– How the judiciary deals with its own mistakes –
The purest form of madness is to leave everything as it is and to hope that something will change.
Albert Einstein
Thomas Fischer, a judge in criminal matters at the Federal Court of Justice (Bundesgerichtshof, BGH), took the Gustl Mollath case as an occasion in August 2013 to reflect on the foundations of criminal law. The former presiding judge of the 2nd Criminal Division (Strafsenat) and probably the most influential criminal lawyer in Germany wrote in the weekly newspaper “Die Zeit” (“Wahn und Willkür”, in: “Die Zeit” of 22 August 2013, p. 13): “Culprits for mistakes are quickly found. For the rules, we are all responsible.”
This assertion is only half true. For in order to find culprits, one first has to recognise the mistakes. And it is precisely this that is systematically prevented in German criminal law.
What the Higher Regional Court of Nuremberg left open
For the public, once the HypoVereinsbank report became known, the matter was quickly settled: Judge Brixner had made himself guilty of perversion of the course of justice (Rechtsbeugung), the experts were corrupt, and one or other of those involved was perhaps even in thrall to the international banking mafia.
All of that is possible, but none of it has been proven. One ought to exercise the same restraint in apportioning blame to those involved in the proceedings as one owes to every defendant in criminal proceedings.
It is true that the application for the reopening of proceedings which the lawyer Gerhard Strate filed with the Regional Court of Regensburg (Landgericht Regensburg) in February 2013 (brief to the Regional Court of Regensburg of 19 February 2013) was based, among other things, on the allegation of perversion of the course of justice by the presiding judge, Brixner. In a meticulous analysis of the case files, numerous errors in the preparation of the proceedings and in the conduct of the main hearing were set out in support of this allegation.
The decision of the Higher Regional Court of Nuremberg (Oberlandesgericht, OLG), which in August 2013 (OLG Nürnberg of 06.08.2013, case no. 1 Ws 354/13 WA) ordered the immediate release of Mr Mollath, was however based exclusively on a chance ambiguity in a medical certificate. One gains the impression that the judges at the Higher Regional Court were glad to have found an innocuous ground for reopening the proceedings. It allowed them to draw, with genteel discretion, a veil of silence over the personal misconduct of those involved in the proceedings which hung in the air.
As welcome as the outcome of the Higher Regional Court’s decision is, the approach taken nevertheless reveals a fundamental problem of German criminal procedure. Instead of making an effort to learn from mistakes, they are covered up and hushed up as far as is humanly possible. It can be assumed that the Regional Court responsible for the retrial will follow this line.
The allegations against those involved in the proceedings, which Mr Mollath’s supporters had assembled in painstaking research, will consequently never be clarified in proceedings governed by the rule of law. There will be no lessons as to how such errors are to be avoided in future. Still less will there be any personal consequences for the individuals involved.
Editorial note (as of September 2026): This article was written in 2013. The reopened proceedings ended on 14 August 2014 with an acquittal by the Regional Court of Regensburg (case no. 6 KLs 151 Js 4111/2013 WA). The fundamental reflections on error culture are unaffected by this.
Legal certainty?
The road to a criminal judgment is long and stony. Before it is reached, many hurdles have to be overcome, hurdles which also serve to avoid wrongful judgments. Whole libraries have been written about the possibilities of defence during ongoing proceedings, such as the right to file motions to take evidence or legal remedies. It is all the more astonishing that, in German criminal procedure, hardly any thought has so far been given to the lessons to be drawn when a wrongful judgment has nevertheless been handed down.
Imagine that a train accident occurred with dozens of people killed and that Deutsche Bahn took the view that, in order not to endanger the public’s confidence in the safety of the rail network, the causes which led to the accident should deliberately remain hidden. Such an attitude would be particularly absurd if it were even forbidden to name or to analyse mistakes made by the engineers and the staff – on the grounds that nobody would then dare to travel by train any more.
It is precisely such a “lid-on mentality” that underlies the German justice system.
