CAS Exam 7 just produced the lowest pass rate in its history. The Spring 2026 sitting passed 152 of 642 candidates, a raw rate of 23.7%, the weakest result the exam has posted since it reached full scale more than a decade ago. And it is not a one-off: Exam 7 has fallen at each of its last three sittings, from 57.1% in Spring 2024 to 49.1%, then to 31.1% in Fall 2025, and now to 23.7%. For an upper-level exam that stands between candidates and Fellowship, that is a steep and fast decline.
The reaction has been predictable, and so has the confusion: people who felt they wrote strong papers failed, and no change to the syllabus explains it. The explanation is structural. Exam 7's pass rate is not a fixed bar you clear by answering a set percentage correctly. It is the output of the CAS's consistent-standard grading, in which the pass mark is set after each sitting to hold the competence bar level. When a sitting is hard, the pass rate falls even though the standard did not move. This piece uses the CAS's own sitting-by-sitting data and its documented grading process to explain the drop, and what it means if you are sitting for Exam 7 next.
What Exam 7 Actually Tests
Exam 7, Advanced Estimation of Claims Liabilities, is an upper-level exam on the FCAS track, and after the CAS Admissions Transformation Plan reworked the fellowship exams it became almost entirely focused on advanced reserving; the current content outline weights it fully toward estimation of claims liabilities. Since the CAS moved Exams 5 through 9 to computer-based testing in 2020, it is delivered through Pearson VUE as a mix of item types: multiple choice and multiple selection, fill-in-the-blank, point-and-click, matching, spreadsheet calculations, and written constructed-response questions, with an essay-response type added for the 2026 sittings. For all that variety, the upper-level material still turns on the spreadsheet and written work, where you show your calculations and justify your selections and a human grader scores what you wrote.
Two features make it a genuine milestone rather than a formality. First, the content synthesizes across multiple readings, so answers that recite one method in isolation fall short of a full-credit response. Second, it is graded to a mastery standard, which is where the pass-rate volatility comes from. Beginning in 2025 the CAS offers Exam 7 twice a year, spring and fall, up from a single annual sitting, so retakes come around faster than they used to. Multiple attempts are common, and treating a first attempt as a likely single try is the wrong mental model for most candidates.
The Pass-Rate Record, Verified
The numbers below come directly from the CAS's own exam statistics, which run from 2011 through the Spring 2026 sitting. They are worth seeing in full, because the recent drop is easy to mistake for a blip when it is actually a trend. Both the raw pass rate and the CAS's effective rate, which screens out papers too incomplete to be realistic passes, are shown. For the interactive version of this history, with the chart, sortable table, and computed statistics, see our Exam 7 pass-rate tracker.
| Sitting | Candidates | Passed | Raw Rate | Effective Rate |
|---|---|---|---|---|
| Spring 2026 | 642 | 152 | 23.7% | 25.6% |
| Fall 2025 | 639 | 199 | 31.1% | 33.7% |
| Spring 2025 | 745 | 366 | 49.1% | 53.7% |
| Spring 2024 | 785 | 448 | 57.1% | 60.8% |
| Spring 2023 | 572 | 250 | 43.7% | 47.3% |
| Spring 2022 | 617 | 348 | 56.4% | 61.1% |
| Highest on record (Spring 2014) | 300 | 173 | 57.7% | 58.8% |
Across the sixteen full-scale sittings from 2012 through Spring 2026, 4,345 of 9,864 candidates have passed, a cumulative raw rate of 44.0%. The effective rate runs a few points higher, but it has fallen just as hard: Spring 2026's effective rate was 25.6%, down from 60.8% only two years earlier. By whichever measure you prefer, this is the weakest stretch the exam has seen.
Against that history, 23.7% is not ordinary statistical noise. The exam's high-water mark is Spring 2014 at 57.7%, and it touched the high 50s as recently as 2024, so Spring 2026 sits at the very bottom of the range, below even the 29.5% of 2013. The standard deviation across these sittings is about 10 percentage points, which is wide, but the more important pattern is the direction: three consecutive declines, with no change to the syllabus or the exam format over that span.
Why It Swings: the Consistent-Standard Pass Mark
The central fact about Exam 7, and the one that explains almost everything else, is that the CAS does not use a fixed passing percentage. In the CAS's own words, the objective is to evaluate candidates against "criteria for demonstrating adequate knowledge that remains constant throughout the lifetime of the exam series," and "preset pass marks," meaning a rule that you pass if you answer some fixed percentage correctly, "are counter to this philosophy."
Instead, the pass mark is set for each individual sitting. A panel of experienced committee members evaluates the exam and determines the score a minimally qualified candidate should be able to reach, and the Syllabus and Examination Committee uses that benchmark to select the final pass mark after the exam has been administered and graded.
That single design choice produces the volatility. If a sitting is comparatively easy, the pass mark rises to hold the standard constant; if it is hard, the mark falls and so does the share of candidates who clear it. The competence bar stays level while the raw pass rate floats up and down around it. So Spring 2026's 23.7% does not mean the CAS suddenly turned harsh, and Spring 2024's 57.1% did not mean it turned generous. It means those two sittings landed in very different places relative to a fixed standard. On a close paper that also means a single question can decide the outcome, because the pass mark is a real threshold rather than a curve that guarantees a fixed share of passes.
