Between 1997 and 2013 there were 24 sovereign bond defaults and debt restructurings in the global economy. According to Moody’s (2013: 6):

“[T]he losses imposed on creditors in sovereign restructurings have frequently been very large… Further, the variation around the average sovereign loss has been extremely high – losses have varied from as low as 5% to as high as 95%.”

What explains these large differences in ‘haircuts’? Why, for example, did investors in Uruguayan bonds suffer a 7% haircut in 2003, while those that had invested in neighbouring Argentina had losses in excess of 75% in 2005?

In a recent paper (Edwards 2015) I use data on 180 debt restructurings – for both sovereign bonds and sovereign syndicated bank loans – to analyse the determinants of recovery rates and haircuts. I use the results from the empirical analysis to evaluate whether some well-known episodes resulted in ‘excessively high’ losses. In particular, I focus on the Argentine restructuring of 2005, an episode that has generated controversy and that resulted in a US Supreme Court decision that is changing the way in which foreign debt contracts are written.

The analysis is in the spirit of the ‘excusable default’ model developed by Grossman and Van Huyck (1988). According to this work, sovereign debt is never repudiated. It is restructured when the debtor faces (very) bad states of the world. In this setting, investors may lose some of their money, but (almost) never all of it. The extent of losses, and the size of the haircut, depends on the severity of external shocks that hit the sovereign debtor.


On 23 December 2001, Argentina defaulted on its debt. Two weeks later the peso was devalued by 30%, and a ten-year experiment with a currency board and a fixed exchange rate ended.[1] In September 2003 the Argentine government made an offer to investors to exchange defaulted bonds for new ones. This proposal became known as the ‘Dubai guidelines’, and implied an average reduction of the face value of the debt of approximately 75%. Investors balked at the stiff losses, and asked for better conditions. Negotiations ensued, and a new offer, very similar to the original one, was formally made in June 2004 under the moniker of ‘Dubai plus’.

When the exchange window closed on 28 February 2005, 76.2% of bondholders had tendered their defaulted bonds and had accepted new bonds in exchange.[2] In 2010, Argentina reopened the bond exchange and offered identical terms as in 2005 to those that had not presented their defaulted securities. An additional group of investors decided to exchange their bonds. But not everyone came into the fold – bondholders representing approximately 7% of the original debt decided to hold on to the old securities and to press for better terms. On 16 June 2014 the US Supreme Court decided to leave in place a lower court ruling mandating Argentina to make a payment to the ‘holdouts’. After this ruling, frantic negotiations between the holdouts and the Argentine government began. By 30 July 2014, the deadline imposed by the Court, no agreement had been reached, and on 1 August 2014 the International Swaps and Derivatives Association declared Argentina to be in default.[3]

Using a ‘net present value’ approach, Cruces and Trebesch (2013) estimated that the average losses (across different bonds) incurred by investors that participated in Argentina’s 2005 exchange amounted to 76.8%.[4] In Edwards (2015) I adjusted this figure to take into account the value of a GDP-linked warrant that was offered with the new bonds. My calculation results in a haircut of 74.8%.

International comparisons

In Table 1 I provide summary statistics on haircuts for the complete sample (180 episodes) and for a number of subsamples. I also include the estimate for the Argentina 2005 haircut.[5] As may be seen, Argentine losses were significantly higher than the mean and median across all episodes (37% and 32% respectively), as well as across any of the subsamples.[6]

Table 1. Summary statistics for haircuts, 1978–2010

Mean Median Standard deviation
All episodes 37.0% 32.1% 27.3%
Bank loans 37.1% 37.6% 21.6%
Bond exchanges 36.9% 31.7% 27.9%
Africa 46.5% 39.5% 29.4%
Asia 32.6% 34.0% 17.9%
Europe 30.0% 19.7% 26.4%

In Figure 1 I present a histogram for the 180 haircuts. The value of the Argentine 2005 haircut (74.8%) is shown with a vertical black line. An analysis of Figure 1 and of the data behind it indicates that the distribution is ‘bimodal’, suggesting that the data may come from two different populations.[7]

Figure 1. Haircuts histogram, all episodes

Explaining haircuts

In Edwards (2015) I use an empirical model in the spirit of Grossman and Van Huyck (1988) to explain the variation of ‘haircuts’ across restructuring episodes. The following covariates were included:

(a) An index of ‘bad states of the world’ computed as the sum of four shocks – wars and coups d’état, output collapses, massive terms of trade declines, and currency crises;

(b) The ratio of debt restructured to GDP;

(c) Whether the country is poor;

(d) A number of global economy variables at the time of the debt exchange (recession, Treasury yields);

(e) Binary variables that capture the nature of the restructuring; and

(f) Regional dummy variables.

