[Updated for CFA Aug 2023 exams – details below. To help our estimates, send your results PDF to email@example.com.]
IMPORTANT: Please do not use our MPS estimate as a basis to plead to CFA Institute to ask for a recheck. Our MPS estimate is an estimate – using our MPS estimate to contest your result won’t get you anywhere. Your individual MPS may also vary.
What is the CFA passing score?
Unfortunately CFA Institute doesn’t disclose the official Minimum Passing Score (MPS), which varies from exam to exam. We get why CFA Institute does this.
That said, we have been estimating the CFA exams’ MPS as an intellectual exercise since 2012, with some help from our readers. We do this as it can be useful for CFA candidates to have a target range of scores to aim for when doing practice exams.
Without further ado, here’s all you need to know about the CFA passing score: what it is, how it is determined, and what our MPS estimates are for all CFA exams since 2012.
- What is CFA exam's minimum passing score (MPS)?
- Do candidates have to beat the MPS for each topic to pass, or just the overall exam?
- How is the CFA passing score determined?
- Why doesn’t CFA Institute publish the minimum passing score for each level?
- CFA passing scores for each level: Our data-driven estimates for 2012-2023
- My results doesn't seem to fit your MPS estimate!
- Help our MPS research, send us your results PDFs!
What is CFA exam’s minimum passing score (MPS)?
Simply put, the Minimum Passing Score (MPS) is the lowest score a candidate can get and still pass their CFA exam.
It is not published by the CFA Institute.
This should not be confused with the CFA pass rate published annually, which is normally lower and represents the percentage of candidates that passed a particular exam.
Do candidates have to beat the MPS for each topic to pass, or just the overall exam?
There are no topic-specific passing scores.
So as long as your overall score beats the Minimum Passing Score, you should pass.
How is the CFA passing score determined?
In the earlier days, the MPS was a simple formula – 70% of the highest scoring candidate for that year. If the highest-scoring candidate scored 90%, the MPS would therefore be 70% * 90% = 63%.
Some candidates still mistakenly believe that this is the method used to determine the MPS today.
Presently, the MPS is determined by a custom methodology devised by CFA Institute, and one of the primary inputs is a series of workshops based on the Angoff model.
Here’s how the MPS for each Level are determined:
- A large, diverse group of CFA charterholders evaluate the entire exam individually question by question and makes a judgment on the performance of the just-competent candidate. In other words, envisioning a candidate they judge to be just good enough to pass, they estimate the probability that this candidate will get this question right.
- Once all the evaluations are done, the charterholders take a look at actual candidate performance and the evaluation process is repeated again.
- Psychometricians oversee the entire process and the result of this workshop is the main reference (but not the final word) for setting the MPS.
- The CFA Institute Board of Governors then puts all information on the table. All available information is considered including recommendations from the Angoff workshop (the most important input), and then a final decision is made on the MPS. The objective is to set a consistent competency level across years.
You therefore pass by being better than what CFA Institute defines as a ‘just-competent’ candidate.
This may be indirectly influenced by the performance of candidates taking the same exam as you, but it definitely is not a primary driver. So remarks such as ‘bringing down the curve’ are not exactly accurate.
Chris Wiese, Head of Exam Development at CFA Institute briefly summarizes the standard setting process:
Why doesn’t CFA Institute publish the minimum passing score for each level?
CFA Institute’s stance is that credentialing programs are designed to measure whether a candidate has the necessary knowledge, skills, and abilities to perform a certain job function.
They are not designed to indicate the relative performance within those skills.
The MPS can vary somewhat from year to year based on the difficulty of the questions asked, it is designed to be a consistent benchmark of the necessary knowledge, skills, and abilities.
Releasing the MPS would focus on beating a specific ‘score’, rather than acquiring the necessary knowledge to function as a competent finance professional.
Is CFA Institute right?
We think CFA Institute’s reasons are valid. The exams are just one part of what CFA Institute are about, and keeping that in mind, the desire to deemphasise comparison of CFA charterholders and various exam years by MPS and/or actual scores makes sense. Of course, it doesn’t stop candidates from wanting to know!