It is true that there is the possibility of subsequently correcting a judgment in favour of a defendant by way of a reopening of proceedings. In the law governing the reopening of proceedings, however, the guiding idea is justice in the individual case. Unless it happens to become known after the event that a judgment is blatantly unjust in its outcome, even the most glaring errors of the courts end up in the graveyard of oblivion.
Under the vacuous slogan of “legal certainty”, it has been systematically prevented since the century before last that errors in court proceedings are named after the event – let alone corrected.
The extremely restrictive way in which the grounds for reopening proceedings are framed in statute – and the far more small-minded handling of the statutory requirements in practice – is justified by the argument that it is in the interest of “legal peace” (Rechtsfrieden) and of “the public’s confidence in the judiciary” that wrongful judgments should as rarely as possible be described as such after the event.
It is astonishing that such a view, which reaches far back into the century before last, has been able to survive into our own time almost unchallenged.
Secrecy of judicial deliberations (Beratungsgeheimnis)
In criminal trials at first instance before the Regional Court (Landgericht), the bench is made up of lay people as well as professional judges. These so-called lay judges (Schöffen) have no legal training. In the court’s secret deliberations, their vote carries the same weight as that of the professional judges.
One of the lay judges who took part in the trial against Mr Mollath in August 2006 described to a television team in 2013 how Judge Brixner reacted at the hearing whenever the ex-wife’s dealings in undeclared money (Schwarzgeld) were raised (see on this, for example, Süddeutsche Zeitung of 15 November 2012, “Nach Unterbringung in Psychiatrie- Schöffe kritisiert Mollath-Verfahren”). The presiding judge would then make clear, in a choleric tone, that the trial was about bodily harm and criminal damage, but not about money laundering.
Such an attitude – assuming here, for the sake of argument, that the lay judge’s account is true – would of course be absurd. For the question whether the sole witness for the prosecution had a motive for “sending her ex-husband to a psychiatric hospital” is of considerable importance for her credibility. Moreover, Mr Mollath’s allegations about undeclared money – which later turned out to be correct in essential respects – were used in the judgment as the point of reference for the defendant’s supposed delusions (see the original judgment from 2006). For that reason alone, the court should not simply have passed over the matter.
The lay judge who has now spoken out after the event did not tell the journalists whether he also voiced his unease about the way the presiding judge conducted the proceedings back in 2006 – during the court’s internal deliberations on the judgment. He is not allowed to say anything about that. He is bound by what is known as the secrecy of judicial deliberations (Beratungsgeheimnis).
Nothing about the content of the deliberations and the outcome of the vote becomes publicly known. Under the law as it stands, these are internal matters of the court which are nobody else’s business.
Under the law as it currently stands, the secrecy of judicial deliberations may be breached only in rare exceptional cases, for instance where a judge involved has committed the offence of perversion of the course of justice. In view of the immensely high legal hurdles that would have to be cleared in order to prove perversion of the course of justice against a judge (in the Mollath case, no court and no public prosecutor’s office has so far accepted that these hurdles were cleared), it does not happen in reality that a judge breaks out of the collective of silence.
In the Mollath case it will never be clarified whether the subject of undeclared money was discussed in the judges’ deliberations. Just as little will the public ever learn whether all four judges agreed that Mr Mollath should be locked away in a psychiatric hospital.
Experience does not always make you wise
In everyday court practice it is extremely rare for an associate judge or lay judge to speak up and explicitly object to the manner in which the presiding judge is conducting the hearing.
It can be assumed that the presiding judge is also the person who exerts the most influence in the internal vote on the judgment. For as a rule he has the most professional experience.
Judge Brixner, the presiding judge of the 7th Criminal Chamber (Strafkammer) which handed down the judgment against Gustl Mollath in 2006, was an experienced – and in a certain way also charismatic – judge.
In a great many professions it is indeed an advantage to have people at work who possess a great deal of experience. This is likely to be the case, for example, with brain surgeons or pilots.