The regression results are satisfactory and are broadly in agreement with the ‘excusable default’ model. They indicate that countries that have suffered very severe shocks – including wars, armed conflicts, coups d’état, output collapses, and major declines in the terms of trade – end up having larger haircuts than countries that have not faced these major disturbances. Very poor countries and nations with larger debt burdens also have larger haircuts. The results are robust to variables’ definitions, periods considered, specifications, and estimation methods (White-corrected least squares or instrumental variables). The fit is quite adequate with the R-squared hovering around 0.6.

Was Argentina’s 2005 haircut excessive? A residuals analysis

These estimates may be used to inquire whether haircuts in particular episodes conformed to the predictions of the model, or if, on the contrary, they were excessively high or excessively low. This is what I do in this section for the Argentine exchange of 2005.

A good starting point is the analysis of fitted values obtained in the regression analysis. For Argentina’s 2005 exchange, the fitted values go from a minimum of 36.0% to a maximum of 60.1%. This range doesn’t include the actual haircut of 74.8%. The mean for fitted values from 20 regressions with different specifications is 47.1%, and the median is 45.7%; the standard deviation is 7.3%. Although these numbers are quite high – indeed, significantly higher than the mean and median for all episodes reported in Table 1 (37% and 32%) – they are still much smaller than the actual haircut imposed by Argentina on investors in the 2005 and 2010 debt exchanges.

I rely on two ‘influence statistics’ to investigate formally whether Argentina 2005 is an outlier in the empirical analysis: I use the R-student standardised test and the DFFITS test. In order to provide some context I also analyse the residuals from debt restructuring episodes in two of Argentina’s neighbours: Chile in 1984–1990 and Uruguay in 2003. In addition, I discuss briefly, and in light of these results, the Greek restructuring of 2012.

I computed 26 ‘influence statistics’ for the Argentine 2005 episode. The results obtained from this analysis are quite revealing. In 21 out of the 26 tests the Argentine 2005 debt restructuring is a statistical outlier.[8] In Figure 2 I present, as an illustration, the two residual tests – including the critical 95% bands – for a particular regression; see Edwards(2015) for details. According to the R-student standardised test, only three episodes are outliers: Argentina 2005, Bosnia and Herzegovina 1997, and the Ukraine 1998. Only the latter has negative residuals and, thus, an ‘unusually low’ haircut. The DFFIT test, on the other hand, identifies four outliers: Argentina 2005, Bosnia and Herzegovina 1997, Cote d’Ivoire 1998, and Iraq 2006. As may be seen, Argentina appears in both lists, indicating that the haircut imposed on investors in in 2005 was ‘excessively high’ from a comparative perspective.

Figure 2. Influence statistics and outliers

The residuals analysis shows that the haircuts in Chile’s (1984–1990) restructurings were ‘appropriate’, in the sense that the fitted values are very close to the actual haircuts. This is also the case for Uruguay’s debt exchange of 2003. In addition, an out-of-sample forecast suggests that for Greece in 2012, a haircut of 63% was consistent with the historical evidence and with the empirical model; this figure is very similar to the actual ‘haircut’ of 64% calculated for Greece by Zettelmeyeret al. (2013).

This article is published in collaboration with VoxEU. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with the Agenda subscribe to our weekly newsletter.

Author: Sebastian Edwards is a Henry Ford II Professor of International Economics at the University of California, Los Angeles

Image: Overview of the Buenos Aires’9 de Julio Avenue with the Obelisk in the background. REUTERS/Enrique Marcarian