Using R in equity research

Thanks Bernhard,
Yes, I'm familiar with that kind of work. Thanks for
that link to research. I've mostly been looking at
papers at http://www.ssrn.com/ . There must be
hundreds of such analyses related to CAPM theory
alone. I personally have been using R to estimate
company costs of equity using the multifactor
Fama-French approach, and that's pretty simple stuff.
(The main question is how long to go back, whether to
use weekly or monthly returns, and whether to use lm,
wle, or other robust regression methods.)

I'm more thinking of single company or industry
research approaches that equity analysts perform
regularly. I'm looking for something that bridges the
gulf between the typical CFA equity practitioner and
the more statistically advanced, PhD-bearing
mathematical finance researcher. There are many of us,
for example, who read CFA magazine, or Journal of
Portfolio Management, and can somewhat follow along
with the discussion of mathematical procedures in the
article, but would have a difficult time performing
similar research ourselves. There's a big gap between
the research shown in the financial journals and the
research shown in typical sell-side research. I'm
trying to build up quantitative skills without going
back for a PhD.

It seems to me that there are numerous tools available
in R that can be put to good use in more mundane
research tasks, things like determining sensitivities
of earnings to economic growth and interest rates, or
the sensitivity of price to sales ratios to
profitability, leverage, and growth, or using ancova
and time series data to estimate quarterly margins for
a given company or industry. I've not yet seen a
"cookbook" for accomplishing this kind of research.

On the other hand, it's pretty easy to find
discussions relating to derivatives and volatility of
univariate time series. By the way, I have Carol
Alexander's "Market Models," and I found it both
difficult and not of much relevance to the company
fundamentals type work I do in finance. (I'm sure it
would be helpful to me if I worked in derivatives or
ran a quant hedge fund, were I able to really
understand it.)   

Regards,
Andrew

For what I know, it is used fairly extensive in

portfolio 

management (for
equity as well as other portfolio) but I don't

have a quick 

reference I
could point you to for more.  

Hello Andrew, Dirk,


a quick Wopec search on "multifactor portfolio*
revealed 11 records.
Although not being an equity analyst some entries
seem promising; but see
for yourself at:

http://netec.mcc.ac.uk/WoPEc/

Regards,
Bernhard

Suggestions, anyone?

Best regards, Dirk

Hi

It seems to me that there are numerous tools available
in R that can be put to good use in more mundane
research tasks, things like determining sensitivities
of earnings to economic growth and interest rates, or
the sensitivity of price to sales ratios to
profitability, leverage, and growth, or using ancova
and time series data to estimate quarterly margins for
a given company or industry. I've not yet seen a
"cookbook" for accomplishing this kind of research.

I think the ultimate goal of this kind of research is still to explain 
the cross section of returns, or to differentiate future winners from 
losers.  While you may be interested in earnings, price sales ratios, 
etc, these are really intermediate results.  Once you have these 
numbers, the question is: now that  have better earnings / sales 
estimates, how should I use them?  I would look at this as more of a 
two step problem: 1) trying to obtain better forecasts of earnings, 
etc; 2) using these forecasts to explain the cross section of returns 
ala fama french, or to form portfolios of winners/losers.

There is a lot of research going along these lines, and it is one of 
the reasons Accounting is such a hot field these days.  Much of the 
accounting research is concerned with using fine grained balance sheet 
items as building blocks to better estimates of earnings, residual 
income, return on assets, etc.

I don't have exact references handy (and a lot of what I've read are 
working papers), but some of the names to check out are

Charles Lee (Cornell) - 
http://www.johnson.cornell.edu/faculty/profiles/lee/

Stephen Penman (Columbia) - there are specific papers by Ou and Penman, 
and Nassem and Penman
	http://www0.gsb.columbia.edu/whoswho/full.cfm?id=55604

David Hirshleifer (Ohio State) - 
http://fisher.osu.edu/fin/faculty/hirshleifer/

I think the Accounting literature is where you will find the gap 
bridged between fundamental "CFA" style analysis and more rigorous 
statistics.

cheers,

Jim M