Will CFA Institute change this policy in the future?
It’s unlikely. We’ve interviewed CFA Institute on this before and they’ve responded: “This, among other things, is to prevent the temptation for people (e.g. employers) to compare candidates based on score, and to preserve a consistent benchmark across years.“
CFA passing scores for each level: Our data-driven estimates for 2012-2023
So why are we estimating the MPS then?
- Because as ex-candidates ourselves, we also understand the need to have a target score to be aiming for when preparing for the CFA exams.
- Because through our awesome readers input, we do have the collective info that no one else has to glean insights like this.
- And because we’re curious.
Prior to CFA Institute’s introduction of the new format of results in June 2018, candidates did not receive their score, but they received a ‘score category’ (0-50%, 51-70%, 71-100%) in each topic area.
To estimate the MPS, we analyzed thousands of actual, detailed, self-reported candidate results. We could deduce the MPS by considering extreme cases, i.e.
- candidates that passed despite showing bad topic area scores, and;
- candidates that failed despite showing good or decent topic area scores
Our method – in our opinion – was quite robust, and increased in accuracy the more cases we had to analyze.
Even though the results are reported differently now, we’ve developed a new method to estimate the MPS. We prefer to keep things close to our chest for now, but we do believe that our new methodology is equally as robust, if not better, than our old method.
So, here we go:
CFA Level 1 passing scores 2012-2023 (estimated)
We estimated that CFA Level 1’s MPS ranged from 56%-74% from 2012-2023, with an 11 year average of 65%. The latest estimated MPS for CFA Level 1 (i.e. Aug 2023 exams) is estimated to be around 67%.
The latest estimated MPS for recent CFA Level 1 exams are:
- Aug 2023: 67%
- May 2023: 69%
- Feb 2023: 69%
Since the introduction of computer-based testing (CBT), our MPS estimates have consistently exceeded 70%, except until recently the May and August 2022 exams.
CFA Institute’s official guidance states that the minimum passing score reflects the score that can be achieved by a ‘just competent’ candidate, which is supposed to be consistent across the years, whether it’s 2010, 2015 or 2020. So if the MPS has indeed experienced a step change to 70+%, this implies that the CBT exam format is easier to score compared to the older paper-based format.
It used to be solid advice for candidates to aim for a 70% score in your mocks. With the latest MPS estimates, we hope that this continues to be the case and that 2021 is an anomaly.
While we have seen some evidence that CFA Level 1’s passing score is heading back to the long term average, we would be cautious of recommending a lower MPS to target for until more data point emerges.
Why are Feb 2021, Mar 2021 and Apr 2021 grouped together?
When we performed the MPS analysis for the March and April 2021 CFA exams, we realized that the MPS, 90th percentile, 10th percentile, and topic weightings were the same as the February 2021 exams, so we have grouped them as one exam. This implies:
- Feb 2021, Mar 2021 and Apr 2021 used the same exam, or the same topic percentage mix, or same pool of questions
- Feb 2021, Mar 2021 and Apr 2021 used the same candidate cohort to determine the MPS, 90th percentile and 10th percentile. It’s likely that Mar 2021 and Apr 2021 simply used Feb 2021’s pool, which statistically should be fine since Feb 2021 has way more candidates.
CFA Level 2 passing scores 2012-2023 (estimated)
We estimated that CFA Level 2’s MPS ranged from 60%-71% from 2012-2023, with an 10-year average of 65%. The latest estimated MPSs for CFA Level 2 are:
- Aug 2023 = 69%
- May 2023 = 70%
- Nov 2022 = 62%
CFA Level 2 MPS estimates have surged upwards and stayed high since 2021. Although this dipped in November 2022, this seems to have returned to its higher levels in May-Aug 2023. It would be interesting to see if this lower MPS threshold continues – my gut-feel is that it will.
