When yesterday we presented [25]the view from CLSA's Chris Wood that the February 29 LTRO could be €1 Trillion (compared to under €500 billion for the December 21 iteration), we snickered, although we knew quite well that the market response, in stocks and gold, today would be precisely as has transpired. However, after reading the report by Credit Suisse's William Porter, we no longer assign a trivial probability to some ridiculous amount hitting the headlines early in the morning on February 29. Why? Because from this moment on, the market will no longer be preoccupied with a €1 trillion LTRO number as the potential headline, one which in itself would be sufficient to send the Euro tumbling, the USD surging, and provoking an immediate in kind response from the Fed. Instead, the new 'possible' number is just a "little" higher, which intuitively would make sense. After all both S&P [26] and now Fitch [27]expect Greece to default on March 20 (just to have the event somewhat "priced in"). Which means that in an attempt to front-run the unprecedented liquidity scramble that will certainly result as nobody has any idea what would happen should Greece default in an orderly fashion, let alone disorderly, the only buffer is having cash. Lots of it. A shock and awe liquidity firewall that will leave everyone stunned. How much. According to Credit Suisse the new LTRO number could be up to a gargantuan, and unprecedented, €10 TRILLION!
Here is how the strawman is now put in place for what may the biggest liquidity injection in modern history in just under two months.
February’s second 3-year LTRO looks set to be extremely large. Really extravagant claims (we have heard reports of €10 tn) are probably wide of the mark because this will not be a complete collateral free-for-all (unless NCBs choose to make it so, which for some of them is admittedly an open question; again, see rational player section below). But the idea of path-finder lightning springs to mind (High-speed cameras reveal that lightning evolves “bang BANG”, essentially); the last LTRO has removed any stigma, making managements who do not exploit the value on offer arguably careless at best. This is, on the face of it, very cheap protection indeed against any possibility of a liquidity crisis for three years.
Naturally, if indeed there is anything even remotely resembling a €10 trillion expansion in the ECB's sub €3 trillion balance sheet, all bets will be off as the ratio of the ECB to the Fed assets [28], a correlation which would imply a sub parity level on the EURUSD would gut corporate earnings in the US, all merely to prevent the disintegration of the Eurozone. And while this event will be welcomed by the Fed initially as it will send stocks exploding to potentially all time highs (and gold to well over $2000/ounce), it will cripple the US manufacturing model unless the Fed immediately responds in kind, and prints outright, and unsterilized, a non-trivial comparable amount. In other words, the world could very well enter the final round of global coordinated currency devaluation, aka FX war, together. Yes, that means coordinated printing by the SNB, BOE, PBoC, BOJ, etc, etc. Simply in a last ditch attempt to preserve the status quo. Which, unfortunately, after the knee jerk reaction, will fail. CS explains why:
But there are many problems, particularly at the systemic level. So, although an eye-catching number may trigger a rally (to be clear: 29 February is an eternity away in the current environment), Greece is a salutary reminder that treating a solvency problem as a liquidity problem, including buying time for solvency to be addressed and other dubious concepts, is a disaster for existing creditors, who are subordinated throughout by the senior rescue funds. It also shows that three years is an eternity and that timescales (such as the possible introduction of resolution regimes) are flexible. It remains unclear to us where the new capital for the European banking system is going to come from if not from existing bondholders and that, in some cases, that must involve senior, in our view. So the LTRO is another example of reducing idiosyncratic risk at the expense of systematizing risk, in our view.
The participating bank needs a plan for refinancing the LTRO; the lack of a plan means that financing will not be forthcoming, in a death spiral. So the massive provision of official liquidity will act as a long-term triage and increases the pressure (albeit over a longer period) to raise capital. It is a free lunch only for stronger banks, in some sense, and at the systemic level is anything but. Rather, it is paid for by those with senior exposure to weaker banks.
In fact, Credit Suisse is openly hunkering down as it now see the "endgame taking shape" - "We do not expect the downgrade of France to have an immediate impact, but it highlights very clearly to us the ultimate issue that has to be tested in the euro area. Will Germany hold together the euro as, when and if France becomes part of the periphery?"
Here is how the European endgame will look, through the prism of game theory.
The Nashing of teeth
We continue to analyse the euro area sovereigns in game theoretic terms [29], with the game looking from outside like a CDO [30]where value allocation is a function of expected losses and of correlation (distribution of those losses in a tranche structure).
In simplistic terms, we have stated that Portugal cannot rescue Greece (i.e., Greece's creditors), Spain cannot rescue Portugal, Italy cannot rescue Spain, France cannot rescue Italy, but Germany can rescue France.
In equally simplistic terms, we have stated many times that “crises are baked into the cake”.
With the CDO and game theory analysis, we have formally modelled both statements. In the game theory, we concluded that the most likely outcome is a series of ever-deeper crises and relief rallies culminating in a definitive moment. Before that definitive moment, our analysis suggests that both core and periphery parties playing hard-ball leads to an escalation of the crisis, but not calamity. Only at the decisive moment does a “collision” in our game of chicken model lead to catastrophe. We remain convinced that the decisive crisis has not been seen yet.
[31]
What we have not done before is put the two statements together. To do so, it strikes us that each “wave” of the series of crises represents the transition of a country or countries to the periphery. We started with the “Greek crisis”, still under way with, at the time of writing, both parties threatening hardball. Then we had the rest of the outer periphery, with Ireland “swerving” rather spectacularly and Portugal not so clear. This led to a set of rescue plans that made the implicit assumption that the periphery would broaden no further.
But the outer periphery was followed by Italy and Spain as the crisis emerged in the summer as truly systemic. The resulting crises have been forestalled, for now, by ECB action (SMP and 3-year LTRO). Now, the (well-flagged) action by S&P hints at the final crisis.
