This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ore, what are we going
to do about this! But keep it in mind in case the CAPM doesn’t
work as you expect it to work. • For a publicly traded company we observe past returns and we can compute covariances with the market return using recent history. Hopefully it will tell
us something about the future covariance. • If it is a new project or ﬁrm, try to ﬁnd comparable
publicly traded ﬁrms with returns you can observe. • Once you have the data (historical returns), how can
we compute beta? • Formally, run a linear regression like this:
rit = αi + β irmt + it e
• Where rit = rit − rF t and rmt = rmt − rF t • rit is the observed return on stock i and time t, and
rmt is the observed return on the market for the
same period. The "market" is usually approximated
with a broad stock index like the S&P 500. • Take T periods: t = 1, ..., T.
• In a linear regression:
(rit − ri )(rmt − rm)
¯e β i = t=1 T
(rmt − rm)2
¯e t=1 • This is equivalent as estimating the covariance and
then dividing that by the market variance. • Even easier: use the SLOPE function in Excel to
compute β i. If you want to compute αi use the
View Full Document
This document was uploaded on 02/27/2014 for the course ECON 1745 at Harvard.
- Fall '08