By Björn Elberling

The first topic I will blog about in some detail, which is a bit of a byproduct of my dissertation, concerns a question of sentencing in international criminal law, namely sentencing in the context of guilty pleas. (The picture to the left shows Stevan Todorovic, one of the first accused to be sentenced after a guilty plea based on a plea agreement) 

Sentencing generally is a very contentious issue, especially for the ICTY, which is often criticized as passing much too lenient sentences – not only, but especially where sentences are the result of a guilty plea based on a plea agreement between prosecutor and accused. The common feeling seems to be that plea bargains have resulted in unduly lenient sentences, that truth and justice have been traded away in return for savings in time and judicial resources, to the detriment of the Tribunals’ legitimacy.

I will try to find out to what extent this is actually the case.

In this first post and the following one, I will deal with the ICTY. A later post will consider the ICTR, and I will conclude with some predictions for future trials, especially before the ICC. Some of this may become rather technical, but I do hope that it will show some interesting tendencies, so please bear with me…

I do not plan at this point to get too deeply involved in the debate whether plea bargaining in trials concerning core crimes is generally a good thing – quite a number of arguments have been made to say that it is not. Such arguments come, e.g., from a historiographical perspective according to which the most important goal of the Tribunals is laying down a historical record of the conflict for posterity. This goal may be hurt by a tendency towards “minimalistic” plea bargains which establish just enough facts as are needed for the conviction of the individual accused. Especially when combined with charge bargaining, i.e. the dropping of some or most charges, this can lead to a very limited consideration of the historical facts (See para. 6 et seq. of the Dissenting Opinion of Judge Schomburg to the Trial Chamber Judgment in Deronjic). Similarly the practice could be criticized from a Defence perspective – I will say something about such possible criticisms when dealing with the ICTR.

However, it seems to me that the perception of excessively lenient sentences is one of the most widespread grounds for criticizing plea bargaining before the ICTY. I was not so sure whether this perception was factually correct, so I set out to establish what sentence reduction an accused may actually hope to receive in exchange for a plea of guilty (and for attendant mitigating factors such as saving of time and judicial ressources, acceptance of responsibility, helping establish the historical record and, frequently though not necessarily, statements of remorse).

Now, trying to establish this “discount” by looking at individual sentences is hard to do as Chambers rarely attach any numerical weight to the severity of the crime and to aggravating and mitigating factors; they usually just list all these factors and then “magically” come up with a sentence (For some exceptions, see the next post).

I was also unsuccessful in trying to find pairs of cases comparable in terms of severity of the crime and general mitigating and aggravating factors and differing only in that one accused had pled guilty while the other had not. This attempt failed, on the one hand, because it is hard to find any two cases that are actually comparable with regard to all these factors (as the Chambers themselves often point out in their judgments). Additionally, sentences before the ICTY probably depend not only on these factors, but also on the national and professional background(s) of the judges involved in sentencing, so even comparable cases will not necessarily lead to comparable sentences.

There is, however, one method which allows at least some tentative remarks about sentencing “discounts”, namely that of comparing the sentences of all accused who pled guilty with those of all accused convicted after a full trial. The results of this comparison are set out in the following table (Some technical remarks and possible caveats at the bottom of the post):

The table shows, first, a very significant overlap in sentencing ranges for the two groups.

Second, the difference in sentences – the tentatively established average sentencing “discount” – is about 15.7%. This number is rather low, not only when considered on its own, but probably also in comparison to other jurisdictions: I do not have any hard data at hand, but I know that the “rule of thumb” under German law – which is traditionally not very keen on plea bargaining and which allows hardly any charge bargaining at all – is a “discount” of between 20 and 30 percent (as stated in the standard commentary on Criminal Procedure by former Supreme Court Judge Lutz Meyer-Gossner (at p. 795) – for our German readers out there: vor § 213, Rn. 16 mwN).

Third, and maybe most importantly, this difference in average sentences is not due to guilty pleas leading to particularly low sentences – the percentage of defendants sentenced to particularly low sentences of less than 10 years is actually lower for those who plead guilty than for those who don’t.
The main difference, however, is that accused who plead guilty only rarely receive a particularly high sentence of 20 years or more. This could be taken to show that plea-bargaining does make a difference for high-ranking defendants who have something to offer the court in terms of admissions which are valuable for the history books, while for the “smaller fish” it may not make all that much of a difference.

So much for now – more in the next post, including a look at some individual “big fish”.

(The promised technical remarks and caveats:

1. As stated, the data set on which this table is based consists of all final sentences as of today. While all sentences based on plea agreements are final, there are at least eight cases where appeals proceedings are ongoing against conviction after a full trial (namely the cases of Mrksic and Sljivancanin, Hadzihasanovic and Kubura, Oric, Martic, Dragomir Milosevic, and Krajisnik). Barring any drastic deviation from the Trial Chamber judgments on Appeal, however, these cases should not skew the picture presented here: If the Appeals Chambers upheld the convictions and sentences in all cases, the average sentence for these eight would be 16.2 years – slightly below the average of final sentences for the group –, with an equal number of sentences well below and well above the median of 15 years for the group.

2. I had to decide what numerical value to attach to the life sentence imposed on General Galic. Given that the ICTY does not consider a life sentence to be an automatic sentence for the remainder of the convicted person’s natural life, I considered this sentence as equivalent to 45 years, i.e. “only” 5 years longer than the highest non-life sentence of 40 years.

3. In some of the guilty plea cases, Chambers also considered mitigating factors which are independent of the guilty plea itself, such as “character before the conflict” or “behavior in the Detention Unit”. Such factors, however, also exist in cases of defendants who had not pled guilty, also they are probably not given too much weight anyway.

4. The main caveat is that this table only presents meaningful data if, on average, the cases leading to conviction after plea bargaining and the cases leading to conviction after full trial are comparable as to the severity of the crimes and other mitigating and aggravating factors. If there is some sort of selection bias in that, e.g., only those accused of lesser crimes tend to plead guilty, any difference in sentences can no longer be clearly attributed to guilty pleas. Keeping in mind what I said above about the problems in comparing cases before the ICTY, I would still say that in general, the two groups should be at least roughly comparable – both contain, e.g., political leaders, high-ranking officers of the Bosnian Serb forces involved in the mass killings at Srebrenica, and low-ranking camp guards convicted for crimes committed against a number of individual victims. So I would tend to hold the two groups to be roughly comparable with regard to crimes committed, at least enough so that the comparison should show some general tendencies. Besides this general impression, two data points from the table also tend, in my view, to show that the groups are roughly comparable: One, there is a significant overlap between the sentencing ranges; two, not only the average, but also the median sentences are rather comparable.

5. Speaking of crimes, one last remark: In at least some of the guilty plea cases, there was substantial charge bargaining, which might be taken to disturb the picture of comparable cases in terms of severity of crimes. However, one can also take this reduction in charges – which should generally also lead to a reduction in sentence – as part of the general sentencing “discount” that I am considering here.)