Skip to content
Get started

Regressing on a noisy signal

MediumStatistics4 min

A return obeys y=2X∗+εy=2X^*+\varepsilon, where the true signal X∗X^* has standard deviation 22 and ε\varepsilon is independent noise. You cannot see X∗X^*; instead you have two independent noisy readings X∗+u1X^*+u_1 and X∗+u2X^*+u_2, each with measurement noise of standard deviation 11 (independent of everything else). You average the two readings and regress yy on that average by OLS.

What slope does the regression estimate in a large sample? Give a decimal to 3 places.

Show hints (2)+
  1. The OLS slope tends to Cov⁡(y,X)/Var⁡(X)\operatorname{Cov}(y,X)/\operatorname{Var}(X).
  2. Averaging two independent readings halves the noise variance.

Answer

Reveal answer →

1.778 (± 0.005)

Create a free account to see the full worked solution and track your progress.

Interview format: Hedge Fund Quant, Systematic & Data-Driven Trading

Related questions