Skip to content

CFA Level II Exam · Market-Based Valuation: Price and Enterprise Value Multiples

Cross-Sectional Regression and Momentum Indicators Explained

Updated 7 October 2026 · Fact-checked

Cross-sectional regression explains a multiple, such as P/E, across many companies using fundamentals like growth and risk, then uses the fitted line to estimate a fair multiple. Momentum indicators compare results to expectations: unexpected earnings, standardized unexpected earnings (SUE) and relative strength. Apply the formula, then interpret the sign and size.

Understand Cross-Sectional Regression and Momentum Indicators

A cross-sectional regression takes one date and many companies. The dependent variable is a valuation multiple, for example P/E or P/B. The independent variables are fundamentals that should drive the multiple, such as expected earnings growth, payout ratio, beta or ROE. The fitted equation gives a predicted multiple for any company with those fundamentals.

To use it, put the target company's fundamentals into the equation. The result is its predicted (justified) multiple. If the actual multiple is below the predicted one, the stock looks undervalued relative to peers. If it is above, it looks overvalued. The signs of the coefficients should make sense: higher growth should raise P/E, higher risk should lower it.

The method has limits. The relationship can change over time, so a line fitted today may not hold next year. Independent variables are often correlated with each other (multicollinearity), which makes individual coefficients unreliable. Outliers can distort the line. The vignette may also hint at these problems through weak R-squared or odd coefficient signs.

Momentum indicators are a different family. They do not ask what a stock is worth from fundamentals. They ask whether news and price trends are pushing the stock up or down. Unexpected earnings is actual earnings minus expected earnings. Standardized unexpected earnings (SUE) divides that surprise by the standard deviation of the surprise, so surprises can be compared across companies. A high positive SUE has historically been linked to continued positive returns, but this is a tendency, not a guarantee.

Relative strength compares a stock's recent price performance with a benchmark or with other stocks. A rising ratio means the stock is outperforming. Momentum indicators are used to screen for candidates or to time entries, and they are often combined with valuation signals.

Key formulas to remember

Cross-sectional regression for a multiple
Predicted multiple = b0 + b1 × (Growth) + b2 × (Risk) + b3 × (Payout) + ...
Plug in the target company's values. Compare the predicted multiple with the actual multiple.
Valuation signal
Actual multiple < Predicted multiple → looks undervalued; Actual > Predicted → looks overvalued
Holds only if the model is reliable and the inputs are comparable.
Unexpected earnings (UE)
UE = Actual EPS − Expected EPS
Expected EPS may come from analyst consensus or a time-series model.
Standardized unexpected earnings (SUE)
SUE = (Actual EPS − Expected EPS) ÷ Standard deviation of unexpected earnings
The standard deviation is taken over a stated historical period of surprises. Scaled surprises allow comparison across firms.
Earnings surprise (percentage)
Surprise % = (Actual EPS − Consensus EPS) ÷ |Consensus EPS|
Use the absolute value of consensus when it is negative or small.
Relative strength
Relative strength = Stock price ÷ Benchmark (or index) value
A rising ratio over time means the stock is outperforming. Can also be a stock's return over a period minus the benchmark return.

How to solve Cross-Sectional Regression and Momentum Indicators questions

Decide first whether the question is about a regression-based multiple or a momentum indicator. Then read the exhibit carefully for the inputs and units.

  1. 1Identify the task: predicted multiple from a regression, SUE or surprise, or relative strength.
  2. 2For a regression, locate the intercept and each coefficient in the exhibit. Match each coefficient to its variable.
  3. 3Check units. Growth of 12% enters as 12 if the exhibit's regression used percent, or 0.12 if it used decimals. Use the form the exhibit states.
  4. 4Compute the predicted multiple, then compare it with the actual multiple to judge over- or undervaluation.
  5. 5For SUE, compute actual minus expected EPS, then divide by the stated standard deviation. Do not use the standard deviation of EPS itself unless told.
  6. 6For relative strength, compute the ratio or return difference and note whether it is rising or falling.
  7. 7State the conclusion in the direction the question asks and mention a limit if the item asks for one (instability, multicollinearity, outliers).

Quickest way: Plug in, compare, read the sign

When to use it: Use when the vignette gives a fitted equation or earnings data and asks for a number or a valuation view.

  1. Underline the coefficients, the target company's inputs and the actual multiple.
  2. Compute the predicted multiple in one line on your scratch paper.
  3. Subtract: actual minus predicted. Negative means cheap, positive means expensive.
  4. For SUE, mentally check that the sign of SUE matches the sign of actual minus expected EPS before you finish.
  5. Eliminate answer options that reverse the direction of the conclusion.

