Updated for the 2026-2027 CFA® Level I curriculum.
Analyst forecasts are not pure math. They run through human judgment, and human judgment carries predictable biases. These biases push forecasts in specific, identifiable directions, even when the underlying model and data are correct. CFA Level I tests whether you can name the bias in a scenario and match it to the right fix.
Quick Answer
Behavioral biases are psychological patterns that distort analyst judgment during forecasting, separate from model or data errors. The main biases are overconfidence, illusion of control, conservatism, representativeness, and confirmation bias, and each one skews forecasts in a specific direction. Remedial actions include using checklists, comparing forecasts to base rates, running scenario analysis, and actively seeking disconfirming evidence. For CFA Level I, you need to connect each bias to its effect and its best remedy.
Key Takeaways
Behavioral biases distort analyst forecasts even when the model and data inputs are accurate.
Overconfidence and illusion of control cause analysts to understate forecast uncertainty.
Conservatism bias causes analysts to underreact to new, contradictory information.
Representativeness bias causes analysts to assume a company will behave like a familiar past case.
Confirmation bias causes analysts to favor evidence supporting an existing view and discount evidence that does not.
Remedial actions such as checklists, base rate comparisons, scenario analysis, and seeking disconfirming evidence reduce these effects.
What You Need to Know for CFA Level I
The LOS asks you to explain how behavioral factors affect forecasts and recommend remedial actions, so expect questions that test both halves.
You must recognize a named bias from a short scenario and identify its effect on the forecast.
You must match the bias to the correct remedial action, not just define the bias in isolation.
Keep this note separate from forecasting mechanics. Pro forma models, inflation adjustments, and Porter's Five Forces analysis cover the inputs and structure of a forecast. This note covers the judgment errors that distort those inputs.
The most common trap is confusing two biases that look similar but push the forecast in different ways, such as overconfidence and conservatism.
Why Analyst Forecasts Need a Behavioral Lens
Every forecast combines data, a model, and a human decision about what the numbers mean. The data and model can be correct and the forecast can still be wrong, because the analyst's judgment introduces a consistent, predictable bias. These biases are not random mistakes. They repeat in the same direction across analysts and across time, which is why CFA Level I treats them as a distinct topic rather than general forecasting error.
The Key Behavioral Biases That Distort Forecasts
Overconfidence and Illusion of Control
Overconfidence causes analysts to overestimate the precision of their forecasts. An analyst might give a narrow range around a point estimate, even when the history of forecast errors is much wider than that range suggests. Illusion of control is a related bias where the analyst believes their analysis gives them more influence over the outcome than it actually does, such as assuming a detailed model reduces uncertainty that is actually driven by factors outside the model.
Remedial action: Compare the stated forecast range to the analyst's own historical forecast errors. Widen the range to match the real track record, and use scenario or sensitivity analysis instead of a single confident point estimate.
Conservatism Bias
Conservatism bias causes analysts to underreact to new information. When fresh data contradicts an existing forecast, a conservative analyst adjusts the forecast only slightly, anchoring too heavily to the original estimate instead of updating it in proportion to the new evidence.
Remedial action: Build a systematic process for weighting new information. Assess the reliability, relevance, and persistence of new evidence, document the resulting update, and challenge any continued reliance on the prior forecast.
Confirmation Bias
Confirmation bias causes analysts to seek out and favor information that supports their existing view, while discounting or ignoring information that contradicts it. This is different from conservatism. Conservatism underweights new information generally. Confirmation bias actively filters which information gets attention based on whether it agrees with the current forecast.
Remedial action: Require the analyst to document the strongest evidence against their own forecast before finalizing it. Assign a colleague to argue the opposite case.
Representativeness Bias
Representativeness bias causes analysts to classify a new situation based on how closely it resembles a familiar pattern. An analyst might assume a smaller competitor will replicate a larger rival's growth path simply because the two companies look similar on the surface, ignoring differences in scale, management, or market position.
Remedial action: Check the forecast against base rates for the broader group of similar companies, not just the one comparable case the analyst has in mind. Test whether the "resemblance" actually predicts the outcome historically.
