The science of haircuts


The Art of Haircuts (illustration by Klaus Meinhardt http://www.klausmeinhardt.de)

In the good old days, haircuts were made in barbershops, not on financial markets. Until the late 80ies, economists writing about haircuts in the “American Economic Review” (AER) referred to real hairdressing. The first economist to use the term in the AER with regard to finance was Herbert Bear, a researcher with the Federal Reserve Bank of Chicago, in 1989. It took another 15 years and the default of Argentina until financial haircuts made it into “The Economist” for the first time, after all.

A haircut means a borrower partly defaults on his debt. When the Greek government owes you 100 Euro but pays back only 60, you’re suffering a haircut of 40 percent.

Courtesy to the Euro zone debt crisis, this meaning of the term has become common knowledge. For several years academic and private sector economists have been arguing that Greece should restructure its debt and partially default. Nouriel Roubini, a notoriously gloomy economist with New York University, thinks a haircut of up to 50 per cent is necessary. A few of days ago, the IMF urged private bondholders to share the burden, and even high level bankers like Martin Blessing, CEO of Germany’s Commerzbank, recommended a haircut of 30 percent. According to media reports, the European leaders are starting to think twice, too. Only the ECB staunchly fights against any kind of partial default by Athens. Anything else than a full and timely payback of debt would cause financial Armageddon, Europe’s central bankers warn.

Who is right? A new paper by Juan Cruces (Universidad Torcuato Di Tella, Argentina) and Christoph Trebesch (Hertie School of Governance, Berlin) sheds some light on this highly important question. While the two economists don’t address the current situation in Greece directly, they’ve analysed historical precedents of sovereign defaults.

Their paper, entitled “Sovereign Defaults: The Price of Haircuts” and presented by Cruces on the world congress of the International Economics Association in Beijing this week, yields a daunting conclusion for Greece: defaulting countries are punished harshly on the financial markets – and international lenders have an elephant’s memory. After a significant haircut, a government will find it extremely difficult to borrow new money on the capital markets, and if they get access again, they’ll have to pay prohibitively high interest rates.

Interestingly, these results are completely at odds with the previously published academic research on the same issue. Until now empirical economists who looked at the costs of haircuts always came to the conclusion that financial markets have an amazingly short memory. “Debts which are forgiven will be forgotten”, Jeremy Bulow and Ken Rogoff once summarised this consensus (for a good summary of the previous research see this  piece by “The Economist”).

According to Cruces and Trebesch, the old results were derived from very crude data and hence are misleading. Until now, no detailed database containing comprehensive information about the scale and scope of individual sovereign defaults existed. Therefore empirical economists trying to assess the impact of debt restructuring on borrowing costs had to make a simplifying assumption: they treated all haircuts equally. As Cruces and Trebesch explain in their paper:

“Previous papers attempting to gauge the effects of defaults on subsequent market access have used a binary default indicator, capturing anymissed payment as explanatory variable for past credit history.”

To overcome these shortcomings, Cruces and Trebesch embarked on a meticulous data collection effort. They relentlessly scraped together information about sovereign defaults between 1970 and 2010 using almost 180 different sources, including the archives of the International Monetary Fund, research by private banks and rating agencies, financial newspapers and 29 different lists on restructuring terms. According to their database, in the late four decades governments in 68 different countries defaulted on debt 182 times.

A brief look at the raw data shows that almost no haircut is like any other. On average investors had to write off 37 percent of their loans. However, below the surface, there is an amazing amount of divergence:

“Recent years have seen a particularly large variation, with some deals involving haircuts as high as 90% and others involving haircuts as low as 5%”, the authors state.

A lot depends on the precise design of the haircut. Only 57 cases involved a real reduction of the face value of the debt – here average losses were 65 percent. The big rest (123 cases) just affected the maturity of the debt and were less painful for investors. They had to write off only 24 percent of their money.

As long as Cruces and Trebesch follow the crude distinction “haircut / no haircut” used in the previous literature, they are able to replicate the old results. However, when they take the scope and scale of the individual haircut into consideration, the picture completely changes. Assessed form that angle, “credit market penalties for non-payment are much larger and longer-lived than previously thought”, the authors conclude.