Current news on Greece raises the question of whether the (existing) core wants to “rescue Greece”, amid the usual nonsensical debate about laziness and doctors’ swimming pools. But our analysis suggests a steady narrowing of the core and broadening of the periphery as we head to the ultimate question. Would Germany want to rescue France? And – see the below section on the rationality assumption – would France want to be rescued? We are not gong to futurize French politics, but we note that we are already in a situation where it is not impossible that Le Pen eliminates Sarkozy in the first round, according to the polls. By definition, France finding itself on the periphery would damage France’s leadership position in Europe and risk a change in politics. Of course, we think the ultimate answers are “yes” and “yes”, but that frisson of doubt, in light of the consequences of the implications, will keep the market well on its toes.
Putting it all together: the rational national player assumption.
News agencies try to reconstruct the internal national debates of the euro area, which are often carried out fairly openly, it has to be said. But a country often has a simple choice to express to the world, and we model this in our game theory as “hard” or “soft”. We make, thereby, a key assumption, that a country acts in a way that appears rational when viewed externally as a single entity in that single action.
We have complained for years that countries cannot be seen as monoliths. We need to consider the possibility that the outcome of the internal game, since it is not fully transparent to the outside, can superficially seem irrational. Circumstances could be envisioned where a player chooses to play hard in a situation where he knows that it will damage national interests, i.e., playing hard will be worse for it than playing soft. For example, it might be rational for a player within France, lets say le Pen or Sarkozy, to play uncooperatively in the European context, if that is beneficial from a French or individual perspective. Even within the European context, a country playing "irrationally" aggressively can be part of a rational strategy of brinkmanship behaviour. For example, by trying to make a threat look more credible. We could be seeing this in German attitudes to Greece, as we explore below. The corollary is that this can only work up to the penultimate stage of the crisis, as only that way one can try to force the other player to do the right thing in the final stage. Overall, we see the risk to be vigilant with the rational monolithic player assumption in the euro area. And, of course, flexing it does not invalidate our conclusions, it merely points out that getting it wrong is a key risk in the analysis. This is framed by the historical context, where Europe has achieved disastrous outcomes under the incentives of the time.
The risk of the simplification can be highlighted by a simple example. The rational action of the College of Cardinals under the post-13th century papal conclave is collectively to fill the vacant Papacy immediately. Individually, as well, the cardinals have the same incentive; no matter how beautiful the ceiling of the Sistine Chapel, they can see it anytime. So any sensible rational expectation of the smoke colour, based on treating the conclave as a single rational entity, is white at all times, including on the first day. That assumption would have led our observer astray by 82 rounds over 50 days between 1830 and 1831. Similarly, as we head into elections in France, already referred to above, and to Germany, the assumption is likely to come under some stress. We think that understanding of the internal “game” is most central in these two countries, particularly Germany. This is because, in our view, the periphery playing “soft”, rationally or irrationally, cannot make the crisis resolve on its own. Or, austerity is not sufficient. The central question, now hoving into view, is whether France and Germany, when the time comes, can co-operate. By then France may be in an embarrassingly inferior position, possibly forcing a Nationalist response that is apparently irrational when viewed externally but rational when viewed in consideration of internal incentives.
Between Greece, French elections and the downgrades, the situation remains as deeply uncertain as ever. The effect of the LTRO may last a while longer, buoyed by a negative market, but we are on the alert for a turn.
Finally, CS' appendix on why what started with Greece, will likely end with it:
We no longer publish views on Greece because the situation is too fluid to permit it, but we would observe that a key is whether sufficient voluntary participation can be induced to enable a CAC to be credibly introduced. In this case, “free-loading” on anything other than the 20 March bond becomes dangerous if a further restructuring is thought necessary, which we regard as the market’s expectation. In a game-theoretic sense, which is still the only way we can look at it, the dominant strategy, if a restructuring on more burdensome terms is expected with probability 1, is to volunteer now. The exception is the basis holders but, as we have pointed out many times, CDS are a tiny fraction of bond outstandings so basis holders are unlikely to tip the balance. Free-loading on March is a much more subtle game given that there is not time for a second round and the authorities would still like to avoid a hard event. What would they be willing to pay for this? How much 20 March is held outside the ECB and banks which are under official influence? etc. As we write, both parties seem to be playing “hard”, suggesting a suboptimal outcome of a hard default. As examined above, at this relatively early stage of the game we should expect a “swerve”, most probably by the core. But continued playing hard might be the outcome of rational behaviour by either party, given the multi-stage nature of the game. What message would paying the 20 March in full on Greece’s behalf give to the rest of the periphery?
The last bolded sentence is precisely what we warned about last week when we said that should Greece devolve into a full out coercive restructuring, the one real question would be: "who is next?"
Needless to say, if Credit Suisse is even 20% correct in its estimates, and the more we think about it, the more plausible it is that 20 days ahead of the Greek default the ECB will bend over to provide every last penny European banks may need, and then some, to firewall exposure fall out (since none except for UniCredit actually did a capital raise and we all saw what happened then), then all bets are truly off. Should the ECB indeed escalate events to this degree, then we are about to leave the paradigm started with the late 2008 bailout of Lehman, and enter one in which every incremental swing in the global socio-economic sinewave could well be the last.
As such, attempting to predict what happens after becomes futile.
Incidentally, those curious what a €10 trillion expansion to the ECB's balance sheet without a proportionate response by the Fed, would do to balance sheet correlation, and implicitly, to the EURUSD pair (the correlation was explained previously here [28]), this presents it vividly.
[32]