Opinion

# Felix Salmon

## Will Grexit topple Obama?

May 25, 2012

One of the hardest questions to answer, when people ask about the European crisis in general and the Greek crisis in particular, is “why should we in the US care?” The simple answer is that well, this is an important part of the world, and it’s big news. But if you only care about news insofar as it directly effects the US, then the answer is harder.

One possible answer — I’ve heard this given in a number of places — is that another major crisis in Europe would spill over into the US, cause serious economic damage here, and could quite possibly make the difference between an Obama and a Romney victory. But just how likely is that? I’m no expert when it comes to assigning probabilities to events, but we can at least come up with a general framework which lets us answer the question.

Let’s start with the fiscal pact. Will all of Europe credibly commit to fiscal austerity going forwards? If so, that increases the chances of crisis and Grexit, since southern European countries in general, and Greece in particular, simply can’t operate under an austerity regime in the way that, say, the Baltics have managed, painfully, to do. On the other hand, everybody seems quite likely to break the fiscal pact in one way or another — which means that there has to be a good chance the pact will end up being honored mostly in the breach. Let’s call the probability of a Europe-wide austerity regime A; my best guess for A is roughly 15%, or 0.15.

So the next question is — what is the probability of Grexit, any time soon? That’s really two questions. First, what is the probability of Grexit if there’s Europe-wide austerity. Let’s call that B, and I’ll peg it at 85%, or 0.85. Second, what’s the probability of Grexit if Europe-wide austerity slips a bit? We’ll call that C, and I’ll say it’s 65%, or 0.65. Overall, we can define the chance of Grexit, D, as A * B + (1-A) * C. If you’re playing along at home, that’s 0.68, or 68%.

But just because Grexit happens, doesn’t mean it will necessarily affect the US election. For one thing, by definition, Grexit can’t affect the US election if it hasn’t happened by the time the election takes place. So the next question is: if there’s Grexit, what are the chances that it will happen by November? The Europeans have proven themselves very good at kicking the can down the road, so even if Grexit is inevitable, it’s still not inevitable by November. In any case, let’s define E as being the conditional probability of Grexit by November, given Grexit. I’ll say that’s 50%. Which means that the overall probability of Grexit by November, F, is D * E, or 34%.

Grexit, if and when it happens, will cause a lot of disruption in European markets, and certain deposit flight out of Spanish and Portuguese banks. Again, there are two ways this can play out. Either it will cause a series of further dominoes to fall, or else it will concentrate the Europeans’ mind and force them to build a large and genuinely effective firewall, drawing a line in the sand and saying “this far, but no further”. Will Europe let the Grexit crisis go to waste? Let’s say the probability of a credible, coordinated and constructive pan-European response to Grexit is G. Then the probability of Grexit causing a big European crisis is 1-G. What’s G? That’s a tough one, but I’ll put it at 35%.

For the purposes of this calculation, we’ll assume that Greece alone is too small to cause a big global crisis: you need contagion, for that. So we’re looking for H, the chance of a big European crisis before the US election. We can calculate that as F * (1-G), or 22% — that’s the chance of Grexit before November, multiplied by the chance of a bigger crisis if Grexit happens. (Note that a big European crisis can happen at any time; the chance of that is D * (1-G), which works out at 44%.)

If we have a big European crisis before the election, then that will certainly send US stocks falling. But will a sharp drop in the US stock market have any effect on the outcome of the election? Probably only if the election is reasonably close — and certainly not if Romney is in the lead. A European crisis, and consequent plunge in US stocks, would only be good for Romney and bad for Obama, just as the crisis in the fall of 2008 was good for Obama and bad for the incumbent Republicans. So what we’re looking for, here, is I, the probability that Obama will have a narrow lead over Romney — one small enough to be erased by a big stock-market plunge. I’ll peg I at 65%.

And thus, finally, we get to the big answer: what is X, the probability of Grexit toppling Obama? That is H * I, which using my off-the-top-of-my-head probabilities, works out at about 14%. But you should work this out for yourself. Come up with your own values for all these:

A: What is the probability of a Europe-wide austerity regime?
B: If we get Europe-wide austerity, what are the chances of Grexit, any time soon?
C: If we don’t get Europe-wide austerity, what are the chances of Grexit, any time soon?
D: What, then, is the probability of Grexit? This is calculated as A * B + (1-A) * C.
E: If we have Grexit, what are the chances it’ll happen before the election?
F: This is the overall probability of Grexit before the election, and is D * E.
G: If we have Grexit, what are the chances of it eliciting a credible, coordinated and constructive pan-European response?
H: This is the probability of a big European crisis before the election, it’s F * (1-G).
I: What is the probability of Obama having a narrow enough lead over Romney that it would be erased by a plunging stock market?

