Carson Wrote: ------------------------------------------------------- > What’s wrong with the calculation? I think he’s > just saying that 59.6%*41%*35% equals the 5 year > average rate. > > Granted I think he is being very creative in his > logic here, but the sums themselves are ok I > think. Do we have any evidence to suggest that the > CFA actually looks at this metric? > > Leaving that minor detail aside, the thought of a > 60% pass rate makes me feel happy inside, so I am > going to give this post a stamp of approval! I think the math is wrong. I would use addition instead of multiplication.
FinanceMBA2011 Wrote: ------------------------------------------------------- > I would say around ~60% > > Reasoning; look at the last three year cycles, > cumulative pass rate. > > For example 2007 Lvl 1, 2008 Lvl 2, 2009 Lvl 3 > pass rates. Multiply these and you get the > percentage of people who made it through, assuming > they passed on the first attempt. > > For 2005 L1, 2006 L2, 2007 L3; 8.40% > For 2006 L1, 2007 L2, 2008 L3; 8.48% > For 2007 L1, 2008 L2, 2009 L3; 8.79% > > If you accept the argument that CFAI limits pass > rates to reflect an 8-9% cumulative pass rate > which is apparent over the last 5 years, then take > my cohort; > > 2008 L1: 35% pass > 2009 L2: 41% pass > > Then take the average cumulative pass rate over > the last 5 years (8.56%), divided by the factor of > last two years pass rates to compute the 2010 L3 > pass rate. > > 2010 L3 pass rate = (.0856)/[(.41)*(.35)]=59.6% > > There’s reason to be optimistic ladies and > gentlemen!! Interesting idea. The arithmetic here is correct as far as I know. The probability of three *independent* events occurring (with replacement) is given by P(A) x P(B) x P©. When a small number of items are selected from a large population without replacement, the probability doesn’t change significantly, and the same formula can still be used to approximate the total probability. However, we don’t really know how independent passing rates are. We would need some data on the number of people taking all the exams in a row (i.e. someone who actually took 2008 L1, 2009 L2, and 2010 L3), verses all of the other re-takers and year-skippers. If you assume *dependent* events, then the formula gets a little more complex: P(A,B,C) = P(A) x P(B) x P(C|A and B) with P(C|A and B) read as ‘the probability that C will occur given both A and B have occurred’ So for 2010, that would be P(C|A and B) = .085 / (.35 x .41) = .59 So, if I am doing this right, then we can say that: * assuming *independence*, the probability that ANYONE who took 2010 L3 will pass is 59% (I find this difficult to believe) * assuming *dependence*, the probability that a person who passed 2008 L1 and 2009 LII will also pass 2010 L3 is 59% (I find this much more believable)
The average charter holder takes 4 years to pass all 3 exams. CFA lore tends to call the second exam “the most failed one” so maybe we should count that one twice.
59 is too optimistic. Mark my word - result would be between 46-54
Dlpicket - I am aiming to be Mr. Average. Passed L1 first time, got L2 at the 2nd attempt and now looking to pass L3 first time around. With a 60% pass rate I’d be confident. 40% I’d be worried. Chances are I’m in a percentile between the two I feel. Of course that’s just gut instinct based on a whim!
Pass rate expected to be around 62%-65% From examination results table statistics in 2009 15,892 cand. passed LII exam Add to them Level III 2009 failing candidates 9,839 –> total expected to sit in 2010 = 25,731 (excluding freelancers). With a variance of 10% effective expected candidates to sit for the LIII exam in 2010 = 23,158. LII candidate expected to sit for 2010 exam should be expected around49,000. (June Level I 2009 passing candidate 21,034+Level II 2009 failed candidates 23,106+ a variance of 5,000 Dec Level 1Passing candidates). Assuming a 40% (avg) of level II pass rate this year => .40*49,000=19,600, and a normal growth in Level III candidates sitting for 2011 exams would be expected to be 27,720 ( in 2010 expected 23,158*1.197) based on historical rates. A little algebra would imply for 2011=> 27,720-19600=8,120 expected failing LIII failing candidates and returning in 2011 Guess pass rate in 2010 expected to be 23,158-8,120/23158 ~ 65% (increasing expected candidates to 28,484 who will sit in 2011 would give a 62% expected pass rate for 2010)
Jeos Wrote: ------------------------------------------------------- > Pass rate expected to be around 62%-65% > > From examination results table statistics in 2009 > 15,892 cand. passed LII exam > Add to them Level III 2009 failing candidates > 9,839 –> total expected to sit in 2010 = 25,731 > (excluding freelancers). With a variance of 10% > effective expected candidates to sit for the LIII > exam in 2010 = 23,158. > > LII candidate expected to sit for 2010 exam should > be expected around49,000. (June Level I 2009 > passing candidate 21,034+Level II 2009 failed > candidates 23,106+ a variance of 5,000 Dec Level > 1Passing candidates). > > Assuming a 40% (avg) of level II pass rate this > year => .40*49,000=19,600, and a normal growth in > Level III candidates sitting for 2011 exams would > be expected to be 27,720 ( in 2010 expected > 23,158*1.197) based on historical rates. > A little algebra would imply for 2011=> > 27,720-19600=8,120 expected failing LIII failing > candidates and returning in 2011 > > Guess pass rate in 2010 expected to be > 23,158-8,120/23158 ~ 65% (increasing expected > candidates to 28,484 who will sit in 2011 would > give a 62% expected pass rate for 2010) Jeos, judging by the thoughtfulness and thoroughness of your answer, I’d say you’d make a great analyst even if you’re rationale is completely wrong.