDS trade and Interest Rate swap into a clearing mechanism or exchange. Bank balance sheets need to be stripped of these massive gross numbers. I don’t believe the numbers are as dangerous as some people do, but I see no reason in this day and age, given all the angst these positions cause the system, why they should be allowed to remain there and continue to create lack of transparency and unnecessary complexity. That might be bad for bank stocks. It will be good for boutiques. It will be great for the economy as the fear of systemic risk will have been decreased. It is certainly negligent that almost 4 years after Bear Stearns, this has not yet been done.
TALF-like programs could be started. TALF was possibly too generous, but it would be a good way to encourage private risk taking.
The initial reaction will be fear. That has to be dealt with as well as possible. Then the rebuilding process can start. People will realize that Italy is not actually at risk because Greece defaulted. Italy has a lot of problems, as do Spain, Ireland, and Portugal, but they are largely their own problems. They really are not that interconnected. Once this sinks in, if Italy had widened additionally in the fear, Italy will be right back to where it is today. Not great, but fixable. Now people will be able to focus their efforts in a much more concentrated way. Investors will know that at the very least there are no hidden Greek time bombs that have not been revealed yet. By this time (a month), the Greek CDS will have settled without any surprise calamities. There might be a couple of calamities, but they won’t be a surprise. Banks are trading at a steep discount to book, because no one believes book value, and this process will be a good step in adjusting book value to reality.
Every single major bank CEO has come out and said they are fine. They have managed their PIIGS exposure. Well, let them prove it. Just imagine a world where Greece defaults, and the banking system actually remains functioning? Then I can see SPX heading to 1500. Yes, there are a lot of “ifs” in my scenario, but except for printing massive amounts of money and handing it to the weakest sovereigns and living with raging inflation, I do not see how we can avoid it. And I would rather do it now while there still are some pillars of strength.