Formulating and Implementing Investment Strategies Using Financial Econometrics 319
■ (^) For backtesting proposed strategies, the sample used can be a set of
observations that have been preselected through time by some common
denominators.
■ (^) Although a sample of a sample should not pose a problem in back-
testing if the subset is selected randomly, this is not the case for most
samples that suffer from survivorship bias.
■ (^) A statistically significant pattern found for a strategy may merely reflect
the underlying common bond that was used to construct the testing
sample.
■ (^) The selection of a methodology for estimating a financial economet-
ric model should satisfy the same quality tests as developing economic
theories and selecting samples. In the absence of strong intuition, the
methodology that needs the least amount of human inputs should be
employed for estimating a model.
■ (^) For both testing and model estimation, the first task is to decide which
and how many explanatory variables should be included.
■ (^) Economic theories underlying a model typically involve abstract con-
cepts that need to be measured by alternative proxies.
■ (^) Selection of the appropriate proxies, while getting dangerously close to
data snooping, makes the determination of both the type and the num-
ber of explanatory variables an art rather than a science. One rule of
thumb is to be parsimonious.
■ (^) To safeguard against data snooping there should be a sequential testing
of the prediction model in the forecast period in order to affirm that
the condition that converts the model of actual returns to the model of
expected returns still produces an acceptable level of performance.
■ (^) Even if the expected return is modeled properly at the individual stock
level, the bottom line of implementable investment strategies is evalu-
ated by an acceptable level of risk-adjusted portfolio excess returns.
■ (^) Because most institutional portfolios are benchmarked, the objective is
to minimize tracking error given some level of portfolio excess return.
For this purpose, risk control becomes technically much more complex
than the conventional efficient portfolio concept.