Their activities are characterised by three factors:
- They work in an environment in which potential dangers are sufficiently regular to be predictable.
- Anyone who makes an individual mistake will recognise it immediately, because he receives instant feedback. There is an opportunity, through many years of practice, to become aware again and again of one’s own proneness to error.
- Every mistake is immediately analysed by a broad public and by the international professional community. Potential sources of error are documented and published.
Cerebral haemorrhages or thunderstorms can, admittedly, never be predicted with certainty. There are, however, a whole series of statistical empirical values which allow a pilot, for example, to watch out for anomalies in the weather and to react to them at an early stage. If on some occasion he fails to do so, his aircraft will get into drastic danger.
Co-pilots are under instructions to intervene immediately if the rules are breached and to document without mercy every failing on the part of the person primarily responsible.
If a crash – or even just a specific dangerous situation – does occur, an army of technicians and scientists all over the world will do everything possible to establish the causes. The more professional experience a pilot brings with him, the more knowledge about the mistakes of colleagues he acquires in the course of his life. This enables him to develop a finer feel for his own human shortcomings.
The situation of a judge is a completely different one. When it comes to assessing the truth of a witness statement, there is no regularity. For years psychologists have been trying to develop tangible criteria by which a witness statement based on facts can be distinguished from one that rests on a deliberate lie or – something that happens by no means rarely – on an error. The criteria developed in this way have in the meantime also found their way into the case law (see, as the leading authority, the judgment of the BGH of 30 July 1999, case no. 1 StR 618/98).
Gut feeling
In the end, however, when assessing a person’s credibility, judges still often listen to their gut feeling (see on this the study by M.K. Dhami, Psychological models of professional decision making, in: Psychological Science (2003), 14, pp. 175-180). Nothing else applies where a prognosis is to be made as to whether a person will in future commit criminal offences so serious that they justify a long-term deprivation of liberty.
The problem with this is: intuition is a poor adviser!
Just as people are easily deceived by optical illusions or conjuring tricks, the processes that take place when one judges another person from the gut are equally treacherous. It is no different from the first encounter with a potential great love. In 90% of all cases it is astonishing how the brain can correctly grasp fundamental characteristics of a fellow human being within fractions of a second. In the remaining 10%, intuition can be led onto thin ice – with disastrous consequences.
When it comes to predicting the dangerousness of a defendant or assessing the credibility of a witness, judges are no more experienced than lay people. For even if they have already handed down numerous judgments, they lack the direct feedback by which they could develop their abilities further (see on this the dissertation by Mark Daniel Schweizer, “Kognitive Täuschungen vor Gericht. Eine empirische Studie” [Cognitive illusions in court. An empirical study] (Zurich 2005), p. 262 f.: “The calibration of experts is typically superior to that of lay people only where the expert receives rapid, unambiguous feedback as to whether his prediction was correct – as is the case, for example, with meteorologists.” With judges this is not the case).
There is nevertheless a significant difference between professionals and lay people: the supposed specialists are far more convinced of their own judgement than the non-experts are. In other words: the more experienced a judge is, the greater the danger that he overestimates his own competence.
A study in which the susceptibility of German judges to the so-called anchoring effect was measured can be found in Englich, “Blind or Biased? Justitia’s Susceptibility to Anchoring Effects in the Courtroom Based on Given Numerical Representations”, LAW&POLICY (2006), pp. 497 ff. (508). None of the judges tested was aware of his own susceptibility. The more experienced a judge was, however, the more he was (mistakenly!) subjectively convinced that his decision had been uninfluenced by the anchor given to him. One speaks of the anchoring effect where an estimate is influenced by a figure that was mentioned shortly beforehand (example: “How many litres of fuel fit into a jumbo jet? Is it more than 3,000 litres” or “Is it more than 300,000 litres?” Depending on the way the question is framed, the probability is very high that the estimate will be influenced by the figure given. It is no different with judgments (“Will the defendant receive a custodial sentence of more than seven years”).