Put these all together, and you can finally come up with a number for:

X: The probability of Grexit toppling Obama. It’s H * I.

I’d be interested to know what results you get, but my guess is that most of them will come up with a number which is low and yet still significant. It’s something to bear in mind, but of course it’s also something which is pretty much entirely out of Obama’s control. That’s the way that crises work: individual politicians are rarely personally responsible for them, but whomever’s in power when they happen nearly always ends up getting the blame.

There’s a small amount of dodginess in there – some of these probabilities aren’t independent. For instance, whether the pan-European response is CCC is likely (negatively?) correlated with whether or not grexit happens before the election.

So you really want to know the probability the response is CCC *conditional on* grexit happening prior to the election, which is probably lower than G.

Posted by absinthe | Report as abusive

First of all, I’d claim the formula you use for calculating F is wrong:

F actually depends solely on the result of the June 17th Greek elections. Either there will be a pro-austerity government after the elections and a Grexit will not happen (or if it ever happens not before the US elections), or the Grexit will happen before the US elections.

A cannot influence F since any serious changes to A won’t happen until after the 2nd round of the French parliament elections, which is also on June 17th (and Hollande is e.g. in his demands for Eurobonds currently vague enough for allowing him to claim later that all he wanted was common EU debt for a few billion Euro in additional EU infrastructure projects).

The Greek have an interesting feature in their elections that one sixth of the seats in parliament are additionally given as bonus to the party that gets most votes. Current polling sees a head-to-head race between ND and SYRIZA, so I’d give ND a 50% chance of getting most votes.

If SYRIZA gets most votes, I don’t see any chance of a government that would avoid a Grexit.

If ND gets more votes than SYRIZA, and PASOK and other parties who might join a coalition don’t have too disastrous results, there will be no Grexit before the US elections. If ND gets more votes than SYRIZA, I’d assume a 90% probability of avoiding a Grexit.

Everyone in the EU is busy using the 6 weeks between the May elections in Greece and the June elections in Greek for planning the case of a Grexit. No matter if it is about the effects on banks in other EU countries or about delivering medication to the Greek people (it would be inhuman not to give medication to a EU citizen only because he and/or his country is bancrupt), the EU planning of the details for coping with a Grexit is already pretty far today. And since the French elections will end on the same day as the Greek elections, Hollande will also be able to act more freely then. I’d therefore assume an 80% probabilty that there will be a a credible, coordinated and constructive pan-EU response to a Grexit.

H is the number I care about here in Europe, I and X are much less relevant for me (and when talking about “something which is pretty much entirely out of Obama’s control” an Israeli attack on Iran has IMHO a higher probability than a big European crisis).

E: ND gets more votes than SYRIZA
My guess: 50%

F: Overall probability of Grexit before the US elections
My guess: 1 – (90% * E)) = 55%

G: If we have Grexit, what are the chances of it eliciting a credible, coordinated and constructive pan-EU response?
My guess: 80%

H: This is the probability of a big European crisis before the election, it’s F * (1-G).
My guess: 11%

Posted by AdrianBunk | Report as abusive

Since this is in its initial stages a financial crisis and not a macro economic one, all your attempts at probability assignment are probably irrelevant. It’s what the financial markets perceive as the risk that is significant.

As you and everyone else has pointed out, they are just not that big a player. If between now and Grexit election the market internalize the risk and factors the discount, the final event will not be significant. Spain and Portugal are a larger canvas that will allow for other options. Greece, in the embrace of chaos, has lost its seat at the table, not so Spain and the rest.

Posted by ARJTurgot2 | Report as abusive

I suggested in the past that if you’re going to play around with numbers, at least learn some math. The problem as you’ve posed it is not multiple independent events as you are treating it, but rather Bayesian conditional likelihoods. Go back and resubmit with a Bayesian solution before you can get credit for the assignment.

However, in general I don’t think it matters. There are simply far more variables to an election than you note, and you’re simply pulling selected probabilities out of thin air. If it’s possible to model election results, I would suggest starting with either a discriminant analysis or a time series. But I don’t think what you’re trying to do is even remotely possible.

Posted by Curmudgeon | Report as abusive

obama shot himself in the foot with sanctions against iran

the blow torch of higher gasoline prices showed he miscalculated

clinton has gone quiet and iran is back at the negotiation table

grexit is only a transitional side dish – eurozone fiscal reform is the main menu

Posted by scythe | Report as abusive

This is a very nice exercise in probability tree modeling, but it makes some assumptions not borne out by empirical evidence. First of all, there is no precedent for an American presidential election being swung by a stock-market drop. More importantly, at least in the modern era, a sitting president has never been toppled because of economic factors alone. For that to happen a strong opposition from the incumbent’s own side had always been necessary. Historically speaking, the president’s reelection seems a fairly safe bet.

