An Analysis of Exams 3 and 4
Glenn G. Meyers, General Officer, Examination CommitteeBeginning with the May 2000 examinations, Exams 3 and 4 have been offered jointly with the Society of Actuaries. Since their introduction, some have expressed concern that life actuaries have an advantage over casualty actuaries on Exam 3, which covers life contingencies. A second concern that has been raised is that the difficulty and/or amount of material on Exams 3 and 4 makes them challenging to pass through self-study. For the past several years, the CAS has been surveying our candidates about their employment and student status. This article's purpose is to report the survey results related to these two concerns.
First, let me give some basic demographics from the Spring 2001 exams.
TABLE 1
Number of
Candidates SittingCategory Exam 3 Exam 4 Full-Time Student 345 144 CAS Worker 166 111 SOA Worker 822 579 Unknown 193 174 Question #1—Do life actuaries have an advantage on Exam 3?
Table 2 provides pass ratios by candidate category.
TABLE 2
Exam 3 CAS
WorkerSOA
WorkerUnknown Full-Time
StudentTotal Spring 2001 33.1% 38.3% 38.3% 59.7% 42.6% Fall 2000 20.4% 30.9% 30.4% 53.8% 36.1% Spring 2000 15.2% 29.5% 37.0% 44.1% 31.9 Table 3 shows the differences between the CAS worker category and the other categories.
TABLE 3
Exam 3 CAS
WorkerSOA
WorkerUnknown Full-Time
StudentTotal Spring 2001 - 5.2% 5.2% 26.6% 9.5% Fall 2000 - 10.5% 10.0% 33.4% 15.6% Spring 2000 - 14.3% 21.8% 28.9% 16.7 These statistics suggest the most significant differences are between the full-time students and everyone else. I believe there are two reasons for this. The first is that classes help students pass exams. The second is that only the very best students will even take Exam 3 while still students.
There are also differences between CAS workers and SOA workers, however, the differences are smaller and they are shrinking. I suggest a reason for this is that there is a higher proportion of SOA workers who are graduates of actuarial science programs. With the advent of joint CAS/SOA Exams 3 and 4, I expect more students will choose to become casualty actuaries in the future.
I also suggest that a reason the differences are shrinking is that the main study aid for those who did not have actuarial science classes is the prior exams. As the prior exams become available as study aids, the playing field is leveled.
It is worth noting that on Exam 4 (See Table 4 below), where life contingency questions are not a factor, the differences among the groups also exist, but are slightly lower.
TABLE 4
Exam 4 CAS
WorkerSOA
WorkerUnknown Full-Time
StudentTotal Spring 2001 35.1% 39.7% 35.6% 54.2% 40.6% Difference - 4.6% 0.5% 19.0% 5.4% Question #2—Can the difference between exam performance of full-time students and workers be attributed to our new joint exams?
The CAS has kept pass ratio statistics by candidate category for Parts 4A and 4B (See Tables 5 and 6).
TABLE 5
Part 4A S1995 F1995 S1996 F1996 S1997 F1997 S1998 F1998 S1999 F1999 Total Full Time Students # of Candidates 60 35 46 15 30 7 24 14 22 17 270 Pass Ratio 30.0% 34.3% 45.7% 33.3% 26.7% 0.0% 62.5% 42.9% 59.1% 58.8% 40.0% Non-Students # of Candidates 388 378 338 348 326 475 404 449 426 384 3916 Pass Ratio 41.0% 32.5% 35.5% 34.5% 35.9% 37.1% 42.3% 25.4% 34.3% 44.8% 36.2% Pass Ratio Difference -11.0% 1.8% 10.2% -1.2% -9.2% -37.1% 20.2% 17.5% 24.8% 14.0% 3.8% TABLE 6
Part 4B S1995 F1995 S1996 F1996 S1997 F1997 S1998 F1998 S1999 F1999 Total Full Time Students # of Candidates 151 78 98 28 116 14 135 51 152 67 890 Pass Ratio 38.4% 24.4% 28.6% 42.9% 48.3% 28.6% 51.9% 43.1% 42.1% 56.7% 41.7% Non-Students with Other CAS Credit # of Candidates 418 413 424 376 323 444 339 450 427 467 4081 Pass Ratio 31.6% 27.4% 37.3% 40.2% 33.1% 19.4% 28.3% 27.3% 30.0% 34.3% 30.7% Student/CAS Difference 6.8% -3.0% -8.7% 2.7% 15.2% 9.2% 23.5% 15.8% 12.1% 22.5% 11.0% Non-Students without Other CAS Credit # of Candidates 331 227 308 318 308 494 465 551 621 654 4277 Pass Ratio 39.6% 38.3% 46.1% 57.6% 35.4% 32.0% 36.8% 29.4% 36.9% 43.0% 38.6% SOA/CAS Difference 8.0% 11.0% 8.8% 17.4% 2.3% 12.6% 8.5% 2.1% 6.9% 8.7% 7.9% Student/SOA Difference -1.2% -14.0% -17.5% -14.7% 12.9% -3.4% 15.1% 13.7% 5.2% 13.8% 3.0% While the results are not uniform, it does appear full-time students have generally outperformed CAS workers on 4A and 4B. I attribute the observed increase in the difference to the rise in casualty actuarial courses in the various universities. The full-time student phenomenon discussed above does not appear to be a new problem.
The old SOA Flexible Education System gave credit for Part 4B. Since the exam statistics prior to 2000 did not distinguish between life and casualty nonstudents, we had to infer affiliation by seeing if the students have taken other CAS exams. By this measure, it appears SOA students outperformed CAS students. Admittedly, there may be a bias in these statistics since SOA students could choose whether or not to take Part 4B.
In summary, it appears SOA students did have an initial advantage on the new exams. However, these differences are shrinking and I expect them to disappear. A likely explanation for the differences we do see is the long-standing difference between students and workers.
I have always believed the exams should be passable by self-study. Seeing the difference in exam performance between students and workers does not disturb me. It takes time to absorb the material in these exams and students in an actuarial science program have a head start. What would disturb me is to find that almost all who pass come from actuarial science programs. So far this is not the case.
What does disturb me is the low number of CAS workers who are taking the new Exams 3 and 4. By "low" I mean low in comparison to the number of students who took Parts 4A and 4B. It appears our current debate on Exams 3 and 4 is discouraging CAS workers from taking these exams. It may take some time to recover from our current strife.
In closing I should state that nothing I have said above should be taken to mean I favor the status quo. I am quite open to changing our exam system. I just hope the changes we do make are for the right reasons.