The fact that those convicted end up in prison or in a psychiatric hospital is not a reliable indicator. For in this respect a judge is in no different position from a doctor who, at the beginning of the 20th century, believed that he could intuitively sense when patients were about to develop typhoid. The doctor checked his diagnosis by palpating the tongues of the suspected patients one after another, without washing his hands in between. After all the patients who had been examined in this way died sooner or later of the deadly infectious disease, the doctor was profoundly convinced that he possessed an unerring diagnostic instinct (the example comes from Lewis Thomas and is quoted by Daniel Kahneman in his bestseller “Thinking, Fast and Slow” (German edition “Schnelles Denken, Langsames Denken”, 2011), p. 297).
It would be no different for a judge who bases his subjectively felt instinct for truth and justice on the fact that everyone he has convicted was indeed locked away.
The other conditions that would help one to grow wise through experience are likewise absent in the case of judges. The “co-pilots” – that is, the associate judges and lay judges – are not merely given no particular encouragement to name a colleague’s mistakes. Under the slogan of “secrecy of judicial deliberations”, they are actually forbidden to do so.
But the worst thing of all is – and this makes the job of a judge one of the most difficult in the world: a judge will receive reliable feedback on whether his decision was right only in rare exceptional cases, if at all.
In “Thinking, Fast and Slow” (2011) Daniel Kahneman summarised the conditions that have to be met for the acquisition of expertise (p. 296): (1) An environment that is sufficiently regular to be predictable. (2) An opportunity to learn these regularities through prolonged practice.
If a pilot makes a mistake, he will notice immediately that the aircraft begins to lurch. A brain surgeon who overestimates his own abilities will be taught better at the latest when the patient dies of a cerebral haemorrhage. A judge, by contrast, will never find out which of the defendants he has convicted in the course of his life were in reality innocent (or only partly guilty).
Anyone who has lost his liberty through a judicial error and is reasonably shrewd will keep his mouth shut. For if, after years of imprisonment, he still insists on his innocence, he has hardly any chance of early release or of a favourable prognosis in the context of a psychiatric assessment.
Nor does any scientific correction of errors take place. Even if judicial misconduct were to become known after the event (as seems likely in the Mollath case), there is no institution that deals systematically with working through and eliminating potential sources of error.
A judge at the Regional Court is, admittedly, subject to review on an appeal on points of law (Revision). Whether a witness statement is correct is, however, not examined by the Federal Court of Justice in the appeal proceedings on points of law. Germany’s highest criminal court confines itself to pointing out technical errors in the written reasons for the judgment. Thus, as his professional experience increases, a judge may possibly become better at presenting skilfully what he personally considers to be right. He will therefore run less and less risk of being overturned by the appeal court. The degree of truth is not increased by this mechanism. Rather the opposite is the case: the more precisely a judge knows what he has to write in order for his judgment to “stand up”, the more skilfully he will take care about what he puts into the reasons for the judgment – and what he does not.
In Thomas Darnstädt’s book “Der Richter und sein Opfer. Wenn die Justiz sich irrt” [The judge and his victim. When the judiciary gets it wrong] (2013) there is an interview with the presiding judge of the 2nd Criminal Division of the Federal Court of Justice, Thomas Fischer, who concedes (p. 278): “In the appeal on points of law it is even less a matter of direct access to the truth than at the fact-finding instance; almost always it is only about how the truth is presented in the judgment – which can be done well or badly – regardless of what actually was.”
A short history of the catastrophe
What follows is a proposal as to how the finding of truth in court could be improved without great effort. It is true that the antiquated concepts of “legal certainty” and “secrecy of judicial deliberations” are so deeply rooted in the minds of German lawyers that the proposal may at first sight appear bold. But even the longest journey begins with the first step.