Posted by mythdesysiphus | Report as abusive

“First of all, there is no precedent for an American presidential election being swung by a stock-market drop.”

It seems to me likely the 1932 election was rather strongly influenced, maybe even swung, by the stock market drop 3 years before.

And what about the 2008 election? Before the bottom fell ut of the stock market, McCain was enjoying a nice post-convention bounce, he and his mavericky VP choice seemed all the rage, lipstick on the pitbull and all.

Posted by Christofurio | Report as abusive

@Curmudg -

Well, .. um.. yeah – but it’s a Friday piece ahead of a long weekend, so, just about anything ….

Posted by MrRFox | Report as abusive

@Everybody –

‘While (Felix) the cat is (holidaying) away, …’ – why not stage a micro-coup d’etat today?

Here are the relevant high-points of it –

“at most other popular sites, the number of people reading the comments is vastly greater than the number of people writing them. But the way they’re presented, they’re not easy to read, …

“the fact is that even though there are many more readers than there are commenters, there are also many more commenters than there are posters. And collectively, those commenters are faster and funnier and more knowledgeable than the staff of any website….

“forums have been around on the internet since the 90s, and no one’s managed to reinvent them yet….

“editorial product is very much moving towards comments and away from posts, … Expect … blog posts to get shorter, in future, and sometimes just be a headline, at least in the first instance, so that the conversation can get going before a pretty post can be put together.”

And KenG adds this – “Speaking of comment systems, can you guys update the one Reuters uses? It’s kind of outdated, ya know.”

With you there, Ken – the archive system is archaic, and not having any search or edit functions at all is a pain. Which raises the question – what’s the right direction for this blog to go?

I sort of like the idea of going semi-forum format, as Felix describes the trend, but the big question is – will Felix be the only one who writes OPs? I kinda hope so on one level. Having seen what happens to formerly intelligent sites that grow in membership (like INTJf) – the quality level suffers a fatal dilution. So, bells and whistles and the ability to write OPs to attract more commenters – I hope not. Felix is half of what makes this blog special, but the character/quality of the commenters is the other (fragile) half.

How to make it better without jeopardizing what makes it special? IDK. You?

How to make money from it? That’s Reuter’s problem.

Posted by MrRFox | Report as abusive

The math you use seems like it might be sort of irrelevant. The problem as I see it is the likelihood that financial markets have not accurately priced in the risk of a breakup of the eurozone. If Greece exits, this mispricing will become apparent. Lehman Brothers wasn’t deemed to be able to truly damage the US economy, but financial markets had something different to say when it was actually allowed to fail. We have been extending and pretending and mark to unicorn-ing for so long… God knows what will happen.

Posted by http | Report as abusive

You should have stopped with “I’m no expert when it comes to assigning probabilities to events” and let it go at that.

Posted by PseudoTurtle | Report as abusive

Posted by GMavros | Report as abusive

Working from AdrianBunk’s reasoning above, but running the “pessimistic” version, assume the Greek election on 17 June goes as it should, with the Greek people saying “no.”

Grant that it is impossible logistically for Greece to exit the Euro by November, but rather that the probability that they will exit in the near term goes up.

But that probability has been approaching 100% for many months: certainly since the second austerity round, in which it became clear that Greece would have to sell off its revenue-producing assets at fire sale prices to get through the (rather long) short-term liquidity problem.

As soon as “clearly illiquid” became “likely insolvent,” the game was over, and everyone who matters from a financial markets perspective knew it. (Look at the cash flows of the HNW Greeks.)

Greece leaving the Euro won’t be a non-event, but it’s not going to cause a European recession. (Germany is quite capable of doing that themselves, with a bit of help from the Brits.) Let us not confuse which is cause and which is effect.

A Greek exit–and, more importantly, a Greek exit plan–would be an indication that Europe in general and the EU in particular sees a way out of their fiscal crisis.

It won’t save Obama–see GHWB in 1992, though without his predecessor’s clear Desire to Do Other Things–but an exit plan that looks as if the light at the end of the tunnel isn’t owned by SNCF or Deutche Bahn would be a positive prospect.

Posted by klhoughton | Report as abusive

To Christofurio:

Between the 1929 stock market crash and the 1932 election there was a small matter intervening, called the Great Depression. It is not controversial to say that the crash was not its cause, but rather a symptom of the developing broad economic collapse. More to the point of my post, Hoover received a very strong challenge from Coolidge and Blaine and his candidacy was heavily bruised going into the ballot.

In 2008, the market went into bear territory a good 6 months before McCain’s bounce, nor was he the sitting president.

Posted by mythdesysiphus | Report as abusive

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/
•