AMF exam: two categories of questions, a threshold with no averaging

The examination certified by France’s financial markets regulator, the Autorité des marchés financiers (AMF), has a reputation as a test you can cram for. It is partly deserved: it is a multiple-choice questionnaire covering a defined, public body of knowledge. But that reputation obscures the feature behind most failures — the pass threshold is not computed on an overall score.

Two sets of knowledge, not a single syllabus

The content of the minimum knowledge is defined by the AMF after an opinion from the Financial Skills Certification Board. It splits into two sets, which share neither their nature nor their mode of learning.

The regulatory and ethical environment. The institutional framework, the players, professional obligations, ethics, anti-money-laundering, client protection, market abuse. This is a legal strand: it is retained through the structure of the texts, not through intuition.

Technical knowledge and financial literacy. Financial instruments and how they work, markets, third-party asset management, savings taxation, post-trade mechanics. This is a practitioner’s strand: it is anchored in daily practice, and an experienced professional often has a comfortable lead.

Questions asked in the test belong to one or other of these two categories. They come from a common examination database shared between certified providers: each provider supplies and maintains its own questions and answers and may in return use the whole database for its activity. The database is updated at least once a year, to keep pace with regulatory and market developments.

The threshold: category by category, no averaging

This is the point to absorb before starting to revise. The pass threshold is assessed on each of the two categories separately. There is no average between them.

The consequence is counter-intuitive for anyone raised on school exams: a high overall score guarantees nothing. A candidate who crushes the technical part and lands just below the threshold on the regulatory side fails, exactly like one who reached the same overall score by being mediocre everywhere.

This is not gratuitous rigidity. The purpose of the scheme is to verify a knowledge base, not to rank candidates: a base, by definition, cannot be averaged. There would be no point certifying an excellent technician who is unaware of client-protection rules.

Candidate profile	Usual strength	Failure risk
From the markets	Instruments, mechanics, post-trade	The regulatory and ethical strand, often encountered systematically for the first time
From compliance or law	Regulatory framework, ethics	Financial instruments, valuation, market mechanics
Recent graduate	Neither acquired through practice	Uneven allocation of revision time

What that changes about preparation

The method that fails most often is to revise “the exam” as a single block, training on mixed questionnaires and watching the overall score rise. That score hides precisely the information that matters: the level reached in the weaker category.

Three habits are enough to correct course.

Treat the two categories as two separate exams. Two revision plans, two sets of practice, two measures of level.

Measure category by category. At every practice session, record two scores, never one. The progress that counts is that of the lower of the two.

Over-invest in the uncomfortable category. This is the opposite of what everyone does instinctively: people revise what they enjoy, because that is where practice scores look good. Revision time must go where the threshold is at risk.

Practising on questionnaires: a method, not just a measure

There is a documented reason to multiply practice tests rather than re-read your materials. Henry Roediger and Jeffrey Karpicke, in “Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention”, published in 2006 in Psychological Science, showed that testing yourself on content produces markedly better retention over time than repeatedly re-reading the same material. The effect grows as the delay between learning and testing lengthens — exactly the situation of a candidate preparing over several weeks.

One precaution applies to multiple-choice questionnaires, though. Andrew Butler and Henry Roediger, in “Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing”, published in 2008 in Memory & Cognition, showed that the format can also install false knowledge by exposing learners to plausible incorrect options — an effect that feedback, immediate or delayed, sharply reduces (see the study).

In practice: training without reading the answer explanations is the worst way to prepare for a multiple-choice test. Time spent on the explanation of a missed question is worth more than the next question.

Where the test itself takes place

The examination is sat with a training provider whose exam has been certified by the AMF and which appears on the list published by the regulator. That provider runs the session under controlled conditions, marks it and issues the pass certificate. The AMF, for its part, certifies the provider’s examination — not individuals.

Our article on the listed functions and the six-month deadline sets out who must be verified, the one on portability of the certificate explains why the certified exam usually beats internal verification, and the one on the sustainable finance module covers the additional test devoted to responsible investment.

Take action

Before your first practice session, split your materials into two piles — regulatory and ethical on one side, technical and financial literacy on the other — and set yourself two separate score targets. Then, after every set, systematically read the explanations for the questions you missed before moving on. The full scheme sheet, with the steps and frequently asked questions, is here: AMF professional certification.