The psychologist and winner of the Nobel Prize in economics Daniel Kahneman describes in his book “Thinking, Fast and Slow” a method by which entrepreneurial decisions can be subjected to critical scrutiny before it is too late.
Daniel Kahneman, “Thinking, Fast and Slow” (2011), p. 326. The so-called pre-mortem method goes back to Gary Klein. See also Kahneman, Lovallo, Sibony: “The Big Idea: Before You Make That Big Decision”, Harvard Business Review, June 2011
One need only engage in a hypothetical thought experiment: “Imagine that we are a year into the future. We have implemented the plan in its present form. The result was a catastrophe. Please take five to ten minutes to write a short history of that catastrophe.”
The starting position when the management board of a public limited company (Aktiengesellschaft) takes a vote is entirely comparable with the one in which the chamber of a Regional Court finds itself when deliberating on its judgment. Several people take a momentous decision. There are things one knows, there are things one does not know, and there are things one does not know that one does not know (the internal audit report (Revisionsbericht) of HypoVereinsbank in the Mollath case in 2006 is likely to count among the latter). A charismatic person in the chair sets the direction, the public expects quick results, and every delay would be expensive.
The method ensures that all the facts are identified and evaluated. Justified doubts are given a hearing; they are not drowned out by group pressure. The danger that faulty information, prejudices or extraneous motives will influence the decision is minimised.
The model can be transferred to court decisions. Every participant – professional judge and lay judge alike – should answer a number of questions in writing and without mercy, on the premise that the judgment which is to be handed down today will turn out in a few years’ time to be a devastating miscarriage of justice. In a thought experiment, then, a kind of catastrophe report (Katastrophen-Protokoll) is to be drawn up, setting out the (so far merely hypothetical) history of a wrongful judgment.
Which facts were ignored or left out of account? Which participant in the proceedings could have prevented the wrong decision, and in what way? Which procedural rules were disregarded? To what extent is one’s own presentation of the evidence incomplete, affected by errors of reasoning or shaped by prejudice?
Imagine that such reports had been drawn up during the deliberations of the 7th Criminal Chamber of the Nuremberg-Fürth Regional Court in August 2006. It does not take much imagination to picture what they might, for example, have contained.
“Even before the trial began, Judge Brixner had made clear to the tax authorities that he regarded Mr Mollath as a ‘crackpot’, and thereby prevented the tax investigators from taking the indications of dealings in undeclared money seriously.”
“Obvious motives on the part of the sole witness for the prosecution for ‘getting her husband into a psychiatric hospital’ were ignored.”
“It simply fitted the picture too well that an awkward person must also be a liar.”
What I have in mind is to make such a report compulsory for judges, lay judges and also public prosecutors. Defendants and defence counsel – but also expert witnesses, accessory prosecutors and witnesses – should in addition have the possibility of drawing up a corresponding report voluntarily.
The reports should be taken into the custody of the court and, initially, should not be accessible to anyone. They take effect simply by virtue of being drawn up at all. And this for two reasons:
First, members of the court with little assertiveness would develop more courage to voice their misgivings.
Secondly, none of those involved would feel unassailable any longer. Everyone would have to reckon with the possibility that the mistakes he makes today might one day catch up with him after all.
Such an awareness would lead to considerably greater care than can be expected of a judge who is able to assume that he will never have to take personal responsibility for his own decision.
The “catastrophe reports” drawn up in this way could be made available to the defendant after a certain period of time has elapsed. It would then be his free decision whether they are destroyed or, for example, made available to a research body.
Ralf Eschelbach, a judge in criminal matters at the Federal Court of Justice, described the rumour that there are “hardly any wrong criminal judgments” as “the judiciary’s life lie (Lebenslüge)” (see Spiegel-Online of 24 April 2013: “Justizirrtümer: Blind vor der Wahrheit”; see also Sabine Rückert, in Die Zeit of 11 July 2011, “Lügen, die man gerne glaubt”).
A scientific evaluation of the “catastrophe reports” could offer an indication of how serious this life lie really is.