A stock’s beta is 1.2. What slope do you get if you regress the market on the stock?

Statistics
Easy
A regression puzzle about the slope of Y on X and the slope of X on Y.
Published

November 16, 2026

NoteQuestion

You regress a stock’s daily returns on the market’s daily returns by least squares and get a slope, the stock’s beta, of 1.2. The stock’s returns have a standard deviation of 2%, and the market’s have 1%. What slope do you get if you regress the market’s returns on the stock’s returns?

Try it first, then open the solution.

The answer is 0.3. Inverting the beta gives 1/1.2 ≈ 0.83, a common wrong answer.

Let \(X\) be the market’s daily return and \(Y\) the stock’s. The beta is the slope of the regression of \(Y\) on \(X\), and the question asks for the slope of the regression of \(X\) on \(Y\). The least-squares slope divides the covariance by the variance of the regressor, \(\sigma_X^2\) for the market and \(\sigma_Y^2\) for the stock:

\[\begin{aligned} b_{Y \mid X} &= \frac{\operatorname{Cov}(X, Y)}{\sigma_X^2}, \\ b_{X \mid Y} &= \frac{\operatorname{Cov}(X, Y)}{\sigma_Y^2}. \end{aligned} \tag{1}\]

Solving both lines of (1) for the covariance, their shared numerator, gives

\[b_{Y \mid X}\, \sigma_X^2 = \operatorname{Cov}(X, Y) = b_{X \mid Y}\, \sigma_Y^2. \tag{2}\]

The question gives three inputs:

Symbol Meaning Value
\(b_{Y \mid X}\) slope of the stock on the market, the beta 1.2
\(\sigma_Y\) standard deviation of the stock’s returns 2%
\(\sigma_X\) standard deviation of the market’s returns 1%

Dividing (2) by \(\sigma_Y^2\) and putting in the inputs gives

\[\begin{aligned} b_{X \mid Y} &= b_{Y \mid X}\, \frac{\sigma_X^2}{\sigma_Y^2} \\ &= 1.2 \times \frac{(1\%)^2}{(2\%)^2} \\ &= 1.2 \times 0.25 = 0.3. \end{aligned} \tag{3}\]

Three wrong answers are common:

Answer The slip
1/1.2 ≈ 0.83 inverts the beta, which treats the two regression lines as one
1.2 assumes the slope is the same in both directions
0.6 uses 1.2 × 1%/2% and leaves out the squares in (3)
import numpy as np

rng = np.random.default_rng(1)
market = rng.normal(0, 1, 100_000)
stock = 1.2 * market + rng.normal(0, 1.6, 100_000)

stock_on_market = np.polyfit(market, stock, 1)[0]
market_on_stock = np.polyfit(stock, market, 1)[0]
by_equation_3 = stock_on_market * market.var() / stock.var()

print(round(stock_on_market, 2), round(market_on_stock, 2))
print(round(by_equation_3, 2))
1.2 0.3
0.3