Modeling a Grexit Is Darn Near Impossible But Let's Do It Anyway

Everyone loves a good financial model, but the problem with the Greece situation is that it's just so darn hard to model.

The union of this many major economies into users of the same currency is unprecedented to start with, so trying to come up with a playbook for what happens when one leaves is about as hard as trying to put together a bunk bed from Ikea without the instructions. (Or heck, even with the instructions.)

This difficulty in modeling explains why smart people are coming to vastly different conclusions about a) the odds of Greece leaving the euro; b) the potential damage that will be done to the rest of the world's markets; and c) why the upcoming long holiday weekend on Wall Street will be a lot less relaxing than previously modeled. Exhibit A: Mohamed El-Erian put the chance of a Grexit at 85 percent; Morgan Stanley sees a 60 percent probability, up from 45 percent previously; BNP Paribas SA economist Luigi Speranza puts the risk at 20 percent.  

But what about these insanely powerful computers we all have now? Can't some whiz kid put a few of them to work to figure out what's going to happen, so Greeks can know if the banks will ever open again and Wall Streeters can know if they'll need to stare at computer screens instead of fireworks this weekend?

Evan Schnidman of Prattle Analytics is one of those smart whiz kids, and he has fed his machines with the words flowing from European Central Bank policy makers to figure out whether or not they have the stomach to allow a Greece exit. The Harvard PhD's software scores the words of central bank speeches and other communications in an attempt to gauge their mood. This type of textual analysis is obviously a relatively new field without a long track record. But Evan's a nice, smart guy and he made a chart, so let's take a look. 


The first reaction to seeing a chart like this is obviously: Cool chart, bro, but what the heck does it mean?

Explains Evan: "Sentiment has plummeted in recent days with ECB sentiment clearly reflecting frustration with absence of a Greek debt resolution. However, sentiment has not reached the same depths it was at even three weeks ago when skepticism of a Greek debt resolution initially took hold. This indicates that European Central Bank leaders are coming to terms with a Greek default and possible Greek exit from the euro zone. Essentially, although ECB sentiment has mirrored the roller coaster of questions about Greek debt repayment, as these ups and downs temper, it seems that ECB leadership is becoming increasingly accepting of a possible `Grexit.'"

Let's assume, then, that the Grexit predictions come true. So what happens in financial markets afterward? This is what Carlo Acerbi, Zsolt Simon, Vivek Sridhar and Thomas Verbraken tried to do for MSCI Inc. They built a stress-test model by analyzing what happened from December 2011 to May 2012, the glory days of the European sovereign debt crisis. You can read all about it on MSCI's web site. This may be an imperfect model, since the ECB's quantitative easing program was not in place during that period and the turbulence from a Greek exit may be much different than what happened then. (But we warned you about this in the headline!) 

The results of MSCI's stress test ain't pretty. The model predicts that all Greek public and private debtors, struggling with a rapidly devaluing new Greek currency,  would default on "virtually all outstanding bonds" and the nation's equities would plunge 80 percent.  

The results of the stress test predict that the impact would not be "devastating" for the rest of Europe, but it would still be pretty ugly: "widening peripheral yields, tightening core country yields, a weakening euro and a downward pressure on equity, in particular financial stocks." Haven assets like U.S., German, U.K. and Japanese bonds would benefit. 

Predicted returns and standard errors of stress-test model. Source: MSCI.

Predicted returns and standard errors of stress-test model. Stocks on left, sovereign bonds on right. (Typo in ACWI, or All-Country World Index, at top!) Source: MSCI.

Among the highlights, it predicts a 25 percent plunge in European bank stocks and surges of 200 basis points in Italian and Portuguese five-year government bonds. German five-year rates would sink 46 basis points, plunging back below zero. 

Something to think about as you model your plans for the weekend. 


Before it's here, it's on the Bloomberg Terminal.