On average the governments don’t have access to financial markets for about five years. “We find that exclusion time increases notably in haircut size”, the authors assert. As long as less than 30 percent of the debt has to be written off, more than 60 percent of countries return to the financial markets within two years and the average exclusion is little less than 2.5 years. However, if the haircut is bigger than 30 percent, on average it takes more than six years until the governments can borrow on the capital markets again. Only one third of those countries regain access within the first two years.

Even if they can return to financial markets, defaulters have to pay significantly higher interest rates. Compared to what governments from other emerging markets have to pay, a one percentage point increase in haircuts raises borrowing costs after one year by 6.8 basis points (bp) and after four years by 3.2 bp.

“This means that a haircut of 40 percent can be associated with 270 bp higher spreads in year one and 127 basis points higher in years four and five”, the economists assert.

Therefore, the conclusion of the paper is straightforward:

“Non-payments can have substantial adverse consequences for governments in the medium run”.

What does this all mean for Greece? The basic message is not that a haircut should be avoided per se. However, the paper shows quite clearly that even aside from the danger of contagion and the risk of another banking crisis in Greece and Europe, a partial default is associated with major risks for Athens and indirectly for the rest of Europe, as well. Greece would be excluded from the financial markets for quite a long time – this would make further austerity measures and possibly another bailout necessary. These costs of a default should be carefully weighted against the benefits (lower debt burden, participation of the private sector). If the conclusion is that a haircut is necessary, the implication of the paper is that Greece should do it as quickly as possible. Kicking the can down the road only exacerbates the misery and would imply even bigger haircuts in the future. The retaliation by the financial markets would be even more severe.

In a sense, the lessons from history are straightforward: Either do a haircut quickly, or keep your hands off completely.

8 Comments

Filed under Financial Crisis

8 Responses to The science of haircuts

  1. genauer

    Typical assumptions for CDS are a 40 % recovery rate, implying a 60 % haircut. In order to arrive at a debt of 60 % GDP would require cutting 60 % of the 150 % debt Greece has now, and if you look up 30 yr Greek bonds (ISIN: GR0138001673) you get a 40.6 % remaining market price (well, if you go a little bit more into details, with rates, this would change a little bit).
    Sooo, bring Greece to net income neutrality, let them default and exit the Euro, to the condition above, and they can somehow survive on their own terms.

  2. Marica Frangakis

    Nobody is actually (and seriously) discussing the eventuality of a haircut. Not even the greek government!! My hunch is ‘life will prevail’ in the end.

  3. Pingback: The science of haircuts » Greece on WEB

  4. Thanks for bringing this interesting analysis to your readers’ attention. It seems only rational for investors to demand higher interest rates from defaulters for years – a kind of credit version of uncovered interest parity. Although my analysis is much more crude, I came to similar a conclusion about inflationary repudiation of debt: http://reservedplace.blogspot.com/2008/01/it-is-often-asserted-eg-in-brad-setsers.html

    If financial markets regain over the long-term the return that they lose in the short-term, default or devaluation becomes merely another form of borrowing, or more pejoratively, kicking the can down the road.

  5. JV

    All the paper’s findings would be convincing… if God woke up one morning, and told one country to do a 10% haircut, and the next one, 50%. Of course, that is not what happens. Countries in worse shape end up with bigger haircuts. And guess what? They also find it harder (and more expensive) to borrow …

  6. Pingback: Financial Haircuts | Stephen Hannah

  7. interesting read about “haircuts” and defaults:

    “The Verboten Story of Argentina’s Post-Default Economic Success”
    http://www.nakedcapitalism.com/2011/10/the-verboten-story-of-argentinas-economic-success.html

    my favorite quotes:
    “Notice the Argentinian example disproves one of the Big Lie about default, that foreign capital will take a hike and the consequences will be dire”
    and
    “Shorter version: sacrificing your economy on the altar of the Bond Gods may not be such a good idea.”
    discussion?

  8. Esther

    Question should not be what does it mean for your acces to financial markets, but what does it mean for you development? Can those countries find alternative growth strategies? Argentina seems to be doing so….