If you’re preparing for CFA Level 2, my advice is to aim for a high passing score. Don’t bank on the MPS being low – these days it seems unlikely.
Permanent change, or Covid/CBT blip? Right now, who knows? We will always be estimating and publishing MPS estimates as they come, so as we get more datapoints, we’ll get a clearer picture on this.
We won’t get as many data points yet as the Level 2 exams occurs thrice a year, but with future exams we’ll see a pattern emerge and see if CFA Institute is right in saying this is temporary. But to be extra safe, do make sure you aim for a slightly higher score for your mocks.
CFA Level 3 passing scores 2012-2023 (estimated)
We estimated that CFA Level 3’s MPS ranged from 55%-63% from 2012-2023, with an 11 year average of 59%.
The latest estimated MPS for CFA Level 3 (i.e. Aug 2023 exams) is estimated to be around 63%. (To help us narrow this down, send your Level 3 result to firstname.lastname@example.org!)
Important note on L3 MPS accuracy: Since the change to computer-based testing (and the fact that CFA Institute only distributes detailed results to failing L3 candidates), estimating the L3 MPS takes a lot more work and has a larger margin for error – there’s just too large a variation of weightings to get an incredibly precise result. Nonetheless, we try our best.
The post-CBT MPS estimates have been a mixed bag. The Aug 2021 MPS estimate shows a strong drop from May 2021 and is an all-time low, but less of an outlier if you compare it with historical data. This seems to have ‘recovered’ by May 2022 – but we just don’t have enough data-points at this stage to solidly conclude.
The pass rates however are a different story: historical Level 3 pass rates plunged to 39-42% (May 2021 – Nov 2021), which indicates that between the change to a computer-based testing format and a revised testing schedule, the Level 3 CFA exams might be harder to pass than before. That said, CFA Level 3 pass rates are also showing a recovery towards the historical average.
We reached out to CFA Institute for a comment and got a response:
“Unfortunately, fewer candidates passed at each level of the May exams than we might have expected given historical pass rate trends. While some impacted candidates were able to pass, we believe the disrupted nature of the deferred candidates’ studies is reflected in the overall passing rate across all three levels.
All examinations are constructed using industry best-practice psychometric data and detailed examination blueprints. The difficulty of individual exams may vary slightly, and we adjust the minimum passing score to compensate, but generally the required adjustments are small—typically just a few points. The degree of difficulty of the May exams was consistent with previous exam administrations, and this is the case whether we look back to paper-based testing or computer-based testing, which we introduced in February this year for Level I.
Computer-based testing is neither harder nor easier than paper-based testing. We use robust statistical analysis and processes to ensure that “the bar” remains unchanged for every exam administration, employing techniques that are well established by psychometricians to ensure that the exam is neither easier nor harder because it is administered in a computer-based setting. In our move to computer-based testing we went to great lengths to ensure the continued integrity of the exams and the CFA charter. In short, computer-based testing is not harder or easier.”CFA Institute on the CFA Level 3 May’21 exam results
My results doesn’t seem to fit your MPS estimate!
The CFA computer-based exams are not the same for every candidate, so the difficulty of the individual exams may vary.
When this happens, the MPS for that particular exam is also adjusted. Our MPS estimate is an aggregate value, so your MPS may be higher or lower than our estimate “typically by a few points”.
See relevant quote from CFA Institute:
All examinations are constructed using industry best-practice psychometric data and detailed examination blueprints. The difficulty of individual exams may vary slightly, and we adjust the minimum passing score to compensate, but generally the required adjustments are small—typically just a few points.
So you might well be right in thinking that the MPS is higher for your case.
Help our MPS research, send us your results PDFs!
To keep this research going, we need your help to estimate future exam’s MPS for each Level.
All you need to do is just send your results PDF to email@example.com. Thanks for your help!
We understand the trust that you are giving us when you do so, and will not be sharing your information with anybody else, and will only use it to perform anonymized analyses that continue to help present and future CFA exam candidates.
Fingers crossed for your results day!
Meanwhile, here are some related articles that may be of interest: