Updated for the 2026-2027 CFA® Level I curriculum.
Bayes’ theorem updates the probability of a condition after new evidence becomes available. In CFA Level I questions, you usually begin with a prior probability, assess how likely the evidence is under each possible condition, and then calculate the revised or posterior probability.
Once the events are labeled clearly, the calculation follows a repeatable sequence. Most errors come from confusing the direction of conditional probabilities or leaving out a possible condition that could have produced the evidence.
Quick Answer
Bayes’ theorem combines a prior probability with the likelihood of observing new evidence to produce a posterior probability. In an investment setting, Bayesian updating allows an analyst to revise an estimate about events such as a recession, default, earnings outcome, or manager skill after receiving a new signal.
Key Takeaways About Bayes’ Formula and Bayesian Updating
A prior probability represents your estimate before observing new evidence.
A likelihood shows how probable the evidence is under a particular condition.
The total probability of the evidence includes every condition that could have produced it.
A posterior probability is the updated estimate after the evidence has been considered.
Evidence has more influence when it is much more likely under one condition than under the alternatives.
The posterior remains a probability estimate and can be updated again when more information becomes available.
What You Need to Know for CFA Level I
When you see a Bayes’ theorem question, you should be able to:
Identify the prior probabilities and their complements.
Recognize the likelihood of the evidence under each possible condition.
Calculate joint probabilities by multiplying along each outcome path.
Add the relevant joint probabilities to find the total probability of the evidence.
Apply the Bayes’ theorem formula correctly.
Interpret the posterior probability in the context of the investment problem.
What Is Bayes’ Theorem?
Bayes’ theorem links two conditional probabilities.
Suppose represents the condition you are interested in and represents new evidence. The expression measures the probability of observing evidence when condition is true.
After observing , Bayes’ theorem calculates , the updated probability that is true.
The basic posterior probability formula is:
The numerator, , is the joint probability that condition is true and evidence occurs.
The denominator, , is the total probability of observing the evidence across all possible conditions.
Bayes’ Theorem Formula for Two Conditions
When the only possible conditions are and its complement , the Bayes’ theorem formula becomes:
For several mutually exclusive and collectively exhaustive conditions, the general formula is:
Where:
= prior probability that condition is true
= likelihood of observing evidence if condition is true
= unconditional or total probability of observing evidence
= posterior probability that condition is true after observing evidence
= probability that condition is false
= number of possible conditions
= index used to represent each possible condition
The numerator, , represents the joint probability that condition is true and evidence occurs.
The denominator adds the joint probabilities for every possible condition that could have produced evidence .
The conditions in the denominator must cover every possible way the evidence could occur. Leaving out an alternative condition understates the total probability of the evidence and overstates the posterior probability.
How Does Bayesian Updating Work?
Bayesian updating can be handled in four steps.
1. List the Possible Conditions
Identify the possible states or outcomes that could exist. They should be mutually exclusive, meaning only one can occur, and collectively exhaustive, meaning they cover every possibility.
2. Assign the Prior Probabilities
Write down the probability of each condition before the new evidence is observed. The prior probabilities must sum to 1.
3. Calculate the Joint Probabilities
Multiply each prior probability by the conditional probability of observing the evidence under that condition.
4. Divide by the Total Probability of the Evidence
Add the joint probabilities for every path that produces the evidence. Then divide the joint probability for the target condition by that total.
A probability tree can help you organize these calculations. The first set of branches shows the prior conditions, and the second set shows the likelihood of the evidence under each condition. Multiplying along a branch gives its joint probability.
Worked Bayes’ Theorem Example
An analyst assigns a 30% prior probability that a portfolio manager has genuine skill. A positive performance signal occurs:
70% of the time when the manager has skill.
25% of the time when the manager does not have skill.
The analyst observes a positive signal. What is the updated probability that the manager has skill?
Step 1: Identify the Prior Probabilities
Step 2: Calculate the Joint Probabilities
The probability of a positive signal from a skilled manager is:
The probability of a positive signal from a manager without skill is:
Step 3: Calculate the Total Probability of a Positive Signal
Step 4: Calculate the Posterior Probability
The posterior probability that the manager has skill is approximately 54.6%.
The positive signal raises the estimate from the 30% prior because it is more common among skilled managers. Positive signals also occur 25% of the time without skill, so the evidence supports a meaningful update rather than a near-certain conclusion.
How Should You Interpret the Posterior Probability?
The posterior probability reflects three parts of the problem:
How likely the condition seemed before the evidence.
How strongly the evidence supports that condition.
How often the same evidence appears under alternative conditions.
In the example, the analyst begins with a relatively low prior probability of skill. The positive signal carries useful information because skilled managers are more likely to produce it. The false-positive rate among managers without skill limits the size of the update.
For exam questions, finish the calculation with a plain-language interpretation. State the updated probability, compare it with the prior, and explain why the evidence changed the estimate.
Common Exam Traps
Common mistakes include:
Reversing and .
Using the likelihood as the posterior probability.
Calculating only the numerator and omitting the total probability in the denominator.
Forgetting to calculate the complement of a prior probability.
Leaving out an alternative condition that can also produce the evidence.
Adding probabilities along one branch instead of multiplying them.
Treating the posterior probability as certainty.
Reporting the answer without explaining how the evidence changed the prior estimate.
Practice Question
An economist assigns a 20% prior probability to a recession. A negative leading indicator occurs with a probability of 60% during a recession and 15% when there is no recession.
After observing a negative indicator, what is the updated probability of a recession?
20%
50%
80%
Correct Answer: B
The probability of a negative indicator occurring during a recession is:
The probability of a negative indicator occurring without a recession is:
The posterior probability is:
The updated probability of a recession is 50%.
Option A uses the original prior probability without incorporating the negative indicator.
Option C excludes the possibility that a negative indicator can occur when there is no recession.
Continue Your CFA Level I Prep With KeyPoint
Use structured lessons, practice questions, mock exams, and progress tracking to focus on the time you have left
FAQs About Bayes’ Theorem and Bayesian Updating
What Is the Difference Between a Prior and Posterior Probability?
A prior probability represents an estimate before new evidence is observed. A posterior probability is the revised estimate after the evidence and its likelihood under each possible condition have been considered.
Why Is the Denominator Needed in Bayes’ Theorem?
The denominator calculates the total probability of observing the evidence. It includes every condition that could have produced that evidence and places the target joint probability within the full set of relevant outcomes.
Can a Posterior Probability Become a New Prior?
Yes. When additional evidence becomes available, the current posterior probability can serve as the prior for the next update. This allows estimates to evolve as analysts receive more information.
Is Bayes’ Theorem the Same as Conditional Probability?
Bayes’ theorem is built from conditional probability. It provides a structured way to calculate when you know the prior probability of and the likelihood of observing under the possible conditions.