SOA FAM Difficulty and Pass Rates
FAM difficulty comes from combining probability, severity and frequency models, credibility, reserving, mortality, present values and premiums in a 210-minute paper.
- SOA reported a 60.6% pass rate for 1,304 candidates in February 2026.
- That historical cohort result describes one sitting, not an individual candidate’s odds.
On this page7 sections
- February 2026 candidates
- 1,304
- Passing
- 790 (60.6%)
- Effective candidates
- 1,209
- Effective pass rate
- 65.3%
A demanding breadth of actuarial models
FAM combines short-term insurance and reinsurance with probability models, estimation, credibility, ratemaking, reserving and option valuation. It then moves to insurance products, mortality, present-value random variables, premiums and policy values. A candidate must switch between distribution theory, financial mathematics and life-contingency notation. The breadth makes retrieval and model selection important alongside algebra.
The February 2026 SOA report gives a useful dated pass-rate snapshot: 790 of 1,304 candidates passed, or 60.6%. Among 1,209 effective candidates, 65.3% passed. These are that sitting’s cohort statistics, not an estimate of a particular person’s chance or a promise about another administration.
Why candidates find it difficult
A problem can require recognizing the model before calculating. A deductible transforms ground-up loss; aggregate risk combines frequency and severity; limited-fluctuation credibility blends experience; a life annuity weights payments by survival and discounting. Similar symbols appear across topics, but the assumptions differ. Most errors happen before arithmetic when a candidate chooses the wrong random variable or payment timing.
A long sitting adds attention demands. In 210 minutes, repeated calculations can create fatigue, especially if a candidate has not practiced writing assumptions and checking units. A good preparation set should mix conceptual selection, derivations and arithmetic. A candidate who only memorizes formulas may struggle when the problem changes the deductible, payment time or conditioning event.
Use the pass rate carefully
The 60.6% February pass rate gives context for one administration. The report identifies the candidate count and effective group, which makes the denominator clearer than an unsourced social-media percentage. It does not mean 60.6% is the probability that any person will pass, nor does it predict the October 2026 or 2027 outcomes. Candidate preparation and form results differ.
Similarly, the 60.0% pass mark from February is not a permanent raw cutoff. Forms are calibrated and scaled for comparable criteria. Keep pass rate, pass mark, result and study score separate. If a candidate wants personal readiness information, use performance on fresh mixed problems and review the source of each error.
Identify your own difficulty pattern
Classify a miss as model choice, notation, algebra, arithmetic, assumptions, timing or concept. If you used the wrong survival conditioning set, revisit life-table probability. If you used an unconditional expected claim payment when the question asks conditional severity, review the atom at zero. If you used a net premium where expenses are included, define the loss random variable and loading.
A second diagnostic compares topic coverage. A high score in a small preferred area cannot compensate for never practicing one of the large ranges. The outline’s percentages are ranges, so use them as broad weights and also account for dependencies. Probability and financial math assumptions feed many later topics.
Switching between insurance models is a major challenge
FAM asks candidates to move between short-term insurance quantities and long-term contingent cash flows. An aggregate claim model uses a frequency distribution and individual severities; a life-contingent value uses survival or mortality probabilities and discounting. The same word “expected” appears in both settings, but the underlying random variable changes. A candidate should write a one-line model statement before computing so that the correct formula follows from the structure.
For example, expected aggregate loss may be E[N]E[X] under the stated independence assumptions, whereas an expected life-insurance benefit sums payment amounts weighted by the probability of death in each interval and the discount factor. Neither is a universal replacement for the other. The difficulty is often in identifying the event and timing, not in the arithmetic. Mixed practice helps because it forces the candidate to choose the model rather than repeat a chapter’s one formula.
Breadth and mathematical depth interact
The syllabus ranges across estimation, distributions, reserving, pricing, options and life contingencies. A candidate can understand each formula in isolation yet lose time when a stem combines two concepts. Censoring can change a likelihood contribution; trend and development can alter how historical claims inform a rate; survival can alter whether a premium is paid. Build concept maps that show assumptions and inputs, not just a list of equations.
Depth matters too. A severity question may ask for a transformed expectation after a deductible or policy limit. A life-contingency question may distinguish complete from curtate future lifetime or require an extra-risk mortality method. Recognizing terminology is only the first step. Practice derives a quantity from definitions and checks whether the result responds plausibly when an assumption changes.
A practical topic-switching drill
Make a small deck of prompts that contains only the quantity and model family, not the formula. One prompt might say “annual aggregate payment with random claim count”; another might say “benefit paid at the end of year of death”; a third might say “likelihood with right-censored observations.” For each, name the random variable and key assumption before opening a solution. This isolates model recognition, a major source of difficulty in a broad syllabus.
After the model is identified, solve a short problem and explain what changes if one assumption changes. If severity and frequency are independent, say so. If the benefit pays immediately, adjust timing. If the data are censored, write the probability contribution. This practice improves transfer among the ten topic areas.
Common questions
What was the February 2026 FAM pass rate?
SOA reported 60.6% for 790 of 1,304 candidates in that sitting.
Does that predict my chance of passing?
No. It is a historical cohort result, not an individual probability.
Is FAM difficult because it is adaptive?
The syllabus describes multiple forms and statistical scaling; the main preparation demand is broad mathematical application.