ALTAM is a 3-hour written-answer exam: six questions worth 60 points in total, one of them answered in Excel. Pass rates have historically landed somewhere around 40 to 55%. It sits above FAM on the associateship track, assumes you already know survival models and basic life contingencies, and pushes into multiple-state models, joint lives, profit testing, pensions, universal life and embedded options. It's offered twice a year, typically in spring and fall sittings.
The format is what people underestimate. You aren't choosing A through E. You set up the model, carry the algebra and present a numerical answer with your work visible. Graders give partial credit, so a clean, labeled solution that stops one step short still earns points, and a bare final number that's wrong earns nothing. That changes how you should study, because you're training to communicate a solution as well as reach one. Since it's graded by people, results take longer than for a computer-scored exam, and the pass mark is set for each sitting.
What's on the syllabus
The SOA's current syllabus has seven topics:
| Topic | Weight | What's in it |
|---|---|---|
| Survival models for contingent cash flows | 10 to 20% | Markov multiple-state models, Kolmogorov's forward equations, multiple decrement tables |
| Premium and policy valuation for long-term coverages | 12 to 20% | Premiums and policy values for state-dependent coverages, Woolhouse approximations, Thiele's differential equation |
| Joint life insurance and annuities | 8 to 16% | Joint-life models, dependent lifetimes, premiums and policy values on joint lives |
| Profit analysis | 10 to 20% | Profit testing, gains by source, profit signature, NPV, IRR, profit margin |
| Pension plans and retirement benefits | 10 to 18% | Replacement ratios, service tables and salary scales, accrued liability and normal cost for DB plans |
| Universal life insurance | 10 to 18% | Account values, no-lapse guarantee reserves, UL profit testing |
| Embedded options in life insurance and annuity products | 10 to 18% | Guaranteed minimum benefits, Black-Scholes valuation, delta hedging |
No topic is weighted above 20%, so you can't skip one. Multiple-state models run through several of them: they're the core of the first topic, the policy values in the second are for state-dependent coverages, and joint lives can be modeled as a multiple-state model. Candidates tend to put them off because they're the hardest block conceptually, but you're better off doing them first. Weights shift between curriculum updates, so pull the current syllabus from the SOA before you start.
Hours, timeline and practice
For a working candidate I'd plan on 300 to 400 hours over 4 to 6 months. If your FAM-level survival models and integral calculus are fresh, you can lean toward the lower end. If it's been a year or two, budget extra time up front for a refresh, because ALTAM problems assume that machinery is automatic.
Here's the timeline I'd run for a typical sitting:
- Weeks 1 to 3: refresh the fundamentals. Re-derive the standard life contingency identities, survival functions and insurance and annuity present values until they're automatic, before you start ALTAM problems.
- Weeks 4 to 10: the core syllabus, topic by topic. Read a topic, then work problems on it right away. Multiple-state models and policy values get the most calendar time.
- Weeks 11 to 15: pensions, profit testing and the remaining topics. These reward organized setup, so write full solutions.
- Weeks 16 to 18: full written-answer practice under time, simulating the 3-hour sitting with past written-answer problems.
- Final week: review, formula recall and light problem sets, so you arrive rested.
Two focused hours most days will do more for you than one big weekend block.
The biggest change from earlier exams is that you have to write full solutions. Work each problem cold, on paper or in the exam's answer format, showing every step. Then grade yourself the way a grader would: did you define your notation, state the model and carry units, and where would you have earned or lost partial credit? Redo any problem you couldn't finish from scratch a day later, since re-solving teaches more than re-reading. Keep an error log tagged by topic and by type (setup error, algebra slip, wrong assumption), and the patterns show up fast.
For volume you want a large, tagged problem set you can drill by topic and then mix. FreeFellow has a free ALTAM question bank. Use it for reps on multiple-state and policy value problems, then move to timed mixed sets once each topic feels routine. Practice writing efficiently too: state your assumptions in one line and move on.
Mistakes I see
The most common one is treating ALTAM like a multiple-choice exam. Candidates who only practice getting the final number get blindsided by the need to present clean work, so practice the presentation from day one.
Skipping the fundamentals refresh is another. ALTAM assumes FAM-level fluency, and whatever time you save by skipping the review, you lose back when basic identities slow you down mid-problem.
Plenty of candidates never do a timed full practice run. Time pressure on a written exam is real, and if your first timed run is exam day, you'll misallocate minutes across the six questions.
When you're stuck on exam day, write the setup, state the method and take whatever partial credit you can. Never leave a problem blank. And learn the derivations along with the formulas. When a formula slips, being able to re-derive it (Thiele, the recursions, Kolmogorov) is what saves the problem.
You can get your reps in with FreeFellow's free ALTAM questions, and check syllabus details and dates against the SOA before you sit.