How the Grading Actually Happens
Because so much of Exam 7 is written and spreadsheet work, grading is a human process, and the CAS has documented how it runs. Before a sitting, a pass-mark panel reviews the questions against a predetermined definition of the minimally qualified candidate. After administration, each written item is graded independently by two graders, who then reconcile their scoring rubrics and their grades. Graders reward responses that show enough work to follow the calculation, justify the selections made, and tie the relevant readings together rather than reciting one in isolation, and that process still governs how the pass mark is set for every sitting.
What has changed is transparency. When the CAS moved these exams to computer-based testing in the fall of 2020, it stopped releasing exam papers, answer keys, pass marks, and the Examiner's Reports that once accompanied each sitting; only materials from before that transition remain posted for study. More recently, beginning with the Fall 2025 administration, the CAS also stopped publishing the list of passing candidates. The upshot is that candidates now see less about how any given sitting was graded than they used to, which is part of why a result like Spring 2026 feels so opaque from the outside.
What This Means for Candidates
The grading design explains the experience Exam 7 candidates describe. Because the pass mark floats with difficulty, strong work on the individual questions does not guarantee a pass, and a paper can land a point or two short on a sitting that felt manageable, because the mark rose to meet it. Failing candidates get score-range feedback rather than a single raw percentage, which helps for aiming a retake but frustrates anyone hoping to reverse-engineer exactly how close they came. The productive question after a fail is not how close you were last time but which topic will move your score the most next time.
Three conclusions follow. First, aim well above the historical pass mark, not at it, because the mark you must clear is set after you sit and can move against you, as Spring 2026 shows. Second, how an exam felt is a poor predictor of the result, since perceived difficulty is already priced into the pass mark. Third, a narrow fail is common and is not evidence that you are far off; it usually means you are one sharper topic or one cleaner justification away, a very different problem from a foundational gap. On a sitting like Spring 2026, a large share of genuinely prepared candidates will have failed by a small margin, which is exactly what a hard exam graded to a held standard produces.
How Exam 7 Compares to the SOA
The CAS is not unusual in grading to a consistent standard; the SOA sets its passing bar the same way, against a fixed competence level rather than a fixed percentage, and its upper-level exams swing too. The sharper contrast now is transparency. The CAS once published unusually candid Examiner's Reports that walked through what earned and lost points, but it discontinued them with the move to computer-based testing, and it no longer releases pass marks or passing-candidate lists. The SOA has moved the other way, cutting fellowship grading time to roughly four weeks and giving failing candidates detailed per-item score reports, as covered in our SOA FSA pathway analysis. A candidate who wants granular feedback on a failed sitting now gets more of it from the SOA than from the CAS.
One structural factor worth naming is the candidate pool. Fellowship-level CAS exams like Exam 7 carry heavier concentrations of repeat takers, since they sit near the end of a long sequence, and a pool weighted toward candidates who have already failed once tends to pull headline pass rates down relative to earlier, first-attempt-heavy exams. The upshot for anyone weighing the two tracks: Exam 7 belongs with the harder SOA fellowship assessments, not with newer, higher-passing modules, and its pass rate should be read in that company.
How to Prepare for the Volatility
You cannot control the pass mark, but you can prepare in a way that makes it irrelevant. The candidates who clear Exam 7 reliably tend to do the same handful of things.
Build for depth, and budget the hours honestly. Upper-level candidates commonly invest several hundred focused hours, and the material rewards working from the source papers rather than relying on a study manual's summary. Reread the primary readings, because the synthesis questions that decide close papers are built from the details manuals compress.
Practice the thing that is actually graded. Exam 7 does not reward knowing a method; it rewards showing and justifying it under time pressure. Because the CAS no longer releases recent exams or solutions, lean on the pre-2020 past exams and Examiner's Reports that remain posted, along with reputable providers' mock exams, and treat them as writing exercises rather than answer checks.
Answer every part, and match your length to the verb. A "briefly describe" prompt earns nothing extra for a long answer and quietly costs you time you need elsewhere. Address every item the question asks for, since a strong response to a question you skipped is worth zero.
Plan for more than one sitting. With Exam 7 now offered twice a year and pass rates this volatile, a study plan that assumes a possible retake is realistic rather than pessimistic. Treat a narrow fail as calibration data, and do not let any single sitting's pass rate, high or low, tell you how hard the exam really is.
The Bottom Line
Exam 7's record-low 23.7% looks alarming, and for the candidates who just failed it is. But it is not a sign that the exam broke or that the graders turned punitive. It is the CAS holding a constant competence bar across a run of difficult sittings, with the raw pass rate falling to keep that bar level. Read that way, the number says less about this year's candidates than about how the exam is built. Aim above the standard rather than at it, justify everything you write, and judge your own result, and the headline pass rate, against a system designed to measure mastery rather than to hit a target.