The key point of the paper by John R. M. Hand and Jeremiah Green (2011)
is to demonstrate that portfolio performance can be improved considerably by
optimally incorporating accounting information. The stated goal of the authors is
to study the importance of accounting information from the perspective of a portfolio
manager. They are not particularly interested in demonstrating the importance
of accounting information for firm valuation (the fundamental analysis
viewpoint adopted by many authors like Penman [2009]), or the prediction of
future abnormal returns (as in Bernard &Thomas [1989, 1990], and many others).
Extending an optimal parametric portfolio policy (PPP) optimization technique
proposed by Brandt, Santa-Clara, and Valkanov (BSCV [2009]), Hand and
Green (2011) show that the incorporation of three accounting variables—namely,
accruals, change in earnings, and asset growth—produces significantly greater
returns than simply using the well-known Fama-French-Carhart (FFC) (Fama &
French [1993]; Carhart [1997]) factors used by BSCV (2009). They also show
that their approach outperforms simple buy-and-hold strategies proposed in the
prior accounting and finance literature. Impressively, they demonstrate that their
findings hold for out-of-sample portfolios in frictionless market settings. Unfortunately,
however, the performance of their model goes down substantially once
short sales constraints and transactions costs are introduced.
Notwithstanding the deteriorating performance of the Hand and Green model
when frictions and constraints are imposed, I view the paper as a contribution for
the academic and the practitioner. Apart from introducing an interesting technique
from the finance literature, the paper shows that accounting variables are
useful for stock picking, and points the way to profitably exploiting these variables
in a practical situation. In addition, the paper shows that even some simple
optimization strategies for picking stocks may dominate commonly used buyand-hold
strategies.