Bias-to-Remedy Quick Reference
Bias | Effect on Forecast | Remedial Action |
|---|---|---|
Overconfidence / illusion of control | Forecast range too narrow relative to real uncertainty | Compare to historical error range, use scenario analysis |
Conservatism | Forecast updates too slowly after new information | Set rules for how much new data should move the forecast |
Confirmation bias | Forecast reflects only supporting evidence | Document and weigh the strongest opposing evidence |
Representativeness | Forecast assumes company matches a familiar pattern | Check base rates across a wider comparable group |
Working Example: Reviewing an EPS Forecast
You are reviewing a sell-side analyst's EPS forecast for Company Z. The analyst forecasts next year's EPS at $2.50, with a stated range of $2.45 to $2.55, a $0.10 spread. You pull the analyst's forecast errors from the last four annual forecasts, measured as the difference between the point forecast and actual reported EPS:
Year 1: missed by $0.30
Year 2: missed by $0.25
Year 3: missed by $0.20
Year 4: missed by $0.35
Step 1: Calculate the average absolute forecast error.
Step 2: Compare this average error to the analyst's stated range.
The interval extends only $0.05 on either side of the $2.50 point estimate. The $0.275 average absolute error for the same forecast horizon is much larger. Four observations are a limited sample, so review broader comparable errors and interval coverage before calibrating a new range.
Step 3: Identify the bias.
The analyst is stating far more precision than the historical record supports. This is overconfidence.
The analyst's point estimate might still be reasonable, but the stated range understates real uncertainty. A reviewer should widen the range to reflect the historical error pattern or ask for a scenario analysis showing a low, base, and high case instead of a single narrow band.
Common Exam Traps
Confusing overconfidence with representativeness
Both can look like "the analyst is too sure of this number," but overconfidence is about an unrealistically narrow range or precision. Representativeness is about assuming the company matches a different, familiar case.
Mixing up conservatism and confirmation bias
Conservatism is a general failure to update enough after new information, regardless of whether that information supports or contradicts the existing view. Confirmation bias is selective, it specifically favors supporting evidence while ignoring contrary evidence.
Assuming "more information" fixes every bias
For overconfidence and illusion of control, adding more data often increases false confidence rather than correcting it. The fix is a structured check against historical accuracy, not simply more inputs.
Treating this LOS as a modeling-mechanics question
This page tests diagnosis of judgment error and the matching remedy. It does not ask you to build or revise a pro forma statement, that belongs to a different note.
Thinking the remedy for conservatism is to discard prior forecasts
The correct remedy is proportional updating based on a defined rule, not throwing out history entirely.
Practice Question
An equity analyst forecasts that a retailer's same-store sales will grow 6% next year. Two months into the fiscal year, the retailer reports same-store sales growth of only 2%. Reliable evidence of a lasting customer loss materially weakens the full-year outlook; seasonality and temporary timing do not explain the shortfall. The analyst revises the full-year forecast down to 5.5%, reasoning that two months is too little data to materially change a full-year estimate. Which bias does this revision most likely reflect?
Confirmation bias
Conservatism bias
Representativeness bias
Correct Answer: Option B
Conservatism bias is the tendency to underweight new information and adjust estimates too slowly. The analyst received clear evidence that growth was running well below the original forecast and adjusted the estimate by only half a percentage point, continuing to anchor heavily to the original 6% figure.
Option A: Confirmation bias is incorrect. Confirmation bias involves actively favoring information that supports an existing belief and dismissing contrary evidence. The analyst acknowledged the weak data rather than ignoring it, the issue is that he underweighted it, not that he filtered it out.
Option C: Representativeness bias is incorrect. Representativeness involves judging a situation based on resemblance to a separate, familiar pattern, such as assuming this retailer will perform like a past comparable company. There is no comparison to another case in this scenario.
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 Behavioral Biases in Analyst Forecasts
Is a behavioral bias the same thing as a forecasting model error?
No. A model error comes from wrong assumptions or incorrect data inputs. A behavioral bias comes from a predictable pattern in the analyst's own judgment, and it can distort a forecast even when the model and data are correct.
Do I need to memorize every bias from behavioral finance for this note?
No. This LOS focuses on the biases most relevant to analyst forecasting: overconfidence, illusion of control, conservatism, representativeness, and confirmation bias. Broader behavioral finance concepts belong to other topic areas.
Is there one best remedial action that works for every bias?
No. CFA Level I expects you to match the remedy to the specific bias. Checklists and scenario analysis help with overconfidence. Seeking out disconfirming evidence helps with confirmation bias. Using the wrong remedy for the bias in the question is a common wrong-answer trap.
Can more than one bias apply to the same forecast scenario?
In practice, yes. On the exam, scenarios are usually written to point clearly to one dominant bias. Focus on which bias the facts fit best rather than trying to apply several at once.