Common mistakes in Cross-Sectional Regression and Momentum Indicators

  • Dividing the earnings surprise by EPS or price instead of the standard deviation of surprises when calculating SUE.

    Percentage surprise and SUE look similar and both scale a surprise.

    Fix: For SUE, the denominator is the standard deviation of unexpected earnings over the stated period. Read the exhibit label.

  • Reading a lower actual multiple than predicted as overvalued.

    Candidates mix up the direction of comparison.

    Fix: Actual below predicted means the market pays less than fundamentals justify, so it looks undervalued.

  • Entering growth as 0.12 into an equation estimated with growth in percent.

    Candidates convert to decimals by habit.

    Fix: Use the same units as the regression. Check the exhibit notes.

  • Treating a regression's predicted multiple as a guaranteed fair value.

    The output looks precise.

    Fix: Remember the relationship can be unstable over time and coefficients can be unreliable when variables are correlated. It is a relative signal.

  • Confusing relative strength with valuation.

    Both are used to pick stocks.

    Fix: Relative strength measures price trend against a benchmark. It says nothing about whether the price is justified by fundamentals.

Worked examples

Example 1

An analyst regresses P/E on expected earnings growth (g, in percent) and beta across 40 companies. The fitted equation is P/E = 8.0 + 1.5 × g − 4.0 × beta. Company X has expected growth of 10%, a beta of 1.2 and an actual P/E of 17.0. (1) What is the predicted P/E? (2) Is X over- or undervalued on this basis? (3) What would a P/E of 17.0 imply for a company with the same beta but growth of 8%?

Show the solution
  1. (1) Predicted P/E = 8.0 + 1.5 × 10 − 4.0 × 1.2 = 8.0 + 15.0 − 4.8 = 18.2.
  2. (2) Actual 17.0 is below predicted 18.2, so X looks undervalued relative to the peer relationship.
  3. (3) With g = 8: predicted P/E = 8.0 + 12.0 − 4.8 = 15.2. An actual P/E of 17.0 is above 15.2, so that company would look overvalued.

Answer: (1) 18.2. (2) Undervalued, since 17.0 < 18.2. (3) Predicted 15.2, so a P/E of 17.0 would look overvalued.

Example 2

Company Y reported EPS of ₹14.50 against consensus expected EPS of ₹12.50. Over the past 16 quarters, the standard deviation of unexpected earnings has been ₹2.50. Company Z reported a surprise of ₹1.50 with a standard deviation of unexpected earnings of ₹0.50. (1) Compute SUE for Y. (2) Compute SUE for Z. (3) Which company had the stronger standardized surprise?

Show the solution
  1. (1) Unexpected earnings for Y = 14.50 − 12.50 = ₹2.00. SUE = 2.00 ÷ 2.50 = 0.80.
  2. (2) SUE for Z = 1.50 ÷ 0.50 = 3.00.
  3. (3) Compare 3.00 with 0.80. Z's surprise is larger relative to its normal variability, even though Y's rupee surprise is larger.

Answer: (1) SUE for Y = 0.80. (2) SUE for Z = 3.00. (3) Z has the stronger standardized surprise.

Exam tips

  • Read the SUE denominator label closely. The exam may offer a distractor that divides by EPS or price.
  • When a regression is given, expect a follow-up on limitations: instability over time, multicollinearity or outliers.
  • Check the units of each variable before you substitute. Percent versus decimal is a common trap.
  • Know the direction rules cold: actual below predicted multiple is undervalued, positive SUE is a favourable surprise.
  • Momentum indicators are signals, not proof. If an option says a signal guarantees future returns, reject it.

Cross-Sectional Regression and Momentum Indicators: frequently asked questions

What is the standardized unexpected earnings formula?

SUE = (Actual EPS − Expected EPS) ÷ Standard deviation of unexpected earnings. The standard deviation is measured over a stated historical period. Dividing by it lets you compare surprises across companies with different earnings volatility.

Why use a cross-sectional regression for valuation multiples?

It links a multiple such as P/E to fundamentals like growth and risk across many firms at one point in time. You can then estimate what a company's multiple should be given its own fundamentals. It gives a relative, data-based benchmark instead of a simple peer average.

What is the difference between earnings surprise and SUE?

Earnings surprise is actual minus expected earnings, often shown in percent. SUE scales that surprise by the standard deviation of past surprises. SUE is better for comparing companies because it adjusts for how noisy each company's earnings normally are.

What does relative strength tell you?

Relative strength shows how a stock's price has performed against a benchmark or other stocks. A rising ratio means it is outperforming. It is a trend indicator and does not tell you whether the stock is cheap or expensive.