240 CHAPTER ◆ 2 6 Plan Performance and Risk Processes
etc. versus the benchmark, but to generate pure alphas without accepting unforeseen risk.
Using portfolio attribution we can understand the performance of a trading/investment
system relative to an index benchmark.
Performance attribution is a technique used to analyze the sources of excess returns of
a trading/investment system relative to a benchmark. Attribution will assess the ability of
a trading/investment system and identify where and how it is adding value. Performance
attribution is the process of decomposition of returns and risk into the position selection
algorithms to measure the value added by the active investment management and to com-
municate the risk components of the trading/investment strategy.
So, to make sure that a trading system consistently outperforms its benchmark we need
to perform attribution analysis on portfolio versus a benchmark. This is done to understand
the sources of active returns in an investment portfolio. Product teams mostly use the most
common method of attribution, the Brinson method, which uses three factors to explain
the active performance of a portfolio:
● Sector allocation
● Stock selection
● Interaction.
The next question is how to select an index benchmark, or construct a benchmark for
strategies that do not have a well-defined index.
26.2.1. Index Benchmark
In most cases, the product team chooses a market index as the benchmark for a trad-
ing/investment system. Generally, the instruments traded will be largely drawn or derived
from the constituents of the index. Because indexes track returns on a buy-and-hold basis
and make no attempt to determine which instruments are the most attractive, they track
passive investment and can provide a good standard against which to compare the per-
formance of a trading/investment system that is actively, though systematically, managed.
Using an index, it is possible to see how much value a trading/investment strategy adds
and from where or through what positions that value comes.
Selecting a specific benchmark is an individual decision, but there are some mini-
mum standards that any benchmark under consideration should meet. According to Carl
Bacon, in his book Practical Portfolio Performance Measurement and Attribution , to be
effective, a benchmark should meet most, if not all, of the following criteria:
● Unambiguous and transparent. The names and weights of securities comprising a
benchmark should be clearly defined.
● Investable. The benchmark should contain securities that an investor can purchase
in the market or easily replicate.
● Priced regularly. The benchmark ’ s return should be calculated daily, weekly, or
monthly.
● Availability of historical data. Past returns of the benchmark should be available
in order to gauge historical returns.
● Low turnover. Ideally, there should not be high turnover in the securities in the
index because it can be difficult to perform monthly portfolio allocation on an index